]> git.proxmox.com Git - mirror_ubuntu-bionic-kernel.git/blob - net/ipv4/fib_trie.c
Kobject: convert remaining kobject_unregister() to kobject_put()
[mirror_ubuntu-bionic-kernel.git] / net / ipv4 / fib_trie.c
1 /*
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.
6 *
7 * Robert Olsson <robert.olsson@its.uu.se> Uppsala Universitet
8 * & Swedish University of Agricultural Sciences.
9 *
10 * Jens Laas <jens.laas@data.slu.se> Swedish University of
11 * Agricultural Sciences.
12 *
13 * Hans Liss <hans.liss@its.uu.se> Uppsala Universitet
14 *
15 * This work is based on the LPC-trie which is originally descibed in:
16 *
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/
20 *
21 *
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
24 *
25 * Version: $Id: fib_trie.c,v 1.3 2005/06/08 14:20:01 robert Exp $
26 *
27 *
28 * Code from fib_hash has been reused which includes the following header:
29 *
30 *
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.
34 *
35 * IPv4 FIB: lookup engine and maintenance routines.
36 *
37 *
38 * Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
39 *
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.
44 *
45 * Substantial contributions to this work comes from:
46 *
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>
51 */
52
53 #define VERSION "0.408"
54
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>
60 #include <linux/mm.h>
61 #include <linux/string.h>
62 #include <linux/socket.h>
63 #include <linux/sockios.h>
64 #include <linux/errno.h>
65 #include <linux/in.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>
77 #include <net/ip.h>
78 #include <net/protocol.h>
79 #include <net/route.h>
80 #include <net/tcp.h>
81 #include <net/sock.h>
82 #include <net/ip_fib.h>
83 #include "fib_lookup.h"
84
85 #undef CONFIG_IP_FIB_TRIE_STATS
86 #define MAX_STAT_DEPTH 32
87
88 #define KEYLENGTH (8*sizeof(t_key))
89
90 typedef unsigned int t_key;
91
92 #define T_TNODE 0
93 #define T_LEAF 1
94 #define NODE_TYPE_MASK 0x1UL
95 #define NODE_TYPE(node) ((node)->parent & NODE_TYPE_MASK)
96
97 #define IS_TNODE(n) (!(n->parent & T_LEAF))
98 #define IS_LEAF(n) (n->parent & T_LEAF)
99
100 struct node {
101 t_key key;
102 unsigned long parent;
103 };
104
105 struct leaf {
106 t_key key;
107 unsigned long parent;
108 struct hlist_head list;
109 struct rcu_head rcu;
110 };
111
112 struct leaf_info {
113 struct hlist_node hlist;
114 struct rcu_head rcu;
115 int plen;
116 struct list_head falh;
117 };
118
119 struct tnode {
120 t_key key;
121 unsigned long parent;
122 unsigned short pos:5; /* 2log(KEYLENGTH) bits needed */
123 unsigned short bits:5; /* 2log(KEYLENGTH) bits needed */
124 unsigned short full_children; /* KEYLENGTH bits needed */
125 unsigned short empty_children; /* KEYLENGTH bits needed */
126 struct rcu_head rcu;
127 struct node *child[0];
128 };
129
130 #ifdef CONFIG_IP_FIB_TRIE_STATS
131 struct trie_use_stats {
132 unsigned int gets;
133 unsigned int backtrack;
134 unsigned int semantic_match_passed;
135 unsigned int semantic_match_miss;
136 unsigned int null_node_hit;
137 unsigned int resize_node_skipped;
138 };
139 #endif
140
141 struct trie_stat {
142 unsigned int totdepth;
143 unsigned int maxdepth;
144 unsigned int tnodes;
145 unsigned int leaves;
146 unsigned int nullpointers;
147 unsigned int nodesizes[MAX_STAT_DEPTH];
148 };
149
150 struct trie {
151 struct node *trie;
152 #ifdef CONFIG_IP_FIB_TRIE_STATS
153 struct trie_use_stats stats;
154 #endif
155 int size;
156 unsigned int revision;
157 };
158
159 static void put_child(struct trie *t, struct tnode *tn, int i, struct node *n);
160 static void tnode_put_child_reorg(struct tnode *tn, int i, struct node *n, int wasfull);
161 static struct node *resize(struct trie *t, struct tnode *tn);
162 static struct tnode *inflate(struct trie *t, struct tnode *tn);
163 static struct tnode *halve(struct trie *t, struct tnode *tn);
164 static void tnode_free(struct tnode *tn);
165
166 static struct kmem_cache *fn_alias_kmem __read_mostly;
167 static struct trie *trie_local = NULL, *trie_main = NULL;
168
169 static inline struct tnode *node_parent(struct node *node)
170 {
171 struct tnode *ret;
172
173 ret = (struct tnode *)(node->parent & ~NODE_TYPE_MASK);
174 return rcu_dereference(ret);
175 }
176
177 static inline void node_set_parent(struct node *node, struct tnode *ptr)
178 {
179 rcu_assign_pointer(node->parent,
180 (unsigned long)ptr | NODE_TYPE(node));
181 }
182
183 /* rcu_read_lock needs to be hold by caller from readside */
184
185 static inline struct node *tnode_get_child(struct tnode *tn, int i)
186 {
187 BUG_ON(i >= 1 << tn->bits);
188
189 return rcu_dereference(tn->child[i]);
190 }
191
192 static inline int tnode_child_length(const struct tnode *tn)
193 {
194 return 1 << tn->bits;
195 }
196
197 static inline t_key mask_pfx(t_key k, unsigned short l)
198 {
199 return (l == 0) ? 0 : k >> (KEYLENGTH-l) << (KEYLENGTH-l);
200 }
201
202 static inline t_key tkey_extract_bits(t_key a, int offset, int bits)
203 {
204 if (offset < KEYLENGTH)
205 return ((t_key)(a << offset)) >> (KEYLENGTH - bits);
206 else
207 return 0;
208 }
209
210 static inline int tkey_equals(t_key a, t_key b)
211 {
212 return a == b;
213 }
214
215 static inline int tkey_sub_equals(t_key a, int offset, int bits, t_key b)
216 {
217 if (bits == 0 || offset >= KEYLENGTH)
218 return 1;
219 bits = bits > KEYLENGTH ? KEYLENGTH : bits;
220 return ((a ^ b) << offset) >> (KEYLENGTH - bits) == 0;
221 }
222
223 static inline int tkey_mismatch(t_key a, int offset, t_key b)
224 {
225 t_key diff = a ^ b;
226 int i = offset;
227
228 if (!diff)
229 return 0;
230 while ((diff << i) >> (KEYLENGTH-1) == 0)
231 i++;
232 return i;
233 }
234
235 /*
236 To understand this stuff, an understanding of keys and all their bits is
237 necessary. Every node in the trie has a key associated with it, but not
238 all of the bits in that key are significant.
239
240 Consider a node 'n' and its parent 'tp'.
241
242 If n is a leaf, every bit in its key is significant. Its presence is
243 necessitated by path compression, since during a tree traversal (when
244 searching for a leaf - unless we are doing an insertion) we will completely
245 ignore all skipped bits we encounter. Thus we need to verify, at the end of
246 a potentially successful search, that we have indeed been walking the
247 correct key path.
248
249 Note that we can never "miss" the correct key in the tree if present by
250 following the wrong path. Path compression ensures that segments of the key
251 that are the same for all keys with a given prefix are skipped, but the
252 skipped part *is* identical for each node in the subtrie below the skipped
253 bit! trie_insert() in this implementation takes care of that - note the
254 call to tkey_sub_equals() in trie_insert().
255
256 if n is an internal node - a 'tnode' here, the various parts of its key
257 have many different meanings.
258
259 Example:
260 _________________________________________________________________
261 | i | i | i | i | i | i | i | N | N | N | S | S | S | S | S | C |
262 -----------------------------------------------------------------
263 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
264
265 _________________________________________________________________
266 | C | C | C | u | u | u | u | u | u | u | u | u | u | u | u | u |
267 -----------------------------------------------------------------
268 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
269
270 tp->pos = 7
271 tp->bits = 3
272 n->pos = 15
273 n->bits = 4
274
275 First, let's just ignore the bits that come before the parent tp, that is
276 the bits from 0 to (tp->pos-1). They are *known* but at this point we do
277 not use them for anything.
278
279 The bits from (tp->pos) to (tp->pos + tp->bits - 1) - "N", above - are the
280 index into the parent's child array. That is, they will be used to find
281 'n' among tp's children.
282
283 The bits from (tp->pos + tp->bits) to (n->pos - 1) - "S" - are skipped bits
284 for the node n.
285
286 All the bits we have seen so far are significant to the node n. The rest
287 of the bits are really not needed or indeed known in n->key.
288
289 The bits from (n->pos) to (n->pos + n->bits - 1) - "C" - are the index into
290 n's child array, and will of course be different for each child.
291
292
293 The rest of the bits, from (n->pos + n->bits) onward, are completely unknown
294 at this point.
