1 // SPDX-License-Identifier: GPL-2.0
2 /*
3  * Randomized tests for eBPF longest-prefix-match maps
4  *
5  * This program runs randomized tests against the lpm-bpf-map. It implements a
6  * "Trivial Longest Prefix Match" (tlpm) based on simple, linear, singly linked
7  * lists. The implementation should be pretty straightforward.
8  *
9  * Based on tlpm, this inserts randomized data into bpf-lpm-maps and verifies
10  * the trie-based bpf-map implementation behaves the same way as tlpm.
11  */
12 
13 #include <assert.h>
14 #include <errno.h>
15 #include <inttypes.h>
16 #include <linux/bpf.h>
17 #include <pthread.h>
18 #include <stdio.h>
19 #include <stdlib.h>
20 #include <string.h>
21 #include <time.h>
22 #include <unistd.h>
23 #include <arpa/inet.h>
24 #include <sys/time.h>
25 
26 #include <bpf/bpf.h>
27 
28 #include "bpf_util.h"
29 #include "bpf_rlimit.h"
30 
31 struct tlpm_node {
32 	struct tlpm_node *next;
33 	size_t n_bits;
34 	uint8_t key[];
35 };
36 
37 static struct tlpm_node *tlpm_match(struct tlpm_node *list,
38 				    const uint8_t *key,
39 				    size_t n_bits);
40 
41 static struct tlpm_node *tlpm_add(struct tlpm_node *list,
42 				  const uint8_t *key,
43 				  size_t n_bits)
44 {
45 	struct tlpm_node *node;
46 	size_t n;
47 
48 	n = (n_bits + 7) / 8;
49 
50 	/* 'overwrite' an equivalent entry if one already exists */
51 	node = tlpm_match(list, key, n_bits);
52 	if (node && node->n_bits == n_bits) {
53 		memcpy(node->key, key, n);
54 		return list;
55 	}
56 
57 	/* add new entry with @key/@n_bits to @list and return new head */
58 
59 	node = malloc(sizeof(*node) + n);
60 	assert(node);
61 
62 	node->next = list;
63 	node->n_bits = n_bits;
64 	memcpy(node->key, key, n);
65 
66 	return node;
67 }
68 
69 static void tlpm_clear(struct tlpm_node *list)
70 {
71 	struct tlpm_node *node;
72 
73 	/* free all entries in @list */
74 
75 	while ((node = list)) {
76 		list = list->next;
77 		free(node);
78 	}
79 }
80 
81 static struct tlpm_node *tlpm_match(struct tlpm_node *list,
82 				    const uint8_t *key,
83 				    size_t n_bits)
84 {
85 	struct tlpm_node *best = NULL;
86 	size_t i;
87 
88 	/* Perform longest prefix-match on @key/@n_bits. That is, iterate all
89 	 * entries and match each prefix against @key. Remember the "best"
90 	 * entry we find (i.e., the longest prefix that matches) and return it
91 	 * to the caller when done.
92 	 */
93 
94 	for ( ; list; list = list->next) {
95 		for (i = 0; i < n_bits && i < list->n_bits; ++i) {
96 			if ((key[i / 8] & (1 << (7 - i % 8))) !=
97 			    (list->key[i / 8] & (1 << (7 - i % 8))))
98 				break;
99 		}
100 
101 		if (i >= list->n_bits) {
102 			if (!best || i > best->n_bits)
103 				best = list;
104 		}
105 	}
106 
107 	return best;
108 }
109 
110 static struct tlpm_node *tlpm_delete(struct tlpm_node *list,
111 				     const uint8_t *key,
112 				     size_t n_bits)
113 {
114 	struct tlpm_node *best = tlpm_match(list, key, n_bits);
115 	struct tlpm_node *node;
116 
117 	if (!best || best->n_bits != n_bits)
118 		return list;
119 
120 	if (best == list) {
121 		node = best->next;
122 		free(best);
123 		return node;
124 	}
125 
