Loading...
1/*
2 * Copyright (c) 2007-2014 Nicira, Inc.
3 *
4 * This program is free software; you can redistribute it and/or
5 * modify it under the terms of version 2 of the GNU General Public
6 * License as published by the Free Software Foundation.
7 *
8 * This program is distributed in the hope that it will be useful, but
9 * WITHOUT ANY WARRANTY; without even the implied warranty of
10 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
11 * General Public License for more details.
12 *
13 * You should have received a copy of the GNU General Public License
14 * along with this program; if not, write to the Free Software
15 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
16 * 02110-1301, USA
17 */
18
19#include "flow.h"
20#include "datapath.h"
21#include "flow_netlink.h"
22#include <linux/uaccess.h>
23#include <linux/netdevice.h>
24#include <linux/etherdevice.h>
25#include <linux/if_ether.h>
26#include <linux/if_vlan.h>
27#include <net/llc_pdu.h>
28#include <linux/kernel.h>
29#include <linux/jhash.h>
30#include <linux/jiffies.h>
31#include <linux/llc.h>
32#include <linux/module.h>
33#include <linux/in.h>
34#include <linux/rcupdate.h>
35#include <linux/cpumask.h>
36#include <linux/if_arp.h>
37#include <linux/ip.h>
38#include <linux/ipv6.h>
39#include <linux/sctp.h>
40#include <linux/tcp.h>
41#include <linux/udp.h>
42#include <linux/icmp.h>
43#include <linux/icmpv6.h>
44#include <linux/rculist.h>
45#include <net/ip.h>
46#include <net/ipv6.h>
47#include <net/ndisc.h>
48
49#define TBL_MIN_BUCKETS 1024
50#define REHASH_INTERVAL (10 * 60 * HZ)
51
52static struct kmem_cache *flow_cache;
53struct kmem_cache *flow_stats_cache __read_mostly;
54
55static u16 range_n_bytes(const struct sw_flow_key_range *range)
56{
57 return range->end - range->start;
58}
59
60void ovs_flow_mask_key(struct sw_flow_key *dst, const struct sw_flow_key *src,
61 bool full, const struct sw_flow_mask *mask)
62{
63 int start = full ? 0 : mask->range.start;
64 int len = full ? sizeof *dst : range_n_bytes(&mask->range);
65 const long *m = (const long *)((const u8 *)&mask->key + start);
66 const long *s = (const long *)((const u8 *)src + start);
67 long *d = (long *)((u8 *)dst + start);
68 int i;
69
70 /* If 'full' is true then all of 'dst' is fully initialized. Otherwise,
71 * if 'full' is false the memory outside of the 'mask->range' is left
72 * uninitialized. This can be used as an optimization when further
73 * operations on 'dst' only use contents within 'mask->range'.
74 */
75 for (i = 0; i < len; i += sizeof(long))
76 *d++ = *s++ & *m++;
77}
78
79struct sw_flow *ovs_flow_alloc(void)
80{
81 struct sw_flow *flow;
82 struct flow_stats *stats;
83
84 flow = kmem_cache_zalloc(flow_cache, GFP_KERNEL);
85 if (!flow)
86 return ERR_PTR(-ENOMEM);
87
88 flow->stats_last_writer = -1;
89
90 /* Initialize the default stat node. */
91 stats = kmem_cache_alloc_node(flow_stats_cache,
92 GFP_KERNEL | __GFP_ZERO,
93 node_online(0) ? 0 : NUMA_NO_NODE);
94 if (!stats)
95 goto err;
96
97 spin_lock_init(&stats->lock);
98
99 RCU_INIT_POINTER(flow->stats[0], stats);
100
101 cpumask_set_cpu(0, &flow->cpu_used_mask);
102
103 return flow;
104err:
105 kmem_cache_free(flow_cache, flow);
106 return ERR_PTR(-ENOMEM);
107}
108
109int ovs_flow_tbl_count(const struct flow_table *table)
110{
111 return table->count;
112}
113
114static struct flex_array *alloc_buckets(unsigned int n_buckets)
115{
116 struct flex_array *buckets;
117 int i, err;
118
119 buckets = flex_array_alloc(sizeof(struct hlist_head),
120 n_buckets, GFP_KERNEL);
121 if (!buckets)
122 return NULL;
123
124 err = flex_array_prealloc(buckets, 0, n_buckets, GFP_KERNEL);
125 if (err) {
126 flex_array_free(buckets);
127 return NULL;
128 }
129
130 for (i = 0; i < n_buckets; i++)
131 INIT_HLIST_HEAD((struct hlist_head *)
132 flex_array_get(buckets, i));
133
134 return buckets;
135}
136
137static void flow_free(struct sw_flow *flow)
138{
139 int cpu;
140
141 if (ovs_identifier_is_key(&flow->id))
142 kfree(flow->id.unmasked_key);
143 if (flow->sf_acts)
144 ovs_nla_free_flow_actions((struct sw_flow_actions __force *)flow->sf_acts);
145 /* We open code this to make sure cpu 0 is always considered */
146 for (cpu = 0; cpu < nr_cpu_ids; cpu = cpumask_next(cpu, &flow->cpu_used_mask))
147 if (flow->stats[cpu])
148 kmem_cache_free(flow_stats_cache,
149 (struct flow_stats __force *)flow->stats[cpu]);
150 kmem_cache_free(flow_cache, flow);
151}
152
153static void rcu_free_flow_callback(struct rcu_head *rcu)
154{
155 struct sw_flow *flow = container_of(rcu, struct sw_flow, rcu);
156
157 flow_free(flow);
158}
159
160void ovs_flow_free(struct sw_flow *flow, bool deferred)
161{
162 if (!flow)
163 return;
164
165 if (deferred)
166 call_rcu(&flow->rcu, rcu_free_flow_callback);
167 else
168 flow_free(flow);
169}
170
171static void free_buckets(struct flex_array *buckets)
172{
173 flex_array_free(buckets);
174}
175
176
177static void __table_instance_destroy(struct table_instance *ti)
178{
179 free_buckets(ti->buckets);
180 kfree(ti);
181}
182
183static struct table_instance *table_instance_alloc(int new_size)
184{
185 struct table_instance *ti = kmalloc(sizeof(*ti), GFP_KERNEL);
186
187 if (!ti)
188 return NULL;
189
190 ti->buckets = alloc_buckets(new_size);
191
192 if (!ti->buckets) {
193 kfree(ti);
194 return NULL;
195 }
