1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16#include <linux/err.h>
17#include <linux/slab.h>
18#include <linux/kmod.h>
19#include <linux/list.h>
20#include <linux/spinlock.h>
21#include <linux/workqueue.h>
22#include <linux/notifier.h>
23#include <linux/netdevice.h>
24#include <linux/netfilter.h>
25#include <linux/module.h>
26#include <linux/cache.h>
27#include <linux/audit.h>
28#include <net/dst.h>
29#include <net/xfrm.h>
30#include <net/ip.h>
31#ifdef CONFIG_XFRM_STATISTICS
32#include <net/snmp.h>
33#endif
34
35#include "xfrm_hash.h"
36
37int sysctl_xfrm_larval_drop __read_mostly = 1;
38
39#ifdef CONFIG_XFRM_STATISTICS
40DEFINE_SNMP_STAT(struct linux_xfrm_mib, xfrm_statistics) __read_mostly;
41EXPORT_SYMBOL(xfrm_statistics);
42#endif
43
44DEFINE_MUTEX(xfrm_cfg_mutex);
45EXPORT_SYMBOL(xfrm_cfg_mutex);
46
47static DEFINE_RWLOCK(xfrm_policy_lock);
48
49static struct list_head xfrm_policy_all;
50unsigned int xfrm_policy_count[XFRM_POLICY_MAX*2];
51EXPORT_SYMBOL(xfrm_policy_count);
52
53static DEFINE_RWLOCK(xfrm_policy_afinfo_lock);
54static struct xfrm_policy_afinfo *xfrm_policy_afinfo[NPROTO];
55
56static struct kmem_cache *xfrm_dst_cache __read_mostly;
57
58static struct work_struct xfrm_policy_gc_work;
59static HLIST_HEAD(xfrm_policy_gc_list);
60static DEFINE_SPINLOCK(xfrm_policy_gc_lock);
61
62static struct xfrm_policy_afinfo *xfrm_policy_get_afinfo(unsigned short family);
63static void xfrm_policy_put_afinfo(struct xfrm_policy_afinfo *afinfo);
64static void xfrm_init_pmtu(struct dst_entry *dst);
65
66static inline int
67__xfrm4_selector_match(struct xfrm_selector *sel, struct flowi *fl)
68{
69 return addr_match(&fl->fl4_dst, &sel->daddr, sel->prefixlen_d) &&
70 addr_match(&fl->fl4_src, &sel->saddr, sel->prefixlen_s) &&
71 !((xfrm_flowi_dport(fl) ^ sel->dport) & sel->dport_mask) &&
72 !((xfrm_flowi_sport(fl) ^ sel->sport) & sel->sport_mask) &&
73 (fl->proto == sel->proto || !sel->proto) &&
74 (fl->oif == sel->ifindex || !sel->ifindex);
75}
76
77static inline int
78__xfrm6_selector_match(struct xfrm_selector *sel, struct flowi *fl)
79{
80 return addr_match(&fl->fl6_dst, &sel->daddr, sel->prefixlen_d) &&
81 addr_match(&fl->fl6_src, &sel->saddr, sel->prefixlen_s) &&
82 !((xfrm_flowi_dport(fl) ^ sel->dport) & sel->dport_mask) &&
83 !((xfrm_flowi_sport(fl) ^ sel->sport) & sel->sport_mask) &&
84 (fl->proto == sel->proto || !sel->proto) &&
85 (fl->oif == sel->ifindex || !sel->ifindex);
86}
87
88int xfrm_selector_match(struct xfrm_selector *sel, struct flowi *fl,
89 unsigned short family)
90{
91 switch (family) {
92 case AF_INET:
93 return __xfrm4_selector_match(sel, fl);
94 case AF_INET6:
95 return __xfrm6_selector_match(sel, fl);
96 }
97 return 0;
98}
99
100static inline struct dst_entry *__xfrm_dst_lookup(int tos,
101 xfrm_address_t *saddr,
102 xfrm_address_t *daddr,
103 int family)
104{
105 struct xfrm_policy_afinfo *afinfo;
106 struct dst_entry *dst;
107
108 afinfo = xfrm_policy_get_afinfo(family);
109 if (unlikely(afinfo == NULL))
110 return ERR_PTR(-EAFNOSUPPORT);
111
112 dst = afinfo->dst_lookup(tos, saddr, daddr);
113
114 xfrm_policy_put_afinfo(afinfo);
115
116 return dst;
117}
118
119static inline struct dst_entry *xfrm_dst_lookup(struct xfrm_state *x, int tos,
120 xfrm_address_t *prev_saddr,
121 xfrm_address_t *prev_daddr,
122 int family)
123{
124 xfrm_address_t *saddr = &x->props.saddr;
125 xfrm_address_t *daddr = &x->id.daddr;
126 struct dst_entry *dst;
127
128 if (x->type->flags & XFRM_TYPE_LOCAL_COADDR) {
129 saddr = x->coaddr;
130 daddr = prev_daddr;
131 }
132 if (x->type->flags & XFRM_TYPE_REMOTE_COADDR) {
133 saddr = prev_saddr;
134 daddr = x->coaddr;
135 }
136
137 dst = __xfrm_dst_lookup(tos, saddr, daddr, family);
138
139 if (!IS_ERR(dst)) {
140 if (prev_saddr != saddr)
141 memcpy(prev_saddr, saddr, sizeof(*prev_saddr));
142 if (prev_daddr != daddr)
143 memcpy(prev_daddr, daddr, sizeof(*prev_daddr));
144 }
145
146 return dst;
147}
148
149static inline unsigned long make_jiffies(long secs)
150{
151 if (secs >= (MAX_SCHEDULE_TIMEOUT-1)/HZ)
152 return MAX_SCHEDULE_TIMEOUT-1;
153 else
154 return secs*HZ;
155}
156
157static void xfrm_policy_timer(unsigned long data)
158{
159 struct xfrm_policy *xp = (struct xfrm_policy*)data;
160 unsigned long now = get_seconds();
161 long next = LONG_MAX;
162 int warn = 0;
163 int dir;
164
165 read_lock(&xp->lock);
166
167 if (xp->walk.dead)
168 goto out;
169
170 dir = xfrm_policy_id2dir(xp->index);
171
172 if (xp->lft.hard_add_expires_seconds) {
173 long tmo = xp->lft.hard_add_expires_seconds +
174 xp->curlft.add_time - now;
175 if (tmo <= 0)
176 goto expired;
177 if (tmo < next)
178 next = tmo;
179 }
180 if (xp->lft.hard_use_expires_seconds) {
181 long tmo = xp->lft.hard_use_expires_seconds +
182 (xp->curlft.use_time ? : xp->curlft.add_time) - now;
183 if (tmo <= 0)
184 goto expired;
185 if (tmo < next)
186 next = tmo;
187 }
188 if (xp->lft.soft_add_expires_seconds) {
189 long tmo = xp->lft.soft_add_expires_seconds +
190 xp->curlft.add_time - now;
191 if (tmo <= 0) {
192 warn = 1;
193 tmo = XFRM_KM_TIMEOUT;
194 }
195 if (tmo < next)
196 next = tmo;
197 }
198 if (xp->lft.soft_use_expires_seconds) {
199 long tmo = xp->lft.soft_use_expires_seconds +
200 (xp->curlft.use_time ? : xp->curlft.add_time) - now;
201 if (tmo <= 0) {
202 warn = 1;
203 tmo = XFRM_KM_TIMEOUT;
204 }
205 if (tmo < next)
206 next = tmo;
207 }
208
209 if (warn)
210 km_policy_expired(xp, dir, 0, 0);
211 if (next != LONG_MAX &&
212 !mod_timer(&xp->timer, jiffies + make_jiffies(next)))
213 xfrm_pol_hold(xp);
214
215out:
216 read_unlock(&xp->lock);
217 xfrm_pol_put(xp);
218 return;
219
220expired:
221 read_unlock(&xp->lock);
222 if (!xfrm_policy_delete(xp, dir))
223 km_policy_expired(xp, dir, 1, 0);
224 xfrm_pol_put(xp);
225}
226
227
228
229
230
231
232struct xfrm_policy *xfrm_policy_alloc(gfp_t gfp)
233{
234 struct xfrm_policy *policy;
235
236 policy = kzalloc(sizeof(struct xfrm_policy), gfp);
237
238 if (policy) {
239 INIT_LIST_HEAD(&policy->walk.all);
240 INIT_HLIST_NODE(&policy->bydst);
241 INIT_HLIST_NODE(&policy->byidx);
242 rwlock_init(&policy->lock);
243 atomic_set(&policy->refcnt, 1);
244 setup_timer(&policy->timer, xfrm_policy_timer,
245 (unsigned long)policy);
246 }
247 return policy;
248}
249EXPORT_SYMBOL(xfrm_policy_alloc);
250
251
252
253void xfrm_policy_destroy(struct xfrm_policy *policy)
254{
255 BUG_ON(!policy->walk.dead);
256
257 BUG_ON(policy->bundles);
258
259 if (del_timer(&policy->timer))
260 BUG();
261
262 security_xfrm_policy_free(policy->security);
263 kfree(policy);
264}
265EXPORT_SYMBOL(xfrm_policy_destroy);
266
267static void xfrm_policy_gc_kill(struct xfrm_policy *policy)
268{
269 struct dst_entry *dst;
270
271 while ((dst = policy->bundles) != NULL) {
272 policy->bundles = dst->next;
273 dst_free(dst);
274 }
275
276 if (del_timer(&policy->timer))
277 atomic_dec(&policy->refcnt);
278
279 if (atomic_read(&policy->refcnt) > 1)
280 flow_cache_flush();
281
282 xfrm_pol_put(policy);
283}
284
285static void xfrm_policy_gc_task(struct work_struct *work)
286{
287 struct xfrm_policy *policy;
288 struct hlist_node *entry, *tmp;
289 struct hlist_head gc_list;
290
291 spin_lock_bh(&xfrm_policy_gc_lock);
292 gc_list.first = xfrm_policy_gc_list.first;
293 INIT_HLIST_HEAD(&xfrm_policy_gc_list);
294 spin_unlock_bh(&xfrm_policy_gc_lock);
295
296 hlist_for_each_entry_safe(policy, entry, tmp, &gc_list, bydst)
297 xfrm_policy_gc_kill(policy);
298}
299
300
301
302
303
304static void xfrm_policy_kill(struct xfrm_policy *policy)
305{
306 int dead;
307
308 write_lock_bh(&policy->lock);
309 dead = policy->walk.dead;
310 policy->walk.dead = 1;
311 write_unlock_bh(&policy->lock);
312
313 if (unlikely(dead)) {
314 WARN_ON(1);
315 return;
316 }
317
318 spin_lock_bh(&xfrm_policy_gc_lock);
319 hlist_add_head(&policy->bydst, &xfrm_policy_gc_list);
320 spin_unlock_bh(&xfrm_policy_gc_lock);
321
322 schedule_work(&xfrm_policy_gc_work);
323}
324
325struct xfrm_policy_hash {
326 struct hlist_head *table;
327 unsigned int hmask;
328};
329
330static struct hlist_head xfrm_policy_inexact[XFRM_POLICY_MAX*2];
331static struct xfrm_policy_hash xfrm_policy_bydst[XFRM_POLICY_MAX*2] __read_mostly;
332static struct hlist_head *xfrm_policy_byidx __read_mostly;
333static unsigned int xfrm_idx_hmask __read_mostly;
334static unsigned int xfrm_policy_hashmax __read_mostly = 1 * 1024 * 1024;
335
336static inline unsigned int idx_hash(u32 index)
337{
338 return __idx_hash(index, xfrm_idx_hmask);
339}
340
341static struct hlist_head *policy_hash_bysel(struct xfrm_selector *sel, unsigned short family, int dir)
342{
343 unsigned int hmask = xfrm_policy_bydst[dir].hmask;
344 unsigned int hash = __sel_hash(sel, family, hmask);
345
346 return (hash == hmask + 1 ?
