1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17#include <linux/module.h>
18#include <linux/types.h>
19#include <linux/kernel.h>
20#include <linux/string.h>
21#include <linux/errno.h>
22#include <linux/skbuff.h>
23#include <linux/init.h>
24#include <linux/kmod.h>
25#include <linux/err.h>
26#include <linux/slab.h>
27#include <net/net_namespace.h>
28#include <net/sock.h>
29#include <net/netlink.h>
30#include <net/pkt_sched.h>
31#include <net/pkt_cls.h>
32
33
34
35static struct tcf_proto_ops *tcf_proto_base __read_mostly;
36
37
38static DEFINE_RWLOCK(cls_mod_lock);
39
40
41
42static const struct tcf_proto_ops *tcf_proto_lookup_ops(struct nlattr *kind)
43{
44 const struct tcf_proto_ops *t = NULL;
45
46 if (kind) {
47 read_lock(&cls_mod_lock);
48 for (t = tcf_proto_base; t; t = t->next) {
49 if (nla_strcmp(kind, t->kind) == 0) {
50 if (!try_module_get(t->owner))
51 t = NULL;
52 break;
53 }
54 }
55 read_unlock(&cls_mod_lock);
56 }
57 return t;
58}
59
60
61
62int register_tcf_proto_ops(struct tcf_proto_ops *ops)
63{
64 struct tcf_proto_ops *t, **tp;
65 int rc = -EEXIST;
66
67 write_lock(&cls_mod_lock);
68 for (tp = &tcf_proto_base; (t = *tp) != NULL; tp = &t->next)
69 if (!strcmp(ops->kind, t->kind))
70 goto out;
71
72 ops->next = NULL;
73 *tp = ops;
74 rc = 0;
75out:
76 write_unlock(&cls_mod_lock);
77 return rc;
78}
79EXPORT_SYMBOL(register_tcf_proto_ops);
80
81int unregister_tcf_proto_ops(struct tcf_proto_ops *ops)
82{
83 struct tcf_proto_ops *t, **tp;
84 int rc = -ENOENT;
85
86 write_lock(&cls_mod_lock);
87 for (tp = &tcf_proto_base; (t = *tp) != NULL; tp = &t->next)
88 if (t == ops)
89 break;
90
91 if (!t)
92 goto out;
93 *tp = t->next;
94 rc = 0;
95out:
96 write_unlock(&cls_mod_lock);
97 return rc;
98}
99EXPORT_SYMBOL(unregister_tcf_proto_ops);
100
101static int tfilter_notify(struct net *net, struct sk_buff *oskb,
102 struct nlmsghdr *n, struct tcf_proto *tp,
103 unsigned long fh, int event);
104
105
106
107
108static inline u32 tcf_auto_prio(struct tcf_proto *tp)
109{
110 u32 first = TC_H_MAKE(0xC0000000U, 0U);
111
112 if (tp)
113 first = tp->prio - 1;
114
115 return first;
116}
117
118
119
120static int tc_ctl_tfilter(struct sk_buff *skb, struct nlmsghdr *n)
121{
122 struct net *net = sock_net(skb->sk);
123 struct nlattr *tca[TCA_MAX + 1];
124 spinlock_t *root_lock;
125 struct tcmsg *t;
126 u32 protocol;
127 u32 prio;
128 u32 nprio;
129 u32 parent;
130 struct net_device *dev;
131 struct Qdisc *q;
132 struct tcf_proto **back, **chain;
133 struct tcf_proto *tp;
134 const struct tcf_proto_ops *tp_ops;
135 const struct Qdisc_class_ops *cops;
136 unsigned long cl;
137 unsigned long fh;
138 int err;
139 int tp_created = 0;
140
141 if ((n->nlmsg_type != RTM_GETTFILTER) && !capable(CAP_NET_ADMIN))
142 return -EPERM;
143
144replay:
145 err = nlmsg_parse(n, sizeof(*t), tca, TCA_MAX, NULL);
146 if (err < 0)
147 return err;
148
149 t = nlmsg_data(n);
150 protocol = TC_H_MIN(t->tcm_info);
151 prio = TC_H_MAJ(t->tcm_info);
152 nprio = prio;
153 parent = t->tcm_parent;
154 cl = 0;
155
156 if (prio == 0) {
157
158 if (n->nlmsg_type != RTM_NEWTFILTER ||
159 !(n->nlmsg_flags & NLM_F_CREATE))
160 return -ENOENT;
161 prio = TC_H_MAKE(0x80000000U, 0U);
162 }
163
164
165
166
167 dev = __dev_get_by_index(net, t->tcm_ifindex);
168 if (dev == NULL)
169 return -ENODEV;
170
171
172 if (!parent) {
173 q = dev->qdisc;
174 parent = q->handle;
175 } else {
176 q = qdisc_lookup(dev, TC_H_MAJ(t->tcm_parent));
177 if (q == NULL)
178 return -EINVAL;
179 }
180
181
182 cops = q->ops->cl_ops;
183 if (!cops)
184 return -EINVAL;
185
186 if (cops->tcf_chain == NULL)
187 return -EOPNOTSUPP;
188
189
190 if (TC_H_MIN(parent)) {
191 cl = cops->get(q, parent);
192 if (cl == 0)
193 return -ENOENT;
194 }
195
196
197 chain = cops->tcf_chain(q, cl);
198 err = -EINVAL;
199 if (chain == NULL)
200 goto errout;
201
202
203 for (back = chain; (tp = *back) != NULL; back = &tp->next) {
204 if (tp->prio >= prio) {
205 if (tp->prio == prio) {
206 if (!nprio ||
207 (tp->protocol != protocol && protocol))
208 goto errout;
209 } else
210 tp = NULL;
211 break;
212 }
213 }
214
215 root_lock = qdisc_root_sleeping_lock(q);
216
217 if (tp == NULL) {
218
219
220 if (tca[TCA_KIND] == NULL || !protocol)
221 goto errout;
222
223 err = -ENOENT;
224 if (n->nlmsg_type != RTM_NEWTFILTER ||
225 !(n->nlmsg_flags & NLM_F_CREATE))
226 goto errout;
227
228
229
230
231 err = -ENOBUFS;
232 tp = kzalloc(sizeof(*tp), GFP_KERNEL);
233 if (tp == NULL)
234 goto errout;
235 err = -ENOENT;
236 tp_ops = tcf_proto_lookup_ops(tca[TCA_KIND]);
237 if (tp_ops == NULL) {
238#ifdef CONFIG_MODULES
239 struct nlattr *kind = tca[TCA_KIND];
240 char name[IFNAMSIZ];
241
242 if (kind != NULL &&
243 nla_strlcpy(name, kind, IFNAMSIZ) < IFNAMSIZ) {
244 rtnl_unlock();
245 request_module("cls_%s", name);
246 rtnl_lock();
247 tp_ops = tcf_proto_lookup_ops(kind);
248
249
250
251
252
253
254 if (tp_ops != NULL) {
255 module_put(tp_ops->owner);
256 err = -EAGAIN;
257 }
258 }
259#endif
260 kfree(tp);
261 goto errout;
262 }
263 tp->ops = tp_ops;
264 tp->protocol = protocol;
265 tp->prio = nprio ? : TC_H_MAJ(tcf_auto_prio(*back));
266 tp->q = q;
267 tp->classify = tp_ops->classify;
268 tp->classid = parent;
269
270 err = tp_ops->init(tp);
271 if (err != 0) {
272 module_put(tp_ops->owner);
273 kfree(tp);
274 goto errout;
275 }
276
277 tp_created = 1;
278
279 } else if (tca[TCA_KIND] && nla_strcmp(tca[TCA_KIND], tp->ops->kind))
280 goto errout;
281
282 fh = tp->ops->get(tp, t->tcm_handle);
283
284 if (fh == 0) {
285 if (n->nlmsg_type == RTM_DELTFILTER && t->tcm_handle == 0) {
286 spin_lock_bh(root_lock);