196 ti->n_buckets = new_size;
197 ti->node_ver = 0;
198 ti->keep_flows = false;
199 get_random_bytes(&ti->hash_seed, sizeof(u32));
200
201 return ti;
202}
203
204int ovs_flow_tbl_init(struct flow_table *table)
205{
206 struct table_instance *ti, *ufid_ti;
207
208 ti = table_instance_alloc(TBL_MIN_BUCKETS);
209
210 if (!ti)
211 return -ENOMEM;
212
213 ufid_ti = table_instance_alloc(TBL_MIN_BUCKETS);
214 if (!ufid_ti)
215 goto free_ti;
216
217 rcu_assign_pointer(table->ti, ti);
218 rcu_assign_pointer(table->ufid_ti, ufid_ti);
219 INIT_LIST_HEAD(&table->mask_list);
220 table->last_rehash = jiffies;
221 table->count = 0;
222 table->ufid_count = 0;
223 return 0;
224
225free_ti:
226 __table_instance_destroy(ti);
227 return -ENOMEM;
228}
229
230static void flow_tbl_destroy_rcu_cb(struct rcu_head *rcu)
231{
232 struct table_instance *ti = container_of(rcu, struct table_instance, rcu);
233
234 __table_instance_destroy(ti);
235}
236
237static void table_instance_destroy(struct table_instance *ti,
238 struct table_instance *ufid_ti,
239 bool deferred)
240{
241 int i;
242
243 if (!ti)
244 return;
245
246 BUG_ON(!ufid_ti);
247 if (ti->keep_flows)
248 goto skip_flows;
249
250 for (i = 0; i < ti->n_buckets; i++) {
251 struct sw_flow *flow;
252 struct hlist_head *head = flex_array_get(ti->buckets, i);
253 struct hlist_node *n;
254 int ver = ti->node_ver;
255 int ufid_ver = ufid_ti->node_ver;
256
257 hlist_for_each_entry_safe(flow, n, head, flow_table.node[ver]) {
258 hlist_del_rcu(&flow->flow_table.node[ver]);
259 if (ovs_identifier_is_ufid(&flow->id))
260 hlist_del_rcu(&flow->ufid_table.node[ufid_ver]);
261 ovs_flow_free(flow, deferred);
262 }
263 }
264
265skip_flows:
266 if (deferred) {
267 call_rcu(&ti->rcu, flow_tbl_destroy_rcu_cb);
268 call_rcu(&ufid_ti->rcu, flow_tbl_destroy_rcu_cb);
269 } else {
270 __table_instance_destroy(ti);
271 __table_instance_destroy(ufid_ti);
272 }
273}
274
275/* No need for locking this function is called from RCU callback or
276 * error path.
277 */
278void ovs_flow_tbl_destroy(struct flow_table *table)
279{
280 struct table_instance *ti = rcu_dereference_raw(table->ti);
281 struct table_instance *ufid_ti = rcu_dereference_raw(table->ufid_ti);
282
283 table_instance_destroy(ti, ufid_ti, false);
284}
285
286struct sw_flow *ovs_flow_tbl_dump_next(struct table_instance *ti,
287 u32 *bucket, u32 *last)
288{
289 struct sw_flow *flow;
290 struct hlist_head *head;
291 int ver;
292 int i;
293
294 ver = ti->node_ver;
295 while (*bucket < ti->n_buckets) {
296 i = 0;
297 head = flex_array_get(ti->buckets, *bucket);
298 hlist_for_each_entry_rcu(flow, head, flow_table.node[ver]) {
299 if (i < *last) {
300 i++;
301 continue;
302 }
303 *last = i + 1;
304 return flow;
305 }
306 (*bucket)++;
307 *last = 0;
308 }
309
310 return NULL;
311}
312
313static struct hlist_head *find_bucket(struct table_instance *ti, u32 hash)
314{
315 hash = jhash_1word(hash, ti->hash_seed);
316 return flex_array_get(ti->buckets,
317 (hash & (ti->n_buckets - 1)));
318}
319
320static void table_instance_insert(struct table_instance *ti,
321 struct sw_flow *flow)
322{
323 struct hlist_head *head;
324
325 head = find_bucket(ti, flow->flow_table.hash);
326 hlist_add_head_rcu(&flow->flow_table.node[ti->node_ver], head);
327}
328
329static void ufid_table_instance_insert(struct table_instance *ti,
330 struct sw_flow *flow)
331{
332 struct hlist_head *head;
333
334 head = find_bucket(ti, flow->ufid_table.hash);
335 hlist_add_head_rcu(&flow->ufid_table.node[ti->node_ver], head);
336}
337
338static void flow_table_copy_flows(struct table_instance *old,
339 struct table_instance *new, bool ufid)
340{
341 int old_ver;
342 int i;
343
344 old_ver = old->node_ver;
345 new->node_ver = !old_ver;
346
347 /* Insert in new table. */
348 for (i = 0; i < old->n_buckets; i++) {
349 struct sw_flow *flow;
350 struct hlist_head *head;
351
352 head = flex_array_get(old->buckets, i);
353
354 if (ufid)
355 hlist_for_each_entry(flow, head,
356 ufid_table.node[old_ver])
357 ufid_table_instance_insert(new, flow);
358 else
359 hlist_for_each_entry(flow, head,
360 flow_table.node[old_ver])
361 table_instance_insert(new, flow);
362 }
363
364 old->keep_flows = true;
365}
366
367static struct table_instance *table_instance_rehash(struct table_instance *ti,
368 int n_buckets, bool ufid)
369{
370 struct table_instance *new_ti;
371
372 new_ti = table_instance_alloc(n_buckets);
373 if (!new_ti)
374 return NULL;
375
376 flow_table_copy_flows(ti, new_ti, ufid);
377
378 return new_ti;
379}
380
381int ovs_flow_tbl_flush(struct flow_table *flow_table)
382{
383 struct table_instance *old_ti, *new_ti;
384 struct table_instance *old_ufid_ti, *new_ufid_ti;
385
386 new_ti = table_instance_alloc(TBL_MIN_BUCKETS);
387 if (!new_ti)
388 return -ENOMEM;
389 new_ufid_ti = table_instance_alloc(TBL_MIN_BUCKETS);
390 if (!new_ufid_ti)
391 goto err_free_ti;
392
393 old_ti = ovsl_dereference(flow_table->ti);
394 old_ufid_ti = ovsl_dereference(flow_table->ufid_ti);
395
396 rcu_assign_pointer(flow_table->ti, new_ti);
397 rcu_assign_pointer(flow_table->ufid_ti, new_ufid_ti);
398 flow_table->last_rehash = jiffies;
399 flow_table->count = 0;
400 flow_table->ufid_count = 0;
401
402 table_instance_destroy(old_ti, old_ufid_ti, true);
403 return 0;
404
405err_free_ti:
406 __table_instance_destroy(new_ti);
