kmemleak.c 60 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
/*
 * mm/kmemleak.c
 *
 * Copyright (C) 2008 ARM Limited
 * Written by Catalin Marinas <catalin.marinas@arm.com>
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as
 * published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
 *
 *
 * For more information on the algorithm and kmemleak usage, please see
22
 * Documentation/dev-tools/kmemleak.rst.
23 24 25 26 27 28 29 30 31
 *
 * Notes on locking
 * ----------------
 *
 * The following locks and mutexes are used by kmemleak:
 *
 * - kmemleak_lock (rwlock): protects the object_list modifications and
 *   accesses to the object_tree_root. The object_list is the main list
 *   holding the metadata (struct kmemleak_object) for the allocated memory
32
 *   blocks. The object_tree_root is a red black tree used to look-up
33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50
 *   metadata based on a pointer to the corresponding memory block.  The
 *   kmemleak_object structures are added to the object_list and
 *   object_tree_root in the create_object() function called from the
 *   kmemleak_alloc() callback and removed in delete_object() called from the
 *   kmemleak_free() callback
 * - kmemleak_object.lock (spinlock): protects a kmemleak_object. Accesses to
 *   the metadata (e.g. count) are protected by this lock. Note that some
 *   members of this structure may be protected by other means (atomic or
 *   kmemleak_lock). This lock is also held when scanning the corresponding
 *   memory block to avoid the kernel freeing it via the kmemleak_free()
 *   callback. This is less heavyweight than holding a global lock like
 *   kmemleak_lock during scanning
 * - scan_mutex (mutex): ensures that only one thread may scan the memory for
 *   unreferenced objects at a time. The gray_list contains the objects which
 *   are already referenced or marked as false positives and need to be
 *   scanned. This list is only modified during a scanning episode when the
 *   scan_mutex is held. At the end of a scan, the gray_list is always empty.
 *   Note that the kmemleak_object.use_count is incremented when an object is
51 52 53 54
 *   added to the gray_list and therefore cannot be freed. This mutex also
 *   prevents multiple users of the "kmemleak" debugfs file together with
 *   modifications to the memory scanning parameters including the scan_thread
 *   pointer
55
 *
56
 * Locks and mutexes are acquired/nested in the following order:
57
 *
58 59 60 61
 *   scan_mutex [-> object->lock] -> kmemleak_lock -> other_object->lock (SINGLE_DEPTH_NESTING)
 *
 * No kmemleak_lock and object->lock nesting is allowed outside scan_mutex
 * regions.
62
 *
63 64 65 66 67 68 69 70
 * The kmemleak_object structures have a use_count incremented or decremented
 * using the get_object()/put_object() functions. When the use_count becomes
 * 0, this count can no longer be incremented and put_object() schedules the
 * kmemleak_object freeing via an RCU callback. All calls to the get_object()
 * function must be protected by rcu_read_lock() to avoid accessing a freed
 * structure.
 */

Joe Perches's avatar
Joe Perches committed
71 72
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

73 74 75
#include <linux/init.h>
#include <linux/kernel.h>
#include <linux/list.h>
76
#include <linux/sched/signal.h>
77
#include <linux/sched/task.h>
78
#include <linux/sched/task_stack.h>
79 80
#include <linux/jiffies.h>
#include <linux/delay.h>
81
#include <linux/export.h>
82
#include <linux/kthread.h>
83
#include <linux/rbtree.h>
84 85 86 87 88
#include <linux/fs.h>
#include <linux/debugfs.h>
#include <linux/seq_file.h>
#include <linux/cpumask.h>
#include <linux/spinlock.h>
89
#include <linux/module.h>
90 91 92 93 94
#include <linux/mutex.h>
#include <linux/rcupdate.h>
#include <linux/stacktrace.h>
#include <linux/cache.h>
#include <linux/percpu.h>
95
#include <linux/memblock.h>
96
#include <linux/pfn.h>
97 98 99 100 101 102 103 104
#include <linux/mmzone.h>
#include <linux/slab.h>
#include <linux/thread_info.h>
#include <linux/err.h>
#include <linux/uaccess.h>
#include <linux/string.h>
#include <linux/nodemask.h>
#include <linux/mm.h>
105
#include <linux/workqueue.h>
106
#include <linux/crc32.h>
107 108 109

#include <asm/sections.h>
#include <asm/processor.h>
Arun Sharma's avatar
Arun Sharma committed
110
#include <linux/atomic.h>
111

112
#include <linux/kasan.h>
113
#include <linux/kmemleak.h>
114
#include <linux/memory_hotplug.h>
115 116 117 118 119 120 121 122

/*
 * Kmemleak configuration and common defines.
 */
#define MAX_TRACE		16	/* stack trace length */
#define MSECS_MIN_AGE		5000	/* minimum object age for reporting */
#define SECS_FIRST_SCAN		60	/* delay before the first scan */
#define SECS_SCAN_WAIT		600	/* subsequent auto scanning delay */
123
#define MAX_SCAN_SIZE		4096	/* maximum size of a scanned block */
124 125 126

#define BYTES_PER_POINTER	sizeof(void *)

127
/* GFP bitmask for kmemleak internal allocations */
128
#define gfp_kmemleak_mask(gfp)	(((gfp) & (GFP_KERNEL | GFP_ATOMIC)) | \
129
				 __GFP_NORETRY | __GFP_NOMEMALLOC | \
130
				 __GFP_NOWARN | __GFP_NOFAIL)
131

132 133 134
/* scanning area inside a memory block */
struct kmemleak_scan_area {
	struct hlist_node node;
135 136
	unsigned long start;
	size_t size;
137 138
};

139 140 141
#define KMEMLEAK_GREY	0
#define KMEMLEAK_BLACK	-1

142 143 144 145
/*
 * Structure holding the metadata for each allocated memory block.
 * Modifications to such objects should be made while holding the
 * object->lock. Insertions or deletions from object_list, gray_list or
146
 * rb_node are already protected by the corresponding locks or mutex (see
147 148 149 150 151
 * the notes on locking above). These objects are reference-counted
 * (use_count) and freed using the RCU mechanism.
 */
struct kmemleak_object {
	spinlock_t lock;
152
	unsigned int flags;		/* object status flags */
153 154
	struct list_head object_list;
	struct list_head gray_list;
155
	struct rb_node rb_node;
156 157 158 159 160
	struct rcu_head rcu;		/* object_list lockless traversal */
	/* object usage count; object freed when use_count == 0 */
	atomic_t use_count;
	unsigned long pointer;
	size_t size;
161 162
	/* pass surplus references to this pointer */
	unsigned long excess_ref;
163 164 165 166
	/* minimum number of a pointers found before it is considered leak */
	int min_count;
	/* the total number of pointers found pointing to this object */
	int count;
167 168
	/* checksum for detecting modified objects */
	u32 checksum;
169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184
	/* memory ranges to be scanned inside an object (empty for all) */
	struct hlist_head area_list;
	unsigned long trace[MAX_TRACE];
	unsigned int trace_len;
	unsigned long jiffies;		/* creation timestamp */
	pid_t pid;			/* pid of the current task */
	char comm[TASK_COMM_LEN];	/* executable name */
};

