lockdep.c 84.8 KB
Newer Older
I
Ingo Molnar 已提交
1 2 3 4 5 6 7
/*
 * kernel/lockdep.c
 *
 * Runtime locking correctness validator
 *
 * Started by Ingo Molnar:
 *
P
Peter Zijlstra 已提交
8 9
 *  Copyright (C) 2006,2007 Red Hat, Inc., Ingo Molnar <mingo@redhat.com>
 *  Copyright (C) 2007 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com>
I
Ingo Molnar 已提交
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
 *
 * this code maps all the lock dependencies as they occur in a live kernel
 * and will warn about the following classes of locking bugs:
 *
 * - lock inversion scenarios
 * - circular lock dependencies
 * - hardirq/softirq safe/unsafe locking bugs
 *
 * Bugs are reported even if the current locking scenario does not cause
 * any deadlock at this point.
 *
 * I.e. if anytime in the past two locks were taken in a different order,
 * even if it happened for another task, even if those were different
 * locks (but of the same class as this lock), this code will detect it.
 *
 * Thanks to Arjan van de Ven for coming up with the initial idea of
 * mapping lock dependencies runtime.
 */
28
#define DISABLE_BRANCH_PROFILING
I
Ingo Molnar 已提交
29 30 31 32 33 34 35 36 37 38 39 40
#include <linux/mutex.h>
#include <linux/sched.h>
#include <linux/delay.h>
#include <linux/module.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.h>
#include <linux/spinlock.h>
#include <linux/kallsyms.h>
#include <linux/interrupt.h>
#include <linux/stacktrace.h>
#include <linux/debug_locks.h>
#include <linux/irqflags.h>
41
#include <linux/utsname.h>
P
Peter Zijlstra 已提交
42
#include <linux/hash.h>
43
#include <linux/ftrace.h>
P
Peter Zijlstra 已提交
44
#include <linux/stringify.h>
I
Ingo Molnar 已提交
45 46 47 48 49

#include <asm/sections.h>

#include "lockdep_internals.h"

50 51 52
#define CREATE_TRACE_POINTS
#include <trace/lockdep.h>

P
Peter Zijlstra 已提交
53 54 55 56 57 58 59 60 61 62 63 64 65 66
#ifdef CONFIG_PROVE_LOCKING
int prove_locking = 1;
module_param(prove_locking, int, 0644);
#else
#define prove_locking 0
#endif

#ifdef CONFIG_LOCK_STAT
int lock_stat = 1;
module_param(lock_stat, int, 0644);
#else
#define lock_stat 0
#endif

I
Ingo Molnar 已提交
67
/*
68 69
 * lockdep_lock: protects the lockdep graph, the hashes and the
 *               class/list/hash allocators.
I
Ingo Molnar 已提交
70 71 72
 *
 * This is one of the rare exceptions where it's justified
 * to use a raw spinlock - we really dont want the spinlock
73
 * code to recurse back into the lockdep code...
I
Ingo Molnar 已提交
74
 */
75 76 77 78 79 80 81 82 83 84 85 86 87 88 89
static raw_spinlock_t lockdep_lock = (raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED;

static int graph_lock(void)
{
	__raw_spin_lock(&lockdep_lock);
	/*
	 * Make sure that if another CPU detected a bug while
	 * walking the graph we dont change it (while the other
	 * CPU is busy printing out stuff with the graph lock
	 * dropped already)
	 */
	if (!debug_locks) {
		__raw_spin_unlock(&lockdep_lock);
		return 0;
	}
90 91
	/* prevent any recursions within lockdep from causing deadlocks */
	current->lockdep_recursion++;
92 93 94 95 96
	return 1;
}

static inline int graph_unlock(void)
{
97 98 99
	if (debug_locks && !__raw_spin_is_locked(&lockdep_lock))
		return DEBUG_LOCKS_WARN_ON(1);

100
	current->lockdep_recursion--;
101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116
	__raw_spin_unlock(&lockdep_lock);
	return 0;
}

/*
 * Turn lock debugging off and return with 0 if it was off already,
 * and also release the graph lock:
 */
static inline int debug_locks_off_graph_unlock(void)
{
	int ret = debug_locks_off();

	__raw_spin_unlock(&lockdep_lock);

	return ret;
}
I
Ingo Molnar 已提交
117 118 119 120 121 122 123 124 125 126 127 128 129 130 131

static int lockdep_initialized;

unsigned long nr_list_entries;
static struct lock_list list_entries[MAX_LOCKDEP_ENTRIES];

/*
 * All data structures here are protected by the global debug_lock.
 *
 * Mutex key structs only get allocated, once during bootup, and never
 * get freed - this significantly simplifies the debugging code.
 */
unsigned long nr_lock_classes;
static struct lock_class lock_classes[MAX_LOCKDEP_KEYS];

D
Dave Jones 已提交
132 133 134 135 136 137 138 139 140
static inline struct lock_class *hlock_class(struct held_lock *hlock)
{
	if (!hlock->class_idx) {
		DEBUG_LOCKS_WARN_ON(1);
		return NULL;
	}
	return lock_classes + hlock->class_idx - 1;
}

P
Peter Zijlstra 已提交
141 142 143
#ifdef CONFIG_LOCK_STAT
static DEFINE_PER_CPU(struct lock_class_stats[MAX_LOCKDEP_KEYS], lock_stats);

P
Peter Zijlstra 已提交
144
static int lock_point(unsigned long points[], unsigned long ip)
P
Peter Zijlstra 已提交
145 146 147
{
	int i;

P
Peter Zijlstra 已提交
148 149 150
	for (i = 0; i < LOCKSTAT_POINTS; i++) {
		if (points[i] == 0) {
			points[i] = ip;
P
Peter Zijlstra 已提交
151 152
			break;
		}
P
Peter Zijlstra 已提交
153
		if (points[i] == ip)
P
Peter Zijlstra 已提交
154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171
			break;
	}

	return i;
}

static void lock_time_inc(struct lock_time *lt, s64 time)
{
	if (time > lt->max)
		lt->max = time;

	if (time < lt->min || !lt->min)
		lt->min = time;

	lt->total += time;
	lt->nr++;
}

172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192
static inline void lock_time_add(struct lock_time *src, struct lock_time *dst)
{
	dst->min += src->min;
	dst->max += src->max;
	dst->total += src->total;
	dst->nr += src->nr;
}

struct lock_class_stats lock_stats(struct lock_class *class)
{
	struct lock_class_stats stats;
	int cpu, i;

	memset(&stats, 0, sizeof(struct lock_class_stats));
	for_each_possible_cpu(cpu) {
		struct lock_class_stats *pcs =
			&per_cpu(lock_stats, cpu)[class - lock_classes];

		for (i = 0; i < ARRAY_SIZE(stats.contention_point); i++)
			stats.contention_point[i] += pcs->contention_point[i];

P
Peter Zijlstra 已提交
193 194 195
		for (i = 0; i < ARRAY_SIZE(stats.contending_point); i++)
			stats.contending_point[i] += pcs->contending_point[i];

196 197 198 199 200
		lock_time_add(&pcs->read_waittime, &stats.read_waittime);
		lock_time_add(&pcs->write_waittime, &stats.write_waittime);

		lock_time_add(&pcs->read_holdtime, &stats.read_holdtime);
		lock_time_add(&pcs->write_holdtime, &stats.write_holdtime);
P
Peter Zijlstra 已提交
201 202 203

		for (i = 0; i < ARRAY_SIZE(stats.bounces); i++)
			stats.bounces[i] += pcs->bounces[i];
204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219
	}

	return stats;
}

void clear_lock_stats(struct lock_class *class)
{
	int cpu;

	for_each_possible_cpu(cpu) {
		struct lock_class_stats *cpu_stats =
			&per_cpu(lock_stats, cpu)[class - lock_classes];

		memset(cpu_stats, 0, sizeof(struct lock_class_stats));
	}
	memset(class->contention_point, 0, sizeof(class->contention_point));
P
Peter Zijlstra 已提交
220
	memset(class->contending_point, 0, sizeof(class->contending_point));
221 222
}

P
Peter Zijlstra 已提交
223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242
static struct lock_class_stats *get_lock_stats(struct lock_class *class)
{
	return &get_cpu_var(lock_stats)[class - lock_classes];
}

static void put_lock_stats(struct lock_class_stats *stats)
{
	put_cpu_var(lock_stats);
}

static void lock_release_holdtime(struct held_lock *hlock)
{
	struct lock_class_stats *stats;
	s64 holdtime;

	if (!lock_stat)
		return;

	holdtime = sched_clock() - hlock->holdtime_stamp;

D
Dave Jones 已提交
243
	stats = get_lock_stats(hlock_class(hlock));
P
Peter Zijlstra 已提交
244 245 246 247 248 249 250 251 252 253 254 255
	if (hlock->read)
		lock_time_inc(&stats->read_holdtime, holdtime);
	else
		lock_time_inc(&stats->write_holdtime, holdtime);
	put_lock_stats(stats);
}
#else
static inline void lock_release_holdtime(struct held_lock *hlock)
{
}
#endif

I
Ingo Molnar 已提交
256 257 258 259 260 261 262 263 264 265 266 267
/*
 * We keep a global list of all lock classes. The list only grows,
 * never shrinks. The list is only accessed with the lockdep
 * spinlock lock held.
 */
LIST_HEAD(all_lock_classes);

/*
 * The lockdep classes are in a hash-table as well, for fast lookup:
 */
#define CLASSHASH_BITS		(MAX_LOCKDEP_KEYS_BITS - 1)
#define CLASSHASH_SIZE		(1UL << CLASSHASH_BITS)
P
Peter Zijlstra 已提交
268
#define __classhashfn(key)	hash_long((unsigned long)key, CLASSHASH_BITS)
I
Ingo Molnar 已提交
269 270 271 272 273 274 275 276 277 278
#define classhashentry(key)	(classhash_table + __classhashfn((key)))

static struct list_head classhash_table[CLASSHASH_SIZE];

/*
 * We put the lock dependency chains into a hash-table as well, to cache
 * their existence:
 */
#define CHAINHASH_BITS		(MAX_LOCKDEP_CHAINS_BITS-1)
#define CHAINHASH_SIZE		(1UL << CHAINHASH_BITS)
P
Peter Zijlstra 已提交
279
#define __chainhashfn(chain)	hash_long(chain, CHAINHASH_BITS)
I
Ingo Molnar 已提交
280 281 282 283 284 285 286 287 288 289 290
#define chainhashentry(chain)	(chainhash_table + __chainhashfn((chain)))

static struct list_head chainhash_table[CHAINHASH_SIZE];

/*
 * The hash key of the lock dependency chains is a hash itself too:
 * it's a hash of all locks taken up to that lock, including that lock.
 * It's a 64-bit hash, because it's important for the keys to be
 * unique.
 */
#define iterate_chain_key(key1, key2) \
291 292
	(((key1) << MAX_LOCKDEP_KEYS_BITS) ^ \
	((key1) >> (64-MAX_LOCKDEP_KEYS_BITS)) ^ \
I
Ingo Molnar 已提交
293 294
	(key2))

295
void lockdep_off(void)
I
Ingo Molnar 已提交
296 297 298 299 300
{
	current->lockdep_recursion++;
}
EXPORT_SYMBOL(lockdep_off);

301
void lockdep_on(void)
I
Ingo Molnar 已提交
302 303 304 305 306 307 308 309 310 311
{
	current->lockdep_recursion--;
}
EXPORT_SYMBOL(lockdep_on);

/*
 * Debugging switches:
 */

#define VERBOSE			0
312
#define VERY_VERBOSE		0
I
Ingo Molnar 已提交
313 314 315 316

#if VERBOSE
# define HARDIRQ_VERBOSE	1
# define SOFTIRQ_VERBOSE	1
317
# define RECLAIM_VERBOSE	1
I
Ingo Molnar 已提交
318 319 320
#else
# define HARDIRQ_VERBOSE	0
# define SOFTIRQ_VERBOSE	0
321
# define RECLAIM_VERBOSE	0
I
Ingo Molnar 已提交
322 323
#endif

324
#if VERBOSE || HARDIRQ_VERBOSE || SOFTIRQ_VERBOSE || RECLAIM_VERBOSE
I
Ingo Molnar 已提交
325 326 327 328 329
/*
 * Quick filtering for interesting events:
 */
static int class_filter(struct lock_class *class)
{
A
Andi Kleen 已提交
330 331
#if 0
	/* Example */
I
Ingo Molnar 已提交
332
	if (class->name_version == 1 &&
A
Andi Kleen 已提交
333
			!strcmp(class->name, "lockname"))
I
Ingo Molnar 已提交
334 335
		return 1;
	if (class->name_version == 1 &&
A
Andi Kleen 已提交
336
			!strcmp(class->name, "&struct->lockfield"))
I
Ingo Molnar 已提交
337
		return 1;
A
Andi Kleen 已提交
338
#endif
339 340
	/* Filter everything else. 1 would be to allow everything else */
	return 0;
I
Ingo Molnar 已提交
341 342 343 344 345 346 347 348 349 350 351 352 353
}
#endif

static int verbose(struct lock_class *class)
{
#if VERBOSE
	return class_filter(class);
#endif
	return 0;
}

/*
 * Stack-trace: tightly packed array of stack backtrace
354
 * addresses. Protected by the graph_lock.
I
Ingo Molnar 已提交
355 356 357 358 359 360 361 362 363 364
 */
unsigned long nr_stack_trace_entries;
static unsigned long stack_trace[MAX_STACK_TRACE_ENTRIES];

static int save_trace(struct stack_trace *trace)
{
	trace->nr_entries = 0;
	trace->max_entries = MAX_STACK_TRACE_ENTRIES - nr_stack_trace_entries;
	trace->entries = stack_trace + nr_stack_trace_entries;

365 366
	trace->skip = 3;

C
Christoph Hellwig 已提交
367
	save_stack_trace(trace);
I
Ingo Molnar 已提交
368 369 370 371 372 373

	trace->max_entries = trace->nr_entries;

	nr_stack_trace_entries += trace->nr_entries;

	if (nr_stack_trace_entries == MAX_STACK_TRACE_ENTRIES) {
374 375 376 377 378 379 380
		if (!debug_locks_off_graph_unlock())
			return 0;

		printk("BUG: MAX_STACK_TRACE_ENTRIES too low!\n");
		printk("turning off the locking correctness validator.\n");
		dump_stack();

I
Ingo Molnar 已提交
381 382 383 384 385 386 387 388 389 390 391 392
		return 0;
	}

	return 1;
}

unsigned int nr_hardirq_chains;
unsigned int nr_softirq_chains;
unsigned int nr_process_chains;
unsigned int max_lockdep_depth;
unsigned int max_recursion_depth;

393 394 395 396 397 398 399 400 401 402 403 404 405
static unsigned int lockdep_dependency_gen_id;

static bool lockdep_dependency_visit(struct lock_class *source,
				     unsigned int depth)
{
	if (!depth)
		lockdep_dependency_gen_id++;
	if (source->dep_gen_id == lockdep_dependency_gen_id)
		return true;
	source->dep_gen_id = lockdep_dependency_gen_id;
	return false;
}

I
Ingo Molnar 已提交
406 407 408 409 410 411 412
#ifdef CONFIG_DEBUG_LOCKDEP
/*
 * We cannot printk in early bootup code. Not even early_printk()
 * might work. So we mark any initialization errors and printk
 * about it later on, in lockdep_info().
 */
static int lockdep_init_error;
413 414 415 416 417
static unsigned long lockdep_init_trace_data[20];
static struct stack_trace lockdep_init_trace = {
	.max_entries = ARRAY_SIZE(lockdep_init_trace_data),
	.entries = lockdep_init_trace_data,
};
I
Ingo Molnar 已提交
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

/*
 * Various lockdep statistics:
 */
atomic_t chain_lookup_hits;
atomic_t chain_lookup_misses;
atomic_t hardirqs_on_events;
atomic_t hardirqs_off_events;
atomic_t redundant_hardirqs_on;
atomic_t redundant_hardirqs_off;
atomic_t softirqs_on_events;
atomic_t softirqs_off_events;
atomic_t redundant_softirqs_on;
atomic_t redundant_softirqs_off;
atomic_t nr_unused_locks;
atomic_t nr_cyclic_checks;
atomic_t nr_cyclic_check_recursions;
atomic_t nr_find_usage_forwards_checks;
atomic_t nr_find_usage_forwards_recursions;
atomic_t nr_find_usage_backwards_checks;
atomic_t nr_find_usage_backwards_recursions;
#endif

