2 * This program is free software; you can redistribute it and/or
3 * modify it under the terms of the GNU General Public License
4 * as published by the Free Software Foundation; either version
5 * 2 of the License, or (at your option) any later version.
7 * Robert Olsson <robert.olsson@its.uu.se> Uppsala Universitet
8 * & Swedish University of Agricultural Sciences.
10 * Jens Laas <jens.laas@data.slu.se> Swedish University of
11 * Agricultural Sciences.
13 * Hans Liss <hans.liss@its.uu.se> Uppsala Universitet
15 * This work is based on the LPC-trie which is originally descibed in:
17 * An experimental study of compression methods for dynamic tries
18 * Stefan Nilsson and Matti Tikkanen. Algorithmica, 33(1):19-33, 2002.
19 * http://www.nada.kth.se/~snilsson/public/papers/dyntrie2/
22 * IP-address lookup using LC-tries. Stefan Nilsson and Gunnar Karlsson
23 * IEEE Journal on Selected Areas in Communications, 17(6):1083-1092, June 1999
25 * Version: $Id: fib_trie.c,v 1.3 2005/06/08 14:20:01 robert Exp $
28 * Code from fib_hash has been reused which includes the following header:
31 * INET An implementation of the TCP/IP protocol suite for the LINUX
32 * operating system. INET is implemented using the BSD Socket
33 * interface as the means of communication with the user level.
35 * IPv4 FIB: lookup engine and maintenance routines.
38 * Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
40 * This program is free software; you can redistribute it and/or
41 * modify it under the terms of the GNU General Public License
42 * as published by the Free Software Foundation; either version
43 * 2 of the License, or (at your option) any later version.
45 * Substantial contributions to this work comes from:
47 * David S. Miller, <davem@davemloft.net>
48 * Stephen Hemminger <shemminger@osdl.org>
49 * Paul E. McKenney <paulmck@us.ibm.com>
50 * Patrick McHardy <kaber@trash.net>
53 #define VERSION "0.408"
55 #include <asm/uaccess.h>
56 #include <asm/system.h>
57 #include <linux/bitops.h>
58 #include <linux/types.h>
59 #include <linux/kernel.h>
61 #include <linux/string.h>
62 #include <linux/socket.h>
63 #include <linux/sockios.h>
64 #include <linux/errno.h>
66 #include <linux/inet.h>
67 #include <linux/inetdevice.h>
68 #include <linux/netdevice.h>
69 #include <linux/if_arp.h>
70 #include <linux/proc_fs.h>
71 #include <linux/rcupdate.h>
72 #include <linux/skbuff.h>
73 #include <linux/netlink.h>
74 #include <linux/init.h>
75 #include <linux/list.h>
76 #include <net/net_namespace.h>
78 #include <net/protocol.h>
79 #include <net/route.h>
82 #include <net/ip_fib.h>
83 #include "fib_lookup.h"
85 #define MAX_STAT_DEPTH 32
87 #define KEYLENGTH (8*sizeof(t_key))
89 typedef unsigned int t_key
;
93 #define NODE_TYPE_MASK 0x1UL
94 #define NODE_TYPE(node) ((node)->parent & NODE_TYPE_MASK)
96 #define IS_TNODE(n) (!(n->parent & T_LEAF))
97 #define IS_LEAF(n) (n->parent & T_LEAF)
100 unsigned long parent
;
105 unsigned long parent
;
107 struct hlist_head list
;
112 struct hlist_node hlist
;
115 struct list_head falh
;
119 unsigned long parent
;
121 unsigned char pos
; /* 2log(KEYLENGTH) bits needed */
122 unsigned char bits
; /* 2log(KEYLENGTH) bits needed */
123 unsigned int full_children
; /* KEYLENGTH bits needed */
124 unsigned int empty_children
; /* KEYLENGTH bits needed */
127 struct work_struct work
;
129 struct node
*child
[0];
132 #ifdef CONFIG_IP_FIB_TRIE_STATS
133 struct trie_use_stats
{
135 unsigned int backtrack
;
136 unsigned int semantic_match_passed
;
137 unsigned int semantic_match_miss
;
138 unsigned int null_node_hit
;
139 unsigned int resize_node_skipped
;
144 unsigned int totdepth
;
145 unsigned int maxdepth
;
148 unsigned int nullpointers
;
149 unsigned int prefixes
;
150 unsigned int nodesizes
[MAX_STAT_DEPTH
];
155 #ifdef CONFIG_IP_FIB_TRIE_STATS
156 struct trie_use_stats stats
;
160 static void put_child(struct trie
*t
, struct tnode
*tn
, int i
, struct node
*n
);
161 static void tnode_put_child_reorg(struct tnode
*tn
, int i
, struct node
*n
,
163 static struct node
*resize(struct trie
*t
, struct tnode
*tn
);
164 static struct tnode
*inflate(struct trie
*t
, struct tnode
*tn
);
165 static struct tnode
*halve(struct trie
*t
, struct tnode
*tn
);
167 static struct kmem_cache
*fn_alias_kmem __read_mostly
;
168 static struct kmem_cache
*trie_leaf_kmem __read_mostly
;
170 static inline struct tnode
*node_parent(struct node
*node
)
172 return (struct tnode
*)(node
->parent
& ~NODE_TYPE_MASK
);
175 static inline struct tnode
*node_parent_rcu(struct node
*node
)
177 struct tnode
*ret
= node_parent(node
);
179 return rcu_dereference(ret
);
182 /* Same as rcu_assign_pointer
183 * but that macro() assumes that value is a pointer.
185 static inline void node_set_parent(struct node
*node
, struct tnode
*ptr
)
188 node
->parent
= (unsigned long)ptr
| NODE_TYPE(node
);
191 static inline struct node
*tnode_get_child(struct tnode
*tn
, unsigned int i
)
193 BUG_ON(i
>= 1U << tn
->bits
);
198 static inline struct node
*tnode_get_child_rcu(struct tnode
*tn
, unsigned int i
)
200 struct node
*ret
= tnode_get_child(tn
, i
);
202 return rcu_dereference(ret
);
205 static inline int tnode_child_length(const struct tnode
*tn
)
207 return 1 << tn
->bits
;
210 static inline t_key
mask_pfx(t_key k
, unsigned short l
)
212 return (l
== 0) ? 0 : k
>> (KEYLENGTH
-l
) << (KEYLENGTH
-l
);
215 static inline t_key
tkey_extract_bits(t_key a
, int offset
, int bits
)
217 if (offset
< KEYLENGTH
)
218 return ((t_key
)(a
<< offset
)) >> (KEYLENGTH
- bits
);
223 static inline int tkey_equals(t_key a
, t_key b
)
228 static inline int tkey_sub_equals(t_key a
, int offset
, int bits
, t_key b
)
230 if (bits
== 0 || offset
>= KEYLENGTH
)
232 bits
= bits
> KEYLENGTH
? KEYLENGTH
: bits
;
233 return ((a
^ b
) << offset
) >> (KEYLENGTH
- bits
) == 0;
236 static inline int tkey_mismatch(t_key a
, int offset
, t_key b
)
243 while ((diff
<< i
) >> (KEYLENGTH
-1) == 0)
249 To understand this stuff, an understanding of keys and all their bits is
250 necessary. Every node in the trie has a key associated with it, but not
251 all of the bits in that key are significant.
253 Consider a node 'n' and its parent 'tp'.
255 If n is a leaf, every bit in its key is significant. Its presence is
256 necessitated by path compression, since during a tree traversal (when
257 searching for a leaf - unless we are doing an insertion) we will completely
258 ignore all skipped bits we encounter. Thus we need to verify, at the end of
259 a potentially successful search, that we have indeed been walking the
262 Note that we can never "miss" the correct key in the tree if present by
263 following the wrong path. Path compression ensures that segments of the key
264 that are the same for all keys with a given prefix are skipped, but the
265 skipped part *is* identical for each node in the subtrie below the skipped
266 bit! trie_insert() in this implementation takes care of that - note the
267 call to tkey_sub_equals() in trie_insert().
269 if n is an internal node - a 'tnode' here, the various parts of its key
270 have many different meanings.
273 _________________________________________________________________
274 | i | i | i | i | i | i | i | N | N | N | S | S | S | S | S | C |
275 -----------------------------------------------------------------
276 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
278 _________________________________________________________________
279 | C | C | C | u | u | u | u | u | u | u | u | u | u | u | u | u |
280 -----------------------------------------------------------------
281 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
288 First, let's just ignore the bits that come before the parent tp, that is
289 the bits from 0 to (tp->pos-1). They are *known* but at this point we do
290 not use them for anything.
292 The bits from (tp->pos) to (tp->pos + tp->bits - 1) - "N", above - are the
293 index into the parent's child array. That is, they will be used to find
294 'n' among tp's children.
296 The bits from (tp->pos + tp->bits) to (n->pos - 1) - "S" - are skipped bits
299 All the bits we have seen so far are significant to the node n. The rest
300 of the bits are really not needed or indeed known in n->key.
302 The bits from (n->pos) to (n->pos + n->bits - 1) - "C" - are the index into
303 n's child array, and will of course be different for each child.