/* flag representing the memory block allocation status */
#define OBJECT_ALLOCATED	(1 << 0)
/* flag set after the first reporting of an unreference object */
#define OBJECT_REPORTED		(1 << 1)
/* flag set to not scan the object */
#define OBJECT_NO_SCAN		(1 << 2)

185
#define HEX_PREFIX		"    "
186 187 188 189 190 191 192 193 194
/* number of bytes to print per line; must be 16 or 32 */
#define HEX_ROW_SIZE		16
/* number of bytes to print at a time (1, 2, 4, 8) */
#define HEX_GROUP_SIZE		1
/* include ASCII after the hex output */
#define HEX_ASCII		1
/* max number of lines to be printed */
#define HEX_MAX_LINES		2

195 196 197 198
/* the list of all allocated objects */
static LIST_HEAD(object_list);
/* the list of gray-colored objects (see color_gray comment below) */
static LIST_HEAD(gray_list);
199 200 201
/* search tree for object boundaries */
static struct rb_root object_tree_root = RB_ROOT;
/* rw_lock protecting the access to object_list and object_tree_root */
202 203 204 205 206 207 208
static DEFINE_RWLOCK(kmemleak_lock);

/* allocation caches for kmemleak internal data */
static struct kmem_cache *object_cache;
static struct kmem_cache *scan_area_cache;

/* set if tracing memory operations is enabled */
209
static int kmemleak_enabled;
210 211
/* same as above but only for the kmemleak_free() callback */
static int kmemleak_free_enabled;
212
/* set in the late_initcall if there were no errors */
213
static int kmemleak_initialized;
214
/* enables or disables early logging of the memory operations */
215
static int kmemleak_early_log = 1;
216
/* set if a kmemleak warning was issued */
217
static int kmemleak_warning;
218
/* set if a fatal kmemleak error has occurred */
219
static int kmemleak_error;
220 221 222 223 224 225

/* minimum and maximum address that may be valid pointers */
static unsigned long min_addr = ULONG_MAX;
static unsigned long max_addr;

static struct task_struct *scan_thread;
226
/* used to avoid reporting of recently allocated objects */
227
static unsigned long jiffies_min_age;
228
static unsigned long jiffies_last_scan;
229 230 231
/* delay between automatic memory scannings */
static signed long jiffies_scan_wait;
/* enables or disables the task stacks scanning */
232
static int kmemleak_stack_scan = 1;
233
/* protects the memory scanning, parameters and debug/kmemleak file access */
234
static DEFINE_MUTEX(scan_mutex);
235 236
/* setting kmemleak=on, will set this var, skipping the disable */
static int kmemleak_skip_disable;
237 238
/* If there are leaks that can be reported */
static bool kmemleak_found_leaks;
239

240 241 242
static bool kmemleak_verbose;
module_param_named(verbose, kmemleak_verbose, bool, 0600);

243
/*
244
 * Early object allocation/freeing logging. Kmemleak is initialized after the
245
 * kernel allocator. However, both the kernel allocator and kmemleak may
246
 * allocate memory blocks which need to be tracked. Kmemleak defines an
247 248 249 250 251 252 253
 * arbitrary buffer to hold the allocation/freeing information before it is
 * fully initialized.
 */

/* kmemleak operation type for early logging */
enum {
	KMEMLEAK_ALLOC,
254
	KMEMLEAK_ALLOC_PERCPU,
255
	KMEMLEAK_FREE,
256
	KMEMLEAK_FREE_PART,
257
	KMEMLEAK_FREE_PERCPU,
258 259 260
	KMEMLEAK_NOT_LEAK,
	KMEMLEAK_IGNORE,
	KMEMLEAK_SCAN_AREA,
261 262
	KMEMLEAK_NO_SCAN,
	KMEMLEAK_SET_EXCESS_REF
263 264 265 266 267 268 269 270
};

/*
 * Structure holding the information passed to kmemleak callbacks during the
 * early logging.
 */
struct early_log {
	int op_type;			/* kmemleak operation type */
271
	int min_count;			/* minimum reference count */
272
	const void *ptr;		/* allocated/freed memory block */
273 274 275 276
	union {
		size_t size;		/* memory block size */
		unsigned long excess_ref; /* surplus reference passing */
	};
277 278
	unsigned long trace[MAX_TRACE];	/* stack trace */
	unsigned int trace_len;		/* stack trace length */
279 280 281
};

/* early logging buffer and current position */
282 283 284
static struct early_log
	early_log[CONFIG_DEBUG_KMEMLEAK_EARLY_LOG_SIZE] __initdata;
static int crt_early_log __initdata;
285 286 287 288 289 290

static void kmemleak_disable(void);

/*
 * Print a warning and dump the stack trace.
 */
291
#define kmemleak_warn(x...)	do {		\
292
	pr_warn(x);				\
293
	dump_stack();				\
294
	kmemleak_warning = 1;			\
295 296 297
} while (0)

/*
Lucas De Marchi's avatar
Lucas De Marchi committed
298
 * Macro invoked when a serious kmemleak condition occurred and cannot be
299
 * recovered from. Kmemleak will be disabled and further allocation/freeing
300 301
 * tracing no longer available.
 */
302
#define kmemleak_stop(x...)	do {	\
303 304 305 306
	kmemleak_warn(x);		\
	kmemleak_disable();		\
} while (0)

307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325
#define warn_or_seq_printf(seq, fmt, ...)	do {	\
	if (seq)					\
		seq_printf(seq, fmt, ##__VA_ARGS__);	\
	else						\
		pr_warn(fmt, ##__VA_ARGS__);		\
} while (0)

static void warn_or_seq_hex_dump(struct seq_file *seq, int prefix_type,
				 int rowsize, int groupsize, const void *buf,
				 size_t len, bool ascii)
{
	if (seq)
		seq_hex_dump(seq, HEX_PREFIX, prefix_type, rowsize, groupsize,
			     buf, len, ascii);
	else
		print_hex_dump(KERN_WARNING, pr_fmt(HEX_PREFIX), prefix_type,
			       rowsize, groupsize, buf, len, ascii);
}