/*
 * Locking printouts:
 */

P
Peter Zijlstra 已提交
445
#define __USAGE(__STATE)						\
P
Peter Zijlstra 已提交
446 447 448 449
	[LOCK_USED_IN_##__STATE] = "IN-"__stringify(__STATE)"-W",	\
	[LOCK_ENABLED_##__STATE] = __stringify(__STATE)"-ON-W",		\
	[LOCK_USED_IN_##__STATE##_READ] = "IN-"__stringify(__STATE)"-R",\
	[LOCK_ENABLED_##__STATE##_READ] = __stringify(__STATE)"-ON-R",
P
Peter Zijlstra 已提交
450

I
Ingo Molnar 已提交
451 452
static const char *usage_str[] =
{
P
Peter Zijlstra 已提交
453 454 455 456
#define LOCKDEP_STATE(__STATE) __USAGE(__STATE)
#include "lockdep_states.h"
#undef LOCKDEP_STATE
	[LOCK_USED] = "INITIAL USE",
I
Ingo Molnar 已提交
457 458 459 460
};

const char * __get_key_name(struct lockdep_subclass_key *key, char *str)
{
A
Alexey Dobriyan 已提交
461
	return kallsyms_lookup((unsigned long)key, NULL, NULL, NULL, str);
I
Ingo Molnar 已提交
462 463
}

464
static inline unsigned long lock_flag(enum lock_usage_bit bit)
I
Ingo Molnar 已提交
465
{
466 467
	return 1UL << bit;
}
I
Ingo Molnar 已提交
468

469 470 471 472 473 474 475 476 477 478
static char get_usage_char(struct lock_class *class, enum lock_usage_bit bit)
{
	char c = '.';

	if (class->usage_mask & lock_flag(bit + 2))
		c = '+';
	if (class->usage_mask & lock_flag(bit)) {
		c = '-';
		if (class->usage_mask & lock_flag(bit + 2))
			c = '?';
I
Ingo Molnar 已提交
479 480
	}

481 482
	return c;
}
483

P
Peter Zijlstra 已提交
484
void get_usage_chars(struct lock_class *class, char usage[LOCK_USAGE_CHARS])
485
{
P
Peter Zijlstra 已提交
486
	int i = 0;
487

P
Peter Zijlstra 已提交
488 489 490 491 492 493 494
#define LOCKDEP_STATE(__STATE) 						\
	usage[i++] = get_usage_char(class, LOCK_USED_IN_##__STATE);	\
	usage[i++] = get_usage_char(class, LOCK_USED_IN_##__STATE##_READ);
#include "lockdep_states.h"
#undef LOCKDEP_STATE

	usage[i] = '\0';
I
Ingo Molnar 已提交
495 496 497 498
}

static void print_lock_name(struct lock_class *class)
{
P
Peter Zijlstra 已提交
499
	char str[KSYM_NAME_LEN], usage[LOCK_USAGE_CHARS];
I
Ingo Molnar 已提交
500 501
	const char *name;

P
Peter Zijlstra 已提交
502
	get_usage_chars(class, usage);
I
Ingo Molnar 已提交
503 504 505 506 507 508 509 510 511 512 513 514

	name = class->name;
	if (!name) {
		name = __get_key_name(class->key, str);
		printk(" (%s", name);
	} else {
		printk(" (%s", name);
		if (class->name_version > 1)
			printk("#%d", class->name_version);
		if (class->subclass)
			printk("/%d", class->subclass);
	}
P
Peter Zijlstra 已提交
515
	printk("){%s}", usage);
I
Ingo Molnar 已提交
516 517 518 519 520
}

static void print_lockdep_cache(struct lockdep_map *lock)
{
	const char *name;
521
	char str[KSYM_NAME_LEN];
I
Ingo Molnar 已提交
522 523 524 525 526 527 528 529 530 531

	name = lock->name;
	if (!name)
		name = __get_key_name(lock->key->subkeys, str);

	printk("%s", name);
}

static void print_lock(struct held_lock *hlock)
{
D
Dave Jones 已提交
532
	print_lock_name(hlock_class(hlock));
I
Ingo Molnar 已提交
533 534 535 536 537 538 539 540 541
	printk(", at: ");
	print_ip_sym(hlock->acquire_ip);
}

static void lockdep_print_held_locks(struct task_struct *curr)
{
	int i, depth = curr->lockdep_depth;

	if (!depth) {
542
		printk("no locks held by %s/%d.\n", curr->comm, task_pid_nr(curr));
I
Ingo Molnar 已提交
543 544 545
		return;
	}
	printk("%d lock%s held by %s/%d:\n",
546
		depth, depth > 1 ? "s" : "", curr->comm, task_pid_nr(curr));
I
Ingo Molnar 已提交
547 548 549 550 551 552 553 554 555 556 557

	for (i = 0; i < depth; i++) {
		printk(" #%d: ", i);
		print_lock(curr->held_locks + i);
	}
}

static void print_lock_class_header(struct lock_class *class, int depth)
{
	int bit;

A
Andi Kleen 已提交
558
	printk("%*s->", depth, "");
I
Ingo Molnar 已提交
559 560 561 562 563 564 565 566
	print_lock_name(class);
	printk(" ops: %lu", class->ops);
	printk(" {\n");

	for (bit = 0; bit < LOCK_USAGE_STATES; bit++) {
		if (class->usage_mask & (1 << bit)) {
			int len = depth;

A
Andi Kleen 已提交
567
			len += printk("%*s   %s", depth, "", usage_str[bit]);
I
Ingo Molnar 已提交
568 569 570 571
			len += printk(" at:\n");
			print_stack_trace(class->usage_traces + bit, len);
		}
	}
A
Andi Kleen 已提交
572
	printk("%*s }\n", depth, "");
I
Ingo Molnar 已提交
573

A
Andi Kleen 已提交
574
	printk("%*s ... key      at: ",depth,"");
I
Ingo Molnar 已提交
575 576 577 578 579 580
	print_ip_sym((unsigned long)class->key);
}

/*
 * printk all lock dependencies starting at <entry>:
 */
581 582
static void __used
print_lock_dependencies(struct lock_class *class, int depth)
I
Ingo Molnar 已提交
583 584 585
{
	struct lock_list *entry;

586 587 588
	if (lockdep_dependency_visit(class, depth))
		return;

I
Ingo Molnar 已提交
589 590 591 592 593 594
	if (DEBUG_LOCKS_WARN_ON(depth >= 20))
		return;

	print_lock_class_header(class, depth);

	list_for_each_entry(entry, &class->locks_after, entry) {
595 596 597
		if (DEBUG_LOCKS_WARN_ON(!entry->class))
			return;

I
Ingo Molnar 已提交
598 599
		print_lock_dependencies(entry->class, depth + 1);

A
Andi Kleen 已提交
600
		printk("%*s ... acquired at:\n",depth,"");
I
Ingo Molnar 已提交
601 602 603 604 605
		print_stack_trace(&entry->trace, 2);
		printk("\n");
	}
}

P
Peter Zijlstra 已提交
606 607 608 609 610 611 612 613 614 615 616 617 618 619 620
static void print_kernel_version(void)
{
	printk("%s %.*s\n", init_utsname()->release,
		(int)strcspn(init_utsname()->version, " "),
		init_utsname()->version);
}

static int very_verbose(struct lock_class *class)
{
#if VERY_VERBOSE
	return class_filter(class);
#endif
	return 0;
}

I
Ingo Molnar 已提交
621
/*
P
Peter Zijlstra 已提交
622
 * Is this the address of a static object:
I
Ingo Molnar 已提交
623
 */
P
Peter Zijlstra 已提交
624
static int static_obj(void *obj)
I
Ingo Molnar 已提交
625
{
P
Peter Zijlstra 已提交
626 627 628 629 630 631 632
	unsigned long start = (unsigned long) &_stext,
		      end   = (unsigned long) &_end,
		      addr  = (unsigned long) obj;
#ifdef CONFIG_SMP
	int i;
#endif

I
Ingo Molnar 已提交
633
	/*
P
Peter Zijlstra 已提交
634
	 * static variable?
I
Ingo Molnar 已提交
635
	 */
P
Peter Zijlstra 已提交
636 637
	if ((addr >= start) && (addr < end))
		return 1;
I
Ingo Molnar 已提交
638

P
Peter Zijlstra 已提交
639
#ifdef CONFIG_SMP
I
Ingo Molnar 已提交
640
	/*
P
Peter Zijlstra 已提交
641
	 * percpu var?
I
Ingo Molnar 已提交
642
	 */
P
Peter Zijlstra 已提交
643 644 645 646
	for_each_possible_cpu(i) {
		start = (unsigned long) &__per_cpu_start + per_cpu_offset(i);
		end   = (unsigned long) &__per_cpu_start + PERCPU_ENOUGH_ROOM
					+ per_cpu_offset(i);
I
Ingo Molnar 已提交
647

P
Peter Zijlstra 已提交
648 649 650
		if ((addr >= start) && (addr < end))
			return 1;
	}
651
#endif
I
Ingo Molnar 已提交
652

P
Peter Zijlstra 已提交
653 654 655 656
	/*
	 * module var?
	 */
	return is_module_address(addr);
657 658
}

I
Ingo Molnar 已提交
659
/*
P
Peter Zijlstra 已提交
660 661
 * To make lock name printouts unique, we calculate a unique
 * class->name_version generation counter:
I
Ingo Molnar 已提交
662
 */
P
Peter Zijlstra 已提交
663
static int count_matching_names(struct lock_class *new_class)
I
Ingo Molnar 已提交
664
{
P
Peter Zijlstra 已提交
665 666
	struct lock_class *class;
	int count = 0;
I
Ingo Molnar 已提交
667

P
Peter Zijlstra 已提交
668
	if (!new_class->name)
I
Ingo Molnar 已提交
669 670
		return 0;

P
Peter Zijlstra 已提交
671 672 673 674 675 676
	list_for_each_entry(class, &all_lock_classes, lock_entry) {
		if (new_class->key - new_class->subclass == class->key)
			return class->name_version;
		if (class->name && !strcmp(class->name, new_class->name))
			count = max(count, class->name_version);
	}
I
Ingo Molnar 已提交
677

P
Peter Zijlstra 已提交
678
	return count + 1;
I
Ingo Molnar 已提交
679 680
}

P
Peter Zijlstra 已提交
681 682 683 684 685 686 687
/*
 * Register a lock's class in the hash-table, if the class is not present
 * yet. Otherwise we look it up. We cache the result in the lock object
 * itself, so actual lookup of the hash should be once per lock object.
 */
static inline struct lock_class *
look_up_lock_class(struct lockdep_map *lock, unsigned int subclass)
I
Ingo Molnar 已提交
688
{
P
Peter Zijlstra 已提交
689 690 691
	struct lockdep_subclass_key *key;
	struct list_head *hash_head;
	struct lock_class *class;
I
Ingo Molnar 已提交
692

P
Peter Zijlstra 已提交
693 694 695 696 697 698 699 700 701
#ifdef CONFIG_DEBUG_LOCKDEP
	/*
	 * If the architecture calls into lockdep before initializing
	 * the hashes then we'll warn about it later. (we cannot printk
	 * right now)
	 */
	if (unlikely(!lockdep_initialized)) {
		lockdep_init();
		lockdep_init_error = 1;
702
		save_stack_trace(&lockdep_init_trace);
P
Peter Zijlstra 已提交
703 704
	}
#endif
I
Ingo Molnar 已提交
705

P
Peter Zijlstra 已提交
706 707 708 709 710 711
	/*
	 * Static locks do not have their class-keys yet - for them the key
	 * is the lock object itself:
	 */
	if (unlikely(!lock->key))
		lock->key = (void *)lock;
I
Ingo Molnar 已提交
712

P
Peter Zijlstra 已提交
713 714 715 716 717 718
	/*
	 * NOTE: the class-key must be unique. For dynamic locks, a static
	 * lock_class_key variable is passed in through the mutex_init()
	 * (or spin_lock_init()) call - which acts as the key. For static
	 * locks we use the lock object itself as the key.
	 */
P
Peter Zijlstra 已提交
719 720
	BUILD_BUG_ON(sizeof(struct lock_class_key) >
			sizeof(struct lockdep_map));
I
Ingo Molnar 已提交
721

P
Peter Zijlstra 已提交
722
	key = lock->key->subkeys + subclass;
723

P
Peter Zijlstra 已提交
724
	hash_head = classhashentry(key);
725

P
Peter Zijlstra 已提交
726 727 728 729
	/*
	 * We can walk the hash lockfree, because the hash only
	 * grows, and we are careful when adding entries to the end:
	 */
P
Peter Zijlstra 已提交
730 731 732
	list_for_each_entry(class, hash_head, hash_entry) {
		if (class->key == key) {
			WARN_ON_ONCE(class->name != lock->name);
P
Peter Zijlstra 已提交
733
			return class;
P
Peter Zijlstra 已提交
734 735
		}
	}
I
Ingo Molnar 已提交
736

P
Peter Zijlstra 已提交
737
	return NULL;
I
Ingo Molnar 已提交
738 739 740
}

/*
P
Peter Zijlstra 已提交
741 742 743
 * Register a lock's class in the hash-table, if the class is not present
 * yet. Otherwise we look it up. We cache the result in the lock object
 * itself, so actual lookup of the hash should be once per lock object.
I
Ingo Molnar 已提交
744
 */
P
Peter Zijlstra 已提交
745 746
static inline struct lock_class *
register_lock_class(struct lockdep_map *lock, unsigned int subclass, int force)
I
Ingo Molnar 已提交
747
{
P
Peter Zijlstra 已提交
748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797
	struct lockdep_subclass_key *key;
	struct list_head *hash_head;
	struct lock_class *class;
	unsigned long flags;

	class = look_up_lock_class(lock, subclass);
	if (likely(class))
		return class;

	/*
	 * Debug-check: all keys must be persistent!
 	 */
	if (!static_obj(lock->key)) {
		debug_locks_off();
		printk("INFO: trying to register non-static key.\n");
		printk("the code is fine but needs lockdep annotation.\n");
		printk("turning off the locking correctness validator.\n");
		dump_stack();

		return NULL;
	}

	key = lock->key->subkeys + subclass;
	hash_head = classhashentry(key);

	raw_local_irq_save(flags);
	if (!graph_lock()) {
		raw_local_irq_restore(flags);
		return NULL;
	}
	/*
	 * We have to do the hash-walk again, to avoid races
	 * with another CPU:
	 */
	list_for_each_entry(class, hash_head, hash_entry)
		if (class->key == key)
			goto out_unlock_set;
	/*
	 * Allocate a new key from the static array, and add it to
	 * the hash:
	 */
	if (nr_lock_classes >= MAX_LOCKDEP_KEYS) {
		if (!debug_locks_off_graph_unlock()) {
			raw_local_irq_restore(flags);
			return NULL;
		}
		raw_local_irq_restore(flags);

		printk("BUG: MAX_LOCKDEP_KEYS too low!\n");
		printk("turning off the locking correctness validator.\n");
798
		dump_stack();
P
Peter Zijlstra 已提交
799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814
		return NULL;
	}
	class = lock_classes + nr_lock_classes++;
	debug_atomic_inc(&nr_unused_locks);
	class->key = key;
	class->name = lock->name;
	class->subclass = subclass;
	INIT_LIST_HEAD(&class->lock_entry);
	INIT_LIST_HEAD(&class->locks_before);
	INIT_LIST_HEAD(&class->locks_after);
	class->name_version = count_matching_names(class);
	/*
	 * We use RCU's safe list-add method to make
	 * parallel walking of the hash-list safe:
	 */
	list_add_tail_rcu(&class->hash_entry, hash_head);
815 816 817 818
	/*
	 * Add it to the global list of classes:
	 */
	list_add_tail_rcu(&class->lock_entry, &all_lock_classes);
P
Peter Zijlstra 已提交
819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861

	if (verbose(class)) {
		graph_unlock();
		raw_local_irq_restore(flags);

		printk("\nnew class %p: %s", class->key, class->name);
		if (class->name_version > 1)
			printk("#%d", class->name_version);
		printk("\n");
		dump_stack();

		raw_local_irq_save(flags);
		if (!graph_lock()) {
			raw_local_irq_restore(flags);
			return NULL;
		}
	}
out_unlock_set:
	graph_unlock();
	raw_local_irq_restore(flags);

	if (!subclass || force)
		lock->class_cache = class;

	if (DEBUG_LOCKS_WARN_ON(class->subclass != subclass))
		return NULL;

	return class;
}

#ifdef CONFIG_PROVE_LOCKING
/*
 * Allocate a lockdep entry. (assumes the graph_lock held, returns
 * with NULL on failure)
 */
static struct lock_list *alloc_list_entry(void)
{
	if (nr_list_entries >= MAX_LOCKDEP_ENTRIES) {
		if (!debug_locks_off_graph_unlock())
			return NULL;

		printk("BUG: MAX_LOCKDEP_ENTRIES too low!\n");
		printk("turning off the locking correctness validator.\n");
862
		dump_stack();
P
Peter Zijlstra 已提交
863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885
		return NULL;
	}
	return list_entries + nr_list_entries++;
}