347 &xfrm_policy_inexact[dir] :
348 xfrm_policy_bydst[dir].table + hash);
349}
350
351static struct hlist_head *policy_hash_direct(xfrm_address_t *daddr, xfrm_address_t *saddr, unsigned short family, int dir)
352{
353 unsigned int hmask = xfrm_policy_bydst[dir].hmask;
354 unsigned int hash = __addr_hash(daddr, saddr, family, hmask);
355
356 return xfrm_policy_bydst[dir].table + hash;
357}
358
359static void xfrm_dst_hash_transfer(struct hlist_head *list,
360 struct hlist_head *ndsttable,
361 unsigned int nhashmask)
362{
363 struct hlist_node *entry, *tmp, *entry0 = NULL;
364 struct xfrm_policy *pol;
365 unsigned int h0 = 0;
366
367redo:
368 hlist_for_each_entry_safe(pol, entry, tmp, list, bydst) {
369 unsigned int h;
370
371 h = __addr_hash(&pol->selector.daddr, &pol->selector.saddr,
372 pol->family, nhashmask);
373 if (!entry0) {
374 hlist_del(entry);
375 hlist_add_head(&pol->bydst, ndsttable+h);
376 h0 = h;
377 } else {
378 if (h != h0)
379 continue;
380 hlist_del(entry);
381 hlist_add_after(entry0, &pol->bydst);
382 }
383 entry0 = entry;
384 }
385 if (!hlist_empty(list)) {
386 entry0 = NULL;
387 goto redo;
388 }
389}
390
391static void xfrm_idx_hash_transfer(struct hlist_head *list,
392 struct hlist_head *nidxtable,
393 unsigned int nhashmask)
394{
395 struct hlist_node *entry, *tmp;
396 struct xfrm_policy *pol;
397
398 hlist_for_each_entry_safe(pol, entry, tmp, list, byidx) {
399 unsigned int h;
400
401 h = __idx_hash(pol->index, nhashmask);
402 hlist_add_head(&pol->byidx, nidxtable+h);
403 }
404}
405
406static unsigned long xfrm_new_hash_mask(unsigned int old_hmask)
407{
408 return ((old_hmask + 1) << 1) - 1;
409}
410
411static void xfrm_bydst_resize(int dir)
412{
413 unsigned int hmask = xfrm_policy_bydst[dir].hmask;
414 unsigned int nhashmask = xfrm_new_hash_mask(hmask);
415 unsigned int nsize = (nhashmask + 1) * sizeof(struct hlist_head);
416 struct hlist_head *odst = xfrm_policy_bydst[dir].table;
417 struct hlist_head *ndst = xfrm_hash_alloc(nsize);
418 int i;
419
420 if (!ndst)
421 return;
422
423 write_lock_bh(&xfrm_policy_lock);
424
425 for (i = hmask; i >= 0; i--)
426 xfrm_dst_hash_transfer(odst + i, ndst, nhashmask);
427
428 xfrm_policy_bydst[dir].table = ndst;
429 xfrm_policy_bydst[dir].hmask = nhashmask;
430
431 write_unlock_bh(&xfrm_policy_lock);
432
433 xfrm_hash_free(odst, (hmask + 1) * sizeof(struct hlist_head));
434}
435
436static void xfrm_byidx_resize(int total)
437{
438 unsigned int hmask = xfrm_idx_hmask;
439 unsigned int nhashmask = xfrm_new_hash_mask(hmask);
440 unsigned int nsize = (nhashmask + 1) * sizeof(struct hlist_head);
441 struct hlist_head *oidx = xfrm_policy_byidx;
442 struct hlist_head *nidx = xfrm_hash_alloc(nsize);
443 int i;
444
445 if (!nidx)
446 return;
447
448 write_lock_bh(&xfrm_policy_lock);
449
450 for (i = hmask; i >= 0; i--)
451 xfrm_idx_hash_transfer(oidx + i, nidx, nhashmask);
452
453 xfrm_policy_byidx = nidx;
454 xfrm_idx_hmask = nhashmask;
455
456 write_unlock_bh(&xfrm_policy_lock);
457
458 xfrm_hash_free(oidx, (hmask + 1) * sizeof(struct hlist_head));
459}
460
461static inline int xfrm_bydst_should_resize(int dir, int *total)
462{
463 unsigned int cnt = xfrm_policy_count[dir];
464 unsigned int hmask = xfrm_policy_bydst[dir].hmask;
465
466 if (total)
467 *total += cnt;
468
469 if ((hmask + 1) < xfrm_policy_hashmax &&
470 cnt > hmask)
471 return 1;
472
473 return 0;
474}
475
476static inline int xfrm_byidx_should_resize(int total)
477{
478 unsigned int hmask = xfrm_idx_hmask;
479
480 if ((hmask + 1) < xfrm_policy_hashmax &&
481 total > hmask)
482 return 1;
483
484 return 0;
485}
486
487void xfrm_spd_getinfo(struct xfrmk_spdinfo *si)
488{
489 read_lock_bh(&xfrm_policy_lock);
490 si->incnt = xfrm_policy_count[XFRM_POLICY_IN];
491 si->outcnt = xfrm_policy_count[XFRM_POLICY_OUT];
492 si->fwdcnt = xfrm_policy_count[XFRM_POLICY_FWD];
493 si->inscnt = xfrm_policy_count[XFRM_POLICY_IN+XFRM_POLICY_MAX];
494 si->outscnt = xfrm_policy_count[XFRM_POLICY_OUT+XFRM_POLICY_MAX];
495 si->fwdscnt = xfrm_policy_count[XFRM_POLICY_FWD+XFRM_POLICY_MAX];
496 si->spdhcnt = xfrm_idx_hmask;
497 si->spdhmcnt = xfrm_policy_hashmax;
498 read_unlock_bh(&xfrm_policy_lock);
499}
500EXPORT_SYMBOL(xfrm_spd_getinfo);
501
502static DEFINE_MUTEX(hash_resize_mutex);
503static void xfrm_hash_resize(struct work_struct *__unused)
504{
505 int dir, total;
506
507 mutex_lock(&hash_resize_mutex);
508
509 total = 0;
510 for (dir = 0; dir < XFRM_POLICY_MAX * 2; dir++) {
511 if (xfrm_bydst_should_resize(dir, &total))
512 xfrm_bydst_resize(dir);
513 }
514 if (xfrm_byidx_should_resize(total))
515 xfrm_byidx_resize(total);
516
517 mutex_unlock(&hash_resize_mutex);
518}
519
520static DECLARE_WORK(xfrm_hash_work, xfrm_hash_resize);
521
522
523
524static u32 xfrm_gen_index(u8 type, int dir)
525{
526 static u32 idx_generator;
527
528 for (;;) {
529 struct hlist_node *entry;
530 struct hlist_head *list;
531 struct xfrm_policy *p;
532 u32 idx;
533 int found;
534
535 idx = (idx_generator | dir);
536 idx_generator += 8;
537 if (idx == 0)
538 idx = 8;
539 list = xfrm_policy_byidx + idx_hash(idx);
540 found = 0;
541 hlist_for_each_entry(p, entry, list, byidx) {
542 if (p->index == idx) {
543 found = 1;
544 break;
545 }
546 }
547 if (!found)
548 return idx;
549 }
550}
551
552static inline int selector_cmp(struct xfrm_selector *s1, struct xfrm_selector *s2)
553{
554 u32 *p1 = (u32 *) s1;
555 u32 *p2 = (u32 *) s2;
556 int len = sizeof(struct xfrm_selector) / sizeof(u32);
557 int i;
558
559 for (i = 0; i < len; i++) {
560 if (p1[i] != p2[i])
561 return 1;
562 }
563
564 return 0;
565}
566
567int xfrm_policy_insert(int dir, struct xfrm_policy *policy, int excl)
568{
569 struct xfrm_policy *pol;
570 struct xfrm_policy *delpol;
571 struct hlist_head *chain;
572 struct hlist_node *entry, *newpos;
573 struct dst_entry *gc_list;
574
575 write_lock_bh(&xfrm_policy_lock);
576 chain = policy_hash_bysel(&policy->selector, policy->family, dir);
577 delpol = NULL;
578 newpos = NULL;
579 hlist_for_each_entry(pol, entry, chain, bydst) {
580 if (pol->type == policy->type &&
581 !selector_cmp(&pol->selector, &policy->selector) &&
582 xfrm_sec_ctx_match(pol->security, policy->security) &&
583 !WARN_ON(delpol)) {
584 if (excl) {
585 write_unlock_bh(&xfrm_policy_lock);
586 return -EEXIST;
587 }
588 delpol = pol;
589 if (policy->priority > pol->priority)
590 continue;
591 } else if (policy->priority >= pol->priority) {
592 newpos = &pol->bydst;
593 continue;
594 }
595 if (delpol)
596 break;
597 }
598 if (newpos)
599 hlist_add_after(newpos, &policy->bydst);
600 else
601 hlist_add_head(&policy->bydst, chain);
602 xfrm_pol_hold(policy);
603 xfrm_policy_count[dir]++;
604 atomic_inc(&flow_cache_genid);
605 if (delpol) {
606 hlist_del(&delpol->bydst);
607 hlist_del(&delpol->byidx);
608 list_del(&delpol->walk.all);
609 xfrm_policy_count[dir]--;
610 }
611 policy->index = delpol ? delpol->index : xfrm_gen_index(policy->type, dir);
612 hlist_add_head(&policy->byidx, xfrm_policy_byidx+idx_hash(policy->index));
613 policy->curlft.add_time = get_seconds();
614 policy->curlft.use_time = 0;
615 if (!mod_timer(&policy->timer, jiffies + HZ))
616 xfrm_pol_hold(policy);
617 list_add(&policy->walk.all, &xfrm_policy_all);
618 write_unlock_bh(&xfrm_policy_lock);
619
620 if (delpol)
621 xfrm_policy_kill(delpol);
622 else if (xfrm_bydst_should_resize(dir, NULL))
623 schedule_work(&xfrm_hash_work);
624
625 read_lock_bh(&xfrm_policy_lock);
626 gc_list = NULL;
627 entry = &policy->bydst;
628 hlist_for_each_entry_continue(policy, entry, bydst) {
629 struct dst_entry *dst;
630
631 write_lock(&policy->lock);
632 dst = policy->bundles;
633 if (dst) {
634 struct dst_entry *tail = dst;
635 while (tail->next)
636 tail = tail->next;
637 tail->next = gc_list;
638 gc_list = dst;
639
640 policy->bundles = NULL;
641 }
642 write_unlock(&policy->lock);
643 }
644 read_unlock_bh(&xfrm_policy_lock);
645
646 while (gc_list) {
647 struct dst_entry *dst = gc_list;
648
649 gc_list = dst->next;
650 dst_free(dst);
651 }
652
653 return 0;
654}
655EXPORT_SYMBOL(xfrm_policy_insert);
656
657struct xfrm_policy *xfrm_policy_bysel_ctx(u8 type, int dir,
658 struct xfrm_selector *sel,
659 struct xfrm_sec_ctx *ctx, int delete,
660 int *err)
661{
662 struct xfrm_policy *pol, *ret;
663 struct hlist_head *chain;
664 struct hlist_node *entry;
665
666 *err = 0;
667 write_lock_bh(&xfrm_policy_lock);
668 chain = policy_hash_bysel(sel, sel->family, dir);
669 ret = NULL;
670 hlist_for_each_entry(pol, entry, chain, bydst) {
671 if (pol->type == type &&
