random.c 46.5 KB
Newer Older
1
// SPDX-License-Identifier: (GPL-2.0 OR BSD-3-Clause)
L
Linus Torvalds 已提交
2
/*
3
 * Copyright (C) 2017-2022 Jason A. Donenfeld <Jason@zx2c4.com>. All Rights Reserved.
4
 * Copyright Matt Mackall <mpm@selenic.com>, 2003, 2004, 2005
L
Linus Torvalds 已提交
5 6 7 8 9 10 11 12
 * Copyright Theodore Ts'o, 1994, 1995, 1996, 1997, 1998, 1999.  All
 * rights reserved.
 */

/*
 * Exported interfaces ---- output
 * ===============================
 *
13
 * There are four exported interfaces; two for use within the kernel,
schspa's avatar
schspa 已提交
14
 * and two for use from userspace.
L
Linus Torvalds 已提交
15
 *
16 17
 * Exported interfaces ---- userspace output
 * -----------------------------------------
L
Linus Torvalds 已提交
18
 *
19
 * The userspace interfaces are two character devices /dev/random and
L
Linus Torvalds 已提交
20 21 22 23 24 25 26 27 28 29 30 31
 * /dev/urandom.  /dev/random is suitable for use when very high
 * quality randomness is desired (for example, for key generation or
 * one-time pads), as it will only return a maximum of the number of
 * bits of randomness (as estimated by the random number generator)
 * contained in the entropy pool.
 *
 * The /dev/urandom device does not have this limit, and will return
 * as many bytes as are requested.  As more and more random bytes are
 * requested without giving time for the entropy pool to recharge,
 * this will result in random numbers that are merely cryptographically
 * strong.  For many applications, however, this is acceptable.
 *
32 33 34
 * Exported interfaces ---- kernel output
 * --------------------------------------
 *
35
 * The primary kernel interfaces are:
36
 *
37
 *	void get_random_bytes(void *buf, size_t nbytes);
38 39 40 41
 *	u32 get_random_u32()
 *	u64 get_random_u64()
 *	unsigned int get_random_int()
 *	unsigned long get_random_long()
42
 *
43 44 45 46 47
 * These interfaces will return the requested number of random bytes
 * into the given buffer or as a return value. This is equivalent to a
 * read from /dev/urandom. The get_random_{u32,u64,int,long}() family
 * of functions may be higher performance for one-off random integers,
 * because they do a bit of buffering.
48 49 50 51 52 53 54 55 56 57 58
 *
 * prandom_u32()
 * -------------
 *
 * For even weaker applications, see the pseudorandom generator
 * prandom_u32(), prandom_max(), and prandom_bytes().  If the random
 * numbers aren't security-critical at all, these are *far* cheaper.
 * Useful for self-tests, random error simulation, randomized backoffs,
 * and any other application where you trust that nobody is trying to
 * maliciously mess with you by guessing the "random" numbers.
 *
L
Linus Torvalds 已提交
59 60 61 62 63 64
 * Exported interfaces ---- input
 * ==============================
 *
 * The current exported interfaces for gathering environmental noise
 * from the devices are:
 *
65
 *	void add_device_randomness(const void *buf, size_t size);
66
 *	void add_input_randomness(unsigned int type, unsigned int code,
L
Linus Torvalds 已提交
67
 *                                unsigned int value);
68
 *	void add_interrupt_randomness(int irq);
69
 *	void add_disk_randomness(struct gendisk *disk);
70
 *	void add_hwgenerator_randomness(const void *buffer, size_t count,
71
 *					size_t entropy);
72
 *	void add_bootloader_randomness(const void *buf, size_t size);
L
Linus Torvalds 已提交
73
 *
74 75 76 77 78 79 80 81
 * add_device_randomness() is for adding data to the random pool that
 * is likely to differ between two devices (or possibly even per boot).
 * This would be things like MAC addresses or serial numbers, or the
 * read-out of the RTC. This does *not* add any actual entropy to the
 * pool, but it initializes the pool to different values for devices
 * that might otherwise be identical and have very little entropy
 * available to them (particularly common in the embedded world).
 *
L
Linus Torvalds 已提交
82 83 84
 * add_input_randomness() uses the input layer interrupt timing, as well as
 * the event type information from the hardware.
 *
85 86 87
 * add_interrupt_randomness() uses the interrupt timing as random
 * inputs to the entropy pool. Using the cycle counters and the irq source
 * as inputs, it feeds the randomness roughly once a second.
88 89 90 91 92 93
 *
 * add_disk_randomness() uses what amounts to the seek time of block
 * layer request events, on a per-disk_devt basis, as input to the
 * entropy pool. Note that high-speed solid state drives with very low
 * seek times do not make for good sources of entropy, as their seek
 * times are usually fairly consistent.
L
Linus Torvalds 已提交
94 95 96 97 98
 *
 * All of these routines try to estimate how many bits of randomness a
 * particular randomness source.  They do this by keeping track of the
 * first and second order deltas of the event timings.
 *
99 100 101 102 103 104 105 106
 * add_hwgenerator_randomness() is for true hardware RNGs, and will credit
 * entropy as specified by the caller. If the entropy pool is full it will
 * block until more entropy is needed.
 *
 * add_bootloader_randomness() is the same as add_hwgenerator_randomness() or
 * add_device_randomness(), depending on whether or not the configuration
 * option CONFIG_RANDOM_TRUST_BOOTLOADER is set.
 *
L
Linus Torvalds 已提交
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
 * Ensuring unpredictability at system startup
 * ============================================
 *
 * When any operating system starts up, it will go through a sequence
 * of actions that are fairly predictable by an adversary, especially
 * if the start-up does not involve interaction with a human operator.
 * This reduces the actual number of bits of unpredictability in the
 * entropy pool below the value in entropy_count.  In order to
 * counteract this effect, it helps to carry information in the
 * entropy pool across shut-downs and start-ups.  To do this, put the
 * following lines an appropriate script which is run during the boot
 * sequence:
 *
 *	echo "Initializing random number generator..."
 *	random_seed=/var/run/random-seed
 *	# Carry a random seed from start-up to start-up
 *	# Load and then save the whole entropy pool
 *	if [ -f $random_seed ]; then
 *		cat $random_seed >/dev/urandom
 *	else
 *		touch $random_seed
 *	fi
 *	chmod 600 $random_seed
 *	dd if=/dev/urandom of=$random_seed count=1 bs=512
 *
 * and the following lines in an appropriate script which is run as
 * the system is shutdown:
 *
 *	# Carry a random seed from shut-down to start-up
 *	# Save the whole entropy pool
 *	echo "Saving random seed..."
 *	random_seed=/var/run/random-seed
 *	touch $random_seed
 *	chmod 600 $random_seed
 *	dd if=/dev/urandom of=$random_seed count=1 bs=512
 *
 * For example, on most modern systems using the System V init
 * scripts, such code fragments would be found in
 * /etc/rc.d/init.d/random.  On older Linux systems, the correct script
 * location might be in /etc/rcb.d/rc.local or /etc/rc.d/rc.0.
 *
 * Effectively, these commands cause the contents of the entropy pool
 * to be saved at shut-down time and reloaded into the entropy pool at
 * start-up.  (The 'dd' in the addition to the bootup script is to
 * make sure that /etc/random-seed is different for every start-up,
 * even if the system crashes without executing rc.0.)  Even with
 * complete knowledge of the start-up activities, predicting the state
 * of the entropy pool requires knowledge of the previous history of
 * the system.
 *
 * Configuring the /dev/random driver under Linux
 * ==============================================
 *
 * The /dev/random driver under Linux uses minor numbers 8 and 9 of
 * the /dev/mem major number (#1).  So if your system does not have
 * /dev/random and /dev/urandom created already, they can be created
 * by using the commands:
 *
165 166
 *	mknod /dev/random c 1 8
 *	mknod /dev/urandom c 1 9
L
Linus Torvalds 已提交
167 168
 */

Y
Yangtao Li 已提交
169 170
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

L
Linus Torvalds 已提交
171 172 173 174 175 176 177 178 179 180 181 182 183
#include <linux/utsname.h>
#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/major.h>
#include <linux/string.h>
#include <linux/fcntl.h>
#include <linux/slab.h>
#include <linux/random.h>
#include <linux/poll.h>
#include <linux/init.h>
#include <linux/fs.h>
#include <linux/genhd.h>
#include <linux/interrupt.h>
184
#include <linux/mm.h>
185
#include <linux/nodemask.h>
L
Linus Torvalds 已提交
186
#include <linux/spinlock.h>
187
#include <linux/kthread.h>
L
Linus Torvalds 已提交
188
#include <linux/percpu.h>
189
#include <linux/ptrace.h>
190
#include <linux/workqueue.h>
191
#include <linux/irq.h>
192
#include <linux/ratelimit.h>
193 194
#include <linux/syscalls.h>
#include <linux/completion.h>
195
#include <linux/uuid.h>
196
#include <crypto/chacha.h>
197
#include <crypto/blake2s.h>
198