407 return -ENOMEM;
408}
409
410static u32 flow_hash(const struct sw_flow_key *key,
411 const struct sw_flow_key_range *range)
412{
413 int key_start = range->start;
414 int key_end = range->end;
415 const u32 *hash_key = (const u32 *)((const u8 *)key + key_start);
416 int hash_u32s = (key_end - key_start) >> 2;
417
418 /* Make sure number of hash bytes are multiple of u32. */
419 BUILD_BUG_ON(sizeof(long) % sizeof(u32));
420
421 return jhash2(hash_key, hash_u32s, 0);
422}
423
424static int flow_key_start(const struct sw_flow_key *key)
425{
426 if (key->tun_proto)
427 return 0;
428 else
429 return rounddown(offsetof(struct sw_flow_key, phy),
430 sizeof(long));
431}
432
433static bool cmp_key(const struct sw_flow_key *key1,
434 const struct sw_flow_key *key2,
435 int key_start, int key_end)
436{
437 const long *cp1 = (const long *)((const u8 *)key1 + key_start);
438 const long *cp2 = (const long *)((const u8 *)key2 + key_start);
439 long diffs = 0;
440 int i;
441
442 for (i = key_start; i < key_end; i += sizeof(long))
443 diffs |= *cp1++ ^ *cp2++;
444
445 return diffs == 0;
446}
447
448static bool flow_cmp_masked_key(const struct sw_flow *flow,
449 const struct sw_flow_key *key,
450 const struct sw_flow_key_range *range)
451{
452 return cmp_key(&flow->key, key, range->start, range->end);
453}
454
455static bool ovs_flow_cmp_unmasked_key(const struct sw_flow *flow,
456 const struct sw_flow_match *match)
457{
458 struct sw_flow_key *key = match->key;
459 int key_start = flow_key_start(key);
460 int key_end = match->range.end;
461
462 BUG_ON(ovs_identifier_is_ufid(&flow->id));
463 return cmp_key(flow->id.unmasked_key, key, key_start, key_end);
464}
465
466static struct sw_flow *masked_flow_lookup(struct table_instance *ti,
467 const struct sw_flow_key *unmasked,
468 const struct sw_flow_mask *mask)
469{
470 struct sw_flow *flow;
471 struct hlist_head *head;
472 u32 hash;
473 struct sw_flow_key masked_key;
474
475 ovs_flow_mask_key(&masked_key, unmasked, false, mask);
476 hash = flow_hash(&masked_key, &mask->range);
477 head = find_bucket(ti, hash);
478 hlist_for_each_entry_rcu(flow, head, flow_table.node[ti->node_ver]) {
479 if (flow->mask == mask && flow->flow_table.hash == hash &&
480 flow_cmp_masked_key(flow, &masked_key, &mask->range))
481 return flow;
482 }
483 return NULL;
484}
485
486struct sw_flow *ovs_flow_tbl_lookup_stats(struct flow_table *tbl,
487 const struct sw_flow_key *key,
488 u32 *n_mask_hit)
489{
490 struct table_instance *ti = rcu_dereference_ovsl(tbl->ti);
491 struct sw_flow_mask *mask;
492 struct sw_flow *flow;
493
494 *n_mask_hit = 0;
495 list_for_each_entry_rcu(mask, &tbl->mask_list, list) {
496 (*n_mask_hit)++;
497 flow = masked_flow_lookup(ti, key, mask);
498 if (flow) /* Found */
499 return flow;
500 }
501 return NULL;
502}
503
504struct sw_flow *ovs_flow_tbl_lookup(struct flow_table *tbl,
505 const struct sw_flow_key *key)
506{
507 u32 __always_unused n_mask_hit;
508
509 return ovs_flow_tbl_lookup_stats(tbl, key, &n_mask_hit);
510}
511
512struct sw_flow *ovs_flow_tbl_lookup_exact(struct flow_table *tbl,
513 const struct sw_flow_match *match)
514{
515 struct table_instance *ti = rcu_dereference_ovsl(tbl->ti);
516 struct sw_flow_mask *mask;
517 struct sw_flow *flow;
518
519 /* Always called under ovs-mutex. */
520 list_for_each_entry(mask, &tbl->mask_list, list) {
521 flow = masked_flow_lookup(ti, match->key, mask);
522 if (flow && ovs_identifier_is_key(&flow->id) &&
523 ovs_flow_cmp_unmasked_key(flow, match))
524 return flow;
525 }
526 return NULL;
527}
528
529static u32 ufid_hash(const struct sw_flow_id *sfid)
530{
531 return jhash(sfid->ufid, sfid->ufid_len, 0);
532}
533
534static bool ovs_flow_cmp_ufid(const struct sw_flow *flow,
535 const struct sw_flow_id *sfid)
536{
537 if (flow->id.ufid_len != sfid->ufid_len)
538 return false;
539
540 return !memcmp(flow->id.ufid, sfid->ufid, sfid->ufid_len);
541}
542
543bool ovs_flow_cmp(const struct sw_flow *flow, const struct sw_flow_match *match)
544{
545 if (ovs_identifier_is_ufid(&flow->id))
546 return flow_cmp_masked_key(flow, match->key, &match->range);
547
548 return ovs_flow_cmp_unmasked_key(flow, match);
549}
550
551struct sw_flow *ovs_flow_tbl_lookup_ufid(struct flow_table *tbl,
552 const struct sw_flow_id *ufid)
553{
554 struct table_instance *ti = rcu_dereference_ovsl(tbl->ufid_ti);
555 struct sw_flow *flow;
556 struct hlist_head *head;
557 u32 hash;
558
559 hash = ufid_hash(ufid);
560 head = find_bucket(ti, hash);
561 hlist_for_each_entry_rcu(flow, head, ufid_table.node[ti->node_ver]) {
562 if (flow->ufid_table.hash == hash &&
563 ovs_flow_cmp_ufid(flow, ufid))
564 return flow;
565 }
566 return NULL;
567}
568
569int ovs_flow_tbl_num_masks(const struct flow_table *table)
570{
571 struct sw_flow_mask *mask;
572 int num = 0;
573
574 list_for_each_entry(mask, &table->mask_list, list)
575 num++;
576
577 return num;
578}
579
580static struct table_instance *table_instance_expand(struct table_instance *ti,
581 bool ufid)
582{
583 return table_instance_rehash(ti, ti->n_buckets * 2, ufid);
584}
585
586/* Remove 'mask' from the mask list, if it is not needed any more. */
587static void flow_mask_remove(struct flow_table *tbl, struct sw_flow_mask *mask)
588{
589 if (mask) {
590 /* ovs-lock is required to protect mask-refcount and
591 * mask list.