306 The rest of the bits, from (n->pos + n->bits) onward, are completely unknown
311 static inline void check_tnode(const struct tnode
*tn
)
313 WARN_ON(tn
&& tn
->pos
+tn
->bits
> 32);
316 static const int halve_threshold
= 25;
317 static const int inflate_threshold
= 50;
318 static const int halve_threshold_root
= 8;
319 static const int inflate_threshold_root
= 15;
322 static void __alias_free_mem(struct rcu_head
*head
)
324 struct fib_alias
*fa
= container_of(head
, struct fib_alias
, rcu
);
325 kmem_cache_free(fn_alias_kmem
, fa
);
328 static inline void alias_free_mem_rcu(struct fib_alias
*fa
)
330 call_rcu(&fa
->rcu
, __alias_free_mem
);
333 static void __leaf_free_rcu(struct rcu_head
*head
)
335 struct leaf
*l
= container_of(head
, struct leaf
, rcu
);
336 kmem_cache_free(trie_leaf_kmem
, l
);
339 static inline void free_leaf(struct leaf
*l
)
341 call_rcu_bh(&l
->rcu
, __leaf_free_rcu
);
344 static void __leaf_info_free_rcu(struct rcu_head
*head
)
346 kfree(container_of(head
, struct leaf_info
, rcu
));
349 static inline void free_leaf_info(struct leaf_info
*leaf
)
351 call_rcu(&leaf
->rcu
, __leaf_info_free_rcu
);
354 static struct tnode
*tnode_alloc(size_t size
)
356 if (size
<= PAGE_SIZE
)
357 return kzalloc(size
, GFP_KERNEL
);
359 return __vmalloc(size
, GFP_KERNEL
| __GFP_ZERO
, PAGE_KERNEL
);
362 static void __tnode_vfree(struct work_struct
*arg
)
364 struct tnode
*tn
= container_of(arg
, struct tnode
, work
);
368 static void __tnode_free_rcu(struct rcu_head
*head
)
370 struct tnode
*tn
= container_of(head
, struct tnode
, rcu
);
371 size_t size
= sizeof(struct tnode
) +
372 (sizeof(struct node
*) << tn
->bits
);
374 if (size
<= PAGE_SIZE
)
377 INIT_WORK(&tn
->work
, __tnode_vfree
);
378 schedule_work(&tn
->work
);
382 static inline void tnode_free(struct tnode
*tn
)
385 free_leaf((struct leaf
*) tn
);
387 call_rcu(&tn
->rcu
, __tnode_free_rcu
);
390 static struct leaf
*leaf_new(void)
392 struct leaf
*l
= kmem_cache_alloc(trie_leaf_kmem
, GFP_KERNEL
);
395 INIT_HLIST_HEAD(&l
->list
);
400 static struct leaf_info
*leaf_info_new(int plen
)
402 struct leaf_info
*li
= kmalloc(sizeof(struct leaf_info
), GFP_KERNEL
);
405 INIT_LIST_HEAD(&li
->falh
);
410 static struct tnode
*tnode_new(t_key key
, int pos
, int bits
)
412 size_t sz
= sizeof(struct tnode
) + (sizeof(struct node
*) << bits
);
413 struct tnode
*tn
= tnode_alloc(sz
);
416 tn
->parent
= T_TNODE
;
420 tn
->full_children
= 0;
421 tn
->empty_children
= 1<<bits
;
424 pr_debug("AT %p s=%u %lu\n", tn
, (unsigned int) sizeof(struct tnode
),
425 (unsigned long) (sizeof(struct node
) << bits
));
430 * Check whether a tnode 'n' is "full", i.e. it is an internal node
431 * and no bits are skipped. See discussion in dyntree paper p. 6
434 static inline int tnode_full(const struct tnode
*tn
, const struct node
*n
)
436 if (n
== NULL
|| IS_LEAF(n
))
439 return ((struct tnode
*) n
)->pos
== tn
->pos
+ tn
->bits
;
442 static inline void put_child(struct trie
*t
, struct tnode
*tn
, int i
,
445 tnode_put_child_reorg(tn
, i
, n
, -1);
449 * Add a child at position i overwriting the old value.
450 * Update the value of full_children and empty_children.
453 static void tnode_put_child_reorg(struct tnode
*tn
, int i
, struct node
*n
,
456 struct node
*chi
= tn
->child
[i
];
459 BUG_ON(i
>= 1<<tn
->bits
);
461 /* update emptyChildren */
462 if (n
== NULL
&& chi
!= NULL
)
463 tn
->empty_children
++;
464 else if (n
!= NULL
&& chi
== NULL
)
465 tn
->empty_children
--;
467 /* update fullChildren */
469 wasfull
= tnode_full(tn
, chi
);
471 isfull
= tnode_full(tn
, n
);
472 if (wasfull
&& !isfull
)
474 else if (!wasfull
&& isfull
)
478 node_set_parent(n
, tn
);
480 rcu_assign_pointer(tn
->child
[i
], n
);
483 static struct node
*resize(struct trie
*t
, struct tnode
*tn
)
487 struct tnode
*old_tn
;
488 int inflate_threshold_use
;
489 int halve_threshold_use
;
495 pr_debug("In tnode_resize %p inflate_threshold=%d threshold=%d\n",
496 tn
, inflate_threshold
, halve_threshold
);
499 if (tn
->empty_children
== tnode_child_length(tn
)) {
504 if (tn
->empty_children
== tnode_child_length(tn
) - 1)
505 for (i
= 0; i
< tnode_child_length(tn
); i
++) {
512 /* compress one level */
513 node_set_parent(n
, NULL
);
518 * Double as long as the resulting node has a number of
519 * nonempty nodes that are above the threshold.
523 * From "Implementing a dynamic compressed trie" by Stefan Nilsson of
524 * the Helsinki University of Technology and Matti Tikkanen of Nokia
525 * Telecommunications, page 6:
526 * "A node is doubled if the ratio of non-empty children to all
527 * children in the *doubled* node is at least 'high'."
529 * 'high' in this instance is the variable 'inflate_threshold'. It
530 * is expressed as a percentage, so we multiply it with
531 * tnode_child_length() and instead of multiplying by 2 (since the
532 * child array will be doubled by inflate()) and multiplying
533 * the left-hand side by 100 (to handle the percentage thing) we
534 * multiply the left-hand side by 50.
536 * The left-hand side may look a bit weird: tnode_child_length(tn)
537 * - tn->empty_children is of course the number of non-null children
538 * in the current node. tn->full_children is the number of "full"
539 * children, that is non-null tnodes with a skip value of 0.
540 * All of those will be doubled in the resulting inflated tnode, so
541 * we just count them one extra time here.
543 * A clearer way to write this would be:
545 * to_be_doubled = tn->full_children;
546 * not_to_be_doubled = tnode_child_length(tn) - tn->empty_children -
549 * new_child_length = tnode_child_length(tn) * 2;
551 * new_fill_factor = 100 * (not_to_be_doubled + 2*to_be_doubled) /
553 * if (new_fill_factor >= inflate_threshold)
555 * ...and so on, tho it would mess up the while () loop.
558 * 100 * (not_to_be_doubled + 2*to_be_doubled) / new_child_length >=
562 * 100 * (not_to_be_doubled + 2*to_be_doubled) >=
563 * inflate_threshold * new_child_length
565 * expand not_to_be_doubled and to_be_doubled, and shorten:
566 * 100 * (tnode_child_length(tn) - tn->empty_children +
567 * tn->full_children) >= inflate_threshold * new_child_length
569 * expand new_child_length:
570 * 100 * (tnode_child_length(tn) - tn->empty_children +
571 * tn->full_children) >=
572 * inflate_threshold * tnode_child_length(tn) * 2
575 * 50 * (tn->full_children + tnode_child_length(tn) -
576 * tn->empty_children) >= inflate_threshold *
577 * tnode_child_length(tn)
583 /* Keep root node larger */
586 inflate_threshold_use
= inflate_threshold_root
;
588 inflate_threshold_use
= inflate_threshold
;
592 while ((tn
->full_children
> 0 && max_resize
-- &&
593 50 * (tn
->full_children
+ tnode_child_length(tn
)
594 - tn
->empty_children
)
595 >= inflate_threshold_use
* tnode_child_length(tn
))) {
602 #ifdef CONFIG_IP_FIB_TRIE_STATS
603 t
->stats
.resize_node_skipped
++;
609 if (max_resize
< 0) {
611 pr_warning("Fix inflate_threshold_root."
612 " Now=%d size=%d bits\n",
613 inflate_threshold_root
, tn
->bits
);
615 pr_warning("Fix inflate_threshold."
616 " Now=%d size=%d bits\n",
617 inflate_threshold
, tn
->bits
);
623 * Halve as long as the number of empty children in this
624 * node is above threshold.
628 /* Keep root node larger */
631 halve_threshold_use
= halve_threshold_root
;
633 halve_threshold_use
= halve_threshold
;
637 while (tn
->bits
> 1 && max_resize
-- &&
638 100 * (tnode_child_length(tn
) - tn
->empty_children
) <
639 halve_threshold_use
* tnode_child_length(tn
)) {
645 #ifdef CONFIG_IP_FIB_TRIE_STATS
646 t
->stats
.resize_node_skipped
++;
652 if (max_resize
< 0) {
654 pr_warning("Fix halve_threshold_root."