326 327 328 329 330 331 332 333 334 335
/*
 * Printing of the objects hex dump to the seq file. The number of lines to be
 * printed is limited to HEX_MAX_LINES to prevent seq file spamming. The
 * actual number of printed bytes depends on HEX_ROW_SIZE. It must be called
 * with the object->lock held.
 */
static void hex_dump_object(struct seq_file *seq,
			    struct kmemleak_object *object)
{
	const u8 *ptr = (const u8 *)object->pointer;
336
	size_t len;
337 338

	/* limit the number of lines to HEX_MAX_LINES */
339
	len = min_t(size_t, object->size, HEX_MAX_LINES * HEX_ROW_SIZE);
340

341
	warn_or_seq_printf(seq, "  hex dump (first %zu bytes):\n", len);
342
	kasan_disable_current();
343 344
	warn_or_seq_hex_dump(seq, DUMP_PREFIX_NONE, HEX_ROW_SIZE,
			     HEX_GROUP_SIZE, ptr, len, HEX_ASCII);
345
	kasan_enable_current();
346 347
}

348 349 350 351 352 353 354 355 356 357
/*
 * Object colors, encoded with count and min_count:
 * - white - orphan object, not enough references to it (count < min_count)
 * - gray  - not orphan, not marked as false positive (min_count == 0) or
 *		sufficient references to it (count >= min_count)
 * - black - ignore, it doesn't contain references (e.g. text section)
 *		(min_count == -1). No function defined for this color.
 * Newly created objects don't have any color assigned (object->count == -1)
 * before the next memory scan when they become white.
 */
358
static bool color_white(const struct kmemleak_object *object)
359
{
360 361
	return object->count != KMEMLEAK_BLACK &&
		object->count < object->min_count;
362 363
}

364
static bool color_gray(const struct kmemleak_object *object)
365
{
366 367
	return object->min_count != KMEMLEAK_BLACK &&
		object->count >= object->min_count;
368 369 370 371 372 373 374
}

/*
 * Objects are considered unreferenced only if their color is white, they have
 * not be deleted and have a minimum age to avoid false positives caused by
 * pointers temporarily stored in CPU registers.
 */
375
static bool unreferenced_object(struct kmemleak_object *object)
376
{
377
	return (color_white(object) && object->flags & OBJECT_ALLOCATED) &&
378 379
		time_before_eq(object->jiffies + jiffies_min_age,
			       jiffies_last_scan);
380 381 382
}

/*
383 384
 * Printing of the unreferenced objects information to the seq file. The
 * print_unreferenced function must be called with the object->lock held.
385 386 387 388 389
 */
static void print_unreferenced(struct seq_file *seq,
			       struct kmemleak_object *object)
{
	int i;
390
	unsigned int msecs_age = jiffies_to_msecs(jiffies - object->jiffies);
391

392
	warn_or_seq_printf(seq, "unreferenced object 0x%08lx (size %zu):\n",
393
		   object->pointer, object->size);
394
	warn_or_seq_printf(seq, "  comm \"%s\", pid %d, jiffies %lu (age %d.%03ds)\n",
395 396
		   object->comm, object->pid, object->jiffies,
		   msecs_age / 1000, msecs_age % 1000);
397
	hex_dump_object(seq, object);
398
	warn_or_seq_printf(seq, "  backtrace:\n");
399 400 401

	for (i = 0; i < object->trace_len; i++) {
		void *ptr = (void *)object->trace[i];
402
		warn_or_seq_printf(seq, "    [<%p>] %pS\n", ptr, ptr);
403 404 405 406 407 408 409 410 411 412 413 414 415 416 417
	}
}

/*
 * Print the kmemleak_object information. This function is used mainly for
 * debugging special cases when kmemleak operations. It must be called with
 * the object->lock held.
 */
static void dump_object_info(struct kmemleak_object *object)
{
	struct stack_trace trace;

	trace.nr_entries = object->trace_len;
	trace.entries = object->trace;

Joe Perches's avatar
Joe Perches committed
418
	pr_notice("Object 0x%08lx (size %zu):\n",
419
		  object->pointer, object->size);
420 421 422 423
	pr_notice("  comm \"%s\", pid %d, jiffies %lu\n",
		  object->comm, object->pid, object->jiffies);
	pr_notice("  min_count = %d\n", object->min_count);
	pr_notice("  count = %d\n", object->count);
424
	pr_notice("  flags = 0x%x\n", object->flags);
425
	pr_notice("  checksum = %u\n", object->checksum);
426 427 428 429 430
	pr_notice("  backtrace:\n");
	print_stack_trace(&trace, 4);
}

/*
431
 * Look-up a memory block metadata (kmemleak_object) in the object search
432 433 434 435 436 437
 * tree based on a pointer value. If alias is 0, only values pointing to the
 * beginning of the memory block are allowed. The kmemleak_lock must be held
 * when calling this function.
 */
static struct kmemleak_object *lookup_object(unsigned long ptr, int alias)
{
438 439 440 441 442 443 444 445 446 447 448 449
	struct rb_node *rb = object_tree_root.rb_node;

	while (rb) {
		struct kmemleak_object *object =
			rb_entry(rb, struct kmemleak_object, rb_node);
		if (ptr < object->pointer)
			rb = object->rb_node.rb_left;
		else if (object->pointer + object->size <= ptr)
			rb = object->rb_node.rb_right;
		else if (object->pointer == ptr || alias)
			return object;
		else {
450 451
			kmemleak_warn("Found object by alias at 0x%08lx\n",
				      ptr);
452
			dump_object_info(object);
453
			break;
454
		}
455 456
	}
	return NULL;
457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474
}

/*
 * Increment the object use_count. Return 1 if successful or 0 otherwise. Note
 * that once an object's use_count reached 0, the RCU freeing was already
 * registered and the object should no longer be used. This function must be
 * called under the protection of rcu_read_lock().
 */
static int get_object(struct kmemleak_object *object)
{
	return atomic_inc_not_zero(&object->use_count);
}

/*
 * RCU callback to free a kmemleak_object.
 */
static void free_object_rcu(struct rcu_head *rcu)
{
475
	struct hlist_node *tmp;
476 477 478 479 480 481 482 483
	struct kmemleak_scan_area *area;
	struct kmemleak_object *object =
		container_of(rcu, struct kmemleak_object, rcu);

	/*
	 * Once use_count is 0 (guaranteed by put_object), there is no other
	 * code accessing this object, hence no need for locking.
	 */
484 485
	hlist_for_each_entry_safe(area, tmp, &object->area_list, node) {
		hlist_del(&area->node);
486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509
		kmem_cache_free(scan_area_cache, area);
	}
	kmem_cache_free(object_cache, object);
}