592 */
593 ASSERT_OVSL();
594 BUG_ON(!mask->ref_count);
595 mask->ref_count--;
596
597 if (!mask->ref_count) {
598 list_del_rcu(&mask->list);
599 kfree_rcu(mask, rcu);
600 }
601 }
602}
603
604/* Must be called with OVS mutex held. */
605void ovs_flow_tbl_remove(struct flow_table *table, struct sw_flow *flow)
606{
607 struct table_instance *ti = ovsl_dereference(table->ti);
608 struct table_instance *ufid_ti = ovsl_dereference(table->ufid_ti);
609
610 BUG_ON(table->count == 0);
611 hlist_del_rcu(&flow->flow_table.node[ti->node_ver]);
612 table->count--;
613 if (ovs_identifier_is_ufid(&flow->id)) {
614 hlist_del_rcu(&flow->ufid_table.node[ufid_ti->node_ver]);
615 table->ufid_count--;
616 }
617
618 /* RCU delete the mask. 'flow->mask' is not NULLed, as it should be
619 * accessible as long as the RCU read lock is held.
620 */
621 flow_mask_remove(table, flow->mask);
622}
623
624static struct sw_flow_mask *mask_alloc(void)
625{
626 struct sw_flow_mask *mask;
627
628 mask = kmalloc(sizeof(*mask), GFP_KERNEL);
629 if (mask)
630 mask->ref_count = 1;
631
632 return mask;
633}
634
635static bool mask_equal(const struct sw_flow_mask *a,
636 const struct sw_flow_mask *b)
637{
638 const u8 *a_ = (const u8 *)&a->key + a->range.start;
639 const u8 *b_ = (const u8 *)&b->key + b->range.start;
640
641 return (a->range.end == b->range.end)
642 && (a->range.start == b->range.start)
643 && (memcmp(a_, b_, range_n_bytes(&a->range)) == 0);
644}
645
646static struct sw_flow_mask *flow_mask_find(const struct flow_table *tbl,
647 const struct sw_flow_mask *mask)
648{
649 struct list_head *ml;
650
651 list_for_each(ml, &tbl->mask_list) {
652 struct sw_flow_mask *m;
653 m = container_of(ml, struct sw_flow_mask, list);
654 if (mask_equal(mask, m))
655 return m;
656 }
657
658 return NULL;
659}
660
661/* Add 'mask' into the mask list, if it is not already there. */
662static int flow_mask_insert(struct flow_table *tbl, struct sw_flow *flow,
663 const struct sw_flow_mask *new)
664{
665 struct sw_flow_mask *mask;
666 mask = flow_mask_find(tbl, new);
667 if (!mask) {
668 /* Allocate a new mask if none exsits. */
669 mask = mask_alloc();
670 if (!mask)
671 return -ENOMEM;
672 mask->key = new->key;
673 mask->range = new->range;
674 list_add_rcu(&mask->list, &tbl->mask_list);
675 } else {
676 BUG_ON(!mask->ref_count);
677 mask->ref_count++;
678 }
679
680 flow->mask = mask;
681 return 0;
682}
683
684/* Must be called with OVS mutex held. */
685static void flow_key_insert(struct flow_table *table, struct sw_flow *flow)
686{
687 struct table_instance *new_ti = NULL;
688 struct table_instance *ti;
689
690 flow->flow_table.hash = flow_hash(&flow->key, &flow->mask->range);
691 ti = ovsl_dereference(table->ti);
692 table_instance_insert(ti, flow);
693 table->count++;
694
695 /* Expand table, if necessary, to make room. */
696 if (table->count > ti->n_buckets)
697 new_ti = table_instance_expand(ti, false);
698 else if (time_after(jiffies, table->last_rehash + REHASH_INTERVAL))
699 new_ti = table_instance_rehash(ti, ti->n_buckets, false);
700
701 if (new_ti) {
702 rcu_assign_pointer(table->ti, new_ti);
703 call_rcu(&ti->rcu, flow_tbl_destroy_rcu_cb);
704 table->last_rehash = jiffies;
705 }
706}
707
708/* Must be called with OVS mutex held. */
709static void flow_ufid_insert(struct flow_table *table, struct sw_flow *flow)
710{
711 struct table_instance *ti;
712
713 flow->ufid_table.hash = ufid_hash(&flow->id);
714 ti = ovsl_dereference(table->ufid_ti);
715 ufid_table_instance_insert(ti, flow);
716 table->ufid_count++;
717
718 /* Expand table, if necessary, to make room. */
719 if (table->ufid_count > ti->n_buckets) {
720 struct table_instance *new_ti;
721
722 new_ti = table_instance_expand(ti, true);
723 if (new_ti) {
724 rcu_assign_pointer(table->ufid_ti, new_ti);
725 call_rcu(&ti->rcu, flow_tbl_destroy_rcu_cb);
726 }
727 }
728}
729
730/* Must be called with OVS mutex held. */
731int ovs_flow_tbl_insert(struct flow_table *table, struct sw_flow *flow,
732 const struct sw_flow_mask *mask)
733{
734 int err;
735
736 err = flow_mask_insert(table, flow, mask);
737 if (err)
738 return err;
739 flow_key_insert(table, flow);
740 if (ovs_identifier_is_ufid(&flow->id))
741 flow_ufid_insert(table, flow);
742
743 return 0;
744}
745
746/* Initializes the flow module.
747 * Returns zero if successful or a negative error code. */
748int ovs_flow_init(void)
749{
750 BUILD_BUG_ON(__alignof__(struct sw_flow_key) % __alignof__(long));
751 BUILD_BUG_ON(sizeof(struct sw_flow_key) % sizeof(long));
752
753 flow_cache = kmem_cache_create("sw_flow", sizeof(struct sw_flow)
754 + (nr_cpu_ids
755 * sizeof(struct flow_stats *)),
756 0, 0, NULL);
757 if (flow_cache == NULL)
758 return -ENOMEM;
759
760 flow_stats_cache
761 = kmem_cache_create("sw_flow_stats", sizeof(struct flow_stats),
762 0, SLAB_HWCACHE_ALIGN, NULL);
763 if (flow_stats_cache == NULL) {
764 kmem_cache_destroy(flow_cache);
765 flow_cache = NULL;
766 return -ENOMEM;
767 }
768
769 return 0;
770}
771
772/* Uninitializes the flow module. */
773void ovs_flow_exit(void)
774{
775 kmem_cache_destroy(flow_stats_cache);
776 kmem_cache_destroy(flow_cache);
777}
1/*
2 * Copyright (c) 2007-2014 Nicira, Inc.
