Training courses

Kernel and Embedded Linux

Bootlin training courses

Embedded Linux, kernel,
Yocto Project, Buildroot, real-time,
graphics, boot time, debugging...

Bootlin logo

Elixir Cross Referencer

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
/*
 * Copyright (c) 2016 Thomas Pornin <pornin@bolet.org>
 *
 * Permission is hereby granted, free of charge, to any person obtaining 
 * a copy of this software and associated documentation files (the
 * "Software"), to deal in the Software without restriction, including
 * without limitation the rights to use, copy, modify, merge, publish,
 * distribute, sublicense, and/or sell copies of the Software, and to
 * permit persons to whom the Software is furnished to do so, subject to
 * the following conditions:
 *
 * The above copyright notice and this permission notice shall be 
 * included in all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, 
 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND 
 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
 * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
 * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 * SOFTWARE.
 */

#include "inner.h"

/*
 * Each entry consists in a fixed number of bytes. Entries are concatenated
 * in the store block. "Addresses" are really offsets in the block,
 * expressed over 32 bits (so the cache may have size at most 4 GB, which
 * "ought to be enough for everyone"). The "null address" is 0xFFFFFFFF.
 * Note that since the storage block alignment is in no way guaranteed, we
 * perform only accesses that can handle unaligned data.
 *
 * Two concurrent data structures are maintained:
 *
 * -- Entries are organised in a doubly-linked list; saved entries are added
 * at the head, and loaded entries are moved to the head. Eviction uses
 * the list tail (this is the LRU algorithm).
 *
 * -- Entries are indexed with a binary tree: all left descendants of a
 * node have a lower session ID (in lexicographic order), while all
 * right descendants have a higher session ID. The tree is heuristically
 * balanced.
 *
 * Entry format:
 *
 *   session ID          32 bytes
 *   master secret       48 bytes
 *   protocol version    2 bytes (big endian)
 *   cipher suite        2 bytes (big endian)
 *   list prev           4 bytes (big endian)
 *   list next           4 bytes (big endian)
 *   tree left child     4 bytes (big endian)
 *   tree right child    4 bytes (big endian)
 *
 * If an entry has a protocol version set to 0, then it is "disabled":
 * it was a session pushed to the cache at some point, but it has
 * been explicitly removed.
 *
 * We need to keep the tree balanced because an attacker could make
 * handshakes, selecting some specific sessions (by reusing them) to
 * try to make us make an imbalanced tree that makes lookups expensive
 * (a denial-of-service attack that would persist as long as the cache
 * remains, i.e. even after the attacker made all his connections).
 * To do that, we replace the session ID (or the start of the session ID)
 * with a HMAC value computed over the replaced part; the hash function
 * implementation and the key are obtained from the server context upon
 * first save() call.
 *
 * Theoretically, an attacker could use the exact timing of the lookup
 * to infer the current tree topology, and try to revive entries to make
 * it as unbalanced as possible. However, since the session ID are
 * chosen randomly by the server, and the attacker cannot see the
 * indexing values and must thus rely on blind selection, it should be
 * exponentially difficult for the attacker to maintain a large
 * imbalance.
 */
#define SESSION_ID_LEN       32
#define MASTER_SECRET_LEN    48

#define SESSION_ID_OFF        0
#define MASTER_SECRET_OFF    32
#define VERSION_OFF          80
#define CIPHER_SUITE_OFF     82
#define LIST_PREV_OFF        84
#define LIST_NEXT_OFF        88
#define TREE_LEFT_OFF        92
#define TREE_RIGHT_OFF       96

#define LRU_ENTRY_LEN       100

#define ADDR_NULL   ((uint32_t)-1)

#define GETSET(name, off) \
static inline uint32_t get_ ## name(br_ssl_session_cache_lru *cc, uint32_t x) \
{ \
	return br_dec32be(cc->store + x + (off)); \
} \
static inline void set_ ## name(br_ssl_session_cache_lru *cc, \
	uint32_t x, uint32_t val) \
{ \
	br_enc32be(cc->store + x + (off), val); \
}

GETSET(prev, LIST_PREV_OFF)
GETSET(next, LIST_NEXT_OFF)
GETSET(left, TREE_LEFT_OFF)
GETSET(right, TREE_RIGHT_OFF)