/*
 * Decrement the object use_count. Once the count is 0, free the object using
 * an RCU callback. Since put_object() may be called via the kmemleak_free() ->
 * delete_object() path, the delayed RCU freeing ensures that there is no
 * recursive call to the kernel allocator. Lock-less RCU object_list traversal
 * is also possible.
 */
static void put_object(struct kmemleak_object *object)
{
	if (!atomic_dec_and_test(&object->use_count))
		return;

	/* should only get here after delete_object was called */
	WARN_ON(object->flags & OBJECT_ALLOCATED);

	call_rcu(&object->rcu, free_object_rcu);
}

/*
510
 * Look up an object in the object search tree and increase its use_count.
511 512 513 514
 */
static struct kmemleak_object *find_and_get_object(unsigned long ptr, int alias)
{
	unsigned long flags;
515
	struct kmemleak_object *object;
516 517 518

	rcu_read_lock();
	read_lock_irqsave(&kmemleak_lock, flags);
519
	object = lookup_object(ptr, alias);
520 521 522 523 524 525 526 527 528 529
	read_unlock_irqrestore(&kmemleak_lock, flags);

	/* check whether the object is still available */
	if (object && !get_object(object))
		object = NULL;
	rcu_read_unlock();

	return object;
}

530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550
/*
 * Look up an object in the object search tree and remove it from both
 * object_tree_root and object_list. The returned object's use_count should be
 * at least 1, as initially set by create_object().
 */
static struct kmemleak_object *find_and_remove_object(unsigned long ptr, int alias)
{
	unsigned long flags;
	struct kmemleak_object *object;

	write_lock_irqsave(&kmemleak_lock, flags);
	object = lookup_object(ptr, alias);
	if (object) {
		rb_erase(&object->rb_node, &object_tree_root);
		list_del_rcu(&object->object_list);
	}
	write_unlock_irqrestore(&kmemleak_lock, flags);

	return object;
}

551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566
/*
 * Save stack trace to the given array of MAX_TRACE size.
 */
static int __save_stack_trace(unsigned long *trace)
{
	struct stack_trace stack_trace;

	stack_trace.max_entries = MAX_TRACE;
	stack_trace.nr_entries = 0;
	stack_trace.entries = trace;
	stack_trace.skip = 2;
	save_stack_trace(&stack_trace);

	return stack_trace.nr_entries;
}

567 568 569 570
/*
 * Create the metadata (struct kmemleak_object) corresponding to an allocated
 * memory block and add it to the object_list and object_tree_root.
 */
571 572
static struct kmemleak_object *create_object(unsigned long ptr, size_t size,
					     int min_count, gfp_t gfp)
573 574
{
	unsigned long flags;
575 576
	struct kmemleak_object *object, *parent;
	struct rb_node **link, *rb_parent;
577

578
	object = kmem_cache_alloc(object_cache, gfp_kmemleak_mask(gfp));
579
	if (!object) {
580
		pr_warn("Cannot allocate a kmemleak_object structure\n");
581
		kmemleak_disable();
582
		return NULL;
583 584 585 586 587 588 589
	}

	INIT_LIST_HEAD(&object->object_list);
	INIT_LIST_HEAD(&object->gray_list);
	INIT_HLIST_HEAD(&object->area_list);
	spin_lock_init(&object->lock);
	atomic_set(&object->use_count, 1);
590
	object->flags = OBJECT_ALLOCATED;
591 592
	object->pointer = ptr;
	object->size = size;
593
	object->excess_ref = 0;
594
	object->min_count = min_count;
595
	object->count = 0;			/* white color initially */
596
	object->jiffies = jiffies;
597
	object->checksum = 0;
598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617

	/* task information */
	if (in_irq()) {
		object->pid = 0;
		strncpy(object->comm, "hardirq", sizeof(object->comm));
	} else if (in_softirq()) {
		object->pid = 0;
		strncpy(object->comm, "softirq", sizeof(object->comm));
	} else {
		object->pid = current->pid;
		/*
		 * There is a small chance of a race with set_task_comm(),
		 * however using get_task_comm() here may cause locking
		 * dependency issues with current->alloc_lock. In the worst
		 * case, the command line is not correct.
		 */
		strncpy(object->comm, current->comm, sizeof(object->comm));
	}

	/* kernel backtrace */
618
	object->trace_len = __save_stack_trace(object->trace);
619 620

	write_lock_irqsave(&kmemleak_lock, flags);
621

622 623
	min_addr = min(min_addr, ptr);
	max_addr = max(max_addr, ptr + size);
624 625 626 627 628 629 630 631 632 633
	link = &object_tree_root.rb_node;
	rb_parent = NULL;
	while (*link) {
		rb_parent = *link;
		parent = rb_entry(rb_parent, struct kmemleak_object, rb_node);
		if (ptr + size <= parent->pointer)
			link = &parent->rb_node.rb_left;
		else if (parent->pointer + parent->size <= ptr)
			link = &parent->rb_node.rb_right;
		else {
Joe Perches's avatar
Joe Perches committed
634
			kmemleak_stop("Cannot insert 0x%lx into the object search tree (overlaps existing)\n",
635
				      ptr);
636 637 638 639 640
			/*
			 * No need for parent->lock here since "parent" cannot
			 * be freed while the kmemleak_lock is held.
			 */
			dump_object_info(parent);
641
			kmem_cache_free(object_cache, object);
642
			object = NULL;
643 644
			goto out;
		}
645
	}
646 647 648
	rb_link_node(&object->rb_node, rb_parent, link);
	rb_insert_color(&object->rb_node, &object_tree_root);

649 650 651
	list_add_tail_rcu(&object->object_list, &object_list);
out:
	write_unlock_irqrestore(&kmemleak_lock, flags);
652
	return object;
653 654 655
}

/*
656
 * Mark the object as not allocated and schedule RCU freeing via put_object().
657
 */
658
static void __delete_object(struct kmemleak_object *object)
659 660 661 662
{
	unsigned long flags;

	WARN_ON(!(object->flags & OBJECT_ALLOCATED));
663
	WARN_ON(atomic_read(&object->use_count) < 1);
664 665 666 667 668 669 670 671 672 673 674

	/*
	 * Locking here also ensures that the corresponding memory block
	 * cannot be freed when it is being scanned.
	 */
	spin_lock_irqsave(&object->lock, flags);
	object->flags &= ~OBJECT_ALLOCATED;
	spin_unlock_irqrestore(&object->lock, flags);
	put_object(object);
}

675 676 677 678 679 680 681 682
/*
 * Look up the metadata (struct kmemleak_object) corresponding to ptr and
 * delete it.
 */
static void delete_object_full(unsigned long ptr)
{
	struct kmemleak_object *object;