655 " Now=%d size=%d bits\n",
656 halve_threshold_root
, tn
->bits
);
658 pr_warning("Fix halve_threshold."
659 " Now=%d size=%d bits\n",
660 halve_threshold
, tn
->bits
);
663 /* Only one child remains */
664 if (tn
->empty_children
== tnode_child_length(tn
) - 1)
665 for (i
= 0; i
< tnode_child_length(tn
); i
++) {
672 /* compress one level */
674 node_set_parent(n
, NULL
);
679 return (struct node
*) tn
;
682 static struct tnode
*inflate(struct trie
*t
, struct tnode
*tn
)
684 struct tnode
*oldtnode
= tn
;
685 int olen
= tnode_child_length(tn
);
688 pr_debug("In inflate\n");
690 tn
= tnode_new(oldtnode
->key
, oldtnode
->pos
, oldtnode
->bits
+ 1);
693 return ERR_PTR(-ENOMEM
);
696 * Preallocate and store tnodes before the actual work so we
697 * don't get into an inconsistent state if memory allocation
698 * fails. In case of failure we return the oldnode and inflate
699 * of tnode is ignored.
702 for (i
= 0; i
< olen
; i
++) {
705 inode
= (struct tnode
*) tnode_get_child(oldtnode
, i
);
708 inode
->pos
== oldtnode
->pos
+ oldtnode
->bits
&&
710 struct tnode
*left
, *right
;
711 t_key m
= ~0U << (KEYLENGTH
- 1) >> inode
->pos
;
713 left
= tnode_new(inode
->key
&(~m
), inode
->pos
+ 1,
718 right
= tnode_new(inode
->key
|m
, inode
->pos
+ 1,
726 put_child(t
, tn
, 2*i
, (struct node
*) left
);
727 put_child(t
, tn
, 2*i
+1, (struct node
*) right
);
731 for (i
= 0; i
< olen
; i
++) {
733 struct node
*node
= tnode_get_child(oldtnode
, i
);
734 struct tnode
*left
, *right
;
741 /* A leaf or an internal node with skipped bits */
743 if (IS_LEAF(node
) || ((struct tnode
*) node
)->pos
>
744 tn
->pos
+ tn
->bits
- 1) {
745 if (tkey_extract_bits(node
->key
,
746 oldtnode
->pos
+ oldtnode
->bits
,
748 put_child(t
, tn
, 2*i
, node
);
750 put_child(t
, tn
, 2*i
+1, node
);
754 /* An internal node with two children */
755 inode
= (struct tnode
*) node
;
757 if (inode
->bits
== 1) {
758 put_child(t
, tn
, 2*i
, inode
->child
[0]);
759 put_child(t
, tn
, 2*i
+1, inode
->child
[1]);
765 /* An internal node with more than two children */
767 /* We will replace this node 'inode' with two new
768 * ones, 'left' and 'right', each with half of the
769 * original children. The two new nodes will have
770 * a position one bit further down the key and this
771 * means that the "significant" part of their keys
772 * (see the discussion near the top of this file)
773 * will differ by one bit, which will be "0" in
774 * left's key and "1" in right's key. Since we are
775 * moving the key position by one step, the bit that
776 * we are moving away from - the bit at position
777 * (inode->pos) - is the one that will differ between
778 * left and right. So... we synthesize that bit in the
780 * The mask 'm' below will be a single "one" bit at
781 * the position (inode->pos)
784 /* Use the old key, but set the new significant
788 left
= (struct tnode
*) tnode_get_child(tn
, 2*i
);
789 put_child(t
, tn
, 2*i
, NULL
);
793 right
= (struct tnode
*) tnode_get_child(tn
, 2*i
+1);
794 put_child(t
, tn
, 2*i
+1, NULL
);
798 size
= tnode_child_length(left
);
799 for (j
= 0; j
< size
; j
++) {
800 put_child(t
, left
, j
, inode
->child
[j
]);
801 put_child(t
, right
, j
, inode
->child
[j
+ size
]);
803 put_child(t
, tn
, 2*i
, resize(t
, left
));
804 put_child(t
, tn
, 2*i
+1, resize(t
, right
));
808 tnode_free(oldtnode
);
812 int size
= tnode_child_length(tn
);
815 for (j
= 0; j
< size
; j
++)
817 tnode_free((struct tnode
*)tn
->child
[j
]);
821 return ERR_PTR(-ENOMEM
);
825 static struct tnode
*halve(struct trie
*t
, struct tnode
*tn
)
827 struct tnode
*oldtnode
= tn
;
828 struct node
*left
, *right
;
830 int olen
= tnode_child_length(tn
);
832 pr_debug("In halve\n");
834 tn
= tnode_new(oldtnode
->key
, oldtnode
->pos
, oldtnode
->bits
- 1);
837 return ERR_PTR(-ENOMEM
);
840 * Preallocate and store tnodes before the actual work so we
841 * don't get into an inconsistent state if memory allocation
842 * fails. In case of failure we return the oldnode and halve
843 * of tnode is ignored.
846 for (i
= 0; i
< olen
; i
+= 2) {
847 left
= tnode_get_child(oldtnode
, i
);
848 right
= tnode_get_child(oldtnode
, i
+1);
850 /* Two nonempty children */
854 newn
= tnode_new(left
->key
, tn
->pos
+ tn
->bits
, 1);
859 put_child(t
, tn
, i
/2, (struct node
*)newn
);
864 for (i
= 0; i
< olen
; i
+= 2) {
865 struct tnode
*newBinNode
;
867 left
= tnode_get_child(oldtnode
, i
);
868 right
= tnode_get_child(oldtnode
, i
+1);
870 /* At least one of the children is empty */
872 if (right
== NULL
) /* Both are empty */
874 put_child(t
, tn
, i
/2, right
);
879 put_child(t
, tn
, i
/2, left
);
883 /* Two nonempty children */
884 newBinNode
= (struct tnode
*) tnode_get_child(tn
, i
/2);
885 put_child(t
, tn
, i
/2, NULL
);
886 put_child(t
, newBinNode
, 0, left
);
887 put_child(t
, newBinNode
, 1, right
);
888 put_child(t
, tn
, i
/2, resize(t
, newBinNode
));
890 tnode_free(oldtnode
);
894 int size
= tnode_child_length(tn
);
897 for (j
= 0; j
< size
; j
++)
899 tnode_free((struct tnode
*)tn
->child
[j
]);
903 return ERR_PTR(-ENOMEM
);
907 /* readside must use rcu_read_lock currently dump routines
908 via get_fa_head and dump */
910 static struct leaf_info
*find_leaf_info(struct leaf
*l
, int plen
)
912 struct hlist_head
*head
= &l
->list
;
913 struct hlist_node
*node
;
914 struct leaf_info
*li
;
916 hlist_for_each_entry_rcu(li
, node
, head
, hlist
)
917 if (li
->plen
== plen
)
923 static inline struct list_head
*get_fa_head(struct leaf
*l
, int plen
)
925 struct leaf_info
*li
= find_leaf_info(l
, plen
);
933 static void insert_leaf_info(struct hlist_head
*head
, struct leaf_info
*new)
935 struct leaf_info
*li
= NULL
, *last
= NULL
;
936 struct hlist_node
*node
;
938 if (hlist_empty(head
)) {
939 hlist_add_head_rcu(&new->hlist
, head
);
941 hlist_for_each_entry(li
, node
, head
, hlist
) {
942 if (new->plen
> li
->plen
)
948 hlist_add_after_rcu(&last
->hlist
, &new->hlist
);
950 hlist_add_before_rcu(&new->hlist
, &li
->hlist
);
954 /* rcu_read_lock needs to be hold by caller from readside */
957 fib_find_node(struct trie
*t
, u32 key
)
964 n
= rcu_dereference(t
->trie
);
966 while (n
!= NULL
&& NODE_TYPE(n
) == T_TNODE
) {
967 tn
= (struct tnode
*) n
;
971 if (tkey_sub_equals(tn
->key
, pos
, tn
->pos
-pos
, key
)) {
972 pos
= tn
->pos
+ tn
->bits
;
973 n
= tnode_get_child_rcu(tn
,
974 tkey_extract_bits(key
,
980 /* Case we have found a leaf. Compare prefixes */
982 if (n
!= NULL
&& IS_LEAF(n
) && tkey_equals(key
, n
->key
))
983 return (struct leaf
*)n
;
988 static struct node
*trie_rebalance(struct trie
*t
, struct tnode
*tn
)
991 t_key cindex
, key
= tn
->key
;
994 while (tn
!= NULL
&& (tp
= node_parent((struct node
*)tn
)) != NULL
) {
995 cindex
= tkey_extract_bits(key
, tp
->pos
, tp
->bits
);
996 wasfull
= tnode_full(tp
, tnode_get_child(tp
, cindex
));
997 tn
= (struct tnode
*) resize(t
, (struct tnode
*)tn
);
999 tnode_put_child_reorg((struct tnode
*)tp
, cindex
,
1000 (struct node
*)tn
, wasfull
);
1002 tp
= node_parent((struct node
*) tn
);
1008 /* Handle last (top) tnode */
1010 tn
= (struct tnode
*)resize(t
, (struct tnode
*)tn
);
1012 return (struct node
*)tn
;
1015 /* only used from updater-side */
1017 static struct list_head
*fib_insert_node(struct trie
*t
, u32 key
, int plen
)
1020 struct tnode
*tp
= NULL
, *tn
= NULL
;
1024 struct list_head
*fa_head
= NULL
;
1025 struct leaf_info
*li
;
1031 /* If we point to NULL, stop. Either the tree is empty and we should
1032 * just put a new leaf in if, or we have reached an empty child slot,
1033 * and we should just put our new leaf in that.