672 !selector_cmp(sel, &pol->selector) &&
673 xfrm_sec_ctx_match(ctx, pol->security)) {
674 xfrm_pol_hold(pol);
675 if (delete) {
676 *err = security_xfrm_policy_delete(
677 pol->security);
678 if (*err) {
679 write_unlock_bh(&xfrm_policy_lock);
680 return pol;
681 }
682 hlist_del(&pol->bydst);
683 hlist_del(&pol->byidx);
684 list_del(&pol->walk.all);
685 xfrm_policy_count[dir]--;
686 }
687 ret = pol;
688 break;
689 }
690 }
691 write_unlock_bh(&xfrm_policy_lock);
692
693 if (ret && delete) {
694 atomic_inc(&flow_cache_genid);
695 xfrm_policy_kill(ret);
696 }
697 return ret;
698}
699EXPORT_SYMBOL(xfrm_policy_bysel_ctx);
700
701struct xfrm_policy *xfrm_policy_byid(u8 type, int dir, u32 id, int delete,
702 int *err)
703{
704 struct xfrm_policy *pol, *ret;
705 struct hlist_head *chain;
706 struct hlist_node *entry;
707
708 *err = -ENOENT;
709 if (xfrm_policy_id2dir(id) != dir)
710 return NULL;
711
712 *err = 0;
713 write_lock_bh(&xfrm_policy_lock);
714 chain = xfrm_policy_byidx + idx_hash(id);
715 ret = NULL;
716 hlist_for_each_entry(pol, entry, chain, byidx) {
717 if (pol->type == type && pol->index == id) {
718 xfrm_pol_hold(pol);
719 if (delete) {
720 *err = security_xfrm_policy_delete(
721 pol->security);
722 if (*err) {
723 write_unlock_bh(&xfrm_policy_lock);
724 return pol;
725 }
726 hlist_del(&pol->bydst);
727 hlist_del(&pol->byidx);
728 list_del(&pol->walk.all);
729 xfrm_policy_count[dir]--;
730 }
731 ret = pol;
732 break;
733 }
734 }
735 write_unlock_bh(&xfrm_policy_lock);
736
737 if (ret && delete) {
738 atomic_inc(&flow_cache_genid);
739 xfrm_policy_kill(ret);
740 }
741 return ret;
742}
743EXPORT_SYMBOL(xfrm_policy_byid);
744
745#ifdef CONFIG_SECURITY_NETWORK_XFRM
746static inline int
747xfrm_policy_flush_secctx_check(u8 type, struct xfrm_audit *audit_info)
748{
749 int dir, err = 0;
750
751 for (dir = 0; dir < XFRM_POLICY_MAX; dir++) {
752 struct xfrm_policy *pol;
753 struct hlist_node *entry;
754 int i;
755
756 hlist_for_each_entry(pol, entry,
757 &xfrm_policy_inexact[dir], bydst) {
758 if (pol->type != type)
759 continue;
760 err = security_xfrm_policy_delete(pol->security);
761 if (err) {
762 xfrm_audit_policy_delete(pol, 0,
763 audit_info->loginuid,
764 audit_info->sessionid,
765 audit_info->secid);
766 return err;
767 }
768 }
769 for (i = xfrm_policy_bydst[dir].hmask; i >= 0; i--) {
770 hlist_for_each_entry(pol, entry,
771 xfrm_policy_bydst[dir].table + i,
772 bydst) {
773 if (pol->type != type)
774 continue;
775 err = security_xfrm_policy_delete(
776 pol->security);
777 if (err) {
778 xfrm_audit_policy_delete(pol, 0,
779 audit_info->loginuid,
780 audit_info->sessionid,
781 audit_info->secid);
782 return err;
783 }
784 }
785 }
786 }
787 return err;
788}
789#else
790static inline int
791xfrm_policy_flush_secctx_check(u8 type, struct xfrm_audit *audit_info)
792{
793 return 0;
794}
795#endif
796
797int xfrm_policy_flush(u8 type, struct xfrm_audit *audit_info)
798{
799 int dir, err = 0;
800
801 write_lock_bh(&xfrm_policy_lock);
802
803 err = xfrm_policy_flush_secctx_check(type, audit_info);
804 if (err)
805 goto out;
806
807 for (dir = 0; dir < XFRM_POLICY_MAX; dir++) {
808 struct xfrm_policy *pol;
809 struct hlist_node *entry;
810 int i, killed;
811
812 killed = 0;
813 again1:
814 hlist_for_each_entry(pol, entry,
815 &xfrm_policy_inexact[dir], bydst) {
816 if (pol->type != type)
817 continue;
818 hlist_del(&pol->bydst);
819 hlist_del(&pol->byidx);
820 list_del(&pol->walk.all);
821 write_unlock_bh(&xfrm_policy_lock);
822
823 xfrm_audit_policy_delete(pol, 1, audit_info->loginuid,
824 audit_info->sessionid,
825 audit_info->secid);
826
827 xfrm_policy_kill(pol);
828 killed++;
829
830 write_lock_bh(&xfrm_policy_lock);
831 goto again1;
832 }
833
834 for (i = xfrm_policy_bydst[dir].hmask; i >= 0; i--) {
835 again2:
836 hlist_for_each_entry(pol, entry,
837 xfrm_policy_bydst[dir].table + i,
838 bydst) {
839 if (pol->type != type)
840 continue;
841 hlist_del(&pol->bydst);
842 hlist_del(&pol->byidx);
843 list_del(&pol->walk.all);
844 write_unlock_bh(&xfrm_policy_lock);
845
846 xfrm_audit_policy_delete(pol, 1,
847 audit_info->loginuid,
848 audit_info->sessionid,
849 audit_info->secid);
850 xfrm_policy_kill(pol);
851 killed++;
852
853 write_lock_bh(&xfrm_policy_lock);
854 goto again2;
855 }
856 }
857
858 xfrm_policy_count[dir] -= killed;
859 }
860 atomic_inc(&flow_cache_genid);
861out:
862 write_unlock_bh(&xfrm_policy_lock);
863 return err;
864}
865EXPORT_SYMBOL(xfrm_policy_flush);
866
867int xfrm_policy_walk(struct xfrm_policy_walk *walk,
868 int (*func)(struct xfrm_policy *, int, int, void*),
869 void *data)
870{
871 struct xfrm_policy *pol;
872 struct xfrm_policy_walk_entry *x;
873 int error = 0;
874
875 if (walk->type >= XFRM_POLICY_TYPE_MAX &&
876 walk->type != XFRM_POLICY_TYPE_ANY)
877 return -EINVAL;
878
879 if (list_empty(&walk->walk.all) && walk->seq != 0)
880 return 0;
881
882 write_lock_bh(&xfrm_policy_lock);
883 if (list_empty(&walk->walk.all))
884 x = list_first_entry(&xfrm_policy_all, struct xfrm_policy_walk_entry, all);
885 else
886 x = list_entry(&walk->walk.all, struct xfrm_policy_walk_entry, all);
887 list_for_each_entry_from(x, &xfrm_policy_all, all) {
888 if (x->dead)
889 continue;
890 pol = container_of(x, struct xfrm_policy, walk);
891 if (walk->type != XFRM_POLICY_TYPE_ANY &&
892 walk->type != pol->type)
893 continue;
894 error = func(pol, xfrm_policy_id2dir(pol->index),
895 walk->seq, data);
896 if (error) {
897 list_move_tail(&walk->walk.all, &x->all);
898 goto out;
899 }
900 walk->seq++;
901 }
902 if (walk->seq == 0) {
903 error = -ENOENT;
904 goto out;
905 }
906 list_del_init(&walk->walk.all);
907out:
908 write_unlock_bh(&xfrm_policy_lock);
909 return error;
910}
911EXPORT_SYMBOL(xfrm_policy_walk);
912
913void xfrm_policy_walk_init(struct xfrm_policy_walk *walk, u8 type)
914{
915 INIT_LIST_HEAD(&walk->walk.all);
916 walk->walk.dead = 1;
917 walk->type = type;
918 walk->seq = 0;
919}
920EXPORT_SYMBOL(xfrm_policy_walk_init);
921
922void xfrm_policy_walk_done(struct xfrm_policy_walk *walk)
923{
924 if (list_empty(&walk->walk.all))
925 return;
926
927 write_lock_bh(&xfrm_policy_lock);
928 list_del(&walk->walk.all);
929 write_unlock_bh(&xfrm_policy_lock);
930}
931EXPORT_SYMBOL(xfrm_policy_walk_done);
932
933
934
935
936
937
938static int xfrm_policy_match(struct xfrm_policy *pol, struct flowi *fl,
939 u8 type, u16 family, int dir)
940{
941 struct xfrm_selector *sel = &pol->selector;
942 int match, ret = -ESRCH;
943
944 if (pol->family != family ||
945 pol->type != type)
946 return ret;
947
948 match = xfrm_selector_match(sel, fl, family);
949 if (match)
950 ret = security_xfrm_policy_lookup(pol->security, fl->secid,
951 dir);
952
953 return ret;
954}
955
956static struct xfrm_policy *xfrm_policy_lookup_bytype(u8 type, struct flowi *fl,
957 u16 family, u8 dir)
958{
959 int err;
960 struct xfrm_policy *pol, *ret;
961 xfrm_address_t *daddr, *saddr;
962 struct hlist_node *entry;
963 struct hlist_head *chain;
964 u32 priority = ~0U;
965
966 daddr = xfrm_flowi_daddr(fl, family);
967 saddr = xfrm_flowi_saddr(fl, family);
968 if (unlikely(!daddr || !saddr))
969 return NULL;
970
971 read_lock_bh(&xfrm_policy_lock);
972 chain = policy_hash_direct(daddr, saddr, family, dir);
973 ret = NULL;
974 hlist_for_each_entry(pol, entry, chain, bydst) {
975 err = xfrm_policy_match(pol, fl, type, family, dir);
976 if (err) {
977 if (err == -ESRCH)
978 continue;
979 else {
980 ret = ERR_PTR(err);
981 goto fail;
982 }
983 } else {
984 ret = pol;
985 priority = ret->priority;
986 break;
987 }
988 }
989 chain = &xfrm_policy_inexact[dir];
990 hlist_for_each_entry(pol, entry, chain, bydst) {
991 err = xfrm_policy_match(pol, fl, type, family, dir);
992 if (err) {
993 if (err == -ESRCH)
994 continue;
995 else {
996 ret = ERR_PTR(err);
997 goto fail;
998 }
999 } else if (pol->priority < priority) {
1000 ret = pol;
1001 break;
1002 }
1003 }
1004 if (ret)
1005 xfrm_pol_hold(ret);
1006fail:
1007 read_unlock_bh(&xfrm_policy_lock);
1008
1009 return ret;
1010}
1011
1012static int xfrm_policy_lookup(struct flowi *fl, u16 family, u8 dir,
1013 void **objp, atomic_t **obj_refp)
1014{
1015 struct xfrm_policy *pol;
1016 int err = 0;
1017
1018#ifdef CONFIG_XFRM_SUB_POLICY
1019 pol = xfrm_policy_lookup_bytype(XFRM_POLICY_TYPE_SUB, fl, family, dir);
1020 if (IS_ERR(pol)) {
1021 err = PTR_ERR(pol);
1022 pol = NULL;
1023 }
1024 if (pol || err)
1025 goto end;
1026#endif
1027 pol = xfrm_policy_lookup_bytype(XFRM_POLICY_TYPE_MAIN, fl, family, dir);
1028 if (IS_ERR(pol)) {