683
	object = find_and_remove_object(ptr, 0);
684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703
	if (!object) {
#ifdef DEBUG
		kmemleak_warn("Freeing unknown object at 0x%08lx\n",
			      ptr);
#endif
		return;
	}
	__delete_object(object);
}

/*
 * Look up the metadata (struct kmemleak_object) corresponding to ptr and
 * delete it. If the memory block is partially freed, the function may create
 * additional metadata for the remaining parts of the block.
 */
static void delete_object_part(unsigned long ptr, size_t size)
{
	struct kmemleak_object *object;
	unsigned long start, end;

704
	object = find_and_remove_object(ptr, 1);
705 706
	if (!object) {
#ifdef DEBUG
Joe Perches's avatar
Joe Perches committed
707 708
		kmemleak_warn("Partially freeing unknown object at 0x%08lx (size %zu)\n",
			      ptr, size);
709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728
#endif
		return;
	}

	/*
	 * Create one or two objects that may result from the memory block
	 * split. Note that partial freeing is only done by free_bootmem() and
	 * this happens before kmemleak_init() is called. The path below is
	 * only executed during early log recording in kmemleak_init(), so
	 * GFP_KERNEL is enough.
	 */
	start = object->pointer;
	end = object->pointer + object->size;
	if (ptr > start)
		create_object(start, ptr - start, object->min_count,
			      GFP_KERNEL);
	if (ptr + size < end)
		create_object(ptr + size, end - ptr - size, object->min_count,
			      GFP_KERNEL);

729
	__delete_object(object);
730
}
731 732 733 734 735 736 737 738 739

static void __paint_it(struct kmemleak_object *object, int color)
{
	object->min_count = color;
	if (color == KMEMLEAK_BLACK)
		object->flags |= OBJECT_NO_SCAN;
}

static void paint_it(struct kmemleak_object *object, int color)
740 741
{
	unsigned long flags;
742 743 744 745 746 747 748 749

	spin_lock_irqsave(&object->lock, flags);
	__paint_it(object, color);
	spin_unlock_irqrestore(&object->lock, flags);
}

static void paint_ptr(unsigned long ptr, int color)
{
750 751 752 753
	struct kmemleak_object *object;

	object = find_and_get_object(ptr, 0);
	if (!object) {
Joe Perches's avatar
Joe Perches committed
754 755
		kmemleak_warn("Trying to color unknown object at 0x%08lx as %s\n",
			      ptr,
756 757
			      (color == KMEMLEAK_GREY) ? "Grey" :
			      (color == KMEMLEAK_BLACK) ? "Black" : "Unknown");
758 759
		return;
	}
760
	paint_it(object, color);
761 762 763
	put_object(object);
}

764
/*
765
 * Mark an object permanently as gray-colored so that it can no longer be
766 767 768 769 770 771 772
 * reported as a leak. This is used in general to mark a false positive.
 */
static void make_gray_object(unsigned long ptr)
{
	paint_ptr(ptr, KMEMLEAK_GREY);
}

773 774 775 776 777 778
/*
 * Mark the object as black-colored so that it is ignored from scans and
 * reporting.
 */
static void make_black_object(unsigned long ptr)
{
779
	paint_ptr(ptr, KMEMLEAK_BLACK);
780 781 782 783 784 785
}

/*
 * Add a scanning area to the object. If at least one such area is added,
 * kmemleak will only scan these ranges rather than the whole memory block.
 */
786
static void add_scan_area(unsigned long ptr, size_t size, gfp_t gfp)
787 788 789 790 791
{
	unsigned long flags;
	struct kmemleak_object *object;
	struct kmemleak_scan_area *area;

792
	object = find_and_get_object(ptr, 1);
793
	if (!object) {
Joe Perches's avatar
Joe Perches committed
794 795
		kmemleak_warn("Adding scan area to unknown object at 0x%08lx\n",
			      ptr);
796 797 798
		return;
	}

799
	area = kmem_cache_alloc(scan_area_cache, gfp_kmemleak_mask(gfp));
800
	if (!area) {
801
		pr_warn("Cannot allocate a scan area\n");
802 803 804 805
		goto out;
	}

	spin_lock_irqsave(&object->lock, flags);
806 807 808
	if (size == SIZE_MAX) {
		size = object->pointer + object->size - ptr;
	} else if (ptr + size > object->pointer + object->size) {
Joe Perches's avatar
Joe Perches committed
809
		kmemleak_warn("Scan area larger than object 0x%08lx\n", ptr);
810 811 812 813 814 815
		dump_object_info(object);
		kmem_cache_free(scan_area_cache, area);
		goto out_unlock;
	}

	INIT_HLIST_NODE(&area->node);
816 817
	area->start = ptr;
	area->size = size;
818 819 820 821 822 823 824 825

	hlist_add_head(&area->node, &object->area_list);
out_unlock:
	spin_unlock_irqrestore(&object->lock, flags);
out:
	put_object(object);
}

826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849
/*
 * Any surplus references (object already gray) to 'ptr' are passed to
 * 'excess_ref'. This is used in the vmalloc() case where a pointer to
 * vm_struct may be used as an alternative reference to the vmalloc'ed object
 * (see free_thread_stack()).
 */
static void object_set_excess_ref(unsigned long ptr, unsigned long excess_ref)
{
	unsigned long flags;
	struct kmemleak_object *object;

	object = find_and_get_object(ptr, 0);
	if (!object) {
		kmemleak_warn("Setting excess_ref on unknown object at 0x%08lx\n",
			      ptr);
		return;
	}

	spin_lock_irqsave(&object->lock, flags);
	object->excess_ref = excess_ref;
	spin_unlock_irqrestore(&object->lock, flags);
	put_object(object);
}

850 851 852 853 854 855 856 857 858 859 860 861
/*
 * Set the OBJECT_NO_SCAN flag for the object corresponding to the give
 * pointer. Such object will not be scanned by kmemleak but references to it
 * are searched.
 */
static void object_no_scan(unsigned long ptr)
{
	unsigned long flags;
	struct kmemleak_object *object;

	object = find_and_get_object(ptr, 0);
	if (!object) {
Joe Perches's avatar
Joe Perches committed
862
		kmemleak_warn("Not scanning unknown object at 0x%08lx\n", ptr);
863 864 865 866 867 868 869 870 871 872 873 874 875
		return;
	}

	spin_lock_irqsave(&object->lock, flags);
	object->flags |= OBJECT_NO_SCAN;
	spin_unlock_irqrestore(&object->lock, flags);
	put_object(object);
}

/*
 * Log an early kmemleak_* call to the early_log buffer. These calls will be
 * processed later once kmemleak is fully initialized.
 */
876
static void __init log_early(int op_type, const void *ptr, size_t size,
877
			     int min_count)
878 879 880 881
{
	unsigned long flags;
	struct early_log *log;