L
Linus Torvalds 已提交
199
#include <asm/processor.h>
200
#include <linux/uaccess.h>
L
Linus Torvalds 已提交
201
#include <asm/irq.h>
202
#include <asm/irq_regs.h>
L
Linus Torvalds 已提交
203 204
#include <asm/io.h>

205
enum {
206
	POOL_BITS = BLAKE2S_HASH_SIZE * 8,
207
	POOL_MIN_BITS = POOL_BITS /* No point in settling for less. */
L
Linus Torvalds 已提交
208 209 210 211 212
};

/*
 * Static global variables
 */
213
static DECLARE_WAIT_QUEUE_HEAD(random_write_wait);
214
static struct fasync_struct *fasync;
L
Linus Torvalds 已提交
215

216 217 218
static DEFINE_SPINLOCK(random_ready_list_lock);
static LIST_HEAD(random_ready_list);

219 220 221 222 223 224 225 226 227
/*
 * crng_init =  0 --> Uninitialized
 *		1 --> Initialized
 *		2 --> Initialized from input_pool
 *
 * crng_init is protected by primary_crng->lock, and only increases
 * its value (from 0->1->2).
 */
static int crng_init = 0;
T
Theodore Ts'o 已提交
228
#define crng_ready() (likely(crng_init > 1))
229 230
static int crng_init_cnt = 0;
static void process_random_ready_list(void);
231
static void _get_random_bytes(void *buf, size_t nbytes);
232

233 234 235 236 237 238 239 240 241 242
static struct ratelimit_state unseeded_warning =
	RATELIMIT_STATE_INIT("warn_unseeded_randomness", HZ, 3);
static struct ratelimit_state urandom_warning =
	RATELIMIT_STATE_INIT("warn_urandom_randomness", HZ, 3);

static int ratelimit_disable __read_mostly;

module_param_named(ratelimit_disable, ratelimit_disable, int, 0644);
MODULE_PARM_DESC(ratelimit_disable, "Disable random ratelimit suppression");

L
Linus Torvalds 已提交
243 244 245 246 247 248 249
/**********************************************************************
 *
 * OS independent entropy store.   Here are the functions which handle
 * storing entropy in an entropy pool.
 *
 **********************************************************************/

250
static struct {
251
	struct blake2s_state hash;
252
	spinlock_t lock;
253
	unsigned int entropy_count;
254
} input_pool = {
255 256 257 258
	.hash.h = { BLAKE2S_IV0 ^ (0x01010000 | BLAKE2S_HASH_SIZE),
		    BLAKE2S_IV1, BLAKE2S_IV2, BLAKE2S_IV3, BLAKE2S_IV4,
		    BLAKE2S_IV5, BLAKE2S_IV6, BLAKE2S_IV7 },
	.hash.outlen = BLAKE2S_HASH_SIZE,
259
	.lock = __SPIN_LOCK_UNLOCKED(input_pool.lock),
L
Linus Torvalds 已提交
260 261
};

262
static void extract_entropy(void *buf, size_t nbytes);
263
static bool drain_entropy(void *buf, size_t nbytes);
264

265
static void crng_reseed(void);
266

L
Linus Torvalds 已提交
267
/*
268
 * This function adds bytes into the entropy "pool".  It does not
L
Linus Torvalds 已提交
269
 * update the entropy estimate.  The caller should call
270
 * credit_entropy_bits if this is appropriate.
L
Linus Torvalds 已提交
271
 */
272
static void _mix_pool_bytes(const void *in, size_t nbytes)
L
Linus Torvalds 已提交
273
{
274
	blake2s_update(&input_pool.hash, in, nbytes);
L
Linus Torvalds 已提交
275 276
}

277
static void mix_pool_bytes(const void *in, size_t nbytes)
L
Linus Torvalds 已提交
278
{
279 280
	unsigned long flags;

281 282 283
	spin_lock_irqsave(&input_pool.lock, flags);
	_mix_pool_bytes(in, nbytes);
	spin_unlock_irqrestore(&input_pool.lock, flags);
L
Linus Torvalds 已提交
284 285
}

286
struct fast_pool {
287 288 289 290
	union {
		u32 pool32[4];
		u64 pool64[2];
	};
291 292 293
	unsigned long last;
	u16 reg_idx;
	u8 count;
294 295 296 297 298 299 300
};

/*
 * This is a fast mixing routine used by the interrupt randomness
 * collector.  It's hardcoded for an 128 bit pool and assumes that any
 * locks that might be needed are taken by the caller.
 */
301
static void fast_mix(u32 pool[4])
302
{
303 304
	u32 a = pool[0],	b = pool[1];
	u32 c = pool[2],	d = pool[3];
305 306

	a += b;			c += d;
G
George Spelvin 已提交
307
	b = rol32(b, 6);	d = rol32(d, 27);
308 309 310
	d ^= a;			b ^= c;

	a += b;			c += d;
G
George Spelvin 已提交
311
	b = rol32(b, 16);	d = rol32(d, 14);
312 313 314
	d ^= a;			b ^= c;

	a += b;			c += d;
G
George Spelvin 已提交
315
	b = rol32(b, 6);	d = rol32(d, 27);
316 317 318
	d ^= a;			b ^= c;

	a += b;			c += d;
G
George Spelvin 已提交
319
	b = rol32(b, 16);	d = rol32(d, 14);
320 321
	d ^= a;			b ^= c;

322 323
	pool[0] = a;  pool[1] = b;
	pool[2] = c;  pool[3] = d;
324 325
}

326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341
static void process_random_ready_list(void)
{
	unsigned long flags;
	struct random_ready_callback *rdy, *tmp;

	spin_lock_irqsave(&random_ready_list_lock, flags);
	list_for_each_entry_safe(rdy, tmp, &random_ready_list, list) {
		struct module *owner = rdy->owner;

		list_del_init(&rdy->list);
		rdy->func(rdy);
		module_put(owner);
	}
	spin_unlock_irqrestore(&random_ready_list_lock, flags);
}

342
static void credit_entropy_bits(size_t nbits)
L
Linus Torvalds 已提交
343
{
344
	unsigned int entropy_count, orig, add;
345

346
	if (!nbits)
347 348
		return;

349
	add = min_t(size_t, nbits, POOL_BITS);
350

351 352
	do {
		orig = READ_ONCE(input_pool.entropy_count);
353
		entropy_count = min_t(unsigned int, POOL_BITS, orig + add);
354
	} while (cmpxchg(&input_pool.entropy_count, orig, entropy_count) != orig);
L
Linus Torvalds 已提交
355

356
	if (crng_init < 2 && entropy_count >= POOL_MIN_BITS)
357
		crng_reseed();
L
Linus Torvalds 已提交
358 359
}

360 361 362 363 364 365
/*********************************************************************
 *
 * CRNG using CHACHA20
 *
 *********************************************************************/

366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389
enum {
	CRNG_RESEED_INTERVAL = 300 * HZ,
	CRNG_INIT_CNT_THRESH = 2 * CHACHA_KEY_SIZE
};

static struct {
	u8 key[CHACHA_KEY_SIZE] __aligned(__alignof__(long));
	unsigned long birth;
	unsigned long generation;
	spinlock_t lock;
} base_crng = {
	.lock = __SPIN_LOCK_UNLOCKED(base_crng.lock)
};

struct crng {
	u8 key[CHACHA_KEY_SIZE];
	unsigned long generation;
	local_lock_t lock;
};

static DEFINE_PER_CPU(struct crng, crngs) = {
	.generation = ULONG_MAX,
	.lock = INIT_LOCAL_LOCK(crngs.lock),
};
390 391 392

static DECLARE_WAIT_QUEUE_HEAD(crng_init_wait);

393 394
/*
 * crng_fast_load() can be called by code in the interrupt service
395 396
 * path.  So we can't afford to dilly-dally. Returns the number of
 * bytes processed from cp.
397
 */
398
static size_t crng_fast_load(const void *cp, size_t len)
399 400
{
	unsigned long flags;
401
	const u8 *src = (const u8 *)cp;
402
	size_t ret = 0;
403