/*
 * Transform the session ID by replacing the first N bytes with a HMAC
 * value computed over these bytes, using the random key K (the HMAC
 * value is truncated if needed). HMAC will use the same hash function
 * as the DRBG in the SSL server context, so with SHA-256, SHA-384,
 * or SHA-1, depending on what is available.
 *
 * The risk of collision is considered too small to be a concern; and
 * the impact of a collision is low (the handshake won't succeed). This
 * risk is much lower than any transmission error, which would lead to
 * the same consequences.
 *
 * Source and destination arrays msut be disjoint.
 */
static void
mask_id(br_ssl_session_cache_lru *cc,
	const unsigned char *src, unsigned char *dst)
{
	br_hmac_key_context hkc;
	br_hmac_context hc;

	memcpy(dst, src, SESSION_ID_LEN);
	br_hmac_key_init(&hkc, cc->hash, cc->index_key, sizeof cc->index_key);
	br_hmac_init(&hc, &hkc, SESSION_ID_LEN);
	br_hmac_update(&hc, src, SESSION_ID_LEN);
	br_hmac_out(&hc, dst);
}

/*
 * Find a node by ID. Returned value is the node address, or ADDR_NULL if
 * the node is not found.
 *
 * If addr_link is not NULL, then '*addr_link' is set to the address of the
 * last followed link. If the found node is the root, or if the tree is
 * empty, then '*addr_link' is set to ADDR_NULL.
 */
static uint32_t
find_node(br_ssl_session_cache_lru *cc, const unsigned char *id,
	uint32_t *addr_link)
{
	uint32_t x, y;

	x = cc->root;
	y = ADDR_NULL;
	while (x != ADDR_NULL) {
		int r;

		r = memcmp(id, cc->store + x + SESSION_ID_OFF, SESSION_ID_LEN);
		if (r < 0) {
			y = x + TREE_LEFT_OFF;
			x = get_left(cc, x);
		} else if (r == 0) {
			if (addr_link != NULL) {
				*addr_link = y;
			}
			return x;
		} else {
			y = x + TREE_RIGHT_OFF;
			x = get_right(cc, x);
		}
	}
	if (addr_link != NULL) {
		*addr_link = y;
	}
	return ADDR_NULL;
}

/*
 * For node x, find its replacement upon removal.
 *
 *  -- If node x has no child, then this returns ADDR_NULL.
 *  -- Otherwise, if node x has a left child, then the replacement is the
 *     rightmost left-descendent.
 *  -- Otherwise, the replacement is the leftmost right-descendent.
 *
 * If a node is returned, then '*al' is set to the address of the field
 * that points to that node. Otherwise (node x has no child), '*al' is
 * set to ADDR_NULL.
 *
 * Note that the replacement node, when found, is always a descendent
 * of node 'x', so it cannot be the tree root. Thus, '*al' can be set
 * to ADDR_NULL only when no node is found and ADDR_NULL is returned.
 */
static uint32_t
find_replacement_node(br_ssl_session_cache_lru *cc, uint32_t x, uint32_t *al)
{
	uint32_t y1, y2;

	y1 = get_left(cc, x);
	if (y1 != ADDR_NULL) {
		y2 = x + TREE_LEFT_OFF;
		for (;;) {
			uint32_t z;

			z = get_right(cc, y1);
			if (z == ADDR_NULL) {
				*al = y2;
				return y1;
			}
			y2 = y1 + TREE_RIGHT_OFF;
			y1 = z;
		}
	}
	y1 = get_right(cc, x);
	if (y1 != ADDR_NULL) {
		y2 = x + TREE_RIGHT_OFF;
		for (;;) {
			uint32_t z;

			z = get_left(cc, y1);
			if (z == ADDR_NULL) {
				*al = y2;
				return y1;
			}
			y2 = y1 + TREE_LEFT_OFF;
			y1 = z;
		}
	}
	*al = ADDR_NULL;
	return ADDR_NULL;
}

/*
 * Set the link at address 'alx' to point to node 'x'. If 'alx' is
 * ADDR_NULL, then this sets the tree root to 'x'.
 */
static inline void
set_link(br_ssl_session_cache_lru *cc, uint32_t alx, uint32_t x)
{
	if (alx == ADDR_NULL) {
		cc->root = x;
	} else {
		br_enc32be(cc->store + alx, x);
	}
}

/*
 * Remove node 'x' from the tree. This function shall not be called if
 * node 'x' is not part of the tree.
 */
static void
remove_node(br_ssl_session_cache_lru *cc, uint32_t x)
{
	uint32_t alx, y, aly;