/*
 * Add a new dependency to the head of the list:
 */
static int add_lock_to_list(struct lock_class *class, struct lock_class *this,
			    struct list_head *head, unsigned long ip, int distance)
{
	struct lock_list *entry;
	/*
	 * Lock not present yet - get a new dependency struct and
	 * add it to the list:
	 */
	entry = alloc_list_entry();
	if (!entry)
		return 0;

	if (!save_trace(&entry->trace))
		return 0;

886 887
	entry->class = this;
	entry->distance = distance;
P
Peter Zijlstra 已提交
888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944
	/*
	 * Since we never remove from the dependency list, the list can
	 * be walked lockless by other CPUs, it's only allocation
	 * that must be protected by the spinlock. But this also means
	 * we must make new entries visible only once writes to the
	 * entry become visible - hence the RCU op:
	 */
	list_add_tail_rcu(&entry->entry, head);

	return 1;
}

/*
 * Recursive, forwards-direction lock-dependency checking, used for
 * both noncyclic checking and for hardirq-unsafe/softirq-unsafe
 * checking.
 *
 * (to keep the stackframe of the recursive functions small we
 *  use these global variables, and we also mark various helper
 *  functions as noinline.)
 */
static struct held_lock *check_source, *check_target;

/*
 * Print a dependency chain entry (this is only done when a deadlock
 * has been detected):
 */
static noinline int
print_circular_bug_entry(struct lock_list *target, unsigned int depth)
{
	if (debug_locks_silent)
		return 0;
	printk("\n-> #%u", depth);
	print_lock_name(target->class);
	printk(":\n");
	print_stack_trace(&target->trace, 6);

	return 0;
}

/*
 * When a circular dependency is detected, print the
 * header first:
 */
static noinline int
print_circular_bug_header(struct lock_list *entry, unsigned int depth)
{
	struct task_struct *curr = current;

	if (!debug_locks_off_graph_unlock() || debug_locks_silent)
		return 0;

	printk("\n=======================================================\n");
	printk(  "[ INFO: possible circular locking dependency detected ]\n");
	print_kernel_version();
	printk(  "-------------------------------------------------------\n");
	printk("%s/%d is trying to acquire lock:\n",
945
		curr->comm, task_pid_nr(curr));
P
Peter Zijlstra 已提交
946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964
	print_lock(check_source);
	printk("\nbut task is already holding lock:\n");
	print_lock(check_target);
	printk("\nwhich lock already depends on the new lock.\n\n");
	printk("\nthe existing dependency chain (in reverse order) is:\n");

	print_circular_bug_entry(entry, depth);

	return 0;
}

static noinline int print_circular_bug_tail(void)
{
	struct task_struct *curr = current;
	struct lock_list this;

	if (debug_locks_silent)
		return 0;

D
Dave Jones 已提交
965
	this.class = hlock_class(check_source);
P
Peter Zijlstra 已提交
966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991
	if (!save_trace(&this.trace))
		return 0;

	print_circular_bug_entry(&this, 0);

	printk("\nother info that might help us debug this:\n\n");
	lockdep_print_held_locks(curr);

	printk("\nstack backtrace:\n");
	dump_stack();

	return 0;
}

#define RECURSION_LIMIT 40

static int noinline print_infinite_recursion_bug(void)
{
	if (!debug_locks_off_graph_unlock())
		return 0;

	WARN_ON(1);

	return 0;
}

992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052
unsigned long __lockdep_count_forward_deps(struct lock_class *class,
					   unsigned int depth)
{
	struct lock_list *entry;
	unsigned long ret = 1;

	if (lockdep_dependency_visit(class, depth))
		return 0;

	/*
	 * Recurse this class's dependency list:
	 */
	list_for_each_entry(entry, &class->locks_after, entry)
		ret += __lockdep_count_forward_deps(entry->class, depth + 1);

	return ret;
}

unsigned long lockdep_count_forward_deps(struct lock_class *class)
{
	unsigned long ret, flags;

	local_irq_save(flags);
	__raw_spin_lock(&lockdep_lock);
	ret = __lockdep_count_forward_deps(class, 0);
	__raw_spin_unlock(&lockdep_lock);
	local_irq_restore(flags);

	return ret;
}

unsigned long __lockdep_count_backward_deps(struct lock_class *class,
					    unsigned int depth)
{
	struct lock_list *entry;
	unsigned long ret = 1;

	if (lockdep_dependency_visit(class, depth))
		return 0;
	/*
	 * Recurse this class's dependency list:
	 */
	list_for_each_entry(entry, &class->locks_before, entry)
		ret += __lockdep_count_backward_deps(entry->class, depth + 1);

	return ret;
}

unsigned long lockdep_count_backward_deps(struct lock_class *class)
{
	unsigned long ret, flags;

	local_irq_save(flags);
	__raw_spin_lock(&lockdep_lock);
	ret = __lockdep_count_backward_deps(class, 0);
	__raw_spin_unlock(&lockdep_lock);
	local_irq_restore(flags);

	return ret;
}

P
Peter Zijlstra 已提交
1053 1054 1055 1056 1057 1058 1059 1060 1061
/*
 * Prove that the dependency graph starting at <entry> can not
 * lead to <target>. Print an error and return 0 if it does.
 */
static noinline int
check_noncircular(struct lock_class *source, unsigned int depth)
{
	struct lock_list *entry;

1062 1063 1064
	if (lockdep_dependency_visit(source, depth))
		return 1;

P
Peter Zijlstra 已提交
1065 1066
	debug_atomic_inc(&nr_cyclic_check_recursions);
	if (depth > max_recursion_depth)
I
Ingo Molnar 已提交
1067
		max_recursion_depth = depth;
1068
	if (depth >= RECURSION_LIMIT)
I
Ingo Molnar 已提交
1069 1070 1071 1072 1073
		return print_infinite_recursion_bug();
	/*
	 * Check this lock's dependency list:
	 */
	list_for_each_entry(entry, &source->locks_after, entry) {
D
Dave Jones 已提交
1074
		if (entry->class == hlock_class(check_target))
I
Ingo Molnar 已提交
1075 1076 1077 1078 1079 1080 1081 1082
			return print_circular_bug_header(entry, depth+1);
		debug_atomic_inc(&nr_cyclic_checks);
		if (!check_noncircular(entry->class, depth+1))
			return print_circular_bug_entry(entry, depth+1);
	}
	return 1;
}

1083
#if defined(CONFIG_TRACE_IRQFLAGS) && defined(CONFIG_PROVE_LOCKING)
I
Ingo Molnar 已提交
1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107
/*
 * Forwards and backwards subgraph searching, for the purposes of
 * proving that two subgraphs can be connected by a new dependency
 * without creating any illegal irq-safe -> irq-unsafe lock dependency.
 */
static enum lock_usage_bit find_usage_bit;
static struct lock_class *forwards_match, *backwards_match;

/*
 * Find a node in the forwards-direction dependency sub-graph starting
 * at <source> that matches <find_usage_bit>.
 *
 * Return 2 if such a node exists in the subgraph, and put that node
 * into <forwards_match>.
 *
 * Return 1 otherwise and keep <forwards_match> unchanged.
 * Return 0 on error.
 */
static noinline int
find_usage_forwards(struct lock_class *source, unsigned int depth)
{
	struct lock_list *entry;
	int ret;

1108 1109 1110
	if (lockdep_dependency_visit(source, depth))
		return 1;

I
Ingo Molnar 已提交
1111 1112
	if (depth > max_recursion_depth)
		max_recursion_depth = depth;
1113
	if (depth >= RECURSION_LIMIT)
I
Ingo Molnar 已提交
1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143
		return print_infinite_recursion_bug();

	debug_atomic_inc(&nr_find_usage_forwards_checks);
	if (source->usage_mask & (1 << find_usage_bit)) {
		forwards_match = source;
		return 2;
	}

	/*
	 * Check this lock's dependency list:
	 */
	list_for_each_entry(entry, &source->locks_after, entry) {
		debug_atomic_inc(&nr_find_usage_forwards_recursions);
		ret = find_usage_forwards(entry->class, depth+1);
		if (ret == 2 || ret == 0)
			return ret;
	}
	return 1;
}

/*
 * Find a node in the backwards-direction dependency sub-graph starting
 * at <source> that matches <find_usage_bit>.
 *
 * Return 2 if such a node exists in the subgraph, and put that node
 * into <backwards_match>.
 *
 * Return 1 otherwise and keep <backwards_match> unchanged.
 * Return 0 on error.
 */
1144
static noinline int
I
Ingo Molnar 已提交
1145 1146 1147 1148 1149
find_usage_backwards(struct lock_class *source, unsigned int depth)
{
	struct lock_list *entry;
	int ret;

1150 1151 1152
	if (lockdep_dependency_visit(source, depth))
		return 1;

1153 1154 1155
	if (!__raw_spin_is_locked(&lockdep_lock))
		return DEBUG_LOCKS_WARN_ON(1);

I
Ingo Molnar 已提交
1156 1157
	if (depth > max_recursion_depth)
		max_recursion_depth = depth;
1158
	if (depth >= RECURSION_LIMIT)
I
Ingo Molnar 已提交
1159 1160 1161 1162 1163 1164 1165 1166
		return print_infinite_recursion_bug();

	debug_atomic_inc(&nr_find_usage_backwards_checks);
	if (source->usage_mask & (1 << find_usage_bit)) {
		backwards_match = source;
		return 2;
	}

D
Dave Jones 已提交
1167 1168 1169 1170 1171
	if (!source && debug_locks_off_graph_unlock()) {
		WARN_ON(1);
		return 0;
	}

I
Ingo Molnar 已提交
1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191
	/*
	 * Check this lock's dependency list:
	 */
	list_for_each_entry(entry, &source->locks_before, entry) {
		debug_atomic_inc(&nr_find_usage_backwards_recursions);
		ret = find_usage_backwards(entry->class, depth+1);
		if (ret == 2 || ret == 0)
			return ret;
	}
	return 1;
}

static int
print_bad_irq_dependency(struct task_struct *curr,
			 struct held_lock *prev,
			 struct held_lock *next,
			 enum lock_usage_bit bit1,
			 enum lock_usage_bit bit2,
			 const char *irqclass)
{
1192
	if (!debug_locks_off_graph_unlock() || debug_locks_silent)
I
Ingo Molnar 已提交
1193 1194 1195 1196 1197
		return 0;

	printk("\n======================================================\n");
	printk(  "[ INFO: %s-safe -> %s-unsafe lock order detected ]\n",
		irqclass, irqclass);
1198
	print_kernel_version();
I
Ingo Molnar 已提交
1199 1200
	printk(  "------------------------------------------------------\n");
	printk("%s/%d [HC%u[%lu]:SC%u[%lu]:HE%u:SE%u] is trying to acquire:\n",
1201
		curr->comm, task_pid_nr(curr),
I
Ingo Molnar 已提交
1202 1203 1204 1205 1206 1207 1208 1209 1210
		curr->hardirq_context, hardirq_count() >> HARDIRQ_SHIFT,
		curr->softirq_context, softirq_count() >> SOFTIRQ_SHIFT,
		curr->hardirqs_enabled,
		curr->softirqs_enabled);
	print_lock(next);

	printk("\nand this task is already holding:\n");
	print_lock(prev);
	printk("which would create a new lock dependency:\n");
D
Dave Jones 已提交
1211
	print_lock_name(hlock_class(prev));
I
Ingo Molnar 已提交
1212
	printk(" ->");
D
Dave Jones 已提交
1213
	print_lock_name(hlock_class(next));
I
Ingo Molnar 已提交
1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253
	printk("\n");

	printk("\nbut this new dependency connects a %s-irq-safe lock:\n",
		irqclass);
	print_lock_name(backwards_match);
	printk("\n... which became %s-irq-safe at:\n", irqclass);

	print_stack_trace(backwards_match->usage_traces + bit1, 1);

	printk("\nto a %s-irq-unsafe lock:\n", irqclass);
	print_lock_name(forwards_match);
	printk("\n... which became %s-irq-unsafe at:\n", irqclass);
	printk("...");

	print_stack_trace(forwards_match->usage_traces + bit2, 1);

	printk("\nother info that might help us debug this:\n\n");
	lockdep_print_held_locks(curr);

	printk("\nthe %s-irq-safe lock's dependencies:\n", irqclass);
	print_lock_dependencies(backwards_match, 0);

	printk("\nthe %s-irq-unsafe lock's dependencies:\n", irqclass);
	print_lock_dependencies(forwards_match, 0);

	printk("\nstack backtrace:\n");
	dump_stack();

	return 0;
}

static int
check_usage(struct task_struct *curr, struct held_lock *prev,
	    struct held_lock *next, enum lock_usage_bit bit_backwards,
	    enum lock_usage_bit bit_forwards, const char *irqclass)
{
	int ret;

	find_usage_bit = bit_backwards;
	/* fills in <backwards_match> */
D
Dave Jones 已提交
1254
	ret = find_usage_backwards(hlock_class(prev), 0);
I
Ingo Molnar 已提交
1255 1256 1257 1258
	if (!ret || ret == 1)
		return ret;

	find_usage_bit = bit_forwards;
D
Dave Jones 已提交
1259
	ret = find_usage_forwards(hlock_class(next), 0);
I
Ingo Molnar 已提交
1260 1261 1262 1263 1264 1265 1266
	if (!ret || ret == 1)
		return ret;
	/* ret == 2 */
	return print_bad_irq_dependency(curr, prev, next,
			bit_backwards, bit_forwards, irqclass);
}

1267 1268
static const char *state_names[] = {
#define LOCKDEP_STATE(__STATE) \
P
Peter Zijlstra 已提交
1269
	__stringify(__STATE),
1270 1271 1272 1273 1274 1275
#include "lockdep_states.h"
#undef LOCKDEP_STATE
};

static const char *state_rnames[] = {
#define LOCKDEP_STATE(__STATE) \
P
Peter Zijlstra 已提交
1276
	__stringify(__STATE)"-READ",
1277 1278 1279 1280 1281
#include "lockdep_states.h"
#undef LOCKDEP_STATE
};

static inline const char *state_name(enum lock_usage_bit bit)
P
Peter Zijlstra 已提交
1282
{
1283 1284
	return (bit & 1) ? state_rnames[bit >> 2] : state_names[bit >> 2];
}
P
Peter Zijlstra 已提交
1285

1286 1287
static int exclusive_bit(int new_bit)
{
P
Peter Zijlstra 已提交
1288
	/*
1289 1290 1291 1292 1293 1294 1295 1296
	 * USED_IN
	 * USED_IN_READ
	 * ENABLED
	 * ENABLED_READ
	 *
	 * bit 0 - write/read
	 * bit 1 - used_in/enabled
	 * bit 2+  state
P
Peter Zijlstra 已提交
1297
	 */
1298 1299 1300

	int state = new_bit & ~3;
	int dir = new_bit & 2;
P
Peter Zijlstra 已提交
1301 1302

	/*
1303
	 * keep state, bit flip the direction and strip read.
P
Peter Zijlstra 已提交
1304
	 */
1305 1306 1307 1308 1309 1310
	return state | (dir ^ 2);
}

static int check_irq_usage(struct task_struct *curr, struct held_lock *prev,
			   struct held_lock *next, enum lock_usage_bit bit)
{
P
Peter Zijlstra 已提交
1311
	/*
1312 1313
	 * Prove that the new dependency does not connect a hardirq-safe
	 * lock with a hardirq-unsafe lock - to achieve this we search
P
Peter Zijlstra 已提交
1314 1315 1316
	 * the backwards-subgraph starting at <prev>, and the
	 * forwards-subgraph starting at <next>:
	 */
1317 1318
	if (!check_usage(curr, prev, next, bit,
			   exclusive_bit(bit), state_name(bit)))
P
Peter Zijlstra 已提交
1319 1320
		return 0;

1321 1322
	bit++; /* _READ */

1323
	/*
1324 1325
	 * Prove that the new dependency does not connect a hardirq-safe-read
	 * lock with a hardirq-unsafe lock - to achieve this we search
1326 1327 1328
	 * the backwards-subgraph starting at <prev>, and the
	 * forwards-subgraph starting at <next>:
	 */
1329 1330
	if (!check_usage(curr, prev, next, bit,
			   exclusive_bit(bit), state_name(bit)))
1331 1332
		return 0;

1333 1334 1335 1336 1337 1338 1339 1340 1341
	return 1;
}

static int
check_prev_add_irq(struct task_struct *curr, struct held_lock *prev,
		struct held_lock *next)
{
#define LOCKDEP_STATE(__STATE)						\
	if (!check_irq_usage(curr, prev, next, LOCK_USED_IN_##__STATE))	\
1342
		return 0;
1343 1344
#include "lockdep_states.h"
#undef LOCKDEP_STATE
1345