295
296 */
297
298 static inline void check_tnode(const struct tnode *tn)
299 {
300 WARN_ON(tn && tn->pos+tn->bits > 32);
301 }
302
303 static int halve_threshold = 25;
304 static int inflate_threshold = 50;
305 static int halve_threshold_root = 8;
306 static int inflate_threshold_root = 15;
307
308
309 static void __alias_free_mem(struct rcu_head *head)
310 {
311 struct fib_alias *fa = container_of(head, struct fib_alias, rcu);
312 kmem_cache_free(fn_alias_kmem, fa);
313 }
314
315 static inline void alias_free_mem_rcu(struct fib_alias *fa)
316 {
317 call_rcu(&fa->rcu, __alias_free_mem);
318 }
319
320 static void __leaf_free_rcu(struct rcu_head *head)
321 {
322 kfree(container_of(head, struct leaf, rcu));
323 }
324
325 static void __leaf_info_free_rcu(struct rcu_head *head)
326 {
327 kfree(container_of(head, struct leaf_info, rcu));
328 }
329
330 static inline void free_leaf_info(struct leaf_info *leaf)
331 {
332 call_rcu(&leaf->rcu, __leaf_info_free_rcu);
333 }
334
335 static struct tnode *tnode_alloc(unsigned int size)
336 {
337 struct page *pages;
338
339 if (size <= PAGE_SIZE)
340 return kcalloc(size, 1, GFP_KERNEL);
341
342 pages = alloc_pages(GFP_KERNEL|__GFP_ZERO, get_order(size));
343 if (!pages)
344 return NULL;
345
346 return page_address(pages);
347 }
348
349 static void __tnode_free_rcu(struct rcu_head *head)
350 {
351 struct tnode *tn = container_of(head, struct tnode, rcu);
352 unsigned int size = sizeof(struct tnode) +
353 (1 << tn->bits) * sizeof(struct node *);
354
355 if (size <= PAGE_SIZE)
356 kfree(tn);
357 else
358 free_pages((unsigned long)tn, get_order(size));
359 }
360
361 static inline void tnode_free(struct tnode *tn)
362 {
363 if (IS_LEAF(tn)) {
364 struct leaf *l = (struct leaf *) tn;
365 call_rcu_bh(&l->rcu, __leaf_free_rcu);
366 } else
367 call_rcu(&tn->rcu, __tnode_free_rcu);
368 }
369
370 static struct leaf *leaf_new(void)
371 {
372 struct leaf *l = kmalloc(sizeof(struct leaf), GFP_KERNEL);
373 if (l) {
374 l->parent = T_LEAF;
375 INIT_HLIST_HEAD(&l->list);
376 }
377 return l;
378 }
379
380 static struct leaf_info *leaf_info_new(int plen)
381 {
382 struct leaf_info *li = kmalloc(sizeof(struct leaf_info), GFP_KERNEL);
383 if (li) {
384 li->plen = plen;
385 INIT_LIST_HEAD(&li->falh);
386 }
387 return li;
388 }
389
390 static struct tnode* tnode_new(t_key key, int pos, int bits)
391 {
392 int nchildren = 1<<bits;
393 int sz = sizeof(struct tnode) + nchildren * sizeof(struct node *);
394 struct tnode *tn = tnode_alloc(sz);
395
396 if (tn) {
397 memset(tn, 0, sz);
398 tn->parent = T_TNODE;
399 tn->pos = pos;
400 tn->bits = bits;
401 tn->key = key;
402 tn->full_children = 0;
403 tn->empty_children = 1<<bits;
404 }
405
406 pr_debug("AT %p s=%u %u\n", tn, (unsigned int) sizeof(struct tnode),
407 (unsigned int) (sizeof(struct node) * 1<<bits));
408 return tn;
409 }
410
411 /*
412 * Check whether a tnode 'n' is "full", i.e. it is an internal node
413 * and no bits are skipped. See discussion in dyntree paper p. 6
414 */
415
416 static inline int tnode_full(const struct tnode *tn, const struct node *n)
417 {
418 if (n == NULL || IS_LEAF(n))
419 return 0;
420
421 return ((struct tnode *) n)->pos == tn->pos + tn->bits;
422 }
423
424 static inline void put_child(struct trie *t, struct tnode *tn, int i, struct node *n)
425 {
426 tnode_put_child_reorg(tn, i, n, -1);
427 }
428
429 /*
430 * Add a child at position i overwriting the old value.
431 * Update the value of full_children and empty_children.
432 */
433
434 static void tnode_put_child_reorg(struct tnode *tn, int i, struct node *n, int wasfull)
435 {
436 struct node *chi = tn->child[i];
437 int isfull;
438
439 BUG_ON(i >= 1<<tn->bits);
440
441
442 /* update emptyChildren */
443 if (n == NULL && chi != NULL)
444 tn->empty_children++;
445 else if (n != NULL && chi == NULL)
446 tn->empty_children--;
447
448 /* update fullChildren */
449 if (wasfull == -1)
450 wasfull = tnode_full(tn, chi);
451
452 isfull = tnode_full(tn, n);
453 if (wasfull && !isfull)
454 tn->full_children--;
455 else if (!wasfull && isfull)
456 tn->full_children++;
457
458 if (n)
459 node_set_parent(n, tn);
460
461 rcu_assign_pointer(tn->child[i], n);
462 }
463
464 static struct node *resize(struct trie *t, struct tnode *tn)
465 {
466 int i;
467 int err = 0;
468 struct tnode *old_tn;
469 int inflate_threshold_use;
470 int halve_threshold_use;
471 int max_resize;
472
473 if (!tn)
474 return NULL;
475
476 pr_debug("In tnode_resize %p inflate_threshold=%d threshold=%d\n",
477 tn, inflate_threshold, halve_threshold);
478
479 /* No children */
480 if (tn->empty_children == tnode_child_length(tn)) {
481 tnode_free(tn);
482 return NULL;
483 }
484 /* One child */
485 if (tn->empty_children == tnode_child_length(tn) - 1)
486 for (i = 0; i < tnode_child_length(tn); i++) {
487 struct node *n;
488
489 n = tn->child[i];
490 if (!n)
491 continue;
492
493 /* compress one level */
494 node_set_parent(n, NULL);
495 tnode_free(tn);
496 return n;
497 }
498 /*
499 * Double as long as the resulting node has a number of
500 * nonempty nodes that are above the threshold.
501 */
502
503 /*
504 * From "Implementing a dynamic compressed trie" by Stefan Nilsson of
505 * the Helsinki University of Technology and Matti Tikkanen of Nokia
506 * Telecommunications, page 6:
507 * "A node is doubled if the ratio of non-empty children to all
508 * children in the *doubled* node is at least 'high'."
509 *
510 * 'high' in this instance is the variable 'inflate_threshold'. It
511 * is expressed as a percentage, so we multiply it with
512 * tnode_child_length() and instead of multiplying by 2 (since the
513 * child array will be doubled by inflate()) and multiplying
514 * the left-hand side by 100 (to handle the percentage thing) we
515 * multiply the left-hand side by 50.
516 *
517 * The left-hand side may look a bit weird: tnode_child_length(tn)
518 * - tn->empty_children is of course the number of non-null children
519 * in the current node. tn->full_children is the number of "full"
520 * children, that is non-null tnodes with a skip value of 0.
521 * All of those will be doubled in the resulting inflated tnode, so
522 * we just count them one extra time here.
523 *
524 * A clearer way to write this would be:
525 *
526 * to_be_doubled = tn->full_children;
527 * not_to_be_doubled = tnode_child_length(tn) - tn->empty_children -
528 * tn->full_children;
529 *
530 * new_child_length = tnode_child_length(tn) * 2;
531 *
532 * new_fill_factor = 100 * (not_to_be_doubled + 2*to_be_doubled) /
533 * new_child_length;
534 * if (new_fill_factor >= inflate_threshold)
535 *
536 * ...and so on, tho it would mess up the while () loop.
537 *
538 * anyway,
539 * 100 * (not_to_be_doubled + 2*to_be_doubled) / new_child_length >=
540 * inflate_threshold
541 *
542 * avoid a division:
543 * 100 * (not_to_be_doubled + 2*to_be_doubled) >=
544 * inflate_threshold * new_child_length
545 *
546 * expand not_to_be_doubled and to_be_doubled, and shorten:
547 * 100 * (tnode_child_length(tn) - tn->empty_children +
548 * tn->full_children) >= inflate_threshold * new_child_length
549 *
550 * expand new_child_length:
551 * 100 * (tnode_child_length(tn) - tn->empty_children +
552 * tn->full_children) >=
553 * inflate_threshold * tnode_child_length(tn) * 2
554 *
555 * shorten again:
556 * 50 * (tn->full_children + tnode_child_length(tn) -
557 * tn->empty_children) >= inflate_threshold *
558 * tnode_child_length(tn)
559 *
560 */
561
562 check_tnode(tn);
563
564 /* Keep root node larger */
565
566 if (!tn->parent)
567 inflate_threshold_use = inflate_threshold_root;
568 else
569 inflate_threshold_use = inflate_threshold;
570
571 err = 0;
572 max_resize = 10;
573 while ((tn->full_children > 0 && max_resize-- &&
574 50 * (tn->full_children + tnode_child_length(tn) - tn->empty_children) >=
575 inflate_threshold_use * tnode_child_length(tn))) {
576
577 old_tn = tn;
578 tn = inflate(t, tn);
579 if (IS_ERR(tn)) {
580 tn = old_tn;
581 #ifdef CONFIG_IP_FIB_TRIE_STATS
582 t->stats.resize_node_skipped++;
583 #endif
584 break;
585 }
586 }
587
588 if (max_resize < 0) {
589 if (!tn->parent)
590 printk(KERN_WARNING "Fix inflate_threshold_root. Now=%d size=%d bits\n",
591 inflate_threshold_root, tn->bits);
592 else
593 printk(KERN_WARNING "Fix inflate_threshold. Now=%d size=%d bits\n",
594 inflate_threshold, tn->bits);
595 }
596
597 check_tnode(tn);
598
599 /*
600 * Halve as long as the number of empty children in this
601 * node is above threshold.
602 */
603
604
605 /* Keep root node larger */
606
607 if (!tn->parent)
608 halve_threshold_use = halve_threshold_root;
609 else
610 halve_threshold_use = halve_threshold;
611
612 err = 0;
613 max_resize = 10;
614 while (tn->bits > 1 && max_resize-- &&
615 100 * (tnode_child_length(tn) - tn->empty_children) <
616 halve_threshold_use * tnode_child_length(tn)) {
617
618 old_tn = tn;
619 tn = halve(t, tn);
620 if (IS_ERR(tn)) {
621 tn = old_tn;
622 #ifdef CONFIG_IP_FIB_TRIE_STATS
623 t->stats.resize_node_skipped++;
624 #endif
625 break;
626 }
627 }
628
629 if (max_resize < 0) {
630 if (!tn->parent)
631 printk(KERN_WARNING "Fix halve_threshold_root. Now=%d size=%d bits\n",
632 halve_threshold_root, tn->bits);
633 else
634 printk(KERN_WARNING "Fix halve_threshold. Now=%d size=%d bits\n",
635 halve_threshold, tn->bits);
636 }
637
638 /* Only one child remains */
639 if (tn->empty_children == tnode_child_length(tn) - 1)
640 for (i = 0; i < tnode_child_length(tn); i++) {
641 struct node *n;
642
643 n = tn->child[i];
644 if (!n)
645 continue;
646
647 /* compress one level */
648
649 node_set_parent(n, NULL);
650 tnode_free(tn);
651 return n;
652 }
653
654 return (struct node *) tn;
655 }
656
657 static struct tnode *inflate(struct trie *t, struct tnode *tn)
658 {
659 struct tnode *inode;
660 struct tnode *oldtnode = tn;
661 int olen = tnode_child_length(tn);
662 int i;
663
664 pr_debug("In inflate\n");
665
666 tn = tnode_new(oldtnode->key, oldtnode->pos, oldtnode->bits + 1);
667
668 if (!tn)
669 return ERR_PTR(-ENOMEM);
670
671 /*
672 * Preallocate and store tnodes before the actual work so we
673 * don't get into an inconsistent state if memory allocation
674 * fails. In case of failure we return the oldnode and inflate
675 * of tnode is ignored.