1029 err = PTR_ERR(pol);
1030 pol = NULL;
1031 }
1032#ifdef CONFIG_XFRM_SUB_POLICY
1033end:
1034#endif
1035 if ((*objp = (void *) pol) != NULL)
1036 *obj_refp = &pol->refcnt;
1037 return err;
1038}
1039
1040static inline int policy_to_flow_dir(int dir)
1041{
1042 if (XFRM_POLICY_IN == FLOW_DIR_IN &&
1043 XFRM_POLICY_OUT == FLOW_DIR_OUT &&
1044 XFRM_POLICY_FWD == FLOW_DIR_FWD)
1045 return dir;
1046 switch (dir) {
1047 default:
1048 case XFRM_POLICY_IN:
1049 return FLOW_DIR_IN;
1050 case XFRM_POLICY_OUT:
1051 return FLOW_DIR_OUT;
1052 case XFRM_POLICY_FWD:
1053 return FLOW_DIR_FWD;
1054 }
1055}
1056
1057static struct xfrm_policy *xfrm_sk_policy_lookup(struct sock *sk, int dir, struct flowi *fl)
1058{
1059 struct xfrm_policy *pol;
1060
1061 read_lock_bh(&xfrm_policy_lock);
1062 if ((pol = sk->sk_policy[dir]) != NULL) {
1063 int match = xfrm_selector_match(&pol->selector, fl,
1064 sk->sk_family);
1065 int err = 0;
1066
1067 if (match) {
1068 err = security_xfrm_policy_lookup(pol->security,
1069 fl->secid,
1070 policy_to_flow_dir(dir));
1071 if (!err)
1072 xfrm_pol_hold(pol);
1073 else if (err == -ESRCH)
1074 pol = NULL;
1075 else
1076 pol = ERR_PTR(err);
1077 } else
1078 pol = NULL;
1079 }
1080 read_unlock_bh(&xfrm_policy_lock);
1081 return pol;
1082}
1083
1084static void __xfrm_policy_link(struct xfrm_policy *pol, int dir)
1085{
1086 struct hlist_head *chain = policy_hash_bysel(&pol->selector,
1087 pol->family, dir);
1088
1089 list_add(&pol->walk.all, &xfrm_policy_all);
1090 hlist_add_head(&pol->bydst, chain);
1091 hlist_add_head(&pol->byidx, xfrm_policy_byidx+idx_hash(pol->index));
1092 xfrm_policy_count[dir]++;
1093 xfrm_pol_hold(pol);
1094
1095 if (xfrm_bydst_should_resize(dir, NULL))
1096 schedule_work(&xfrm_hash_work);
1097}
1098
1099static struct xfrm_policy *__xfrm_policy_unlink(struct xfrm_policy *pol,
1100 int dir)
1101{
1102 if (hlist_unhashed(&pol->bydst))
1103 return NULL;
1104
1105 hlist_del(&pol->bydst);
1106 hlist_del(&pol->byidx);
1107 list_del(&pol->walk.all);
1108 xfrm_policy_count[dir]--;
1109
1110 return pol;
1111}
1112
1113int xfrm_policy_delete(struct xfrm_policy *pol, int dir)
1114{
1115 write_lock_bh(&xfrm_policy_lock);
1116 pol = __xfrm_policy_unlink(pol, dir);
1117 write_unlock_bh(&xfrm_policy_lock);
1118 if (pol) {
1119 if (dir < XFRM_POLICY_MAX)
1120 atomic_inc(&flow_cache_genid);
1121 xfrm_policy_kill(pol);
1122 return 0;
1123 }
1124 return -ENOENT;
1125}
1126EXPORT_SYMBOL(xfrm_policy_delete);
1127
1128int xfrm_sk_policy_insert(struct sock *sk, int dir, struct xfrm_policy *pol)
1129{
1130 struct xfrm_policy *old_pol;
1131
1132#ifdef CONFIG_XFRM_SUB_POLICY
1133 if (pol && pol->type != XFRM_POLICY_TYPE_MAIN)
1134 return -EINVAL;
1135#endif
1136
1137 write_lock_bh(&xfrm_policy_lock);
1138 old_pol = sk->sk_policy[dir];
1139 sk->sk_policy[dir] = pol;
1140 if (pol) {
1141 pol->curlft.add_time = get_seconds();
1142 pol->index = xfrm_gen_index(pol->type, XFRM_POLICY_MAX+dir);
1143 __xfrm_policy_link(pol, XFRM_POLICY_MAX+dir);
1144 }
1145 if (old_pol)
1146 __xfrm_policy_unlink(old_pol, XFRM_POLICY_MAX+dir);
1147 write_unlock_bh(&xfrm_policy_lock);
1148
1149 if (old_pol) {
1150 xfrm_policy_kill(old_pol);
1151 }
1152 return 0;
1153}
1154
1155static struct xfrm_policy *clone_policy(struct xfrm_policy *old, int dir)
1156{
1157 struct xfrm_policy *newp = xfrm_policy_alloc(GFP_ATOMIC);
1158
1159 if (newp) {
1160 newp->selector = old->selector;
1161 if (security_xfrm_policy_clone(old->security,
1162 &newp->security)) {
1163 kfree(newp);
1164 return NULL;
1165 }
1166 newp->lft = old->lft;
1167 newp->curlft = old->curlft;
1168 newp->action = old->action;
1169 newp->flags = old->flags;
1170 newp->xfrm_nr = old->xfrm_nr;
1171 newp->index = old->index;
1172 newp->type = old->type;
1173 memcpy(newp->xfrm_vec, old->xfrm_vec,
1174 newp->xfrm_nr*sizeof(struct xfrm_tmpl));
1175 write_lock_bh(&xfrm_policy_lock);
1176 __xfrm_policy_link(newp, XFRM_POLICY_MAX+dir);
1177 write_unlock_bh(&xfrm_policy_lock);
1178 xfrm_pol_put(newp);
1179 }
1180 return newp;
1181}
1182
1183int __xfrm_sk_clone_policy(struct sock *sk)
1184{
1185 struct xfrm_policy *p0 = sk->sk_policy[0],
1186 *p1 = sk->sk_policy[1];
1187
1188 sk->sk_policy[0] = sk->sk_policy[1] = NULL;
1189 if (p0 && (sk->sk_policy[0] = clone_policy(p0, 0)) == NULL)
1190 return -ENOMEM;
1191 if (p1 && (sk->sk_policy[1] = clone_policy(p1, 1)) == NULL)
1192 return -ENOMEM;
1193 return 0;
1194}
1195
1196static int
1197xfrm_get_saddr(xfrm_address_t *local, xfrm_address_t *remote,
1198 unsigned short family)
1199{
1200 int err;
1201 struct xfrm_policy_afinfo *afinfo = xfrm_policy_get_afinfo(family);
1202
1203 if (unlikely(afinfo == NULL))
1204 return -EINVAL;
1205 err = afinfo->get_saddr(local, remote);
1206 xfrm_policy_put_afinfo(afinfo);
1207 return err;
1208}
1209
1210
1211
1212static int
1213xfrm_tmpl_resolve_one(struct xfrm_policy *policy, struct flowi *fl,
1214 struct xfrm_state **xfrm,
1215 unsigned short family)
1216{
1217 int nx;
1218 int i, error;
1219 xfrm_address_t *daddr = xfrm_flowi_daddr(fl, family);
1220 xfrm_address_t *saddr = xfrm_flowi_saddr(fl, family);
1221 xfrm_address_t tmp;
1222
1223 for (nx=0, i = 0; i < policy->xfrm_nr; i++) {
1224 struct xfrm_state *x;
1225 xfrm_address_t *remote = daddr;
1226 xfrm_address_t *local = saddr;
1227 struct xfrm_tmpl *tmpl = &policy->xfrm_vec[i];
1228
1229 if (tmpl->mode == XFRM_MODE_TUNNEL ||
1230 tmpl->mode == XFRM_MODE_BEET) {
1231 remote = &tmpl->id.daddr;
1232 local = &tmpl->saddr;
1233 family = tmpl->encap_family;
1234 if (xfrm_addr_any(local, family)) {
1235 error = xfrm_get_saddr(&tmp, remote, family);
1236 if (error)
1237 goto fail;
1238 local = &tmp;
1239 }
1240 }
1241
1242 x = xfrm_state_find(remote, local, fl, tmpl, policy, &error, family);
1243
1244 if (x && x->km.state == XFRM_STATE_VALID) {
1245 xfrm[nx++] = x;
1246 daddr = remote;
1247 saddr = local;
1248 continue;
1249 }
1250 if (x) {
1251 error = (x->km.state == XFRM_STATE_ERROR ?
1252 -EINVAL : -EAGAIN);
1253 xfrm_state_put(x);
1254 }
1255 else if (error == -ESRCH)
1256 error = -EAGAIN;
1257
1258 if (!tmpl->optional)
1259 goto fail;
1260 }
1261 return nx;
1262
1263fail:
1264 for (nx--; nx>=0; nx--)
1265 xfrm_state_put(xfrm[nx]);
1266 return error;
1267}
1268
1269static int
1270xfrm_tmpl_resolve(struct xfrm_policy **pols, int npols, struct flowi *fl,
1271 struct xfrm_state **xfrm,
1272 unsigned short family)
1273{
1274 struct xfrm_state *tp[XFRM_MAX_DEPTH];
1275 struct xfrm_state **tpp = (npols > 1) ? tp : xfrm;
1276 int cnx = 0;
1277 int error;
1278 int ret;
1279 int i;
1280
1281 for (i = 0; i < npols; i++) {
1282 if (cnx + pols[i]->xfrm_nr >= XFRM_MAX_DEPTH) {
1283 error = -ENOBUFS;
1284 goto fail;
1285 }
1286
1287 ret = xfrm_tmpl_resolve_one(pols[i], fl, &tpp[cnx], family);
1288 if (ret < 0) {
1289 error = ret;
1290 goto fail;
1291 } else
1292 cnx += ret;
1293 }
1294
1295
1296 if (npols > 1)
1297 xfrm_state_sort(xfrm, tpp, cnx, family);
1298
1299 return cnx;
1300
1301 fail:
1302 for (cnx--; cnx>=0; cnx--)
1303 xfrm_state_put(tpp[cnx]);
1304 return error;
1305
1306}
1307
1308
1309
1310
1311
1312static struct dst_entry *
1313xfrm_find_bundle(struct flowi *fl, struct xfrm_policy *policy, unsigned short family)
1314{
1315 struct dst_entry *x;
1316 struct xfrm_policy_afinfo *afinfo = xfrm_policy_get_afinfo(family);
1317 if (unlikely(afinfo == NULL))
1318 return ERR_PTR(-EINVAL);
1319 x = afinfo->find_bundle(fl, policy);
1320 xfrm_policy_put_afinfo(afinfo);
1321 return x;
1322}
1323
1324static inline int xfrm_get_tos(struct flowi *fl, int family)
1325{
1326 struct xfrm_policy_afinfo *afinfo = xfrm_policy_get_afinfo(family);
1327 int tos;
1328
1329 if (!afinfo)
1330 return -EINVAL;
1331
1332 tos = afinfo->get_tos(fl);
1333
1334 xfrm_policy_put_afinfo(afinfo);
1335
1336 return tos;
1337}
1338
1339static inline struct xfrm_dst *xfrm_alloc_dst(int family)
1340{
1341 struct xfrm_policy_afinfo *afinfo = xfrm_policy_get_afinfo(family);
1342 struct xfrm_dst *xdst;
1343
1344 if (!afinfo)
1345 return ERR_PTR(-EINVAL);
1346
1347 xdst = dst_alloc(afinfo->dst_ops) ?: ERR_PTR(-ENOBUFS);
1348
1349 xfrm_policy_put_afinfo(afinfo);
1350
1351 return xdst;
1352}
1353
1354static inline int xfrm_init_path(struct xfrm_dst *path, struct dst_entry *dst,
1355 int nfheader_len)
1356{
1357 struct xfrm_policy_afinfo *afinfo =
1358 xfrm_policy_get_afinfo(dst->ops->family);
1359 int err;
1360
1361 if (!afinfo)
1362 return -EINVAL;
1363
1364 err = afinfo->init_path(path, dst, nfheader_len);
1365