P
Peter Zijlstra 已提交
1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374
	return 1;
}

static void inc_chains(void)
{
	if (current->hardirq_context)
		nr_hardirq_chains++;
	else {
		if (current->softirq_context)
			nr_softirq_chains++;
		else
			nr_process_chains++;
	}
}

#else

static inline int
check_prev_add_irq(struct task_struct *curr, struct held_lock *prev,
		struct held_lock *next)
{
	return 1;
}

static inline void inc_chains(void)
{
	nr_process_chains++;
}

I
Ingo Molnar 已提交
1375 1376 1377 1378 1379 1380
#endif

static int
print_deadlock_bug(struct task_struct *curr, struct held_lock *prev,
		   struct held_lock *next)
{
1381
	if (!debug_locks_off_graph_unlock() || debug_locks_silent)
I
Ingo Molnar 已提交
1382 1383 1384 1385
		return 0;

	printk("\n=============================================\n");
	printk(  "[ INFO: possible recursive locking detected ]\n");
1386
	print_kernel_version();
I
Ingo Molnar 已提交
1387 1388
	printk(  "---------------------------------------------\n");
	printk("%s/%d is trying to acquire lock:\n",
1389
		curr->comm, task_pid_nr(curr));
I
Ingo Molnar 已提交
1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415
	print_lock(next);
	printk("\nbut task is already holding lock:\n");
	print_lock(prev);

	printk("\nother info that might help us debug this:\n");
	lockdep_print_held_locks(curr);

	printk("\nstack backtrace:\n");
	dump_stack();

	return 0;
}

/*
 * Check whether we are holding such a class already.
 *
 * (Note that this has to be done separately, because the graph cannot
 * detect such classes of deadlocks.)
 *
 * Returns: 0 on deadlock detected, 1 on OK, 2 on recursive read
 */
static int
check_deadlock(struct task_struct *curr, struct held_lock *next,
	       struct lockdep_map *next_instance, int read)
{
	struct held_lock *prev;
P
Peter Zijlstra 已提交
1416
	struct held_lock *nest = NULL;
I
Ingo Molnar 已提交
1417 1418 1419 1420
	int i;

	for (i = 0; i < curr->lockdep_depth; i++) {
		prev = curr->held_locks + i;
P
Peter Zijlstra 已提交
1421 1422 1423 1424

		if (prev->instance == next->nest_lock)
			nest = prev;

D
Dave Jones 已提交
1425
		if (hlock_class(prev) != hlock_class(next))
I
Ingo Molnar 已提交
1426
			continue;
P
Peter Zijlstra 已提交
1427

I
Ingo Molnar 已提交
1428 1429
		/*
		 * Allow read-after-read recursion of the same
1430
		 * lock class (i.e. read_lock(lock)+read_lock(lock)):
I
Ingo Molnar 已提交
1431
		 */
1432
		if ((read == 2) && prev->read)
I
Ingo Molnar 已提交
1433
			return 2;
P
Peter Zijlstra 已提交
1434 1435 1436 1437 1438 1439 1440 1441

		/*
		 * We're holding the nest_lock, which serializes this lock's
		 * nesting behaviour.
		 */
		if (nest)
			return 2;

I
Ingo Molnar 已提交
1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470
		return print_deadlock_bug(curr, prev, next);
	}
	return 1;
}

/*
 * There was a chain-cache miss, and we are about to add a new dependency
 * to a previous lock. We recursively validate the following rules:
 *
 *  - would the adding of the <prev> -> <next> dependency create a
 *    circular dependency in the graph? [== circular deadlock]
 *
 *  - does the new prev->next dependency connect any hardirq-safe lock
 *    (in the full backwards-subgraph starting at <prev>) with any
 *    hardirq-unsafe lock (in the full forwards-subgraph starting at
 *    <next>)? [== illegal lock inversion with hardirq contexts]
 *
 *  - does the new prev->next dependency connect any softirq-safe lock
 *    (in the full backwards-subgraph starting at <prev>) with any
 *    softirq-unsafe lock (in the full forwards-subgraph starting at
 *    <next>)? [== illegal lock inversion with softirq contexts]
 *
 * any of these scenarios could lead to a deadlock.
 *
 * Then if all the validations pass, we add the forwards and backwards
 * dependency.
 */
static int
check_prev_add(struct task_struct *curr, struct held_lock *prev,
1471
	       struct held_lock *next, int distance)
I
Ingo Molnar 已提交
1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486
{
	struct lock_list *entry;
	int ret;

	/*
	 * Prove that the new <prev> -> <next> dependency would not
	 * create a circular dependency in the graph. (We do this by
	 * forward-recursing into the graph starting at <next>, and
	 * checking whether we can reach <prev>.)
	 *
	 * We are using global variables to control the recursion, to
	 * keep the stackframe size of the recursive functions low:
	 */
	check_source = next;
	check_target = prev;
D
Dave Jones 已提交
1487
	if (!(check_noncircular(hlock_class(next), 0)))
I
Ingo Molnar 已提交
1488 1489
		return print_circular_bug_tail();

P
Peter Zijlstra 已提交
1490
	if (!check_prev_add_irq(curr, prev, next))
I
Ingo Molnar 已提交
1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510
		return 0;

	/*
	 * For recursive read-locks we do all the dependency checks,
	 * but we dont store read-triggered dependencies (only
	 * write-triggered dependencies). This ensures that only the
	 * write-side dependencies matter, and that if for example a
	 * write-lock never takes any other locks, then the reads are
	 * equivalent to a NOP.
	 */
	if (next->read == 2 || prev->read == 2)
		return 1;
	/*
	 * Is the <prev> -> <next> dependency already present?
	 *
	 * (this may occur even though this is a new chain: consider
	 *  e.g. the L1 -> L2 -> L3 -> L4 and the L5 -> L1 -> L2 -> L3
	 *  chains - the second one will be new, but L1 already has
	 *  L2 added to its dependency list, due to the first chain.)
	 */
D
Dave Jones 已提交
1511 1512
	list_for_each_entry(entry, &hlock_class(prev)->locks_after, entry) {
		if (entry->class == hlock_class(next)) {
1513 1514
			if (distance == 1)
				entry->distance = 1;
I
Ingo Molnar 已提交
1515
			return 2;
1516
		}
I
Ingo Molnar 已提交
1517 1518 1519 1520 1521 1522
	}

	/*
	 * Ok, all validations passed, add the new lock
	 * to the previous lock's dependency list:
	 */
D
Dave Jones 已提交
1523 1524 1525
	ret = add_lock_to_list(hlock_class(prev), hlock_class(next),
			       &hlock_class(prev)->locks_after,
			       next->acquire_ip, distance);
1526

I
Ingo Molnar 已提交
1527 1528
	if (!ret)
		return 0;
1529

D
Dave Jones 已提交
1530 1531 1532
	ret = add_lock_to_list(hlock_class(next), hlock_class(prev),
			       &hlock_class(next)->locks_before,
			       next->acquire_ip, distance);
1533 1534
	if (!ret)
		return 0;
I
Ingo Molnar 已提交
1535 1536

	/*
P
Peter Zijlstra 已提交
1537 1538
	 * Debugging printouts:
	 */
D
Dave Jones 已提交
1539
	if (verbose(hlock_class(prev)) || verbose(hlock_class(next))) {
P
Peter Zijlstra 已提交
1540 1541
		graph_unlock();
		printk("\n new dependency: ");
D
Dave Jones 已提交
1542
		print_lock_name(hlock_class(prev));
P
Peter Zijlstra 已提交
1543
		printk(" => ");
D
Dave Jones 已提交
1544
		print_lock_name(hlock_class(next));
P
Peter Zijlstra 已提交
1545
		printk("\n");
I
Ingo Molnar 已提交
1546
		dump_stack();
P
Peter Zijlstra 已提交
1547
		return graph_lock();
I
Ingo Molnar 已提交
1548
	}
P
Peter Zijlstra 已提交
1549 1550
	return 1;
}
I
Ingo Molnar 已提交
1551

P
Peter Zijlstra 已提交
1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562
/*
 * Add the dependency to all directly-previous locks that are 'relevant'.
 * The ones that are relevant are (in increasing distance from curr):
 * all consecutive trylock entries and the final non-trylock entry - or
 * the end of this context's lock-chain - whichever comes first.
 */
static int
check_prevs_add(struct task_struct *curr, struct held_lock *next)
{
	int depth = curr->lockdep_depth;
	struct held_lock *hlock;
1563

I
Ingo Molnar 已提交
1564
	/*
P
Peter Zijlstra 已提交
1565 1566 1567
	 * Debugging checks.
	 *
	 * Depth must not be zero for a non-head lock:
I
Ingo Molnar 已提交
1568
	 */
P
Peter Zijlstra 已提交
1569 1570
	if (!depth)
		goto out_bug;
I
Ingo Molnar 已提交
1571
	/*
P
Peter Zijlstra 已提交
1572 1573
	 * At least two relevant locks must exist for this
	 * to be a head:
I
Ingo Molnar 已提交
1574
	 */
P
Peter Zijlstra 已提交
1575 1576 1577
	if (curr->held_locks[depth].irq_context !=
			curr->held_locks[depth-1].irq_context)
		goto out_bug;
1578

P
Peter Zijlstra 已提交
1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596
	for (;;) {
		int distance = curr->lockdep_depth - depth + 1;
		hlock = curr->held_locks + depth-1;
		/*
		 * Only non-recursive-read entries get new dependencies
		 * added:
		 */
		if (hlock->read != 2) {
			if (!check_prev_add(curr, hlock, next, distance))
				return 0;
			/*
			 * Stop after the first non-trylock entry,
			 * as non-trylock entries have added their
			 * own direct dependencies already, so this
			 * lock is connected to them indirectly:
			 */
			if (!hlock->trylock)
				break;
1597
		}
P
Peter Zijlstra 已提交
1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609
		depth--;
		/*
		 * End of lock-stack?
		 */
		if (!depth)
			break;
		/*
		 * Stop the search if we cross into another context:
		 */
		if (curr->held_locks[depth].irq_context !=
				curr->held_locks[depth-1].irq_context)
			break;
I
Ingo Molnar 已提交
1610
	}
P
Peter Zijlstra 已提交
1611 1612 1613 1614
	return 1;
out_bug:
	if (!debug_locks_off_graph_unlock())
		return 0;
I
Ingo Molnar 已提交
1615

P
Peter Zijlstra 已提交
1616
	WARN_ON(1);
I
Ingo Molnar 已提交
1617

P
Peter Zijlstra 已提交
1618
	return 0;
I
Ingo Molnar 已提交
1619 1620
}

P
Peter Zijlstra 已提交
1621
unsigned long nr_lock_chains;
1622
struct lock_chain lock_chains[MAX_LOCKDEP_CHAINS];
1623
int nr_chain_hlocks;
1624 1625 1626 1627 1628 1629
static u16 chain_hlocks[MAX_LOCKDEP_CHAIN_HLOCKS];

struct lock_class *lock_chain_get_class(struct lock_chain *chain, int i)
{
	return lock_classes + chain_hlocks[chain->base + i];
}
P
Peter Zijlstra 已提交
1630

I
Ingo Molnar 已提交
1631 1632
/*
 * Look up a dependency chain. If the key is not present yet then
1633 1634 1635
 * add it and return 1 - in this case the new dependency chain is
 * validated. If the key is already hashed, return 0.
 * (On return with 1 graph_lock is held.)
I
Ingo Molnar 已提交
1636
 */
1637 1638 1639
static inline int lookup_chain_cache(struct task_struct *curr,
				     struct held_lock *hlock,
				     u64 chain_key)
I
Ingo Molnar 已提交
1640
{
D
Dave Jones 已提交
1641
	struct lock_class *class = hlock_class(hlock);
I
Ingo Molnar 已提交
1642 1643
	struct list_head *hash_head = chainhashentry(chain_key);
	struct lock_chain *chain;
1644
	struct held_lock *hlock_curr, *hlock_next;
1645
	int i, j, n, cn;
I
Ingo Molnar 已提交
1646

1647 1648
	if (DEBUG_LOCKS_WARN_ON(!irqs_disabled()))
		return 0;
I
Ingo Molnar 已提交
1649 1650 1651 1652 1653 1654 1655 1656
	/*
	 * We can walk it lock-free, because entries only get added
	 * to the hash:
	 */
	list_for_each_entry(chain, hash_head, entry) {
		if (chain->chain_key == chain_key) {
cache_hit:
			debug_atomic_inc(&chain_lookup_hits);
1657
			if (very_verbose(class))
1658 1659 1660 1661
				printk("\nhash chain already cached, key: "
					"%016Lx tail class: [%p] %s\n",
					(unsigned long long)chain_key,
					class->key, class->name);
I
Ingo Molnar 已提交
1662 1663 1664
			return 0;
		}
	}
1665
	if (very_verbose(class))
1666 1667
		printk("\nnew hash chain, key: %016Lx tail class: [%p] %s\n",
			(unsigned long long)chain_key, class->key, class->name);
I
Ingo Molnar 已提交
1668 1669 1670 1671
	/*
	 * Allocate a new chain entry from the static array, and add
	 * it to the hash:
	 */
1672 1673
	if (!graph_lock())
		return 0;
I
Ingo Molnar 已提交
1674 1675 1676 1677 1678
	/*
	 * We have to walk the chain again locked - to avoid duplicates:
	 */
	list_for_each_entry(chain, hash_head, entry) {
		if (chain->chain_key == chain_key) {
1679
			graph_unlock();
I
Ingo Molnar 已提交
1680 1681 1682 1683
			goto cache_hit;
		}
	}
	if (unlikely(nr_lock_chains >= MAX_LOCKDEP_CHAINS)) {
1684 1685 1686
		if (!debug_locks_off_graph_unlock())
			return 0;

I
Ingo Molnar 已提交
1687 1688
		printk("BUG: MAX_LOCKDEP_CHAINS too low!\n");
		printk("turning off the locking correctness validator.\n");
1689
		dump_stack();
I
Ingo Molnar 已提交
1690 1691 1692 1693
		return 0;
	}
	chain = lock_chains + nr_lock_chains++;
	chain->chain_key = chain_key;
1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704
	chain->irq_context = hlock->irq_context;
	/* Find the first held_lock of current chain */
	hlock_next = hlock;
	for (i = curr->lockdep_depth - 1; i >= 0; i--) {
		hlock_curr = curr->held_locks + i;
		if (hlock_curr->irq_context != hlock_next->irq_context)
			break;
		hlock_next = hlock;
	}
	i++;
	chain->depth = curr->lockdep_depth + 1 - i;
1705 1706 1707 1708 1709 1710 1711 1712 1713
	cn = nr_chain_hlocks;
	while (cn + chain->depth <= MAX_LOCKDEP_CHAIN_HLOCKS) {
		n = cmpxchg(&nr_chain_hlocks, cn, cn + chain->depth);
		if (n == cn)
			break;
		cn = n;
	}
	if (likely(cn + chain->depth <= MAX_LOCKDEP_CHAIN_HLOCKS)) {
		chain->base = cn;
1714
		for (j = 0; j < chain->depth - 1; j++, i++) {
D
Dave Jones 已提交
1715
			int lock_id = curr->held_locks[i].class_idx - 1;
1716 1717 1718 1719
			chain_hlocks[chain->base + j] = lock_id;
		}
		chain_hlocks[chain->base + j] = class - lock_classes;
	}
I
Ingo Molnar 已提交
1720 1721
	list_add_tail_rcu(&chain->entry, hash_head);
	debug_atomic_inc(&chain_lookup_misses);
P
Peter Zijlstra 已提交
1722 1723 1724 1725 1726 1727
	inc_chains();

	return 1;
}

static int validate_chain(struct task_struct *curr, struct lockdep_map *lock,
1728
		struct held_lock *hlock, int chain_head, u64 chain_key)
P
Peter Zijlstra 已提交
1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740
{
	/*
	 * Trylock needs to maintain the stack of held locks, but it
	 * does not add new dependencies, because trylock can be done
	 * in any order.
	 *
	 * We look up the chain_key and do the O(N^2) check and update of
	 * the dependencies only if this is a new dependency chain.
	 * (If lookup_chain_cache() returns with 1 it acquires
	 * graph_lock for us)
	 */
	if (!hlock->trylock && (hlock->check == 2) &&
1741
	    lookup_chain_cache(curr, hlock, chain_key)) {
P
Peter Zijlstra 已提交
1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776
		/*
		 * Check whether last held lock:
		 *
		 * - is irq-safe, if this lock is irq-unsafe
		 * - is softirq-safe, if this lock is hardirq-unsafe
		 *
		 * And check whether the new lock's dependency graph
		 * could lead back to the previous lock.
		 *
		 * any of these scenarios could lead to a deadlock. If
		 * All validations
		 */
		int ret = check_deadlock(curr, hlock, lock, hlock->read);

		if (!ret)
			return 0;
		/*
		 * Mark recursive read, as we jump over it when
		 * building dependencies (just like we jump over
		 * trylock entries):
		 */
		if (ret == 2)
			hlock->read = 2;
		/*
		 * Add dependency only if this lock is not the head
		 * of the chain, and if it's not a secondary read-lock:
		 */
		if (!chain_head && ret != 2)
			if (!check_prevs_add(curr, hlock))
				return 0;
		graph_unlock();
	} else
		/* after lookup_chain_cache(): */
		if (unlikely(!debug_locks))
			return 0;
I
Ingo Molnar 已提交
1777 1778 1779

	return 1;
}
P
Peter Zijlstra 已提交
1780 1781 1782
#else
static inline int validate_chain(struct task_struct *curr,
	       	struct lockdep_map *lock, struct held_lock *hlock,
1783
		int chain_head, u64 chain_key)
P
Peter Zijlstra 已提交
1784 1785 1786
{
	return 1;
}
1787
#endif
I
Ingo Molnar 已提交
1788 1789 1790 1791 1792