404
	if (!spin_trylock_irqsave(&base_crng.lock, flags))
405
		return 0;
T
Theodore Ts'o 已提交
406
	if (crng_init != 0) {
407
		spin_unlock_irqrestore(&base_crng.lock, flags);
408 409 410
		return 0;
	}
	while (len > 0 && crng_init_cnt < CRNG_INIT_CNT_THRESH) {
411 412
		base_crng.key[crng_init_cnt % sizeof(base_crng.key)] ^= *src;
		src++; crng_init_cnt++; len--; ret++;
413 414
	}
	if (crng_init_cnt >= CRNG_INIT_CNT_THRESH) {
415
		++base_crng.generation;
416 417
		crng_init = 1;
	}
418
	spin_unlock_irqrestore(&base_crng.lock, flags);
419 420
	if (crng_init == 1)
		pr_notice("fast init done\n");
421
	return ret;
422 423
}

424 425 426 427 428 429 430
/*
 * crng_slow_load() is called by add_device_randomness, which has two
 * attributes.  (1) We can't trust the buffer passed to it is
 * guaranteed to be unpredictable (so it might not have any entropy at
 * all), and (2) it doesn't have the performance constraints of
 * crng_fast_load().
 *
431 432 433 434
 * So, we simply hash the contents in with the current key. Finally,
 * we do *not* advance crng_init_cnt since buffer we may get may be
 * something like a fixed DMI table (for example), which might very
 * well be unique to the machine, but is otherwise unvarying.
435
 */
436
static void crng_slow_load(const void *cp, size_t len)
437
{
438
	unsigned long flags;
439 440 441
	struct blake2s_state hash;

	blake2s_init(&hash, sizeof(base_crng.key));
442

443
	if (!spin_trylock_irqsave(&base_crng.lock, flags))
444
		return;
445
	if (crng_init != 0) {
446
		spin_unlock_irqrestore(&base_crng.lock, flags);
447
		return;
448
	}
449 450 451 452 453

	blake2s_update(&hash, base_crng.key, sizeof(base_crng.key));
	blake2s_update(&hash, cp, len);
	blake2s_final(&hash, base_crng.key);

454
	spin_unlock_irqrestore(&base_crng.lock, flags);
455 456
}

457
static void crng_reseed(void)
458
{
459
	unsigned long flags;
460 461
	unsigned long next_gen;
	u8 key[CHACHA_KEY_SIZE];
462
	bool finalize_init = false;
463

464 465 466
	/* Only reseed if we can, to prevent brute forcing a small amount of new bits. */
	if (!drain_entropy(key, sizeof(key)))
		return;
467

468 469 470 471 472 473 474 475 476 477 478 479 480
	/*
	 * We copy the new key into the base_crng, overwriting the old one,
	 * and update the generation counter. We avoid hitting ULONG_MAX,
	 * because the per-cpu crngs are initialized to ULONG_MAX, so this
	 * forces new CPUs that come online to always initialize.
	 */
	spin_lock_irqsave(&base_crng.lock, flags);
	memcpy(base_crng.key, key, sizeof(base_crng.key));
	next_gen = base_crng.generation + 1;
	if (next_gen == ULONG_MAX)
		++next_gen;
	WRITE_ONCE(base_crng.generation, next_gen);
	WRITE_ONCE(base_crng.birth, jiffies);
481 482
	if (crng_init < 2) {
		crng_init = 2;
483 484 485 486 487
		finalize_init = true;
	}
	spin_unlock_irqrestore(&base_crng.lock, flags);
	memzero_explicit(key, sizeof(key));
	if (finalize_init) {
488 489 490 491 492 493 494 495 496 497 498 499 500 501 502
		process_random_ready_list();
		wake_up_interruptible(&crng_init_wait);
		kill_fasync(&fasync, SIGIO, POLL_IN);
		pr_notice("crng init done\n");
		if (unseeded_warning.missed) {
			pr_notice("%d get_random_xx warning(s) missed due to ratelimiting\n",
				  unseeded_warning.missed);
			unseeded_warning.missed = 0;
		}
		if (urandom_warning.missed) {
			pr_notice("%d urandom warning(s) missed due to ratelimiting\n",
				  urandom_warning.missed);
			urandom_warning.missed = 0;
		}
	}
503 504
}

505 506 507 508 509 510 511 512 513 514 515 516
/*
 * The general form here is based on a "fast key erasure RNG" from
 * <https://blog.cr.yp.to/20170723-random.html>. It generates a ChaCha
 * block using the provided key, and then immediately overwites that
 * key with half the block. It returns the resultant ChaCha state to the
 * user, along with the second half of the block containing 32 bytes of
 * random data that may be used; random_data_len may not be greater than
 * 32.
 */
static void crng_fast_key_erasure(u8 key[CHACHA_KEY_SIZE],
				  u32 chacha_state[CHACHA_STATE_WORDS],
				  u8 *random_data, size_t random_data_len)
517
{
518
	u8 first_block[CHACHA_BLOCK_SIZE];
519

520 521 522 523 524 525 526 527 528 529
	BUG_ON(random_data_len > 32);

	chacha_init_consts(chacha_state);
	memcpy(&chacha_state[4], key, CHACHA_KEY_SIZE);
	memset(&chacha_state[12], 0, sizeof(u32) * 4);
	chacha20_block(chacha_state, first_block);

	memcpy(key, first_block, CHACHA_KEY_SIZE);
	memcpy(random_data, first_block + CHACHA_KEY_SIZE, random_data_len);
	memzero_explicit(first_block, sizeof(first_block));
530 531
}

532
/*
533 534 535
 * This function returns a ChaCha state that you may use for generating
 * random data. It also returns up to 32 bytes on its own of random data
 * that may be used; random_data_len may not be greater than 32.
536
 */
537 538
static void crng_make_state(u32 chacha_state[CHACHA_STATE_WORDS],
			    u8 *random_data, size_t random_data_len)
539
{
540
	unsigned long flags;
541
	struct crng *crng;
542

543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561
	BUG_ON(random_data_len > 32);

	/*
	 * For the fast path, we check whether we're ready, unlocked first, and
	 * then re-check once locked later. In the case where we're really not
	 * ready, we do fast key erasure with the base_crng directly, because
	 * this is what crng_{fast,slow}_load mutate during early init.
	 */
	if (unlikely(!crng_ready())) {
		bool ready;

		spin_lock_irqsave(&base_crng.lock, flags);
		ready = crng_ready();
		if (!ready)
			crng_fast_key_erasure(base_crng.key, chacha_state,
					      random_data, random_data_len);
		spin_unlock_irqrestore(&base_crng.lock, flags);
		if (!ready)
			return;
562
	}
563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596

	/*
	 * If the base_crng is more than 5 minutes old, we reseed, which
	 * in turn bumps the generation counter that we check below.
	 */
	if (unlikely(time_after(jiffies, READ_ONCE(base_crng.birth) + CRNG_RESEED_INTERVAL)))
		crng_reseed();

	local_lock_irqsave(&crngs.lock, flags);
	crng = raw_cpu_ptr(&crngs);

	/*
	 * If our per-cpu crng is older than the base_crng, then it means
	 * somebody reseeded the base_crng. In that case, we do fast key
	 * erasure on the base_crng, and use its output as the new key
	 * for our per-cpu crng. This brings us up to date with base_crng.
	 */
	if (unlikely(crng->generation != READ_ONCE(base_crng.generation))) {
		spin_lock(&base_crng.lock);
		crng_fast_key_erasure(base_crng.key, chacha_state,
				      crng->key, sizeof(crng->key));
		crng->generation = base_crng.generation;
		spin_unlock(&base_crng.lock);
	}

	/*
	 * Finally, when we've made it this far, our per-cpu crng has an up
	 * to date key, and we can do fast key erasure with it to produce
	 * some random data and a ChaCha state for the caller. All other
	 * branches of this function are "unlikely", so most of the time we
	 * should wind up here immediately.
	 */
	crng_fast_key_erasure(crng->key, chacha_state, random_data, random_data_len);
	local_unlock_irqrestore(&crngs.lock, flags);
597 598
}

599
static ssize_t get_random_bytes_user(void __user *buf, size_t nbytes)
600
{
601
	bool large_request = nbytes > 256;
602 603
	ssize_t ret = 0;
	size_t len;
604 605 606 607 608 609
	u32 chacha_state[CHACHA_STATE_WORDS];
	u8 output[CHACHA_BLOCK_SIZE];

	if (!nbytes)
		return 0;

