Version:  2.0.40 2.2.26 2.4.37 3.5 3.6 3.7 3.8 3.9 3.10 3.11 3.12 3.13 3.14 3.15 3.16 3.17 3.18 3.19 4.0 4.1

Linux/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 described 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.csc.kth.se/~snilsson/software/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  *
 26  * Code from fib_hash has been reused which includes the following header:
 27  *
 28  *
 29  * INET         An implementation of the TCP/IP protocol suite for the LINUX
 30  *              operating system.  INET is implemented using the  BSD Socket
 31  *              interface as the means of communication with the user level.
 32  *
 33  *              IPv4 FIB: lookup engine and maintenance routines.
 34  *
 35  *
 36  * Authors:     Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
 37  *
 38  *              This program is free software; you can redistribute it and/or
 39  *              modify it under the terms of the GNU General Public License
 40  *              as published by the Free Software Foundation; either version
 41  *              2 of the License, or (at your option) any later version.
 42  *
 43  * Substantial contributions to this work comes from:
 44  *
 45  *              David S. Miller, <davem@davemloft.net>
 46  *              Stephen Hemminger <shemminger@osdl.org>
 47  *              Paul E. McKenney <paulmck@us.ibm.com>
 48  *              Patrick McHardy <kaber@trash.net>
 49  */
 50 
 51 #define VERSION "0.409"
 52 
 53 #include <asm/uaccess.h>
 54 #include <linux/bitops.h>
 55 #include <linux/types.h>
 56 #include <linux/kernel.h>
 57 #include <linux/mm.h>
 58 #include <linux/string.h>
 59 #include <linux/socket.h>
 60 #include <linux/sockios.h>
 61 #include <linux/errno.h>
 62 #include <linux/in.h>
 63 #include <linux/inet.h>
 64 #include <linux/inetdevice.h>
 65 #include <linux/netdevice.h>
 66 #include <linux/if_arp.h>
 67 #include <linux/proc_fs.h>
 68 #include <linux/rcupdate.h>
 69 #include <linux/skbuff.h>
 70 #include <linux/netlink.h>
 71 #include <linux/init.h>
 72 #include <linux/list.h>
 73 #include <linux/slab.h>
 74 #include <linux/export.h>
 75 #include <net/net_namespace.h>
 76 #include <net/ip.h>
 77 #include <net/protocol.h>
 78 #include <net/route.h>
 79 #include <net/tcp.h>
 80 #include <net/sock.h>
 81 #include <net/ip_fib.h>
 82 #include <net/switchdev.h>
 83 #include "fib_lookup.h"
 84 
 85 #define MAX_STAT_DEPTH 32
 86 
 87 #define KEYLENGTH       (8*sizeof(t_key))
 88 #define KEY_MAX         ((t_key)~0)
 89 
 90 typedef unsigned int t_key;
 91 
 92 #define IS_TRIE(n)      ((n)->pos >= KEYLENGTH)
 93 #define IS_TNODE(n)     ((n)->bits)
 94 #define IS_LEAF(n)      (!(n)->bits)
 95 
 96 struct key_vector {
 97         t_key key;
 98         unsigned char pos;              /* 2log(KEYLENGTH) bits needed */
 99         unsigned char bits;             /* 2log(KEYLENGTH) bits needed */
100         unsigned char slen;
101         union {
102                 /* This list pointer if valid if (pos | bits) == 0 (LEAF) */
103                 struct hlist_head leaf;
104                 /* This array is valid if (pos | bits) > 0 (TNODE) */
105                 struct key_vector __rcu *tnode[0];
106         };
107 };
108 
109 struct tnode {
110         struct rcu_head rcu;
111         t_key empty_children;           /* KEYLENGTH bits needed */
112         t_key full_children;            /* KEYLENGTH bits needed */
113         struct key_vector __rcu *parent;
114         struct key_vector kv[1];
115 #define tn_bits kv[0].bits
116 };
117 
118 #define TNODE_SIZE(n)   offsetof(struct tnode, kv[0].tnode[n])
119 #define LEAF_SIZE       TNODE_SIZE(1)
120 
121 #ifdef CONFIG_IP_FIB_TRIE_STATS
122 struct trie_use_stats {
123         unsigned int gets;
124         unsigned int backtrack;
125         unsigned int semantic_match_passed;
126         unsigned int semantic_match_miss;
127         unsigned int null_node_hit;
128         unsigned int resize_node_skipped;
129 };
130 #endif
131 
132 struct trie_stat {
133         unsigned int totdepth;
134         unsigned int maxdepth;
135         unsigned int tnodes;
136         unsigned int leaves;
137         unsigned int nullpointers;
138         unsigned int prefixes;
139         unsigned int nodesizes[MAX_STAT_DEPTH];
140 };
141 
142 struct trie {
143         struct key_vector kv[1];
144 #ifdef CONFIG_IP_FIB_TRIE_STATS
145         struct trie_use_stats __percpu *stats;
146 #endif
147 };
148 
149 static struct key_vector *resize(struct trie *t, struct key_vector *tn);
150 static size_t tnode_free_size;
151 
152 /*
153  * synchronize_rcu after call_rcu for that many pages; it should be especially
154  * useful before resizing the root node with PREEMPT_NONE configs; the value was
155  * obtained experimentally, aiming to avoid visible slowdown.
156  */
157 static const int sync_pages = 128;
158 
159 static struct kmem_cache *fn_alias_kmem __read_mostly;
160 static struct kmem_cache *trie_leaf_kmem __read_mostly;
161 
162 static inline struct tnode *tn_info(struct key_vector *kv)
163 {
164         return container_of(kv, struct tnode, kv[0]);
165 }
166 
167 /* caller must hold RTNL */
168 #define node_parent(tn) rtnl_dereference(tn_info(tn)->parent)
169 #define get_child(tn, i) rtnl_dereference((tn)->tnode[i])
170 
171 /* caller must hold RCU read lock or RTNL */
172 #define node_parent_rcu(tn) rcu_dereference_rtnl(tn_info(tn)->parent)
173 #define get_child_rcu(tn, i) rcu_dereference_rtnl((tn)->tnode[i])
174 
175 /* wrapper for rcu_assign_pointer */
176 static inline void node_set_parent(struct key_vector *n, struct key_vector *tp)
177 {
178         if (n)
179                 rcu_assign_pointer(tn_info(n)->parent, tp);
180 }
181 
182 #define NODE_INIT_PARENT(n, p) RCU_INIT_POINTER(tn_info(n)->parent, p)
183 
184 /* This provides us with the number of children in this node, in the case of a
185  * leaf this will return 0 meaning none of the children are accessible.
186  */
187 static inline unsigned long child_length(const struct key_vector *tn)
188 {
189         return (1ul << tn->bits) & ~(1ul);
190 }
191 
192 #define get_cindex(key, kv) (((key) ^ (kv)->key) >> (kv)->pos)
193 
194 static inline unsigned long get_index(t_key key, struct key_vector *kv)
195 {
196         unsigned long index = key ^ kv->key;
197 
198         if ((BITS_PER_LONG <= KEYLENGTH) && (KEYLENGTH == kv->pos))
199                 return 0;
200 
201         return index >> kv->pos;
202 }
203 
204 /* To understand this stuff, an understanding of keys and all their bits is
205  * necessary. Every node in the trie has a key associated with it, but not
206  * all of the bits in that key are significant.
207  *
208  * Consider a node 'n' and its parent 'tp'.
209  *
210  * If n is a leaf, every bit in its key is significant. Its presence is
211  * necessitated by path compression, since during a tree traversal (when
212  * searching for a leaf - unless we are doing an insertion) we will completely
213  * ignore all skipped bits we encounter. Thus we need to verify, at the end of
214  * a potentially successful search, that we have indeed been walking the
215  * correct key path.
216  *
217  * Note that we can never "miss" the correct key in the tree if present by
218  * following the wrong path. Path compression ensures that segments of the key
219  * that are the same for all keys with a given prefix are skipped, but the
220  * skipped part *is* identical for each node in the subtrie below the skipped
221  * bit! trie_insert() in this implementation takes care of that.
222  *
223  * if n is an internal node - a 'tnode' here, the various parts of its key
224  * have many different meanings.
225  *
226  * Example:
227  * _________________________________________________________________
228  * | i | i | i | i | i | i | i | N | N | N | S | S | S | S | S | C |
229  * -----------------------------------------------------------------
230  *  31  30  29  28  27  26  25  24  23  22  21  20  19  18  17  16
231  *
232  * _________________________________________________________________
233  * | C | C | C | u | u | u | u | u | u | u | u | u | u | u | u | u |
234  * -----------------------------------------------------------------
235  *  15  14  13  12  11  10   9   8   7   6   5   4   3   2   1   0
236  *
237  * tp->pos = 22
238  * tp->bits = 3
239  * n->pos = 13
240  * n->bits = 4
241  *
242  * First, let's just ignore the bits that come before the parent tp, that is
243  * the bits from (tp->pos + tp->bits) to 31. They are *known* but at this
244  * point we do not use them for anything.
245  *
246  * The bits from (tp->pos) to (tp->pos + tp->bits - 1) - "N", above - are the
247  * index into the parent's child array. That is, they will be used to find
248  * 'n' among tp's children.
249  *
250  * The bits from (n->pos + n->bits) to (tn->pos - 1) - "S" - are skipped bits
251  * for the node n.
252  *
253  * All the bits we have seen so far are significant to the node n. The rest
254  * of the bits are really not needed or indeed known in n->key.
255  *
256  * The bits from (n->pos) to (n->pos + n->bits - 1) - "C" - are the index into
257  * n's child array, and will of course be different for each child.
258  *
259  * The rest of the bits, from 0 to (n->pos + n->bits), are completely unknown
260  * at this point.
261  */
262 
263 static const int halve_threshold = 25;
264 static const int inflate_threshold = 50;
265 static const int halve_threshold_root = 15;
266 static const int inflate_threshold_root = 30;
267 
268 static void __alias_free_mem(struct rcu_head *head)
269 {
270         struct fib_alias *fa = container_of(head, struct fib_alias, rcu);
271         kmem_cache_free(fn_alias_kmem, fa);
272 }
273 
274 static inline void alias_free_mem_rcu(struct fib_alias *fa)
275 {
276         call_rcu(&fa->rcu, __alias_free_mem);
277 }
278 
279 #define TNODE_KMALLOC_MAX \
280         ilog2((PAGE_SIZE - TNODE_SIZE(0)) / sizeof(struct key_vector *))
281 #define TNODE_VMALLOC_MAX \
282         ilog2((SIZE_MAX - TNODE_SIZE(0)) / sizeof(struct key_vector *))
283 
284 static void __node_free_rcu(struct rcu_head *head)
285 {
286         struct tnode *n = container_of(head, struct tnode, rcu);
287 
288         if (!n->tn_bits)
289                 kmem_cache_free(trie_leaf_kmem, n);
290         else if (n->tn_bits <= TNODE_KMALLOC_MAX)
291                 kfree(n);
292         else
293                 vfree(n);
294 }
295 
296 #define node_free(n) call_rcu(&tn_info(n)->rcu, __node_free_rcu)
297 
298 static struct tnode *tnode_alloc(int bits)
299 {
300         size_t size;
301 
302         /* verify bits is within bounds */
303         if (bits > TNODE_VMALLOC_MAX)
304                 return NULL;
305 
306         /* determine size and verify it is non-zero and didn't overflow */
307         size = TNODE_SIZE(1ul << bits);
308 
309         if (size <= PAGE_SIZE)
310                 return kzalloc(size, GFP_KERNEL);
311         else
312                 return vzalloc(size);
313 }
314 
315 static inline void empty_child_inc(struct key_vector *n)
316 {
317         ++tn_info(n)->empty_children ? : ++tn_info(n)->full_children;
318 }
319 
320 static inline void empty_child_dec(struct key_vector *n)
321 {
322         tn_info(n)->empty_children-- ? : tn_info(n)->full_children--;
323 }
324 
325 static struct key_vector *leaf_new(t_key key, struct fib_alias *fa)
326 {
327         struct tnode *kv = kmem_cache_alloc(trie_leaf_kmem, GFP_KERNEL);
328         struct key_vector *l = kv->kv;
329 
330         if (!kv)
331                 return NULL;
332 
333         /* initialize key vector */
334         l->key = key;
335         l->pos = 0;
336         l->bits = 0;
337         l->slen = fa->fa_slen;
338 
339         /* link leaf to fib alias */
340         INIT_HLIST_HEAD(&l->leaf);
341         hlist_add_head(&fa->fa_list, &l->leaf);
342 
343         return l;
344 }
345 
346 static struct key_vector *tnode_new(t_key key, int pos, int bits)
347 {
348         struct tnode *tnode = tnode_alloc(bits);
349         unsigned int shift = pos + bits;
350         struct key_vector *tn = tnode->kv;
351 
352         /* verify bits and pos their msb bits clear and values are valid */
353         BUG_ON(!bits || (shift > KEYLENGTH));
354 
355         pr_debug("AT %p s=%zu %zu\n", tnode, TNODE_SIZE(0),
356                  sizeof(struct key_vector *) << bits);
357 
358         if (!tnode)
359                 return NULL;
360 
361         if (bits == KEYLENGTH)
362                 tnode->full_children = 1;
363         else
364                 tnode->empty_children = 1ul << bits;
365 
366         tn->key = (shift < KEYLENGTH) ? (key >> shift) << shift : 0;
367         tn->pos = pos;
368         tn->bits = bits;
369         tn->slen = pos;
370 
371         return tn;
372 }
373 
374 /* Check whether a tnode 'n' is "full", i.e. it is an internal node
375  * and no bits are skipped. See discussion in dyntree paper p. 6
376  */
377 static inline int tnode_full(struct key_vector *tn, struct key_vector *n)
378 {
379         return n && ((n->pos + n->bits) == tn->pos) && IS_TNODE(n);
380 }
381 
382 /* Add a child at position i overwriting the old value.
