xref: /netbsd-src/sys/net/npf/npf_tableset.c (revision b757af438b42b93f8c6571f026d8b8ef3eaf5fc9)
1 /*	$NetBSD: npf_tableset.c,v 1.10 2012/02/20 00:18:20 rmind Exp $	*/
2 
3 /*-
4  * Copyright (c) 2009-2012 The NetBSD Foundation, Inc.
5  * All rights reserved.
6  *
7  * This material is based upon work partially supported by The
8  * NetBSD Foundation under a contract with Mindaugas Rasiukevicius.
9  *
10  * Redistribution and use in source and binary forms, with or without
11  * modification, are permitted provided that the following conditions
12  * are met:
13  * 1. Redistributions of source code must retain the above copyright
14  *    notice, this list of conditions and the following disclaimer.
15  * 2. Redistributions in binary form must reproduce the above copyright
16  *    notice, this list of conditions and the following disclaimer in the
17  *    documentation and/or other materials provided with the distribution.
18  *
19  * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
20  * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
21  * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
22  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
23  * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
24  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
25  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
26  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
27  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
28  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
29  * POSSIBILITY OF SUCH DAMAGE.
30  */
31 
32 /*
33  * NPF tableset module.
34  *
35  * TODO:
36  * - Currently, code is modeled to handle IPv4 CIDR blocks.
37  * - Dynamic hash growing/shrinking (i.e. re-hash functionality), maybe?
38  * - Dynamic array resize.
39  */
40 
41 #include <sys/cdefs.h>
42 __KERNEL_RCSID(0, "$NetBSD: npf_tableset.c,v 1.10 2012/02/20 00:18:20 rmind Exp $");
43 
44 #include <sys/param.h>
45 #include <sys/types.h>
46 
47 #include <sys/atomic.h>
48 #include <sys/hash.h>
49 #include <sys/kmem.h>
50 #include <sys/pool.h>
51 #include <sys/queue.h>
52 #include <sys/rwlock.h>
53 #include <sys/systm.h>
54 #include <sys/types.h>
55 
56 #include "npf_impl.h"
57 
58 /* Table entry structure. */
59 struct npf_tblent {
60 	/* Hash/tree entry. */
61 	union {
62 		LIST_ENTRY(npf_tblent)	hashq;
63 		rb_node_t		rbnode;
64 	} te_entry;
65 	/* IPv4 CIDR block. */
66 	npf_addr_t			te_addr;
67 	npf_netmask_t			te_mask;
68 };
69 
70 LIST_HEAD(npf_hashl, npf_tblent);
71 
72 /* Table structure. */
73 struct npf_table {
74 	char				t_name[16];
75 	/* Lock and reference count. */
76 	krwlock_t			t_lock;
77 	u_int				t_refcnt;
78 	/* Table ID. */
79 	u_int				t_id;
80 	/* The storage type can be: 1. Hash 2. RB-tree. */
81 	int				t_type;
82 	struct npf_hashl *		t_hashl;
83 	u_long				t_hashmask;
84 	rb_tree_t			t_rbtree;
85 };
86 
87 static pool_cache_t			tblent_cache	__read_mostly;
88 
89 /*
90  * npf_table_sysinit: initialise tableset structures.
91  */
92 void
93 npf_tableset_sysinit(void)
94 {
95 
96 	tblent_cache = pool_cache_init(sizeof(npf_tblent_t), coherency_unit,
97 	    0, 0, "npftenpl", NULL, IPL_NONE, NULL, NULL, NULL);
98 }
99 
100 void
101 npf_tableset_sysfini(void)
102 {
103 
104 	pool_cache_destroy(tblent_cache);
105 }
106 
107 npf_tableset_t *
108 npf_tableset_create(void)
109 {
110 	const size_t sz = NPF_TABLE_SLOTS * sizeof(npf_table_t *);
111 
112 	return kmem_zalloc(sz, KM_SLEEP);
113 }
114 
115 void
116 npf_tableset_destroy(npf_tableset_t *tblset)
117 {
118 	const size_t sz = NPF_TABLE_SLOTS * sizeof(npf_table_t *);
119 	npf_table_t *t;
120 	u_int tid;
121 
122 	/*
123 	 * Destroy all tables (no references should be held, as ruleset
124 	 * should be destroyed before).