1034 * If we point to a T_TNODE, check if it matches our key. Note that
1035 * a T_TNODE might be skipping any number of bits - its 'pos' need
1036 * not be the parent's 'pos'+'bits'!
1038 * If it does match the current key, get pos/bits from it, extract
1039 * the index from our key, push the T_TNODE and walk the tree.
1041 * If it doesn't, we have to replace it with a new T_TNODE.
1043 * If we point to a T_LEAF, it might or might not have the same key
1044 * as we do. If it does, just change the value, update the T_LEAF's
1045 * value, and return it.
1046 * If it doesn't, we need to replace it with a T_TNODE.
1049 while (n
!= NULL
&& NODE_TYPE(n
) == T_TNODE
) {
1050 tn
= (struct tnode
*) n
;
1054 if (tkey_sub_equals(tn
->key
, pos
, tn
->pos
-pos
, key
)) {
1056 pos
= tn
->pos
+ tn
->bits
;
1057 n
= tnode_get_child(tn
,
1058 tkey_extract_bits(key
,
1062 BUG_ON(n
&& node_parent(n
) != tn
);
1068 * n ----> NULL, LEAF or TNODE
1070 * tp is n's (parent) ----> NULL or TNODE
1073 BUG_ON(tp
&& IS_LEAF(tp
));
1075 /* Case 1: n is a leaf. Compare prefixes */
1077 if (n
!= NULL
&& IS_LEAF(n
) && tkey_equals(key
, n
->key
)) {
1078 l
= (struct leaf
*) n
;
1079 li
= leaf_info_new(plen
);
1084 fa_head
= &li
->falh
;
1085 insert_leaf_info(&l
->list
, li
);
1094 li
= leaf_info_new(plen
);
1101 fa_head
= &li
->falh
;
1102 insert_leaf_info(&l
->list
, li
);
1104 if (t
->trie
&& n
== NULL
) {
1105 /* Case 2: n is NULL, and will just insert a new leaf */
1107 node_set_parent((struct node
*)l
, tp
);
1109 cindex
= tkey_extract_bits(key
, tp
->pos
, tp
->bits
);
1110 put_child(t
, (struct tnode
*)tp
, cindex
, (struct node
*)l
);
1112 /* Case 3: n is a LEAF or a TNODE and the key doesn't match. */
1114 * Add a new tnode here
1115 * first tnode need some special handling
1119 pos
= tp
->pos
+tp
->bits
;
1124 newpos
= tkey_mismatch(key
, pos
, n
->key
);
1125 tn
= tnode_new(n
->key
, newpos
, 1);
1128 tn
= tnode_new(key
, newpos
, 1); /* First tnode */
1137 node_set_parent((struct node
*)tn
, tp
);
1139 missbit
= tkey_extract_bits(key
, newpos
, 1);
1140 put_child(t
, tn
, missbit
, (struct node
*)l
);
1141 put_child(t
, tn
, 1-missbit
, n
);
1144 cindex
= tkey_extract_bits(key
, tp
->pos
, tp
->bits
);
1145 put_child(t
, (struct tnode
*)tp
, cindex
,
1148 rcu_assign_pointer(t
->trie
, (struct node
*)tn
);
1153 if (tp
&& tp
->pos
+ tp
->bits
> 32)
1154 pr_warning("fib_trie"
1155 " tp=%p pos=%d, bits=%d, key=%0x plen=%d\n",
1156 tp
, tp
->pos
, tp
->bits
, key
, plen
);
1158 /* Rebalance the trie */
1160 rcu_assign_pointer(t
->trie
, trie_rebalance(t
, tp
));
1166 * Caller must hold RTNL.
1168 static int fn_trie_insert(struct fib_table
*tb
, struct fib_config
*cfg
)
1170 struct trie
*t
= (struct trie
*) tb
->tb_data
;
1171 struct fib_alias
*fa
, *new_fa
;
1172 struct list_head
*fa_head
= NULL
;
1173 struct fib_info
*fi
;
1174 int plen
= cfg
->fc_dst_len
;
1175 u8 tos
= cfg
->fc_tos
;
1183 key
= ntohl(cfg
->fc_dst
);
1185 pr_debug("Insert table=%u %08x/%d\n", tb
->tb_id
, key
, plen
);
1187 mask
= ntohl(inet_make_mask(plen
));
1194 fi
= fib_create_info(cfg
);
1200 l
= fib_find_node(t
, key
);
1204 fa_head
= get_fa_head(l
, plen
);
1205 fa
= fib_find_alias(fa_head
, tos
, fi
->fib_priority
);
1208 /* Now fa, if non-NULL, points to the first fib alias
1209 * with the same keys [prefix,tos,priority], if such key already
1210 * exists or to the node before which we will insert new one.
1212 * If fa is NULL, we will need to allocate a new one and
1213 * insert to the head of f.
1215 * If f is NULL, no fib node matched the destination key
1216 * and we need to allocate a new one of those as well.
1219 if (fa
&& fa
->fa_tos
== tos
&&
1220 fa
->fa_info
->fib_priority
== fi
->fib_priority
) {
1221 struct fib_alias
*fa_first
, *fa_match
;
1224 if (cfg
->fc_nlflags
& NLM_F_EXCL
)
1228 * 1. Find exact match for type, scope, fib_info to avoid
1230 * 2. Find next 'fa' (or head), NLM_F_APPEND inserts before it
1234 fa
= list_entry(fa
->fa_list
.prev
, struct fib_alias
, fa_list
);
1235 list_for_each_entry_continue(fa
, fa_head
, fa_list
) {
1236 if (fa
->fa_tos
!= tos
)
1238 if (fa
->fa_info
->fib_priority
!= fi
->fib_priority
)
1240 if (fa
->fa_type
== cfg
->fc_type
&&
1241 fa
->fa_scope
== cfg
->fc_scope
&&
1242 fa
->fa_info
== fi
) {
1248 if (cfg
->fc_nlflags
& NLM_F_REPLACE
) {
1249 struct fib_info
*fi_drop
;
1259 new_fa
= kmem_cache_alloc(fn_alias_kmem
, GFP_KERNEL
);
1263 fi_drop
= fa
->fa_info
;
1264 new_fa
->fa_tos
= fa
->fa_tos
;
1265 new_fa
->fa_info
= fi
;
1266 new_fa
->fa_type
= cfg
->fc_type
;
1267 new_fa
->fa_scope
= cfg
->fc_scope
;
1268 state
= fa
->fa_state
;
1269 new_fa
->fa_state
= state
& ~FA_S_ACCESSED
;
1271 list_replace_rcu(&fa
->fa_list
, &new_fa
->fa_list
);
1272 alias_free_mem_rcu(fa
);
1274 fib_release_info(fi_drop
);
1275 if (state
& FA_S_ACCESSED
)
1277 rtmsg_fib(RTM_NEWROUTE
, htonl(key
), new_fa
, plen
,
1278 tb
->tb_id
, &cfg
->fc_nlinfo
, NLM_F_REPLACE
);
1282 /* Error if we find a perfect match which
1283 * uses the same scope, type, and nexthop
1289 if (!(cfg
->fc_nlflags
& NLM_F_APPEND
))
1293 if (!(cfg
->fc_nlflags
& NLM_F_CREATE
))
1297 new_fa
= kmem_cache_alloc(fn_alias_kmem
, GFP_KERNEL
);
1301 new_fa
->fa_info
= fi
;
1302 new_fa
->fa_tos
= tos
;
1303 new_fa
->fa_type
= cfg
->fc_type
;
1304 new_fa
->fa_scope
= cfg
->fc_scope
;
1305 new_fa
->fa_state
= 0;
1307 * Insert new entry to the list.