383  * Update the value of full_children and empty_children.
384  */
385 static void put_child(struct key_vector *tn, unsigned long i,
386                       struct key_vector *n)
387 {
388         struct key_vector *chi = get_child(tn, i);
389         int isfull, wasfull;
390 
391         BUG_ON(i >= child_length(tn));
392 
393         /* update emptyChildren, overflow into fullChildren */
394         if (!n && chi)
395                 empty_child_inc(tn);
396         if (n && !chi)
397                 empty_child_dec(tn);
398 
399         /* update fullChildren */
400         wasfull = tnode_full(tn, chi);
401         isfull = tnode_full(tn, n);
402 
403         if (wasfull && !isfull)
404                 tn_info(tn)->full_children--;
405         else if (!wasfull && isfull)
406                 tn_info(tn)->full_children++;
407 
408         if (n && (tn->slen < n->slen))
409                 tn->slen = n->slen;
410 
411         rcu_assign_pointer(tn->tnode[i], n);
412 }
413 
414 static void update_children(struct key_vector *tn)
415 {
416         unsigned long i;
417 
418         /* update all of the child parent pointers */
419         for (i = child_length(tn); i;) {
420                 struct key_vector *inode = get_child(tn, --i);
421 
422                 if (!inode)
423                         continue;
424 
425                 /* Either update the children of a tnode that
426                  * already belongs to us or update the child
427                  * to point to ourselves.
428                  */
429                 if (node_parent(inode) == tn)
430                         update_children(inode);
431                 else
432                         node_set_parent(inode, tn);
433         }
434 }
435 
436 static inline void put_child_root(struct key_vector *tp, t_key key,
437                                   struct key_vector *n)
438 {
439         if (IS_TRIE(tp))
440                 rcu_assign_pointer(tp->tnode[0], n);
441         else
442                 put_child(tp, get_index(key, tp), n);
443 }
444 
445 static inline void tnode_free_init(struct key_vector *tn)
446 {
447         tn_info(tn)->rcu.next = NULL;
448 }
449 
450 static inline void tnode_free_append(struct key_vector *tn,
451                                      struct key_vector *n)
452 {
453         tn_info(n)->rcu.next = tn_info(tn)->rcu.next;
454         tn_info(tn)->rcu.next = &tn_info(n)->rcu;
455 }
456 
457 static void tnode_free(struct key_vector *tn)
458 {
459         struct callback_head *head = &tn_info(tn)->rcu;
460 
461         while (head) {
462                 head = head->next;
463                 tnode_free_size += TNODE_SIZE(1ul << tn->bits);
464                 node_free(tn);
465 
466                 tn = container_of(head, struct tnode, rcu)->kv;
467         }
468 
469         if (tnode_free_size >= PAGE_SIZE * sync_pages) {
470                 tnode_free_size = 0;
471                 synchronize_rcu();
472         }
473 }
474 
475 static struct key_vector *replace(struct trie *t,
476                                   struct key_vector *oldtnode,
477                                   struct key_vector *tn)
478 {
479         struct key_vector *tp = node_parent(oldtnode);
480         unsigned long i;
481 
482         /* setup the parent pointer out of and back into this node */
483         NODE_INIT_PARENT(tn, tp);
484         put_child_root(tp, tn->key, tn);
485 
486         /* update all of the child parent pointers */
487         update_children(tn);
488 
489         /* all pointers should be clean so we are done */
490         tnode_free(oldtnode);
491 
492         /* resize children now that oldtnode is freed */
493         for (i = child_length(tn); i;) {
494                 struct key_vector *inode = get_child(tn, --i);
495 
496                 /* resize child node */
497                 if (tnode_full(tn, inode))
498                         tn = resize(t, inode);
499         }
500 
501         return tp;
502 }
503 
504 static struct key_vector *inflate(struct trie *t,
505                                   struct key_vector *oldtnode)
506 {
507         struct key_vector *tn;
508         unsigned long i;
509         t_key m;
510 
511         pr_debug("In inflate\n");
512 
513         tn = tnode_new(oldtnode->key, oldtnode->pos - 1, oldtnode->bits + 1);
514         if (!tn)
515                 goto notnode;
516 
517         /* prepare oldtnode to be freed */
518         tnode_free_init(oldtnode);
519 
520         /* Assemble all of the pointers in our cluster, in this case that
521          * represents all of the pointers out of our allocated nodes that
522          * point to existing tnodes and the links between our allocated
523          * nodes.
524          */
525         for (i = child_length(oldtnode), m = 1u << tn->pos; i;) {
526                 struct key_vector *inode = get_child(oldtnode, --i);
527                 struct key_vector *node0, *node1;
528                 unsigned long j, k;
529 
530                 /* An empty child */
531                 if (!inode)
532                         continue;
533 
534                 /* A leaf or an internal node with skipped bits */
535                 if (!tnode_full(oldtnode, inode)) {
536                         put_child(tn, get_index(inode->key, tn), inode);
537                         continue;
538                 }
539 
540                 /* drop the node in the old tnode free list */
541                 tnode_free_append(oldtnode, inode);
542 
543                 /* An internal node with two children */
544                 if (inode->bits == 1) {
545                         put_child(tn, 2 * i + 1, get_child(inode, 1));
546                         put_child(tn, 2 * i, get_child(inode, 0));
547                         continue;
548                 }
549 
550                 /* We will replace this node 'inode' with two new
551                  * ones, 'node0' and 'node1', each with half of the
552                  * original children. The two new nodes will have
553                  * a position one bit further down the key and this
554                  * means that the "significant" part of their keys
555                  * (see the discussion near the top of this file)
556                  * will differ by one bit, which will be "" in
557                  * node0's key and "1" in node1's key. Since we are
558                  * moving the key position by one step, the bit that
559                  * we are moving away from - the bit at position
560                  * (tn->pos) - is the one that will differ between
561                  * node0 and node1. So... we synthesize that bit in the
562                  * two new keys.
563                  */
564                 node1 = tnode_new(inode->key | m, inode->pos, inode->bits - 1);
565                 if (!node1)
566                         goto nomem;
567                 node0 = tnode_new(inode->key, inode->pos, inode->bits - 1);
568 
569                 tnode_free_append(tn, node1);
570                 if (!node0)
571                         goto nomem;
572                 tnode_free_append(tn, node0);
573 
574                 /* populate child pointers in new nodes */
575                 for (k = child_length(inode), j = k / 2; j;) {
576                         put_child(node1, --j, get_child(inode, --k));
577                         put_child(node0, j, get_child(inode, j));
578                         put_child(node1, --j, get_child(inode, --k));
579                         put_child(node0, j, get_child(inode, j));
580                 }
581 
582                 /* link new nodes to parent */
583                 NODE_INIT_PARENT(node1, tn);
584                 NODE_INIT_PARENT(node0, tn);
585 
586                 /* link parent to nodes */
587                 put_child(tn, 2 * i + 1, node1);
588                 put_child(tn, 2 * i, node0);
589         }
590 
591         /* setup the parent pointers into and out of this node */
592         return replace(t, oldtnode, tn);
593 nomem:
594         /* all pointers should be clean so we are done */
595         tnode_free(tn);
596 notnode:
597         return NULL;
598 }
599 
600 static struct key_vector *halve(struct trie *t,
601                                 struct key_vector *oldtnode)
602 {
603         struct key_vector *tn;
604         unsigned long i;
605 
606         pr_debug("In halve\n");
607 
608         tn = tnode_new(oldtnode->key, oldtnode->pos + 1, oldtnode->bits - 1);
609         if (!tn)
610                 goto notnode;
611 
612         /* prepare oldtnode to be freed */
613         tnode_free_init(oldtnode);
614 
615         /* Assemble all of the pointers in our cluster, in this case that
616          * represents all of the pointers out of our allocated nodes that
617          * point to existing tnodes and the links between our allocated
618          * nodes.
619          */
620         for (i = child_length(oldtnode); i;) {
621                 struct key_vector *node1 = get_child(oldtnode, --i);
622                 struct key_vector *node0 = get_child(oldtnode, --i);
623                 struct key_vector *inode;
624 
625                 /* At least one of the children is empty */
626                 if (!node1 || !node0) {
627                         put_child(tn, i / 2, node1 ? : node0);
628                         continue;
629                 }
630 
631                 /* Two nonempty children */
632                 inode = tnode_new(node0->key, oldtnode->pos, 1);
633                 if (!inode)
634                         goto nomem;
635                 tnode_free_append(tn, inode);
636 
637                 /* initialize pointers out of node */
638                 put_child(inode, 1, node1);
639                 put_child(inode, 0, node0);
640                 NODE_INIT_PARENT(inode, tn);
641 
642                 /* link parent to node */
643                 put_child(tn, i / 2, inode);
644         }
645 
646         /* setup the parent pointers into and out of this node */
647         return replace(t, oldtnode, tn);
648 nomem:
649         /* all pointers should be clean so we are done */
650         tnode_free(tn);
651 notnode:
652         return NULL;
653 }
654 
655 static struct key_vector *collapse(struct trie *t,
656                                    struct key_vector *oldtnode)
657 {
658         struct key_vector *n, *tp;
659         unsigned long i;
660 
661         /* scan the tnode looking for that one child that might still exist */
662         for (n = NULL, i = child_length(oldtnode); !n && i;)
663                 n = get_child(oldtnode, --i);
664 
665         /* compress one level */
666         tp = node_parent(oldtnode);
667         put_child_root(tp, oldtnode->key, n);
668         node_set_parent(n, tp);
669 
670         /* drop dead node */
671         node_free(oldtnode);
672 
673         return tp;
674 }
675 
676 static unsigned char update_suffix(struct key_vector *tn)
677 {
678         unsigned char slen = tn->pos;
679         unsigned long stride, i;
680 
681         /* search though the list of children looking for nodes that might
682          * have a suffix greater than the one we currently have.  This is
683          * why we start with a stride of 2 since a stride of 1 would
684          * represent the nodes with suffix length equal to tn->pos
685          */
686         for (i = 0, stride = 0x2ul ; i < child_length(tn); i += stride) {
687                 struct key_vector *n = get_child(tn, i);
688 
689                 if (!n || (n->slen <= slen))
690                         continue;
691 
692                 /* update stride and slen based on new value */
693                 stride <<= (n->slen - slen);
694                 slen = n->slen;
695                 i &= ~(stride - 1);
696 
697                 /* if slen covers all but the last bit we can stop here
698                  * there will be nothing longer than that since only node
699                  * 0 and 1 << (bits - 1) could have that as their suffix
700                  * length.