676 */
677
678 for (i = 0; i < olen; i++) {
679 struct tnode *inode = (struct tnode *) tnode_get_child(oldtnode, i);
680
681 if (inode &&
682 IS_TNODE(inode) &&
683 inode->pos == oldtnode->pos + oldtnode->bits &&
684 inode->bits > 1) {
685 struct tnode *left, *right;
686 t_key m = ~0U << (KEYLENGTH - 1) >> inode->pos;
687
688 left = tnode_new(inode->key&(~m), inode->pos + 1,
689 inode->bits - 1);
690 if (!left)
691 goto nomem;
692
693 right = tnode_new(inode->key|m, inode->pos + 1,
694 inode->bits - 1);
695
696 if (!right) {
697 tnode_free(left);
698 goto nomem;
699 }
700
701 put_child(t, tn, 2*i, (struct node *) left);
702 put_child(t, tn, 2*i+1, (struct node *) right);
703 }
704 }
705
706 for (i = 0; i < olen; i++) {
707 struct node *node = tnode_get_child(oldtnode, i);
708 struct tnode *left, *right;
709 int size, j;
710
711 /* An empty child */
712 if (node == NULL)
713 continue;
714
715 /* A leaf or an internal node with skipped bits */
716
717 if (IS_LEAF(node) || ((struct tnode *) node)->pos >
718 tn->pos + tn->bits - 1) {
719 if (tkey_extract_bits(node->key, oldtnode->pos + oldtnode->bits,
720 1) == 0)
721 put_child(t, tn, 2*i, node);
722 else
723 put_child(t, tn, 2*i+1, node);
724 continue;
725 }
726
727 /* An internal node with two children */
728 inode = (struct tnode *) node;
729
730 if (inode->bits == 1) {
731 put_child(t, tn, 2*i, inode->child[0]);
732 put_child(t, tn, 2*i+1, inode->child[1]);
733
734 tnode_free(inode);
735 continue;
736 }
737
738 /* An internal node with more than two children */
739
740 /* We will replace this node 'inode' with two new
741 * ones, 'left' and 'right', each with half of the
742 * original children. The two new nodes will have
743 * a position one bit further down the key and this
744 * means that the "significant" part of their keys
745 * (see the discussion near the top of this file)
746 * will differ by one bit, which will be "0" in
747 * left's key and "1" in right's key. Since we are
748 * moving the key position by one step, the bit that
749 * we are moving away from - the bit at position
750 * (inode->pos) - is the one that will differ between
751 * left and right. So... we synthesize that bit in the
752 * two new keys.
753 * The mask 'm' below will be a single "one" bit at
754 * the position (inode->pos)
755 */
756
757 /* Use the old key, but set the new significant
758 * bit to zero.
759 */
760
761 left = (struct tnode *) tnode_get_child(tn, 2*i);
762 put_child(t, tn, 2*i, NULL);
763
764 BUG_ON(!left);
765
766 right = (struct tnode *) tnode_get_child(tn, 2*i+1);
767 put_child(t, tn, 2*i+1, NULL);
768
769 BUG_ON(!right);
770
771 size = tnode_child_length(left);
772 for (j = 0; j < size; j++) {
773 put_child(t, left, j, inode->child[j]);
774 put_child(t, right, j, inode->child[j + size]);
775 }
776 put_child(t, tn, 2*i, resize(t, left));
777 put_child(t, tn, 2*i+1, resize(t, right));
778
779 tnode_free(inode);
780 }
781 tnode_free(oldtnode);
782 return tn;
783 nomem:
784 {
785 int size = tnode_child_length(tn);
786 int j;
787
788 for (j = 0; j < size; j++)
789 if (tn->child[j])
790 tnode_free((struct tnode *)tn->child[j]);
791
792 tnode_free(tn);
793
794 return ERR_PTR(-ENOMEM);
795 }
796 }
797
798 static struct tnode *halve(struct trie *t, struct tnode *tn)
799 {
800 struct tnode *oldtnode = tn;
801 struct node *left, *right;
802 int i;
803 int olen = tnode_child_length(tn);
804
805 pr_debug("In halve\n");
806
807 tn = tnode_new(oldtnode->key, oldtnode->pos, oldtnode->bits - 1);
808
809 if (!tn)
810 return ERR_PTR(-ENOMEM);
811
812 /*
813 * Preallocate and store tnodes before the actual work so we
814 * don't get into an inconsistent state if memory allocation
815 * fails. In case of failure we return the oldnode and halve
816 * of tnode is ignored.
817 */
818
819 for (i = 0; i < olen; i += 2) {
820 left = tnode_get_child(oldtnode, i);
821 right = tnode_get_child(oldtnode, i+1);
822
823 /* Two nonempty children */
824 if (left && right) {
825 struct tnode *newn;
826
827 newn = tnode_new(left->key, tn->pos + tn->bits, 1);
828
829 if (!newn)
830 goto nomem;
831
832 put_child(t, tn, i/2, (struct node *)newn);
833 }
834
835 }
836
837 for (i = 0; i < olen; i += 2) {
838 struct tnode *newBinNode;
839
840 left = tnode_get_child(oldtnode, i);
841 right = tnode_get_child(oldtnode, i+1);
842
843 /* At least one of the children is empty */
844 if (left == NULL) {
845 if (right == NULL) /* Both are empty */
846 continue;
847 put_child(t, tn, i/2, right);
848 continue;
849 }
850
851 if (right == NULL) {
852 put_child(t, tn, i/2, left);
853 continue;
854 }
855
856 /* Two nonempty children */
857 newBinNode = (struct tnode *) tnode_get_child(tn, i/2);
858 put_child(t, tn, i/2, NULL);
859 put_child(t, newBinNode, 0, left);
860 put_child(t, newBinNode, 1, right);
861 put_child(t, tn, i/2, resize(t, newBinNode));
862 }
863 tnode_free(oldtnode);
864 return tn;
865 nomem:
866 {
867 int size = tnode_child_length(tn);
868 int j;
869
870 for (j = 0; j < size; j++)
871 if (tn->child[j])
872 tnode_free((struct tnode *)tn->child[j]);
873
874 tnode_free(tn);
875
876 return ERR_PTR(-ENOMEM);
877 }
878 }
879
880 static void trie_init(struct trie *t)
881 {
882 if (!t)
883 return;
884
885 t->size = 0;
886 rcu_assign_pointer(t->trie, NULL);
887 t->revision = 0;
888 #ifdef CONFIG_IP_FIB_TRIE_STATS
889 memset(&t->stats, 0, sizeof(struct trie_use_stats));
890 #endif
891 }
892
893 /* readside must use rcu_read_lock currently dump routines
894 via get_fa_head and dump */
895
896 static struct leaf_info *find_leaf_info(struct leaf *l, int plen)
897 {
898 struct hlist_head *head = &l->list;
899 struct hlist_node *node;
900 struct leaf_info *li;
901
902 hlist_for_each_entry_rcu(li, node, head, hlist)
903 if (li->plen == plen)
904 return li;
905
906 return NULL;
907 }
908
909 static inline struct list_head * get_fa_head(struct leaf *l, int plen)
910 {
911 struct leaf_info *li = find_leaf_info(l, plen);
912
913 if (!li)
914 return NULL;
915
916 return &li->falh;
917 }
918
919 static void insert_leaf_info(struct hlist_head *head, struct leaf_info *new)
920 {
921 struct leaf_info *li = NULL, *last = NULL;
922 struct hlist_node *node;
923
924 if (hlist_empty(head)) {
925 hlist_add_head_rcu(&new->hlist, head);
926 } else {
927 hlist_for_each_entry(li, node, head, hlist) {
928 if (new->plen > li->plen)
929 break;
930
931 last = li;
932 }
933 if (last)
934 hlist_add_after_rcu(&last->hlist, &new->hlist);
935 else
936 hlist_add_before_rcu(&new->hlist, &li->hlist);
937 }
938 }
939
940 /* rcu_read_lock needs to be hold by caller from readside */
941
942 static struct leaf *
943 fib_find_node(struct trie *t, u32 key)
944 {
945 int pos;
946 struct tnode *tn;
947 struct node *n;
948
949 pos = 0;
950 n = rcu_dereference(t->trie);
951
952 while (n != NULL && NODE_TYPE(n) == T_TNODE) {
953 tn = (struct tnode *) n;
954
955 check_tnode(tn);
956
957 if (tkey_sub_equals(tn->key, pos, tn->pos-pos, key)) {
958 pos = tn->pos + tn->bits;
959 n = tnode_get_child(tn, tkey_extract_bits(key, tn->pos, tn->bits));
960 } else
961 break;
962 }
963 /* Case we have found a leaf. Compare prefixes */
964
965 if (n != NULL && IS_LEAF(n) && tkey_equals(key, n->key))
966 return (struct leaf *)n;
967
968 return NULL;
969 }
970
971 static struct node *trie_rebalance(struct trie *t, struct tnode *tn)
972 {
973 int wasfull;
974 t_key cindex, key = tn->key;
975 struct tnode *tp;
976
977 while (tn != NULL && (tp = node_parent((struct node *)tn)) != NULL) {
978 cindex = tkey_extract_bits(key, tp->pos, tp->bits);
979 wasfull = tnode_full(tp, tnode_get_child(tp, cindex));
980 tn = (struct tnode *) resize (t, (struct tnode *)tn);
981 tnode_put_child_reorg((struct tnode *)tp, cindex,(struct node*)tn, wasfull);
982
983 tp = node_parent((struct node *) tn);
984 if (!tp)
985 break;
986 tn = tp;
987 }
988
989 /* Handle last (top) tnode */
990 if (IS_TNODE(tn))
991 tn = (struct tnode*) resize(t, (struct tnode *)tn);
992
993 return (struct node*) tn;
994 }
995
996 /* only used from updater-side */
997
998 static struct list_head *
999 fib_insert_node(struct trie *t, int *err, u32 key, int plen)
1000 {
1001 int pos, newpos;
1002 struct tnode *tp = NULL, *tn = NULL;
1003 struct node *n;
1004 struct leaf *l;
1005 int missbit;
1006 struct list_head *fa_head = NULL;
1007 struct leaf_info *li;
1008 t_key cindex;
1009
1010 pos = 0;
1011 n = t->trie;
1012
1013 /* If we point to NULL, stop. Either the tree is empty and we should
1014 * just put a new leaf in if, or we have reached an empty child slot,
1015 * and we should just put our new leaf in that.
1016 * If we point to a T_TNODE, check if it matches our key. Note that
1017 * a T_TNODE might be skipping any number of bits - its 'pos' need
1018 * not be the parent's 'pos'+'bits'!