	/*
	 * Removal algorithm:
	 * ------------------
	 *
	 * - If we remove the root, then the tree becomes empty.
	 *
	 * - If the removed node has no child, then we can simply remove
	 *   it, with nothing else to do.
	 *
	 * - Otherwise, the removed node must be replaced by either its
	 *   rightmost left-descendent, or its leftmost right-descendent.
	 *   The replacement node itself must be removed from its current
	 *   place. By definition, that replacement node has either no
	 *   child, or at most a single child that will replace it in the
	 *   tree.
	 */

	/*
	 * Find node back and its ancestor link. If the node was the
	 * root, then alx is set to ADDR_NULL.
	 */
	find_node(cc, cc->store + x + SESSION_ID_OFF, &alx);

	/*
	 * Find replacement node 'y', and 'aly' is set to the address of
	 * the link to that replacement node. If the removed node has no
	 * child, then both 'y' and 'aly' are set to ADDR_NULL.
	 */
	y = find_replacement_node(cc, x, &aly);

	if (y != ADDR_NULL) {
		uint32_t z;

		/*
		 * The unlinked replacement node may have one child (but
		 * not two) that takes its place.
		 */
		z = get_left(cc, y);
		if (z == ADDR_NULL) {
			z = get_right(cc, y);
		}
		set_link(cc, aly, z);

		/*
		 * Link the replacement node in its new place, overwriting
		 * the current link to the node 'x' (which removes 'x').
		 */
		set_link(cc, alx, y);

		/*
		 * The replacement node adopts the left and right children
		 * of the removed node. Note that this also works even if
		 * the replacement node was a direct descendent of the
		 * removed node, since we unlinked it previously.
		 */
		set_left(cc, y, get_left(cc, x));
		set_right(cc, y, get_right(cc, x));
	} else {
		/*
		 * No replacement, we simply unlink the node 'x'.
		 */
		set_link(cc, alx, ADDR_NULL);
	}
}

static void
lru_save(const br_ssl_session_cache_class **ctx,
	br_ssl_server_context *server_ctx,
	const br_ssl_session_parameters *params)
{
	br_ssl_session_cache_lru *cc;
	unsigned char id[SESSION_ID_LEN];
	uint32_t x, alx;

	cc = (br_ssl_session_cache_lru *)ctx;

	/*
	 * If the buffer is too small, we don't record anything. This
	 * test avoids problems in subsequent code.
	 */
	if (cc->store_len < LRU_ENTRY_LEN) {
		return;
	}

	/*
	 * Upon the first save in a session cache instance, we obtain
	 * a random key for our indexing.
	 */
	if (!cc->init_done) {
		br_hmac_drbg_generate(&server_ctx->eng.rng,
			cc->index_key, sizeof cc->index_key);
		cc->hash = br_hmac_drbg_get_hash(&server_ctx->eng.rng);
		cc->init_done = 1;
	}
	mask_id(cc, params->session_id, id);

	/*
	 * Look for the node in the tree. If the same ID is already used,
	 * then reject it. This is a collision event, which should be
	 * exceedingly rare.
	 * Note: we do NOT record the emplacement here, because the
	 * removal of an entry may change the tree topology.
	 */
	if (find_node(cc, id, NULL) != ADDR_NULL) {
		return;
	}

	/*
	 * Find some room for the new parameters. If the cache is not
	 * full yet, add it to the end of the area and bump the pointer up.
	 * Otherwise, evict the list tail entry. Note that we already
	 * filtered out the case of a ridiculously small buffer that
	 * cannot hold any entry at all; thus, if there is no room for an
	 * extra entry, then the cache cannot be empty.
	 */
	if (cc->store_ptr > (cc->store_len - LRU_ENTRY_LEN)) {
		/*
		 * Evict tail. If the buffer has room for a single entry,
		 * then this may also be the head.
		 */
		x = cc->tail;
		cc->tail = get_prev(cc, x);
		if (cc->tail == ADDR_NULL) {
			cc->head = ADDR_NULL;
		} else {
			set_next(cc, cc->tail, ADDR_NULL);
		}

		/*
		 * Remove the node from the tree.
		 */
		remove_node(cc, x);
	} else {
		/*
		 * Allocate room for new node.
		 */
		x = cc->store_ptr;
		cc->store_ptr += LRU_ENTRY_LEN;
	}