701                  */
702                 if ((slen + 1) >= (tn->pos + tn->bits))
703                         break;
704         }
705 
706         tn->slen = slen;
707 
708         return slen;
709 }
710 
711 /* From "Implementing a dynamic compressed trie" by Stefan Nilsson of
712  * the Helsinki University of Technology and Matti Tikkanen of Nokia
713  * Telecommunications, page 6:
714  * "A node is doubled if the ratio of non-empty children to all
715  * children in the *doubled* node is at least 'high'."
716  *
717  * 'high' in this instance is the variable 'inflate_threshold'. It
718  * is expressed as a percentage, so we multiply it with
719  * child_length() and instead of multiplying by 2 (since the
720  * child array will be doubled by inflate()) and multiplying
721  * the left-hand side by 100 (to handle the percentage thing) we
722  * multiply the left-hand side by 50.
723  *
724  * The left-hand side may look a bit weird: child_length(tn)
725  * - tn->empty_children is of course the number of non-null children
726  * in the current node. tn->full_children is the number of "full"
727  * children, that is non-null tnodes with a skip value of 0.
728  * All of those will be doubled in the resulting inflated tnode, so
729  * we just count them one extra time here.
730  *
731  * A clearer way to write this would be:
732  *
733  * to_be_doubled = tn->full_children;
734  * not_to_be_doubled = child_length(tn) - tn->empty_children -
735  *     tn->full_children;
736  *
737  * new_child_length = child_length(tn) * 2;
738  *
739  * new_fill_factor = 100 * (not_to_be_doubled + 2*to_be_doubled) /
740  *      new_child_length;
741  * if (new_fill_factor >= inflate_threshold)
742  *
743  * ...and so on, tho it would mess up the while () loop.
744  *
745  * anyway,
746  * 100 * (not_to_be_doubled + 2*to_be_doubled) / new_child_length >=
747  *      inflate_threshold
748  *
749  * avoid a division:
750  * 100 * (not_to_be_doubled + 2*to_be_doubled) >=
751  *      inflate_threshold * new_child_length
752  *
753  * expand not_to_be_doubled and to_be_doubled, and shorten:
754  * 100 * (child_length(tn) - tn->empty_children +
755  *    tn->full_children) >= inflate_threshold * new_child_length
756  *
757  * expand new_child_length:
758  * 100 * (child_length(tn) - tn->empty_children +
759  *    tn->full_children) >=
760  *      inflate_threshold * child_length(tn) * 2
761  *
762  * shorten again:
763  * 50 * (tn->full_children + child_length(tn) -
764  *    tn->empty_children) >= inflate_threshold *
765  *    child_length(tn)
766  *
767  */
768 static inline bool should_inflate(struct key_vector *tp, struct key_vector *tn)
769 {
770         unsigned long used = child_length(tn);
771         unsigned long threshold = used;
772 
773         /* Keep root node larger */
774         threshold *= IS_TRIE(tp) ? inflate_threshold_root : inflate_threshold;
775         used -= tn_info(tn)->empty_children;
776         used += tn_info(tn)->full_children;
777 
778         /* if bits == KEYLENGTH then pos = 0, and will fail below */
779 
780         return (used > 1) && tn->pos && ((50 * used) >= threshold);
781 }
782 
783 static inline bool should_halve(struct key_vector *tp, struct key_vector *tn)
784 {
785         unsigned long used = child_length(tn);
786         unsigned long threshold = used;
787 
788         /* Keep root node larger */
789         threshold *= IS_TRIE(tp) ? halve_threshold_root : halve_threshold;
790         used -= tn_info(tn)->empty_children;
791 
792         /* if bits == KEYLENGTH then used = 100% on wrap, and will fail below */
793 
794         return (used > 1) && (tn->bits > 1) && ((100 * used) < threshold);
795 }
796 
797 static inline bool should_collapse(struct key_vector *tn)
798 {
799         unsigned long used = child_length(tn);
800 
801         used -= tn_info(tn)->empty_children;
802 
803         /* account for bits == KEYLENGTH case */
804         if ((tn->bits == KEYLENGTH) && tn_info(tn)->full_children)
805                 used -= KEY_MAX;
806 
807         /* One child or none, time to drop us from the trie */
808         return used < 2;
809 }
810 
811 #define MAX_WORK 10
812 static struct key_vector *resize(struct trie *t, struct key_vector *tn)
813 {
814 #ifdef CONFIG_IP_FIB_TRIE_STATS
815         struct trie_use_stats __percpu *stats = t->stats;
816 #endif
817         struct key_vector *tp = node_parent(tn);
818         unsigned long cindex = get_index(tn->key, tp);
819         int max_work = MAX_WORK;
820 
821         pr_debug("In tnode_resize %p inflate_threshold=%d threshold=%d\n",
822                  tn, inflate_threshold, halve_threshold);
823 
824         /* track the tnode via the pointer from the parent instead of
825          * doing it ourselves.  This way we can let RCU fully do its
826          * thing without us interfering
827          */
828         BUG_ON(tn != get_child(tp, cindex));
829 
830         /* Double as long as the resulting node has a number of
831          * nonempty nodes that are above the threshold.
832          */
833         while (should_inflate(tp, tn) && max_work) {
834                 tp = inflate(t, tn);
835                 if (!tp) {
836 #ifdef CONFIG_IP_FIB_TRIE_STATS
837                         this_cpu_inc(stats->resize_node_skipped);
838 #endif
839                         break;
840                 }
841 
842                 max_work--;
843                 tn = get_child(tp, cindex);
844         }
845 
846         /* update parent in case inflate failed */
847         tp = node_parent(tn);
848 
849         /* Return if at least one inflate is run */
850         if (max_work != MAX_WORK)
851                 return tp;
852 
853         /* Halve as long as the number of empty children in this
854          * node is above threshold.
855          */
856         while (should_halve(tp, tn) && max_work) {
857                 tp = halve(t, tn);
858                 if (!tp) {
859 #ifdef CONFIG_IP_FIB_TRIE_STATS
860                         this_cpu_inc(stats->resize_node_skipped);
861 #endif
862                         break;
863                 }
864 
865                 max_work--;
866                 tn = get_child(tp, cindex);
867         }
868 
869         /* Only one child remains */
870         if (should_collapse(tn))
871                 return collapse(t, tn);
872 
873         /* update parent in case halve failed */
874         tp = node_parent(tn);
875 
876         /* Return if at least one deflate was run */
877         if (max_work != MAX_WORK)
878                 return tp;
879 
880         /* push the suffix length to the parent node */
881         if (tn->slen > tn->pos) {
882                 unsigned char slen = update_suffix(tn);
883 
884                 if (slen > tp->slen)
885                         tp->slen = slen;
886         }
887 
888         return tp;
889 }
890 
891 static void leaf_pull_suffix(struct key_vector *tp, struct key_vector *l)
892 {
893         while ((tp->slen > tp->pos) && (tp->slen > l->slen)) {
894                 if (update_suffix(tp) > l->slen)
895                         break;
896                 tp = node_parent(tp);
897         }
898 }
899 
900 static void leaf_push_suffix(struct key_vector *tn, struct key_vector *l)
901 {
902         /* if this is a new leaf then tn will be NULL and we can sort
903          * out parent suffix lengths as a part of trie_rebalance
904          */
905         while (tn->slen < l->slen) {
906                 tn->slen = l->slen;
907                 tn = node_parent(tn);
908         }
909 }
910 
911 /* rcu_read_lock needs to be hold by caller from readside */
912 static struct key_vector *fib_find_node(struct trie *t,
913                                         struct key_vector **tp, u32 key)
914 {
915         struct key_vector *pn, *n = t->kv;
916         unsigned long index = 0;
917 
918         do {
919                 pn = n;
920                 n = get_child_rcu(n, index);
921 
922                 if (!n)
923                         break;
924 
925                 index = get_cindex(key, n);
926 
927                 /* This bit of code is a bit tricky but it combines multiple
928                  * checks into a single check.  The prefix consists of the
929                  * prefix plus zeros for the bits in the cindex. The index
930                  * is the difference between the key and this value.  From
931                  * this we can actually derive several pieces of data.
932                  *   if (index >= (1ul << bits))
933                  *     we have a mismatch in skip bits and failed
934                  *   else
935                  *     we know the value is cindex
936                  *
937                  * This check is safe even if bits == KEYLENGTH due to the
938                  * fact that we can only allocate a node with 32 bits if a
939                  * long is greater than 32 bits.
940                  */
941                 if (index >= (1ul << n->bits)) {
942                         n = NULL;
943                         break;
944                 }
945 
946                 /* keep searching until we find a perfect match leaf or NULL */
947         } while (IS_TNODE(n));
948 
949         *tp = pn;
950 
951         return n;
952 }
953 
954 /* Return the first fib alias matching TOS with
955  * priority less than or equal to PRIO.
956  */
957 static struct fib_alias *fib_find_alias(struct hlist_head *fah, u8 slen,
958                                         u8 tos, u32 prio, u32 tb_id)
959 {
960         struct fib_alias *fa;
961 
962         if (!fah)
963                 return NULL;
964 
965         hlist_for_each_entry(fa, fah, fa_list) {
966                 if (fa->fa_slen < slen)
967                         continue;
968                 if (fa->fa_slen != slen)
969                         break;
970                 if (fa->tb_id > tb_id)
971                         continue;
972                 if (fa->tb_id != tb_id)
973                         break;
974                 if (fa->fa_tos > tos)
975                         continue;
976                 if (fa->fa_info->fib_priority >= prio || fa->fa_tos < tos)
977                         return fa;
978         }
979 
980         return NULL;
981 }
982 
983 static void trie_rebalance(struct trie *t, struct key_vector *tn)
984 {
985         while (!IS_TRIE(tn))
986                 tn = resize(t, tn);
987 }
988 
989 static int fib_insert_node(struct trie *t, struct key_vector *tp,
990                            struct fib_alias *new, t_key key)
991 {
992         struct key_vector *n, *l;
993 
994         l = leaf_new(key, new);
995         if (!l)
996                 goto noleaf;
997 
998         /* retrieve child from parent node */
999         n = get_child(tp, get_index(key, tp));
1000 
1001         /* Case 2: n is a LEAF or a TNODE and the key doesn't match.