610
	len = min_t(size_t, 32, nbytes);
611 612 613 614 615 616 617
	crng_make_state(chacha_state, output, len);

	if (copy_to_user(buf, output, len))
		return -EFAULT;
	nbytes -= len;
	buf += len;
	ret += len;
618 619 620

	while (nbytes) {
		if (large_request && need_resched()) {
621
			if (signal_pending(current))
622 623 624 625
				break;
			schedule();
		}

626 627 628 629
		chacha20_block(chacha_state, output);
		if (unlikely(chacha_state[12] == 0))
			++chacha_state[13];

630
		len = min_t(size_t, nbytes, CHACHA_BLOCK_SIZE);
631
		if (copy_to_user(buf, output, len)) {
632 633 634 635
			ret = -EFAULT;
			break;
		}

636 637 638
		nbytes -= len;
		buf += len;
		ret += len;
639 640
	}

641 642
	memzero_explicit(chacha_state, sizeof(chacha_state));
	memzero_explicit(output, sizeof(output));
643 644 645
	return ret;
}

L
Linus Torvalds 已提交
646 647 648 649 650 651 652 653 654
/*********************************************************************
 *
 * Entropy input management
 *
 *********************************************************************/

/* There is one of these per entropy source */
struct timer_rand_state {
	cycles_t last_time;
655
	long last_delta, last_delta2;
L
Linus Torvalds 已提交
656 657
};

658 659
#define INIT_TIMER_RAND_STATE { INITIAL_JIFFIES, };

660
/*
661 662
 * Add device- or boot-specific data to the input pool to help
 * initialize it.
663
 *
664 665 666
 * None of this adds any entropy; it is meant to avoid the problem of
 * the entropy pool having similar initial state across largely
 * identical devices.
667
 */
668
void add_device_randomness(const void *buf, size_t size)
669
{
670
	unsigned long time = random_get_entropy() ^ jiffies;
671
	unsigned long flags;
672

673 674
	if (!crng_ready() && size)
		crng_slow_load(buf, size);
675

676
	spin_lock_irqsave(&input_pool.lock, flags);
677 678
	_mix_pool_bytes(buf, size);
	_mix_pool_bytes(&time, sizeof(time));
679
	spin_unlock_irqrestore(&input_pool.lock, flags);
680 681 682
}
EXPORT_SYMBOL(add_device_randomness);

683
static struct timer_rand_state input_timer_state = INIT_TIMER_RAND_STATE;
684

L
Linus Torvalds 已提交
685 686 687 688 689 690 691 692 693 694
/*
 * This function adds entropy to the entropy "pool" by using timing
 * delays.  It uses the timer_rand_state structure to make an estimate
 * of how many bits of entropy this call has added to the pool.
 *
 * The number "num" is also added to the pool - it should somehow describe
 * the type of event which just happened.  This is currently 0-255 for
 * keyboard scan codes, and 256 upwards for interrupts.
 *
 */
695
static void add_timer_randomness(struct timer_rand_state *state, unsigned int num)
L
Linus Torvalds 已提交
696 697 698
{
	struct {
		long jiffies;
699 700
		unsigned int cycles;
		unsigned int num;
L
Linus Torvalds 已提交
701 702 703 704
	} sample;
	long delta, delta2, delta3;

	sample.jiffies = jiffies;
705
	sample.cycles = random_get_entropy();
L
Linus Torvalds 已提交
706
	sample.num = num;
707
	mix_pool_bytes(&sample, sizeof(sample));
L
Linus Torvalds 已提交
708 709 710 711 712 713

	/*
	 * Calculate number of bits of randomness we probably added.
	 * We take into account the first, second and third-order deltas
	 * in order to make our estimate.
	 */
714 715
	delta = sample.jiffies - READ_ONCE(state->last_time);
	WRITE_ONCE(state->last_time, sample.jiffies);
716

717 718
	delta2 = delta - READ_ONCE(state->last_delta);
	WRITE_ONCE(state->last_delta, delta);
719

720 721
	delta3 = delta2 - READ_ONCE(state->last_delta2);
	WRITE_ONCE(state->last_delta2, delta2);
722 723 724 725 726 727 728 729 730 731 732

	if (delta < 0)
		delta = -delta;
	if (delta2 < 0)
		delta2 = -delta2;
	if (delta3 < 0)
		delta3 = -delta3;
	if (delta > delta2)
		delta = delta2;
	if (delta > delta3)
		delta = delta3;
L
Linus Torvalds 已提交
733

734 735 736
	/*
	 * delta is now minimum absolute delta.
	 * Round down by 1 bit on general principles,
737
	 * and limit entropy estimate to 12 bits.
738
	 */
739
	credit_entropy_bits(min_t(unsigned int, fls(delta >> 1), 11));
L
Linus Torvalds 已提交
740 741
}

742
void add_input_randomness(unsigned int type, unsigned int code,
743
			  unsigned int value)
L
Linus Torvalds 已提交
744 745 746 747 748 749 750 751 752 753 754
{
	static unsigned char last_value;

	/* ignore autorepeat and the like */
	if (value == last_value)
		return;

	last_value = value;
	add_timer_randomness(&input_timer_state,
			     (type << 4) ^ code ^ (code >> 4) ^ value);
}
755
EXPORT_SYMBOL_GPL(add_input_randomness);
L
Linus Torvalds 已提交
756

757 758
static DEFINE_PER_CPU(struct fast_pool, irq_randomness);

759
static u32 get_reg(struct fast_pool *f, struct pt_regs *regs)
760
{
761
	u32 *ptr = (u32 *)regs;
762
	unsigned int idx;
763 764 765

	if (regs == NULL)
		return 0;
766
	idx = READ_ONCE(f->reg_idx);
767
	if (idx >= sizeof(struct pt_regs) / sizeof(u32))
768 769 770
		idx = 0;
	ptr += idx++;
	WRITE_ONCE(f->reg_idx, idx);
771
	return *ptr;
772 773
}

774
void add_interrupt_randomness(int irq)
L
Linus Torvalds 已提交
775
{
776 777 778 779
	struct fast_pool *fast_pool = this_cpu_ptr(&irq_randomness);
	struct pt_regs *regs = get_irq_regs();
	unsigned long now = jiffies;
	cycles_t cycles = random_get_entropy();
780

781 782
	if (cycles == 0)
		cycles = get_reg(fast_pool, regs);
783

784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799
	if (sizeof(cycles) == 8)
		fast_pool->pool64[0] ^= cycles ^ rol64(now, 32) ^ irq;
	else {
		fast_pool->pool32[0] ^= cycles ^ irq;
		fast_pool->pool32[1] ^= now;
	}

	if (sizeof(unsigned long) == 8)
		fast_pool->pool64[1] ^= regs ? instruction_pointer(regs) : _RET_IP_;
	else {
		fast_pool->pool32[2] ^= regs ? instruction_pointer(regs) : _RET_IP_;
		fast_pool->pool32[3] ^= get_reg(fast_pool, regs);
	}

	fast_mix(fast_pool->pool32);
	++fast_pool->count;
800

T
Theodore Ts'o 已提交
801
	if (unlikely(crng_init == 0)) {
802
		if (fast_pool->count >= 64 &&
803
		    crng_fast_load(fast_pool->pool32, sizeof(fast_pool->pool32)) > 0) {
804 805
			fast_pool->count = 0;
			fast_pool->last = now;
806
			if (spin_trylock(&input_pool.lock)) {
807
				_mix_pool_bytes(&fast_pool->pool32, sizeof(fast_pool->pool32));
808 809
				spin_unlock(&input_pool.lock);
			}
810 811 812 813
		}
		return;
	}

814
	if ((fast_pool->count < 64) && !time_after(now, fast_pool->last + HZ))
L
Linus Torvalds 已提交
815 816
		return;

817
	if (!spin_trylock(&input_pool.lock))
818
		return;
819

820
	fast_pool->last = now;
821
	_mix_pool_bytes(&fast_pool->pool32, sizeof(fast_pool->pool32));
822
	spin_unlock(&input_pool.lock);
823

824
	fast_pool->count = 0;
825

826
	/* award one bit for the contents of the fast pool */
827
	credit_entropy_bits(1);
L
Linus Torvalds 已提交
828
}
829
EXPORT_SYMBOL_GPL(add_interrupt_randomness);
L
Linus Torvalds 已提交
830