287 *back = tp->next;
288 spin_unlock_bh(root_lock);
289
290 tfilter_notify(net, skb, n, tp, fh, RTM_DELTFILTER);
291 tcf_destroy(tp);
292 err = 0;
293 goto errout;
294 }
295
296 err = -ENOENT;
297 if (n->nlmsg_type != RTM_NEWTFILTER ||
298 !(n->nlmsg_flags & NLM_F_CREATE))
299 goto errout;
300 } else {
301 switch (n->nlmsg_type) {
302 case RTM_NEWTFILTER:
303 err = -EEXIST;
304 if (n->nlmsg_flags & NLM_F_EXCL) {
305 if (tp_created)
306 tcf_destroy(tp);
307 goto errout;
308 }
309 break;
310 case RTM_DELTFILTER:
311 err = tp->ops->delete(tp, fh);
312 if (err == 0)
313 tfilter_notify(net, skb, n, tp, fh, RTM_DELTFILTER);
314 goto errout;
315 case RTM_GETTFILTER:
316 err = tfilter_notify(net, skb, n, tp, fh, RTM_NEWTFILTER);
317 goto errout;
318 default:
319 err = -EINVAL;
320 goto errout;
321 }
322 }
323
324 err = tp->ops->change(net, skb, tp, cl, t->tcm_handle, tca, &fh);
325 if (err == 0) {
326 if (tp_created) {
327 spin_lock_bh(root_lock);
328 tp->next = *back;
329 *back = tp;
330 spin_unlock_bh(root_lock);
331 }
332 tfilter_notify(net, skb, n, tp, fh, RTM_NEWTFILTER);
333 } else {
334 if (tp_created)
335 tcf_destroy(tp);
336 }
337
338errout:
339 if (cl)
340 cops->put(q, cl);
341 if (err == -EAGAIN)
342
343 goto replay;
344 return err;
345}
346
347static int tcf_fill_node(struct sk_buff *skb, struct tcf_proto *tp,
348 unsigned long fh, u32 portid, u32 seq, u16 flags, int event)
349{
350 struct tcmsg *tcm;
351 struct nlmsghdr *nlh;
352 unsigned char *b = skb_tail_pointer(skb);
353
354 nlh = nlmsg_put(skb, portid, seq, event, sizeof(*tcm), flags);
355 if (!nlh)
356 goto out_nlmsg_trim;
357 tcm = nlmsg_data(nlh);
358 tcm->tcm_family = AF_UNSPEC;
359 tcm->tcm__pad1 = 0;
360 tcm->tcm__pad2 = 0;
361 tcm->tcm_ifindex = qdisc_dev(tp->q)->ifindex;
362 tcm->tcm_parent = tp->classid;
363 tcm->tcm_info = TC_H_MAKE(tp->prio, tp->protocol);
364 if (nla_put_string(skb, TCA_KIND, tp->ops->kind))
365 goto nla_put_failure;
366 tcm->tcm_handle = fh;
367 if (RTM_DELTFILTER != event) {
368 tcm->tcm_handle = 0;
369 if (tp->ops->dump && tp->ops->dump(tp, fh, skb, tcm) < 0)
370 goto nla_put_failure;
371 }
372 nlh->nlmsg_len = skb_tail_pointer(skb) - b;
373 return skb->len;
374
375out_nlmsg_trim:
376nla_put_failure:
377 nlmsg_trim(skb, b);
378 return -1;
379}
380
381static int tfilter_notify(struct net *net, struct sk_buff *oskb,
382 struct nlmsghdr *n, struct tcf_proto *tp,
383 unsigned long fh, int event)
384{
385 struct sk_buff *skb;
386 u32 portid = oskb ? NETLINK_CB(oskb).portid : 0;
387
388 skb = alloc_skb(NLMSG_GOODSIZE, GFP_KERNEL);
389 if (!skb)
390 return -ENOBUFS;
391
392 if (tcf_fill_node(skb, tp, fh, portid, n->nlmsg_seq, 0, event) <= 0) {
393 kfree_skb(skb);
394 return -EINVAL;
395 }
396
397 return rtnetlink_send(skb, net, portid, RTNLGRP_TC,