1366 xfrm_policy_put_afinfo(afinfo);
1367
1368 return err;
1369}
1370
1371static inline int xfrm_fill_dst(struct xfrm_dst *xdst, struct net_device *dev)
1372{
1373 struct xfrm_policy_afinfo *afinfo =
1374 xfrm_policy_get_afinfo(xdst->u.dst.ops->family);
1375 int err;
1376
1377 if (!afinfo)
1378 return -EINVAL;
1379
1380 err = afinfo->fill_dst(xdst, dev);
1381
1382 xfrm_policy_put_afinfo(afinfo);
1383
1384 return err;
1385}
1386
1387
1388
1389
1390
1391static struct dst_entry *xfrm_bundle_create(struct xfrm_policy *policy,
1392 struct xfrm_state **xfrm, int nx,
1393 struct flowi *fl,
1394 struct dst_entry *dst)
1395{
1396 unsigned long now = jiffies;
1397 struct net_device *dev;
1398 struct dst_entry *dst_prev = NULL;
1399 struct dst_entry *dst0 = NULL;
1400 int i = 0;
1401 int err;
1402 int header_len = 0;
1403 int nfheader_len = 0;
1404 int trailer_len = 0;
1405 int tos;
1406 int family = policy->selector.family;
1407 xfrm_address_t saddr, daddr;
1408
1409 xfrm_flowi_addr_get(fl, &saddr, &daddr, family);
1410
1411 tos = xfrm_get_tos(fl, family);
1412 err = tos;
1413 if (tos < 0)
1414 goto put_states;
1415
1416 dst_hold(dst);
1417
1418 for (; i < nx; i++) {
1419 struct xfrm_dst *xdst = xfrm_alloc_dst(family);
1420 struct dst_entry *dst1 = &xdst->u.dst;
1421
1422 err = PTR_ERR(xdst);
1423 if (IS_ERR(xdst)) {
1424 dst_release(dst);
1425 goto put_states;
1426 }
1427
1428 if (!dst_prev)
1429 dst0 = dst1;
1430 else {
1431 dst_prev->child = dst_clone(dst1);
1432 dst1->flags |= DST_NOHASH;
1433 }
1434
1435 xdst->route = dst;
1436 memcpy(&dst1->metrics, &dst->metrics, sizeof(dst->metrics));
1437
1438 if (xfrm[i]->props.mode != XFRM_MODE_TRANSPORT) {
1439 family = xfrm[i]->props.family;
1440 dst = xfrm_dst_lookup(xfrm[i], tos, &saddr, &daddr,
1441 family);
1442 err = PTR_ERR(dst);
1443 if (IS_ERR(dst))
1444 goto put_states;
1445 } else
1446 dst_hold(dst);
1447
1448 dst1->xfrm = xfrm[i];
1449 xdst->genid = xfrm[i]->genid;
1450
1451 dst1->obsolete = -1;
1452 dst1->flags |= DST_HOST;
1453 dst1->lastuse = now;
1454
1455 dst1->input = dst_discard;
1456 dst1->output = xfrm[i]->outer_mode->afinfo->output;
1457
1458 dst1->next = dst_prev;
1459 dst_prev = dst1;
1460
1461 header_len += xfrm[i]->props.header_len;
1462 if (xfrm[i]->type->flags & XFRM_TYPE_NON_FRAGMENT)
1463 nfheader_len += xfrm[i]->props.header_len;
1464 trailer_len += xfrm[i]->props.trailer_len;
1465 }
1466
1467 dst_prev->child = dst;
1468 dst0->path = dst;
1469
1470 err = -ENODEV;
1471 dev = dst->dev;
1472 if (!dev)
1473 goto free_dst;
1474
1475
1476 dst0->neighbour = neigh_clone(dst->neighbour);
1477
1478 xfrm_init_path((struct xfrm_dst *)dst0, dst, nfheader_len);
1479 xfrm_init_pmtu(dst_prev);
1480
1481 for (dst_prev = dst0; dst_prev != dst; dst_prev = dst_prev->child) {
1482 struct xfrm_dst *xdst = (struct xfrm_dst *)dst_prev;
1483
1484 err = xfrm_fill_dst(xdst, dev);
1485 if (err)
1486 goto free_dst;
1487
1488 dst_prev->header_len = header_len;
1489 dst_prev->trailer_len = trailer_len;
1490 header_len -= xdst->u.dst.xfrm->props.header_len;
1491 trailer_len -= xdst->u.dst.xfrm->props.trailer_len;
1492 }
1493
1494out:
1495 return dst0;
1496
1497put_states:
1498 for (; i < nx; i++)
1499 xfrm_state_put(xfrm[i]);
1500free_dst:
1501 if (dst0)
1502 dst_free(dst0);
1503 dst0 = ERR_PTR(err);
1504 goto out;
1505}
1506
1507static int inline
1508xfrm_dst_alloc_copy(void **target, void *src, int size)
1509{
1510 if (!*target) {
1511 *target = kmalloc(size, GFP_ATOMIC);
1512 if (!*target)
1513 return -ENOMEM;
1514 }
1515 memcpy(*target, src, size);
1516 return 0;
1517}
1518
1519static int inline
1520xfrm_dst_update_parent(struct dst_entry *dst, struct xfrm_selector *sel)
1521{
1522#ifdef CONFIG_XFRM_SUB_POLICY
1523 struct xfrm_dst *xdst = (struct xfrm_dst *)dst;
1524 return xfrm_dst_alloc_copy((void **)&(xdst->partner),
1525 sel, sizeof(*sel));
1526#else
1527 return 0;
1528#endif
1529}
1530
1531static int inline
1532xfrm_dst_update_origin(struct dst_entry *dst, struct flowi *fl)
1533{
1534#ifdef CONFIG_XFRM_SUB_POLICY
1535 struct xfrm_dst *xdst = (struct xfrm_dst *)dst;
1536 return xfrm_dst_alloc_copy((void **)&(xdst->origin), fl, sizeof(*fl));
1537#else
1538 return 0;
1539#endif
1540}
1541
1542static int stale_bundle(struct dst_entry *dst);
1543
1544
1545
1546
1547
1548
1549int __xfrm_lookup(struct dst_entry **dst_p, struct flowi *fl,
1550 struct sock *sk, int flags)
1551{
1552 struct xfrm_policy *policy;
1553 struct xfrm_policy *pols[XFRM_POLICY_TYPE_MAX];
1554 int npols;
1555 int pol_dead;
1556 int xfrm_nr;
1557 int pi;
1558 struct xfrm_state *xfrm[XFRM_MAX_DEPTH];
1559 struct dst_entry *dst, *dst_orig = *dst_p;
1560 int nx = 0;
1561 int err;
1562 u32 genid;
1563 u16 family;
1564 u8 dir = policy_to_flow_dir(XFRM_POLICY_OUT);
1565
1566restart:
1567 genid = atomic_read(&flow_cache_genid);
1568 policy = NULL;
1569 for (pi = 0; pi < ARRAY_SIZE(pols); pi++)
1570 pols[pi] = NULL;
1571 npols = 0;
1572 pol_dead = 0;
1573 xfrm_nr = 0;
1574
1575 if (sk && sk->sk_policy[XFRM_POLICY_OUT]) {
1576 policy = xfrm_sk_policy_lookup(sk, XFRM_POLICY_OUT, fl);
1577 err = PTR_ERR(policy);
1578 if (IS_ERR(policy)) {
1579 XFRM_INC_STATS(LINUX_MIB_XFRMOUTPOLERROR);
1580 goto dropdst;
1581 }
1582 }
1583
1584 if (!policy) {
1585
1586 if ((dst_orig->flags & DST_NOXFRM) ||
1587 !xfrm_policy_count[XFRM_POLICY_OUT])
1588 goto nopol;
1589
1590 policy = flow_cache_lookup(fl, dst_orig->ops->family,
1591 dir, xfrm_policy_lookup);
1592 err = PTR_ERR(policy);
1593 if (IS_ERR(policy)) {
1594 XFRM_INC_STATS(LINUX_MIB_XFRMOUTPOLERROR);
1595 goto dropdst;
1596 }
1597 }
1598
1599 if (!policy)
1600 goto nopol;
1601
1602 family = dst_orig->ops->family;
1603 pols[0] = policy;
1604 npols ++;
1605 xfrm_nr += pols[0]->xfrm_nr;
1606
1607 err = -ENOENT;
1608 if ((flags & XFRM_LOOKUP_ICMP) && !(policy->flags & XFRM_POLICY_ICMP))
1609 goto error;
1610
1611 policy->curlft.use_time = get_seconds();
1612
1613 switch (policy->action) {
1614 default:
1615 case XFRM_POLICY_BLOCK:
1616
1617 XFRM_INC_STATS(LINUX_MIB_XFRMOUTPOLBLOCK);
1618 err = -EPERM;
1619 goto error;
1620
1621 case XFRM_POLICY_ALLOW:
1622#ifndef CONFIG_XFRM_SUB_POLICY
1623 if (policy->xfrm_nr == 0) {
1624
1625 xfrm_pol_put(policy);
1626 return 0;
1627 }
1628#endif
1629
1630
1631
1632
1633
1634
1635 dst = xfrm_find_bundle(fl, policy, family);
1636 if (IS_ERR(dst)) {
1637 XFRM_INC_STATS(LINUX_MIB_XFRMOUTBUNDLECHECKERROR);
1638 err = PTR_ERR(dst);
1639 goto error;
1640 }
1641
1642 if (dst)
1643 break;
1644
1645#ifdef CONFIG_XFRM_SUB_POLICY
1646 if (pols[0]->type != XFRM_POLICY_TYPE_MAIN) {
1647 pols[1] = xfrm_policy_lookup_bytype(XFRM_POLICY_TYPE_MAIN,
1648 fl, family,
1649 XFRM_POLICY_OUT);
1650 if (pols[1]) {
1651 if (IS_ERR(pols[1])) {
1652 XFRM_INC_STATS(LINUX_MIB_XFRMOUTPOLERROR);
1653 err = PTR_ERR(pols[1]);
1654 goto error;
1655 }
1656 if (pols[1]->action == XFRM_POLICY_BLOCK) {
1657 XFRM_INC_STATS(LINUX_MIB_XFRMOUTPOLBLOCK);
1658 err = -EPERM;
1659 goto error;
1660 }
1661 npols ++;
1662 xfrm_nr += pols[1]->xfrm_nr;
1663 }
1664 }
1665
1666
1667
1668
1669
1670
1671
1672
1673 if (xfrm_nr == 0) {
1674
1675 xfrm_pols_put(pols, npols);
1676 return 0;
1677 }
1678
1679#endif
1680 nx = xfrm_tmpl_resolve(pols, npols, fl, xfrm, family);
1681
1682 if (unlikely(nx<0)) {
1683 err = nx;
1684 if (err == -EAGAIN && sysctl_xfrm_larval_drop) {
1685
1686
1687
1688 XFRM_INC_STATS(LINUX_MIB_XFRMOUTNOSTATES);
1689 xfrm_pol_put(policy);
1690 return -EREMOTE;
1691 }
1692 if (err == -EAGAIN && (flags & XFRM_LOOKUP_WAIT)) {
1693 DECLARE_WAITQUEUE(wait, current);
1694
1695 add_wait_queue(&km_waitq, &wait);
1696 set_current_state(TASK_INTERRUPTIBLE);
1697 schedule();
1698 set_current_state(TASK_RUNNING);
1699 remove_wait_queue(&km_waitq, &wait);
1700
1701 nx = xfrm_tmpl_resolve(pols, npols, fl, xfrm, family);
1702
1703 if (nx == -EAGAIN && signal_pending(current)) {
1704 XFRM_INC_STATS(LINUX_MIB_XFRMOUTNOSTATES);
1705 err = -ERESTART;
1706 goto error;
1707 }
1708 if (nx == -EAGAIN ||
1709 genid != atomic_read(&flow_cache_genid)) {
1710 xfrm_pols_put(pols, npols);
1711 goto restart;
1712 }
1713 err = nx;
1714 }
1715 if (err < 0) {
1716 XFRM_INC_STATS(LINUX_MIB_XFRMOUTNOSTATES);
1717 goto error;
1718 }
1719 }
1720 if (nx == 0) {
1721
1722 xfrm_pols_put(pols, npols);
1723 return 0;
1724 }
1725
1726 dst = xfrm_bundle_create(policy, xfrm, nx, fl, dst_orig);
1727 err = PTR_ERR(dst);
1728 if (IS_ERR(dst)) {
1729 XFRM_INC_STATS(LINUX_MIB_XFRMOUTBUNDLEGENERROR);
1730 goto error;
1731 }
1732
1733 for (pi = 0; pi < npols; pi++) {