831
#ifdef CONFIG_BLOCK
L
Linus Torvalds 已提交
832 833 834 835 836
void add_disk_randomness(struct gendisk *disk)
{
	if (!disk || !disk->random)
		return;
	/* first major is 1, so we get >= 0x200 here */
837
	add_timer_randomness(disk->random, 0x100 + disk_devt(disk));
L
Linus Torvalds 已提交
838
}
839
EXPORT_SYMBOL_GPL(add_disk_randomness);
840
#endif
L
Linus Torvalds 已提交
841 842 843 844 845 846 847

/*********************************************************************
 *
 * Entropy extraction routines
 *
 *********************************************************************/

G
Greg Price 已提交
848
/*
849 850
 * This is an HKDF-like construction for using the hashed collected entropy
 * as a PRF key, that's then expanded block-by-block.
G
Greg Price 已提交
851
 */
852
static void extract_entropy(void *buf, size_t nbytes)
L
Linus Torvalds 已提交
853
{
854
	unsigned long flags;
855 856
	u8 seed[BLAKE2S_HASH_SIZE], next_key[BLAKE2S_HASH_SIZE];
	struct {
857
		unsigned long rdseed[32 / sizeof(long)];
858 859 860 861
		size_t counter;
	} block;
	size_t i;

862 863 864 865
	for (i = 0; i < ARRAY_SIZE(block.rdseed); ++i) {
		if (!arch_get_random_seed_long(&block.rdseed[i]) &&
		    !arch_get_random_long(&block.rdseed[i]))
			block.rdseed[i] = random_get_entropy();
866 867
	}

868
	spin_lock_irqsave(&input_pool.lock, flags);
869

870 871
	/* seed = HASHPRF(last_key, entropy_input) */
	blake2s_final(&input_pool.hash, seed);
L
Linus Torvalds 已提交
872

873
	/* next_key = HASHPRF(seed, RDSEED || 0) */
874 875 876
	block.counter = 0;
	blake2s(next_key, (u8 *)&block, seed, sizeof(next_key), sizeof(block), sizeof(seed));
	blake2s_init_key(&input_pool.hash, BLAKE2S_HASH_SIZE, next_key, sizeof(next_key));
L
Linus Torvalds 已提交
877

878 879
	spin_unlock_irqrestore(&input_pool.lock, flags);
	memzero_explicit(next_key, sizeof(next_key));
880 881

	while (nbytes) {
882
		i = min_t(size_t, nbytes, BLAKE2S_HASH_SIZE);
883
		/* output = HASHPRF(seed, RDSEED || ++counter) */
884 885
		++block.counter;
		blake2s(buf, (u8 *)&block, seed, i, sizeof(block), sizeof(seed));
886 887 888 889
		nbytes -= i;
		buf += i;
	}

890 891
	memzero_explicit(seed, sizeof(seed));
	memzero_explicit(&block, sizeof(block));
892 893
}

894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912
/*
 * First we make sure we have POOL_MIN_BITS of entropy in the pool, and then we
 * set the entropy count to zero (but don't actually touch any data). Only then
 * can we extract a new key with extract_entropy().
 */
static bool drain_entropy(void *buf, size_t nbytes)
{
	unsigned int entropy_count;
	do {
		entropy_count = READ_ONCE(input_pool.entropy_count);
		if (entropy_count < POOL_MIN_BITS)
			return false;
	} while (cmpxchg(&input_pool.entropy_count, entropy_count, 0) != entropy_count);
	extract_entropy(buf, nbytes);
	wake_up_interruptible(&random_write_wait);
	kill_fasync(&fasync, SIGIO, POLL_OUT);
	return true;
}

913
#define warn_unseeded_randomness(previous) \
914
	_warn_unseeded_randomness(__func__, (void *)_RET_IP_, (previous))
915

916
static void _warn_unseeded_randomness(const char *func_name, void *caller, void **previous)
917 918 919 920 921 922 923
{
#ifdef CONFIG_WARN_ALL_UNSEEDED_RANDOM
	const bool print_once = false;
#else
	static bool print_once __read_mostly;
#endif

924
	if (print_once || crng_ready() ||
925 926 927 928 929 930
	    (previous && (caller == READ_ONCE(*previous))))
		return;
	WRITE_ONCE(*previous, caller);
#ifndef CONFIG_WARN_ALL_UNSEEDED_RANDOM
	print_once = true;
#endif
931
	if (__ratelimit(&unseeded_warning))
932 933
		printk_deferred(KERN_NOTICE "random: %s called from %pS with crng_init=%d\n",
				func_name, caller, crng_init);
934 935
}

L
Linus Torvalds 已提交
936 937
/*
 * This function is the exported kernel interface.  It returns some
938
 * number of good random numbers, suitable for key generation, seeding
939 940
 * TCP sequence numbers, etc.  It does not rely on the hardware random
 * number generator.  For random bytes direct from the hardware RNG
941 942 943 944
 * (when available), use get_random_bytes_arch(). In order to ensure
 * that the randomness provided by this function is okay, the function
 * wait_for_random_bytes() should be called and return 0 at least once
 * at any point prior.
L
Linus Torvalds 已提交
945
 */
946
static void _get_random_bytes(void *buf, size_t nbytes)
947
{
948 949
	u32 chacha_state[CHACHA_STATE_WORDS];
	u8 tmp[CHACHA_BLOCK_SIZE];
950
	size_t len;
951

952 953 954
	if (!nbytes)
		return;

955
	len = min_t(size_t, 32, nbytes);
956 957 958 959 960 961 962 963 964 965 966 967 968 969 970
	crng_make_state(chacha_state, buf, len);
	nbytes -= len;
	buf += len;

	while (nbytes) {
		if (nbytes < CHACHA_BLOCK_SIZE) {
			chacha20_block(chacha_state, tmp);
			memcpy(buf, tmp, nbytes);
			memzero_explicit(tmp, sizeof(tmp));
			break;
		}

		chacha20_block(chacha_state, buf);
		if (unlikely(chacha_state[12] == 0))
			++chacha_state[13];
971
		nbytes -= CHACHA_BLOCK_SIZE;
972
		buf += CHACHA_BLOCK_SIZE;
973 974
	}

975
	memzero_explicit(chacha_state, sizeof(chacha_state));
976
}
977

978
void get_random_bytes(void *buf, size_t nbytes)
979 980 981 982 983 984
{
	static void *previous;

	warn_unseeded_randomness(&previous);
	_get_random_bytes(buf, nbytes);
}
985 986
EXPORT_SYMBOL(get_random_bytes);

987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001
/*
 * Each time the timer fires, we expect that we got an unpredictable
 * jump in the cycle counter. Even if the timer is running on another
 * CPU, the timer activity will be touching the stack of the CPU that is
 * generating entropy..
 *
 * Note that we don't re-arm the timer in the timer itself - we are
 * happy to be scheduled away, since that just makes the load more
 * complex, but we do not want the timer to keep ticking unless the
 * entropy loop is running.
 *
 * So the re-arming always happens in the entropy loop itself.
 */
static void entropy_timer(struct timer_list *t)
{
1002
	credit_entropy_bits(1);
1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024
}

/*
 * If we have an actual cycle counter, see if we can
 * generate enough entropy with timing noise
 */
static void try_to_generate_entropy(void)
{
	struct {
		unsigned long now;
		struct timer_list timer;
	} stack;

	stack.now = random_get_entropy();

	/* Slow counter - or none. Don't even bother */
	if (stack.now == random_get_entropy())
		return;

	timer_setup_on_stack(&stack.timer, entropy_timer, 0);
	while (!crng_ready()) {
		if (!timer_pending(&stack.timer))
1025
			mod_timer(&stack.timer, jiffies + 1);
1026
		mix_pool_bytes(&stack.now, sizeof(stack.now));
1027 1028 1029 1030 1031 1032
		schedule();
		stack.now = random_get_entropy();
	}

	del_timer_sync(&stack.timer);
	destroy_timer_on_stack(&stack.timer);
1033
	mix_pool_bytes(&stack.now, sizeof(stack.now));
1034 1035
}

1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049
/*
 * Wait for the urandom pool to be seeded and thus guaranteed to supply
 * cryptographically secure random numbers. This applies to: the /dev/urandom
 * device, the get_random_bytes function, and the get_random_{u32,u64,int,long}
 * family of functions. Using any of these functions without first calling
 * this function forfeits the guarantee of security.
 *
 * Returns: 0 if the urandom pool has been seeded.
 *          -ERESTARTSYS if the function was interrupted by a signal.
 */
int wait_for_random_bytes(void)
{
	if (likely(crng_ready()))
		return 0;
1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060

	do {
		int ret;
		ret = wait_event_interruptible_timeout(crng_init_wait, crng_ready(), HZ);
		if (ret)
			return ret > 0 ? 0 : ret;

		try_to_generate_entropy();
	} while (!crng_ready());

	return 0;
1061 1062 1063
}
EXPORT_SYMBOL(wait_for_random_bytes);