1019 *
1020 * If it does match the current key, get pos/bits from it, extract
1021 * the index from our key, push the T_TNODE and walk the tree.
1022 *
1023 * If it doesn't, we have to replace it with a new T_TNODE.
1024 *
1025 * If we point to a T_LEAF, it might or might not have the same key
1026 * as we do. If it does, just change the value, update the T_LEAF's
1027 * value, and return it.
1028 * If it doesn't, we need to replace it with a T_TNODE.
1029 */
1030
1031 while (n != NULL && NODE_TYPE(n) == T_TNODE) {
1032 tn = (struct tnode *) n;
1033
1034 check_tnode(tn);
1035
1036 if (tkey_sub_equals(tn->key, pos, tn->pos-pos, key)) {
1037 tp = tn;
1038 pos = tn->pos + tn->bits;
1039 n = tnode_get_child(tn, tkey_extract_bits(key, tn->pos, tn->bits));
1040
1041 BUG_ON(n && node_parent(n) != tn);
1042 } else
1043 break;
1044 }
1045
1046 /*
1047 * n ----> NULL, LEAF or TNODE
1048 *
1049 * tp is n's (parent) ----> NULL or TNODE
1050 */
1051
1052 BUG_ON(tp && IS_LEAF(tp));
1053
1054 /* Case 1: n is a leaf. Compare prefixes */
1055
1056 if (n != NULL && IS_LEAF(n) && tkey_equals(key, n->key)) {
1057 struct leaf *l = (struct leaf *) n;
1058
1059 li = leaf_info_new(plen);
1060
1061 if (!li) {
1062 *err = -ENOMEM;
1063 goto err;
1064 }
1065
1066 fa_head = &li->falh;
1067 insert_leaf_info(&l->list, li);
1068 goto done;
1069 }
1070 t->size++;
1071 l = leaf_new();
1072
1073 if (!l) {
1074 *err = -ENOMEM;
1075 goto err;
1076 }
1077
1078 l->key = key;
1079 li = leaf_info_new(plen);
1080
1081 if (!li) {
1082 tnode_free((struct tnode *) l);
1083 *err = -ENOMEM;
1084 goto err;
1085 }
1086
1087 fa_head = &li->falh;
1088 insert_leaf_info(&l->list, li);
1089
1090 if (t->trie && n == NULL) {
1091 /* Case 2: n is NULL, and will just insert a new leaf */
1092
1093 node_set_parent((struct node *)l, tp);
1094
1095 cindex = tkey_extract_bits(key, tp->pos, tp->bits);
1096 put_child(t, (struct tnode *)tp, cindex, (struct node *)l);
1097 } else {
1098 /* Case 3: n is a LEAF or a TNODE and the key doesn't match. */
1099 /*
1100 * Add a new tnode here
1101 * first tnode need some special handling
1102 */
1103
1104 if (tp)
1105 pos = tp->pos+tp->bits;
1106 else
1107 pos = 0;
1108
1109 if (n) {
1110 newpos = tkey_mismatch(key, pos, n->key);
1111 tn = tnode_new(n->key, newpos, 1);
1112 } else {
1113 newpos = 0;
1114 tn = tnode_new(key, newpos, 1); /* First tnode */
1115 }
1116
1117 if (!tn) {
1118 free_leaf_info(li);
1119 tnode_free((struct tnode *) l);
1120 *err = -ENOMEM;
1121 goto err;
1122 }
1123
1124 node_set_parent((struct node *)tn, tp);
1125
1126 missbit = tkey_extract_bits(key, newpos, 1);
1127 put_child(t, tn, missbit, (struct node *)l);
1128 put_child(t, tn, 1-missbit, n);
1129
1130 if (tp) {
1131 cindex = tkey_extract_bits(key, tp->pos, tp->bits);
1132 put_child(t, (struct tnode *)tp, cindex, (struct node *)tn);
1133 } else {
1134 rcu_assign_pointer(t->trie, (struct node *)tn); /* First tnode */
1135 tp = tn;
1136 }
1137 }
1138
1139 if (tp && tp->pos + tp->bits > 32)
1140 printk(KERN_WARNING "fib_trie tp=%p pos=%d, bits=%d, key=%0x plen=%d\n",
1141 tp, tp->pos, tp->bits, key, plen);
1142
1143 /* Rebalance the trie */
1144
1145 rcu_assign_pointer(t->trie, trie_rebalance(t, tp));
1146 done:
1147 t->revision++;
1148 err:
1149 return fa_head;
1150 }
1151
1152 /*
1153 * Caller must hold RTNL.
1154 */
1155 static int fn_trie_insert(struct fib_table *tb, struct fib_config *cfg)
1156 {
1157 struct trie *t = (struct trie *) tb->tb_data;
1158 struct fib_alias *fa, *new_fa;
1159 struct list_head *fa_head = NULL;
1160 struct fib_info *fi;
1161 int plen = cfg->fc_dst_len;
1162 u8 tos = cfg->fc_tos;
1163 u32 key, mask;
1164 int err;
1165 struct leaf *l;
1166
1167 if (plen > 32)
1168 return -EINVAL;
1169
1170 key = ntohl(cfg->fc_dst);
1171
1172 pr_debug("Insert table=%u %08x/%d\n", tb->tb_id, key, plen);
1173
1174 mask = ntohl(inet_make_mask(plen));
1175
1176 if (key & ~mask)
1177 return -EINVAL;
1178
1179 key = key & mask;
1180
1181 fi = fib_create_info(cfg);
1182 if (IS_ERR(fi)) {
1183 err = PTR_ERR(fi);
1184 goto err;
1185 }
1186
1187 l = fib_find_node(t, key);
1188 fa = NULL;
1189
1190 if (l) {
1191 fa_head = get_fa_head(l, plen);
1192 fa = fib_find_alias(fa_head, tos, fi->fib_priority);
1193 }
1194
1195 /* Now fa, if non-NULL, points to the first fib alias
1196 * with the same keys [prefix,tos,priority], if such key already
1197 * exists or to the node before which we will insert new one.
1198 *
1199 * If fa is NULL, we will need to allocate a new one and
1200 * insert to the head of f.
1201 *
1202 * If f is NULL, no fib node matched the destination key
1203 * and we need to allocate a new one of those as well.
1204 */
1205
1206 if (fa && fa->fa_info->fib_priority == fi->fib_priority) {
1207 struct fib_alias *fa_orig;
1208
1209 err = -EEXIST;
1210 if (cfg->fc_nlflags & NLM_F_EXCL)
1211 goto out;
1212
1213 if (cfg->fc_nlflags & NLM_F_REPLACE) {
1214 struct fib_info *fi_drop;
1215 u8 state;
1216
1217 if (fi->fib_treeref > 1)
1218 goto out;
1219
1220 err = -ENOBUFS;
1221 new_fa = kmem_cache_alloc(fn_alias_kmem, GFP_KERNEL);
1222 if (new_fa == NULL)
1223 goto out;
1224
1225 fi_drop = fa->fa_info;
1226 new_fa->fa_tos = fa->fa_tos;
1227 new_fa->fa_info = fi;
1228 new_fa->fa_type = cfg->fc_type;
1229 new_fa->fa_scope = cfg->fc_scope;
1230 state = fa->fa_state;
1231 new_fa->fa_state &= ~FA_S_ACCESSED;
1232
1233 list_replace_rcu(&fa->fa_list, &new_fa->fa_list);
1234 alias_free_mem_rcu(fa);
1235
1236 fib_release_info(fi_drop);
1237 if (state & FA_S_ACCESSED)
1238 rt_cache_flush(-1);
1239 rtmsg_fib(RTM_NEWROUTE, htonl(key), new_fa, plen,
1240 tb->tb_id, &cfg->fc_nlinfo, NLM_F_REPLACE);
1241
1242 goto succeeded;
1243 }
1244 /* Error if we find a perfect match which
1245 * uses the same scope, type, and nexthop
1246 * information.
1247 */
1248 fa_orig = fa;
1249 list_for_each_entry(fa, fa_orig->fa_list.prev, fa_list) {
1250 if (fa->fa_tos != tos)
1251 break;
1252 if (fa->fa_info->fib_priority != fi->fib_priority)
1253 break;
1254 if (fa->fa_type == cfg->fc_type &&
1255 fa->fa_scope == cfg->fc_scope &&
1256 fa->fa_info == fi) {
1257 goto out;
1258 }
1259 }
1260 if (!(cfg->fc_nlflags & NLM_F_APPEND))
1261 fa = fa_orig;
1262 }
1263 err = -ENOENT;
1264 if (!(cfg->fc_nlflags & NLM_F_CREATE))
1265 goto out;
1266
1267 err = -ENOBUFS;
1268 new_fa = kmem_cache_alloc(fn_alias_kmem, GFP_KERNEL);
1269 if (new_fa == NULL)
1270 goto out;
1271
1272 new_fa->fa_info = fi;
1273 new_fa->fa_tos = tos;
1274 new_fa->fa_type = cfg->fc_type;
1275 new_fa->fa_scope = cfg->fc_scope;
1276 new_fa->fa_state = 0;
1277 /*
1278 * Insert new entry to the list.