882
	if (kmemleak_error) {
883 884 885 886 887
		/* kmemleak stopped recording, just count the requests */
		crt_early_log++;
		return;
	}

888
	if (crt_early_log >= ARRAY_SIZE(early_log)) {
889
		crt_early_log++;
890
		kmemleak_disable();
891 892 893 894 895 896 897 898 899 900 901 902 903
		return;
	}

	/*
	 * There is no need for locking since the kernel is still in UP mode
	 * at this stage. Disabling the IRQs is enough.
	 */
	local_irq_save(flags);
	log = &early_log[crt_early_log];
	log->op_type = op_type;
	log->ptr = ptr;
	log->size = size;
	log->min_count = min_count;
904
	log->trace_len = __save_stack_trace(log->trace);
905 906 907 908
	crt_early_log++;
	local_irq_restore(flags);
}

909 910 911 912 913 914 915 916 917
/*
 * Log an early allocated block and populate the stack trace.
 */
static void early_alloc(struct early_log *log)
{
	struct kmemleak_object *object;
	unsigned long flags;
	int i;

918
	if (!kmemleak_enabled || !log->ptr || IS_ERR(log->ptr))
919 920 921 922 923 924 925
		return;

	/*
	 * RCU locking needed to ensure object is not freed via put_object().
	 */
	rcu_read_lock();
	object = create_object((unsigned long)log->ptr, log->size,
926
			       log->min_count, GFP_ATOMIC);
927 928
	if (!object)
		goto out;
929 930 931 932 933
	spin_lock_irqsave(&object->lock, flags);
	for (i = 0; i < log->trace_len; i++)
		object->trace[i] = log->trace[i];
	object->trace_len = log->trace_len;
	spin_unlock_irqrestore(&object->lock, flags);
934
out:
935 936 937
	rcu_read_unlock();
}

938 939 940 941 942 943 944 945 946 947 948 949 950 951
/*
 * Log an early allocated block and populate the stack trace.
 */
static void early_alloc_percpu(struct early_log *log)
{
	unsigned int cpu;
	const void __percpu *ptr = log->ptr;

	for_each_possible_cpu(cpu) {
		log->ptr = per_cpu_ptr(ptr, cpu);
		early_alloc(log);
	}
}

952 953 954 955 956 957 958 959 960 961 962 963
/**
 * kmemleak_alloc - register a newly allocated object
 * @ptr:	pointer to beginning of the object
 * @size:	size of the object
 * @min_count:	minimum number of references to this object. If during memory
 *		scanning a number of references less than @min_count is found,
 *		the object is reported as a memory leak. If @min_count is 0,
 *		the object is never reported as a leak. If @min_count is -1,
 *		the object is ignored (not scanned and not reported as a leak)
 * @gfp:	kmalloc() flags used for kmemleak internal memory allocations
 *
 * This function is called from the kernel allocators when a new object
964
 * (memory block) is allocated (kmem_cache_alloc, kmalloc etc.).
965
 */
966 967
void __ref kmemleak_alloc(const void *ptr, size_t size, int min_count,
			  gfp_t gfp)
968 969 970
{
	pr_debug("%s(0x%p, %zu, %d)\n", __func__, ptr, size, min_count);

971
	if (kmemleak_enabled && ptr && !IS_ERR(ptr))
972
		create_object((unsigned long)ptr, size, min_count, gfp);
973
	else if (kmemleak_early_log)
974
		log_early(KMEMLEAK_ALLOC, ptr, size, min_count);
975 976 977
}
EXPORT_SYMBOL_GPL(kmemleak_alloc);

978 979 980 981
/**
 * kmemleak_alloc_percpu - register a newly allocated __percpu object
 * @ptr:	__percpu pointer to beginning of the object
 * @size:	size of the object
982
 * @gfp:	flags used for kmemleak internal memory allocations
983 984
 *
 * This function is called from the kernel percpu allocator when a new object
985
 * (memory block) is allocated (alloc_percpu).
986
 */
987 988
void __ref kmemleak_alloc_percpu(const void __percpu *ptr, size_t size,
				 gfp_t gfp)
989 990 991 992 993 994 995 996 997
{
	unsigned int cpu;

	pr_debug("%s(0x%p, %zu)\n", __func__, ptr, size);

	/*
	 * Percpu allocations are only scanned and not reported as leaks
	 * (min_count is set to 0).
	 */
998
	if (kmemleak_enabled && ptr && !IS_ERR(ptr))
999 1000
		for_each_possible_cpu(cpu)
			create_object((unsigned long)per_cpu_ptr(ptr, cpu),
1001
				      size, 0, gfp);
1002
	else if (kmemleak_early_log)
1003 1004 1005 1006
		log_early(KMEMLEAK_ALLOC_PERCPU, ptr, size, 0);
}
EXPORT_SYMBOL_GPL(kmemleak_alloc_percpu);

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
/**
 * kmemleak_vmalloc - register a newly vmalloc'ed object
 * @area:	pointer to vm_struct
 * @size:	size of the object
 * @gfp:	__vmalloc() flags used for kmemleak internal memory allocations
 *
 * This function is called from the vmalloc() kernel allocator when a new
 * object (memory block) is allocated.
 */
void __ref kmemleak_vmalloc(const struct vm_struct *area, size_t size, gfp_t gfp)
{
	pr_debug("%s(0x%p, %zu)\n", __func__, area, size);

	/*
	 * A min_count = 2 is needed because vm_struct contains a reference to
	 * the virtual address of the vmalloc'ed block.
	 */
	if (kmemleak_enabled) {
		create_object((unsigned long)area->addr, size, 2, gfp);
		object_set_excess_ref((unsigned long)area,
				      (unsigned long)area->addr);
	} else if (kmemleak_early_log) {
		log_early(KMEMLEAK_ALLOC, area->addr, size, 2);
		/* reusing early_log.size for storing area->addr */
		log_early(KMEMLEAK_SET_EXCESS_REF,
			  area, (unsigned long)area->addr, 0);
	}
}
EXPORT_SYMBOL_GPL(kmemleak_vmalloc);

1037 1038 1039 1040 1041 1042
/**
 * kmemleak_free - unregister a previously registered object
 * @ptr:	pointer to beginning of the object
 *
 * This function is called from the kernel allocators when an object (memory
 * block) is freed (kmem_cache_free, kfree, vfree etc.).
1043
 */
1044
void __ref kmemleak_free(const void *ptr)
1045 1046 1047
{
	pr_debug("%s(0x%p)\n", __func__, ptr);