1311 fa_head
= fib_insert_node(t
, key
, plen
);
1312 if (unlikely(!fa_head
)) {
1314 goto out_free_new_fa
;
1318 list_add_tail_rcu(&new_fa
->fa_list
,
1319 (fa
? &fa
->fa_list
: fa_head
));
1322 rtmsg_fib(RTM_NEWROUTE
, htonl(key
), new_fa
, plen
, tb
->tb_id
,
1323 &cfg
->fc_nlinfo
, 0);
1328 kmem_cache_free(fn_alias_kmem
, new_fa
);
1330 fib_release_info(fi
);
1335 /* should be called with rcu_read_lock */
1336 static int check_leaf(struct trie
*t
, struct leaf
*l
,
1337 t_key key
, const struct flowi
*flp
,
1338 struct fib_result
*res
)
1340 struct leaf_info
*li
;
1341 struct hlist_head
*hhead
= &l
->list
;
1342 struct hlist_node
*node
;
1344 hlist_for_each_entry_rcu(li
, node
, hhead
, hlist
) {
1346 int plen
= li
->plen
;
1347 __be32 mask
= inet_make_mask(plen
);
1349 if (l
->key
!= (key
& ntohl(mask
)))
1352 err
= fib_semantic_match(&li
->falh
, flp
, res
,
1353 htonl(l
->key
), mask
, plen
);
1355 #ifdef CONFIG_IP_FIB_TRIE_STATS
1357 t
->stats
.semantic_match_passed
++;
1359 t
->stats
.semantic_match_miss
++;
1368 static int fn_trie_lookup(struct fib_table
*tb
, const struct flowi
*flp
,
1369 struct fib_result
*res
)
1371 struct trie
*t
= (struct trie
*) tb
->tb_data
;
1376 t_key key
= ntohl(flp
->fl4_dst
);
1379 int current_prefix_length
= KEYLENGTH
;
1381 t_key node_prefix
, key_prefix
, pref_mismatch
;
1386 n
= rcu_dereference(t
->trie
);
1390 #ifdef CONFIG_IP_FIB_TRIE_STATS
1396 plen
= check_leaf(t
, (struct leaf
*)n
, key
, flp
, res
);
1403 pn
= (struct tnode
*) n
;
1411 cindex
= tkey_extract_bits(mask_pfx(key
, current_prefix_length
),
1414 n
= tnode_get_child(pn
, cindex
);
1417 #ifdef CONFIG_IP_FIB_TRIE_STATS
1418 t
->stats
.null_node_hit
++;
1424 plen
= check_leaf(t
, (struct leaf
*)n
, key
, flp
, res
);
1432 cn
= (struct tnode
*)n
;
1435 * It's a tnode, and we can do some extra checks here if we
1436 * like, to avoid descending into a dead-end branch.
1437 * This tnode is in the parent's child array at index
1438 * key[p_pos..p_pos+p_bits] but potentially with some bits
1439 * chopped off, so in reality the index may be just a
1440 * subprefix, padded with zero at the end.
1441 * We can also take a look at any skipped bits in this
1442 * tnode - everything up to p_pos is supposed to be ok,
1443 * and the non-chopped bits of the index (se previous
1444 * paragraph) are also guaranteed ok, but the rest is
1445 * considered unknown.
1447 * The skipped bits are key[pos+bits..cn->pos].
1450 /* If current_prefix_length < pos+bits, we are already doing
1451 * actual prefix matching, which means everything from
1452 * pos+(bits-chopped_off) onward must be zero along some
1453 * branch of this subtree - otherwise there is *no* valid
1454 * prefix present. Here we can only check the skipped
1455 * bits. Remember, since we have already indexed into the
1456 * parent's child array, we know that the bits we chopped of
1460 /* NOTA BENE: Checking only skipped bits
1461 for the new node here */
1463 if (current_prefix_length
< pos
+bits
) {
1464 if (tkey_extract_bits(cn
->key
, current_prefix_length
,
1465 cn
->pos
- current_prefix_length
)
1471 * If chopped_off=0, the index is fully validated and we
1472 * only need to look at the skipped bits for this, the new,
1473 * tnode. What we actually want to do is to find out if
1474 * these skipped bits match our key perfectly, or if we will
1475 * have to count on finding a matching prefix further down,
1476 * because if we do, we would like to have some way of
1477 * verifying the existence of such a prefix at this point.
1480 /* The only thing we can do at this point is to verify that
1481 * any such matching prefix can indeed be a prefix to our
1482 * key, and if the bits in the node we are inspecting that
1483 * do not match our key are not ZERO, this cannot be true.
1484 * Thus, find out where there is a mismatch (before cn->pos)
1485 * and verify that all the mismatching bits are zero in the
1490 * Note: We aren't very concerned about the piece of
1491 * the key that precede pn->pos+pn->bits, since these
1492 * have already been checked. The bits after cn->pos
1493 * aren't checked since these are by definition
1494 * "unknown" at this point. Thus, what we want to see
1495 * is if we are about to enter the "prefix matching"
1496 * state, and in that case verify that the skipped
1497 * bits that will prevail throughout this subtree are
1498 * zero, as they have to be if we are to find a
1502 node_prefix
= mask_pfx(cn
->key
, cn
->pos
);
1503 key_prefix
= mask_pfx(key
, cn
->pos
);
1504 pref_mismatch
= key_prefix
^node_prefix
;
1508 * In short: If skipped bits in this node do not match
1509 * the search key, enter the "prefix matching"
1512 if (pref_mismatch
) {
1513 while (!(pref_mismatch
& (1<<(KEYLENGTH
-1)))) {
1515 pref_mismatch
= pref_mismatch
<< 1;
1517 key_prefix
= tkey_extract_bits(cn
->key
, mp
, cn
->pos
-mp
);
1519 if (key_prefix
!= 0)
1522 if (current_prefix_length
>= cn
->pos
)
1523 current_prefix_length
= mp
;
1526 pn
= (struct tnode
*)n
; /* Descend */
1533 /* As zero don't change the child key (cindex) */
1534 while ((chopped_off
<= pn
->bits
)
1535 && !(cindex
& (1<<(chopped_off
-1))))
1538 /* Decrease current_... with bits chopped off */
1539 if (current_prefix_length
> pn
->pos
+ pn
->bits
- chopped_off
)
1540 current_prefix_length
= pn
->pos
+ pn
->bits
1544 * Either we do the actual chop off according or if we have
1545 * chopped off all bits in this tnode walk up to our parent.
1548 if (chopped_off
<= pn
->bits
) {
1549 cindex
&= ~(1 << (chopped_off
-1));
1551 struct tnode
*parent
= node_parent((struct node
*) pn
);
1555 /* Get Child's index */
1556 cindex
= tkey_extract_bits(pn
->key
, parent
->pos
, parent
->bits
);
1560 #ifdef CONFIG_IP_FIB_TRIE_STATS
1561 t
->stats
.backtrack
++;
1574 * Remove the leaf and return parent.
1576 static void trie_leaf_remove(struct trie
*t
, struct leaf
*l
)
1578 struct tnode
*tp
= node_parent((struct node
*) l
);
1580 pr_debug("entering trie_leaf_remove(%p)\n", l
);
1583 t_key cindex
= tkey_extract_bits(l
->key
, tp
->pos
, tp
->bits
);
1584 put_child(t
, (struct tnode
*)tp
, cindex
, NULL
);
1585 rcu_assign_pointer(t
->trie
, trie_rebalance(t
, tp
));
1587 rcu_assign_pointer(t
->trie
, NULL
);
1593 * Caller must hold RTNL.
1595 static int fn_trie_delete(struct fib_table
*tb
, struct fib_config
*cfg
)
1597 struct trie
*t
= (struct trie
*) tb
->tb_data
;
1599 int plen
= cfg
->fc_dst_len
;
1600 u8 tos
= cfg
->fc_tos
;
1601 struct fib_alias
*fa
, *fa_to_delete
;
1602 struct list_head
*fa_head
;
1604 struct leaf_info
*li
;
1609 key
= ntohl(cfg
->fc_dst
);
1610 mask
= ntohl(inet_make_mask(plen
));
1616 l
= fib_find_node(t
, key
);
1621 fa_head
= get_fa_head(l
, plen
);
1622 fa
= fib_find_alias(fa_head
, tos
, 0);
1627 pr_debug("Deleting %08x/%d tos=%d t=%p\n", key
, plen
, tos
, t
);
1629 fa_to_delete
= NULL
;
1630 fa
= list_entry(fa
->fa_list
.prev
, struct fib_alias
, fa_list
);
1631 list_for_each_entry_continue(fa
, fa_head
, fa_list
) {
1632 struct fib_info
*fi
= fa
->fa_info
;
1634 if (fa
->fa_tos
!= tos
)
1637 if ((!cfg
->fc_type
|| fa
->fa_type
== cfg
->fc_type
) &&
1638 (cfg
->fc_scope
== RT_SCOPE_NOWHERE
||
1639 fa
->fa_scope
== cfg
->fc_scope
) &&
1640 (!cfg
->fc_protocol
||
1641 fi
->fib_protocol
== cfg
->fc_protocol
) &&
1642 fib_nh_match(cfg
, fi
) == 0) {
1652 rtmsg_fib(RTM_DELROUTE
, htonl(key
), fa
, plen
, tb
->tb_id
,
1653 &cfg
->fc_nlinfo
, 0);
1655 l
= fib_find_node(t
, key
);
1656 li
= find_leaf_info(l
, plen
);
1658 list_del_rcu(&fa
->fa_list
);
1660 if (list_empty(fa_head
)) {
1661 hlist_del_rcu(&li
->hlist
);
1665 if (hlist_empty(&l
->list
))
1666 trie_leaf_remove(t
, l
);
1668 if (fa
->fa_state
& FA_S_ACCESSED
)
1671 fib_release_info(fa
->fa_info
);
1672 alias_free_mem_rcu(fa
);
1676 static int trie_flush_list(struct list_head
*head
)
1678 struct fib_alias
*fa
, *fa_node
;
1681 list_for_each_entry_safe(fa
, fa_node
, head
, fa_list
) {
1682 struct fib_info
*fi
= fa
->fa_info
;
1684 if (fi
&& (fi
->fib_flags
& RTNH_F_DEAD
)) {
1685 list_del_rcu(&fa
->fa_list
);
1686 fib_release_info(fa
->fa_info
);
1687 alias_free_mem_rcu(fa
);
1694 static int trie_flush_leaf(struct leaf
*l
)
1697 struct hlist_head
*lih
= &l
->list
;
1698 struct hlist_node
*node
, *tmp
;
1699 struct leaf_info
*li
= NULL
;
1701 hlist_for_each_entry_safe(li
, node
, tmp
, lih
, hlist
) {
1702 found
+= trie_flush_list(&li
->falh
);
1704 if (list_empty(&li
->falh
)) {
1705 hlist_del_rcu(&li
->hlist
);
1713 * Scan for the next right leaf starting at node p->child[idx]
1714 * Since we have back pointer, no recursion necessary.