1279 */
1280
1281 if (!fa_head) {
1282 err = 0;
1283 fa_head = fib_insert_node(t, &err, key, plen);
1284 if (err)
1285 goto out_free_new_fa;
1286 }
1287
1288 list_add_tail_rcu(&new_fa->fa_list,
1289 (fa ? &fa->fa_list : fa_head));
1290
1291 rt_cache_flush(-1);
1292 rtmsg_fib(RTM_NEWROUTE, htonl(key), new_fa, plen, tb->tb_id,
1293 &cfg->fc_nlinfo, 0);
1294 succeeded:
1295 return 0;
1296
1297 out_free_new_fa:
1298 kmem_cache_free(fn_alias_kmem, new_fa);
1299 out:
1300 fib_release_info(fi);
1301 err:
1302 return err;
1303 }
1304
1305
1306 /* should be called with rcu_read_lock */
1307 static inline int check_leaf(struct trie *t, struct leaf *l,
1308 t_key key, int *plen, const struct flowi *flp,
1309 struct fib_result *res)
1310 {
1311 int err, i;
1312 __be32 mask;
1313 struct leaf_info *li;
1314 struct hlist_head *hhead = &l->list;
1315 struct hlist_node *node;
1316
1317 hlist_for_each_entry_rcu(li, node, hhead, hlist) {
1318 i = li->plen;
1319 mask = inet_make_mask(i);
1320 if (l->key != (key & ntohl(mask)))
1321 continue;
1322
1323 if ((err = fib_semantic_match(&li->falh, flp, res, htonl(l->key), mask, i)) <= 0) {
1324 *plen = i;
1325 #ifdef CONFIG_IP_FIB_TRIE_STATS
1326 t->stats.semantic_match_passed++;
1327 #endif
1328 return err;
1329 }
1330 #ifdef CONFIG_IP_FIB_TRIE_STATS
1331 t->stats.semantic_match_miss++;
1332 #endif
1333 }
1334 return 1;
1335 }
1336
1337 static int
1338 fn_trie_lookup(struct fib_table *tb, const struct flowi *flp, struct fib_result *res)
1339 {
1340 struct trie *t = (struct trie *) tb->tb_data;
1341 int plen, ret = 0;
1342 struct node *n;
1343 struct tnode *pn;
1344 int pos, bits;
1345 t_key key = ntohl(flp->fl4_dst);
1346 int chopped_off;
1347 t_key cindex = 0;
1348 int current_prefix_length = KEYLENGTH;
1349 struct tnode *cn;
1350 t_key node_prefix, key_prefix, pref_mismatch;
1351 int mp;
1352
1353 rcu_read_lock();
1354
1355 n = rcu_dereference(t->trie);
1356 if (!n)
1357 goto failed;
1358
1359 #ifdef CONFIG_IP_FIB_TRIE_STATS
1360 t->stats.gets++;
1361 #endif
1362
1363 /* Just a leaf? */
1364 if (IS_LEAF(n)) {
1365 if ((ret = check_leaf(t, (struct leaf *)n, key, &plen, flp, res)) <= 0)
1366 goto found;
1367 goto failed;
1368 }
1369 pn = (struct tnode *) n;
1370 chopped_off = 0;
1371
1372 while (pn) {
1373 pos = pn->pos;
1374 bits = pn->bits;
1375
1376 if (!chopped_off)
1377 cindex = tkey_extract_bits(mask_pfx(key, current_prefix_length),
1378 pos, bits);
1379
1380 n = tnode_get_child(pn, cindex);
1381
1382 if (n == NULL) {
1383 #ifdef CONFIG_IP_FIB_TRIE_STATS
1384 t->stats.null_node_hit++;
1385 #endif
1386 goto backtrace;
1387 }
1388
1389 if (IS_LEAF(n)) {
1390 if ((ret = check_leaf(t, (struct leaf *)n, key, &plen, flp, res)) <= 0)
1391 goto found;
1392 else
1393 goto backtrace;
1394 }
1395
1396 #define HL_OPTIMIZE
1397 #ifdef HL_OPTIMIZE
1398 cn = (struct tnode *)n;
1399
1400 /*
1401 * It's a tnode, and we can do some extra checks here if we
1402 * like, to avoid descending into a dead-end branch.
1403 * This tnode is in the parent's child array at index
1404 * key[p_pos..p_pos+p_bits] but potentially with some bits
1405 * chopped off, so in reality the index may be just a
1406 * subprefix, padded with zero at the end.
1407 * We can also take a look at any skipped bits in this
1408 * tnode - everything up to p_pos is supposed to be ok,
1409 * and the non-chopped bits of the index (se previous
1410 * paragraph) are also guaranteed ok, but the rest is
1411 * considered unknown.
1412 *
1413 * The skipped bits are key[pos+bits..cn->pos].
1414 */
1415
1416 /* If current_prefix_length < pos+bits, we are already doing
1417 * actual prefix matching, which means everything from
1418 * pos+(bits-chopped_off) onward must be zero along some
1419 * branch of this subtree - otherwise there is *no* valid
1420 * prefix present. Here we can only check the skipped
1421 * bits. Remember, since we have already indexed into the
1422 * parent's child array, we know that the bits we chopped of
1423 * *are* zero.
1424 */
1425
1426 /* NOTA BENE: CHECKING ONLY SKIPPED BITS FOR THE NEW NODE HERE */
1427
1428 if (current_prefix_length < pos+bits) {
1429 if (tkey_extract_bits(cn->key, current_prefix_length,
1430 cn->pos - current_prefix_length) != 0 ||
1431 !(cn->child[0]))
1432 goto backtrace;
1433 }
1434
1435 /*
1436 * If chopped_off=0, the index is fully validated and we
1437 * only need to look at the skipped bits for this, the new,
1438 * tnode. What we actually want to do is to find out if
1439 * these skipped bits match our key perfectly, or if we will
1440 * have to count on finding a matching prefix further down,
1441 * because if we do, we would like to have some way of
1442 * verifying the existence of such a prefix at this point.
1443 */
1444
1445 /* The only thing we can do at this point is to verify that
1446 * any such matching prefix can indeed be a prefix to our
1447 * key, and if the bits in the node we are inspecting that
1448 * do not match our key are not ZERO, this cannot be true.
1449 * Thus, find out where there is a mismatch (before cn->pos)
1450 * and verify that all the mismatching bits are zero in the
1451 * new tnode's key.
1452 */
1453
1454 /* Note: We aren't very concerned about the piece of the key
1455 * that precede pn->pos+pn->bits, since these have already been
1456 * checked. The bits after cn->pos aren't checked since these are
1457 * by definition "unknown" at this point. Thus, what we want to
1458 * see is if we are about to enter the "prefix matching" state,
1459 * and in that case verify that the skipped bits that will prevail
1460 * throughout this subtree are zero, as they have to be if we are
1461 * to find a matching prefix.
1462 */
1463
1464 node_prefix = mask_pfx(cn->key, cn->pos);
1465 key_prefix = mask_pfx(key, cn->pos);
1466 pref_mismatch = key_prefix^node_prefix;
1467 mp = 0;
1468
1469 /* In short: If skipped bits in this node do not match the search
1470 * key, enter the "prefix matching" state.directly.
1471 */
1472 if (pref_mismatch) {
1473 while (!(pref_mismatch & (1<<(KEYLENGTH-1)))) {
1474 mp++;
1475 pref_mismatch = pref_mismatch <<1;
1476 }
1477 key_prefix = tkey_extract_bits(cn->key, mp, cn->pos-mp);
1478
1479 if (key_prefix != 0)
1480 goto backtrace;
1481
1482 if (current_prefix_length >= cn->pos)
1483 current_prefix_length = mp;
1484 }
1485 #endif
1486 pn = (struct tnode *)n; /* Descend */
1487 chopped_off = 0;
1488 continue;
1489
1490 backtrace:
1491 chopped_off++;
1492
1493 /* As zero don't change the child key (cindex) */
1494 while ((chopped_off <= pn->bits) && !(cindex & (1<<(chopped_off-1))))
1495 chopped_off++;
1496
1497 /* Decrease current_... with bits chopped off */
1498 if (current_prefix_length > pn->pos + pn->bits - chopped_off)
1499 current_prefix_length = pn->pos + pn->bits - chopped_off;
1500
1501 /*
1502 * Either we do the actual chop off according or if we have
1503 * chopped off all bits in this tnode walk up to our parent.
1504 */
1505
1506 if (chopped_off <= pn->bits) {
1507 cindex &= ~(1 << (chopped_off-1));
1508 } else {
1509 struct tnode *parent = node_parent((struct node *) pn);
1510 if (!parent)
1511 goto failed;
1512
1513 /* Get Child's index */
1514 cindex = tkey_extract_bits(pn->key, parent->pos, parent->bits);
1515 pn = parent;
1516 chopped_off = 0;
1517
1518 #ifdef CONFIG_IP_FIB_TRIE_STATS
1519 t->stats.backtrack++;
1520 #endif
1521 goto backtrace;
1522 }
1523 }
1524 failed:
1525 ret = 1;
1526 found:
1527 rcu_read_unlock();
1528 return ret;
1529 }
1530
1531 /* only called from updater side */
1532 static int trie_leaf_remove(struct trie *t, t_key key)
1533 {
1534 t_key cindex;
1535 struct tnode *tp = NULL;
1536 struct node *n = t->trie;
1537 struct leaf *l;
1538
1539 pr_debug("entering trie_leaf_remove(%p)\n", n);
1540
1541 /* Note that in the case skipped bits, those bits are *not* checked!
1542 * When we finish this, we will have NULL or a T_LEAF, and the
1543 * T_LEAF may or may not match our key.
1544 */
1545
1546 while (n != NULL && IS_TNODE(n)) {
1547 struct tnode *tn = (struct tnode *) n;
1548 check_tnode(tn);
1549 n = tnode_get_child(tn ,tkey_extract_bits(key, tn->pos, tn->bits));
1550
1551 BUG_ON(n && node_parent(n) != tn);
1552 }
1553 l = (struct leaf *) n;
1554
1555 if (!n || !tkey_equals(l->key, key))
1556 return 0;
1557
1558 /*
1559 * Key found.
1560 * Remove the leaf and rebalance the tree
1561 */
1562
1563 t->revision++;
1564 t->size--;
1565
1566 tp = node_parent(n);
1567 tnode_free((struct tnode *) n);
1568
1569 if (tp) {
1570 cindex = tkey_extract_bits(key, tp->pos, tp->bits);
1571 put_child(t, (struct tnode *)tp, cindex, NULL);
1572 rcu_assign_pointer(t->trie, trie_rebalance(t, tp));
1573 } else
1574 rcu_assign_pointer(t->trie, NULL);
1575
1576 return 1;
1577 }
1578
1579 /*
1580 * Caller must hold RTNL.