1002          *
1003          *  Add a new tnode here
1004          *  first tnode need some special handling
1005          *  leaves us in position for handling as case 3
1006          */
1007         if (n) {
1008                 struct key_vector *tn;
1009 
1010                 tn = tnode_new(key, __fls(key ^ n->key), 1);
1011                 if (!tn)
1012                         goto notnode;
1013 
1014                 /* initialize routes out of node */
1015                 NODE_INIT_PARENT(tn, tp);
1016                 put_child(tn, get_index(key, tn) ^ 1, n);
1017 
1018                 /* start adding routes into the node */
1019                 put_child_root(tp, key, tn);
1020                 node_set_parent(n, tn);
1021 
1022                 /* parent now has a NULL spot where the leaf can go */
1023                 tp = tn;
1024         }
1025 
1026         /* Case 3: n is NULL, and will just insert a new leaf */
1027         NODE_INIT_PARENT(l, tp);
1028         put_child_root(tp, key, l);
1029         trie_rebalance(t, tp);
1030 
1031         return 0;
1032 notnode:
1033         node_free(l);
1034 noleaf:
1035         return -ENOMEM;
1036 }
1037 
1038 static int fib_insert_alias(struct trie *t, struct key_vector *tp,
1039                             struct key_vector *l, struct fib_alias *new,
1040                             struct fib_alias *fa, t_key key)
1041 {
1042         if (!l)
1043                 return fib_insert_node(t, tp, new, key);
1044 
1045         if (fa) {
1046                 hlist_add_before_rcu(&new->fa_list, &fa->fa_list);
1047         } else {
1048                 struct fib_alias *last;
1049 
1050                 hlist_for_each_entry(last, &l->leaf, fa_list) {
1051                         if (new->fa_slen < last->fa_slen)
1052                                 break;
1053                         if ((new->fa_slen == last->fa_slen) &&
1054                             (new->tb_id > last->tb_id))
1055                                 break;
1056                         fa = last;
1057                 }
1058 
1059                 if (fa)
1060                         hlist_add_behind_rcu(&new->fa_list, &fa->fa_list);
1061                 else
1062                         hlist_add_head_rcu(&new->fa_list, &l->leaf);
1063         }
1064 
1065         /* if we added to the tail node then we need to update slen */
1066         if (l->slen < new->fa_slen) {
1067                 l->slen = new->fa_slen;
1068                 leaf_push_suffix(tp, l);
1069         }
1070 
1071         return 0;
1072 }
1073 
1074 /* Caller must hold RTNL. */
1075 int fib_table_insert(struct fib_table *tb, struct fib_config *cfg)
1076 {
1077         struct trie *t = (struct trie *)tb->tb_data;
1078         struct fib_alias *fa, *new_fa;
1079         struct key_vector *l, *tp;
1080         struct fib_info *fi;
1081         u8 plen = cfg->fc_dst_len;
1082         u8 slen = KEYLENGTH - plen;
1083         u8 tos = cfg->fc_tos;
1084         u32 key;
1085         int err;
1086 
1087         if (plen > KEYLENGTH)
1088                 return -EINVAL;
1089 
1090         key = ntohl(cfg->fc_dst);
1091 
1092         pr_debug("Insert table=%u %08x/%d\n", tb->tb_id, key, plen);
1093 
1094         if ((plen < KEYLENGTH) && (key << plen))
1095                 return -EINVAL;
1096 
1097         fi = fib_create_info(cfg);
1098         if (IS_ERR(fi)) {
1099                 err = PTR_ERR(fi);
1100                 goto err;
1101         }
1102 
1103         l = fib_find_node(t, &tp, key);
1104         fa = l ? fib_find_alias(&l->leaf, slen, tos, fi->fib_priority,
1105                                 tb->tb_id) : NULL;
1106 
1107         /* Now fa, if non-NULL, points to the first fib alias
1108          * with the same keys [prefix,tos,priority], if such key already
1109          * exists or to the node before which we will insert new one.
1110          *
1111          * If fa is NULL, we will need to allocate a new one and
1112          * insert to the tail of the section matching the suffix length
1113          * of the new alias.
1114          */
1115 
1116         if (fa && fa->fa_tos == tos &&
1117             fa->fa_info->fib_priority == fi->fib_priority) {
1118                 struct fib_alias *fa_first, *fa_match;
1119 
1120                 err = -EEXIST;
1121                 if (cfg->fc_nlflags & NLM_F_EXCL)
1122                         goto out;
1123 
1124                 /* We have 2 goals:
1125                  * 1. Find exact match for type, scope, fib_info to avoid
1126                  * duplicate routes
1127                  * 2. Find next 'fa' (or head), NLM_F_APPEND inserts before it
1128                  */
1129                 fa_match = NULL;
1130                 fa_first = fa;
1131                 hlist_for_each_entry_from(fa, fa_list) {
1132                         if ((fa->fa_slen != slen) ||
1133                             (fa->tb_id != tb->tb_id) ||
1134                             (fa->fa_tos != tos))
1135                                 break;
1136                         if (fa->fa_info->fib_priority != fi->fib_priority)
1137                                 break;
1138                         if (fa->fa_type == cfg->fc_type &&
1139                             fa->fa_info == fi) {
1140                                 fa_match = fa;
1141                                 break;
1142                         }
1143                 }
1144 
1145                 if (cfg->fc_nlflags & NLM_F_REPLACE) {
1146                         struct fib_info *fi_drop;
1147                         u8 state;
1148 
1149                         fa = fa_first;
1150                         if (fa_match) {
1151                                 if (fa == fa_match)
1152                                         err = 0;
1153                                 goto out;
1154                         }
1155                         err = -ENOBUFS;
1156                         new_fa = kmem_cache_alloc(fn_alias_kmem, GFP_KERNEL);
1157                         if (!new_fa)
1158                                 goto out;
1159 
1160                         fi_drop = fa->fa_info;
1161                         new_fa->fa_tos = fa->fa_tos;
1162                         new_fa->fa_info = fi;
1163                         new_fa->fa_type = cfg->fc_type;
1164                         state = fa->fa_state;
1165                         new_fa->fa_state = state & ~FA_S_ACCESSED;
1166                         new_fa->fa_slen = fa->fa_slen;
1167                         new_fa->tb_id = tb->tb_id;
1168 
1169                         err = netdev_switch_fib_ipv4_add(key, plen, fi,
1170                                                          new_fa->fa_tos,
1171                                                          cfg->fc_type,
1172                                                          cfg->fc_nlflags,
1173                                                          tb->tb_id);
1174                         if (err) {
1175                                 netdev_switch_fib_ipv4_abort(fi);
1176                                 kmem_cache_free(fn_alias_kmem, new_fa);
1177                                 goto out;
1178                         }
1179 
1180                         hlist_replace_rcu(&fa->fa_list, &new_fa->fa_list);
1181 
1182                         alias_free_mem_rcu(fa);
1183 
1184                         fib_release_info(fi_drop);
1185                         if (state & FA_S_ACCESSED)
1186                                 rt_cache_flush(cfg->fc_nlinfo.nl_net);
1187                         rtmsg_fib(RTM_NEWROUTE, htonl(key), new_fa, plen,
1188                                 tb->tb_id, &cfg->fc_nlinfo, NLM_F_REPLACE);
1189 
1190                         goto succeeded;
1191                 }
1192                 /* Error if we find a perfect match which
1193                  * uses the same scope, type, and nexthop
1194                  * information.
1195                  */
1196                 if (fa_match)
1197                         goto out;
1198 
1199                 if (!(cfg->fc_nlflags & NLM_F_APPEND))
1200                         fa = fa_first;
1201         }
1202         err = -ENOENT;
1203         if (!(cfg->fc_nlflags & NLM_F_CREATE))
1204                 goto out;
1205 
1206         err = -ENOBUFS;
1207         new_fa = kmem_cache_alloc(fn_alias_kmem, GFP_KERNEL);
1208         if (!new_fa)
1209                 goto out;
1210 
1211         new_fa->fa_info = fi;
1212         new_fa->fa_tos = tos;
1213         new_fa->fa_type = cfg->fc_type;
1214         new_fa->fa_state = 0;
1215         new_fa->fa_slen = slen;
1216         new_fa->tb_id = tb->tb_id;
1217 
1218         /* (Optionally) offload fib entry to switch hardware. */
1219         err = netdev_switch_fib_ipv4_add(key, plen, fi, tos,
1220                                          cfg->fc_type,
1221                                          cfg->fc_nlflags,
1222                                          tb->tb_id);
1223         if (err) {
1224                 netdev_switch_fib_ipv4_abort(fi);
1225                 goto out_free_new_fa;
1226         }
1227 
1228         /* Insert new entry to the list. */
1229         err = fib_insert_alias(t, tp, l, new_fa, fa, key);
1230         if (err)
1231                 goto out_sw_fib_del;
1232 
1233         if (!plen)
1234                 tb->tb_num_default++;
1235 
1236         rt_cache_flush(cfg->fc_nlinfo.nl_net);
1237         rtmsg_fib(RTM_NEWROUTE, htonl(key), new_fa, plen, new_fa->tb_id,
1238                   &cfg->fc_nlinfo, 0);
1239 succeeded:
1240         return 0;
1241 
1242 out_sw_fib_del:
1243         netdev_switch_fib_ipv4_del(key, plen, fi, tos, cfg->fc_type, tb->tb_id);
1244 out_free_new_fa:
1245         kmem_cache_free(fn_alias_kmem, new_fa);
1246 out:
1247         fib_release_info(fi);
1248 err:
1249         return err;
1250 }
1251 
1252 static inline t_key prefix_mismatch(t_key key, struct key_vector *n)
1253 {
1254         t_key prefix = n->key;
1255 
1256         return (key ^ prefix) & (prefix | -prefix);
1257 }
1258 
1259 /* should be called with rcu_read_lock */
1260 int fib_table_lookup(struct fib_table *tb, const struct flowi4 *flp,
1261                      struct fib_result *res, int fib_flags)
1262 {
1263         struct trie *t = (struct trie *) tb->tb_data;
1264 #ifdef CONFIG_IP_FIB_TRIE_STATS
1265         struct trie_use_stats __percpu *stats = t->stats;
1266 #endif
1267         const t_key key = ntohl(flp->daddr);
1268         struct key_vector *n, *pn;
1269         struct fib_alias *fa;
1270         unsigned long index;
1271         t_key cindex;
1272 
1273         pn = t->kv;
1274         cindex = 0;
1275 
1276         n = get_child_rcu(pn, cindex);
1277         if (!n)
1278                 return -EAGAIN;
1279 
1280 #ifdef CONFIG_IP_FIB_TRIE_STATS
1281         this_cpu_inc(stats->gets);
1282 #endif
1283 
1284         /* Step 1: Travel to the longest prefix match in the trie */
1285         for (;;) {
1286                 index = get_cindex(key, n);
1287 
1288                 /* This bit of code is a bit tricky but it combines multiple
1289                  * checks into a single check.  The prefix consists of the
1290                  * prefix plus zeros for the "bits" in the prefix. The index
1291                  * is the difference between the key and this value.  From
1292                  * this we can actually derive several pieces of data.
1293                  *   if (index >= (1ul << bits))
1294                  *     we have a mismatch in skip bits and failed
1295                  *   else
1296                  *     we know the value is cindex
1297                  *
1298                  * This check is safe even if bits == KEYLENGTH due to the
1299                  * fact that we can only allocate a node with 32 bits if a
1300                  * long is greater than 32 bits.
1301                  */
1302                 if (index >= (1ul << n->bits))
1303                         break;
1304 
1305                 /* we have found a leaf. Prefixes have already been compared */
1306                 if (IS_LEAF(n))
1307                         goto found;
1308 
1309                 /* only record pn and cindex if we are going to be chopping
1310                  * bits later.  Otherwise we are just wasting cycles.
1311                  */
1312                 if (n->slen > n->pos) {
1313                         pn = n;
1314                         cindex = index;
1315                 }
1316 
1317                 n = get_child_rcu(n, index);
1318                 if (unlikely(!n))
1319                         goto backtrace;
1320         }
1321 
1322         /* Step 2: Sort out leaves and begin backtracing for longest prefix */
1323         for (;;) {
1324                 /* record the pointer where our next node pointer is stored */
1325                 struct key_vector __rcu **cptr = n->tnode;
1326 
1327                 /* This test verifies that none of the bits that differ
1328                  * between the key and the prefix exist in the region of
1329                  * the lsb and higher in the prefix.
1330                  */
1331                 if (unlikely(prefix_mismatch(key, n)) || (n->slen == n->pos))
1332                         goto backtrace;
1333 
1334                 /* exit out and process leaf */
1335                 if (unlikely(IS_LEAF(n)))
1336                         break;
1337 
1338                 /* Don't bother recording parent info.  Since we are in
1339                  * prefix match mode we will have to come back to wherever
1340                  * we started this traversal anyway
1341                  */
1342 
1343                 while ((n = rcu_dereference(*cptr)) == NULL) {
1344 backtrace:
1345 #ifdef CONFIG_IP_FIB_TRIE_STATS
1346                         if (!n)
1347                                 this_cpu_inc(stats->null_node_hit);
1348 #endif
1349                         /* If we are at cindex 0 there are no more bits for
1350                          * us to strip at this level so we must ascend back
1351                          * up one level to see if there are any more bits to
1352                          * be stripped there.