/*
 * We are building curr_chain_key incrementally, so double-check
 * it from scratch, to make sure that it's done correctly:
 */
1793
static void check_chain_key(struct task_struct *curr)
I
Ingo Molnar 已提交
1794 1795 1796 1797 1798 1799 1800 1801 1802 1803
{
#ifdef CONFIG_DEBUG_LOCKDEP
	struct held_lock *hlock, *prev_hlock = NULL;
	unsigned int i, id;
	u64 chain_key = 0;

	for (i = 0; i < curr->lockdep_depth; i++) {
		hlock = curr->held_locks + i;
		if (chain_key != hlock->prev_chain_key) {
			debug_locks_off();
1804
			WARN(1, "hm#1, depth: %u [%u], %016Lx != %016Lx\n",
I
Ingo Molnar 已提交
1805 1806 1807 1808 1809
				curr->lockdep_depth, i,
				(unsigned long long)chain_key,
				(unsigned long long)hlock->prev_chain_key);
			return;
		}
D
Dave Jones 已提交
1810
		id = hlock->class_idx - 1;
1811 1812 1813
		if (DEBUG_LOCKS_WARN_ON(id >= MAX_LOCKDEP_KEYS))
			return;

I
Ingo Molnar 已提交
1814 1815 1816 1817 1818 1819 1820 1821
		if (prev_hlock && (prev_hlock->irq_context !=
							hlock->irq_context))
			chain_key = 0;
		chain_key = iterate_chain_key(chain_key, id);
		prev_hlock = hlock;
	}
	if (chain_key != curr->curr_chain_key) {
		debug_locks_off();
1822
		WARN(1, "hm#2, depth: %u [%u], %016Lx != %016Lx\n",
I
Ingo Molnar 已提交
1823 1824 1825 1826 1827 1828 1829
			curr->lockdep_depth, i,
			(unsigned long long)chain_key,
			(unsigned long long)curr->curr_chain_key);
	}
#endif
}

P
Peter Zijlstra 已提交
1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845
static int
print_usage_bug(struct task_struct *curr, struct held_lock *this,
		enum lock_usage_bit prev_bit, enum lock_usage_bit new_bit)
{
	if (!debug_locks_off_graph_unlock() || debug_locks_silent)
		return 0;

	printk("\n=================================\n");
	printk(  "[ INFO: inconsistent lock state ]\n");
	print_kernel_version();
	printk(  "---------------------------------\n");

	printk("inconsistent {%s} -> {%s} usage.\n",
		usage_str[prev_bit], usage_str[new_bit]);

	printk("%s/%d [HC%u[%lu]:SC%u[%lu]:HE%u:SE%u] takes:\n",
1846
		curr->comm, task_pid_nr(curr),
P
Peter Zijlstra 已提交
1847 1848 1849 1850 1851 1852 1853
		trace_hardirq_context(curr), hardirq_count() >> HARDIRQ_SHIFT,
		trace_softirq_context(curr), softirq_count() >> SOFTIRQ_SHIFT,
		trace_hardirqs_enabled(curr),
		trace_softirqs_enabled(curr));
	print_lock(this);

	printk("{%s} state was registered at:\n", usage_str[prev_bit]);
D
Dave Jones 已提交
1854
	print_stack_trace(hlock_class(this)->usage_traces + prev_bit, 1);
P
Peter Zijlstra 已提交
1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872

	print_irqtrace_events(curr);
	printk("\nother info that might help us debug this:\n");
	lockdep_print_held_locks(curr);

	printk("\nstack backtrace:\n");
	dump_stack();

	return 0;
}

/*
 * Print out an error if an invalid bit is set:
 */
static inline int
valid_state(struct task_struct *curr, struct held_lock *this,
	    enum lock_usage_bit new_bit, enum lock_usage_bit bad_bit)
{
D
Dave Jones 已提交
1873
	if (unlikely(hlock_class(this)->usage_mask & (1 << bad_bit)))
P
Peter Zijlstra 已提交
1874 1875 1876 1877 1878 1879 1880
		return print_usage_bug(curr, this, bad_bit, new_bit);
	return 1;
}

static int mark_lock(struct task_struct *curr, struct held_lock *this,
		     enum lock_usage_bit new_bit);

1881
#if defined(CONFIG_TRACE_IRQFLAGS) && defined(CONFIG_PROVE_LOCKING)
I
Ingo Molnar 已提交
1882 1883 1884 1885 1886 1887 1888 1889 1890

/*
 * print irq inversion bug:
 */
static int
print_irq_inversion_bug(struct task_struct *curr, struct lock_class *other,
			struct held_lock *this, int forwards,
			const char *irqclass)
{
1891
	if (!debug_locks_off_graph_unlock() || debug_locks_silent)
I
Ingo Molnar 已提交
1892 1893 1894 1895
		return 0;

	printk("\n=========================================================\n");
	printk(  "[ INFO: possible irq lock inversion dependency detected ]\n");
1896
	print_kernel_version();
I
Ingo Molnar 已提交
1897 1898
	printk(  "---------------------------------------------------------\n");
	printk("%s/%d just changed the state of lock:\n",
1899
		curr->comm, task_pid_nr(curr));
I
Ingo Molnar 已提交
1900 1901
	print_lock(this);
	if (forwards)
1902
		printk("but this lock took another, %s-unsafe lock in the past:\n", irqclass);
I
Ingo Molnar 已提交
1903
	else
1904
		printk("but this lock was taken by another, %s-safe lock in the past:\n", irqclass);
I
Ingo Molnar 已提交
1905 1906 1907 1908 1909 1910 1911
	print_lock_name(other);
	printk("\n\nand interrupts could create inverse lock ordering between them.\n\n");

	printk("\nother info that might help us debug this:\n");
	lockdep_print_held_locks(curr);

	printk("\nthe first lock's dependencies:\n");
D
Dave Jones 已提交
1912
	print_lock_dependencies(hlock_class(this), 0);
I
Ingo Molnar 已提交
1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934

	printk("\nthe second lock's dependencies:\n");
	print_lock_dependencies(other, 0);

	printk("\nstack backtrace:\n");
	dump_stack();

	return 0;
}

/*
 * Prove that in the forwards-direction subgraph starting at <this>
 * there is no lock matching <mask>:
 */
static int
check_usage_forwards(struct task_struct *curr, struct held_lock *this,
		     enum lock_usage_bit bit, const char *irqclass)
{
	int ret;

	find_usage_bit = bit;
	/* fills in <forwards_match> */
D
Dave Jones 已提交
1935
	ret = find_usage_forwards(hlock_class(this), 0);
I
Ingo Molnar 已提交
1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953
	if (!ret || ret == 1)
		return ret;

	return print_irq_inversion_bug(curr, forwards_match, this, 1, irqclass);
}

/*
 * Prove that in the backwards-direction subgraph starting at <this>
 * there is no lock matching <mask>:
 */
static int
check_usage_backwards(struct task_struct *curr, struct held_lock *this,
		      enum lock_usage_bit bit, const char *irqclass)
{
	int ret;

	find_usage_bit = bit;
	/* fills in <backwards_match> */
D
Dave Jones 已提交
1954
	ret = find_usage_backwards(hlock_class(this), 0);
I
Ingo Molnar 已提交
1955 1956 1957 1958 1959 1960
	if (!ret || ret == 1)
		return ret;

	return print_irq_inversion_bug(curr, backwards_match, this, 0, irqclass);
}

1961
void print_irqtrace_events(struct task_struct *curr)
I
Ingo Molnar 已提交
1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973
{
	printk("irq event stamp: %u\n", curr->irq_events);
	printk("hardirqs last  enabled at (%u): ", curr->hardirq_enable_event);
	print_ip_sym(curr->hardirq_enable_ip);
	printk("hardirqs last disabled at (%u): ", curr->hardirq_disable_event);
	print_ip_sym(curr->hardirq_disable_ip);
	printk("softirqs last  enabled at (%u): ", curr->softirq_enable_event);
	print_ip_sym(curr->softirq_enable_ip);
	printk("softirqs last disabled at (%u): ", curr->softirq_disable_event);
	print_ip_sym(curr->softirq_disable_ip);
}

1974
static int HARDIRQ_verbose(struct lock_class *class)
I
Ingo Molnar 已提交
1975
{
P
Peter Zijlstra 已提交
1976 1977 1978
#if HARDIRQ_VERBOSE
	return class_filter(class);
#endif
I
Ingo Molnar 已提交
1979 1980 1981
	return 0;
}

1982
static int SOFTIRQ_verbose(struct lock_class *class)
I
Ingo Molnar 已提交
1983
{
P
Peter Zijlstra 已提交
1984 1985 1986 1987
#if SOFTIRQ_VERBOSE
	return class_filter(class);
#endif
	return 0;
I
Ingo Molnar 已提交
1988 1989
}

1990
static int RECLAIM_FS_verbose(struct lock_class *class)
1991 1992 1993 1994 1995 1996 1997
{
#if RECLAIM_VERBOSE
	return class_filter(class);
#endif
	return 0;
}

I
Ingo Molnar 已提交
1998 1999
#define STRICT_READ_CHECKS	1

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
static int (*state_verbose_f[])(struct lock_class *class) = {
#define LOCKDEP_STATE(__STATE) \
	__STATE##_verbose,
#include "lockdep_states.h"
#undef LOCKDEP_STATE
};

static inline int state_verbose(enum lock_usage_bit bit,
				struct lock_class *class)
{
	return state_verbose_f[bit >> 2](class);
}

2013 2014 2015
typedef int (*check_usage_f)(struct task_struct *, struct held_lock *,
			     enum lock_usage_bit bit, const char *name);

2016
static int
2017 2018
mark_lock_irq(struct task_struct *curr, struct held_lock *this,
		enum lock_usage_bit new_bit)
2019
{
2020
	int excl_bit = exclusive_bit(new_bit);
2021
	int read = new_bit & 1;
2022 2023
	int dir = new_bit & 2;

2024 2025 2026 2027 2028 2029 2030
	/*
	 * mark USED_IN has to look forwards -- to ensure no dependency
	 * has ENABLED state, which would allow recursion deadlocks.
	 *
	 * mark ENABLED has to look backwards -- to ensure no dependee
	 * has USED_IN state, which, again, would allow  recursion deadlocks.
	 */
2031 2032
	check_usage_f usage = dir ?
		check_usage_backwards : check_usage_forwards;
2033

2034 2035 2036 2037
	/*
	 * Validate that this particular lock does not have conflicting
	 * usage states.
	 */
2038 2039
	if (!valid_state(curr, this, new_bit, excl_bit))
		return 0;
2040

2041 2042 2043 2044 2045
	/*
	 * Validate that the lock dependencies don't have conflicting usage
	 * states.
	 */
	if ((!read || !dir || STRICT_READ_CHECKS) &&
2046
			!usage(curr, this, excl_bit, state_name(new_bit & ~1)))
2047
		return 0;
2048

2049 2050 2051 2052 2053 2054 2055 2056
	/*
	 * Check for read in write conflicts
	 */
	if (!read) {
		if (!valid_state(curr, this, new_bit, excl_bit + 1))
			return 0;

		if (STRICT_READ_CHECKS &&
2057 2058
			!usage(curr, this, excl_bit + 1,
				state_name(new_bit + 1)))
2059 2060
			return 0;
	}
2061

2062
	if (state_verbose(new_bit, hlock_class(this)))
2063 2064 2065 2066 2067
		return 2;

	return 1;
}

2068
enum mark_type {
2069 2070 2071
#define LOCKDEP_STATE(__STATE)	__STATE,
#include "lockdep_states.h"
#undef LOCKDEP_STATE
2072 2073
};

I
Ingo Molnar 已提交
2074 2075 2076
/*
 * Mark all held locks with a usage bit:
 */
2077
static int
2078
mark_held_locks(struct task_struct *curr, enum mark_type mark)
I
Ingo Molnar 已提交
2079 2080 2081 2082 2083 2084 2085 2086
{
	enum lock_usage_bit usage_bit;
	struct held_lock *hlock;
	int i;

	for (i = 0; i < curr->lockdep_depth; i++) {
		hlock = curr->held_locks + i;

2087 2088 2089 2090 2091
		usage_bit = 2 + (mark << 2); /* ENABLED */
		if (hlock->read)
			usage_bit += 1; /* READ */

		BUG_ON(usage_bit >= LOCK_USAGE_STATES);
2092

2093
		if (!mark_lock(curr, hlock, usage_bit))
I
Ingo Molnar 已提交
2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118
			return 0;
	}

	return 1;
}

/*
 * Debugging helper: via this flag we know that we are in
 * 'early bootup code', and will warn about any invalid irqs-on event:
 */
static int early_boot_irqs_enabled;

void early_boot_irqs_off(void)
{
	early_boot_irqs_enabled = 0;
}

void early_boot_irqs_on(void)
{
	early_boot_irqs_enabled = 1;
}

/*
 * Hardirqs will be enabled:
 */
2119
void trace_hardirqs_on_caller(unsigned long ip)
I
Ingo Molnar 已提交
2120 2121 2122
{
	struct task_struct *curr = current;

2123
	time_hardirqs_on(CALLER_ADDR0, ip);
2124

I
Ingo Molnar 已提交
2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145
	if (unlikely(!debug_locks || current->lockdep_recursion))
		return;

	if (DEBUG_LOCKS_WARN_ON(unlikely(!early_boot_irqs_enabled)))
		return;

	if (unlikely(curr->hardirqs_enabled)) {
		debug_atomic_inc(&redundant_hardirqs_on);
		return;
	}
	/* we'll do an OFF -> ON transition: */
	curr->hardirqs_enabled = 1;

	if (DEBUG_LOCKS_WARN_ON(!irqs_disabled()))
		return;
	if (DEBUG_LOCKS_WARN_ON(current->hardirq_context))
		return;
	/*
	 * We are going to turn hardirqs on, so set the
	 * usage bit for all held locks:
	 */
2146
	if (!mark_held_locks(curr, HARDIRQ))
I
Ingo Molnar 已提交
2147 2148 2149 2150 2151 2152 2153
		return;
	/*
	 * If we have softirqs enabled, then set the usage
	 * bit for all held locks. (disabled hardirqs prevented
	 * this bit from being set before)
	 */
	if (curr->softirqs_enabled)
2154
		if (!mark_held_locks(curr, SOFTIRQ))
I
Ingo Molnar 已提交
2155 2156
			return;

P
Peter Zijlstra 已提交
2157 2158 2159 2160
	curr->hardirq_enable_ip = ip;
	curr->hardirq_enable_event = ++curr->irq_events;
	debug_atomic_inc(&hardirqs_on_events);
}
2161
EXPORT_SYMBOL(trace_hardirqs_on_caller);
P
Peter Zijlstra 已提交
2162

2163
void trace_hardirqs_on(void)
2164 2165 2166
{
	trace_hardirqs_on_caller(CALLER_ADDR0);
}
P
Peter Zijlstra 已提交
2167 2168 2169 2170 2171
EXPORT_SYMBOL(trace_hardirqs_on);

/*
 * Hardirqs were disabled:
 */
2172
void trace_hardirqs_off_caller(unsigned long ip)
P
Peter Zijlstra 已提交
2173 2174 2175
{
	struct task_struct *curr = current;