1048
	if (kmemleak_free_enabled && ptr && !IS_ERR(ptr))
1049
		delete_object_full((unsigned long)ptr);
1050
	else if (kmemleak_early_log)
1051
		log_early(KMEMLEAK_FREE, ptr, 0, 0);
1052 1053 1054
}
EXPORT_SYMBOL_GPL(kmemleak_free);

1055 1056 1057 1058 1059 1060 1061 1062
/**
 * kmemleak_free_part - partially unregister a previously registered object
 * @ptr:	pointer to the beginning or inside the object. This also
 *		represents the start of the range to be freed
 * @size:	size to be unregistered
 *
 * This function is called when only a part of a memory block is freed
 * (usually from the bootmem allocator).
1063
 */
1064
void __ref kmemleak_free_part(const void *ptr, size_t size)
1065 1066 1067
{
	pr_debug("%s(0x%p)\n", __func__, ptr);

1068
	if (kmemleak_enabled && ptr && !IS_ERR(ptr))
1069
		delete_object_part((unsigned long)ptr, size);
1070
	else if (kmemleak_early_log)
1071
		log_early(KMEMLEAK_FREE_PART, ptr, size, 0);
1072 1073 1074
}
EXPORT_SYMBOL_GPL(kmemleak_free_part);

1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087
/**
 * kmemleak_free_percpu - unregister a previously registered __percpu object
 * @ptr:	__percpu pointer to beginning of the object
 *
 * This function is called from the kernel percpu allocator when an object
 * (memory block) is freed (free_percpu).
 */
void __ref kmemleak_free_percpu(const void __percpu *ptr)
{
	unsigned int cpu;

	pr_debug("%s(0x%p)\n", __func__, ptr);

1088
	if (kmemleak_free_enabled && ptr && !IS_ERR(ptr))
1089 1090 1091
		for_each_possible_cpu(cpu)
			delete_object_full((unsigned long)per_cpu_ptr(ptr,
								      cpu));
1092
	else if (kmemleak_early_log)
1093 1094 1095 1096
		log_early(KMEMLEAK_FREE_PERCPU, ptr, 0, 0);
}
EXPORT_SYMBOL_GPL(kmemleak_free_percpu);

1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130
/**
 * kmemleak_update_trace - update object allocation stack trace
 * @ptr:	pointer to beginning of the object
 *
 * Override the object allocation stack trace for cases where the actual
 * allocation place is not always useful.
 */
void __ref kmemleak_update_trace(const void *ptr)
{
	struct kmemleak_object *object;
	unsigned long flags;

	pr_debug("%s(0x%p)\n", __func__, ptr);

	if (!kmemleak_enabled || IS_ERR_OR_NULL(ptr))
		return;

	object = find_and_get_object((unsigned long)ptr, 1);
	if (!object) {
#ifdef DEBUG
		kmemleak_warn("Updating stack trace for unknown object at %p\n",
			      ptr);
#endif
		return;
	}

	spin_lock_irqsave(&object->lock, flags);
	object->trace_len = __save_stack_trace(object->trace);
	spin_unlock_irqrestore(&object->lock, flags);

	put_object(object);
}
EXPORT_SYMBOL(kmemleak_update_trace);

1131 1132 1133 1134 1135 1136
/**
 * kmemleak_not_leak - mark an allocated object as false positive
 * @ptr:	pointer to beginning of the object
 *
 * Calling this function on an object will cause the memory block to no longer
 * be reported as leak and always be scanned.
1137
 */
1138
void __ref kmemleak_not_leak(const void *ptr)
1139 1140 1141
{
	pr_debug("%s(0x%p)\n", __func__, ptr);

1142
	if (kmemleak_enabled && ptr && !IS_ERR(ptr))
1143
		make_gray_object((unsigned long)ptr);
1144
	else if (kmemleak_early_log)
1145
		log_early(KMEMLEAK_NOT_LEAK, ptr, 0, 0);
1146 1147 1148
}
EXPORT_SYMBOL(kmemleak_not_leak);

1149 1150 1151 1152 1153 1154 1155 1156
/**
 * kmemleak_ignore - ignore an allocated object
 * @ptr:	pointer to beginning of the object
 *
 * Calling this function on an object will cause the memory block to be
 * ignored (not scanned and not reported as a leak). This is usually done when
 * it is known that the corresponding block is not a leak and does not contain
 * any references to other allocated memory blocks.
1157
 */
1158
void __ref kmemleak_ignore(const void *ptr)
1159 1160 1161
{
	pr_debug("%s(0x%p)\n", __func__, ptr);

1162
	if (kmemleak_enabled && ptr && !IS_ERR(ptr))
1163
		make_black_object((unsigned long)ptr);
1164
	else if (kmemleak_early_log)
1165
		log_early(KMEMLEAK_IGNORE, ptr, 0, 0);
1166 1167 1168
}
EXPORT_SYMBOL(kmemleak_ignore);

1169 1170 1171 1172 1173 1174 1175 1176 1177 1178
/**
 * kmemleak_scan_area - limit the range to be scanned in an allocated object
 * @ptr:	pointer to beginning or inside the object. This also
 *		represents the start of the scan area
 * @size:	size of the scan area
 * @gfp:	kmalloc() flags used for kmemleak internal memory allocations
 *
 * This function is used when it is known that only certain parts of an object
 * contain references to other objects. Kmemleak will only scan these areas
 * reducing the number false negatives.
1179
 */
1180
void __ref kmemleak_scan_area(const void *ptr, size_t size, gfp_t gfp)
1181 1182 1183
{
	pr_debug("%s(0x%p)\n", __func__, ptr);

1184
	if (kmemleak_enabled && ptr && size && !IS_ERR(ptr))
1185
		add_scan_area((unsigned long)ptr, size, gfp);
1186
	else if (kmemleak_early_log)
1187
		log_early(KMEMLEAK_SCAN_AREA, ptr, size, 0);
1188 1189 1190
}
EXPORT_SYMBOL(kmemleak_scan_area);

1191 1192 1193 1194 1195 1196 1197 1198
/**
 * kmemleak_no_scan - do not scan an allocated object
 * @ptr:	pointer to beginning of the object
 *
 * This function notifies kmemleak not to scan the given memory block. Useful
 * in situations where it is known that the given object does not contain any
 * references to other objects. Kmemleak will not scan such objects reducing
 * the number of false negatives.
1199
 */
1200
void __ref kmemleak_no_scan(const void *ptr)
1201 1202 1203
{
	pr_debug("%s(0x%p)\n", __func__, ptr);

1204
	if (kmemleak_enabled && ptr && !IS_ERR(ptr))
1205
		object_no_scan((unsigned long)ptr);
1206
	else if (kmemleak_early_log)
1207
		log_early(KMEMLEAK_NO_SCAN, ptr, 0, 0);
1208 1209 1210
}
EXPORT_SYMBOL(kmemleak_no_scan);