3 *
4 * This program is free software; you can redistribute it and/or
5 * modify it under the terms of version 2 of the GNU General Public
6 * License as published by the Free Software Foundation.
7 *
8 * This program is distributed in the hope that it will be useful, but
9 * WITHOUT ANY WARRANTY; without even the implied warranty of
10 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
11 * General Public License for more details.
12 *
13 * You should have received a copy of the GNU General Public License
14 * along with this program; if not, write to the Free Software
15 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
16 * 02110-1301, USA
17 */
18
19#include "flow.h"
20#include "datapath.h"
21#include "flow_netlink.h"
22#include <linux/uaccess.h>
23#include <linux/netdevice.h>
24#include <linux/etherdevice.h>
25#include <linux/if_ether.h>
26#include <linux/if_vlan.h>
27#include <net/llc_pdu.h>
28#include <linux/kernel.h>
29#include <linux/jhash.h>
30#include <linux/jiffies.h>
31#include <linux/llc.h>
32#include <linux/module.h>
33#include <linux/in.h>
34#include <linux/rcupdate.h>
35#include <linux/if_arp.h>
36#include <linux/ip.h>
37#include <linux/ipv6.h>
38#include <linux/sctp.h>
39#include <linux/tcp.h>
40#include <linux/udp.h>
41#include <linux/icmp.h>
42#include <linux/icmpv6.h>
43#include <linux/rculist.h>
44#include <net/ip.h>
45#include <net/ipv6.h>
46#include <net/ndisc.h>
47
48#define TBL_MIN_BUCKETS 1024
49#define REHASH_INTERVAL (10 * 60 * HZ)
50
51static struct kmem_cache *flow_cache;
52struct kmem_cache *flow_stats_cache __read_mostly;
53
54static u16 range_n_bytes(const struct sw_flow_key_range *range)
55{
56 return range->end - range->start;
57}
58
59void ovs_flow_mask_key(struct sw_flow_key *dst, const struct sw_flow_key *src,
60 bool full, const struct sw_flow_mask *mask)
61{
62 int start = full ? 0 : mask->range.start;
63 int len = full ? sizeof *dst : range_n_bytes(&mask->range);
64 const long *m = (const long *)((const u8 *)&mask->key + start);
65 const long *s = (const long *)((const u8 *)src + start);
66 long *d = (long *)((u8 *)dst + start);
67 int i;
68
69 /* If 'full' is true then all of 'dst' is fully initialized. Otherwise,
70 * if 'full' is false the memory outside of the 'mask->range' is left
71 * uninitialized. This can be used as an optimization when further
72 * operations on 'dst' only use contents within 'mask->range'.
73 */
74 for (i = 0; i < len; i += sizeof(long))
75 *d++ = *s++ & *m++;
76}
77
78struct sw_flow *ovs_flow_alloc(void)
79{
80 struct sw_flow *flow;
81 struct flow_stats *stats;
82 int node;
83
84 flow = kmem_cache_alloc(flow_cache, GFP_KERNEL);
85 if (!flow)
86 return ERR_PTR(-ENOMEM);
87
88 flow->sf_acts = NULL;
89 flow->mask = NULL;
90 flow->id.unmasked_key = NULL;
91 flow->id.ufid_len = 0;
92 flow->stats_last_writer = NUMA_NO_NODE;
93
94 /* Initialize the default stat node. */
95 stats = kmem_cache_alloc_node(flow_stats_cache,
96 GFP_KERNEL | __GFP_ZERO,
97 node_online(0) ? 0 : NUMA_NO_NODE);
98 if (!stats)
99 goto err;
100
101 spin_lock_init(&stats->lock);
102
103 RCU_INIT_POINTER(flow->stats[0], stats);
104
105 for_each_node(node)
106 if (node != 0)
107 RCU_INIT_POINTER(flow->stats[node], NULL);
108
109 return flow;
110err:
111 kmem_cache_free(flow_cache, flow);
112 return ERR_PTR(-ENOMEM);
113}
114
115int ovs_flow_tbl_count(const struct flow_table *table)
116{
117 return table->count;
118}
119
120static struct flex_array *alloc_buckets(unsigned int n_buckets)
121{
122 struct flex_array *buckets;
123 int i, err;
124
125 buckets = flex_array_alloc(sizeof(struct hlist_head),
126 n_buckets, GFP_KERNEL);
127 if (!buckets)
128 return NULL;
129
130 err = flex_array_prealloc(buckets, 0, n_buckets, GFP_KERNEL);
131 if (err) {
132 flex_array_free(buckets);
133 return NULL;
134 }
135
136 for (i = 0; i < n_buckets; i++)
137 INIT_HLIST_HEAD((struct hlist_head *)
138 flex_array_get(buckets, i));
139
140 return buckets;
141}
142
143static void flow_free(struct sw_flow *flow)
144{
145 int node;
146
147 if (ovs_identifier_is_key(&flow->id))
148 kfree(flow->id.unmasked_key);
149 if (flow->sf_acts)
150 ovs_nla_free_flow_actions((struct sw_flow_actions __force *)flow->sf_acts);
151 for_each_node(node)
152 if (flow->stats[node])
153 kmem_cache_free(flow_stats_cache,
154 (struct flow_stats __force *)flow->stats[node]);
155 kmem_cache_free(flow_cache, flow);
156}
157
158static void rcu_free_flow_callback(struct rcu_head *rcu)
159{
160 struct sw_flow *flow = container_of(rcu, struct sw_flow, rcu);
161
162 flow_free(flow);
163}
164
165void ovs_flow_free(struct sw_flow *flow, bool deferred)
166{
167 if (!flow)
168 return;
169
170 if (deferred)
171 call_rcu(&flow->rcu, rcu_free_flow_callback);
172 else
173 flow_free(flow);
174}
175
176static void free_buckets(struct flex_array *buckets)
177{
178 flex_array_free(buckets);
179}
180
181
182static void __table_instance_destroy(struct table_instance *ti)
183{
184 free_buckets(ti->buckets);
185 kfree(ti);
186}
187
188static struct table_instance *table_instance_alloc(int new_size)
189{
190 struct table_instance *ti = kmalloc(sizeof(*ti), GFP_KERNEL);
191
192 if (!ti)
193 return NULL;
194
195 ti->buckets = alloc_buckets(new_size);
196
197 if (!ti->buckets) {
198 kfree(ti);
199 return NULL;
200 }
201 ti->n_buckets = new_size;