1734 read_lock_bh(&pols[pi]->lock);
1735 pol_dead |= pols[pi]->walk.dead;
1736 read_unlock_bh(&pols[pi]->lock);
1737 }
1738
1739 write_lock_bh(&policy->lock);
1740 if (unlikely(pol_dead || stale_bundle(dst))) {
1741
1742
1743
1744
1745
1746 write_unlock_bh(&policy->lock);
1747 dst_free(dst);
1748
1749 if (pol_dead)
1750 XFRM_INC_STATS(LINUX_MIB_XFRMOUTPOLDEAD);
1751 else
1752 XFRM_INC_STATS(LINUX_MIB_XFRMOUTBUNDLECHECKERROR);
1753 err = -EHOSTUNREACH;
1754 goto error;
1755 }
1756
1757 if (npols > 1)
1758 err = xfrm_dst_update_parent(dst, &pols[1]->selector);
1759 else
1760 err = xfrm_dst_update_origin(dst, fl);
1761 if (unlikely(err)) {
1762 write_unlock_bh(&policy->lock);
1763 dst_free(dst);
1764 XFRM_INC_STATS(LINUX_MIB_XFRMOUTBUNDLECHECKERROR);
1765 goto error;
1766 }
1767
1768 dst->next = policy->bundles;
1769 policy->bundles = dst;
1770 dst_hold(dst);
1771 write_unlock_bh(&policy->lock);
1772 }
1773 *dst_p = dst;
1774 dst_release(dst_orig);
1775 xfrm_pols_put(pols, npols);
1776 return 0;
1777
1778error:
1779 xfrm_pols_put(pols, npols);
1780dropdst:
1781 dst_release(dst_orig);
1782 *dst_p = NULL;
1783 return err;
1784
1785nopol:
1786 err = -ENOENT;
1787 if (flags & XFRM_LOOKUP_ICMP)
1788 goto dropdst;
1789 return 0;
1790}
1791EXPORT_SYMBOL(__xfrm_lookup);
1792
1793int xfrm_lookup(struct dst_entry **dst_p, struct flowi *fl,
1794 struct sock *sk, int flags)
1795{
1796 int err = __xfrm_lookup(dst_p, fl, sk, flags);
1797
1798 if (err == -EREMOTE) {
1799 dst_release(*dst_p);
1800 *dst_p = NULL;
1801 err = -EAGAIN;
1802 }
1803
1804 return err;
1805}
1806EXPORT_SYMBOL(xfrm_lookup);
1807
1808static inline int
1809xfrm_secpath_reject(int idx, struct sk_buff *skb, struct flowi *fl)
1810{
1811 struct xfrm_state *x;
1812
1813 if (!skb->sp || idx < 0 || idx >= skb->sp->len)
1814 return 0;
1815 x = skb->sp->xvec[idx];
1816 if (!x->type->reject)
1817 return 0;
1818 return x->type->reject(x, skb, fl);
1819}
1820
1821
1822
1823
1824
1825
1826
1827static inline int
1828xfrm_state_ok(struct xfrm_tmpl *tmpl, struct xfrm_state *x,
1829 unsigned short family)
1830{
1831 if (xfrm_state_kern(x))
1832 return tmpl->optional && !xfrm_state_addr_cmp(tmpl, x, tmpl->encap_family);
1833 return x->id.proto == tmpl->id.proto &&
1834 (x->id.spi == tmpl->id.spi || !tmpl->id.spi) &&
1835 (x->props.reqid == tmpl->reqid || !tmpl->reqid) &&
1836 x->props.mode == tmpl->mode &&
1837 (tmpl->allalgs || (tmpl->aalgos & (1<<x->props.aalgo)) ||
1838 !(xfrm_id_proto_match(tmpl->id.proto, IPSEC_PROTO_ANY))) &&
1839 !(x->props.mode != XFRM_MODE_TRANSPORT &&
1840 xfrm_state_addr_cmp(tmpl, x, family));
1841}
1842
1843
1844
1845
1846
1847
1848
1849
1850static inline int
1851xfrm_policy_ok(struct xfrm_tmpl *tmpl, struct sec_path *sp, int start,
1852 unsigned short family)
1853{
1854 int idx = start;
1855
1856 if (tmpl->optional) {
1857 if (tmpl->mode == XFRM_MODE_TRANSPORT)
1858 return start;
1859 } else
1860 start = -1;
1861 for (; idx < sp->len; idx++) {
1862 if (xfrm_state_ok(tmpl, sp->xvec[idx], family))
1863 return ++idx;
1864 if (sp->xvec[idx]->props.mode != XFRM_MODE_TRANSPORT) {
1865 if (start == -1)
1866 start = -2-idx;
1867 break;
1868 }
1869 }
1870 return start;
1871}
1872
1873int __xfrm_decode_session(struct sk_buff *skb, struct flowi *fl,
1874 unsigned int family, int reverse)
1875{
1876 struct xfrm_policy_afinfo *afinfo = xfrm_policy_get_afinfo(family);
1877 int err;
1878
1879 if (unlikely(afinfo == NULL))
1880 return -EAFNOSUPPORT;
1881
1882 afinfo->decode_session(skb, fl, reverse);
1883 err = security_xfrm_decode_session(skb, &fl->secid);
1884 xfrm_policy_put_afinfo(afinfo);
1885 return err;
1886}
1887EXPORT_SYMBOL(__xfrm_decode_session);
1888
1889static inline int secpath_has_nontransport(struct sec_path *sp, int k, int *idxp)
1890{
1891 for (; k < sp->len; k++) {
1892 if (sp->xvec[k]->props.mode != XFRM_MODE_TRANSPORT) {
1893 *idxp = k;
1894 return 1;
1895 }
1896 }
1897
1898 return 0;
1899}
1900
1901int __xfrm_policy_check(struct sock *sk, int dir, struct sk_buff *skb,
1902 unsigned short family)
1903{
1904 struct xfrm_policy *pol;
1905 struct xfrm_policy *pols[XFRM_POLICY_TYPE_MAX];
1906 int npols = 0;
1907 int xfrm_nr;
1908 int pi;
1909 int reverse;
1910 struct flowi fl;
1911 u8 fl_dir;
1912 int xerr_idx = -1;
1913
1914 reverse = dir & ~XFRM_POLICY_MASK;
1915 dir &= XFRM_POLICY_MASK;
1916 fl_dir = policy_to_flow_dir(dir);
1917
1918 if (__xfrm_decode_session(skb, &fl, family, reverse) < 0) {
1919 XFRM_INC_STATS(LINUX_MIB_XFRMINHDRERROR);
1920 return 0;
1921 }
1922
1923 nf_nat_decode_session(skb, &fl, family);
1924
1925
1926 if (skb->sp) {
1927 int i;
1928
1929 for (i=skb->sp->len-1; i>=0; i--) {
1930 struct xfrm_state *x = skb->sp->xvec[i];
1931 if (!xfrm_selector_match(&x->sel, &fl, family)) {
1932 XFRM_INC_STATS(LINUX_MIB_XFRMINSTATEMISMATCH);
1933 return 0;
1934 }
1935 }
1936 }
1937
1938 pol = NULL;
1939 if (sk && sk->sk_policy[dir]) {
1940 pol = xfrm_sk_policy_lookup(sk, dir, &fl);
1941 if (IS_ERR(pol)) {
1942 XFRM_INC_STATS(LINUX_MIB_XFRMINPOLERROR);
1943 return 0;
1944 }
1945 }
1946
1947 if (!pol)
1948 pol = flow_cache_lookup(&fl, family, fl_dir,
1949 xfrm_policy_lookup);
1950
1951 if (IS_ERR(pol)) {
1952 XFRM_INC_STATS(LINUX_MIB_XFRMINPOLERROR);
1953 return 0;
1954 }
1955
1956 if (!pol) {
1957 if (skb->sp && secpath_has_nontransport(skb->sp, 0, &xerr_idx)) {
1958 xfrm_secpath_reject(xerr_idx, skb, &fl);
1959 XFRM_INC_STATS(LINUX_MIB_XFRMINNOPOLS);
1960 return 0;
1961 }
1962 return 1;
1963 }
1964
1965 pol->curlft.use_time = get_seconds();
1966
1967 pols[0] = pol;
1968 npols ++;
1969#ifdef CONFIG_XFRM_SUB_POLICY
1970 if (pols[0]->type != XFRM_POLICY_TYPE_MAIN) {
1971 pols[1] = xfrm_policy_lookup_bytype(XFRM_POLICY_TYPE_MAIN,
1972 &fl, family,
1973 XFRM_POLICY_IN);
1974 if (pols[1]) {
1975 if (IS_ERR(pols[1])) {
1976 XFRM_INC_STATS(LINUX_MIB_XFRMINPOLERROR);
1977 return 0;
1978 }
1979 pols[1]->curlft.use_time = get_seconds();
1980 npols ++;
1981 }
1982 }
1983#endif
1984
1985 if (pol->action == XFRM_POLICY_ALLOW) {
1986 struct sec_path *sp;
1987 static struct sec_path dummy;
1988 struct xfrm_tmpl *tp[XFRM_MAX_DEPTH];
1989 struct xfrm_tmpl *stp[XFRM_MAX_DEPTH];
1990 struct xfrm_tmpl **tpp = tp;
1991 int ti = 0;
1992 int i, k;
1993
1994 if ((sp = skb->sp) == NULL)
1995 sp = &dummy;
1996
1997 for (pi = 0; pi < npols; pi++) {
1998 if (pols[pi] != pol &&
1999 pols[pi]->action != XFRM_POLICY_ALLOW) {
2000 XFRM_INC_STATS(LINUX_MIB_XFRMINPOLBLOCK);
2001 goto reject;
2002 }
2003 if (ti + pols[pi]->xfrm_nr >= XFRM_MAX_DEPTH) {
2004 XFRM_INC_STATS(LINUX_MIB_XFRMINBUFFERERROR);
2005 goto reject_error;
2006 }
2007 for (i = 0; i < pols[pi]->xfrm_nr; i++)
2008 tpp[ti++] = &pols[pi]->xfrm_vec[i];
2009 }
2010 xfrm_nr = ti;
2011 if (npols > 1) {
2012 xfrm_tmpl_sort(stp, tpp, xfrm_nr, family);
2013 tpp = stp;
2014 }
2015
2016
2017
2018
2019
2020
2021
2022 for (i = xfrm_nr-1, k = 0; i >= 0; i--) {
2023 k = xfrm_policy_ok(tpp[i], sp, k, family);
2024 if (k < 0) {
2025 if (k < -1)
2026
2027 xerr_idx = -(2+k);
2028 XFRM_INC_STATS(LINUX_MIB_XFRMINTMPLMISMATCH);
2029 goto reject;
2030 }
2031 }
2032
2033 if (secpath_has_nontransport(sp, k, &xerr_idx)) {
2034 XFRM_INC_STATS(LINUX_MIB_XFRMINTMPLMISMATCH);
2035 goto reject;
2036 }
2037
2038 xfrm_pols_put(pols, npols);
2039 return 1;
2040 }
2041 XFRM_INC_STATS(LINUX_MIB_XFRMINPOLBLOCK);
2042
2043reject:
2044 xfrm_secpath_reject(xerr_idx, skb, &fl);
2045reject_error:
2046 xfrm_pols_put(pols, npols);
2047 return 0;
2048}
2049EXPORT_SYMBOL(__xfrm_policy_check);
2050
2051int __xfrm_route_forward(struct sk_buff *skb, unsigned short family)
2052{
2053 struct flowi fl;
2054
2055 if (xfrm_decode_session(skb, &fl, family) < 0) {
2056
2057 XFRM_INC_STATS(LINUX_MIB_XFRMINHDRERROR);
2058 return 0;
2059 }
2060
2061 return xfrm_lookup(&skb->dst, &fl, NULL, 0) == 0;
2062}
2063EXPORT_SYMBOL(__xfrm_route_forward);
2064
2065
2066
2067static struct dst_entry *xfrm_dst_check(struct dst_entry *dst, u32 cookie)
2068{
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089 if (dst->obsolete < 0 && !stale_bundle(dst))
2090 return dst;
2091
2092 return NULL;
2093}
2094
2095static int stale_bundle(struct dst_entry *dst)
2096{
2097 return !xfrm_bundle_ok(NULL, (struct xfrm_dst *)dst, NULL, AF_UNSPEC, 0);
2098}
2099
2100void xfrm_dst_ifdown(struct dst_entry *dst, struct net_device *dev)
2101{
2102 while ((dst = dst->child) && dst->xfrm && dst->dev == dev) {
2103 dst->dev = dev_net(dev)->loopback_dev;
2104 dev_hold(dst->dev);
2105 dev_put(dev);
2106 }
2107}
2108EXPORT_SYMBOL(xfrm_dst_ifdown);