1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078
/*
 * Returns whether or not the urandom pool has been seeded and thus guaranteed
 * to supply cryptographically secure random numbers. This applies to: the
 * /dev/urandom device, the get_random_bytes function, and the get_random_{u32,
 * ,u64,int,long} family of functions.
 *
 * Returns: true if the urandom pool has been seeded.
 *          false if the urandom pool has not been seeded.
 */
bool rng_is_initialized(void)
{
	return crng_ready();
}
EXPORT_SYMBOL(rng_is_initialized);

1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092
/*
 * Add a callback function that will be invoked when the nonblocking
 * pool is initialised.
 *
 * returns: 0 if callback is successfully added
 *	    -EALREADY if pool is already initialised (callback not called)
 *	    -ENOENT if module for callback is not alive
 */
int add_random_ready_callback(struct random_ready_callback *rdy)
{
	struct module *owner;
	unsigned long flags;
	int err = -EALREADY;

1093
	if (crng_ready())
1094 1095 1096 1097 1098 1099 1100
		return err;

	owner = rdy->owner;
	if (!try_module_get(owner))
		return -ENOENT;

	spin_lock_irqsave(&random_ready_list_lock, flags);
1101
	if (crng_ready())
1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136
		goto out;

	owner = NULL;

	list_add(&rdy->list, &random_ready_list);
	err = 0;

out:
	spin_unlock_irqrestore(&random_ready_list_lock, flags);

	module_put(owner);

	return err;
}
EXPORT_SYMBOL(add_random_ready_callback);

/*
 * Delete a previously registered readiness callback function.
 */
void del_random_ready_callback(struct random_ready_callback *rdy)
{
	unsigned long flags;
	struct module *owner = NULL;

	spin_lock_irqsave(&random_ready_list_lock, flags);
	if (!list_empty(&rdy->list)) {
		list_del_init(&rdy->list);
		owner = rdy->owner;
	}
	spin_unlock_irqrestore(&random_ready_list_lock, flags);

	module_put(owner);
}
EXPORT_SYMBOL(del_random_ready_callback);

1137 1138
/*
 * This function will use the architecture-specific hardware random
1139 1140 1141
 * number generator if it is available. It is not recommended for
 * use. Use get_random_bytes() instead. It returns the number of
 * bytes filled in.
1142
 */
1143
size_t __must_check get_random_bytes_arch(void *buf, size_t nbytes)
L
Linus Torvalds 已提交
1144
{
1145
	size_t left = nbytes;
1146
	u8 *p = buf;
1147

1148
	while (left) {
1149
		unsigned long v;
1150
		size_t chunk = min_t(size_t, left, sizeof(unsigned long));
1151

1152 1153
		if (!arch_get_random_long(&v))
			break;
1154

L
Luck, Tony 已提交
1155
		memcpy(p, &v, chunk);
1156
		p += chunk;
1157
		left -= chunk;
1158 1159
	}

1160
	return nbytes - left;
L
Linus Torvalds 已提交
1161
}
1162 1163
EXPORT_SYMBOL(get_random_bytes_arch);

1164 1165 1166 1167 1168 1169 1170
static bool trust_cpu __ro_after_init = IS_ENABLED(CONFIG_RANDOM_TRUST_CPU);
static int __init parse_trust_cpu(char *arg)
{
	return kstrtobool(arg, &trust_cpu);
}
early_param("random.trust_cpu", parse_trust_cpu);

L
Linus Torvalds 已提交
1171
/*
1172 1173 1174 1175 1176 1177 1178 1179
 * Note that setup_arch() may call add_device_randomness()
 * long before we get here. This allows seeding of the pools
 * with some platform dependent data very early in the boot
 * process. But it limits our options here. We must use
 * statically allocated structures that already have all
 * initializations complete at compile time. We should also
 * take care not to overwrite the precious per platform data
 * we were given.
L
Linus Torvalds 已提交
1180
 */
1181
int __init rand_initialize(void)
L
Linus Torvalds 已提交
1182
{
1183
	size_t i;
1184
	ktime_t now = ktime_get_real();
1185
	bool arch_init = true;
1186
	unsigned long rv;
L
Linus Torvalds 已提交
1187

1188
	for (i = 0; i < BLAKE2S_BLOCK_SIZE; i += sizeof(rv)) {
1189 1190 1191 1192 1193
		if (!arch_get_random_seed_long_early(&rv) &&
		    !arch_get_random_long_early(&rv)) {
			rv = random_get_entropy();
			arch_init = false;
		}
1194
		mix_pool_bytes(&rv, sizeof(rv));
1195
	}
1196 1197 1198
	mix_pool_bytes(&now, sizeof(now));
	mix_pool_bytes(utsname(), sizeof(*(utsname())));

1199
	extract_entropy(base_crng.key, sizeof(base_crng.key));
1200 1201
	++base_crng.generation;

1202 1203 1204 1205 1206
	if (arch_init && trust_cpu && crng_init < 2) {
		crng_init = 2;
		pr_notice("crng init done (trusting CPU's manufacturer)\n");
	}

1207 1208 1209 1210
	if (ratelimit_disable) {
		urandom_warning.interval = 0;
		unseeded_warning.interval = 0;
	}
L
Linus Torvalds 已提交
1211 1212 1213
	return 0;
}

1214
#ifdef CONFIG_BLOCK
L
Linus Torvalds 已提交
1215 1216 1217 1218 1219
void rand_initialize_disk(struct gendisk *disk)
{
	struct timer_rand_state *state;

	/*
1220
	 * If kzalloc returns null, we just won't use that entropy
L
Linus Torvalds 已提交
1221 1222
	 * source.
	 */
1223
	state = kzalloc(sizeof(struct timer_rand_state), GFP_KERNEL);
1224 1225
	if (state) {
		state->last_time = INITIAL_JIFFIES;
L
Linus Torvalds 已提交
1226
		disk->random = state;
1227
	}
L
Linus Torvalds 已提交
1228
}
1229
#endif
L
Linus Torvalds 已提交
1230

1231 1232
static ssize_t urandom_read(struct file *file, char __user *buf, size_t nbytes,
			    loff_t *ppos)
L
Linus Torvalds 已提交
1233
{
1234
	static int maxwarn = 10;
1235

1236
	if (!crng_ready() && maxwarn > 0) {
1237
		maxwarn--;
1238
		if (__ratelimit(&urandom_warning))
Y
Yangtao Li 已提交
1239 1240
			pr_notice("%s: uninitialized urandom read (%zd bytes read)\n",
				  current->comm, nbytes);
1241
	}
1242

1243
	return get_random_bytes_user(buf, nbytes);
L
Linus Torvalds 已提交
1244 1245
}

1246 1247
static ssize_t random_read(struct file *file, char __user *buf, size_t nbytes,
			   loff_t *ppos)
1248 1249 1250 1251 1252 1253
{
	int ret;

	ret = wait_for_random_bytes();
	if (ret != 0)
		return ret;
1254
	return get_random_bytes_user(buf, nbytes);
1255 1256
}

1257
static __poll_t random_poll(struct file *file, poll_table *wait)
L
Linus Torvalds 已提交
1258
{
1259
	__poll_t mask;
L
Linus Torvalds 已提交
1260

1261
	poll_wait(file, &crng_init_wait, wait);
1262 1263
	poll_wait(file, &random_write_wait, wait);
	mask = 0;
1264
	if (crng_ready())
1265
		mask |= EPOLLIN | EPOLLRDNORM;
1266
	if (input_pool.entropy_count < POOL_MIN_BITS)
1267
		mask |= EPOLLOUT | EPOLLWRNORM;
L
Linus Torvalds 已提交
1268 1269 1270
	return mask;
}

1271
static int write_pool(const char __user *ubuf, size_t count)
L
Linus Torvalds 已提交
1272
{
1273
	size_t len;
1274
	int ret = 0;
1275
	u8 block[BLAKE2S_BLOCK_SIZE];
L
Linus Torvalds 已提交
1276

1277 1278
	while (count) {
		len = min(count, sizeof(block));
1279 1280 1281 1282
		if (copy_from_user(block, ubuf, len)) {
			ret = -EFAULT;
			goto out;
		}
1283 1284 1285
		count -= len;
		ubuf += len;
		mix_pool_bytes(block, len);
1286
		cond_resched();
L
Linus Torvalds 已提交
1287
	}
1288