1581 */
1582 static int fn_trie_delete(struct fib_table *tb, struct fib_config *cfg)
1583 {
1584 struct trie *t = (struct trie *) tb->tb_data;
1585 u32 key, mask;
1586 int plen = cfg->fc_dst_len;
1587 u8 tos = cfg->fc_tos;
1588 struct fib_alias *fa, *fa_to_delete;
1589 struct list_head *fa_head;
1590 struct leaf *l;
1591 struct leaf_info *li;
1592
1593 if (plen > 32)
1594 return -EINVAL;
1595
1596 key = ntohl(cfg->fc_dst);
1597 mask = ntohl(inet_make_mask(plen));
1598
1599 if (key & ~mask)
1600 return -EINVAL;
1601
1602 key = key & mask;
1603 l = fib_find_node(t, key);
1604
1605 if (!l)
1606 return -ESRCH;
1607
1608 fa_head = get_fa_head(l, plen);
1609 fa = fib_find_alias(fa_head, tos, 0);
1610
1611 if (!fa)
1612 return -ESRCH;
1613
1614 pr_debug("Deleting %08x/%d tos=%d t=%p\n", key, plen, tos, t);
1615
1616 fa_to_delete = NULL;
1617 fa_head = fa->fa_list.prev;
1618
1619 list_for_each_entry(fa, fa_head, fa_list) {
1620 struct fib_info *fi = fa->fa_info;
1621
1622 if (fa->fa_tos != tos)
1623 break;
1624
1625 if ((!cfg->fc_type || fa->fa_type == cfg->fc_type) &&
1626 (cfg->fc_scope == RT_SCOPE_NOWHERE ||
1627 fa->fa_scope == cfg->fc_scope) &&
1628 (!cfg->fc_protocol ||
1629 fi->fib_protocol == cfg->fc_protocol) &&
1630 fib_nh_match(cfg, fi) == 0) {
1631 fa_to_delete = fa;
1632 break;
1633 }
1634 }
1635
1636 if (!fa_to_delete)
1637 return -ESRCH;
1638
1639 fa = fa_to_delete;
1640 rtmsg_fib(RTM_DELROUTE, htonl(key), fa, plen, tb->tb_id,
1641 &cfg->fc_nlinfo, 0);
1642
1643 l = fib_find_node(t, key);
1644 li = find_leaf_info(l, plen);
1645
1646 list_del_rcu(&fa->fa_list);
1647
1648 if (list_empty(fa_head)) {
1649 hlist_del_rcu(&li->hlist);
1650 free_leaf_info(li);
1651 }
1652
1653 if (hlist_empty(&l->list))
1654 trie_leaf_remove(t, key);
1655
1656 if (fa->fa_state & FA_S_ACCESSED)
1657 rt_cache_flush(-1);
1658
1659 fib_release_info(fa->fa_info);
1660 alias_free_mem_rcu(fa);
1661 return 0;
1662 }
1663
1664 static int trie_flush_list(struct trie *t, struct list_head *head)
1665 {
1666 struct fib_alias *fa, *fa_node;
1667 int found = 0;
1668
1669 list_for_each_entry_safe(fa, fa_node, head, fa_list) {
1670 struct fib_info *fi = fa->fa_info;
1671
1672 if (fi && (fi->fib_flags & RTNH_F_DEAD)) {
1673 list_del_rcu(&fa->fa_list);
1674 fib_release_info(fa->fa_info);
1675 alias_free_mem_rcu(fa);
1676 found++;
1677 }
1678 }
1679 return found;
1680 }
1681
1682 static int trie_flush_leaf(struct trie *t, struct leaf *l)
1683 {
1684 int found = 0;
1685 struct hlist_head *lih = &l->list;
1686 struct hlist_node *node, *tmp;
1687 struct leaf_info *li = NULL;
1688
1689 hlist_for_each_entry_safe(li, node, tmp, lih, hlist) {
1690 found += trie_flush_list(t, &li->falh);
1691
1692 if (list_empty(&li->falh)) {
1693 hlist_del_rcu(&li->hlist);
1694 free_leaf_info(li);
1695 }
1696 }
1697 return found;
1698 }
1699
1700 /* rcu_read_lock needs to be hold by caller from readside */
1701
1702 static struct leaf *nextleaf(struct trie *t, struct leaf *thisleaf)
1703 {
1704 struct node *c = (struct node *) thisleaf;
1705 struct tnode *p;
1706 int idx;
1707 struct node *trie = rcu_dereference(t->trie);
1708
1709 if (c == NULL) {
1710 if (trie == NULL)
1711 return NULL;
1712
1713 if (IS_LEAF(trie)) /* trie w. just a leaf */
1714 return (struct leaf *) trie;
1715
1716 p = (struct tnode*) trie; /* Start */
1717 } else
1718 p = node_parent(c);
1719
1720 while (p) {
1721 int pos, last;
1722
1723 /* Find the next child of the parent */
1724 if (c)
1725 pos = 1 + tkey_extract_bits(c->key, p->pos, p->bits);
1726 else
1727 pos = 0;
1728
1729 last = 1 << p->bits;
1730 for (idx = pos; idx < last ; idx++) {
1731 c = rcu_dereference(p->child[idx]);
1732
1733 if (!c)
1734 continue;
1735
1736 /* Decend if tnode */
1737 while (IS_TNODE(c)) {
1738 p = (struct tnode *) c;
1739 idx = 0;
1740
1741 /* Rightmost non-NULL branch */
1742 if (p && IS_TNODE(p))
1743 while (!(c = rcu_dereference(p->child[idx]))
1744 && idx < (1<<p->bits)) idx++;
1745
1746 /* Done with this tnode? */
1747 if (idx >= (1 << p->bits) || !c)
1748 goto up;
1749 }
1750 return (struct leaf *) c;
1751 }
1752 up:
1753 /* No more children go up one step */
1754 c = (struct node *) p;
1755 p = node_parent(c);
1756 }
1757 return NULL; /* Ready. Root of trie */
1758 }
1759
1760 /*
1761 * Caller must hold RTNL.
1762 */
1763 static int fn_trie_flush(struct fib_table *tb)
1764 {
1765 struct trie *t = (struct trie *) tb->tb_data;
1766 struct leaf *ll = NULL, *l = NULL;
1767 int found = 0, h;
1768
1769 t->revision++;
1770
1771 for (h = 0; (l = nextleaf(t, l)) != NULL; h++) {
1772 found += trie_flush_leaf(t, l);
1773
1774 if (ll && hlist_empty(&ll->list))
1775 trie_leaf_remove(t, ll->key);
1776 ll = l;
1777 }
1778
1779 if (ll && hlist_empty(&ll->list))
1780 trie_leaf_remove(t, ll->key);
1781
1782 pr_debug("trie_flush found=%d\n", found);
1783 return found;
1784 }
1785
1786 static int trie_last_dflt = -1;
1787
1788 static void
1789 fn_trie_select_default(struct fib_table *tb, const struct flowi *flp, struct fib_result *res)
1790 {
1791 struct trie *t = (struct trie *) tb->tb_data;
1792 int order, last_idx;
1793 struct fib_info *fi = NULL;
1794 struct fib_info *last_resort;
1795 struct fib_alias *fa = NULL;
1796 struct list_head *fa_head;
1797 struct leaf *l;
1798
1799 last_idx = -1;
1800 last_resort = NULL;
1801 order = -1;
1802
1803 rcu_read_lock();
1804
1805 l = fib_find_node(t, 0);
1806 if (!l)
1807 goto out;
1808
1809 fa_head = get_fa_head(l, 0);
1810 if (!fa_head)
1811 goto out;
1812
1813 if (list_empty(fa_head))
1814 goto out;
1815
1816 list_for_each_entry_rcu(fa, fa_head, fa_list) {
1817 struct fib_info *next_fi = fa->fa_info;
1818
1819 if (fa->fa_scope != res->scope ||
1820 fa->fa_type != RTN_UNICAST)
1821 continue;
1822
1823 if (next_fi->fib_priority > res->fi->fib_priority)
1824 break;
1825 if (!next_fi->fib_nh[0].nh_gw ||
1826 next_fi->fib_nh[0].nh_scope != RT_SCOPE_LINK)
1827 continue;
1828 fa->fa_state |= FA_S_ACCESSED;
1829
1830 if (fi == NULL) {
1831 if (next_fi != res->fi)
1832 break;
1833 } else if (!fib_detect_death(fi, order, &last_resort,
1834 &last_idx, &trie_last_dflt)) {
1835 if (res->fi)
1836 fib_info_put(res->fi);
1837 res->fi = fi;
1838 atomic_inc(&fi->fib_clntref);
1839 trie_last_dflt = order;
1840 goto out;
1841 }
1842 fi = next_fi;
1843 order++;
1844 }
1845 if (order <= 0 || fi == NULL) {
1846 trie_last_dflt = -1;
1847 goto out;
1848 }
1849
1850 if (!fib_detect_death(fi, order, &last_resort, &last_idx, &trie_last_dflt)) {
1851 if (res->fi)
1852 fib_info_put(res->fi);
1853 res->fi = fi;
1854 atomic_inc(&fi->fib_clntref);
1855 trie_last_dflt = order;
1856 goto out;
1857 }
1858 if (last_idx >= 0) {
1859 if (res->fi)
1860 fib_info_put(res->fi);
1861 res->fi = last_resort;
1862 if (last_resort)
1863 atomic_inc(&last_resort->fib_clntref);
1864 }
1865 trie_last_dflt = last_idx;
1866 out:;
1867 rcu_read_unlock();
1868 }
1869
1870 static int fn_trie_dump_fa(t_key key, int plen, struct list_head *fah, struct fib_table *tb,
1871 struct sk_buff *skb, struct netlink_callback *cb)
1872 {
1873 int i, s_i;
1874 struct fib_alias *fa;
1875
1876 __be32 xkey = htonl(key);
1877
1878 s_i = cb->args[4];
1879 i = 0;
1880
1881 /* rcu_read_lock is hold by caller */
1882
1883 list_for_each_entry_rcu(fa, fah, fa_list) {
1884 if (i < s_i) {
1885 i++;
1886 continue;
1887 }
1888 BUG_ON(!fa->fa_info);
1889
1890 if (fib_dump_info(skb, NETLINK_CB(cb->skb).pid,
1891 cb->nlh->nlmsg_seq,
1892 RTM_NEWROUTE,
1893 tb->tb_id,
1894 fa->fa_type,
1895 fa->fa_scope,
1896 xkey,
1897 plen,
1898 fa->fa_tos,
1899 fa->fa_info, 0) < 0) {
1900 cb->args[4] = i;
1901 return -1;
1902 }
1903 i++;
1904 }
1905 cb->args[4] = i;
1906 return skb->len;
1907 }
1908
1909 static int fn_trie_dump_plen(struct trie *t, int plen, struct fib_table *tb, struct sk_buff *skb,
1910 struct netlink_callback *cb)
1911 {
1912 int h, s_h;
1913 struct list_head *fa_head;
1914 struct leaf *l = NULL;
1915