126 	for (node = list; node; node = node->next) {
127 		if (node->next == best) {
128 			node->next = best->next;
129 			free(best);
130 			return list;
131 		}
132 	}
133 	/* should never get here */
134 	assert(0);
135 	return list;
136 }
137 
138 static void test_lpm_basic(void)
139 {
140 	struct tlpm_node *list = NULL, *t1, *t2;
141 
142 	/* very basic, static tests to verify tlpm works as expected */
143 
144 	assert(!tlpm_match(list, (uint8_t[]){ 0xff }, 8));
145 
146 	t1 = list = tlpm_add(list, (uint8_t[]){ 0xff }, 8);
147 	assert(t1 == tlpm_match(list, (uint8_t[]){ 0xff }, 8));
148 	assert(t1 == tlpm_match(list, (uint8_t[]){ 0xff, 0xff }, 16));
149 	assert(t1 == tlpm_match(list, (uint8_t[]){ 0xff, 0x00 }, 16));
150 	assert(!tlpm_match(list, (uint8_t[]){ 0x7f }, 8));
151 	assert(!tlpm_match(list, (uint8_t[]){ 0xfe }, 8));
152 	assert(!tlpm_match(list, (uint8_t[]){ 0xff }, 7));
153 
154 	t2 = list = tlpm_add(list, (uint8_t[]){ 0xff, 0xff }, 16);
155 	assert(t1 == tlpm_match(list, (uint8_t[]){ 0xff }, 8));
156 	assert(t2 == tlpm_match(list, (uint8_t[]){ 0xff, 0xff }, 16));
157 	assert(t1 == tlpm_match(list, (uint8_t[]){ 0xff, 0xff }, 15));
158 	assert(!tlpm_match(list, (uint8_t[]){ 0x7f, 0xff }, 16));
159 
160 	list = tlpm_delete(list, (uint8_t[]){ 0xff, 0xff }, 16);
161 	assert(t1 == tlpm_match(list, (uint8_t[]){ 0xff }, 8));
162 	assert(t1 == tlpm_match(list, (uint8_t[]){ 0xff, 0xff }, 16));
163 
164 	list = tlpm_delete(list, (uint8_t[]){ 0xff }, 8);
165 	assert(!tlpm_match(list, (uint8_t[]){ 0xff }, 8));
166 
167 	tlpm_clear(list);
168 }
169 
170 static void test_lpm_order(void)
171 {
172 	struct tlpm_node *t1, *t2, *l1 = NULL, *l2 = NULL;
173 	size_t i, j;
174 
175 	/* Verify the tlpm implementation works correctly regardless of the
176 	 * order of entries. Insert a random set of entries into @l1, and copy
177 	 * the same data in reverse order into @l2. Then verify a lookup of
178 	 * random keys will yield the same result in both sets.
179 	 */
180 
181 	for (i = 0; i < (1 << 12); ++i)
182 		l1 = tlpm_add(l1, (uint8_t[]){
183 					rand() % 0xff,
184 					rand() % 0xff,
185 				}, rand() % 16 + 1);
186 
187 	for (t1 = l1; t1; t1 = t1->next)
188 		l2 = tlpm_add(l2, t1->key, t1->n_bits);
189 
190 	for (i = 0; i < (1 << 8); ++i) {
191 		uint8_t key[] = { rand() % 0xff, rand() % 0xff };
192 
193 		t1 = tlpm_match(l1, key, 16);
194 		t2 = tlpm_match(l2, key, 16);
195 
196 		assert(!t1 == !t2);
197 		if (t1) {
198 			assert(t1->n_bits == t2->n_bits);
199 			for (j = 0; j < t1->n_bits; ++j)
200 				assert((t1->key[j / 8] & (1 << (7 - j % 8))) ==
201 				       (t2->key[j / 8] & (1 << (7 - j % 8))));
202 		}
203 	}
204 
205 	tlpm_clear(l1);
206 	tlpm_clear(l2);
207 }
208 
209 static void test_lpm_map(int keysize)
210 {
211 	size_t i, j, n_matches, n_matches_after_delete, n_nodes, n_lookups;
212 	struct tlpm_node *t, *list = NULL;
213 	struct bpf_lpm_trie_key *key;
214 	uint8_t *data, *value;
215 	int r, map;
216 
217 	/* Compare behavior of tlpm vs. bpf-lpm. Create a randomized set of
218 	 * prefixes and insert it into both tlpm and bpf-lpm. Then run some
219 	 * randomized lookups and verify both maps return the same result.