125 	 */
126 	for (tid = 0; tid < NPF_TABLE_SLOTS; tid++) {
127 		t = tblset[tid];
128 		if (t != NULL) {
129 			npf_table_destroy(t);
130 		}
131 	}
132 	kmem_free(tblset, sz);
133 }
134 
135 /*
136  * npf_tableset_insert: insert the table into the specified tableset.
137  *
138  * => Returns 0 on success, fails and returns errno if ID is already used.
139  */
140 int
141 npf_tableset_insert(npf_tableset_t *tblset, npf_table_t *t)
142 {
143 	const u_int tid = t->t_id;
144 	int error;
145 
146 	KASSERT((u_int)tid < NPF_TABLE_SLOTS);
147 
148 	if (tblset[tid] == NULL) {
149 		tblset[tid] = t;
150 		error = 0;
151 	} else {
152 		error = EEXIST;
153 	}
154 	return error;
155 }
156 
157 /*
158  * Red-black tree storage.
159  */
160 
161 static signed int
162 table_rbtree_cmp_nodes(void *ctx, const void *n1, const void *n2)
163 {
164 	const npf_tblent_t * const te1 = n1;
165 	const npf_tblent_t * const te2 = n2;
166 
167 	return npf_compare_cidr(&te1->te_addr, te1->te_mask,
168 	    &te2->te_addr, te2->te_mask);
169 }
170 
171 static signed int
172 table_rbtree_cmp_key(void *ctx, const void *n1, const void *key)
173 {
174 	const npf_tblent_t * const te = n1;
175 	const npf_addr_t *t2 = key;
176 
177 	return npf_compare_cidr(&te->te_addr, te->te_mask, t2, NPF_NO_NETMASK);
178 }
179 
180 static const rb_tree_ops_t table_rbtree_ops = {
181 	.rbto_compare_nodes = table_rbtree_cmp_nodes,
182 	.rbto_compare_key = table_rbtree_cmp_key,
183 	.rbto_node_offset = offsetof(npf_tblent_t, te_entry.rbnode),
184 	.rbto_context = NULL
185 };
186 
187 /*
188  * Hash helper routine.
189  */
190 
191 static inline struct npf_hashl *
192 table_hash_bucket(npf_table_t *t, const void *buf, size_t sz)
193 {
194 	const uint32_t hidx = hash32_buf(buf, sz, HASH32_BUF_INIT);
195 
196 	return &t->t_hashl[hidx & t->t_hashmask];
197 }
198 
199 /*
200  * npf_table_create: create table with a specified ID.
201  */
202 npf_table_t *
203 npf_table_create(u_int tid, int type, size_t hsize)
204 {
205 	npf_table_t *t;
206 
207 	KASSERT((u_int)tid < NPF_TABLE_SLOTS);
208 
209 	t = kmem_zalloc(sizeof(npf_table_t), KM_SLEEP);
210 	switch (type) {
211 	case NPF_TABLE_TREE:
212 		rb_tree_init(&t->t_rbtree, &table_rbtree_ops);
213 		break;
214 	case NPF_TABLE_HASH:
215 		t->t_hashl = hashinit(hsize, HASH_LIST, true, &t->t_hashmask);
216 		if (t->t_hashl == NULL) {
217 			kmem_free(t, sizeof(npf_table_t));
218 			return NULL;
219 		}
220 		break;
221 	default:
222 		KASSERT(false);
223 	}
224 	rw_init(&t->t_lock);
225 	t->t_type = type;
226 	t->t_refcnt = 1;
227 	t->t_id = tid;
228 	return t;
229 }
230 
231 /*
232  * npf_table_destroy: free all table entries and table itself.
233  */
234 void
235 npf_table_destroy(npf_table_t *t)
236 {
237 	npf_tblent_t *e;
238 	u_int n;
239 
240 	switch (t->t_type) {
241 	case NPF_TABLE_HASH:
242 		for (n = 0; n <= t->t_hashmask; n++) {
243 			while ((e = LIST_FIRST(&t->t_hashl[n])) != NULL) {
244 				LIST_REMOVE(e, te_entry.hashq);
245 				pool_cache_put(tblent_cache, e);
246 			}
247 		}
248 		hashdone(t->t_hashl, HASH_LIST, t->t_hashmask);
249 		break;
250 	case NPF_TABLE_TREE:
251 		while ((e = rb_tree_iterate(&t->t_rbtree, NULL,
252 		    RB_DIR_LEFT)) != NULL) {
253 			rb_tree_remove_node(&t->t_rbtree, e);
254 			pool_cache_put(tblent_cache, e);
255 		}
256 		break;
257 	default:
258 		KASSERT(false);
259 	}
260 	rw_destroy(&t->t_lock);
261 	kmem_free(t, sizeof(npf_table_t));
262 }
263 
264 /*
265  * npf_table_ref: holds the reference on table.