1916 s_h = cb->args[3];
1917
1918 for (h = 0; (l = nextleaf(t, l)) != NULL; h++) {
1919 if (h < s_h)
1920 continue;
1921 if (h > s_h)
1922 memset(&cb->args[4], 0,
1923 sizeof(cb->args) - 4*sizeof(cb->args[0]));
1924
1925 fa_head = get_fa_head(l, plen);
1926
1927 if (!fa_head)
1928 continue;
1929
1930 if (list_empty(fa_head))
1931 continue;
1932
1933 if (fn_trie_dump_fa(l->key, plen, fa_head, tb, skb, cb)<0) {
1934 cb->args[3] = h;
1935 return -1;
1936 }
1937 }
1938 cb->args[3] = h;
1939 return skb->len;
1940 }
1941
1942 static int fn_trie_dump(struct fib_table *tb, struct sk_buff *skb, struct netlink_callback *cb)
1943 {
1944 int m, s_m;
1945 struct trie *t = (struct trie *) tb->tb_data;
1946
1947 s_m = cb->args[2];
1948
1949 rcu_read_lock();
1950 for (m = 0; m <= 32; m++) {
1951 if (m < s_m)
1952 continue;
1953 if (m > s_m)
1954 memset(&cb->args[3], 0,
1955 sizeof(cb->args) - 3*sizeof(cb->args[0]));
1956
1957 if (fn_trie_dump_plen(t, 32-m, tb, skb, cb)<0) {
1958 cb->args[2] = m;
1959 goto out;
1960 }
1961 }
1962 rcu_read_unlock();
1963 cb->args[2] = m;
1964 return skb->len;
1965 out:
1966 rcu_read_unlock();
1967 return -1;
1968 }
1969
1970 /* Fix more generic FIB names for init later */
1971
1972 #ifdef CONFIG_IP_MULTIPLE_TABLES
1973 struct fib_table * fib_hash_init(u32 id)
1974 #else
1975 struct fib_table * __init fib_hash_init(u32 id)
1976 #endif
1977 {
1978 struct fib_table *tb;
1979 struct trie *t;
1980
1981 if (fn_alias_kmem == NULL)
1982 fn_alias_kmem = kmem_cache_create("ip_fib_alias",
1983 sizeof(struct fib_alias),
1984 0, SLAB_HWCACHE_ALIGN,
1985 NULL);
1986
1987 tb = kmalloc(sizeof(struct fib_table) + sizeof(struct trie),
1988 GFP_KERNEL);
1989 if (tb == NULL)
1990 return NULL;
1991
1992 tb->tb_id = id;
1993 tb->tb_lookup = fn_trie_lookup;
1994 tb->tb_insert = fn_trie_insert;
1995 tb->tb_delete = fn_trie_delete;
1996 tb->tb_flush = fn_trie_flush;
1997 tb->tb_select_default = fn_trie_select_default;
1998 tb->tb_dump = fn_trie_dump;
1999 memset(tb->tb_data, 0, sizeof(struct trie));
2000
2001 t = (struct trie *) tb->tb_data;
2002
2003 trie_init(t);
2004
2005 if (id == RT_TABLE_LOCAL)
2006 trie_local = t;
2007 else if (id == RT_TABLE_MAIN)
2008 trie_main = t;
2009
2010 if (id == RT_TABLE_LOCAL)
2011 printk(KERN_INFO "IPv4 FIB: Using LC-trie version %s\n", VERSION);
2012
2013 return tb;
2014 }
2015
2016 #ifdef CONFIG_PROC_FS
2017 /* Depth first Trie walk iterator */
2018 struct fib_trie_iter {
2019 struct tnode *tnode;
2020 struct trie *trie;
2021 unsigned index;
2022 unsigned depth;
2023 };
2024
2025 static struct node *fib_trie_get_next(struct fib_trie_iter *iter)
2026 {
2027 struct tnode *tn = iter->tnode;
2028 unsigned cindex = iter->index;
2029 struct tnode *p;
2030
2031 /* A single entry routing table */
2032 if (!tn)
2033 return NULL;
2034
2035 pr_debug("get_next iter={node=%p index=%d depth=%d}\n",
2036 iter->tnode, iter->index, iter->depth);
2037 rescan:
2038 while (cindex < (1<<tn->bits)) {
2039 struct node *n = tnode_get_child(tn, cindex);
2040
2041 if (n) {
2042 if (IS_LEAF(n)) {
2043 iter->tnode = tn;
2044 iter->index = cindex + 1;
2045 } else {
2046 /* push down one level */
2047 iter->tnode = (struct tnode *) n;
2048 iter->index = 0;
2049 ++iter->depth;
2050 }
2051 return n;
2052 }
2053
2054 ++cindex;
2055 }
2056
2057 /* Current node exhausted, pop back up */
2058 p = node_parent((struct node *)tn);
2059 if (p) {
2060 cindex = tkey_extract_bits(tn->key, p->pos, p->bits)+1;
2061 tn = p;
2062 --iter->depth;
2063 goto rescan;
2064 }
2065
2066 /* got root? */
2067 return NULL;
2068 }
2069
2070 static struct node *fib_trie_get_first(struct fib_trie_iter *iter,
2071 struct trie *t)
2072 {
2073 struct node *n ;
2074
2075 if (!t)
2076 return NULL;
2077
2078 n = rcu_dereference(t->trie);
2079
2080 if (!iter)
2081 return NULL;
2082
2083 if (n) {
2084 if (IS_TNODE(n)) {
2085 iter->tnode = (struct tnode *) n;
2086 iter->trie = t;
2087 iter->index = 0;
2088 iter->depth = 1;
2089 } else {
2090 iter->tnode = NULL;
2091 iter->trie = t;
2092 iter->index = 0;
2093 iter->depth = 0;
2094 }
2095 return n;
2096 }
2097 return NULL;
2098 }
2099
2100 static void trie_collect_stats(struct trie *t, struct trie_stat *s)
2101 {
2102 struct node *n;
2103 struct fib_trie_iter iter;
2104
2105 memset(s, 0, sizeof(*s));
2106
2107 rcu_read_lock();
2108 for (n = fib_trie_get_first(&iter, t); n;
2109 n = fib_trie_get_next(&iter)) {
2110 if (IS_LEAF(n)) {
2111 s->leaves++;
2112 s->totdepth += iter.depth;
2113 if (iter.depth > s->maxdepth)
2114 s->maxdepth = iter.depth;
2115 } else {
2116 const struct tnode *tn = (const struct tnode *) n;
2117 int i;
2118
2119 s->tnodes++;
2120 if (tn->bits < MAX_STAT_DEPTH)
2121 s->nodesizes[tn->bits]++;
2122
2123 for (i = 0; i < (1<<tn->bits); i++)
2124 if (!tn->child[i])
2125 s->nullpointers++;
2126 }
2127 }
2128 rcu_read_unlock();
2129 }
2130
2131 /*
2132 * This outputs /proc/net/fib_triestats
2133 */
2134 static void trie_show_stats(struct seq_file *seq, struct trie_stat *stat)
2135 {
2136 unsigned i, max, pointers, bytes, avdepth;
2137
2138 if (stat->leaves)
2139 avdepth = stat->totdepth*100 / stat->leaves;
2140 else
2141 avdepth = 0;
2142
2143 seq_printf(seq, "\tAver depth: %d.%02d\n", avdepth / 100, avdepth % 100 );
2144 seq_printf(seq, "\tMax depth: %u\n", stat->maxdepth);
2145
2146 seq_printf(seq, "\tLeaves: %u\n", stat->leaves);
2147
2148 bytes = sizeof(struct leaf) * stat->leaves;
2149 seq_printf(seq, "\tInternal nodes: %d\n\t", stat->tnodes);
2150 bytes += sizeof(struct tnode) * stat->tnodes;
2151
2152 max = MAX_STAT_DEPTH;
2153 while (max > 0 && stat->nodesizes[max-1] == 0)
2154 max--;
2155
2156 pointers = 0;
2157 for (i = 1; i <= max; i++)
2158 if (stat->nodesizes[i] != 0) {
2159 seq_printf(seq, " %d: %d", i, stat->nodesizes[i]);
2160 pointers += (1<<i) * stat->nodesizes[i];
2161 }
2162 seq_putc(seq, '\n');
2163 seq_printf(seq, "\tPointers: %d\n", pointers);
2164
2165 bytes += sizeof(struct node *) * pointers;
2166 seq_printf(seq, "Null ptrs: %d\n", stat->nullpointers);
2167 seq_printf(seq, "Total size: %d kB\n", (bytes + 1023) / 1024);
2168
2169 #ifdef CONFIG_IP_FIB_TRIE_STATS
2170 seq_printf(seq, "Counters:\n---------\n");
2171 seq_printf(seq,"gets = %d\n", t->stats.gets);
2172 seq_printf(seq,"backtracks = %d\n", t->stats.backtrack);
2173 seq_printf(seq,"semantic match passed = %d\n", t->stats.semantic_match_passed);
2174 seq_printf(seq,"semantic match miss = %d\n", t->stats.semantic_match_miss);
2175 seq_printf(seq,"null node hit= %d\n", t->stats.null_node_hit);
2176 seq_printf(seq,"skipped node resize = %d\n", t->stats.resize_node_skipped);
2177 #ifdef CLEAR_STATS
2178 memset(&(t->stats), 0, sizeof(t->stats));
2179 #endif
2180 #endif /* CONFIG_IP_FIB_TRIE_STATS */
2181 }
2182
2183 static int fib_triestat_seq_show(struct seq_file *seq, void *v)
2184 {
2185 struct trie_stat *stat;
2186
2187 stat = kmalloc(sizeof(*stat), GFP_KERNEL);
2188 if (!stat)
2189 return -ENOMEM;
2190
2191 seq_printf(seq, "Basic info: size of leaf: %Zd bytes, size of tnode: %Zd bytes.\n",
2192 sizeof(struct leaf), sizeof(struct tnode));
2193
2194 if (trie_local) {
2195 seq_printf(seq, "Local:\n");
2196 trie_collect_stats(trie_local, stat);
2197 trie_show_stats(seq, stat);
2198 }
2199
2200 if (trie_main) {
2201 seq_printf(seq, "Main:\n");
2202 trie_collect_stats(trie_main, stat);
2203 trie_show_stats(seq, stat);
2204 }
2205 kfree(stat);
2206
2207 return 0;
2208 }
2209
2210 static int fib_triestat_seq_open(struct inode *inode, struct file *file)
2211 {
2212 return single_open(file, fib_triestat_seq_show, NULL);
2213 }
2214
2215 static const struct file_operations fib_triestat_fops = {
2216 .owner = THIS_MODULE,
2217 .open = fib_triestat_seq_open,
2218 .read = seq_read,
2219 .llseek = seq_lseek,
2220 .release = single_release,
2221 };
2222
2223 static struct node *fib_trie_get_idx(struct fib_trie_iter *iter,
2224 loff_t pos)
2225 {
2226 loff_t idx = 0;