220 	 */
221 
222 	n_matches = 0;
223 	n_matches_after_delete = 0;
224 	n_nodes = 1 << 8;
225 	n_lookups = 1 << 16;
226 
227 	data = alloca(keysize);
228 	memset(data, 0, keysize);
229 
230 	value = alloca(keysize + 1);
231 	memset(value, 0, keysize + 1);
232 
233 	key = alloca(sizeof(*key) + keysize);
234 	memset(key, 0, sizeof(*key) + keysize);
235 
236 	map = bpf_create_map(BPF_MAP_TYPE_LPM_TRIE,
237 			     sizeof(*key) + keysize,
238 			     keysize + 1,
239 			     4096,
240 			     BPF_F_NO_PREALLOC);
241 	assert(map >= 0);
242 
243 	for (i = 0; i < n_nodes; ++i) {
244 		for (j = 0; j < keysize; ++j)
245 			value[j] = rand() & 0xff;
246 		value[keysize] = rand() % (8 * keysize + 1);
247 
248 		list = tlpm_add(list, value, value[keysize]);
249 
250 		key->prefixlen = value[keysize];
251 		memcpy(key->data, value, keysize);
252 		r = bpf_map_update_elem(map, key, value, 0);
253 		assert(!r);
254 	}
255 
256 	for (i = 0; i < n_lookups; ++i) {
257 		for (j = 0; j < keysize; ++j)
258 			data[j] = rand() & 0xff;
259 
260 		t = tlpm_match(list, data, 8 * keysize);
261 
262 		key->prefixlen = 8 * keysize;
263 		memcpy(key->data, data, keysize);
264 		r = bpf_map_lookup_elem(map, key, value);
265 		assert(!r || errno == ENOENT);
266 		assert(!t == !!r);
267 
268 		if (t) {
269 			++n_matches;
270 			assert(t->n_bits == value[keysize]);
271 			for (j = 0; j < t->n_bits; ++j)
272 				assert((t->key[j / 8] & (1 << (7 - j % 8))) ==
273 				       (value[j / 8] & (1 << (7 - j % 8))));
274 		}
275 	}
276 
277 	/* Remove the first half of the elements in the tlpm and the
278 	 * corresponding nodes from the bpf-lpm.  Then run the same
279 	 * large number of random lookups in both and make sure they match.
280 	 * Note: we need to count the number of nodes actually inserted
281 	 * since there may have been duplicates.
282 	 */
283 	for (i = 0, t = list; t; i++, t = t->next)
284 		;
285 	for (j = 0; j < i / 2; ++j) {
286 		key->prefixlen = list->n_bits;
287 		memcpy(key->data, list->key, keysize);
288 		r = bpf_map_delete_elem(map, key);
289 		assert(!r);
290 		list = tlpm_delete(list, list->key, list->n_bits);
291 		assert(list);
292 	}
293 	for (i = 0; i < n_lookups; ++i) {
294 		for (j = 0; j < keysize; ++j)
295 			data[j] = rand() & 0xff;
296 
297 		t = tlpm_match(list, data, 8 * keysize);
298 
299 		key->prefixlen = 8 * keysize;
300 		memcpy(key->data, data, keysize);
301 		r = bpf_map_lookup_elem(map, key, value);
302 		assert(!r || errno == ENOENT);
303 		assert(!t == !!r);
304 
305 		if (t) {
306 			++n_matches_after_delete;
307 			assert(t->n_bits == value[keysize]);
308 			for (j = 0; j < t->n_bits; ++j)
309 				assert((t->key[j / 8] & (1 << (7 - j % 8))) ==
310 				       (value[j / 8] & (1 << (7 - j % 8))));
311 		}
312 	}
313 
314 	close(map);
315 	tlpm_clear(list);
316 
317 	/* With 255 random nodes in the map, we are pretty likely to match
318 	 * something on every lookup. For statistics, use this:
319 	 *
320 	 *     printf("          nodes: %zu\n"
321 	 *            "        lookups: %zu\n"
322 	 *            "        matches: %zu\n"
323 	 *            "matches(delete): %zu\n",