2176
	time_hardirqs_off(CALLER_ADDR0, ip);
2177

P
Peter Zijlstra 已提交
2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188
	if (unlikely(!debug_locks || current->lockdep_recursion))
		return;

	if (DEBUG_LOCKS_WARN_ON(!irqs_disabled()))
		return;

	if (curr->hardirqs_enabled) {
		/*
		 * We have done an ON -> OFF transition:
		 */
		curr->hardirqs_enabled = 0;
2189
		curr->hardirq_disable_ip = ip;
P
Peter Zijlstra 已提交
2190 2191 2192 2193 2194
		curr->hardirq_disable_event = ++curr->irq_events;
		debug_atomic_inc(&hardirqs_off_events);
	} else
		debug_atomic_inc(&redundant_hardirqs_off);
}
2195
EXPORT_SYMBOL(trace_hardirqs_off_caller);
P
Peter Zijlstra 已提交
2196

2197
void trace_hardirqs_off(void)
2198 2199 2200
{
	trace_hardirqs_off_caller(CALLER_ADDR0);
}
P
Peter Zijlstra 已提交
2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233
EXPORT_SYMBOL(trace_hardirqs_off);

/*
 * Softirqs will be enabled:
 */
void trace_softirqs_on(unsigned long ip)
{
	struct task_struct *curr = current;

	if (unlikely(!debug_locks))
		return;

	if (DEBUG_LOCKS_WARN_ON(!irqs_disabled()))
		return;

	if (curr->softirqs_enabled) {
		debug_atomic_inc(&redundant_softirqs_on);
		return;
	}

	/*
	 * We'll do an OFF -> ON transition:
	 */
	curr->softirqs_enabled = 1;
	curr->softirq_enable_ip = ip;
	curr->softirq_enable_event = ++curr->irq_events;
	debug_atomic_inc(&softirqs_on_events);
	/*
	 * We are going to turn softirqs on, so set the
	 * usage bit for all held locks, if hardirqs are
	 * enabled too:
	 */
	if (curr->hardirqs_enabled)
2234
		mark_held_locks(curr, SOFTIRQ);
P
Peter Zijlstra 已提交
2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262
}

/*
 * Softirqs were disabled:
 */
void trace_softirqs_off(unsigned long ip)
{
	struct task_struct *curr = current;

	if (unlikely(!debug_locks))
		return;

	if (DEBUG_LOCKS_WARN_ON(!irqs_disabled()))
		return;

	if (curr->softirqs_enabled) {
		/*
		 * We have done an ON -> OFF transition:
		 */
		curr->softirqs_enabled = 0;
		curr->softirq_disable_ip = ip;
		curr->softirq_disable_event = ++curr->irq_events;
		debug_atomic_inc(&softirqs_off_events);
		DEBUG_LOCKS_WARN_ON(!softirq_count());
	} else
		debug_atomic_inc(&redundant_softirqs_off);
}

2263
static void __lockdep_trace_alloc(gfp_t gfp_mask, unsigned long flags)
2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281
{
	struct task_struct *curr = current;

	if (unlikely(!debug_locks))
		return;

	/* no reclaim without waiting on it */
	if (!(gfp_mask & __GFP_WAIT))
		return;

	/* this guy won't enter reclaim */
	if ((curr->flags & PF_MEMALLOC) && !(gfp_mask & __GFP_NOMEMALLOC))
		return;

	/* We're only interested __GFP_FS allocations for now */
	if (!(gfp_mask & __GFP_FS))
		return;

2282
	if (DEBUG_LOCKS_WARN_ON(irqs_disabled_flags(flags)))
2283 2284 2285 2286 2287
		return;

	mark_held_locks(curr, RECLAIM_FS);
}

2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304
static void check_flags(unsigned long flags);

void lockdep_trace_alloc(gfp_t gfp_mask)
{
	unsigned long flags;

	if (unlikely(current->lockdep_recursion))
		return;

	raw_local_irq_save(flags);
	check_flags(flags);
	current->lockdep_recursion = 1;
	__lockdep_trace_alloc(gfp_mask, flags);
	current->lockdep_recursion = 0;
	raw_local_irq_restore(flags);
}

P
Peter Zijlstra 已提交
2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332
static int mark_irqflags(struct task_struct *curr, struct held_lock *hlock)
{
	/*
	 * If non-trylock use in a hardirq or softirq context, then
	 * mark the lock as used in these contexts:
	 */
	if (!hlock->trylock) {
		if (hlock->read) {
			if (curr->hardirq_context)
				if (!mark_lock(curr, hlock,
						LOCK_USED_IN_HARDIRQ_READ))
					return 0;
			if (curr->softirq_context)
				if (!mark_lock(curr, hlock,
						LOCK_USED_IN_SOFTIRQ_READ))
					return 0;
		} else {
			if (curr->hardirq_context)
				if (!mark_lock(curr, hlock, LOCK_USED_IN_HARDIRQ))
					return 0;
			if (curr->softirq_context)
				if (!mark_lock(curr, hlock, LOCK_USED_IN_SOFTIRQ))
					return 0;
		}
	}
	if (!hlock->hardirqs_off) {
		if (hlock->read) {
			if (!mark_lock(curr, hlock,
P
Peter Zijlstra 已提交
2333
					LOCK_ENABLED_HARDIRQ_READ))
P
Peter Zijlstra 已提交
2334 2335 2336
				return 0;
			if (curr->softirqs_enabled)
				if (!mark_lock(curr, hlock,
P
Peter Zijlstra 已提交
2337
						LOCK_ENABLED_SOFTIRQ_READ))
P
Peter Zijlstra 已提交
2338 2339 2340
					return 0;
		} else {
			if (!mark_lock(curr, hlock,
P
Peter Zijlstra 已提交
2341
					LOCK_ENABLED_HARDIRQ))
P
Peter Zijlstra 已提交
2342 2343 2344
				return 0;
			if (curr->softirqs_enabled)
				if (!mark_lock(curr, hlock,
P
Peter Zijlstra 已提交
2345
						LOCK_ENABLED_SOFTIRQ))
P
Peter Zijlstra 已提交
2346 2347 2348 2349
					return 0;
		}
	}

2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365
	/*
	 * We reuse the irq context infrastructure more broadly as a general
	 * context checking code. This tests GFP_FS recursion (a lock taken
	 * during reclaim for a GFP_FS allocation is held over a GFP_FS
	 * allocation).
	 */
	if (!hlock->trylock && (curr->lockdep_reclaim_gfp & __GFP_FS)) {
		if (hlock->read) {
			if (!mark_lock(curr, hlock, LOCK_USED_IN_RECLAIM_FS_READ))
					return 0;
		} else {
			if (!mark_lock(curr, hlock, LOCK_USED_IN_RECLAIM_FS))
					return 0;
		}
	}

P
Peter Zijlstra 已提交
2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391
	return 1;
}

static int separate_irq_context(struct task_struct *curr,
		struct held_lock *hlock)
{
	unsigned int depth = curr->lockdep_depth;

	/*
	 * Keep track of points where we cross into an interrupt context:
	 */
	hlock->irq_context = 2*(curr->hardirq_context ? 1 : 0) +
				curr->softirq_context;
	if (depth) {
		struct held_lock *prev_hlock;

		prev_hlock = curr->held_locks + depth-1;
		/*
		 * If we cross into another context, reset the
		 * hash key (this also prevents the checking and the
		 * adding of the dependency to 'prev'):
		 */
		if (prev_hlock->irq_context != hlock->irq_context)
			return 1;
	}
	return 0;
I
Ingo Molnar 已提交
2392 2393
}

P
Peter Zijlstra 已提交
2394
#else
I
Ingo Molnar 已提交
2395

P
Peter Zijlstra 已提交
2396 2397 2398
static inline
int mark_lock_irq(struct task_struct *curr, struct held_lock *this,
		enum lock_usage_bit new_bit)
I
Ingo Molnar 已提交
2399
{
P
Peter Zijlstra 已提交
2400 2401 2402
	WARN_ON(1);
	return 1;
}
I
Ingo Molnar 已提交
2403

P
Peter Zijlstra 已提交
2404 2405 2406 2407 2408
static inline int mark_irqflags(struct task_struct *curr,
		struct held_lock *hlock)
{
	return 1;
}
I
Ingo Molnar 已提交
2409

P
Peter Zijlstra 已提交
2410 2411 2412 2413
static inline int separate_irq_context(struct task_struct *curr,
		struct held_lock *hlock)
{
	return 0;
I
Ingo Molnar 已提交
2414 2415
}

2416 2417 2418 2419
void lockdep_trace_alloc(gfp_t gfp_mask)
{
}

P
Peter Zijlstra 已提交
2420
#endif
I
Ingo Molnar 已提交
2421 2422

/*
P
Peter Zijlstra 已提交
2423
 * Mark a lock with a usage bit, and validate the state transition:
I
Ingo Molnar 已提交
2424
 */
2425
static int mark_lock(struct task_struct *curr, struct held_lock *this,
2426
			     enum lock_usage_bit new_bit)
I
Ingo Molnar 已提交
2427
{
P
Peter Zijlstra 已提交
2428
	unsigned int new_mask = 1 << new_bit, ret = 1;
I
Ingo Molnar 已提交
2429 2430

	/*
P
Peter Zijlstra 已提交
2431 2432
	 * If already set then do not dirty the cacheline,
	 * nor do any checks:
I
Ingo Molnar 已提交
2433
	 */
D
Dave Jones 已提交
2434
	if (likely(hlock_class(this)->usage_mask & new_mask))
P
Peter Zijlstra 已提交
2435 2436 2437 2438
		return 1;

	if (!graph_lock())
		return 0;
I
Ingo Molnar 已提交
2439
	/*
P
Peter Zijlstra 已提交
2440
	 * Make sure we didnt race:
I
Ingo Molnar 已提交
2441
	 */
D
Dave Jones 已提交
2442
	if (unlikely(hlock_class(this)->usage_mask & new_mask)) {
P
Peter Zijlstra 已提交
2443 2444 2445
		graph_unlock();
		return 1;
	}
I
Ingo Molnar 已提交
2446

D
Dave Jones 已提交
2447
	hlock_class(this)->usage_mask |= new_mask;
I
Ingo Molnar 已提交
2448

D
Dave Jones 已提交
2449
	if (!save_trace(hlock_class(this)->usage_traces + new_bit))
P
Peter Zijlstra 已提交
2450
		return 0;
I
Ingo Molnar 已提交
2451

P
Peter Zijlstra 已提交
2452
	switch (new_bit) {
P
Peter Zijlstra 已提交
2453 2454 2455 2456 2457 2458 2459
#define LOCKDEP_STATE(__STATE)			\
	case LOCK_USED_IN_##__STATE:		\
	case LOCK_USED_IN_##__STATE##_READ:	\
	case LOCK_ENABLED_##__STATE:		\
	case LOCK_ENABLED_##__STATE##_READ:
#include "lockdep_states.h"
#undef LOCKDEP_STATE
P
Peter Zijlstra 已提交
2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472
		ret = mark_lock_irq(curr, this, new_bit);
		if (!ret)
			return 0;
		break;
	case LOCK_USED:
		debug_atomic_dec(&nr_unused_locks);
		break;
	default:
		if (!debug_locks_off_graph_unlock())
			return 0;
		WARN_ON(1);
		return 0;
	}
I
Ingo Molnar 已提交
2473

P
Peter Zijlstra 已提交
2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487
	graph_unlock();

	/*
	 * We must printk outside of the graph_lock:
	 */
	if (ret == 2) {
		printk("\nmarked lock as {%s}:\n", usage_str[new_bit]);
		print_lock(this);
		print_irqtrace_events(curr);
		dump_stack();
	}

	return ret;
}
I
Ingo Molnar 已提交
2488 2489 2490 2491 2492

/*
 * Initialize a lock instance's lock-class mapping info:
 */
void lockdep_init_map(struct lockdep_map *lock, const char *name,
2493
		      struct lock_class_key *key, int subclass)
I
Ingo Molnar 已提交
2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511
{
	if (unlikely(!debug_locks))
		return;

	if (DEBUG_LOCKS_WARN_ON(!key))
		return;
	if (DEBUG_LOCKS_WARN_ON(!name))
		return;
	/*
	 * Sanity check, the lock-class key must be persistent:
	 */
	if (!static_obj(key)) {
		printk("BUG: key %p not in .data!\n", key);
		DEBUG_LOCKS_WARN_ON(1);
		return;
	}
	lock->name = name;
	lock->key = key;
2512
	lock->class_cache = NULL;
P
Peter Zijlstra 已提交
2513 2514 2515
#ifdef CONFIG_LOCK_STAT
	lock->cpu = raw_smp_processor_id();
#endif
2516 2517
	if (subclass)
		register_lock_class(lock, subclass, 1);
I
Ingo Molnar 已提交
2518 2519 2520 2521 2522 2523 2524 2525 2526
}
EXPORT_SYMBOL_GPL(lockdep_init_map);

/*
 * This gets called for every mutex_lock*()/spin_lock*() operation.
 * We maintain the dependency maps and validate the locking attempt:
 */
static int __lock_acquire(struct lockdep_map *lock, unsigned int subclass,
			  int trylock, int read, int check, int hardirqs_off,
P
Peter Zijlstra 已提交
2527
			  struct lockdep_map *nest_lock, unsigned long ip)
I
Ingo Molnar 已提交
2528 2529
{
	struct task_struct *curr = current;
2530
	struct lock_class *class = NULL;
I
Ingo Molnar 已提交
2531 2532 2533 2534 2535
	struct held_lock *hlock;
	unsigned int depth, id;
	int chain_head = 0;
	u64 chain_key;

P
Peter Zijlstra 已提交
2536 2537 2538
	if (!prove_locking)
		check = 1;

I
Ingo Molnar 已提交
2539 2540 2541 2542 2543 2544 2545 2546 2547 2548
	if (unlikely(!debug_locks))
		return 0;

	if (DEBUG_LOCKS_WARN_ON(!irqs_disabled()))
		return 0;

	if (unlikely(subclass >= MAX_LOCKDEP_SUBCLASSES)) {
		debug_locks_off();
		printk("BUG: MAX_LOCKDEP_SUBCLASSES too low!\n");
		printk("turning off the locking correctness validator.\n");
2549
		dump_stack();
I
Ingo Molnar 已提交
2550 2551 2552
		return 0;
	}

2553 2554 2555 2556 2557
	if (!subclass)
		class = lock->class_cache;
	/*
	 * Not cached yet or subclass?
	 */
I
Ingo Molnar 已提交
2558
	if (unlikely(!class)) {
2559
		class = register_lock_class(lock, subclass, 0);
I
Ingo Molnar 已提交
2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581
		if (!class)
			return 0;
	}
	debug_atomic_inc((atomic_t *)&class->ops);
	if (very_verbose(class)) {
		printk("\nacquire class [%p] %s", class->key, class->name);
		if (class->name_version > 1)
			printk("#%d", class->name_version);
		printk("\n");
		dump_stack();
	}

	/*
	 * Add the lock to the list of currently held locks.
	 * (we dont increase the depth just yet, up until the
	 * dependency checks are done)
	 */
	depth = curr->lockdep_depth;
	if (DEBUG_LOCKS_WARN_ON(depth >= MAX_LOCK_DEPTH))
		return 0;

	hlock = curr->held_locks + depth;
D
Dave Jones 已提交
2582 2583 2584
	if (DEBUG_LOCKS_WARN_ON(!class))
		return 0;
	hlock->class_idx = class - lock_classes + 1;
I
Ingo Molnar 已提交
2585 2586
	hlock->acquire_ip = ip;
	hlock->instance = lock;
P
Peter Zijlstra 已提交
2587
	hlock->nest_lock = nest_lock;
I
Ingo Molnar 已提交
2588 2589 2590
	hlock->trylock = trylock;
	hlock->read = read;
	hlock->check = check;
2591
	hlock->hardirqs_off = !!hardirqs_off;
P
Peter Zijlstra 已提交
2592 2593 2594 2595
#ifdef CONFIG_LOCK_STAT
	hlock->waittime_stamp = 0;
	hlock->holdtime_stamp = sched_clock();
#endif
I
Ingo Molnar 已提交
2596

P
Peter Zijlstra 已提交
2597 2598 2599
	if (check == 2 && !mark_irqflags(curr, hlock))
		return 0;

I
Ingo Molnar 已提交
2600
	/* mark it as used: */
2601
	if (!mark_lock(curr, hlock, LOCK_USED))
I
Ingo Molnar 已提交
2602
		return 0;
P
Peter Zijlstra 已提交
2603