398 n->nlmsg_flags & NLM_F_ECHO);
399}
400
401struct tcf_dump_args {
402 struct tcf_walker w;
403 struct sk_buff *skb;
404 struct netlink_callback *cb;
405};
406
407static int tcf_node_dump(struct tcf_proto *tp, unsigned long n,
408 struct tcf_walker *arg)
409{
410 struct tcf_dump_args *a = (void *)arg;
411
412 return tcf_fill_node(a->skb, tp, n, NETLINK_CB(a->cb->skb).portid,
413 a->cb->nlh->nlmsg_seq, NLM_F_MULTI, RTM_NEWTFILTER);
414}
415
416
417static int tc_dump_tfilter(struct sk_buff *skb, struct netlink_callback *cb)
418{
419 struct net *net = sock_net(skb->sk);
420 int t;
421 int s_t;
422 struct net_device *dev;
423 struct Qdisc *q;
424 struct tcf_proto *tp, **chain;
425 struct tcmsg *tcm = nlmsg_data(cb->nlh);
426 unsigned long cl = 0;
427 const struct Qdisc_class_ops *cops;
428 struct tcf_dump_args arg;
429
430 if (nlmsg_len(cb->nlh) < sizeof(*tcm))
431 return skb->len;
432 dev = __dev_get_by_index(net, tcm->tcm_ifindex);
433 if (!dev)
434 return skb->len;
435
436 if (!tcm->tcm_parent)
437 q = dev->qdisc;
438 else
439 q = qdisc_lookup(dev, TC_H_MAJ(tcm->tcm_parent));
440 if (!q)
441 goto out;
442 cops = q->ops->cl_ops;
443 if (!cops)
444 goto errout;
445 if (cops->tcf_chain == NULL)
446 goto errout;
447 if (TC_H_MIN(tcm->tcm_parent)) {
448 cl = cops->get(q, tcm->tcm_parent);
449 if (cl == 0)
450 goto errout;
451 }
452 chain = cops->tcf_chain(q, cl);
453 if (chain == NULL)
454 goto errout;
455
456 s_t = cb->args[0];
457
458 for (tp = *chain, t = 0; tp; tp = tp->next, t++) {
459 if (t < s_t)
460 continue;
461 if (TC_H_MAJ(tcm->tcm_info) &&
462 TC_H_MAJ(tcm->tcm_info) != tp->prio)
463 continue;
464 if (TC_H_MIN(tcm->tcm_info) &&
465 TC_H_MIN(tcm->tcm_info) != tp->protocol)
466 continue;
467 if (t > s_t)
468 memset(&cb->args[1], 0, sizeof(cb->args)-sizeof(cb->args[0]));
469 if (cb->args[1] == 0) {
470 if (tcf_fill_node(skb, tp, 0, NETLINK_CB(cb->skb).portid,
471 cb->nlh->nlmsg_seq, NLM_F_MULTI,
472 RTM_NEWTFILTER) <= 0)
473 break;
474
475 cb->args[1] = 1;
476 }
477 if (tp->ops->walk == NULL)
478 continue;
479 arg.w.fn = tcf_node_dump;
480 arg.skb = skb;
481 arg.cb = cb;
482 arg.w.stop = 0;
483 arg.w.skip = cb->args[1] - 1;
484 arg.w.count = 0;
485 tp->ops->walk(tp, &arg.w);
486 cb->args[1] = arg.w.count + 1;
487 if (arg.w.stop)
488 break;
489 }
490
491 cb->args[0] = t;
492
493errout:
494 if (cl)
495 cops->put(q, cl);
496out:
497 return skb->len;
498}
499
500void tcf_exts_destroy(struct tcf_proto *tp, struct tcf_exts *exts)
501{
502#ifdef CONFIG_NET_CLS_ACT
503 if (exts->action) {
504 tcf_action_destroy(exts->action, TCA_ACT_UNBIND);
505 exts->action = NULL;
506 }
507#endif
508}
509EXPORT_SYMBOL(tcf_exts_destroy);
510
511int tcf_exts_validate(struct net *net, struct tcf_proto *tp, struct nlattr **tb,
512 struct nlattr *rate_tlv, struct tcf_exts *exts,