1353                          */
1354                         while (!cindex) {
1355                                 t_key pkey = pn->key;
1356 
1357                                 /* If we don't have a parent then there is
1358                                  * nothing for us to do as we do not have any
1359                                  * further nodes to parse.
1360                                  */
1361                                 if (IS_TRIE(pn))
1362                                         return -EAGAIN;
1363 #ifdef CONFIG_IP_FIB_TRIE_STATS
1364                                 this_cpu_inc(stats->backtrack);
1365 #endif
1366                                 /* Get Child's index */
1367                                 pn = node_parent_rcu(pn);
1368                                 cindex = get_index(pkey, pn);
1369                         }
1370 
1371                         /* strip the least significant bit from the cindex */
1372                         cindex &= cindex - 1;
1373 
1374                         /* grab pointer for next child node */
1375                         cptr = &pn->tnode[cindex];
1376                 }
1377         }
1378 
1379 found:
1380         /* this line carries forward the xor from earlier in the function */
1381         index = key ^ n->key;
1382 
1383         /* Step 3: Process the leaf, if that fails fall back to backtracing */
1384         hlist_for_each_entry_rcu(fa, &n->leaf, fa_list) {
1385                 struct fib_info *fi = fa->fa_info;
1386                 int nhsel, err;
1387 
1388                 if ((index >= (1ul << fa->fa_slen)) &&
1389                     ((BITS_PER_LONG > KEYLENGTH) || (fa->fa_slen != KEYLENGTH)))
1390                         continue;
1391                 if (fa->fa_tos && fa->fa_tos != flp->flowi4_tos)
1392                         continue;
1393                 if (fi->fib_dead)
1394                         continue;
1395                 if (fa->fa_info->fib_scope < flp->flowi4_scope)
1396                         continue;
1397                 fib_alias_accessed(fa);
1398                 err = fib_props[fa->fa_type].error;
1399                 if (unlikely(err < 0)) {
1400 #ifdef CONFIG_IP_FIB_TRIE_STATS
1401                         this_cpu_inc(stats->semantic_match_passed);
1402 #endif
1403                         return err;
1404                 }
1405                 if (fi->fib_flags & RTNH_F_DEAD)
1406                         continue;
1407                 for (nhsel = 0; nhsel < fi->fib_nhs; nhsel++) {
1408                         const struct fib_nh *nh = &fi->fib_nh[nhsel];
1409 
1410                         if (nh->nh_flags & RTNH_F_DEAD)
1411                                 continue;
1412                         if (flp->flowi4_oif && flp->flowi4_oif != nh->nh_oif)
1413                                 continue;
1414 
1415                         if (!(fib_flags & FIB_LOOKUP_NOREF))
1416                                 atomic_inc(&fi->fib_clntref);
1417 
1418                         res->prefixlen = KEYLENGTH - fa->fa_slen;
1419                         res->nh_sel = nhsel;
1420                         res->type = fa->fa_type;
1421                         res->scope = fi->fib_scope;
1422                         res->fi = fi;
1423                         res->table = tb;
1424                         res->fa_head = &n->leaf;
1425 #ifdef CONFIG_IP_FIB_TRIE_STATS
1426                         this_cpu_inc(stats->semantic_match_passed);
1427 #endif
1428                         return err;
1429                 }
1430         }
1431 #ifdef CONFIG_IP_FIB_TRIE_STATS
1432         this_cpu_inc(stats->semantic_match_miss);
1433 #endif
1434         goto backtrace;
1435 }
1436 EXPORT_SYMBOL_GPL(fib_table_lookup);
1437 
1438 static void fib_remove_alias(struct trie *t, struct key_vector *tp,
1439                              struct key_vector *l, struct fib_alias *old)
1440 {
1441         /* record the location of the previous list_info entry */
1442         struct hlist_node **pprev = old->fa_list.pprev;
1443         struct fib_alias *fa = hlist_entry(pprev, typeof(*fa), fa_list.next);
1444 
1445         /* remove the fib_alias from the list */
1446         hlist_del_rcu(&old->fa_list);
1447 
1448         /* if we emptied the list this leaf will be freed and we can sort
1449          * out parent suffix lengths as a part of trie_rebalance
1450          */
1451         if (hlist_empty(&l->leaf)) {
1452                 put_child_root(tp, l->key, NULL);
1453                 node_free(l);
1454                 trie_rebalance(t, tp);
1455                 return;
1456         }
1457 
1458         /* only access fa if it is pointing at the last valid hlist_node */
1459         if (*pprev)
1460                 return;
1461 
1462         /* update the trie with the latest suffix length */
1463         l->slen = fa->fa_slen;
1464         leaf_pull_suffix(tp, l);
1465 }
1466 
1467 /* Caller must hold RTNL. */
1468 int fib_table_delete(struct fib_table *tb, struct fib_config *cfg)
1469 {
1470         struct trie *t = (struct trie *) tb->tb_data;
1471         struct fib_alias *fa, *fa_to_delete;
1472         struct key_vector *l, *tp;
1473         u8 plen = cfg->fc_dst_len;
1474         u8 slen = KEYLENGTH - plen;
1475         u8 tos = cfg->fc_tos;
1476         u32 key;
1477 
1478         if (plen > KEYLENGTH)
1479                 return -EINVAL;
1480 
1481         key = ntohl(cfg->fc_dst);
1482 
1483         if ((plen < KEYLENGTH) && (key << plen))
1484                 return -EINVAL;
1485 
1486         l = fib_find_node(t, &tp, key);
1487         if (!l)
1488                 return -ESRCH;
1489 
1490         fa = fib_find_alias(&l->leaf, slen, tos, 0, tb->tb_id);
1491         if (!fa)
1492                 return -ESRCH;
1493 
1494         pr_debug("Deleting %08x/%d tos=%d t=%p\n", key, plen, tos, t);
1495 
1496         fa_to_delete = NULL;
1497         hlist_for_each_entry_from(fa, fa_list) {
1498                 struct fib_info *fi = fa->fa_info;
1499 
1500                 if ((fa->fa_slen != slen) ||
1501                     (fa->tb_id != tb->tb_id) ||
1502                     (fa->fa_tos != tos))
1503                         break;
1504 
1505                 if ((!cfg->fc_type || fa->fa_type == cfg->fc_type) &&
1506                     (cfg->fc_scope == RT_SCOPE_NOWHERE ||
1507                      fa->fa_info->fib_scope == cfg->fc_scope) &&
1508                     (!cfg->fc_prefsrc ||
1509                      fi->fib_prefsrc == cfg->fc_prefsrc) &&
1510                     (!cfg->fc_protocol ||
1511                      fi->fib_protocol == cfg->fc_protocol) &&
1512                     fib_nh_match(cfg, fi) == 0) {
1513                         fa_to_delete = fa;
1514                         break;
1515                 }
1516         }
1517 
1518         if (!fa_to_delete)
1519                 return -ESRCH;
1520 
1521         netdev_switch_fib_ipv4_del(key, plen, fa_to_delete->fa_info, tos,
1522                                    cfg->fc_type, tb->tb_id);
1523 
1524         rtmsg_fib(RTM_DELROUTE, htonl(key), fa_to_delete, plen, tb->tb_id,
1525                   &cfg->fc_nlinfo, 0);
1526 
1527         if (!plen)
1528                 tb->tb_num_default--;
1529 
1530         fib_remove_alias(t, tp, l, fa_to_delete);
1531 
1532         if (fa_to_delete->fa_state & FA_S_ACCESSED)
1533                 rt_cache_flush(cfg->fc_nlinfo.nl_net);
1534 
1535         fib_release_info(fa_to_delete->fa_info);
1536         alias_free_mem_rcu(fa_to_delete);
1537         return 0;
1538 }
1539 
1540 /* Scan for the next leaf starting at the provided key value */
1541 static struct key_vector *leaf_walk_rcu(struct key_vector **tn, t_key key)
1542 {
1543         struct key_vector *pn, *n = *tn;
1544         unsigned long cindex;
1545 
1546         /* this loop is meant to try and find the key in the trie */
1547         do {
1548                 /* record parent and next child index */
1549                 pn = n;
1550                 cindex = key ? get_index(key, pn) : 0;
1551 
1552                 if (cindex >> pn->bits)
1553                         break;
1554 
1555                 /* descend into the next child */
1556                 n = get_child_rcu(pn, cindex++);
1557                 if (!n)
1558                         break;
1559 
1560                 /* guarantee forward progress on the keys */
1561                 if (IS_LEAF(n) && (n->key >= key))
1562                         goto found;
1563         } while (IS_TNODE(n));
1564 
1565         /* this loop will search for the next leaf with a greater key */
1566         while (!IS_TRIE(pn)) {
1567                 /* if we exhausted the parent node we will need to climb */
1568                 if (cindex >= (1ul << pn->bits)) {
1569                         t_key pkey = pn->key;
1570 
1571                         pn = node_parent_rcu(pn);
1572                         cindex = get_index(pkey, pn) + 1;
1573                         continue;
1574                 }
1575 
1576                 /* grab the next available node */
1577                 n = get_child_rcu(pn, cindex++);
1578                 if (!n)
1579                         continue;
1580 
1581                 /* no need to compare keys since we bumped the index */
1582                 if (IS_LEAF(n))
1583                         goto found;
1584 
1585                 /* Rescan start scanning in new node */
1586                 pn = n;
1587                 cindex = 0;
1588         }
1589 
1590         *tn = pn;
1591         return NULL; /* Root of trie */
1592 found:
1593         /* if we are at the limit for keys just return NULL for the tnode */
1594         *tn = pn;
1595         return n;
1596 }
1597 
1598 static void fib_trie_free(struct fib_table *tb)
1599 {
1600         struct trie *t = (struct trie *)tb->tb_data;
1601         struct key_vector *pn = t->kv;
1602         unsigned long cindex = 1;
1603         struct hlist_node *tmp;
1604         struct fib_alias *fa;
1605 
1606         /* walk trie in reverse order and free everything */
1607         for (;;) {
1608                 struct key_vector *n;
1609 
1610                 if (!(cindex--)) {
1611                         t_key pkey = pn->key;
1612 
1613                         if (IS_TRIE(pn))
1614                                 break;
1615 
1616                         n = pn;
1617                         pn = node_parent(pn);
1618 
1619                         /* drop emptied tnode */
1620                         put_child_root(pn, n->key, NULL);
1621                         node_free(n);
1622 
1623                         cindex = get_index(pkey, pn);
1624 
1625                         continue;
1626                 }
1627 
1628                 /* grab the next available node */
1629                 n = get_child(pn, cindex);
1630                 if (!n)
1631                         continue;
1632 