I
Ingo Molnar 已提交
2604
	/*
2605
	 * Calculate the chain hash: it's the combined hash of all the
I
Ingo Molnar 已提交
2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625
	 * lock keys along the dependency chain. We save the hash value
	 * at every step so that we can get the current hash easily
	 * after unlock. The chain hash is then used to cache dependency
	 * results.
	 *
	 * The 'key ID' is what is the most compact key value to drive
	 * the hash, not class->key.
	 */
	id = class - lock_classes;
	if (DEBUG_LOCKS_WARN_ON(id >= MAX_LOCKDEP_KEYS))
		return 0;

	chain_key = curr->curr_chain_key;
	if (!depth) {
		if (DEBUG_LOCKS_WARN_ON(chain_key != 0))
			return 0;
		chain_head = 1;
	}

	hlock->prev_chain_key = chain_key;
P
Peter Zijlstra 已提交
2626 2627 2628
	if (separate_irq_context(curr, hlock)) {
		chain_key = 0;
		chain_head = 1;
I
Ingo Molnar 已提交
2629 2630 2631
	}
	chain_key = iterate_chain_key(chain_key, id);

2632
	if (!validate_chain(curr, lock, hlock, chain_head, chain_key))
P
Peter Zijlstra 已提交
2633
		return 0;
2634

2635
	curr->curr_chain_key = chain_key;
I
Ingo Molnar 已提交
2636 2637
	curr->lockdep_depth++;
	check_chain_key(curr);
2638 2639 2640 2641
#ifdef CONFIG_DEBUG_LOCKDEP
	if (unlikely(!debug_locks))
		return 0;
#endif
I
Ingo Molnar 已提交
2642 2643 2644 2645
	if (unlikely(curr->lockdep_depth >= MAX_LOCK_DEPTH)) {
		debug_locks_off();
		printk("BUG: MAX_LOCK_DEPTH too low!\n");
		printk("turning off the locking correctness validator.\n");
2646
		dump_stack();
I
Ingo Molnar 已提交
2647 2648
		return 0;
	}
2649

I
Ingo Molnar 已提交
2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668
	if (unlikely(curr->lockdep_depth > max_lockdep_depth))
		max_lockdep_depth = curr->lockdep_depth;

	return 1;
}

static int
print_unlock_inbalance_bug(struct task_struct *curr, struct lockdep_map *lock,
			   unsigned long ip)
{
	if (!debug_locks_off())
		return 0;
	if (debug_locks_silent)
		return 0;

	printk("\n=====================================\n");
	printk(  "[ BUG: bad unlock balance detected! ]\n");
	printk(  "-------------------------------------\n");
	printk("%s/%d is trying to release lock (",
2669
		curr->comm, task_pid_nr(curr));
I
Ingo Molnar 已提交
2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699
	print_lockdep_cache(lock);
	printk(") at:\n");
	print_ip_sym(ip);
	printk("but there are no more locks to release!\n");
	printk("\nother info that might help us debug this:\n");
	lockdep_print_held_locks(curr);

	printk("\nstack backtrace:\n");
	dump_stack();

	return 0;
}

/*
 * Common debugging checks for both nested and non-nested unlock:
 */
static int check_unlock(struct task_struct *curr, struct lockdep_map *lock,
			unsigned long ip)
{
	if (unlikely(!debug_locks))
		return 0;
	if (DEBUG_LOCKS_WARN_ON(!irqs_disabled()))
		return 0;

	if (curr->lockdep_depth <= 0)
		return print_unlock_inbalance_bug(curr, lock, ip);

	return 1;
}

2700
static int
2701 2702 2703
__lock_set_class(struct lockdep_map *lock, const char *name,
		 struct lock_class_key *key, unsigned int subclass,
		 unsigned long ip)
2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729
{
	struct task_struct *curr = current;
	struct held_lock *hlock, *prev_hlock;
	struct lock_class *class;
	unsigned int depth;
	int i;

	depth = curr->lockdep_depth;
	if (DEBUG_LOCKS_WARN_ON(!depth))
		return 0;

	prev_hlock = NULL;
	for (i = depth-1; i >= 0; i--) {
		hlock = curr->held_locks + i;
		/*
		 * We must not cross into another context:
		 */
		if (prev_hlock && prev_hlock->irq_context != hlock->irq_context)
			break;
		if (hlock->instance == lock)
			goto found_it;
		prev_hlock = hlock;
	}
	return print_unlock_inbalance_bug(curr, lock, ip);

found_it:
2730
	lockdep_init_map(lock, name, key, 0);
2731
	class = register_lock_class(lock, subclass, 0);
D
Dave Jones 已提交
2732
	hlock->class_idx = class - lock_classes + 1;
2733 2734 2735 2736 2737 2738 2739

	curr->lockdep_depth = i;
	curr->curr_chain_key = hlock->prev_chain_key;

	for (; i < depth; i++) {
		hlock = curr->held_locks + i;
		if (!__lock_acquire(hlock->instance,
D
Dave Jones 已提交
2740
			hlock_class(hlock)->subclass, hlock->trylock,
2741
				hlock->read, hlock->check, hlock->hardirqs_off,
P
Peter Zijlstra 已提交
2742
				hlock->nest_lock, hlock->acquire_ip))
2743 2744 2745 2746 2747 2748 2749 2750
			return 0;
	}

	if (DEBUG_LOCKS_WARN_ON(curr->lockdep_depth != depth))
		return 0;
	return 1;
}

I
Ingo Molnar 已提交
2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787
/*
 * Remove the lock to the list of currently held locks in a
 * potentially non-nested (out of order) manner. This is a
 * relatively rare operation, as all the unlock APIs default
 * to nested mode (which uses lock_release()):
 */
static int
lock_release_non_nested(struct task_struct *curr,
			struct lockdep_map *lock, unsigned long ip)
{
	struct held_lock *hlock, *prev_hlock;
	unsigned int depth;
	int i;

	/*
	 * Check whether the lock exists in the current stack
	 * of held locks:
	 */
	depth = curr->lockdep_depth;
	if (DEBUG_LOCKS_WARN_ON(!depth))
		return 0;

	prev_hlock = NULL;
	for (i = depth-1; i >= 0; i--) {
		hlock = curr->held_locks + i;
		/*
		 * We must not cross into another context:
		 */
		if (prev_hlock && prev_hlock->irq_context != hlock->irq_context)
			break;
		if (hlock->instance == lock)
			goto found_it;
		prev_hlock = hlock;
	}
	return print_unlock_inbalance_bug(curr, lock, ip);

found_it:
P
Peter Zijlstra 已提交
2788 2789
	lock_release_holdtime(hlock);

I
Ingo Molnar 已提交
2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800
	/*
	 * We have the right lock to unlock, 'hlock' points to it.
	 * Now we remove it from the stack, and add back the other
	 * entries (if any), recalculating the hash along the way:
	 */
	curr->lockdep_depth = i;
	curr->curr_chain_key = hlock->prev_chain_key;

	for (i++; i < depth; i++) {
		hlock = curr->held_locks + i;
		if (!__lock_acquire(hlock->instance,
D
Dave Jones 已提交
2801
			hlock_class(hlock)->subclass, hlock->trylock,
I
Ingo Molnar 已提交
2802
				hlock->read, hlock->check, hlock->hardirqs_off,
P
Peter Zijlstra 已提交
2803
				hlock->nest_lock, hlock->acquire_ip))
I
Ingo Molnar 已提交
2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841
			return 0;
	}

	if (DEBUG_LOCKS_WARN_ON(curr->lockdep_depth != depth - 1))
		return 0;
	return 1;
}

/*
 * Remove the lock to the list of currently held locks - this gets
 * called on mutex_unlock()/spin_unlock*() (or on a failed
 * mutex_lock_interruptible()). This is done for unlocks that nest
 * perfectly. (i.e. the current top of the lock-stack is unlocked)
 */
static int lock_release_nested(struct task_struct *curr,
			       struct lockdep_map *lock, unsigned long ip)
{
	struct held_lock *hlock;
	unsigned int depth;

	/*
	 * Pop off the top of the lock stack:
	 */
	depth = curr->lockdep_depth - 1;
	hlock = curr->held_locks + depth;

	/*
	 * Is the unlock non-nested:
	 */
	if (hlock->instance != lock)
		return lock_release_non_nested(curr, lock, ip);
	curr->lockdep_depth--;

	if (DEBUG_LOCKS_WARN_ON(!depth && (hlock->prev_chain_key != 0)))
		return 0;

	curr->curr_chain_key = hlock->prev_chain_key;

P
Peter Zijlstra 已提交
2842 2843
	lock_release_holdtime(hlock);

I
Ingo Molnar 已提交
2844 2845
#ifdef CONFIG_DEBUG_LOCKDEP
	hlock->prev_chain_key = 0;
D
Dave Jones 已提交
2846
	hlock->class_idx = 0;
I
Ingo Molnar 已提交
2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880
	hlock->acquire_ip = 0;
	hlock->irq_context = 0;
#endif
	return 1;
}

/*
 * Remove the lock to the list of currently held locks - this gets
 * called on mutex_unlock()/spin_unlock*() (or on a failed
 * mutex_lock_interruptible()). This is done for unlocks that nest
 * perfectly. (i.e. the current top of the lock-stack is unlocked)
 */
static void
__lock_release(struct lockdep_map *lock, int nested, unsigned long ip)
{
	struct task_struct *curr = current;

	if (!check_unlock(curr, lock, ip))
		return;

	if (nested) {
		if (!lock_release_nested(curr, lock, ip))
			return;
	} else {
		if (!lock_release_non_nested(curr, lock, ip))
			return;
	}

	check_chain_key(curr);
}

/*
 * Check whether we follow the irq-flags state precisely:
 */
2881
static void check_flags(unsigned long flags)
I
Ingo Molnar 已提交
2882
{
2883 2884
#if defined(CONFIG_PROVE_LOCKING) && defined(CONFIG_DEBUG_LOCKDEP) && \
    defined(CONFIG_TRACE_IRQFLAGS)
I
Ingo Molnar 已提交
2885 2886 2887
	if (!debug_locks)
		return;

2888 2889 2890 2891 2892 2893 2894 2895 2896
	if (irqs_disabled_flags(flags)) {
		if (DEBUG_LOCKS_WARN_ON(current->hardirqs_enabled)) {
			printk("possible reason: unannotated irqs-off.\n");
		}
	} else {
		if (DEBUG_LOCKS_WARN_ON(!current->hardirqs_enabled)) {
			printk("possible reason: unannotated irqs-on.\n");
		}
	}
I
Ingo Molnar 已提交
2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914

	/*
	 * We dont accurately track softirq state in e.g.
	 * hardirq contexts (such as on 4KSTACKS), so only
	 * check if not in hardirq contexts:
	 */
	if (!hardirq_count()) {
		if (softirq_count())
			DEBUG_LOCKS_WARN_ON(current->softirqs_enabled);
		else
			DEBUG_LOCKS_WARN_ON(!current->softirqs_enabled);
	}

	if (!debug_locks)
		print_irqtrace_events(current);
#endif
}

2915 2916 2917
void lock_set_class(struct lockdep_map *lock, const char *name,
		    struct lock_class_key *key, unsigned int subclass,
		    unsigned long ip)
2918 2919 2920 2921 2922 2923 2924 2925 2926
{
	unsigned long flags;

	if (unlikely(current->lockdep_recursion))
		return;

	raw_local_irq_save(flags);
	current->lockdep_recursion = 1;
	check_flags(flags);
2927
	if (__lock_set_class(lock, name, key, subclass, ip))
2928 2929 2930 2931
		check_chain_key(current);
	current->lockdep_recursion = 0;
	raw_local_irq_restore(flags);
}
2932
EXPORT_SYMBOL_GPL(lock_set_class);
2933

I
Ingo Molnar 已提交
2934 2935 2936 2937
/*
 * We are not always called with irqs disabled - do that here,
 * and also avoid lockdep recursion:
 */
2938
void lock_acquire(struct lockdep_map *lock, unsigned int subclass,
P
Peter Zijlstra 已提交
2939 2940
			  int trylock, int read, int check,
			  struct lockdep_map *nest_lock, unsigned long ip)
I
Ingo Molnar 已提交
2941 2942 2943
{
	unsigned long flags;

2944 2945
	trace_lock_acquire(lock, subclass, trylock, read, check, nest_lock, ip);

I
Ingo Molnar 已提交
2946 2947 2948 2949 2950 2951 2952 2953
	if (unlikely(current->lockdep_recursion))
		return;

	raw_local_irq_save(flags);
	check_flags(flags);

	current->lockdep_recursion = 1;
	__lock_acquire(lock, subclass, trylock, read, check,
P
Peter Zijlstra 已提交
2954
		       irqs_disabled_flags(flags), nest_lock, ip);
I
Ingo Molnar 已提交
2955 2956 2957 2958 2959
	current->lockdep_recursion = 0;
	raw_local_irq_restore(flags);
}
EXPORT_SYMBOL_GPL(lock_acquire);

2960
void lock_release(struct lockdep_map *lock, int nested,
2961
			  unsigned long ip)
I
Ingo Molnar 已提交
2962 2963 2964
{
	unsigned long flags;

2965 2966
	trace_lock_release(lock, nested, ip);

I
Ingo Molnar 已提交
2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978
	if (unlikely(current->lockdep_recursion))
		return;

	raw_local_irq_save(flags);
	check_flags(flags);
	current->lockdep_recursion = 1;
	__lock_release(lock, nested, ip);
	current->lockdep_recursion = 0;
	raw_local_irq_restore(flags);
}
EXPORT_SYMBOL_GPL(lock_release);

2979 2980 2981 2982 2983 2984 2985 2986 2987 2988
void lockdep_set_current_reclaim_state(gfp_t gfp_mask)
{
	current->lockdep_reclaim_gfp = gfp_mask;
}

void lockdep_clear_current_reclaim_state(void)
{
	current->lockdep_reclaim_gfp = 0;
}

P
Peter Zijlstra 已提交
2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002
#ifdef CONFIG_LOCK_STAT
static int
print_lock_contention_bug(struct task_struct *curr, struct lockdep_map *lock,
			   unsigned long ip)
{
	if (!debug_locks_off())
		return 0;
	if (debug_locks_silent)
		return 0;

	printk("\n=================================\n");
	printk(  "[ BUG: bad contention detected! ]\n");
	printk(  "---------------------------------\n");
	printk("%s/%d is trying to contend lock (",
3003
		curr->comm, task_pid_nr(curr));
P
Peter Zijlstra 已提交
3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023
	print_lockdep_cache(lock);
	printk(") at:\n");
	print_ip_sym(ip);
	printk("but there are no locks held!\n");
	printk("\nother info that might help us debug this:\n");
	lockdep_print_held_locks(curr);

	printk("\nstack backtrace:\n");
	dump_stack();

	return 0;
}

static void
__lock_contended(struct lockdep_map *lock, unsigned long ip)
{
	struct task_struct *curr = current;
	struct held_lock *hlock, *prev_hlock;
	struct lock_class_stats *stats;
	unsigned int depth;
P
Peter Zijlstra 已提交
3024
	int i, contention_point, contending_point;
P
Peter Zijlstra 已提交
3025 3026 3027 3028 3029 3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047

	depth = curr->lockdep_depth;
	if (DEBUG_LOCKS_WARN_ON(!depth))
		return;

	prev_hlock = NULL;
	for (i = depth-1; i >= 0; i--) {
		hlock = curr->held_locks + i;
		/*
		 * We must not cross into another context:
		 */
		if (prev_hlock && prev_hlock->irq_context != hlock->irq_context)
			break;
		if (hlock->instance == lock)
			goto found_it;
		prev_hlock = hlock;
	}
	print_lock_contention_bug(curr, lock, ip);
	return;

found_it:
	hlock->waittime_stamp = sched_clock();

P
Peter Zijlstra 已提交
3048 3049 3050
	contention_point = lock_point(hlock_class(hlock)->contention_point, ip);
	contending_point = lock_point(hlock_class(hlock)->contending_point,
				      lock->ip);
P
Peter Zijlstra 已提交
3051

D
Dave Jones 已提交
3052
	stats = get_lock_stats(hlock_class(hlock));
P
Peter Zijlstra 已提交
3053 3054 3055 3056
	if (contention_point < LOCKSTAT_POINTS)
		stats->contention_point[contention_point]++;
	if (contending_point < LOCKSTAT_POINTS)
		stats->contending_point[contending_point]++;
P
Peter Zijlstra 已提交
3057 3058
	if (lock->cpu != smp_processor_id())
		stats->bounces[bounce_contended + !!hlock->read]++;
P
Peter Zijlstra 已提交
3059 3060 3061 3062
	put_lock_stats(stats);
}

static void
P
Peter Zijlstra 已提交
3063
__lock_acquired(struct lockdep_map *lock, unsigned long ip)
P
Peter Zijlstra 已提交
3064 3065 3066 3067 3068 3069
{
	struct task_struct *curr = current;
	struct held_lock *hlock, *prev_hlock;
	struct lock_class_stats *stats;
	unsigned int depth;
	u64 now;
P
Peter Zijlstra 已提交
3070 3071
	s64 waittime = 0;
	int i, cpu;
P
Peter Zijlstra 已提交
3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092

	depth = curr->lockdep_depth;
	if (DEBUG_LOCKS_WARN_ON(!depth))
		return;

	prev_hlock = NULL;
	for (i = depth-1; i >= 0; i--) {
		hlock = curr->held_locks + i;
		/*
		 * We must not cross into another context:
		 */
		if (prev_hlock && prev_hlock->irq_context != hlock->irq_context)
			break;
		if (hlock->instance == lock)
			goto found_it;
		prev_hlock = hlock;
	}
	print_lock_contention_bug(curr, lock, _RET_IP_);
	return;

found_it:
P
Peter Zijlstra 已提交
3093 3094 3095 3096 3097 3098
	cpu = smp_processor_id();
	if (hlock->waittime_stamp) {
		now = sched_clock();
		waittime = now - hlock->waittime_stamp;
		hlock->holdtime_stamp = now;
	}
P
Peter Zijlstra 已提交
3099