266  *
267  * => Table must be locked.
268  */
269 void
270 npf_table_ref(npf_table_t *t)
271 {
272 
273 	KASSERT(rw_lock_held(&t->t_lock));
274 	atomic_inc_uint(&t->t_refcnt);
275 }
276 
277 /*
278  * npf_table_unref: drop reference from the table and destroy the table if
279  * it is the last reference.
280  */
281 void
282 npf_table_unref(npf_table_t *t)
283 {
284 
285 	if (atomic_dec_uint_nv(&t->t_refcnt) != 0) {
286 		return;
287 	}
288 	npf_table_destroy(t);
289 }
290 
291 /*
292  * npf_table_get: find the table according to ID and "get it" by locking it.
293  */
294 npf_table_t *
295 npf_table_get(npf_tableset_t *tset, u_int tid)
296 {
297 	npf_table_t *t;
298 
299 	KASSERT(tset != NULL);
300 
301 	if ((u_int)tid >= NPF_TABLE_SLOTS) {
302 		return NULL;
303 	}
304 	t = tset[tid];
305 	if (t != NULL) {
306 		rw_enter(&t->t_lock, RW_READER);
307 	}
308 	return t;
309 }
310 
311 /*
312  * npf_table_put: "put table back" by unlocking it.
313  */
314 void
315 npf_table_put(npf_table_t *t)
316 {
317 
318 	rw_exit(&t->t_lock);
319 }
320 
321 /*
322  * npf_table_check: validate ID and type.
323  * */
324 int
325 npf_table_check(npf_tableset_t *tset, u_int tid, int type)
326 {
327 
328 	if ((u_int)tid >= NPF_TABLE_SLOTS) {
329 		return EINVAL;
330 	}
331 	if (tset[tid] != NULL) {
332 		return EEXIST;
333 	}
334 	if (type != NPF_TABLE_TREE && type != NPF_TABLE_HASH) {
335 		return EINVAL;
336 	}
337 	return 0;
338 }
339 
340 /*
341  * npf_table_add_cidr: add an IPv4 or IPv6 CIDR into the table.
342  */
343 int
344 npf_table_add_cidr(npf_tableset_t *tset, u_int tid,
345     const npf_addr_t *addr, const npf_netmask_t mask)
346 {
347 	struct npf_hashl *htbl;
348 	npf_tblent_t *e, *it;
349 	npf_table_t *t;
350 	npf_addr_t val;
351 	int error = 0;
352 
353 	if (mask > NPF_MAX_NETMASK) {
354 		return EINVAL;
355 	}
356 	e = pool_cache_get(tblent_cache, PR_WAITOK);
357 	memcpy(&e->te_addr, addr, sizeof(npf_addr_t));
358 	e->te_mask = mask;
359 
360 	/* Get the table (acquire the lock). */
361 	t = npf_table_get(tset, tid);
362 	if (t == NULL) {
363 		pool_cache_put(tblent_cache, e);
364 		return EINVAL;
365 	}
366 	switch (t->t_type) {
367 	case NPF_TABLE_HASH:
368 		/* Generate hash value from: address & mask. */
369 		npf_calculate_masked_addr(&val, addr, mask);
370 		htbl = table_hash_bucket(t, &val, sizeof(npf_addr_t));
371 		/* Lookup to check for duplicates. */
372 		LIST_FOREACH(it, htbl, te_entry.hashq) {
373 			if (it->te_mask != mask) {
374 				continue;
375 			}
376 			if (!memcmp(&it->te_addr, addr, sizeof(npf_addr_t))) {
377 				break;
378 			}
379 		}
380 		/* If no duplicate - insert entry. */
381 		if (__predict_true(it == NULL)) {
382 			LIST_INSERT_HEAD(htbl, e, te_entry.hashq);
383 		} else {
384 			error = EEXIST;
385 		}
386 		break;
387 	case NPF_TABLE_TREE:
388 		/* Insert entry.  Returns false, if duplicate. */
389 		if (rb_tree_insert_node(&t->t_rbtree, e) != e) {
390 			error = EEXIST;
391 		}
392 		break;
393 	default:
394 		KASSERT(false);
395 	}
396 	npf_table_put(t);
397 
398 	if (error) {
399 		pool_cache_put(tblent_cache, e);
400 	}
401 	return error;
402 }
403 
404 /*
405  * npf_table_rem_v4cidr: remove an IPv4 CIDR from the table.