1716 static struct leaf
*leaf_walk_rcu(struct tnode
*p
, struct node
*c
)
1722 idx
= tkey_extract_bits(c
->key
, p
->pos
, p
->bits
) + 1;
1726 while (idx
< 1u << p
->bits
) {
1727 c
= tnode_get_child_rcu(p
, idx
++);
1732 prefetch(p
->child
[idx
]);
1733 return (struct leaf
*) c
;
1736 /* Rescan start scanning in new node */
1737 p
= (struct tnode
*) c
;
1741 /* Node empty, walk back up to parent */
1742 c
= (struct node
*) p
;
1743 } while ( (p
= node_parent_rcu(c
)) != NULL
);
1745 return NULL
; /* Root of trie */
1748 static struct leaf
*trie_firstleaf(struct trie
*t
)
1750 struct tnode
*n
= (struct tnode
*) rcu_dereference(t
->trie
);
1755 if (IS_LEAF(n
)) /* trie is just a leaf */
1756 return (struct leaf
*) n
;
1758 return leaf_walk_rcu(n
, NULL
);
1761 static struct leaf
*trie_nextleaf(struct leaf
*l
)
1763 struct node
*c
= (struct node
*) l
;
1764 struct tnode
*p
= node_parent(c
);
1767 return NULL
; /* trie with just one leaf */
1769 return leaf_walk_rcu(p
, c
);
1772 static struct leaf
*trie_leafindex(struct trie
*t
, int index
)
1774 struct leaf
*l
= trie_firstleaf(t
);
1776 while (l
&& index
-- > 0)
1777 l
= trie_nextleaf(l
);
1784 * Caller must hold RTNL.
1786 static int fn_trie_flush(struct fib_table
*tb
)
1788 struct trie
*t
= (struct trie
*) tb
->tb_data
;
1789 struct leaf
*l
, *ll
= NULL
;
1792 for (l
= trie_firstleaf(t
); l
; l
= trie_nextleaf(l
)) {
1793 found
+= trie_flush_leaf(l
);
1795 if (ll
&& hlist_empty(&ll
->list
))
1796 trie_leaf_remove(t
, ll
);
1800 if (ll
&& hlist_empty(&ll
->list
))
1801 trie_leaf_remove(t
, ll
);
1803 pr_debug("trie_flush found=%d\n", found
);
1807 static void fn_trie_select_default(struct fib_table
*tb
,
1808 const struct flowi
*flp
,
1809 struct fib_result
*res
)
1811 struct trie
*t
= (struct trie
*) tb
->tb_data
;
1812 int order
, last_idx
;
1813 struct fib_info
*fi
= NULL
;
1814 struct fib_info
*last_resort
;
1815 struct fib_alias
*fa
= NULL
;
1816 struct list_head
*fa_head
;
1825 l
= fib_find_node(t
, 0);
1829 fa_head
= get_fa_head(l
, 0);
1833 if (list_empty(fa_head
))
1836 list_for_each_entry_rcu(fa
, fa_head
, fa_list
) {
1837 struct fib_info
*next_fi
= fa
->fa_info
;
1839 if (fa
->fa_scope
!= res
->scope
||
1840 fa
->fa_type
!= RTN_UNICAST
)
1843 if (next_fi
->fib_priority
> res
->fi
->fib_priority
)
1845 if (!next_fi
->fib_nh
[0].nh_gw
||
1846 next_fi
->fib_nh
[0].nh_scope
!= RT_SCOPE_LINK
)
1848 fa
->fa_state
|= FA_S_ACCESSED
;
1851 if (next_fi
!= res
->fi
)
1853 } else if (!fib_detect_death(fi
, order
, &last_resort
,
1854 &last_idx
, tb
->tb_default
)) {
1855 fib_result_assign(res
, fi
);
1856 tb
->tb_default
= order
;
1862 if (order
<= 0 || fi
== NULL
) {
1863 tb
->tb_default
= -1;
1867 if (!fib_detect_death(fi
, order
, &last_resort
, &last_idx
,
1869 fib_result_assign(res
, fi
);
1870 tb
->tb_default
= order
;
1874 fib_result_assign(res
, last_resort
);
1875 tb
->tb_default
= last_idx
;
1880 static int fn_trie_dump_fa(t_key key
, int plen
, struct list_head
*fah
,
1881 struct fib_table
*tb
,
1882 struct sk_buff
*skb
, struct netlink_callback
*cb
)
1885 struct fib_alias
*fa
;
1886 __be32 xkey
= htonl(key
);
1891 /* rcu_read_lock is hold by caller */
1893 list_for_each_entry_rcu(fa
, fah
, fa_list
) {
1899 if (fib_dump_info(skb
, NETLINK_CB(cb
->skb
).pid
,
1908 fa
->fa_info
, NLM_F_MULTI
) < 0) {
1918 static int fn_trie_dump_leaf(struct leaf
*l
, struct fib_table
*tb
,
1919 struct sk_buff
*skb
, struct netlink_callback
*cb
)
1921 struct leaf_info
*li
;
1922 struct hlist_node
*node
;
1928 /* rcu_read_lock is hold by caller */
1929 hlist_for_each_entry_rcu(li
, node
, &l
->list
, hlist
) {
1938 if (list_empty(&li
->falh
))
1941 if (fn_trie_dump_fa(l
->key
, li
->plen
, &li
->falh
, tb
, skb
, cb
) < 0) {
1952 static int fn_trie_dump(struct fib_table
*tb
, struct sk_buff
*skb
,
1953 struct netlink_callback
*cb
)
1956 struct trie
*t
= (struct trie
*) tb
->tb_data
;
1957 t_key key
= cb
->args
[2];
1958 int count
= cb
->args
[3];
1961 /* Dump starting at last key.
1962 * Note: 0.0.0.0/0 (ie default) is first key.
1965 l
= trie_firstleaf(t
);
1967 /* Normally, continue from last key, but if that is missing
1968 * fallback to using slow rescan
1970 l
= fib_find_node(t
, key
);
1972 l
= trie_leafindex(t
, count
);
1976 cb
->args
[2] = l
->key
;
1977 if (fn_trie_dump_leaf(l
, tb
, skb
, cb
) < 0) {
1978 cb
->args
[3] = count
;
1984 l
= trie_nextleaf(l
);
1985 memset(&cb
->args
[4], 0,
1986 sizeof(cb
->args
) - 4*sizeof(cb
->args
[0]));
1988 cb
->args
[3] = count
;
1994 void __init
fib_hash_init(void)
1996 fn_alias_kmem
= kmem_cache_create("ip_fib_alias",
1997 sizeof(struct fib_alias
),
1998 0, SLAB_PANIC
, NULL
);
2000 trie_leaf_kmem
= kmem_cache_create("ip_fib_trie",
2001 max(sizeof(struct leaf
),
2002 sizeof(struct leaf_info
)),
2003 0, SLAB_PANIC
, NULL
);
2007 /* Fix more generic FIB names for init later */
2008 struct fib_table
*fib_hash_table(u32 id
)
2010 struct fib_table
*tb
;
2013 tb
= kmalloc(sizeof(struct fib_table
) + sizeof(struct trie
),
2019 tb
->tb_default
= -1;
2020 tb
->tb_lookup
= fn_trie_lookup
;
2021 tb
->tb_insert
= fn_trie_insert
;
2022 tb
->tb_delete
= fn_trie_delete
;
2023 tb
->tb_flush
= fn_trie_flush
;
2024 tb
->tb_select_default
= fn_trie_select_default
;
2025 tb
->tb_dump
= fn_trie_dump
;
2027 t
= (struct trie
*) tb
->tb_data
;
2028 memset(t
, 0, sizeof(*t
));
2030 if (id
== RT_TABLE_LOCAL
)
2031 pr_info("IPv4 FIB: Using LC-trie version %s\n", VERSION
);
2036 #ifdef CONFIG_PROC_FS
2037 /* Depth first Trie walk iterator */
2038 struct fib_trie_iter
{
2039 struct seq_net_private p
;
2040 struct fib_table
*tb
;
2041 struct tnode
*tnode
;
2046 static struct node
*fib_trie_get_next(struct fib_trie_iter
*iter
)
2048 struct tnode
*tn
= iter
->tnode
;
2049 unsigned cindex
= iter
->index
;
2052 /* A single entry routing table */
2056 pr_debug("get_next iter={node=%p index=%d depth=%d}\n",
2057 iter
->tnode
, iter
->index
, iter
->depth
);
2059 while (cindex
< (1<<tn
->bits
)) {
2060 struct node
*n
= tnode_get_child_rcu(tn
, cindex
);
2065 iter
->index
= cindex
+ 1;
2067 /* push down one level */
2068 iter
->tnode
= (struct tnode
*) n
;
2078 /* Current node exhausted, pop back up */
2079 p
= node_parent_rcu((struct node
*)tn
);
2081 cindex
= tkey_extract_bits(tn
->key
, p
->pos
, p
->bits
)+1;
2091 static struct node