202 ti->node_ver = 0;
203 ti->keep_flows = false;
204 get_random_bytes(&ti->hash_seed, sizeof(u32));
205
206 return ti;
207}
208
209int ovs_flow_tbl_init(struct flow_table *table)
210{
211 struct table_instance *ti, *ufid_ti;
212
213 ti = table_instance_alloc(TBL_MIN_BUCKETS);
214
215 if (!ti)
216 return -ENOMEM;
217
218 ufid_ti = table_instance_alloc(TBL_MIN_BUCKETS);
219 if (!ufid_ti)
220 goto free_ti;
221
222 rcu_assign_pointer(table->ti, ti);
223 rcu_assign_pointer(table->ufid_ti, ufid_ti);
224 INIT_LIST_HEAD(&table->mask_list);
225 table->last_rehash = jiffies;
226 table->count = 0;
227 table->ufid_count = 0;
228 return 0;
229
230free_ti:
231 __table_instance_destroy(ti);
232 return -ENOMEM;
233}
234
235static void flow_tbl_destroy_rcu_cb(struct rcu_head *rcu)
236{
237 struct table_instance *ti = container_of(rcu, struct table_instance, rcu);
238
239 __table_instance_destroy(ti);
240}
241
242static void table_instance_destroy(struct table_instance *ti,
243 struct table_instance *ufid_ti,
244 bool deferred)
245{
246 int i;
247
248 if (!ti)
249 return;
250
251 BUG_ON(!ufid_ti);
252 if (ti->keep_flows)
253 goto skip_flows;
254
255 for (i = 0; i < ti->n_buckets; i++) {
256 struct sw_flow *flow;
257 struct hlist_head *head = flex_array_get(ti->buckets, i);
258 struct hlist_node *n;
259 int ver = ti->node_ver;
260 int ufid_ver = ufid_ti->node_ver;
261
262 hlist_for_each_entry_safe(flow, n, head, flow_table.node[ver]) {
263 hlist_del_rcu(&flow->flow_table.node[ver]);
264 if (ovs_identifier_is_ufid(&flow->id))
265 hlist_del_rcu(&flow->ufid_table.node[ufid_ver]);
266 ovs_flow_free(flow, deferred);
267 }
268 }
269
270skip_flows:
271 if (deferred) {
272 call_rcu(&ti->rcu, flow_tbl_destroy_rcu_cb);
273 call_rcu(&ufid_ti->rcu, flow_tbl_destroy_rcu_cb);
274 } else {
275 __table_instance_destroy(ti);
276 __table_instance_destroy(ufid_ti);
277 }
278}
279
280/* No need for locking this function is called from RCU callback or
281 * error path.
282 */
283void ovs_flow_tbl_destroy(struct flow_table *table)
284{
285 struct table_instance *ti = rcu_dereference_raw(table->ti);
286 struct table_instance *ufid_ti = rcu_dereference_raw(table->ufid_ti);
287
288 table_instance_destroy(ti, ufid_ti, false);
289}
290
291struct sw_flow *ovs_flow_tbl_dump_next(struct table_instance *ti,
292 u32 *bucket, u32 *last)
293{
294 struct sw_flow *flow;
295 struct hlist_head *head;
296 int ver;
297 int i;
298
299 ver = ti->node_ver;
300 while (*bucket < ti->n_buckets) {
301 i = 0;
302 head = flex_array_get(ti->buckets, *bucket);
303 hlist_for_each_entry_rcu(flow, head, flow_table.node[ver]) {
304 if (i < *last) {
305 i++;
306 continue;
307 }
308 *last = i + 1;
309 return flow;
310 }
311 (*bucket)++;
312 *last = 0;
313 }
314
315 return NULL;
316}
317
318static struct hlist_head *find_bucket(struct table_instance *ti, u32 hash)
319{
320 hash = jhash_1word(hash, ti->hash_seed);
321 return flex_array_get(ti->buckets,
322 (hash & (ti->n_buckets - 1)));
323}
324
325static void table_instance_insert(struct table_instance *ti,
326 struct sw_flow *flow)
327{
328 struct hlist_head *head;
329
330 head = find_bucket(ti, flow->flow_table.hash);
331 hlist_add_head_rcu(&flow->flow_table.node[ti->node_ver], head);
332}
333
334static void ufid_table_instance_insert(struct table_instance *ti,
335 struct sw_flow *flow)
336{
337 struct hlist_head *head;
338
339 head = find_bucket(ti, flow->ufid_table.hash);
340 hlist_add_head_rcu(&flow->ufid_table.node[ti->node_ver], head);
341}
342
343static void flow_table_copy_flows(struct table_instance *old,
344 struct table_instance *new, bool ufid)
345{
346 int old_ver;
347 int i;
348
349 old_ver = old->node_ver;
350 new->node_ver = !old_ver;
351
352 /* Insert in new table. */
353 for (i = 0; i < old->n_buckets; i++) {
354 struct sw_flow *flow;
355 struct hlist_head *head;
356
357 head = flex_array_get(old->buckets, i);
358
359 if (ufid)
360 hlist_for_each_entry(flow, head,
361 ufid_table.node[old_ver])
362 ufid_table_instance_insert(new, flow);
363 else
364 hlist_for_each_entry(flow, head,
365 flow_table.node[old_ver])
366 table_instance_insert(new, flow);
367 }
368
369 old->keep_flows = true;
370}
371
372static struct table_instance *table_instance_rehash(struct table_instance *ti,
373 int n_buckets, bool ufid)
374{
375 struct table_instance *new_ti;
376
377 new_ti = table_instance_alloc(n_buckets);
378 if (!new_ti)
379 return NULL;
380
381 flow_table_copy_flows(ti, new_ti, ufid);
382
383 return new_ti;
384}
385
386int ovs_flow_tbl_flush(struct flow_table *flow_table)
387{
388 struct table_instance *old_ti, *new_ti;
389 struct table_instance *old_ufid_ti, *new_ufid_ti;
390
391 new_ti = table_instance_alloc(TBL_MIN_BUCKETS);
392 if (!new_ti)
393 return -ENOMEM;
394 new_ufid_ti = table_instance_alloc(TBL_MIN_BUCKETS);
395 if (!new_ufid_ti)
396 goto err_free_ti;
397
398 old_ti = ovsl_dereference(flow_table->ti);
399 old_ufid_ti = ovsl_dereference(flow_table->ufid_ti);
400
401 rcu_assign_pointer(flow_table->ti, new_ti);
402 rcu_assign_pointer(flow_table->ufid_ti, new_ufid_ti);
403 flow_table->last_rehash = jiffies;
404 flow_table->count = 0;
405 flow_table->ufid_count = 0;
406
407 table_instance_destroy(old_ti, old_ufid_ti, true);
408 return 0;
409
410err_free_ti:
411 __table_instance_destroy(new_ti);
412 return -ENOMEM;