1633                 if (IS_TNODE(n)) {
1634                         /* record pn and cindex for leaf walking */
1635                         pn = n;
1636                         cindex = 1ul << n->bits;
1637 
1638                         continue;
1639                 }
1640 
1641                 hlist_for_each_entry_safe(fa, tmp, &n->leaf, fa_list) {
1642                         hlist_del_rcu(&fa->fa_list);
1643                         alias_free_mem_rcu(fa);
1644                 }
1645 
1646                 put_child_root(pn, n->key, NULL);
1647                 node_free(n);
1648         }
1649 
1650 #ifdef CONFIG_IP_FIB_TRIE_STATS
1651         free_percpu(t->stats);
1652 #endif
1653         kfree(tb);
1654 }
1655 
1656 struct fib_table *fib_trie_unmerge(struct fib_table *oldtb)
1657 {
1658         struct trie *ot = (struct trie *)oldtb->tb_data;
1659         struct key_vector *l, *tp = ot->kv;
1660         struct fib_table *local_tb;
1661         struct fib_alias *fa;
1662         struct trie *lt;
1663         t_key key = 0;
1664 
1665         if (oldtb->tb_data == oldtb->__data)
1666                 return oldtb;
1667 
1668         local_tb = fib_trie_table(RT_TABLE_LOCAL, NULL);
1669         if (!local_tb)
1670                 return NULL;
1671 
1672         lt = (struct trie *)local_tb->tb_data;
1673 
1674         while ((l = leaf_walk_rcu(&tp, key)) != NULL) {
1675                 struct key_vector *local_l = NULL, *local_tp;
1676 
1677                 hlist_for_each_entry_rcu(fa, &l->leaf, fa_list) {
1678                         struct fib_alias *new_fa;
1679 
1680                         if (local_tb->tb_id != fa->tb_id)
1681                                 continue;
1682 
1683                         /* clone fa for new local table */
1684                         new_fa = kmem_cache_alloc(fn_alias_kmem, GFP_KERNEL);
1685                         if (!new_fa)
1686                                 goto out;
1687 
1688                         memcpy(new_fa, fa, sizeof(*fa));
1689 
1690                         /* insert clone into table */
1691                         if (!local_l)
1692                                 local_l = fib_find_node(lt, &local_tp, l->key);
1693 
1694                         if (fib_insert_alias(lt, local_tp, local_l, new_fa,
1695                                              NULL, l->key))
1696                                 goto out;
1697                 }
1698 
1699                 /* stop loop if key wrapped back to 0 */
1700                 key = l->key + 1;
1701                 if (key < l->key)
1702                         break;
1703         }
1704 
1705         return local_tb;
1706 out:
1707         fib_trie_free(local_tb);
1708 
1709         return NULL;
1710 }
1711 
1712 /* Caller must hold RTNL */
1713 void fib_table_flush_external(struct fib_table *tb)
1714 {
1715         struct trie *t = (struct trie *)tb->tb_data;
1716         struct key_vector *pn = t->kv;
1717         unsigned long cindex = 1;
1718         struct hlist_node *tmp;
1719         struct fib_alias *fa;
1720 
1721         /* walk trie in reverse order */
1722         for (;;) {
1723                 unsigned char slen = 0;
1724                 struct key_vector *n;
1725 
1726                 if (!(cindex--)) {
1727                         t_key pkey = pn->key;
1728 
1729                         /* cannot resize the trie vector */
1730                         if (IS_TRIE(pn))
1731                                 break;
1732 
1733                         /* resize completed node */
1734                         pn = resize(t, pn);
1735                         cindex = get_index(pkey, pn);
1736 
1737                         continue;
1738                 }
1739 
1740                 /* grab the next available node */
1741                 n = get_child(pn, cindex);
1742                 if (!n)
1743                         continue;
1744 
1745                 if (IS_TNODE(n)) {
1746                         /* record pn and cindex for leaf walking */
1747                         pn = n;
1748                         cindex = 1ul << n->bits;
1749 
1750                         continue;
1751                 }
1752 
1753                 hlist_for_each_entry_safe(fa, tmp, &n->leaf, fa_list) {
1754                         struct fib_info *fi = fa->fa_info;
1755 
1756                         /* if alias was cloned to local then we just
1757                          * need to remove the local copy from main
1758                          */
1759                         if (tb->tb_id != fa->tb_id) {
1760                                 hlist_del_rcu(&fa->fa_list);
1761                                 alias_free_mem_rcu(fa);
1762                                 continue;
1763                         }
1764 
1765                         /* record local slen */
1766                         slen = fa->fa_slen;
1767 
1768                         if (!fi || !(fi->fib_flags & RTNH_F_OFFLOAD))
1769                                 continue;
1770 
1771                         netdev_switch_fib_ipv4_del(n->key,
1772                                                    KEYLENGTH - fa->fa_slen,
1773                                                    fi, fa->fa_tos,
1774                                                    fa->fa_type, tb->tb_id);
1775                 }
1776 
1777                 /* update leaf slen */
1778                 n->slen = slen;
1779 
1780                 if (hlist_empty(&n->leaf)) {
1781                         put_child_root(pn, n->key, NULL);
1782                         node_free(n);
1783                 } else {
1784                         leaf_pull_suffix(pn, n);
1785                 }
1786         }
1787 }
1788 
1789 /* Caller must hold RTNL. */
1790 int fib_table_flush(struct fib_table *tb)
1791 {
1792         struct trie *t = (struct trie *)tb->tb_data;
1793         struct key_vector *pn = t->kv;
1794         unsigned long cindex = 1;
1795         struct hlist_node *tmp;
1796         struct fib_alias *fa;
1797         int found = 0;
1798 
1799         /* walk trie in reverse order */
1800         for (;;) {
1801                 unsigned char slen = 0;
1802                 struct key_vector *n;
1803 
1804                 if (!(cindex--)) {
1805                         t_key pkey = pn->key;
1806 
1807                         /* cannot resize the trie vector */
1808                         if (IS_TRIE(pn))
1809                                 break;
1810 
1811                         /* resize completed node */
1812                         pn = resize(t, pn);
1813                         cindex = get_index(pkey, pn);
1814 
1815                         continue;
1816                 }
1817 
1818                 /* grab the next available node */
1819                 n = get_child(pn, cindex);
1820                 if (!n)
1821                         continue;
1822 
1823                 if (IS_TNODE(n)) {
1824                         /* record pn and cindex for leaf walking */
1825                         pn = n;
1826                         cindex = 1ul << n->bits;
1827 
1828                         continue;
1829                 }
1830 
1831                 hlist_for_each_entry_safe(fa, tmp, &n->leaf, fa_list) {
1832                         struct fib_info *fi = fa->fa_info;
1833 
1834                         if (!fi || !(fi->fib_flags & RTNH_F_DEAD)) {
1835                                 slen = fa->fa_slen;
1836                                 continue;
1837                         }
1838 
1839                         netdev_switch_fib_ipv4_del(n->key,
1840                                                    KEYLENGTH - fa->fa_slen,
1841                                                    fi, fa->fa_tos,
1842                                                    fa->fa_type, tb->tb_id);
1843                         hlist_del_rcu(&fa->fa_list);
1844                         fib_release_info(fa->fa_info);
1845                         alias_free_mem_rcu(fa);
1846                         found++;
1847                 }
1848 
1849                 /* update leaf slen */
1850                 n->slen = slen;
1851 
1852                 if (hlist_empty(&n->leaf)) {
1853                         put_child_root(pn, n->key, NULL);
1854                         node_free(n);
1855                 } else {
1856                         leaf_pull_suffix(pn, n);
1857                 }
1858         }
1859 
1860         pr_debug("trie_flush found=%d\n", found);
1861         return found;
1862 }
1863 
1864 static void __trie_free_rcu(struct rcu_head *head)
1865 {
1866         struct fib_table *tb = container_of(head, struct fib_table, rcu);
1867 #ifdef CONFIG_IP_FIB_TRIE_STATS
1868         struct trie *t = (struct trie *)tb->tb_data;
1869 
1870         if (tb->tb_data == tb->__data)
1871                 free_percpu(t->stats);
1872 #endif /* CONFIG_IP_FIB_TRIE_STATS */
1873         kfree(tb);
1874 }
1875 
1876 void fib_free_table(struct fib_table *tb)
1877 {
1878         call_rcu(&tb->rcu, __trie_free_rcu);
1879 }
1880 
1881 static int fn_trie_dump_leaf(struct key_vector *l, struct fib_table *tb,
1882                              struct sk_buff *skb, struct netlink_callback *cb)
1883 {
1884         __be32 xkey = htonl(l->key);
1885         struct fib_alias *fa;
1886         int i, s_i;
1887 
1888         s_i = cb->args[4];
1889         i = 0;
1890 
1891         /* rcu_read_lock is hold by caller */
1892         hlist_for_each_entry_rcu(fa, &l->leaf, fa_list) {
1893                 if (i < s_i) {
1894                         i++;
1895                         continue;
1896                 }
1897 
1898                 if (tb->tb_id != fa->tb_id) {
1899                         i++;
1900                         continue;
1901                 }
1902 
1903                 if (fib_dump_info(skb, NETLINK_CB(cb->skb).portid,
1904                                   cb->nlh->nlmsg_seq,
1905                                   RTM_NEWROUTE,
1906                                   tb->tb_id,
1907                                   fa->fa_type,
1908                                   xkey,
1909                                   KEYLENGTH - fa->fa_slen,
1910                                   fa->fa_tos,
1911                                   fa->fa_info, NLM_F_MULTI) < 0) {
1912                         cb->args[4] = i;
1913                         return -1;
1914                 }
1915                 i++;
1916         }
1917 
1918         cb->args[4] = i;
1919         return skb->len;
1920 }
1921 
1922 /* rcu_read_lock needs to be hold by caller from readside */
1923 int fib_table_dump(struct fib_table *tb, struct sk_buff *skb,
1924                    struct netlink_callback *cb)
1925 {
1926         struct trie *t = (struct trie *)tb->tb_data;
1927         struct key_vector *l, *tp = t->kv;
1928         /* Dump starting at last key.
1929          * Note: 0.0.0.0/0 (ie default) is first key.