1289 1290 1291
out:
	memzero_explicit(block, sizeof(block));
	return ret;
1292 1293
}

1294 1295
static ssize_t random_write(struct file *file, const char __user *buffer,
			    size_t count, loff_t *ppos)
1296
{
1297
	int ret;
1298

1299
	ret = write_pool(buffer, count);
1300 1301 1302 1303
	if (ret)
		return ret;

	return (ssize_t)count;
L
Linus Torvalds 已提交
1304 1305
}

M
Matt Mackall 已提交
1306
static long random_ioctl(struct file *f, unsigned int cmd, unsigned long arg)
L
Linus Torvalds 已提交
1307 1308 1309 1310 1311 1312 1313
{
	int size, ent_count;
	int __user *p = (int __user *)arg;
	int retval;

	switch (cmd) {
	case RNDGETENTCNT:
M
Matt Mackall 已提交
1314
		/* inherently racy, no point locking */
1315
		if (put_user(input_pool.entropy_count, p))
L
Linus Torvalds 已提交
1316 1317 1318 1319 1320 1321 1322
			return -EFAULT;
		return 0;
	case RNDADDTOENTCNT:
		if (!capable(CAP_SYS_ADMIN))
			return -EPERM;
		if (get_user(ent_count, p))
			return -EFAULT;
1323 1324 1325 1326
		if (ent_count < 0)
			return -EINVAL;
		credit_entropy_bits(ent_count);
		return 0;
L
Linus Torvalds 已提交
1327 1328 1329 1330 1331 1332 1333 1334 1335
	case RNDADDENTROPY:
		if (!capable(CAP_SYS_ADMIN))
			return -EPERM;
		if (get_user(ent_count, p++))
			return -EFAULT;
		if (ent_count < 0)
			return -EINVAL;
		if (get_user(size, p++))
			return -EFAULT;
1336
		retval = write_pool((const char __user *)p, size);
L
Linus Torvalds 已提交
1337 1338
		if (retval < 0)
			return retval;
1339 1340
		credit_entropy_bits(ent_count);
		return 0;
L
Linus Torvalds 已提交
1341 1342
	case RNDZAPENTCNT:
	case RNDCLEARPOOL:
1343 1344 1345 1346
		/*
		 * Clear the entropy pool counters. We no longer clear
		 * the entropy pool, as that's silly.
		 */
L
Linus Torvalds 已提交
1347 1348
		if (!capable(CAP_SYS_ADMIN))
			return -EPERM;
1349
		if (xchg(&input_pool.entropy_count, 0)) {
1350 1351 1352
			wake_up_interruptible(&random_write_wait);
			kill_fasync(&fasync, SIGIO, POLL_OUT);
		}
L
Linus Torvalds 已提交
1353
		return 0;
1354 1355 1356 1357 1358
	case RNDRESEEDCRNG:
		if (!capable(CAP_SYS_ADMIN))
			return -EPERM;
		if (crng_init < 2)
			return -ENODATA;
1359
		crng_reseed();
1360
		return 0;
L
Linus Torvalds 已提交
1361 1362 1363 1364 1365
	default:
		return -EINVAL;
	}
}

1366 1367 1368 1369 1370
static int random_fasync(int fd, struct file *filp, int on)
{
	return fasync_helper(fd, filp, on, &fasync);
}

1371
const struct file_operations random_fops = {
1372
	.read = random_read,
L
Linus Torvalds 已提交
1373
	.write = random_write,
1374
	.poll = random_poll,
M
Matt Mackall 已提交
1375
	.unlocked_ioctl = random_ioctl,
1376
	.compat_ioctl = compat_ptr_ioctl,
1377
	.fasync = random_fasync,
1378
	.llseek = noop_llseek,
L
Linus Torvalds 已提交
1379 1380
};

1381
const struct file_operations urandom_fops = {
1382
	.read = urandom_read,
L
Linus Torvalds 已提交
1383
	.write = random_write,
M
Matt Mackall 已提交
1384
	.unlocked_ioctl = random_ioctl,
1385
	.compat_ioctl = compat_ptr_ioctl,
1386
	.fasync = random_fasync,
1387
	.llseek = noop_llseek,
L
Linus Torvalds 已提交
1388 1389
};

1390 1391
SYSCALL_DEFINE3(getrandom, char __user *, buf, size_t, count, unsigned int,
		flags)
1392
{
1393
	if (flags & ~(GRND_NONBLOCK | GRND_RANDOM | GRND_INSECURE))
1394 1395 1396 1397 1398 1399
		return -EINVAL;

	/*
	 * Requesting insecure and blocking randomness at the same time makes
	 * no sense.
	 */
1400
	if ((flags & (GRND_INSECURE | GRND_RANDOM)) == (GRND_INSECURE | GRND_RANDOM))
1401 1402 1403 1404 1405
		return -EINVAL;

	if (count > INT_MAX)
		count = INT_MAX;

1406
	if (!(flags & GRND_INSECURE) && !crng_ready()) {
1407 1408
		int ret;

1409 1410
		if (flags & GRND_NONBLOCK)
			return -EAGAIN;
1411 1412 1413
		ret = wait_for_random_bytes();
		if (unlikely(ret))
			return ret;
1414
	}
1415
	return get_random_bytes_user(buf, count);
1416 1417
}

L
Linus Torvalds 已提交
1418 1419 1420 1421 1422 1423 1424 1425 1426 1427
/********************************************************************
 *
 * Sysctl interface
 *
 ********************************************************************/

#ifdef CONFIG_SYSCTL

#include <linux/sysctl.h>

1428
static int random_min_urandom_seed = 60;
1429 1430
static int random_write_wakeup_bits = POOL_MIN_BITS;
static int sysctl_poolsize = POOL_BITS;
L
Linus Torvalds 已提交
1431 1432 1433
static char sysctl_bootid[16];

/*
G
Greg Price 已提交
1434
 * This function is used to return both the bootid UUID, and random
L
Linus Torvalds 已提交
1435 1436 1437
 * UUID.  The difference is in whether table->data is NULL; if it is,
 * then a new UUID is generated and returned to the user.
 *
G
Greg Price 已提交
1438 1439 1440
 * If the user accesses this via the proc interface, the UUID will be
 * returned as an ASCII string in the standard UUID format; if via the
 * sysctl system call, as 16 bytes of binary data.
L
Linus Torvalds 已提交
1441
 */
1442 1443
static int proc_do_uuid(struct ctl_table *table, int write, void *buffer,
			size_t *lenp, loff_t *ppos)
L
Linus Torvalds 已提交
1444
{
1445
	struct ctl_table fake_table;
L
Linus Torvalds 已提交
1446 1447 1448 1449 1450 1451
	unsigned char buf[64], tmp_uuid[16], *uuid;

	uuid = table->data;
	if (!uuid) {
		uuid = tmp_uuid;
		generate_random_uuid(uuid);
1452 1453 1454 1455 1456 1457 1458 1459
	} else {
		static DEFINE_SPINLOCK(bootid_spinlock);

		spin_lock(&bootid_spinlock);
		if (!uuid[8])
			generate_random_uuid(uuid);
		spin_unlock(&bootid_spinlock);
	}
L
Linus Torvalds 已提交
1460

J
Joe Perches 已提交
1461 1462
	sprintf(buf, "%pU", uuid);

L
Linus Torvalds 已提交
1463 1464 1465
	fake_table.data = buf;
	fake_table.maxlen = sizeof(buf);

1466
	return proc_dostring(&fake_table, write, buffer, lenp, ppos);
L
Linus Torvalds 已提交
1467 1468
}

1469
static struct ctl_table random_table[] = {
L
Linus Torvalds 已提交
1470 1471 1472 1473 1474
	{
		.procname	= "poolsize",
		.data		= &sysctl_poolsize,
		.maxlen		= sizeof(int),
		.mode		= 0444,
1475
		.proc_handler	= proc_dointvec,
L
Linus Torvalds 已提交
1476 1477 1478
	},
	{
		.procname	= "entropy_avail",
1479
		.data		= &input_pool.entropy_count,
L
Linus Torvalds 已提交
1480 1481
		.maxlen		= sizeof(int),
		.mode		= 0444,
1482
		.proc_handler	= proc_dointvec,
L
Linus Torvalds 已提交
1483 1484 1485
	},
	{
		.procname	= "write_wakeup_threshold",
1486
		.data		= &random_write_wakeup_bits,
L
Linus Torvalds 已提交
1487 1488
		.maxlen		= sizeof(int),
		.mode		= 0644,
1489
		.proc_handler	= proc_dointvec,
L
Linus Torvalds 已提交
1490
	},
1491 1492 1493 1494 1495 1496 1497
	{
		.procname	= "urandom_min_reseed_secs",
		.data		= &random_min_urandom_seed,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
L
Linus Torvalds 已提交
1498 1499 1500 1501 1502
	{
		.procname	= "boot_id",
		.data		= &sysctl_bootid,
		.maxlen		= 16,
		.mode		= 0444,
1503
		.proc_handler	= proc_do_uuid,
L
Linus Torvalds 已提交
1504 1505 1506 1507 1508
	},
	{
		.procname	= "uuid",
		.maxlen		= 16,
		.mode		= 0444,
1509
		.proc_handler	= proc_do_uuid,
L
Linus Torvalds 已提交
1510
	},
1511
	{ }
L
Linus Torvalds 已提交
1512
};
1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523