2109
2110static void xfrm_link_failure(struct sk_buff *skb)
2111{
2112
2113 return;
2114}
2115
2116static struct dst_entry *xfrm_negative_advice(struct dst_entry *dst)
2117{
2118 if (dst) {
2119 if (dst->obsolete) {
2120 dst_release(dst);
2121 dst = NULL;
2122 }
2123 }
2124 return dst;
2125}
2126
2127static void prune_one_bundle(struct xfrm_policy *pol, int (*func)(struct dst_entry *), struct dst_entry **gc_list_p)
2128{
2129 struct dst_entry *dst, **dstp;
2130
2131 write_lock(&pol->lock);
2132 dstp = &pol->bundles;
2133 while ((dst=*dstp) != NULL) {
2134 if (func(dst)) {
2135 *dstp = dst->next;
2136 dst->next = *gc_list_p;
2137 *gc_list_p = dst;
2138 } else {
2139 dstp = &dst->next;
2140 }
2141 }
2142 write_unlock(&pol->lock);
2143}
2144
2145static void xfrm_prune_bundles(int (*func)(struct dst_entry *))
2146{
2147 struct dst_entry *gc_list = NULL;
2148 int dir;
2149
2150 read_lock_bh(&xfrm_policy_lock);
2151 for (dir = 0; dir < XFRM_POLICY_MAX * 2; dir++) {
2152 struct xfrm_policy *pol;
2153 struct hlist_node *entry;
2154 struct hlist_head *table;
2155 int i;
2156
2157 hlist_for_each_entry(pol, entry,
2158 &xfrm_policy_inexact[dir], bydst)
2159 prune_one_bundle(pol, func, &gc_list);
2160
2161 table = xfrm_policy_bydst[dir].table;
2162 for (i = xfrm_policy_bydst[dir].hmask; i >= 0; i--) {
2163 hlist_for_each_entry(pol, entry, table + i, bydst)
2164 prune_one_bundle(pol, func, &gc_list);
2165 }
2166 }
2167 read_unlock_bh(&xfrm_policy_lock);
2168
2169 while (gc_list) {
2170 struct dst_entry *dst = gc_list;
2171 gc_list = dst->next;
2172 dst_free(dst);
2173 }
2174}
2175
2176static int unused_bundle(struct dst_entry *dst)
2177{
2178 return !atomic_read(&dst->__refcnt);
2179}
2180
2181static void __xfrm_garbage_collect(void)
2182{
2183 xfrm_prune_bundles(unused_bundle);
2184}
2185
2186static int xfrm_flush_bundles(void)
2187{
2188 xfrm_prune_bundles(stale_bundle);
2189 return 0;
2190}
2191
2192static void xfrm_init_pmtu(struct dst_entry *dst)
2193{
2194 do {
2195 struct xfrm_dst *xdst = (struct xfrm_dst *)dst;
2196 u32 pmtu, route_mtu_cached;
2197
2198 pmtu = dst_mtu(dst->child);
2199 xdst->child_mtu_cached = pmtu;
2200
2201 pmtu = xfrm_state_mtu(dst->xfrm, pmtu);
2202
2203 route_mtu_cached = dst_mtu(xdst->route);
2204 xdst->route_mtu_cached = route_mtu_cached;
2205
2206 if (pmtu > route_mtu_cached)
2207 pmtu = route_mtu_cached;
2208
2209 dst->metrics[RTAX_MTU-1] = pmtu;
2210 } while ((dst = dst->next));
2211}
2212
2213
2214
2215
2216
2217int xfrm_bundle_ok(struct xfrm_policy *pol, struct xfrm_dst *first,
2218 struct flowi *fl, int family, int strict)
2219{
2220 struct dst_entry *dst = &first->u.dst;
2221 struct xfrm_dst *last;
2222 u32 mtu;
2223
2224 if (!dst_check(dst->path, ((struct xfrm_dst *)dst)->path_cookie) ||
2225 (dst->dev && !netif_running(dst->dev)))
2226 return 0;
2227#ifdef CONFIG_XFRM_SUB_POLICY
2228 if (fl) {
2229 if (first->origin && !flow_cache_uli_match(first->origin, fl))
2230 return 0;
2231 if (first->partner &&
2232 !xfrm_selector_match(first->partner, fl, family))
2233 return 0;
2234 }
2235#endif
2236
2237 last = NULL;
2238
2239 do {
2240 struct xfrm_dst *xdst = (struct xfrm_dst *)dst;
2241
2242 if (fl && !xfrm_selector_match(&dst->xfrm->sel, fl, family))
2243 return 0;
2244 if (fl && pol &&
2245 !security_xfrm_state_pol_flow_match(dst->xfrm, pol, fl))
2246 return 0;
2247 if (dst->xfrm->km.state != XFRM_STATE_VALID)
2248 return 0;
2249 if (xdst->genid != dst->xfrm->genid)
2250 return 0;
2251
2252 if (strict && fl &&
2253 !(dst->xfrm->outer_mode->flags & XFRM_MODE_FLAG_TUNNEL) &&
2254 !xfrm_state_addr_flow_check(dst->xfrm, fl, family))
2255 return 0;
2256
2257 mtu = dst_mtu(dst->child);
2258 if (xdst->child_mtu_cached != mtu) {
2259 last = xdst;
2260 xdst->child_mtu_cached = mtu;
2261 }
2262
2263 if (!dst_check(xdst->route, xdst->route_cookie))
2264 return 0;
2265 mtu = dst_mtu(xdst->route);
2266 if (xdst->route_mtu_cached != mtu) {
2267 last = xdst;
2268 xdst->route_mtu_cached = mtu;
2269 }
2270
2271 dst = dst->child;
2272 } while (dst->xfrm);
2273
2274 if (likely(!last))
2275 return 1;
2276
2277 mtu = last->child_mtu_cached;
2278 for (;;) {
2279 dst = &last->u.dst;
2280
2281 mtu = xfrm_state_mtu(dst->xfrm, mtu);
2282 if (mtu > last->route_mtu_cached)
2283 mtu = last->route_mtu_cached;
2284 dst->metrics[RTAX_MTU-1] = mtu;
2285
2286 if (last == first)
2287 break;
2288
2289 last = (struct xfrm_dst *)last->u.dst.next;
2290 last->child_mtu_cached = mtu;
2291 }
2292
2293 return 1;
2294}
2295
2296EXPORT_SYMBOL(xfrm_bundle_ok);
2297
2298int xfrm_policy_register_afinfo(struct xfrm_policy_afinfo *afinfo)
2299{
2300 int err = 0;
2301 if (unlikely(afinfo == NULL))
2302 return -EINVAL;
2303 if (unlikely(afinfo->family >= NPROTO))
2304 return -EAFNOSUPPORT;
2305 write_lock_bh(&xfrm_policy_afinfo_lock);
2306 if (unlikely(xfrm_policy_afinfo[afinfo->family] != NULL))
2307 err = -ENOBUFS;
2308 else {
2309 struct dst_ops *dst_ops = afinfo->dst_ops;
2310 if (likely(dst_ops->kmem_cachep == NULL))
2311 dst_ops->kmem_cachep = xfrm_dst_cache;
2312 if (likely(dst_ops->check == NULL))
2313 dst_ops->check = xfrm_dst_check;
2314 if (likely(dst_ops->negative_advice == NULL))
2315 dst_ops->negative_advice = xfrm_negative_advice;
2316 if (likely(dst_ops->link_failure == NULL))
2317 dst_ops->link_failure = xfrm_link_failure;
2318 if (likely(afinfo->garbage_collect == NULL))
2319 afinfo->garbage_collect = __xfrm_garbage_collect;
2320 xfrm_policy_afinfo[afinfo->family] = afinfo;
2321 }
2322 write_unlock_bh(&xfrm_policy_afinfo_lock);
2323 return err;
2324}
2325EXPORT_SYMBOL(xfrm_policy_register_afinfo);
2326
2327int xfrm_policy_unregister_afinfo(struct xfrm_policy_afinfo *afinfo)
2328{
2329 int err = 0;
2330 if (unlikely(afinfo == NULL))
2331 return -EINVAL;
2332 if (unlikely(afinfo->family >= NPROTO))
2333 return -EAFNOSUPPORT;
2334 write_lock_bh(&xfrm_policy_afinfo_lock);
2335 if (likely(xfrm_policy_afinfo[afinfo->family] != NULL)) {
2336 if (unlikely(xfrm_policy_afinfo[afinfo->family] != afinfo))
2337 err = -EINVAL;
2338 else {
2339 struct dst_ops *dst_ops = afinfo->dst_ops;
2340 xfrm_policy_afinfo[afinfo->family] = NULL;
2341 dst_ops->kmem_cachep = NULL;
2342 dst_ops->check = NULL;
2343 dst_ops->negative_advice = NULL;
2344 dst_ops->link_failure = NULL;
2345 afinfo->garbage_collect = NULL;
2346 }
2347 }
2348 write_unlock_bh(&xfrm_policy_afinfo_lock);
2349 return err;
2350}
2351EXPORT_SYMBOL(xfrm_policy_unregister_afinfo);
2352
2353static struct xfrm_policy_afinfo *xfrm_policy_get_afinfo(unsigned short family)
2354{
2355 struct xfrm_policy_afinfo *afinfo;
2356 if (unlikely(family >= NPROTO))
2357 return NULL;
2358 read_lock(&xfrm_policy_afinfo_lock);
2359 afinfo = xfrm_policy_afinfo[family];
2360 if (unlikely(!afinfo))
2361 read_unlock(&xfrm_policy_afinfo_lock);
2362 return afinfo;
2363}
2364
2365static void xfrm_policy_put_afinfo(struct xfrm_policy_afinfo *afinfo)
2366{
2367 read_unlock(&xfrm_policy_afinfo_lock);
2368}
2369
2370static int xfrm_dev_event(struct notifier_block *this, unsigned long event, void *ptr)
2371{
2372 struct net_device *dev = ptr;
2373
2374 if (!net_eq(dev_net(dev), &init_net))
2375 return NOTIFY_DONE;
2376
2377 switch (event) {
2378 case NETDEV_DOWN:
2379 xfrm_flush_bundles();
2380 }
2381 return NOTIFY_DONE;
2382}
2383
2384static struct notifier_block xfrm_dev_notifier = {
2385 xfrm_dev_event,
2386 NULL,
2387 0
2388};
2389
2390#ifdef CONFIG_XFRM_STATISTICS
2391static int __init xfrm_statistics_init(void)
2392{
2393 if (snmp_mib_init((void **)xfrm_statistics,
2394 sizeof(struct linux_xfrm_mib)) < 0)
2395 return -ENOMEM;
2396 return 0;
2397}
2398#endif
2399
2400static void __init xfrm_policy_init(void)
2401{
2402 unsigned int hmask, sz;
2403 int dir;
2404
2405 xfrm_dst_cache = kmem_cache_create("xfrm_dst_cache",
2406 sizeof(struct xfrm_dst),
2407 0, SLAB_HWCACHE_ALIGN|SLAB_PANIC,
2408 NULL);
2409
2410 hmask = 8 - 1;
2411 sz = (hmask+1) * sizeof(struct hlist_head);
2412
2413 xfrm_policy_byidx = xfrm_hash_alloc(sz);
2414 xfrm_idx_hmask = hmask;
2415 if (!xfrm_policy_byidx)
2416 panic("XFRM: failed to allocate byidx hash\n");
2417
2418 for (dir = 0; dir < XFRM_POLICY_MAX * 2; dir++) {
2419 struct xfrm_policy_hash *htab;
2420
2421 INIT_HLIST_HEAD(&xfrm_policy_inexact[dir]);
2422
2423 htab = &xfrm_policy_bydst[dir];
2424 htab->table = xfrm_hash_alloc(sz);
2425 htab->hmask = hmask;
2426 if (!htab->table)
2427 panic("XFRM: failed to allocate bydst hash\n");
2428 }
2429
2430 INIT_LIST_HEAD(&xfrm_policy_all);