324 	 *            n_nodes, n_lookups, n_matches, n_matches_after_delete);
325 	 */
326 }
327 
328 /* Test the implementation with some 'real world' examples */
329 
330 static void test_lpm_ipaddr(void)
331 {
332 	struct bpf_lpm_trie_key *key_ipv4;
333 	struct bpf_lpm_trie_key *key_ipv6;
334 	size_t key_size_ipv4;
335 	size_t key_size_ipv6;
336 	int map_fd_ipv4;
337 	int map_fd_ipv6;
338 	__u64 value;
339 
340 	key_size_ipv4 = sizeof(*key_ipv4) + sizeof(__u32);
341 	key_size_ipv6 = sizeof(*key_ipv6) + sizeof(__u32) * 4;
342 	key_ipv4 = alloca(key_size_ipv4);
343 	key_ipv6 = alloca(key_size_ipv6);
344 
345 	map_fd_ipv4 = bpf_create_map(BPF_MAP_TYPE_LPM_TRIE,
346 				     key_size_ipv4, sizeof(value),
347 				     100, BPF_F_NO_PREALLOC);
348 	assert(map_fd_ipv4 >= 0);
349 
350 	map_fd_ipv6 = bpf_create_map(BPF_MAP_TYPE_LPM_TRIE,
351 				     key_size_ipv6, sizeof(value),
352 				     100, BPF_F_NO_PREALLOC);
353 	assert(map_fd_ipv6 >= 0);
354 
355 	/* Fill data some IPv4 and IPv6 address ranges */
356 	value = 1;
357 	key_ipv4->prefixlen = 16;
358 	inet_pton(AF_INET, "192.168.0.0", key_ipv4->data);
359 	assert(bpf_map_update_elem(map_fd_ipv4, key_ipv4, &value, 0) == 0);
360 
361 	value = 2;
362 	key_ipv4->prefixlen = 24;
363 	inet_pton(AF_INET, "192.168.0.0", key_ipv4->data);
364 	assert(bpf_map_update_elem(map_fd_ipv4, key_ipv4, &value, 0) == 0);
365 
366 	value = 3;
367 	key_ipv4->prefixlen = 24;
368 	inet_pton(AF_INET, "192.168.128.0", key_ipv4->data);
369 	assert(bpf_map_update_elem(map_fd_ipv4, key_ipv4, &value, 0) == 0);
370 
371 	value = 5;
372 	key_ipv4->prefixlen = 24;
373 	inet_pton(AF_INET, "192.168.1.0", key_ipv4->data);
374 	assert(bpf_map_update_elem(map_fd_ipv4, key_ipv4, &value, 0) == 0);
375 
376 	value = 4;
377 	key_ipv4->prefixlen = 23;
378 	inet_pton(AF_INET, "192.168.0.0", key_ipv4->data);
379 	assert(bpf_map_update_elem(map_fd_ipv4, key_ipv4, &value, 0) == 0);
380 
381 	value = 0xdeadbeef;
382 	key_ipv6->prefixlen = 64;
383 	inet_pton(AF_INET6, "2a00:1450:4001:814::200e", key_ipv6->data);
384 	assert(bpf_map_update_elem(map_fd_ipv6, key_ipv6, &value, 0) == 0);
385 
386 	/* Set tprefixlen to maximum for lookups */
387 	key_ipv4->prefixlen = 32;
388 	key_ipv6->prefixlen = 128;
389 
390 	/* Test some lookups that should come back with a value */
391 	inet_pton(AF_INET, "192.168.128.23", key_ipv4->data);
392 	assert(bpf_map_lookup_elem(map_fd_ipv4, key_ipv4, &value) == 0);
393 	assert(value == 3);
394 
395 	inet_pton(AF_INET, "192.168.0.1", key_ipv4->data);
396 	assert(bpf_map_lookup_elem(map_fd_ipv4, key_ipv4, &value) == 0);
397 	assert(value == 2);
398 
399 	inet_pton(AF_INET6, "2a00:1450:4001:814::", key_ipv6->data);
400 	assert(bpf_map_lookup_elem(map_fd_ipv6, key_ipv6, &value) == 0);
401 	assert(value == 0xdeadbeef);
402 
403 	inet_pton(AF_INET6, "2a00:1450:4001:814::1", key_ipv6->data);
404 	assert(bpf_map_lookup_elem(map_fd_ipv6, key_ipv6, &value) == 0);
405 	assert(value == 0xdeadbeef);
406 
407 	/* Test some lookups that should not match any entry */
408 	inet_pton(AF_INET, "10.0.0.1", key_ipv4->data);
409 	assert(bpf_map_lookup_elem(map_fd_ipv4, key_ipv4, &value) == -1 &&