/*
 * rand_initialize() is called before sysctl_init(),
 * so we cannot call register_sysctl_init() in rand_initialize()
 */
static int __init random_sysctls_init(void)
{
	register_sysctl_init("kernel/random", random_table);
	return 0;
}
device_initcall(random_sysctls_init);
1524
#endif	/* CONFIG_SYSCTL */
L
Linus Torvalds 已提交
1525

1526 1527
struct batched_entropy {
	union {
1528 1529 1530 1531 1532 1533 1534 1535 1536
		/*
		 * We make this 1.5x a ChaCha block, so that we get the
		 * remaining 32 bytes from fast key erasure, plus one full
		 * block from the detached ChaCha state. We can increase
		 * the size of this later if needed so long as we keep the
		 * formula of (integer_blocks + 0.5) * CHACHA_BLOCK_SIZE.
		 */
		u64 entropy_u64[CHACHA_BLOCK_SIZE * 3 / (2 * sizeof(u64))];
		u32 entropy_u32[CHACHA_BLOCK_SIZE * 3 / (2 * sizeof(u32))];
1537
	};
1538
	local_lock_t lock;
1539
	unsigned long generation;
1540 1541
	unsigned int position;
};
1542

L
Linus Torvalds 已提交
1543
/*
1544
 * Get a random word for internal kernel use only. The quality of the random
1545 1546 1547
 * number is good as /dev/urandom. In order to ensure that the randomness
 * provided by this function is okay, the function wait_for_random_bytes()
 * should be called and return 0 at least once at any point prior.
L
Linus Torvalds 已提交
1548
 */
1549
static DEFINE_PER_CPU(struct batched_entropy, batched_entropy_u64) = {
1550 1551
	.lock = INIT_LOCAL_LOCK(batched_entropy_u64.lock),
	.position = UINT_MAX
1552 1553
};

1554
u64 get_random_u64(void)
L
Linus Torvalds 已提交
1555
{
1556
	u64 ret;
1557
	unsigned long flags;
1558
	struct batched_entropy *batch;
1559
	static void *previous;
1560
	unsigned long next_gen;
1561

1562
	warn_unseeded_randomness(&previous);
1563

1564
	local_lock_irqsave(&batched_entropy_u64.lock, flags);
1565
	batch = raw_cpu_ptr(&batched_entropy_u64);
1566

1567
	next_gen = READ_ONCE(base_crng.generation);
1568
	if (batch->position >= ARRAY_SIZE(batch->entropy_u64) ||
1569
	    next_gen != batch->generation) {
1570
		_get_random_bytes(batch->entropy_u64, sizeof(batch->entropy_u64));
1571
		batch->position = 0;
1572
		batch->generation = next_gen;
1573
	}
1574

1575 1576 1577
	ret = batch->entropy_u64[batch->position];
	batch->entropy_u64[batch->position] = 0;
	++batch->position;
1578
	local_unlock_irqrestore(&batched_entropy_u64.lock, flags);
1579
	return ret;
L
Linus Torvalds 已提交
1580
}
1581
EXPORT_SYMBOL(get_random_u64);
L
Linus Torvalds 已提交
1582

1583
static DEFINE_PER_CPU(struct batched_entropy, batched_entropy_u32) = {
1584 1585
	.lock = INIT_LOCAL_LOCK(batched_entropy_u32.lock),
	.position = UINT_MAX
1586
};
1587

1588
u32 get_random_u32(void)
1589
{
1590
	u32 ret;
1591
	unsigned long flags;
1592
	struct batched_entropy *batch;
1593
	static void *previous;
1594
	unsigned long next_gen;
1595

1596
	warn_unseeded_randomness(&previous);
1597

1598
	local_lock_irqsave(&batched_entropy_u32.lock, flags);
1599
	batch = raw_cpu_ptr(&batched_entropy_u32);
1600

1601
	next_gen = READ_ONCE(base_crng.generation);
1602
	if (batch->position >= ARRAY_SIZE(batch->entropy_u32) ||
1603
	    next_gen != batch->generation) {
1604
		_get_random_bytes(batch->entropy_u32, sizeof(batch->entropy_u32));
1605
		batch->position = 0;
1606
		batch->generation = next_gen;
1607
	}
1608

1609 1610 1611
	ret = batch->entropy_u32[batch->position];
	batch->entropy_u32[batch->position] = 0;
	++batch->position;
1612
	local_unlock_irqrestore(&batched_entropy_u32.lock, flags);
1613 1614
	return ret;
}
1615
EXPORT_SYMBOL(get_random_u32);
1616

1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630
/**
 * randomize_page - Generate a random, page aligned address
 * @start:	The smallest acceptable address the caller will take.
 * @range:	The size of the area, starting at @start, within which the
 *		random address must fall.
 *
 * If @start + @range would overflow, @range is capped.
 *
 * NOTE: Historical use of randomize_range, which this replaces, presumed that
 * @start was already page aligned.  We now align it regardless.
 *
 * Return: A page aligned address within [start, start + range).  On error,
 * @start is returned.
 */
1631
unsigned long randomize_page(unsigned long start, unsigned long range)
1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648
{
	if (!PAGE_ALIGNED(start)) {
		range -= PAGE_ALIGN(start) - start;
		start = PAGE_ALIGN(start);
	}

	if (start > ULONG_MAX - range)
		range = ULONG_MAX - start;

	range >>= PAGE_SHIFT;

	if (range == 0)
		return start;

	return start + (get_random_long() % range << PAGE_SHIFT);
}

1649 1650 1651 1652
/* Interface for in-kernel drivers of true hardware RNGs.
 * Those devices may produce endless random bits and will be throttled
 * when our pool is full.
 */
1653
void add_hwgenerator_randomness(const void *buffer, size_t count,
1654 1655
				size_t entropy)
{
T
Theodore Ts'o 已提交
1656
	if (unlikely(crng_init == 0)) {
1657
		size_t ret = crng_fast_load(buffer, count);
1658
		mix_pool_bytes(buffer, ret);
1659 1660 1661 1662
		count -= ret;
		buffer += ret;
		if (!count || crng_init == 0)
			return;
1663
	}
1664

1665
	/* Throttle writing if we're above the trickle threshold.
1666 1667 1668
	 * We'll be woken up again once below POOL_MIN_BITS, when
	 * the calling thread is about to terminate, or once
	 * CRNG_RESEED_INTERVAL has elapsed.
1669
	 */
1670
	wait_event_interruptible_timeout(random_write_wait,
1671
			!system_wq || kthread_should_stop() ||
1672
			input_pool.entropy_count < POOL_MIN_BITS,
1673
			CRNG_RESEED_INTERVAL);
1674 1675
	mix_pool_bytes(buffer, count);
	credit_entropy_bits(entropy);
1676 1677
}
EXPORT_SYMBOL_GPL(add_hwgenerator_randomness);
H
Hsin-Yi Wang 已提交
1678 1679 1680 1681 1682 1683

/* Handle random seed passed by bootloader.
 * If the seed is trustworthy, it would be regarded as hardware RNGs. Otherwise
 * it would be regarded as device data.
 * The decision is controlled by CONFIG_RANDOM_TRUST_BOOTLOADER.
 */
1684
void add_bootloader_randomness(const void *buf, size_t size)
H
Hsin-Yi Wang 已提交
1685 1686 1687 1688 1689 1690
{
	if (IS_ENABLED(CONFIG_RANDOM_TRUST_BOOTLOADER))
		add_hwgenerator_randomness(buf, size, size * 8);
	else
		add_device_randomness(buf, size);
}
1691
EXPORT_SYMBOL_GPL(add_bootloader_randomness);