3100 3101
	trace_lock_acquired(lock, ip, waittime);

D
Dave Jones 已提交
3102
	stats = get_lock_stats(hlock_class(hlock));
P
Peter Zijlstra 已提交
3103 3104 3105 3106 3107 3108 3109 3110
	if (waittime) {
		if (hlock->read)
			lock_time_inc(&stats->read_waittime, waittime);
		else
			lock_time_inc(&stats->write_waittime, waittime);
	}
	if (lock->cpu != cpu)
		stats->bounces[bounce_acquired + !!hlock->read]++;
P
Peter Zijlstra 已提交
3111
	put_lock_stats(stats);
P
Peter Zijlstra 已提交
3112 3113

	lock->cpu = cpu;
P
Peter Zijlstra 已提交
3114
	lock->ip = ip;
P
Peter Zijlstra 已提交
3115 3116 3117 3118 3119 3120
}

void lock_contended(struct lockdep_map *lock, unsigned long ip)
{
	unsigned long flags;

3121 3122
	trace_lock_contended(lock, ip);

P
Peter Zijlstra 已提交
3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137
	if (unlikely(!lock_stat))
		return;

	if (unlikely(current->lockdep_recursion))
		return;

	raw_local_irq_save(flags);
	check_flags(flags);
	current->lockdep_recursion = 1;
	__lock_contended(lock, ip);
	current->lockdep_recursion = 0;
	raw_local_irq_restore(flags);
}
EXPORT_SYMBOL_GPL(lock_contended);

P
Peter Zijlstra 已提交
3138
void lock_acquired(struct lockdep_map *lock, unsigned long ip)
P
Peter Zijlstra 已提交
3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 3150
{
	unsigned long flags;

	if (unlikely(!lock_stat))
		return;

	if (unlikely(current->lockdep_recursion))
		return;

	raw_local_irq_save(flags);
	check_flags(flags);
	current->lockdep_recursion = 1;
P
Peter Zijlstra 已提交
3151
	__lock_acquired(lock, ip);
P
Peter Zijlstra 已提交
3152 3153 3154 3155 3156 3157
	current->lockdep_recursion = 0;
	raw_local_irq_restore(flags);
}
EXPORT_SYMBOL_GPL(lock_acquired);
#endif

I
Ingo Molnar 已提交
3158 3159 3160 3161 3162 3163 3164 3165
/*
 * Used by the testsuite, sanitize the validator state
 * after a simulated failure:
 */

void lockdep_reset(void)
{
	unsigned long flags;
3166
	int i;
I
Ingo Molnar 已提交
3167 3168 3169 3170 3171 3172 3173 3174 3175 3176

	raw_local_irq_save(flags);
	current->curr_chain_key = 0;
	current->lockdep_depth = 0;
	current->lockdep_recursion = 0;
	memset(current->held_locks, 0, MAX_LOCK_DEPTH*sizeof(struct held_lock));
	nr_hardirq_chains = 0;
	nr_softirq_chains = 0;
	nr_process_chains = 0;
	debug_locks = 1;
3177 3178
	for (i = 0; i < CHAINHASH_SIZE; i++)
		INIT_LIST_HEAD(chainhash_table + i);
I
Ingo Molnar 已提交
3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199
	raw_local_irq_restore(flags);
}

static void zap_class(struct lock_class *class)
{
	int i;

	/*
	 * Remove all dependencies this lock is
	 * involved in:
	 */
	for (i = 0; i < nr_list_entries; i++) {
		if (list_entries[i].class == class)
			list_del_rcu(&list_entries[i].entry);
	}
	/*
	 * Unhash the class and remove it from the all_lock_classes list:
	 */
	list_del_rcu(&class->hash_entry);
	list_del_rcu(&class->lock_entry);

3200
	class->key = NULL;
I
Ingo Molnar 已提交
3201 3202
}

3203
static inline int within(const void *addr, void *start, unsigned long size)
I
Ingo Molnar 已提交
3204 3205 3206 3207 3208 3209 3210 3211 3212 3213
{
	return addr >= start && addr < start + size;
}

void lockdep_free_key_range(void *start, unsigned long size)
{
	struct lock_class *class, *next;
	struct list_head *head;
	unsigned long flags;
	int i;
3214
	int locked;
I
Ingo Molnar 已提交
3215 3216

	raw_local_irq_save(flags);
3217
	locked = graph_lock();
I
Ingo Molnar 已提交
3218 3219 3220 3221 3222 3223 3224 3225

	/*
	 * Unhash all classes that were created by this module:
	 */
	for (i = 0; i < CLASSHASH_SIZE; i++) {
		head = classhash_table + i;
		if (list_empty(head))
			continue;
3226
		list_for_each_entry_safe(class, next, head, hash_entry) {
I
Ingo Molnar 已提交
3227 3228
			if (within(class->key, start, size))
				zap_class(class);
3229 3230 3231
			else if (within(class->name, start, size))
				zap_class(class);
		}
I
Ingo Molnar 已提交
3232 3233
	}

3234 3235
	if (locked)
		graph_unlock();
I
Ingo Molnar 已提交
3236 3237 3238 3239 3240
	raw_local_irq_restore(flags);
}

void lockdep_reset_lock(struct lockdep_map *lock)
{
3241
	struct lock_class *class, *next;
I
Ingo Molnar 已提交
3242 3243 3244
	struct list_head *head;
	unsigned long flags;
	int i, j;
3245
	int locked;
I
Ingo Molnar 已提交
3246 3247 3248 3249

	raw_local_irq_save(flags);

	/*
3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262
	 * Remove all classes this lock might have:
	 */
	for (j = 0; j < MAX_LOCKDEP_SUBCLASSES; j++) {
		/*
		 * If the class exists we look it up and zap it:
		 */
		class = look_up_lock_class(lock, j);
		if (class)
			zap_class(class);
	}
	/*
	 * Debug check: in the end all mapped classes should
	 * be gone.
I
Ingo Molnar 已提交
3263
	 */
3264
	locked = graph_lock();
I
Ingo Molnar 已提交
3265 3266 3267 3268 3269
	for (i = 0; i < CLASSHASH_SIZE; i++) {
		head = classhash_table + i;
		if (list_empty(head))
			continue;
		list_for_each_entry_safe(class, next, head, hash_entry) {
3270
			if (unlikely(class == lock->class_cache)) {
3271 3272
				if (debug_locks_off_graph_unlock())
					WARN_ON(1);
3273
				goto out_restore;
I
Ingo Molnar 已提交
3274 3275 3276
			}
		}
	}
3277 3278
	if (locked)
		graph_unlock();
3279 3280

out_restore:
I
Ingo Molnar 已提交
3281 3282 3283
	raw_local_irq_restore(flags);
}

3284
void lockdep_init(void)
I
Ingo Molnar 已提交
3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308 3309
{
	int i;

	/*
	 * Some architectures have their own start_kernel()
	 * code which calls lockdep_init(), while we also
	 * call lockdep_init() from the start_kernel() itself,
	 * and we want to initialize the hashes only once:
	 */
	if (lockdep_initialized)
		return;

	for (i = 0; i < CLASSHASH_SIZE; i++)
		INIT_LIST_HEAD(classhash_table + i);

	for (i = 0; i < CHAINHASH_SIZE; i++)
		INIT_LIST_HEAD(chainhash_table + i);

	lockdep_initialized = 1;
}

void __init lockdep_info(void)
{
	printk("Lock dependency validator: Copyright (c) 2006 Red Hat, Inc., Ingo Molnar\n");

3310
	printk("... MAX_LOCKDEP_SUBCLASSES:  %lu\n", MAX_LOCKDEP_SUBCLASSES);
I
Ingo Molnar 已提交
3311 3312
	printk("... MAX_LOCK_DEPTH:          %lu\n", MAX_LOCK_DEPTH);
	printk("... MAX_LOCKDEP_KEYS:        %lu\n", MAX_LOCKDEP_KEYS);
3313
	printk("... CLASSHASH_SIZE:          %lu\n", CLASSHASH_SIZE);
I
Ingo Molnar 已提交
3314 3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328
	printk("... MAX_LOCKDEP_ENTRIES:     %lu\n", MAX_LOCKDEP_ENTRIES);
	printk("... MAX_LOCKDEP_CHAINS:      %lu\n", MAX_LOCKDEP_CHAINS);
	printk("... CHAINHASH_SIZE:          %lu\n", CHAINHASH_SIZE);

	printk(" memory used by lock dependency info: %lu kB\n",
		(sizeof(struct lock_class) * MAX_LOCKDEP_KEYS +
		sizeof(struct list_head) * CLASSHASH_SIZE +
		sizeof(struct lock_list) * MAX_LOCKDEP_ENTRIES +
		sizeof(struct lock_chain) * MAX_LOCKDEP_CHAINS +
		sizeof(struct list_head) * CHAINHASH_SIZE) / 1024);

	printk(" per task-struct memory footprint: %lu bytes\n",
		sizeof(struct held_lock) * MAX_LOCK_DEPTH);

#ifdef CONFIG_DEBUG_LOCKDEP
3329 3330 3331 3332 3333
	if (lockdep_init_error) {
		printk("WARNING: lockdep init error! Arch code didn't call lockdep_init() early enough?\n");
		printk("Call stack leading to lockdep invocation was:\n");
		print_stack_trace(&lockdep_init_trace, 0);
	}
I
Ingo Molnar 已提交
3334 3335 3336 3337 3338
#endif
}

static void
print_freed_lock_bug(struct task_struct *curr, const void *mem_from,
3339
		     const void *mem_to, struct held_lock *hlock)
I
Ingo Molnar 已提交
3340 3341 3342 3343 3344 3345 3346 3347 3348 3349
{
	if (!debug_locks_off())
		return;
	if (debug_locks_silent)
		return;

	printk("\n=========================\n");
	printk(  "[ BUG: held lock freed! ]\n");
	printk(  "-------------------------\n");
	printk("%s/%d is freeing memory %p-%p, with a lock still held there!\n",
3350
		curr->comm, task_pid_nr(curr), mem_from, mem_to-1);
3351
	print_lock(hlock);
I
Ingo Molnar 已提交
3352 3353 3354 3355 3356 3357
	lockdep_print_held_locks(curr);

	printk("\nstack backtrace:\n");
	dump_stack();
}

O
Oleg Nesterov 已提交
3358 3359 3360 3361 3362 3363 3364
static inline int not_in_range(const void* mem_from, unsigned long mem_len,
				const void* lock_from, unsigned long lock_len)
{
	return lock_from + lock_len <= mem_from ||
		mem_from + mem_len <= lock_from;
}

I
Ingo Molnar 已提交
3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377 3378 3379 3380 3381 3382 3383
/*
 * Called when kernel memory is freed (or unmapped), or if a lock
 * is destroyed or reinitialized - this code checks whether there is
 * any held lock in the memory range of <from> to <to>:
 */
void debug_check_no_locks_freed(const void *mem_from, unsigned long mem_len)
{
	struct task_struct *curr = current;
	struct held_lock *hlock;
	unsigned long flags;
	int i;

	if (unlikely(!debug_locks))
		return;

	local_irq_save(flags);
	for (i = 0; i < curr->lockdep_depth; i++) {
		hlock = curr->held_locks + i;

O
Oleg Nesterov 已提交
3384 3385
		if (not_in_range(mem_from, mem_len, hlock->instance,
					sizeof(*hlock->instance)))
I
Ingo Molnar 已提交
3386 3387
			continue;

O
Oleg Nesterov 已提交
3388
		print_freed_lock_bug(curr, mem_from, mem_from + mem_len, hlock);
I
Ingo Molnar 已提交
3389 3390 3391 3392
		break;
	}
	local_irq_restore(flags);
}
3393
EXPORT_SYMBOL_GPL(debug_check_no_locks_freed);
I
Ingo Molnar 已提交
3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405

static void print_held_locks_bug(struct task_struct *curr)
{
	if (!debug_locks_off())
		return;
	if (debug_locks_silent)
		return;

	printk("\n=====================================\n");
	printk(  "[ BUG: lock held at task exit time! ]\n");
	printk(  "-------------------------------------\n");
	printk("%s/%d is exiting with locks still held!\n",
3406
		curr->comm, task_pid_nr(curr));
I
Ingo Molnar 已提交
3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424
	lockdep_print_held_locks(curr);

	printk("\nstack backtrace:\n");
	dump_stack();
}

void debug_check_no_locks_held(struct task_struct *task)
{
	if (unlikely(task->lockdep_depth > 0))
		print_held_locks_bug(task);
}

void debug_show_all_locks(void)
{
	struct task_struct *g, *p;
	int count = 10;
	int unlock = 1;

3425 3426 3427 3428
	if (unlikely(!debug_locks)) {
		printk("INFO: lockdep is turned off.\n");
		return;
	}
I
Ingo Molnar 已提交
3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448
	printk("\nShowing all locks held in the system:\n");

	/*
	 * Here we try to get the tasklist_lock as hard as possible,
	 * if not successful after 2 seconds we ignore it (but keep
	 * trying). This is to enable a debug printout even if a
	 * tasklist_lock-holding task deadlocks or crashes.
	 */
retry:
	if (!read_trylock(&tasklist_lock)) {
		if (count == 10)
			printk("hm, tasklist_lock locked, retrying... ");
		if (count) {
			count--;
			printk(" #%d", 10-count);
			mdelay(200);
			goto retry;
		}
		printk(" ignoring it.\n");
		unlock = 0;
3449 3450 3451
	} else {
		if (count != 10)
			printk(KERN_CONT " locked it.\n");
I
Ingo Molnar 已提交
3452 3453 3454
	}

	do_each_thread(g, p) {
I
Ingo Molnar 已提交
3455 3456 3457 3458 3459 3460 3461
		/*
		 * It's not reliable to print a task's held locks
		 * if it's not sleeping (or if it's not the current
		 * task):
		 */
		if (p->state == TASK_RUNNING && p != current)
			continue;
I
Ingo Molnar 已提交
3462 3463 3464 3465 3466 3467 3468 3469 3470 3471 3472 3473 3474 3475 3476
		if (p->lockdep_depth)
			lockdep_print_held_locks(p);
		if (!unlock)
			if (read_trylock(&tasklist_lock))
				unlock = 1;
	} while_each_thread(g, p);

	printk("\n");
	printk("=============================================\n\n");

	if (unlock)
		read_unlock(&tasklist_lock);
}
EXPORT_SYMBOL_GPL(debug_show_all_locks);

3477 3478 3479 3480 3481
/*
 * Careful: only use this function if you are sure that
 * the task cannot run in parallel!
 */
void __debug_show_held_locks(struct task_struct *task)
I
Ingo Molnar 已提交
3482
{
3483 3484 3485 3486
	if (unlikely(!debug_locks)) {
		printk("INFO: lockdep is turned off.\n");
		return;
	}
I
Ingo Molnar 已提交
3487 3488
	lockdep_print_held_locks(task);
}
3489 3490 3491 3492 3493 3494
EXPORT_SYMBOL_GPL(__debug_show_held_locks);

void debug_show_held_locks(struct task_struct *task)
{
		__debug_show_held_locks(task);
}
I
Ingo Molnar 已提交
3495
EXPORT_SYMBOL_GPL(debug_show_held_locks);
P
Peter Zijlstra 已提交
3496 3497 3498 3499 3500 3501 3502 3503 3504 3505 3506 3507 3508 3509 3510 3511

void lockdep_sys_exit(void)
{
	struct task_struct *curr = current;

	if (unlikely(curr->lockdep_depth)) {
		if (!debug_locks_off())
			return;
		printk("\n================================================\n");
		printk(  "[ BUG: lock held when returning to user space! ]\n");
		printk(  "------------------------------------------------\n");
		printk("%s/%d is leaving the kernel with locks still held!\n",
				curr->comm, curr->pid);
		lockdep_print_held_locks(curr);
	}
}