1211 1212 1213
/**
 * kmemleak_alloc_phys - similar to kmemleak_alloc but taking a physical
 *			 address argument
1214 1215 1216 1217 1218
 * @phys:	physical address of the object
 * @size:	size of the object
 * @min_count:	minimum number of references to this object.
 *              See kmemleak_alloc()
 * @gfp:	kmalloc() flags used for kmemleak internal memory allocations
1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230
 */
void __ref kmemleak_alloc_phys(phys_addr_t phys, size_t size, int min_count,
			       gfp_t gfp)
{
	if (!IS_ENABLED(CONFIG_HIGHMEM) || PHYS_PFN(phys) < max_low_pfn)
		kmemleak_alloc(__va(phys), size, min_count, gfp);
}
EXPORT_SYMBOL(kmemleak_alloc_phys);

/**
 * kmemleak_free_part_phys - similar to kmemleak_free_part but taking a
 *			     physical address argument
1231 1232 1233
 * @phys:	physical address if the beginning or inside an object. This
 *		also represents the start of the range to be freed
 * @size:	size to be unregistered
1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244
 */
void __ref kmemleak_free_part_phys(phys_addr_t phys, size_t size)
{
	if (!IS_ENABLED(CONFIG_HIGHMEM) || PHYS_PFN(phys) < max_low_pfn)
		kmemleak_free_part(__va(phys), size);
}
EXPORT_SYMBOL(kmemleak_free_part_phys);

/**
 * kmemleak_not_leak_phys - similar to kmemleak_not_leak but taking a physical
 *			    address argument
1245
 * @phys:	physical address of the object
1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256
 */
void __ref kmemleak_not_leak_phys(phys_addr_t phys)
{
	if (!IS_ENABLED(CONFIG_HIGHMEM) || PHYS_PFN(phys) < max_low_pfn)
		kmemleak_not_leak(__va(phys));
}
EXPORT_SYMBOL(kmemleak_not_leak_phys);

/**
 * kmemleak_ignore_phys - similar to kmemleak_ignore but taking a physical
 *			  address argument
1257
 * @phys:	physical address of the object
1258 1259 1260 1261 1262 1263 1264 1265
 */
void __ref kmemleak_ignore_phys(phys_addr_t phys)
{
	if (!IS_ENABLED(CONFIG_HIGHMEM) || PHYS_PFN(phys) < max_low_pfn)
		kmemleak_ignore(__va(phys));
}
EXPORT_SYMBOL(kmemleak_ignore_phys);

1266 1267 1268 1269 1270 1271 1272
/*
 * Update an object's checksum and return true if it was modified.
 */
static bool update_checksum(struct kmemleak_object *object)
{
	u32 old_csum = object->checksum;

1273
	kasan_disable_current();
1274
	object->checksum = crc32(0, (void *)object->pointer, object->size);
1275 1276
	kasan_enable_current();

1277 1278 1279
	return object->checksum != old_csum;
}

1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303
/*
 * Update an object's references. object->lock must be held by the caller.
 */
static void update_refs(struct kmemleak_object *object)
{
	if (!color_white(object)) {
		/* non-orphan, ignored or new */
		return;
	}

	/*
	 * Increase the object's reference count (number of pointers to the
	 * memory block). If this count reaches the required minimum, the
	 * object's color will become gray and it will be added to the
	 * gray_list.
	 */
	object->count++;
	if (color_gray(object)) {
		/* put_object() called when removing from gray_list */
		WARN_ON(!get_object(object));
		list_add_tail(&object->gray_list, &gray_list);
	}
}

1304 1305
/*
 * Memory scanning is a long process and it needs to be interruptable. This
Lucas De Marchi's avatar
Lucas De Marchi committed
1306
 * function checks whether such interrupt condition occurred.
1307 1308 1309
 */
static int scan_should_stop(void)
{
1310
	if (!kmemleak_enabled)
1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329
		return 1;

	/*
	 * This function may be called from either process or kthread context,
	 * hence the need to check for both stop conditions.
	 */
	if (current->mm)
		return signal_pending(current);
	else
		return kthread_should_stop();

	return 0;
}

/*
 * Scan a memory block (exclusive range) for valid pointers and add those
 * found to the gray list.
 */
static void scan_block(void *_start, void *_end,
1330
		       struct kmemleak_object *scanned)
1331 1332 1333 1334
{
	unsigned long *ptr;
	unsigned long *start = PTR_ALIGN(_start, BYTES_PER_POINTER);
	unsigned long *end = _end - (BYTES_PER_POINTER - 1);
1335
	unsigned long flags;
1336

1337
	read_lock_irqsave(&kmemleak_lock, flags);
1338 1339
	for (ptr = start; ptr < end; ptr++) {
		struct kmemleak_object *object;
1340
		unsigned long pointer;
1341
		unsigned long excess_ref;
1342 1343 1344 1345

		if (scan_should_stop())
			break;

1346
		kasan_disable_current();
1347
		pointer = *ptr;
1348
		kasan_enable_current();
1349

1350 1351 1352 1353 1354 1355 1356 1357 1358 1359
		if (pointer < min_addr || pointer >= max_addr)
			continue;

		/*
		 * No need for get_object() here since we hold kmemleak_lock.
		 * object->use_count cannot be dropped to 0 while the object
		 * is still present in object_tree_root and object_list
		 * (with updates protected by kmemleak_lock).
		 */
		object = lookup_object(pointer, 1);
1360 1361
		if (!object)
			continue;
1362
		if (object == scanned)
1363 1364 1365 1366 1367 1368 1369 1370
			/* self referenced, ignore */
			continue;

		/*
		 * Avoid the lockdep recursive warning on object->lock being
		 * previously acquired in scan_object(). These locks are
		 * enclosed by scan_mutex.
		 */
1371
		spin_lock_nested(&object->lock, SINGLE_DEPTH_NESTING);
1372 1373 1374 1375 1376 1377 1378 1379
		/* only pass surplus references (object already gray) */
		if (color_gray(object)) {
			excess_ref = object->excess_ref;
			/* no need for update_refs() if object already gray */
		} else {
			excess_ref = 0;
			update_refs(object);
		}
1380
		spin_unlock(&object->lock);
1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392

		if (excess_ref) {
			object = lookup_object(excess_ref, 0);
			if (!object)
				continue;
			if (object == scanned)
				/* circular reference, ignore */
				continue;
			spin_lock_nested(&object->lock, SINGLE_DEPTH_NESTING);
			update_refs(object);
			spin_unlock(&object->lock);
		}
1393 1394 1395