2431 INIT_WORK(&xfrm_policy_gc_work, xfrm_policy_gc_task);
2432 register_netdevice_notifier(&xfrm_dev_notifier);
2433}
2434
2435void __init xfrm_init(void)
2436{
2437#ifdef CONFIG_XFRM_STATISTICS
2438 xfrm_statistics_init();
2439#endif
2440 xfrm_state_init();
2441 xfrm_policy_init();
2442 xfrm_input_init();
2443#ifdef CONFIG_XFRM_STATISTICS
2444 xfrm_proc_init();
2445#endif
2446}
2447
2448#ifdef CONFIG_AUDITSYSCALL
2449static void xfrm_audit_common_policyinfo(struct xfrm_policy *xp,
2450 struct audit_buffer *audit_buf)
2451{
2452 struct xfrm_sec_ctx *ctx = xp->security;
2453 struct xfrm_selector *sel = &xp->selector;
2454
2455 if (ctx)
2456 audit_log_format(audit_buf, " sec_alg=%u sec_doi=%u sec_obj=%s",
2457 ctx->ctx_alg, ctx->ctx_doi, ctx->ctx_str);
2458
2459 switch(sel->family) {
2460 case AF_INET:
2461 audit_log_format(audit_buf, " src=" NIPQUAD_FMT,
2462 NIPQUAD(sel->saddr.a4));
2463 if (sel->prefixlen_s != 32)
2464 audit_log_format(audit_buf, " src_prefixlen=%d",
2465 sel->prefixlen_s);
2466 audit_log_format(audit_buf, " dst=" NIPQUAD_FMT,
2467 NIPQUAD(sel->daddr.a4));
2468 if (sel->prefixlen_d != 32)
2469 audit_log_format(audit_buf, " dst_prefixlen=%d",
2470 sel->prefixlen_d);
2471 break;
2472 case AF_INET6:
2473 audit_log_format(audit_buf, " src=" NIP6_FMT,
2474 NIP6(*(struct in6_addr *)sel->saddr.a6));
2475 if (sel->prefixlen_s != 128)
2476 audit_log_format(audit_buf, " src_prefixlen=%d",
2477 sel->prefixlen_s);
2478 audit_log_format(audit_buf, " dst=" NIP6_FMT,
2479 NIP6(*(struct in6_addr *)sel->daddr.a6));
2480 if (sel->prefixlen_d != 128)
2481 audit_log_format(audit_buf, " dst_prefixlen=%d",
2482 sel->prefixlen_d);
2483 break;
2484 }
2485}
2486
2487void xfrm_audit_policy_add(struct xfrm_policy *xp, int result,
2488 uid_t auid, u32 sessionid, u32 secid)
2489{
2490 struct audit_buffer *audit_buf;
2491
2492 audit_buf = xfrm_audit_start("SPD-add");
2493 if (audit_buf == NULL)
2494 return;
2495 xfrm_audit_helper_usrinfo(auid, sessionid, secid, audit_buf);
2496 audit_log_format(audit_buf, " res=%u", result);
2497 xfrm_audit_common_policyinfo(xp, audit_buf);
2498 audit_log_end(audit_buf);
2499}
2500EXPORT_SYMBOL_GPL(xfrm_audit_policy_add);
2501
2502void xfrm_audit_policy_delete(struct xfrm_policy *xp, int result,
2503 uid_t auid, u32 sessionid, u32 secid)
2504{
2505 struct audit_buffer *audit_buf;
2506
2507 audit_buf = xfrm_audit_start("SPD-delete");
2508 if (audit_buf == NULL)
2509 return;
2510 xfrm_audit_helper_usrinfo(auid, sessionid, secid, audit_buf);
2511 audit_log_format(audit_buf, " res=%u", result);
2512 xfrm_audit_common_policyinfo(xp, audit_buf);
2513 audit_log_end(audit_buf);
2514}
2515EXPORT_SYMBOL_GPL(xfrm_audit_policy_delete);
2516#endif
2517
2518#ifdef CONFIG_XFRM_MIGRATE
2519static int xfrm_migrate_selector_match(struct xfrm_selector *sel_cmp,
2520 struct xfrm_selector *sel_tgt)
2521{
2522 if (sel_cmp->proto == IPSEC_ULPROTO_ANY) {
2523 if (sel_tgt->family == sel_cmp->family &&
2524 xfrm_addr_cmp(&sel_tgt->daddr, &sel_cmp->daddr,
2525 sel_cmp->family) == 0 &&
2526 xfrm_addr_cmp(&sel_tgt->saddr, &sel_cmp->saddr,
2527 sel_cmp->family) == 0 &&
2528 sel_tgt->prefixlen_d == sel_cmp->prefixlen_d &&
2529 sel_tgt->prefixlen_s == sel_cmp->prefixlen_s) {
2530 return 1;
2531 }
2532 } else {
2533 if (memcmp(sel_tgt, sel_cmp, sizeof(*sel_tgt)) == 0) {
2534 return 1;
2535 }
2536 }
2537 return 0;
2538}
2539
2540static struct xfrm_policy * xfrm_migrate_policy_find(struct xfrm_selector *sel,
2541 u8 dir, u8 type)
2542{
2543 struct xfrm_policy *pol, *ret = NULL;
2544 struct hlist_node *entry;
2545 struct hlist_head *chain;
2546 u32 priority = ~0U;
2547
2548 read_lock_bh(&xfrm_policy_lock);
2549 chain = policy_hash_direct(&sel->daddr, &sel->saddr, sel->family, dir);
2550 hlist_for_each_entry(pol, entry, chain, bydst) {
2551 if (xfrm_migrate_selector_match(sel, &pol->selector) &&
2552 pol->type == type) {
2553 ret = pol;
2554 priority = ret->priority;
2555 break;
2556 }
2557 }
2558 chain = &xfrm_policy_inexact[dir];
2559 hlist_for_each_entry(pol, entry, chain, bydst) {
2560 if (xfrm_migrate_selector_match(sel, &pol->selector) &&
2561 pol->type == type &&
2562 pol->priority < priority) {
2563 ret = pol;
2564 break;
2565 }
2566 }
2567
2568 if (ret)
2569 xfrm_pol_hold(ret);
2570
2571 read_unlock_bh(&xfrm_policy_lock);
2572
2573 return ret;
2574}
2575
2576static int migrate_tmpl_match(struct xfrm_migrate *m, struct xfrm_tmpl *t)
2577{
2578 int match = 0;
2579
2580 if (t->mode == m->mode && t->id.proto == m->proto &&
2581 (m->reqid == 0 || t->reqid == m->reqid)) {
2582 switch (t->mode) {
2583 case XFRM_MODE_TUNNEL:
2584 case XFRM_MODE_BEET:
2585 if (xfrm_addr_cmp(&t->id.daddr, &m->old_daddr,
2586 m->old_family) == 0 &&
2587 xfrm_addr_cmp(&t->saddr, &m->old_saddr,
2588 m->old_family) == 0) {
2589 match = 1;
2590 }
2591 break;
2592 case XFRM_MODE_TRANSPORT:
2593
2594
2595
2596 match = 1;
2597 break;
2598 default:
2599 break;
2600 }
2601 }
2602 return match;
2603}
2604
2605
2606static int xfrm_policy_migrate(struct xfrm_policy *pol,
2607 struct xfrm_migrate *m, int num_migrate)
2608{
2609 struct xfrm_migrate *mp;
2610 struct dst_entry *dst;
2611 int i, j, n = 0;
2612
2613 write_lock_bh(&pol->lock);
2614 if (unlikely(pol->walk.dead)) {
2615
2616 write_unlock_bh(&pol->lock);
2617 return -ENOENT;
2618 }
2619
2620 for (i = 0; i < pol->xfrm_nr; i++) {
2621 for (j = 0, mp = m; j < num_migrate; j++, mp++) {
2622 if (!migrate_tmpl_match(mp, &pol->xfrm_vec[i]))
2623 continue;
2624 n++;
2625 if (pol->xfrm_vec[i].mode != XFRM_MODE_TUNNEL &&
2626 pol->xfrm_vec[i].mode != XFRM_MODE_BEET)
2627 continue;
2628
2629 memcpy(&pol->xfrm_vec[i].id.daddr, &mp->new_daddr,
2630 sizeof(pol->xfrm_vec[i].id.daddr));
2631 memcpy(&pol->xfrm_vec[i].saddr, &mp->new_saddr,
2632 sizeof(pol->xfrm_vec[i].saddr));
2633 pol->xfrm_vec[i].encap_family = mp->new_family;
2634
2635 while ((dst = pol->bundles) != NULL) {
2636 pol->bundles = dst->next;
2637 dst_free(dst);
2638 }
2639 }
2640 }
2641
2642 write_unlock_bh(&pol->lock);
2643
2644 if (!n)
2645 return -ENODATA;
2646
2647 return 0;
2648}
2649
2650static int xfrm_migrate_check(struct xfrm_migrate *m, int num_migrate)
2651{
2652 int i, j;
2653
2654 if (num_migrate < 1 || num_migrate > XFRM_MAX_DEPTH)
2655 return -EINVAL;
2656
2657 for (i = 0; i < num_migrate; i++) {
2658 if ((xfrm_addr_cmp(&m[i].old_daddr, &m[i].new_daddr,
2659 m[i].old_family) == 0) &&
2660 (xfrm_addr_cmp(&m[i].old_saddr, &m[i].new_saddr,
2661 m[i].old_family) == 0))
2662 return -EINVAL;
2663 if (xfrm_addr_any(&m[i].new_daddr, m[i].new_family) ||
2664 xfrm_addr_any(&m[i].new_saddr, m[i].new_family))
2665 return -EINVAL;
2666
2667
2668 for (j = i + 1; j < num_migrate; j++) {
2669 if (!memcmp(&m[i].old_daddr, &m[j].old_daddr,
2670 sizeof(m[i].old_daddr)) &&
2671 !memcmp(&m[i].old_saddr, &m[j].old_saddr,
2672 sizeof(m[i].old_saddr)) &&
2673 m[i].proto == m[j].proto &&
2674 m[i].mode == m[j].mode &&
2675 m[i].reqid == m[j].reqid &&
2676 m[i].old_family == m[j].old_family)
2677 return -EINVAL;
2678 }
2679 }
2680
2681 return 0;
2682}
2683
2684int xfrm_migrate(struct xfrm_selector *sel, u8 dir, u8 type,
2685 struct xfrm_migrate *m, int num_migrate,
2686 struct xfrm_kmaddress *k)
2687{
2688 int i, err, nx_cur = 0, nx_new = 0;
2689 struct xfrm_policy *pol = NULL;
2690 struct xfrm_state *x, *xc;
2691 struct xfrm_state *x_cur[XFRM_MAX_DEPTH];
2692 struct xfrm_state *x_new[XFRM_MAX_DEPTH];
2693 struct xfrm_migrate *mp;
2694
2695 if ((err = xfrm_migrate_check(m, num_migrate)) < 0)
2696 goto out;
2697
2698
2699 if ((pol = xfrm_migrate_policy_find(sel, dir, type)) == NULL) {
2700 err = -ENOENT;
2701 goto out;
2702 }
2703
2704
2705 for (i = 0, mp = m; i < num_migrate; i++, mp++) {
2706 if ((x = xfrm_migrate_state_find(mp))) {
2707 x_cur[nx_cur] = x;
2708 nx_cur++;
2709 if ((xc = xfrm_state_migrate(x, mp))) {
2710 x_new[nx_new] = xc;
2711 nx_new++;
2712 } else {
2713 err = -ENODATA;
2714 goto restore_state;
2715 }
2716 }
2717 }
2718
2719
2720 if ((err = xfrm_policy_migrate(pol, m, num_migrate)) < 0)
2721 goto restore_state;
2722
2723
2724 if (nx_cur) {
2725 xfrm_states_put(x_cur, nx_cur);
2726 xfrm_states_delete(x_cur, nx_cur);
2727 }
2728
2729
2730 km_migrate(sel, dir, type, m, num_migrate, k);
2731
2732 xfrm_pol_put(pol);
2733
2734 return 0;
2735out:
2736 return err;
2737
2738restore_state:
2739 if (pol)
2740 xfrm_pol_put(pol);
2741 if (nx_cur)
2742 xfrm_states_put(x_cur, nx_cur);
2743 if (nx_new)
2744 xfrm_states_delete(x_new, nx_new);
2745
2746 return err;
2747}
2748EXPORT_SYMBOL(xfrm_migrate);
2749#endif