	/*
	 * Find the emplacement for the new node, and link it.
	 */
	find_node(cc, id, &alx);
	set_link(cc, alx, x);
	set_left(cc, x, ADDR_NULL);
	set_right(cc, x, ADDR_NULL);

	/*
	 * New entry becomes new list head. It may also become the list
	 * tail if the cache was empty at that point.
	 */
	if (cc->head == ADDR_NULL) {
		cc->tail = x;
	} else {
		set_prev(cc, cc->head, x);
	}
	set_prev(cc, x, ADDR_NULL);
	set_next(cc, x, cc->head);
	cc->head = x;

	/*
	 * Fill data in the entry.
	 */
	memcpy(cc->store + x + SESSION_ID_OFF, id, SESSION_ID_LEN);
	memcpy(cc->store + x + MASTER_SECRET_OFF,
		params->master_secret, MASTER_SECRET_LEN);
	br_enc16be(cc->store + x + VERSION_OFF, params->version);
	br_enc16be(cc->store + x + CIPHER_SUITE_OFF, params->cipher_suite);
}

static int
lru_load(const br_ssl_session_cache_class **ctx,
	br_ssl_server_context *server_ctx,
	br_ssl_session_parameters *params)
{
	br_ssl_session_cache_lru *cc;
	unsigned char id[SESSION_ID_LEN];
	uint32_t x;

	(void)server_ctx;
	cc = (br_ssl_session_cache_lru *)ctx;
	if (!cc->init_done) {
		return 0;
	}
	mask_id(cc, params->session_id, id);
	x = find_node(cc, id, NULL);
	if (x != ADDR_NULL) {
		unsigned version;

		version = br_dec16be(cc->store + x + VERSION_OFF);
		if (version == 0) {
			/*
			 * Entry is disabled, we pretend we did not find it.
			 * Notably, we don't move it to the front of the
			 * LRU list.
			 */
			return 0;
		}
		params->version = version;
		params->cipher_suite = br_dec16be(
			cc->store + x + CIPHER_SUITE_OFF);
		memcpy(params->master_secret,
			cc->store + x + MASTER_SECRET_OFF,
			MASTER_SECRET_LEN);
		if (x != cc->head) {
			/*
			 * Found node is not at list head, so move
			 * it to the head.
			 */
			uint32_t p, n;

			p = get_prev(cc, x);
			n = get_next(cc, x);
			set_next(cc, p, n);
			if (n == ADDR_NULL) {
				cc->tail = p;
			} else {
				set_prev(cc, n, p);
			}
			set_prev(cc, cc->head, x);
			set_next(cc, x, cc->head);
			set_prev(cc, x, ADDR_NULL);
			cc->head = x;
		}
		return 1;
	}
	return 0;
}

static const br_ssl_session_cache_class lru_class = {
	sizeof(br_ssl_session_cache_lru),
	&lru_save,
	&lru_load
};

/* see inner.h */
void
br_ssl_session_cache_lru_init(br_ssl_session_cache_lru *cc,
	unsigned char *store, size_t store_len)
{
	cc->vtable = &lru_class;
	cc->store = store;
	cc->store_len = store_len;
	cc->store_ptr = 0;
	cc->init_done = 0;
	cc->head = ADDR_NULL;
	cc->tail = ADDR_NULL;
	cc->root = ADDR_NULL;
}

/* see bearssl_ssl.h */
void br_ssl_session_cache_lru_forget(
	br_ssl_session_cache_lru *cc, const unsigned char *id)
{
	unsigned char mid[SESSION_ID_LEN];
	uint32_t addr;

	/*
	 * If the cache is not initialised yet, then it is empty, and
	 * there is nothing to forget.
	 */
	if (!cc->init_done) {
		return;
	}

	/*
	 * Look for the node in the tree. If found, the entry is marked
	 * as "disabled"; it will be reused in due course, as it ages
	 * through the list.
	 *
	 * We do not go through the complex moves of actually releasing
	 * the entry right away because explicitly forgetting sessions
	 * should be a rare event, meant mostly for testing purposes,
	 * so this is not worth the extra code size.
	 */
	mask_id(cc, id, mid);
	addr = find_node(cc, mid, NULL);
	if (addr != ADDR_NULL) {
		br_enc16be(cc->store + addr + VERSION_OFF, 0);
	}
}