413}
414
415static u32 flow_hash(const struct sw_flow_key *key,
416 const struct sw_flow_key_range *range)
417{
418 int key_start = range->start;
419 int key_end = range->end;
420 const u32 *hash_key = (const u32 *)((const u8 *)key + key_start);
421 int hash_u32s = (key_end - key_start) >> 2;
422
423 /* Make sure number of hash bytes are multiple of u32. */
424 BUILD_BUG_ON(sizeof(long) % sizeof(u32));
425
426 return jhash2(hash_key, hash_u32s, 0);
427}
428
429static int flow_key_start(const struct sw_flow_key *key)
430{
431 if (key->tun_proto)
432 return 0;
433 else
434 return rounddown(offsetof(struct sw_flow_key, phy),
435 sizeof(long));
436}
437
438static bool cmp_key(const struct sw_flow_key *key1,
439 const struct sw_flow_key *key2,
440 int key_start, int key_end)
441{
442 const long *cp1 = (const long *)((const u8 *)key1 + key_start);
443 const long *cp2 = (const long *)((const u8 *)key2 + key_start);
444 long diffs = 0;
445 int i;
446
447 for (i = key_start; i < key_end; i += sizeof(long))
448 diffs |= *cp1++ ^ *cp2++;
449
450 return diffs == 0;
451}
452
453static bool flow_cmp_masked_key(const struct sw_flow *flow,
454 const struct sw_flow_key *key,
455 const struct sw_flow_key_range *range)
456{
457 return cmp_key(&flow->key, key, range->start, range->end);
458}
459
460static bool ovs_flow_cmp_unmasked_key(const struct sw_flow *flow,
461 const struct sw_flow_match *match)
462{
463 struct sw_flow_key *key = match->key;
464 int key_start = flow_key_start(key);
465 int key_end = match->range.end;
466
467 BUG_ON(ovs_identifier_is_ufid(&flow->id));
468 return cmp_key(flow->id.unmasked_key, key, key_start, key_end);
469}
470
471static struct sw_flow *masked_flow_lookup(struct table_instance *ti,
472 const struct sw_flow_key *unmasked,
473 const struct sw_flow_mask *mask)
474{
475 struct sw_flow *flow;
476 struct hlist_head *head;
477 u32 hash;
478 struct sw_flow_key masked_key;
479
480 ovs_flow_mask_key(&masked_key, unmasked, false, mask);
481 hash = flow_hash(&masked_key, &mask->range);
482 head = find_bucket(ti, hash);
483 hlist_for_each_entry_rcu(flow, head, flow_table.node[ti->node_ver]) {
484 if (flow->mask == mask && flow->flow_table.hash == hash &&
485 flow_cmp_masked_key(flow, &masked_key, &mask->range))
486 return flow;
487 }
488 return NULL;
489}
490
491struct sw_flow *ovs_flow_tbl_lookup_stats(struct flow_table *tbl,
492 const struct sw_flow_key *key,
493 u32 *n_mask_hit)
494{
495 struct table_instance *ti = rcu_dereference_ovsl(tbl->ti);
496 struct sw_flow_mask *mask;
497 struct sw_flow *flow;
498
499 *n_mask_hit = 0;
500 list_for_each_entry_rcu(mask, &tbl->mask_list, list) {
501 (*n_mask_hit)++;
502 flow = masked_flow_lookup(ti, key, mask);
503 if (flow) /* Found */
504 return flow;
505 }
506 return NULL;
507}
508
509struct sw_flow *ovs_flow_tbl_lookup(struct flow_table *tbl,
510 const struct sw_flow_key *key)
511{
512 u32 __always_unused n_mask_hit;
513
514 return ovs_flow_tbl_lookup_stats(tbl, key, &n_mask_hit);
515}
516
517struct sw_flow *ovs_flow_tbl_lookup_exact(struct flow_table *tbl,
518 const struct sw_flow_match *match)
519{
520 struct table_instance *ti = rcu_dereference_ovsl(tbl->ti);
521 struct sw_flow_mask *mask;
522 struct sw_flow *flow;
523
524 /* Always called under ovs-mutex. */
525 list_for_each_entry(mask, &tbl->mask_list, list) {
526 flow = masked_flow_lookup(ti, match->key, mask);
527 if (flow && ovs_identifier_is_key(&flow->id) &&
528 ovs_flow_cmp_unmasked_key(flow, match))
529 return flow;
530 }
531 return NULL;
532}
533
534static u32 ufid_hash(const struct sw_flow_id *sfid)
535{
536 return jhash(sfid->ufid, sfid->ufid_len, 0);
537}
538
539static bool ovs_flow_cmp_ufid(const struct sw_flow *flow,
540 const struct sw_flow_id *sfid)
541{
542 if (flow->id.ufid_len != sfid->ufid_len)
543 return false;
544
545 return !memcmp(flow->id.ufid, sfid->ufid, sfid->ufid_len);
546}
547
548bool ovs_flow_cmp(const struct sw_flow *flow, const struct sw_flow_match *match)
549{
550 if (ovs_identifier_is_ufid(&flow->id))
551 return flow_cmp_masked_key(flow, match->key, &match->range);
552
553 return ovs_flow_cmp_unmasked_key(flow, match);
554}
555
556struct sw_flow *ovs_flow_tbl_lookup_ufid(struct flow_table *tbl,
557 const struct sw_flow_id *ufid)
558{
559 struct table_instance *ti = rcu_dereference_ovsl(tbl->ufid_ti);
560 struct sw_flow *flow;
561 struct hlist_head *head;
562 u32 hash;
563
564 hash = ufid_hash(ufid);
565 head = find_bucket(ti, hash);
566 hlist_for_each_entry_rcu(flow, head, ufid_table.node[ti->node_ver]) {
567 if (flow->ufid_table.hash == hash &&
568 ovs_flow_cmp_ufid(flow, ufid))
569 return flow;
570 }
571 return NULL;
572}
573
574int ovs_flow_tbl_num_masks(const struct flow_table *table)
575{
576 struct sw_flow_mask *mask;
577 int num = 0;
578
579 list_for_each_entry(mask, &table->mask_list, list)
580 num++;
581
582 return num;
583}
584
585static struct table_instance *table_instance_expand(struct table_instance *ti,
586 bool ufid)
587{
588 return table_instance_rehash(ti, ti->n_buckets * 2, ufid);
589}
590
591/* Remove 'mask' from the mask list, if it is not needed any more. */
592static void flow_mask_remove(struct flow_table *tbl, struct sw_flow_mask *mask)
593{
594 if (mask) {
595 /* ovs-lock is required to protect mask-refcount and
596 * mask list.