410 	       errno == ENOENT);
411 
412 	inet_pton(AF_INET, "11.11.11.11", key_ipv4->data);
413 	assert(bpf_map_lookup_elem(map_fd_ipv4, key_ipv4, &value) == -1 &&
414 	       errno == ENOENT);
415 
416 	inet_pton(AF_INET6, "2a00:ffff::", key_ipv6->data);
417 	assert(bpf_map_lookup_elem(map_fd_ipv6, key_ipv6, &value) == -1 &&
418 	       errno == ENOENT);
419 
420 	close(map_fd_ipv4);
421 	close(map_fd_ipv6);
422 }
423 
424 static void test_lpm_delete(void)
425 {
426 	struct bpf_lpm_trie_key *key;
427 	size_t key_size;
428 	int map_fd;
429 	__u64 value;
430 
431 	key_size = sizeof(*key) + sizeof(__u32);
432 	key = alloca(key_size);
433 
434 	map_fd = bpf_create_map(BPF_MAP_TYPE_LPM_TRIE,
435 				key_size, sizeof(value),
436 				100, BPF_F_NO_PREALLOC);
437 	assert(map_fd >= 0);
438 
439 	/* Add nodes:
440 	 * 192.168.0.0/16   (1)
441 	 * 192.168.0.0/24   (2)
442 	 * 192.168.128.0/24 (3)
443 	 * 192.168.1.0/24   (4)
444 	 *
445 	 *         (1)
446 	 *        /   \
447          *     (IM)    (3)
448 	 *    /   \
449          *   (2)  (4)
450 	 */
451 	value = 1;
452 	key->prefixlen = 16;
453 	inet_pton(AF_INET, "192.168.0.0", key->data);
454 	assert(bpf_map_update_elem(map_fd, key, &value, 0) == 0);
455 
456 	value = 2;
457 	key->prefixlen = 24;
458 	inet_pton(AF_INET, "192.168.0.0", key->data);
459 	assert(bpf_map_update_elem(map_fd, key, &value, 0) == 0);
460 
461 	value = 3;
462 	key->prefixlen = 24;
463 	inet_pton(AF_INET, "192.168.128.0", key->data);
464 	assert(bpf_map_update_elem(map_fd, key, &value, 0) == 0);
465 
466 	value = 4;
467 	key->prefixlen = 24;
468 	inet_pton(AF_INET, "192.168.1.0", key->data);
469 	assert(bpf_map_update_elem(map_fd, key, &value, 0) == 0);
470 
471 	/* remove non-existent node */
472 	key->prefixlen = 32;
473 	inet_pton(AF_INET, "10.0.0.1", key->data);
474 	assert(bpf_map_lookup_elem(map_fd, key, &value) == -1 &&
475 		errno == ENOENT);
476 
477 	key->prefixlen = 30; // unused prefix so far
478 	inet_pton(AF_INET, "192.255.0.0", key->data);
479 	assert(bpf_map_delete_elem(map_fd, key) == -1 &&
480 		errno == ENOENT);
481 
482 	key->prefixlen = 16; // same prefix as the root node
483 	inet_pton(AF_INET, "192.255.0.0", key->data);
484 	assert(bpf_map_delete_elem(map_fd, key) == -1 &&
485 		errno == ENOENT);
486 
487 	/* assert initial lookup */
488 	key->prefixlen = 32;
489 	inet_pton(AF_INET, "192.168.0.1", key->data);
490 	assert(bpf_map_lookup_elem(map_fd, key, &value) == 0);
491 	assert(value == 2);
492 
493 	/* remove leaf node */
494 	key->prefixlen = 24;
495 	inet_pton(AF_INET, "192.168.0.0", key->data);
496 	assert(bpf_map_delete_elem(map_fd, key) == 0);
497 
498 	key->prefixlen = 32;
499 	inet_pton(AF_INET, "192.168.0.1", key->data);
500 	assert(bpf_map_lookup_elem(map_fd, key, &value) == 0);
501 	assert(value == 1);
502 
503 	/* remove leaf (and intermediary) node */
504 	key->prefixlen = 24;
505 	inet_pton(AF_INET, "192.168.1.0", key->data);
506 	assert(bpf_map_delete_elem(map_fd, key) == 0);
507 
508 	key->prefixlen = 32;
509 	inet_pton(AF_INET, "192.168.1.1", key->data);
510 	assert(bpf_map_lookup_elem(map_fd, key, &value) == 0);
511 	assert(value == 1);
512 