1930          */
1931         int count = cb->args[2];
1932         t_key key = cb->args[3];
1933 
1934         while ((l = leaf_walk_rcu(&tp, key)) != NULL) {
1935                 if (fn_trie_dump_leaf(l, tb, skb, cb) < 0) {
1936                         cb->args[3] = key;
1937                         cb->args[2] = count;
1938                         return -1;
1939                 }
1940 
1941                 ++count;
1942                 key = l->key + 1;
1943 
1944                 memset(&cb->args[4], 0,
1945                        sizeof(cb->args) - 4*sizeof(cb->args[0]));
1946 
1947                 /* stop loop if key wrapped back to 0 */
1948                 if (key < l->key)
1949                         break;
1950         }
1951 
1952         cb->args[3] = key;
1953         cb->args[2] = count;
1954 
1955         return skb->len;
1956 }
1957 
1958 void __init fib_trie_init(void)
1959 {
1960         fn_alias_kmem = kmem_cache_create("ip_fib_alias",
1961                                           sizeof(struct fib_alias),
1962                                           0, SLAB_PANIC, NULL);
1963 
1964         trie_leaf_kmem = kmem_cache_create("ip_fib_trie",
1965                                            LEAF_SIZE,
1966                                            0, SLAB_PANIC, NULL);
1967 }
1968 
1969 struct fib_table *fib_trie_table(u32 id, struct fib_table *alias)
1970 {
1971         struct fib_table *tb;
1972         struct trie *t;
1973         size_t sz = sizeof(*tb);
1974 
1975         if (!alias)
1976                 sz += sizeof(struct trie);
1977 
1978         tb = kzalloc(sz, GFP_KERNEL);
1979         if (!tb)
1980                 return NULL;
1981 
1982         tb->tb_id = id;
1983         tb->tb_default = -1;
1984         tb->tb_num_default = 0;
1985         tb->tb_data = (alias ? alias->__data : tb->__data);
1986 
1987         if (alias)
1988                 return tb;
1989 
1990         t = (struct trie *) tb->tb_data;
1991         t->kv[0].pos = KEYLENGTH;
1992         t->kv[0].slen = KEYLENGTH;
1993 #ifdef CONFIG_IP_FIB_TRIE_STATS
1994         t->stats = alloc_percpu(struct trie_use_stats);
1995         if (!t->stats) {
1996                 kfree(tb);
1997                 tb = NULL;
1998         }
1999 #endif
2000 
2001         return tb;
2002 }
2003 
2004 #ifdef CONFIG_PROC_FS
2005 /* Depth first Trie walk iterator */
2006 struct fib_trie_iter {
2007         struct seq_net_private p;
2008         struct fib_table *tb;
2009         struct key_vector *tnode;
2010         unsigned int index;
2011         unsigned int depth;
2012 };
2013 
2014 static struct key_vector *fib_trie_get_next(struct fib_trie_iter *iter)
2015 {
2016         unsigned long cindex = iter->index;
2017         struct key_vector *pn = iter->tnode;
2018         t_key pkey;
2019 
2020         pr_debug("get_next iter={node=%p index=%d depth=%d}\n",
2021                  iter->tnode, iter->index, iter->depth);
2022 
2023         while (!IS_TRIE(pn)) {
2024                 while (cindex < child_length(pn)) {
2025                         struct key_vector *n = get_child_rcu(pn, cindex++);
2026 
2027                         if (!n)
2028                                 continue;
2029 
2030                         if (IS_LEAF(n)) {
2031                                 iter->tnode = pn;
2032                                 iter->index = cindex;
2033                         } else {
2034                                 /* push down one level */
2035                                 iter->tnode = n;
2036                                 iter->index = 0;
2037                                 ++iter->depth;
2038                         }
2039 
2040                         return n;
2041                 }
2042 
2043                 /* Current node exhausted, pop back up */
2044                 pkey = pn->key;
2045                 pn = node_parent_rcu(pn);
2046                 cindex = get_index(pkey, pn) + 1;
2047                 --iter->depth;
2048         }
2049 
2050         /* record root node so further searches know we are done */
2051         iter->tnode = pn;
2052         iter->index = 0;
2053 
2054         return NULL;
2055 }
2056 
2057 static struct key_vector *fib_trie_get_first(struct fib_trie_iter *iter,
2058                                              struct trie *t)
2059 {
2060         struct key_vector *n, *pn = t->kv;
2061 
2062         if (!t)
2063                 return NULL;
2064 
2065         n = rcu_dereference(pn->tnode[0]);
2066         if (!n)
2067                 return NULL;
2068 
2069         if (IS_TNODE(n)) {
2070                 iter->tnode = n;
2071                 iter->index = 0;
2072                 iter->depth = 1;
2073         } else {
2074                 iter->tnode = pn;
2075                 iter->index = 0;
2076                 iter->depth = 0;
2077         }
2078 
2079         return n;
2080 }
2081 
2082 static void trie_collect_stats(struct trie *t, struct trie_stat *s)
2083 {
2084         struct key_vector *n;
2085         struct fib_trie_iter iter;
2086 
2087         memset(s, 0, sizeof(*s));
2088 
2089         rcu_read_lock();
2090         for (n = fib_trie_get_first(&iter, t); n; n = fib_trie_get_next(&iter)) {
2091                 if (IS_LEAF(n)) {
2092                         struct fib_alias *fa;
2093 
2094                         s->leaves++;
2095                         s->totdepth += iter.depth;
2096                         if (iter.depth > s->maxdepth)
2097                                 s->maxdepth = iter.depth;
2098 
2099                         hlist_for_each_entry_rcu(fa, &n->leaf, fa_list)
2100                                 ++s->prefixes;
2101                 } else {
2102                         s->tnodes++;
2103                         if (n->bits < MAX_STAT_DEPTH)
2104                                 s->nodesizes[n->bits]++;
2105                         s->nullpointers += tn_info(n)->empty_children;
2106                 }
2107         }
2108         rcu_read_unlock();
2109 }
2110 
2111 /*
2112  *      This outputs /proc/net/fib_triestats
2113  */
2114 static void trie_show_stats(struct seq_file *seq, struct trie_stat *stat)
2115 {
2116         unsigned int i, max, pointers, bytes, avdepth;
2117 
2118         if (stat->leaves)
2119                 avdepth = stat->totdepth*100 / stat->leaves;
2120         else
2121                 avdepth = 0;
2122 
2123         seq_printf(seq, "\tAver depth:     %u.%02d\n",
2124                    avdepth / 100, avdepth % 100);
2125         seq_printf(seq, "\tMax depth:      %u\n", stat->maxdepth);
2126 
2127         seq_printf(seq, "\tLeaves:         %u\n", stat->leaves);
2128         bytes = LEAF_SIZE * stat->leaves;
2129 
2130         seq_printf(seq, "\tPrefixes:       %u\n", stat->prefixes);
2131         bytes += sizeof(struct fib_alias) * stat->prefixes;
2132 
2133         seq_printf(seq, "\tInternal nodes: %u\n\t", stat->tnodes);
2134         bytes += TNODE_SIZE(0) * stat->tnodes;
2135 
2136         max = MAX_STAT_DEPTH;
2137         while (max > 0 && stat->nodesizes[max-1] == 0)
2138                 max--;
2139 
2140         pointers = 0;
2141         for (i = 1; i < max; i++)
2142                 if (stat->nodesizes[i] != 0) {
2143                         seq_printf(seq, "  %u: %u",  i, stat->nodesizes[i]);
2144                         pointers += (1<<i) * stat->nodesizes[i];
2145                 }
2146         seq_putc(seq, '\n');
2147         seq_printf(seq, "\tPointers: %u\n", pointers);
2148 
2149         bytes += sizeof(struct key_vector *) * pointers;
2150         seq_printf(seq, "Null ptrs: %u\n", stat->nullpointers);
2151         seq_printf(seq, "Total size: %u  kB\n", (bytes + 1023) / 1024);
2152 }
2153 
2154 #ifdef CONFIG_IP_FIB_TRIE_STATS
2155 static void trie_show_usage(struct seq_file *seq,
2156                             const struct trie_use_stats __percpu *stats)
2157 {
2158         struct trie_use_stats s = { 0 };
2159         int cpu;
2160 
2161         /* loop through all of the CPUs and gather up the stats */
2162         for_each_possible_cpu(cpu) {
2163                 const struct trie_use_stats *pcpu = per_cpu_ptr(stats, cpu);
2164 
2165                 s.gets += pcpu->gets;
2166                 s.backtrack += pcpu->backtrack;
2167                 s.semantic_match_passed += pcpu->semantic_match_passed;
2168                 s.semantic_match_miss += pcpu->semantic_match_miss;
2169                 s.null_node_hit += pcpu->null_node_hit;
2170                 s.resize_node_skipped += pcpu->resize_node_skipped;
2171         }
2172 
2173         seq_printf(seq, "\nCounters:\n---------\n");
2174         seq_printf(seq, "gets = %u\n", s.gets);
2175         seq_printf(seq, "backtracks = %u\n", s.backtrack);
2176         seq_printf(seq, "semantic match passed = %u\n",
2177                    s.semantic_match_passed);
2178         seq_printf(seq, "semantic match miss = %u\n", s.semantic_match_miss);
2179         seq_printf(seq, "null node hit= %u\n", s.null_node_hit);
2180         seq_printf(seq, "skipped node resize = %u\n\n", s.resize_node_skipped);
2181 }
2182 #endif /*  CONFIG_IP_FIB_TRIE_STATS */
2183 
2184 static void fib_table_print(struct seq_file *seq, struct fib_table *tb)
2185 {
2186         if (tb->tb_id == RT_TABLE_LOCAL)
2187                 seq_puts(seq, "Local:\n");
2188         else if (tb->tb_id == RT_TABLE_MAIN)
2189                 seq_puts(seq, "Main:\n");
2190         else
2191                 seq_printf(seq, "Id %d:\n", tb->tb_id);
2192 }
2193 
2194 
2195 static int fib_triestat_seq_show(struct seq_file *seq, void *v)
2196 {
2197         struct net *net = (struct net *)seq->private;
2198         unsigned int h;
2199 
2200         seq_printf(seq,
2201                    "Basic info: size of leaf:"
2202                    " %Zd bytes, size of tnode: %Zd bytes.\n",
2203                    LEAF_SIZE, TNODE_SIZE(0));
2204 
2205         for (h = 0; h < FIB_TABLE_HASHSZ; h++) {
2206                 struct hlist_head *head = &net->ipv4.fib_table_hash[h];
2207                 struct fib_table *tb;
2208 
2209                 hlist_for_each_entry_rcu(tb, head, tb_hlist) {
2210                         struct trie *t = (struct trie *) tb->tb_data;
2211                         struct trie_stat stat;
2212 
2213                         if (!t)
2214                                 continue;
2215 
2216                         fib_table_print(seq, tb);
2217 
2218                         trie_collect_stats(t, &stat);
2219                         trie_show_stats(seq, &stat);
2220 #ifdef CONFIG_IP_FIB_TRIE_STATS
2221                         trie_show_usage(seq, t->stats);
2222 #endif
2223                 }
2224         }
2225 
2226         return 0;
2227 }
2228 
2229 static int fib_triestat_seq_open(struct inode *inode, struct file *file)
2230 {
2231         return single_open_net(inode, file, fib_triestat_seq_show);
2232 }
2233 
2234 static const struct file_operations fib_triestat_fops = {
2235         .owner  = THIS_MODULE,
2236         .open   = fib_triestat_seq_open,
2237         .read   = seq_read,
2238         .llseek = seq_lseek,
2239         .release = single_release_net,
2240 };
2241 
2242 static struct key_vector *fib_trie_get_idx(struct seq_file *seq, loff_t pos)
2243 {
2244         struct fib_trie_iter *iter = seq->private;
2245         struct net *net = seq_file_net(seq);
2246         loff_t idx = 0;
2247         unsigned int h;
2248 
2249         for (h = 0; h < FIB_TABLE_HASHSZ; h++) {
2250                 struct hlist_head *head = &net->ipv4.fib_table_hash[h];
2251                 struct fib_table *tb;
2252 
2253                 hlist_for_each_entry_rcu(tb, head, tb_hlist) {
2254                         struct key_vector *n;
2255 
2256                         for (n = fib_trie_get_first(iter,
2257                                                     (struct trie *) tb->tb_data);
2258                              n; n = fib_trie_get_next(iter))
2259                                 if (pos == idx++) {
2260                                         iter->tb = tb;
2261                                         return n;
2262                                 }
2263                 }
2264         }
2265 
2266         return NULL;
2267 }
2268 
2269 static void *fib_trie_seq_start(struct seq_file *seq, loff_t *pos)
2270         __acquires(RCU)
2271 {
2272         rcu_read_lock();
2273         return fib_trie_get_idx(seq, *pos);
2274 }
2275 
2276 static void *fib_trie_seq_next(struct seq_file *seq, void *v, loff_t *pos)
2277 {
2278         struct fib_trie_iter *iter = seq->private;
2279         struct net *net = seq_file_net(seq);
2280         struct fib_table *tb = iter->tb;
2281         struct hlist_node *tb_node;
2282         unsigned int h;
2283         struct key_vector *n;
2284 
2285         ++*pos;
2286         /* next node in same table */
2287         n = fib_trie_get_next(iter);
2288         if (n)
2289                 return n;
2290 
2291         /* walk rest of this hash chain */
2292         h = tb->tb_id & (FIB_TABLE_HASHSZ - 1);