*fib_trie_get_first(struct fib_trie_iter
*iter
,
2099 n
= rcu_dereference(t
->trie
);
2104 iter
->tnode
= (struct tnode
*) n
;
2116 static void trie_collect_stats(struct trie
*t
, struct trie_stat
*s
)
2119 struct fib_trie_iter iter
;
2121 memset(s
, 0, sizeof(*s
));
2124 for (n
= fib_trie_get_first(&iter
, t
); n
; n
= fib_trie_get_next(&iter
)) {
2126 struct leaf
*l
= (struct leaf
*)n
;
2127 struct leaf_info
*li
;
2128 struct hlist_node
*tmp
;
2131 s
->totdepth
+= iter
.depth
;
2132 if (iter
.depth
> s
->maxdepth
)
2133 s
->maxdepth
= iter
.depth
;
2135 hlist_for_each_entry_rcu(li
, tmp
, &l
->list
, hlist
)
2138 const struct tnode
*tn
= (const struct tnode
*) n
;
2142 if (tn
->bits
< MAX_STAT_DEPTH
)
2143 s
->nodesizes
[tn
->bits
]++;
2145 for (i
= 0; i
< (1<<tn
->bits
); i
++)
2154 * This outputs /proc/net/fib_triestats
2156 static void trie_show_stats(struct seq_file
*seq
, struct trie_stat
*stat
)
2158 unsigned i
, max
, pointers
, bytes
, avdepth
;
2161 avdepth
= stat
->totdepth
*100 / stat
->leaves
;
2165 seq_printf(seq
, "\tAver depth: %u.%02d\n",
2166 avdepth
/ 100, avdepth
% 100);
2167 seq_printf(seq
, "\tMax depth: %u\n", stat
->maxdepth
);
2169 seq_printf(seq
, "\tLeaves: %u\n", stat
->leaves
);
2170 bytes
= sizeof(struct leaf
) * stat
->leaves
;
2172 seq_printf(seq
, "\tPrefixes: %u\n", stat
->prefixes
);
2173 bytes
+= sizeof(struct leaf_info
) * stat
->prefixes
;
2175 seq_printf(seq
, "\tInternal nodes: %u\n\t", stat
->tnodes
);
2176 bytes
+= sizeof(struct tnode
) * stat
->tnodes
;
2178 max
= MAX_STAT_DEPTH
;
2179 while (max
> 0 && stat
->nodesizes
[max
-1] == 0)
2183 for (i
= 1; i
<= max
; i
++)
2184 if (stat
->nodesizes
[i
] != 0) {
2185 seq_printf(seq
, " %u: %u", i
, stat
->nodesizes
[i
]);
2186 pointers
+= (1<<i
) * stat
->nodesizes
[i
];
2188 seq_putc(seq
, '\n');
2189 seq_printf(seq
, "\tPointers: %u\n", pointers
);
2191 bytes
+= sizeof(struct node
*) * pointers
;
2192 seq_printf(seq
, "Null ptrs: %u\n", stat
->nullpointers
);
2193 seq_printf(seq
, "Total size: %u kB\n", (bytes
+ 1023) / 1024);
2196 #ifdef CONFIG_IP_FIB_TRIE_STATS
2197 static void trie_show_usage(struct seq_file
*seq
,
2198 const struct trie_use_stats
*stats
)
2200 seq_printf(seq
, "\nCounters:\n---------\n");
2201 seq_printf(seq
, "gets = %u\n", stats
->gets
);
2202 seq_printf(seq
, "backtracks = %u\n", stats
->backtrack
);
2203 seq_printf(seq
, "semantic match passed = %u\n",
2204 stats
->semantic_match_passed
);
2205 seq_printf(seq
, "semantic match miss = %u\n",
2206 stats
->semantic_match_miss
);
2207 seq_printf(seq
, "null node hit= %u\n", stats
->null_node_hit
);
2208 seq_printf(seq
, "skipped node resize = %u\n\n",
2209 stats
->resize_node_skipped
);
2211 #endif /* CONFIG_IP_FIB_TRIE_STATS */
2213 static void fib_table_print(struct seq_file
*seq
, struct fib_table
*tb
)
2215 if (tb
->tb_id
== RT_TABLE_LOCAL
)
2216 seq_puts(seq
, "Local:\n");
2217 else if (tb
->tb_id
== RT_TABLE_MAIN
)
2218 seq_puts(seq
, "Main:\n");
2220 seq_printf(seq
, "Id %d:\n", tb
->tb_id
);
2224 static int fib_triestat_seq_show(struct seq_file
*seq
, void *v
)
2226 struct net
*net
= (struct net
*)seq
->private;
2230 "Basic info: size of leaf:"
2231 " %Zd bytes, size of tnode: %Zd bytes.\n",
2232 sizeof(struct leaf
), sizeof(struct tnode
));
2234 for (h
= 0; h
< FIB_TABLE_HASHSZ
; h
++) {
2235 struct hlist_head
*head
= &net
->ipv4
.fib_table_hash
[h
];
2236 struct hlist_node
*node
;
2237 struct fib_table
*tb
;
2239 hlist_for_each_entry_rcu(tb
, node
, head
, tb_hlist
) {
2240 struct trie
*t
= (struct trie
*) tb
->tb_data
;
2241 struct trie_stat stat
;
2246 fib_table_print(seq
, tb
);
2248 trie_collect_stats(t
, &stat
);
2249 trie_show_stats(seq
, &stat
);
2250 #ifdef CONFIG_IP_FIB_TRIE_STATS
2251 trie_show_usage(seq
, &t
->stats
);
2259 static int fib_triestat_seq_open(struct inode
*inode
, struct file
*file
)
2264 net
= get_proc_net(inode
);
2267 err
= single_open(file
, fib_triestat_seq_show
, net
);
2275 static int fib_triestat_seq_release(struct inode
*ino
, struct file
*f
)
2277 struct seq_file
*seq
= f
->private_data
;
2278 put_net(seq
->private);
2279 return single_release(ino
, f
);
2282 static const struct file_operations fib_triestat_fops
= {
2283 .owner
= THIS_MODULE
,
2284 .open
= fib_triestat_seq_open
,
2286 .llseek
= seq_lseek
,
2287 .release
= fib_triestat_seq_release
,
2290 static struct node
*fib_trie_get_idx(struct seq_file
*seq
, loff_t pos
)
2292 struct fib_trie_iter
*iter
= seq
->private;
2293 struct net
*net
= seq_file_net(seq
);
2297 for (h
= 0; h
< FIB_TABLE_HASHSZ
; h
++) {
2298 struct hlist_head
*head
= &net
->ipv4
.fib_table_hash
[h
];
2299 struct hlist_node
*node
;
2300 struct fib_table
*tb
;
2302 hlist_for_each_entry_rcu(tb
, node
, head
, tb_hlist
) {
2305 for (n
= fib_trie_get_first(iter
,
2306 (struct trie
*) tb
->tb_data
);
2307 n
; n
= fib_trie_get_next(iter
))
2318 static void *fib_trie_seq_start(struct seq_file
*seq
, loff_t
*pos
)
2322 return fib_trie_get_idx(seq
, *pos
);
2325 static void *fib_trie_seq_next(struct seq_file
*seq
, void *v
, loff_t
*pos
)
2327 struct fib_trie_iter
*iter
= seq
->private;
2328 struct net
*net
= seq_file_net(seq
);
2329 struct fib_table
*tb
= iter
->tb
;
2330 struct hlist_node
*tb_node
;
2335 /* next node in same table */
2336 n
= fib_trie_get_next(iter
);
2340 /* walk rest of this hash chain */
2341 h
= tb
->tb_id
& (FIB_TABLE_HASHSZ
- 1);
2342 while ( (tb_node
= rcu_dereference(tb
->tb_hlist
.next
)) ) {
2343 tb
= hlist_entry(tb_node
, struct fib_table
, tb_hlist
);
2344 n
= fib_trie_get_first(iter
, (struct trie
*) tb
->tb_data
);
2349 /* new hash chain */
2350 while (++h
< FIB_TABLE_HASHSZ
) {
2351 struct hlist_head
*head
= &net
->ipv4
.fib_table_hash
[h
];
2352 hlist_for_each_entry_rcu(tb
, tb_node
, head
, tb_hlist
) {
2353 n
= fib_trie_get_first(iter
, (struct trie
*) tb
->tb_data
);
2365 static void fib_trie_seq_stop(struct seq_file
*seq
, void *v
)
2371 static void seq_indent(struct seq_file
*seq
, int n
)
2373 while (n
-- > 0) seq_puts(seq
, " ");
2376 static inline const char *rtn_scope(char *buf
, size_t len
, enum rt_scope_t s
)
2379 case RT_SCOPE_UNIVERSE
: return "universe";
2380 case RT_SCOPE_SITE
: return "site";
2381 case RT_SCOPE_LINK
: return "link";
2382 case RT_SCOPE_HOST
: return "host";
2383 case RT_SCOPE_NOWHERE
: return "nowhere";
2385 snprintf(buf
, len