597 */
598 ASSERT_OVSL();
599 BUG_ON(!mask->ref_count);
600 mask->ref_count--;
601
602 if (!mask->ref_count) {
603 list_del_rcu(&mask->list);
604 kfree_rcu(mask, rcu);
605 }
606 }
607}
608
609/* Must be called with OVS mutex held. */
610void ovs_flow_tbl_remove(struct flow_table *table, struct sw_flow *flow)
611{
612 struct table_instance *ti = ovsl_dereference(table->ti);
613 struct table_instance *ufid_ti = ovsl_dereference(table->ufid_ti);
614
615 BUG_ON(table->count == 0);
616 hlist_del_rcu(&flow->flow_table.node[ti->node_ver]);
617 table->count--;
618 if (ovs_identifier_is_ufid(&flow->id)) {
619 hlist_del_rcu(&flow->ufid_table.node[ufid_ti->node_ver]);
620 table->ufid_count--;
621 }
622
623 /* RCU delete the mask. 'flow->mask' is not NULLed, as it should be
624 * accessible as long as the RCU read lock is held.
625 */
626 flow_mask_remove(table, flow->mask);
627}
628
629static struct sw_flow_mask *mask_alloc(void)
630{
631 struct sw_flow_mask *mask;
632
633 mask = kmalloc(sizeof(*mask), GFP_KERNEL);
634 if (mask)
635 mask->ref_count = 1;
636
637 return mask;
638}
639
640static bool mask_equal(const struct sw_flow_mask *a,
641 const struct sw_flow_mask *b)
642{
643 const u8 *a_ = (const u8 *)&a->key + a->range.start;
644 const u8 *b_ = (const u8 *)&b->key + b->range.start;
645
646 return (a->range.end == b->range.end)
647 && (a->range.start == b->range.start)
648 && (memcmp(a_, b_, range_n_bytes(&a->range)) == 0);
649}
650
651static struct sw_flow_mask *flow_mask_find(const struct flow_table *tbl,
652 const struct sw_flow_mask *mask)
653{
654 struct list_head *ml;
655
656 list_for_each(ml, &tbl->mask_list) {
657 struct sw_flow_mask *m;
658 m = container_of(ml, struct sw_flow_mask, list);
659 if (mask_equal(mask, m))
660 return m;
661 }
662
663 return NULL;
664}
665
666/* Add 'mask' into the mask list, if it is not already there. */
667static int flow_mask_insert(struct flow_table *tbl, struct sw_flow *flow,
668 const struct sw_flow_mask *new)
669{
670 struct sw_flow_mask *mask;
671 mask = flow_mask_find(tbl, new);
672 if (!mask) {
673 /* Allocate a new mask if none exsits. */
674 mask = mask_alloc();
675 if (!mask)
676 return -ENOMEM;
677 mask->key = new->key;
678 mask->range = new->range;
679 list_add_rcu(&mask->list, &tbl->mask_list);
680 } else {
681 BUG_ON(!mask->ref_count);
682 mask->ref_count++;
683 }
684
685 flow->mask = mask;
686 return 0;
687}
688
689/* Must be called with OVS mutex held. */
690static void flow_key_insert(struct flow_table *table, struct sw_flow *flow)
691{
692 struct table_instance *new_ti = NULL;
693 struct table_instance *ti;
694
695 flow->flow_table.hash = flow_hash(&flow->key, &flow->mask->range);
696 ti = ovsl_dereference(table->ti);
697 table_instance_insert(ti, flow);
698 table->count++;
699
700 /* Expand table, if necessary, to make room. */
701 if (table->count > ti->n_buckets)
702 new_ti = table_instance_expand(ti, false);
703 else if (time_after(jiffies, table->last_rehash + REHASH_INTERVAL))
704 new_ti = table_instance_rehash(ti, ti->n_buckets, false);
705
706 if (new_ti) {
707 rcu_assign_pointer(table->ti, new_ti);
708 call_rcu(&ti->rcu, flow_tbl_destroy_rcu_cb);
709 table->last_rehash = jiffies;
710 }
711}
712
713/* Must be called with OVS mutex held. */
714static void flow_ufid_insert(struct flow_table *table, struct sw_flow *flow)
715{
716 struct table_instance *ti;
717
718 flow->ufid_table.hash = ufid_hash(&flow->id);
719 ti = ovsl_dereference(table->ufid_ti);
720 ufid_table_instance_insert(ti, flow);
721 table->ufid_count++;
722
723 /* Expand table, if necessary, to make room. */
724 if (table->ufid_count > ti->n_buckets) {
725 struct table_instance *new_ti;
726
727 new_ti = table_instance_expand(ti, true);
728 if (new_ti) {
729 rcu_assign_pointer(table->ufid_ti, new_ti);
730 call_rcu(&ti->rcu, flow_tbl_destroy_rcu_cb);
731 }
732 }
733}
734
735/* Must be called with OVS mutex held. */
736int ovs_flow_tbl_insert(struct flow_table *table, struct sw_flow *flow,
737 const struct sw_flow_mask *mask)
738{
739 int err;
740
741 err = flow_mask_insert(table, flow, mask);
742 if (err)
743 return err;
744 flow_key_insert(table, flow);
745 if (ovs_identifier_is_ufid(&flow->id))
746 flow_ufid_insert(table, flow);
747
748 return 0;
749}
750
751/* Initializes the flow module.
752 * Returns zero if successful or a negative error code. */
753int ovs_flow_init(void)
754{
755 BUILD_BUG_ON(__alignof__(struct sw_flow_key) % __alignof__(long));
756 BUILD_BUG_ON(sizeof(struct sw_flow_key) % sizeof(long));
757
758 flow_cache = kmem_cache_create("sw_flow", sizeof(struct sw_flow)
759 + (nr_node_ids
760 * sizeof(struct flow_stats *)),
761 0, 0, NULL);
762 if (flow_cache == NULL)
763 return -ENOMEM;
764
765 flow_stats_cache
766 = kmem_cache_create("sw_flow_stats", sizeof(struct flow_stats),
767 0, SLAB_HWCACHE_ALIGN, NULL);
768 if (flow_stats_cache == NULL) {
769 kmem_cache_destroy(flow_cache);
770 flow_cache = NULL;
771 return -ENOMEM;
772 }
773
774 return 0;
775}
776
777/* Uninitializes the flow module. */
778void ovs_flow_exit(void)
779{
780 kmem_cache_destroy(flow_stats_cache);
781 kmem_cache_destroy(flow_cache);
782}