406  */
407 int
408 npf_table_rem_cidr(npf_tableset_t *tset, u_int tid,
409     const npf_addr_t *addr, const npf_netmask_t mask)
410 {
411 	struct npf_hashl *htbl;
412 	npf_tblent_t *e;
413 	npf_table_t *t;
414 	npf_addr_t val;
415 	int error;
416 
417 	if (mask > NPF_MAX_NETMASK) {
418 		return EINVAL;
419 	}
420 
421 	/* Get the table (acquire the lock). */
422 	t = npf_table_get(tset, tid);
423 	if (__predict_false(t == NULL)) {
424 		return EINVAL;
425 	}
426 	e = NULL;
427 
428 	switch (t->t_type) {
429 	case NPF_TABLE_HASH:
430 		/* Generate hash value from: (address & mask). */
431 		npf_calculate_masked_addr(&val, addr, mask);
432 		htbl = table_hash_bucket(t, &val, sizeof(npf_addr_t));
433 		LIST_FOREACH(e, htbl, te_entry.hashq) {
434 			if (e->te_mask != mask) {
435 				continue;
436 			}
437 			if (!memcmp(&e->te_addr, addr, sizeof(npf_addr_t))) {
438 				break;
439 			}
440 		}
441 		if (__predict_true(e != NULL)) {
442 			LIST_REMOVE(e, te_entry.hashq);
443 		} else {
444 			error = ESRCH;
445 		}
446 		break;
447 	case NPF_TABLE_TREE:
448 		/* Key: (address & mask). */
449 		npf_calculate_masked_addr(&val, addr, mask);
450 		e = rb_tree_find_node(&t->t_rbtree, &val);
451 		if (__predict_true(e != NULL)) {
452 			rb_tree_remove_node(&t->t_rbtree, e);
453 		} else {
454 			error = ESRCH;
455 		}
456 		break;
457 	default:
458 		KASSERT(false);
459 	}
460 	npf_table_put(t);
461 
462 	if (e == NULL) {
463 		return ENOENT;
464 	}
465 	pool_cache_put(tblent_cache, e);
466 	return 0;
467 }
468 
469 /*
470  * npf_table_match_addr: find the table according to ID, lookup and
471  * match the contents with specified IPv4 address.
472  */
473 int
474 npf_table_match_addr(npf_tableset_t *tset, u_int tid, const npf_addr_t *addr)
475 {
476 	struct npf_hashl *htbl;
477 	npf_tblent_t *e = NULL;
478 	npf_table_t *t;
479 
480 	/* Get the table (acquire the lock). */
481 	t = npf_table_get(tset, tid);
482 	if (__predict_false(t == NULL)) {
483 		return EINVAL;
484 	}
485 	switch (t->t_type) {
486 	case NPF_TABLE_HASH:
487 		htbl = table_hash_bucket(t, addr, sizeof(npf_addr_t));
488 		LIST_FOREACH(e, htbl, te_entry.hashq) {
489 			if (npf_compare_cidr(addr, e->te_mask, &e->te_addr,
490 			    NPF_NO_NETMASK) == 0)
491 				break;
492 		}
493 		break;
494 	case NPF_TABLE_TREE:
495 		e = rb_tree_find_node(&t->t_rbtree, addr);
496 		KASSERT(e && npf_compare_cidr(addr, e->te_mask, &e->te_addr,
497 		    NPF_NO_NETMASK) == 0);
498 		break;
499 	default:
500 		KASSERT(false);
501 	}
502 	npf_table_put(t);
503 
504 	return e ? 0 : ENOENT;
505 }
506