513 	/* remove root node */
514 	key->prefixlen = 16;
515 	inet_pton(AF_INET, "192.168.0.0", key->data);
516 	assert(bpf_map_delete_elem(map_fd, key) == 0);
517 
518 	key->prefixlen = 32;
519 	inet_pton(AF_INET, "192.168.128.1", key->data);
520 	assert(bpf_map_lookup_elem(map_fd, key, &value) == 0);
521 	assert(value == 3);
522 
523 	/* remove last node */
524 	key->prefixlen = 24;
525 	inet_pton(AF_INET, "192.168.128.0", key->data);
526 	assert(bpf_map_delete_elem(map_fd, key) == 0);
527 
528 	key->prefixlen = 32;
529 	inet_pton(AF_INET, "192.168.128.1", key->data);
530 	assert(bpf_map_lookup_elem(map_fd, key, &value) == -1 &&
531 		errno == ENOENT);
532 
533 	close(map_fd);
534 }
535 
536 static void test_lpm_get_next_key(void)
537 {
538 	struct bpf_lpm_trie_key *key_p, *next_key_p;
539 	size_t key_size;
540 	__u32 value = 0;
541 	int map_fd;
542 
543 	key_size = sizeof(*key_p) + sizeof(__u32);
544 	key_p = alloca(key_size);
545 	next_key_p = alloca(key_size);
546 
547 	map_fd = bpf_create_map(BPF_MAP_TYPE_LPM_TRIE, key_size, sizeof(value),
548 				100, BPF_F_NO_PREALLOC);
549 	assert(map_fd >= 0);
550 
551 	/* empty tree. get_next_key should return ENOENT */
552 	assert(bpf_map_get_next_key(map_fd, NULL, key_p) == -1 &&
553 	       errno == ENOENT);
554 
555 	/* get and verify the first key, get the second one should fail. */
556 	key_p->prefixlen = 16;
557 	inet_pton(AF_INET, "192.168.0.0", key_p->data);
558 	assert(bpf_map_update_elem(map_fd, key_p, &value, 0) == 0);
559 
560 	memset(key_p, 0, key_size);
561 	assert(bpf_map_get_next_key(map_fd, NULL, key_p) == 0);
562 	assert(key_p->prefixlen == 16 && key_p->data[0] == 192 &&
563 	       key_p->data[1] == 168);
564 
565 	assert(bpf_map_get_next_key(map_fd, key_p, next_key_p) == -1 &&
566 	       errno == ENOENT);
567 
568 	/* no exact matching key should get the first one in post order. */
569 	key_p->prefixlen = 8;
570 	assert(bpf_map_get_next_key(map_fd, NULL, key_p) == 0);
571 	assert(key_p->prefixlen == 16 && key_p->data[0] == 192 &&
572 	       key_p->data[1] == 168);
573 
574 	/* add one more element (total two) */
575 	key_p->prefixlen = 24;
576 	inet_pton(AF_INET, "192.168.0.0", key_p->data);
577 	assert(bpf_map_update_elem(map_fd, key_p, &value, 0) == 0);
578 
579 	memset(key_p, 0, key_size);
580 	assert(bpf_map_get_next_key(map_fd, NULL, key_p) == 0);
581 	assert(key_p->prefixlen == 24 && key_p->data[0] == 192 &&
582 	       key_p->data[1] == 168 && key_p->data[2] == 0);
583 
584 	memset(next_key_p, 0, key_size);
585 	assert(bpf_map_get_next_key(map_fd, key_p, next_key_p) == 0);
586 	assert(next_key_p->prefixlen == 16 && next_key_p->data[0] == 192 &&
587 	       next_key_p->data[1] == 168);
588 
589 	memcpy(key_p, next_key_p, key_size);
590 	assert(bpf_map_get_next_key(map_fd, key_p, next_key_p) == -1 &&
591 	       errno == ENOENT);
592 
593 	/* Add one more element (total three) */
594 	key_p->prefixlen = 24;
595 	inet_pton(AF_INET, "192.168.128.0", key_p->data);
596 	assert(bpf_map_update_elem(map_fd, key_p, &value, 0) == 0);
597 
598 	memset(key_p, 0, key_size);
599 	assert(bpf_map_get_next_key(map_fd, NULL, key_p) == 0);