, "scope=%d", s
);
2390 static const char *rtn_type_names
[__RTN_MAX
] = {
2391 [RTN_UNSPEC
] = "UNSPEC",
2392 [RTN_UNICAST
] = "UNICAST",
2393 [RTN_LOCAL
] = "LOCAL",
2394 [RTN_BROADCAST
] = "BROADCAST",
2395 [RTN_ANYCAST
] = "ANYCAST",
2396 [RTN_MULTICAST
] = "MULTICAST",
2397 [RTN_BLACKHOLE
] = "BLACKHOLE",
2398 [RTN_UNREACHABLE
] = "UNREACHABLE",
2399 [RTN_PROHIBIT
] = "PROHIBIT",
2400 [RTN_THROW
] = "THROW",
2402 [RTN_XRESOLVE
] = "XRESOLVE",
2405 static inline const char *rtn_type(char *buf
, size_t len
, unsigned t
)
2407 if (t
< __RTN_MAX
&& rtn_type_names
[t
])
2408 return rtn_type_names
[t
];
2409 snprintf(buf
, len
, "type %u", t
);
2413 /* Pretty print the trie */
2414 static int fib_trie_seq_show(struct seq_file
*seq
, void *v
)
2416 const struct fib_trie_iter
*iter
= seq
->private;
2419 if (!node_parent_rcu(n
))
2420 fib_table_print(seq
, iter
->tb
);
2423 struct tnode
*tn
= (struct tnode
*) n
;
2424 __be32 prf
= htonl(mask_pfx(tn
->key
, tn
->pos
));
2426 seq_indent(seq
, iter
->depth
-1);
2427 seq_printf(seq
, " +-- " NIPQUAD_FMT
"/%d %d %d %d\n",
2428 NIPQUAD(prf
), tn
->pos
, tn
->bits
, tn
->full_children
,
2429 tn
->empty_children
);
2432 struct leaf
*l
= (struct leaf
*) n
;
2433 struct leaf_info
*li
;
2434 struct hlist_node
*node
;
2435 __be32 val
= htonl(l
->key
);
2437 seq_indent(seq
, iter
->depth
);
2438 seq_printf(seq
, " |-- " NIPQUAD_FMT
"\n", NIPQUAD(val
));
2440 hlist_for_each_entry_rcu(li
, node
, &l
->list
, hlist
) {
2441 struct fib_alias
*fa
;
2443 list_for_each_entry_rcu(fa
, &li
->falh
, fa_list
) {
2444 char buf1
[32], buf2
[32];
2446 seq_indent(seq
, iter
->depth
+1);
2447 seq_printf(seq
, " /%d %s %s", li
->plen
,
2448 rtn_scope(buf1
, sizeof(buf1
),
2450 rtn_type(buf2
, sizeof(buf2
),
2453 seq_printf(seq
, " tos=%d", fa
->fa_tos
);
2454 seq_putc(seq
, '\n');
2462 static const struct seq_operations fib_trie_seq_ops
= {
2463 .start
= fib_trie_seq_start
,
2464 .next
= fib_trie_seq_next
,
2465 .stop
= fib_trie_seq_stop
,
2466 .show
= fib_trie_seq_show
,
2469 static int fib_trie_seq_open(struct inode
*inode
, struct file
*file
)
2471 return seq_open_net(inode
, file
, &fib_trie_seq_ops
,
2472 sizeof(struct fib_trie_iter
));
2475 static const struct file_operations fib_trie_fops
= {
2476 .owner
= THIS_MODULE
,
2477 .open
= fib_trie_seq_open
,
2479 .llseek
= seq_lseek
,
2480 .release
= seq_release_net
,
2483 struct fib_route_iter
{
2484 struct seq_net_private p
;
2485 struct trie
*main_trie
;
2490 static struct leaf
*fib_route_get_idx(struct fib_route_iter
*iter
, loff_t pos
)
2492 struct leaf
*l
= NULL
;
2493 struct trie
*t
= iter
->main_trie
;
2495 /* use cache location of last found key */
2496 if (iter
->pos
> 0 && pos
>= iter
->pos
&& (l
= fib_find_node(t
, iter
->key
)))
2500 l
= trie_firstleaf(t
);
2503 while (l
&& pos
-- > 0) {
2505 l
= trie_nextleaf(l
);
2509 iter
->key
= pos
; /* remember it */
2511 iter
->pos
= 0; /* forget it */
2516 static void *fib_route_seq_start(struct seq_file
*seq
, loff_t
*pos
)
2519 struct fib_route_iter
*iter
= seq
->private;
2520 struct fib_table
*tb
;
2523 tb
= fib_get_table(seq_file_net(seq
), RT_TABLE_MAIN
);
2527 iter
->main_trie
= (struct trie
*) tb
->tb_data
;
2529 return SEQ_START_TOKEN
;
2531 return fib_route_get_idx(iter
, *pos
- 1);
2534 static void *fib_route_seq_next(struct seq_file
*seq
, void *v
, loff_t
*pos
)
2536 struct fib_route_iter
*iter
= seq
->private;
2540 if (v
== SEQ_START_TOKEN
) {
2542 l
= trie_firstleaf(iter
->main_trie
);
2545 l
= trie_nextleaf(l
);
2555 static void fib_route_seq_stop(struct seq_file
*seq
, void *v
)
2561 static unsigned fib_flag_trans(int type
, __be32 mask
, const struct fib_info
*fi
)
2563 static unsigned type2flags
[RTN_MAX
+ 1] = {
2564 [7] = RTF_REJECT
, [8] = RTF_REJECT
,
2566 unsigned flags
= type2flags
[type
];
2568 if (fi
&& fi
->fib_nh
->nh_gw
)
2569 flags
|= RTF_GATEWAY
;
2570 if (mask
== htonl(0xFFFFFFFF))
2577 * This outputs /proc/net/route.
2578 * The format of the file is not supposed to be changed
2579 * and needs to be same as fib_hash output to avoid breaking
2582 static int fib_route_seq_show(struct seq_file
*seq
, void *v
)
2585 struct leaf_info
*li
;
2586 struct hlist_node
*node
;
2588 if (v
== SEQ_START_TOKEN
) {
2589 seq_printf(seq
, "%-127s\n", "Iface\tDestination\tGateway "
2590 "\tFlags\tRefCnt\tUse\tMetric\tMask\t\tMTU"
2595 hlist_for_each_entry_rcu(li
, node
, &l
->list
, hlist
) {
2596 struct fib_alias
*fa
;
2597 __be32 mask
, prefix
;
2599 mask
= inet_make_mask(li
->plen
);
2600 prefix
= htonl(l
->key
);
2602 list_for_each_entry_rcu(fa
, &li
->falh
, fa_list
) {
2603 const struct fib_info
*fi
= fa
->fa_info
;
2604 unsigned flags
= fib_flag_trans(fa
->fa_type
, mask
, fi
);
2607 if (fa
->fa_type
== RTN_BROADCAST
2608 || fa
->fa_type
== RTN_MULTICAST
)
2613 "%s\t%08X\t%08X\t%04X\t%d\t%u\t"
2614 "%d\t%08X\t%d\t%u\t%u%n",
2615 fi
->fib_dev
? fi
->fib_dev
->name
: "*",
2617 fi
->fib_nh
->nh_gw
, flags
, 0, 0,
2621 fi
->fib_advmss
+ 40 : 0),
2623 fi
->fib_rtt
>> 3, &len
);
2626 "*\t%08X\t%08X\t%04X\t%d\t%u\t"
2627 "%d\t%08X\t%d\t%u\t%u%n",
2628 prefix
, 0, flags
, 0, 0, 0,
2629 mask
, 0, 0, 0, &len
);
2631 seq_printf(seq
, "%*s\n", 127 - len
, "");
2638 static const struct seq_operations fib_route_seq_ops
= {
2639 .start
= fib_route_seq_start
,
2640 .next
= fib_route_seq_next
,
2641 .stop
= fib_route_seq_stop
,
2642 .show
= fib_route_seq_show
,
2645 static int fib_route_seq_open(struct inode
*inode
, struct file
*file
)
2647 return seq_open_net(inode
, file
, &fib_route_seq_ops
,
2648 sizeof(struct fib_route_iter
));
2651 static const struct file_operations fib_route_fops
= {
2652 .owner
= THIS_MODULE
,
2653 .open
= fib_route_seq_open
,
2655 .llseek
= seq_lseek
,
2656 .release
= seq_release_net
,
2659 int __net_init
fib_proc_init(struct net
*net
)
2661 if (!proc_net_fops_create(net
, "fib_trie", S_IRUGO
, &fib_trie_fops
))
2664 if (!proc_net_fops_create(net
, "fib_triestat", S_IRUGO
,
2665 &fib_triestat_fops
))
2668 if (!proc_net_fops_create(net
, "route", S_IRUGO
, &fib_route_fops
))
2674 proc_net_remove(net
, "fib_triestat");
2676 proc_net_remove(net
, "fib_trie");
2681 void __net_exit
fib_proc_exit(struct net
*net
)
2683 proc_net_remove(net
, "fib_trie");
2684 proc_net_remove(net
, "fib_triestat");
2685 proc_net_remove(net
, "route");
2688 #endif /* CONFIG_PROC_FS */