2293         while ((tb_node = rcu_dereference(hlist_next_rcu(&tb->tb_hlist)))) {
2294                 tb = hlist_entry(tb_node, struct fib_table, tb_hlist);
2295                 n = fib_trie_get_first(iter, (struct trie *) tb->tb_data);
2296                 if (n)
2297                         goto found;
2298         }
2299 
2300         /* new hash chain */
2301         while (++h < FIB_TABLE_HASHSZ) {
2302                 struct hlist_head *head = &net->ipv4.fib_table_hash[h];
2303                 hlist_for_each_entry_rcu(tb, head, tb_hlist) {
2304                         n = fib_trie_get_first(iter, (struct trie *) tb->tb_data);
2305                         if (n)
2306                                 goto found;
2307                 }
2308         }
2309         return NULL;
2310 
2311 found:
2312         iter->tb = tb;
2313         return n;
2314 }
2315 
2316 static void fib_trie_seq_stop(struct seq_file *seq, void *v)
2317         __releases(RCU)
2318 {
2319         rcu_read_unlock();
2320 }
2321 
2322 static void seq_indent(struct seq_file *seq, int n)
2323 {
2324         while (n-- > 0)
2325                 seq_puts(seq, "   ");
2326 }
2327 
2328 static inline const char *rtn_scope(char *buf, size_t len, enum rt_scope_t s)
2329 {
2330         switch (s) {
2331         case RT_SCOPE_UNIVERSE: return "universe";
2332         case RT_SCOPE_SITE:     return "site";
2333         case RT_SCOPE_LINK:     return "link";
2334         case RT_SCOPE_HOST:     return "host";
2335         case RT_SCOPE_NOWHERE:  return "nowhere";
2336         default:
2337                 snprintf(buf, len, "scope=%d", s);
2338                 return buf;
2339         }
2340 }
2341 
2342 static const char *const rtn_type_names[__RTN_MAX] = {
2343         [RTN_UNSPEC] = "UNSPEC",
2344         [RTN_UNICAST] = "UNICAST",
2345         [RTN_LOCAL] = "LOCAL",
2346         [RTN_BROADCAST] = "BROADCAST",
2347         [RTN_ANYCAST] = "ANYCAST",
2348         [RTN_MULTICAST] = "MULTICAST",
2349         [RTN_BLACKHOLE] = "BLACKHOLE",
2350         [RTN_UNREACHABLE] = "UNREACHABLE",
2351         [RTN_PROHIBIT] = "PROHIBIT",
2352         [RTN_THROW] = "THROW",
2353         [RTN_NAT] = "NAT",
2354         [RTN_XRESOLVE] = "XRESOLVE",
2355 };
2356 
2357 static inline const char *rtn_type(char *buf, size_t len, unsigned int t)
2358 {
2359         if (t < __RTN_MAX && rtn_type_names[t])
2360                 return rtn_type_names[t];
2361         snprintf(buf, len, "type %u", t);
2362         return buf;
2363 }
2364 
2365 /* Pretty print the trie */
2366 static int fib_trie_seq_show(struct seq_file *seq, void *v)
2367 {
2368         const struct fib_trie_iter *iter = seq->private;
2369         struct key_vector *n = v;
2370 
2371         if (IS_TRIE(node_parent_rcu(n)))
2372                 fib_table_print(seq, iter->tb);
2373 
2374         if (IS_TNODE(n)) {
2375                 __be32 prf = htonl(n->key);
2376 
2377                 seq_indent(seq, iter->depth-1);
2378                 seq_printf(seq, "  +-- %pI4/%zu %u %u %u\n",
2379                            &prf, KEYLENGTH - n->pos - n->bits, n->bits,
2380                            tn_info(n)->full_children,
2381                            tn_info(n)->empty_children);
2382         } else {
2383                 __be32 val = htonl(n->key);
2384                 struct fib_alias *fa;
2385 
2386                 seq_indent(seq, iter->depth);
2387                 seq_printf(seq, "  |-- %pI4\n", &val);
2388 
2389                 hlist_for_each_entry_rcu(fa, &n->leaf, fa_list) {
2390                         char buf1[32], buf2[32];
2391 
2392                         seq_indent(seq, iter->depth + 1);
2393                         seq_printf(seq, "  /%zu %s %s",
2394                                    KEYLENGTH - fa->fa_slen,
2395                                    rtn_scope(buf1, sizeof(buf1),
2396                                              fa->fa_info->fib_scope),
2397                                    rtn_type(buf2, sizeof(buf2),
2398                                             fa->fa_type));
2399                         if (fa->fa_tos)
2400                                 seq_printf(seq, " tos=%d", fa->fa_tos);
2401                         seq_putc(seq, '\n');
2402                 }
2403         }
2404 
2405         return 0;
2406 }
2407 
2408 static const struct seq_operations fib_trie_seq_ops = {
2409         .start  = fib_trie_seq_start,
2410         .next   = fib_trie_seq_next,
2411         .stop   = fib_trie_seq_stop,
2412         .show   = fib_trie_seq_show,
2413 };
2414 
2415 static int fib_trie_seq_open(struct inode *inode, struct file *file)
2416 {
2417         return seq_open_net(inode, file, &fib_trie_seq_ops,
2418                             sizeof(struct fib_trie_iter));
2419 }
2420 
2421 static const struct file_operations fib_trie_fops = {
2422         .owner  = THIS_MODULE,
2423         .open   = fib_trie_seq_open,
2424         .read   = seq_read,
2425         .llseek = seq_lseek,
2426         .release = seq_release_net,
2427 };
2428 
2429 struct fib_route_iter {
2430         struct seq_net_private p;
2431         struct fib_table *main_tb;
2432         struct key_vector *tnode;
2433         loff_t  pos;
2434         t_key   key;
2435 };
2436 
2437 static struct key_vector *fib_route_get_idx(struct fib_route_iter *iter,
2438                                             loff_t pos)
2439 {
2440         struct fib_table *tb = iter->main_tb;
2441         struct key_vector *l, **tp = &iter->tnode;
2442         struct trie *t;
2443         t_key key;
2444 
2445         /* use cache location of next-to-find key */
2446         if (iter->pos > 0 && pos >= iter->pos) {
2447                 pos -= iter->pos;
2448                 key = iter->key;
2449         } else {
2450                 t = (struct trie *)tb->tb_data;
2451                 iter->tnode = t->kv;
2452                 iter->pos = 0;
2453                 key = 0;
2454         }
2455 
2456         while ((l = leaf_walk_rcu(tp, key)) != NULL) {
2457                 key = l->key + 1;
2458                 iter->pos++;
2459 
2460                 if (pos-- <= 0)
2461                         break;
2462 
2463                 l = NULL;
2464 
2465                 /* handle unlikely case of a key wrap */
2466                 if (!key)
2467                         break;
2468         }
2469 
2470         if (l)
2471                 iter->key = key;        /* remember it */
2472         else
2473                 iter->pos = 0;          /* forget it */
2474 
2475         return l;
2476 }
2477 
2478 static void *fib_route_seq_start(struct seq_file *seq, loff_t *pos)
2479         __acquires(RCU)
2480 {
2481         struct fib_route_iter *iter = seq->private;
2482         struct fib_table *tb;
2483         struct trie *t;
2484 
2485         rcu_read_lock();
2486 
2487         tb = fib_get_table(seq_file_net(seq), RT_TABLE_MAIN);
2488         if (!tb)
2489                 return NULL;
2490 
2491         iter->main_tb = tb;
2492 
2493         if (*pos != 0)
2494                 return fib_route_get_idx(iter, *pos);
2495 
2496         t = (struct trie *)tb->tb_data;
2497         iter->tnode = t->kv;
2498         iter->pos = 0;
2499         iter->key = 0;
2500 
2501         return SEQ_START_TOKEN;
2502 }
2503 
2504 static void *fib_route_seq_next(struct seq_file *seq, void *v, loff_t *pos)
2505 {
2506         struct fib_route_iter *iter = seq->private;
2507         struct key_vector *l = NULL;
2508         t_key key = iter->key;
2509 
2510         ++*pos;
2511 
2512         /* only allow key of 0 for start of sequence */
2513         if ((v == SEQ_START_TOKEN) || key)
2514                 l = leaf_walk_rcu(&iter->tnode, key);
2515 
2516         if (l) {
2517                 iter->key = l->key + 1;
2518                 iter->pos++;
2519         } else {
2520                 iter->pos = 0;
2521         }
2522 
2523         return l;
2524 }
2525 
2526 static void fib_route_seq_stop(struct seq_file *seq, void *v)
2527         __releases(RCU)
2528 {
2529         rcu_read_unlock();
2530 }
2531 
2532 static unsigned int fib_flag_trans(int type, __be32 mask, const struct fib_info *fi)
2533 {
2534         unsigned int flags = 0;
2535 
2536         if (type == RTN_UNREACHABLE || type == RTN_PROHIBIT)
2537                 flags = RTF_REJECT;
2538         if (fi && fi->fib_nh->nh_gw)
2539                 flags |= RTF_GATEWAY;
2540         if (mask == htonl(0xFFFFFFFF))
2541                 flags |= RTF_HOST;
2542         flags |= RTF_UP;
2543         return flags;
2544 }
2545 
2546 /*
2547  *      This outputs /proc/net/route.
2548  *      The format of the file is not supposed to be changed
2549  *      and needs to be same as fib_hash output to avoid breaking
2550  *      legacy utilities
2551  */
2552 static int fib_route_seq_show(struct seq_file *seq, void *v)
2553 {
2554         struct fib_route_iter *iter = seq->private;
2555         struct fib_table *tb = iter->main_tb;
2556         struct fib_alias *fa;
2557         struct key_vector *l = v;
2558         __be32 prefix;
2559 
2560         if (v == SEQ_START_TOKEN) {
2561                 seq_printf(seq, "%-127s\n", "Iface\tDestination\tGateway "
2562                            "\tFlags\tRefCnt\tUse\tMetric\tMask\t\tMTU"
2563                            "\tWindow\tIRTT");
2564                 return 0;
2565         }
2566 
2567         prefix = htonl(l->key);
2568 
2569         hlist_for_each_entry_rcu(fa, &l->leaf, fa_list) {
2570                 const struct fib_info *fi = fa->fa_info;
2571                 __be32 mask = inet_make_mask(KEYLENGTH - fa->fa_slen);
2572                 unsigned int flags = fib_flag_trans(fa->fa_type, mask, fi);
2573 
2574                 if ((fa->fa_type == RTN_BROADCAST) ||
2575                     (fa->fa_type == RTN_MULTICAST))
2576                         continue;
2577 
2578                 if (fa->tb_id != tb->tb_id)
2579                         continue;
2580 
2581                 seq_setwidth(seq, 127);
2582 
2583                 if (fi)
2584                         seq_printf(seq,
2585                                    "%s\t%08X\t%08X\t%04X\t%d\t%u\t"
2586                                    "%d\t%08X\t%d\t%u\t%u",
2587                                    fi->fib_dev ? fi->fib_dev->name : "*",
2588                                    prefix,
2589                                    fi->fib_nh->nh_gw, flags, 0, 0,
2590                                    fi->fib_priority,
2591                                    mask,
2592                                    (fi->fib_advmss ?
2593                                     fi->fib_advmss + 40 : 0),
2594                                    fi->fib_window,
2595                                    fi->fib_rtt >> 3);
2596                 else
2597                         seq_printf(seq,
2598                                    "*\t%08X\t%08X\t%04X\t%d\t%u\t"
2599                                    "%d\t%08X\t%d\t%u\t%u",
2600                                    prefix, 0, flags, 0, 0, 0,
2601                                    mask, 0, 0, 0);
2602 
2603                 seq_pad(seq, '\n');
2604         }
2605 
2606         return 0;
2607 }
2608 
2609 static const struct seq_operations fib_route_seq_ops = {
2610         .start  = fib_route_seq_start,
2611         .next   = fib_route_seq_next,
2612         .stop   = fib_route_seq_stop,
2613         .show   = fib_route_seq_show,
2614 };
2615 
2616 static int fib_route_seq_open(struct inode *inode, struct file *file)
2617 {
2618         return seq_open_net(inode, file, &fib_route_seq_ops,
2619                             sizeof(struct fib_route_iter));
2620 }
2621 
2622 static const struct file_operations fib_route_fops = {
2623         .owner  = THIS_MODULE,
2624         .open   = fib_route_seq_open,
2625         .read   = seq_read,
2626         .llseek = seq_lseek,
2627         .release = seq_release_net,
2628 };
2629 
2630 int __net_init fib_proc_init(struct net *net)
2631 {
2632         if (!proc_create("fib_trie", S_IRUGO, net->proc_net, &fib_trie_fops))
2633                 goto out1;
2634 
2635         if (!proc_create("fib_triestat", S_IRUGO, net->proc_net,
2636                          &fib_triestat_fops))
2637                 goto out2;
2638 
2639         if (!proc_create("route", S_IRUGO, net->proc_net, &fib_route_fops))
2640                 goto out3;
2641 
2642         return 0;
2643 
2644 out3:
2645         remove_proc_entry("fib_triestat", net->proc_net);
2646 out2:
2647         remove_proc_entry("fib_trie", net->proc_net);
2648 out1:
2649         return -ENOMEM;
2650 }
2651 
2652 void __net_exit fib_proc_exit(struct net *net)
2653 {
2654         remove_proc_entry("fib_trie", net->proc_net);
2655         remove_proc_entry("fib_triestat", net->proc_net);
2656         remove_proc_entry("route", net->proc_net);
2657 }
2658 
2659 #endif /* CONFIG_PROC_FS */
2660 

This page was automatically generated by LXR 0.3.1 (source).  •  Linux is a registered trademark of Linus Torvalds  •  Contact us