2227 struct node *n;
2228
2229 for (n = fib_trie_get_first(iter, trie_local);
2230 n; ++idx, n = fib_trie_get_next(iter)) {
2231 if (pos == idx)
2232 return n;
2233 }
2234
2235 for (n = fib_trie_get_first(iter, trie_main);
2236 n; ++idx, n = fib_trie_get_next(iter)) {
2237 if (pos == idx)
2238 return n;
2239 }
2240 return NULL;
2241 }
2242
2243 static void *fib_trie_seq_start(struct seq_file *seq, loff_t *pos)
2244 {
2245 rcu_read_lock();
2246 if (*pos == 0)
2247 return SEQ_START_TOKEN;
2248 return fib_trie_get_idx(seq->private, *pos - 1);
2249 }
2250
2251 static void *fib_trie_seq_next(struct seq_file *seq, void *v, loff_t *pos)
2252 {
2253 struct fib_trie_iter *iter = seq->private;
2254 void *l = v;
2255
2256 ++*pos;
2257 if (v == SEQ_START_TOKEN)
2258 return fib_trie_get_idx(iter, 0);
2259
2260 v = fib_trie_get_next(iter);
2261 BUG_ON(v == l);
2262 if (v)
2263 return v;
2264
2265 /* continue scan in next trie */
2266 if (iter->trie == trie_local)
2267 return fib_trie_get_first(iter, trie_main);
2268
2269 return NULL;
2270 }
2271
2272 static void fib_trie_seq_stop(struct seq_file *seq, void *v)
2273 {
2274 rcu_read_unlock();
2275 }
2276
2277 static void seq_indent(struct seq_file *seq, int n)
2278 {
2279 while (n-- > 0) seq_puts(seq, " ");
2280 }
2281
2282 static inline const char *rtn_scope(enum rt_scope_t s)
2283 {
2284 static char buf[32];
2285
2286 switch (s) {
2287 case RT_SCOPE_UNIVERSE: return "universe";
2288 case RT_SCOPE_SITE: return "site";
2289 case RT_SCOPE_LINK: return "link";
2290 case RT_SCOPE_HOST: return "host";
2291 case RT_SCOPE_NOWHERE: return "nowhere";
2292 default:
2293 snprintf(buf, sizeof(buf), "scope=%d", s);
2294 return buf;
2295 }
2296 }
2297
2298 static const char *rtn_type_names[__RTN_MAX] = {
2299 [RTN_UNSPEC] = "UNSPEC",
2300 [RTN_UNICAST] = "UNICAST",
2301 [RTN_LOCAL] = "LOCAL",
2302 [RTN_BROADCAST] = "BROADCAST",
2303 [RTN_ANYCAST] = "ANYCAST",
2304 [RTN_MULTICAST] = "MULTICAST",
2305 [RTN_BLACKHOLE] = "BLACKHOLE",
2306 [RTN_UNREACHABLE] = "UNREACHABLE",
2307 [RTN_PROHIBIT] = "PROHIBIT",
2308 [RTN_THROW] = "THROW",
2309 [RTN_NAT] = "NAT",
2310 [RTN_XRESOLVE] = "XRESOLVE",
2311 };
2312
2313 static inline const char *rtn_type(unsigned t)
2314 {
2315 static char buf[32];
2316
2317 if (t < __RTN_MAX && rtn_type_names[t])
2318 return rtn_type_names[t];
2319 snprintf(buf, sizeof(buf), "type %d", t);
2320 return buf;
2321 }
2322
2323 /* Pretty print the trie */
2324 static int fib_trie_seq_show(struct seq_file *seq, void *v)
2325 {
2326 const struct fib_trie_iter *iter = seq->private;
2327 struct node *n = v;
2328
2329 if (v == SEQ_START_TOKEN)
2330 return 0;
2331
2332 if (!node_parent(n)) {
2333 if (iter->trie == trie_local)
2334 seq_puts(seq, "<local>:\n");
2335 else
2336 seq_puts(seq, "<main>:\n");
2337 }
2338
2339 if (IS_TNODE(n)) {
2340 struct tnode *tn = (struct tnode *) n;
2341 __be32 prf = htonl(mask_pfx(tn->key, tn->pos));
2342
2343 seq_indent(seq, iter->depth-1);
2344 seq_printf(seq, " +-- %d.%d.%d.%d/%d %d %d %d\n",
2345 NIPQUAD(prf), tn->pos, tn->bits, tn->full_children,
2346 tn->empty_children);
2347
2348 } else {
2349 struct leaf *l = (struct leaf *) n;
2350 int i;
2351 __be32 val = htonl(l->key);
2352
2353 seq_indent(seq, iter->depth);
2354 seq_printf(seq, " |-- %d.%d.%d.%d\n", NIPQUAD(val));
2355 for (i = 32; i >= 0; i--) {
2356 struct leaf_info *li = find_leaf_info(l, i);
2357 if (li) {
2358 struct fib_alias *fa;
2359 list_for_each_entry_rcu(fa, &li->falh, fa_list) {
2360 seq_indent(seq, iter->depth+1);
2361 seq_printf(seq, " /%d %s %s", i,
2362 rtn_scope(fa->fa_scope),
2363 rtn_type(fa->fa_type));
2364 if (fa->fa_tos)
2365 seq_printf(seq, "tos =%d\n",
2366 fa->fa_tos);
2367 seq_putc(seq, '\n');
2368 }
2369 }
2370 }
2371 }
2372
2373 return 0;
2374 }
2375
2376 static const struct seq_operations fib_trie_seq_ops = {
2377 .start = fib_trie_seq_start,
2378 .next = fib_trie_seq_next,
2379 .stop = fib_trie_seq_stop,
2380 .show = fib_trie_seq_show,
2381 };
2382
2383 static int fib_trie_seq_open(struct inode *inode, struct file *file)
2384 {
2385 return seq_open_private(file, &fib_trie_seq_ops,
2386 sizeof(struct fib_trie_iter));
2387 }
2388
2389 static const struct file_operations fib_trie_fops = {
2390 .owner = THIS_MODULE,
2391 .open = fib_trie_seq_open,
2392 .read = seq_read,
2393 .llseek = seq_lseek,
2394 .release = seq_release_private,
2395 };
2396
2397 static unsigned fib_flag_trans(int type, __be32 mask, const struct fib_info *fi)
2398 {
2399 static unsigned type2flags[RTN_MAX + 1] = {
2400 [7] = RTF_REJECT, [8] = RTF_REJECT,
2401 };
2402 unsigned flags = type2flags[type];
2403
2404 if (fi && fi->fib_nh->nh_gw)
2405 flags |= RTF_GATEWAY;
2406 if (mask == htonl(0xFFFFFFFF))
2407 flags |= RTF_HOST;
2408 flags |= RTF_UP;
2409 return flags;
2410 }
2411
2412 /*
2413 * This outputs /proc/net/route.
2414 * The format of the file is not supposed to be changed
2415 * and needs to be same as fib_hash output to avoid breaking
2416 * legacy utilities
2417 */
2418 static int fib_route_seq_show(struct seq_file *seq, void *v)
2419 {
2420 const struct fib_trie_iter *iter = seq->private;
2421 struct leaf *l = v;
2422 int i;
2423 char bf[128];
2424
2425 if (v == SEQ_START_TOKEN) {
2426 seq_printf(seq, "%-127s\n", "Iface\tDestination\tGateway "
2427 "\tFlags\tRefCnt\tUse\tMetric\tMask\t\tMTU"
2428 "\tWindow\tIRTT");
2429 return 0;
2430 }
2431
2432 if (iter->trie == trie_local)
2433 return 0;
2434 if (IS_TNODE(l))
2435 return 0;
2436
2437 for (i=32; i>=0; i--) {
2438 struct leaf_info *li = find_leaf_info(l, i);
2439 struct fib_alias *fa;
2440 __be32 mask, prefix;
2441
2442 if (!li)
2443 continue;
2444
2445 mask = inet_make_mask(li->plen);
2446 prefix = htonl(l->key);
2447
2448 list_for_each_entry_rcu(fa, &li->falh, fa_list) {
2449 const struct fib_info *fi = fa->fa_info;
2450 unsigned flags = fib_flag_trans(fa->fa_type, mask, fi);
2451
2452 if (fa->fa_type == RTN_BROADCAST
2453 || fa->fa_type == RTN_MULTICAST)
2454 continue;
2455
2456 if (fi)
2457 snprintf(bf, sizeof(bf),
2458 "%s\t%08X\t%08X\t%04X\t%d\t%u\t%d\t%08X\t%d\t%u\t%u",
2459 fi->fib_dev ? fi->fib_dev->name : "*",
2460 prefix,
2461 fi->fib_nh->nh_gw, flags, 0, 0,
2462 fi->fib_priority,
2463 mask,
2464 (fi->fib_advmss ? fi->fib_advmss + 40 : 0),
2465 fi->fib_window,
2466 fi->fib_rtt >> 3);
2467 else
2468 snprintf(bf, sizeof(bf),
2469 "*\t%08X\t%08X\t%04X\t%d\t%u\t%d\t%08X\t%d\t%u\t%u",
2470 prefix, 0, flags, 0, 0, 0,
2471 mask, 0, 0, 0);
2472
2473 seq_printf(seq, "%-127s\n", bf);
2474 }
2475 }
2476
2477 return 0;
2478 }
2479
2480 static const struct seq_operations fib_route_seq_ops = {
2481 .start = fib_trie_seq_start,
2482 .next = fib_trie_seq_next,
2483 .stop = fib_trie_seq_stop,
2484 .show = fib_route_seq_show,
2485 };
2486
2487 static int fib_route_seq_open(struct inode *inode, struct file *file)
2488 {
2489 return seq_open_private(file, &fib_route_seq_ops,
2490 sizeof(struct fib_trie_iter));
2491 }
2492
2493 static const struct file_operations fib_route_fops = {
2494 .owner = THIS_MODULE,
2495 .open = fib_route_seq_open,
2496 .read = seq_read,
2497 .llseek = seq_lseek,
2498 .release = seq_release_private,
2499 };
2500
2501 int __init fib_proc_init(void)
2502 {
2503 if (!proc_net_fops_create(&init_net, "fib_trie", S_IRUGO, &fib_trie_fops))
2504 goto out1;
2505
2506 if (!proc_net_fops_create(&init_net, "fib_triestat", S_IRUGO, &fib_triestat_fops))
2507 goto out2;
2508
2509 if (!proc_net_fops_create(&init_net, "route", S_IRUGO, &fib_route_fops))
2510 goto out3;
2511
2512 return 0;
2513
2514 out3:
2515 proc_net_remove(&init_net, "fib_triestat");
2516 out2:
2517 proc_net_remove(&init_net, "fib_trie");
2518 out1:
2519 return -ENOMEM;
2520 }
2521
2522 void __init fib_proc_exit(void)
2523 {
2524 proc_net_remove(&init_net, "fib_trie");
2525 proc_net_remove(&init_net, "fib_triestat");
2526 proc_net_remove(&init_net, "route");
2527 }
2528
2529 #endif /* CONFIG_PROC_FS */