600 	assert(key_p->prefixlen == 24 && key_p->data[0] == 192 &&
601 	       key_p->data[1] == 168 && key_p->data[2] == 0);
602 
603 	memset(next_key_p, 0, key_size);
604 	assert(bpf_map_get_next_key(map_fd, key_p, next_key_p) == 0);
605 	assert(next_key_p->prefixlen == 24 && next_key_p->data[0] == 192 &&
606 	       next_key_p->data[1] == 168 && next_key_p->data[2] == 128);
607 
608 	memcpy(key_p, next_key_p, key_size);
609 	assert(bpf_map_get_next_key(map_fd, key_p, next_key_p) == 0);
610 	assert(next_key_p->prefixlen == 16 && next_key_p->data[0] == 192 &&
611 	       next_key_p->data[1] == 168);
612 
613 	memcpy(key_p, next_key_p, key_size);
614 	assert(bpf_map_get_next_key(map_fd, key_p, next_key_p) == -1 &&
615 	       errno == ENOENT);
616 
617 	/* Add one more element (total four) */
618 	key_p->prefixlen = 24;
619 	inet_pton(AF_INET, "192.168.1.0", key_p->data);
620 	assert(bpf_map_update_elem(map_fd, key_p, &value, 0) == 0);
621 
622 	memset(key_p, 0, key_size);
623 	assert(bpf_map_get_next_key(map_fd, NULL, key_p) == 0);
624 	assert(key_p->prefixlen == 24 && key_p->data[0] == 192 &&
625 	       key_p->data[1] == 168 && key_p->data[2] == 0);
626 
627 	memset(next_key_p, 0, key_size);
628 	assert(bpf_map_get_next_key(map_fd, key_p, next_key_p) == 0);
629 	assert(next_key_p->prefixlen == 24 && next_key_p->data[0] == 192 &&
630 	       next_key_p->data[1] == 168 && next_key_p->data[2] == 1);
631 
632 	memcpy(key_p, next_key_p, key_size);
633 	assert(bpf_map_get_next_key(map_fd, key_p, next_key_p) == 0);
634 	assert(next_key_p->prefixlen == 24 && next_key_p->data[0] == 192 &&
635 	       next_key_p->data[1] == 168 && next_key_p->data[2] == 128);
636 
637 	memcpy(key_p, next_key_p, key_size);
638 	assert(bpf_map_get_next_key(map_fd, key_p, next_key_p) == 0);
639 	assert(next_key_p->prefixlen == 16 && next_key_p->data[0] == 192 &&
640 	       next_key_p->data[1] == 168);
641 
642 	memcpy(key_p, next_key_p, key_size);
643 	assert(bpf_map_get_next_key(map_fd, key_p, next_key_p) == -1 &&
644 	       errno == ENOENT);
645 
646 	/* no exact matching key should return the first one in post order */
647 	key_p->prefixlen = 22;
648 	inet_pton(AF_INET, "192.168.1.0", key_p->data);
649 	assert(bpf_map_get_next_key(map_fd, key_p, next_key_p) == 0);
650 	assert(next_key_p->prefixlen == 24 && next_key_p->data[0] == 192 &&
651 	       next_key_p->data[1] == 168 && next_key_p->data[2] == 0);
652 
653 	close(map_fd);
654 }
655 
656 #define MAX_TEST_KEYS	4
657 struct lpm_mt_test_info {
658 	int cmd; /* 0: update, 1: delete, 2: lookup, 3: get_next_key */
659 	int iter;
660 	int map_fd;
661 	struct {
662 		__u32 prefixlen;
663 		__u32 data;
664 	} key[MAX_TEST_KEYS];
665 };
666 
667 static void *lpm_test_command(void *arg)
668 {
669 	int i, j, ret, iter, key_size;
670 	struct lpm_mt_test_info *info = arg;
671 	struct bpf_lpm_trie_key *key_p;
672 
673 	key_size = sizeof(struct bpf_lpm_trie_key) + sizeof(__u32);
674 	key_p = alloca(key_size);
675 	for (iter = 0; iter < info->iter; iter++)
676 		for (i = 0; i < MAX_TEST_KEYS; i++) {
677 			/* first half of iterations in forward order,
678 			 * and second half in backward order.