513 const struct tcf_ext_map *map)
514{
515 memset(exts, 0, sizeof(*exts));
516
517#ifdef CONFIG_NET_CLS_ACT
518 {
519 struct tc_action *act;
520
521 if (map->police && tb[map->police]) {
522 act = tcf_action_init_1(net, tb[map->police], rate_tlv,
523 "police", TCA_ACT_NOREPLACE,
524 TCA_ACT_BIND);
525 if (IS_ERR(act))
526 return PTR_ERR(act);
527
528 act->type = TCA_OLD_COMPAT;
529 exts->action = act;
530 } else if (map->action && tb[map->action]) {
531 act = tcf_action_init(net, tb[map->action], rate_tlv,
532 NULL, TCA_ACT_NOREPLACE,
533 TCA_ACT_BIND);
534 if (IS_ERR(act))
535 return PTR_ERR(act);
536
537 exts->action = act;
538 }
539 }
540#else
541 if ((map->action && tb[map->action]) ||
542 (map->police && tb[map->police]))
543 return -EOPNOTSUPP;
544#endif
545
546 return 0;
547}
548EXPORT_SYMBOL(tcf_exts_validate);
549
550void tcf_exts_change(struct tcf_proto *tp, struct tcf_exts *dst,
551 struct tcf_exts *src)
552{
553#ifdef CONFIG_NET_CLS_ACT
554 if (src->action) {
555 struct tc_action *act;
556 tcf_tree_lock(tp);
557 act = dst->action;
558 dst->action = src->action;
559 tcf_tree_unlock(tp);
560 if (act)
561 tcf_action_destroy(act, TCA_ACT_UNBIND);
562 }
563#endif
564}
565EXPORT_SYMBOL(tcf_exts_change);
566
567int tcf_exts_dump(struct sk_buff *skb, struct tcf_exts *exts,
568 const struct tcf_ext_map *map)
569{
570#ifdef CONFIG_NET_CLS_ACT
571 if (map->action && exts->action) {
572
573
574
575
576
577 struct nlattr *nest;
578
579 if (exts->action->type != TCA_OLD_COMPAT) {
580 nest = nla_nest_start(skb, map->action);
581 if (nest == NULL)
582 goto nla_put_failure;
583 if (tcf_action_dump(skb, exts->action, 0, 0) < 0)
584 goto nla_put_failure;
585 nla_nest_end(skb, nest);
586 } else if (map->police) {
587 nest = nla_nest_start(skb, map->police);
588 if (nest == NULL)
589 goto nla_put_failure;
590 if (tcf_action_dump_old(skb, exts->action, 0, 0) < 0)
591 goto nla_put_failure;
592 nla_nest_end(skb, nest);
593 }
594 }
595#endif
596 return 0;
597nla_put_failure: __attribute__ ((unused))
598 return -1;
599}
600EXPORT_SYMBOL(tcf_exts_dump);
601
602
603int tcf_exts_dump_stats(struct sk_buff *skb, struct tcf_exts *exts,
604 const struct tcf_ext_map *map)
605{
606#ifdef CONFIG_NET_CLS_ACT
607 if (exts->action)
608 if (tcf_action_copy_stats(skb, exts->action, 1) < 0)
609 goto nla_put_failure;
610#endif
611 return 0;
612nla_put_failure: __attribute__ ((unused))
613 return -1;
614}
615EXPORT_SYMBOL(tcf_exts_dump_stats);
616
617static int __init tc_filter_init(void)
618{
619 rtnl_register(PF_UNSPEC, RTM_NEWTFILTER, tc_ctl_tfilter, NULL, NULL);
620 rtnl_register(PF_UNSPEC, RTM_DELTFILTER, tc_ctl_tfilter, NULL, NULL);
621 rtnl_register(PF_UNSPEC, RTM_GETTFILTER, tc_ctl_tfilter,
622 tc_dump_tfilter, NULL);
623
624 return 0;
625}
626
627subsys_initcall(tc_filter_init);
628