679 			 */
680 			j = (iter < (info->iter / 2)) ? i : MAX_TEST_KEYS - i - 1;
681 			key_p->prefixlen = info->key[j].prefixlen;
682 			memcpy(key_p->data, &info->key[j].data, sizeof(__u32));
683 			if (info->cmd == 0) {
684 				__u32 value = j;
685 				/* update must succeed */
686 				assert(bpf_map_update_elem(info->map_fd, key_p, &value, 0) == 0);
687 			} else if (info->cmd == 1) {
688 				ret = bpf_map_delete_elem(info->map_fd, key_p);
689 				assert(ret == 0 || errno == ENOENT);
690 			} else if (info->cmd == 2) {
691 				__u32 value;
692 				ret = bpf_map_lookup_elem(info->map_fd, key_p, &value);
693 				assert(ret == 0 || errno == ENOENT);
694 			} else {
695 				struct bpf_lpm_trie_key *next_key_p = alloca(key_size);
696 				ret = bpf_map_get_next_key(info->map_fd, key_p, next_key_p);
697 				assert(ret == 0 || errno == ENOENT || errno == ENOMEM);
698 			}
699 		}
700 
701 	// Pass successful exit info back to the main thread
702 	pthread_exit((void *)info);
703 }
704 
705 static void setup_lpm_mt_test_info(struct lpm_mt_test_info *info, int map_fd)
706 {
707 	info->iter = 2000;
708 	info->map_fd = map_fd;
709 	info->key[0].prefixlen = 16;
710 	inet_pton(AF_INET, "192.168.0.0", &info->key[0].data);
711 	info->key[1].prefixlen = 24;
712 	inet_pton(AF_INET, "192.168.0.0", &info->key[1].data);
713 	info->key[2].prefixlen = 24;
714 	inet_pton(AF_INET, "192.168.128.0", &info->key[2].data);
715 	info->key[3].prefixlen = 24;
716 	inet_pton(AF_INET, "192.168.1.0", &info->key[3].data);
717 }
718 
719 static void test_lpm_multi_thread(void)
720 {
721 	struct lpm_mt_test_info info[4];
722 	size_t key_size, value_size;
723 	pthread_t thread_id[4];
724 	int i, map_fd;
725 	void *ret;
726 
727 	/* create a trie */
728 	value_size = sizeof(__u32);
729 	key_size = sizeof(struct bpf_lpm_trie_key) + value_size;
730 	map_fd = bpf_create_map(BPF_MAP_TYPE_LPM_TRIE, key_size, value_size,
731 				100, BPF_F_NO_PREALLOC);
732 
733 	/* create 4 threads to test update, delete, lookup and get_next_key */
734 	setup_lpm_mt_test_info(&info[0], map_fd);
735 	for (i = 0; i < 4; i++) {
736 		if (i != 0)
737 			memcpy(&info[i], &info[0], sizeof(info[i]));
738 		info[i].cmd = i;
739 		assert(pthread_create(&thread_id[i], NULL, &lpm_test_command, &info[i]) == 0);
740 	}
741 
742 	for (i = 0; i < 4; i++)
743 		assert(pthread_join(thread_id[i], &ret) == 0 && ret == (void *)&info[i]);
744 
745 	close(map_fd);
746 }
747 
748 int main(void)
749 {
750 	int i;
751 
752 	/* we want predictable, pseudo random tests */
753 	srand(0xf00ba1);
754 
755 	test_lpm_basic();
756 	test_lpm_order();
757 
758 	/* Test with 8, 16, 24, 32, ... 128 bit prefix length */
759 	for (i = 1; i <= 16; ++i)
760 		test_lpm_map(i);
761 
762 	test_lpm_ipaddr();
763 	test_lpm_delete();
764 	test_lpm_get_next_key();
765 	test_lpm_multi_thread();
766 
767 	printf("test_lpm: OK\n");
768 	return 0;
769 }
770