slub.c 109 KB
Newer Older
Christoph Lameter's avatar
Christoph Lameter committed
1
2
3
4
5
6
7
/*
 * SLUB: A slab allocator that limits cache line use instead of queuing
 * objects in per cpu and per node lists.
 *
 * The allocator synchronizes using per slab locks and only
 * uses a centralized lock to manage a pool of partial slabs.
 *
Christoph Lameter's avatar
Christoph Lameter committed
8
 * (C) 2007 SGI, Christoph Lameter
Christoph Lameter's avatar
Christoph Lameter committed
9
10
11
 */

#include <linux/mm.h>
Nick Piggin's avatar
Nick Piggin committed
12
#include <linux/swap.h> /* struct reclaim_state */
Christoph Lameter's avatar
Christoph Lameter committed
13
14
15
16
17
#include <linux/module.h>
#include <linux/bit_spinlock.h>
#include <linux/interrupt.h>
#include <linux/bitops.h>
#include <linux/slab.h>
18
#include <linux/proc_fs.h>
Christoph Lameter's avatar
Christoph Lameter committed
19
#include <linux/seq_file.h>
Vegard Nossum's avatar
Vegard Nossum committed
20
#include <linux/kmemcheck.h>
Christoph Lameter's avatar
Christoph Lameter committed
21
22
23
24
#include <linux/cpu.h>
#include <linux/cpuset.h>
#include <linux/mempolicy.h>
#include <linux/ctype.h>
25
#include <linux/debugobjects.h>
Christoph Lameter's avatar
Christoph Lameter committed
26
#include <linux/kallsyms.h>
27
#include <linux/memory.h>
Roman Zippel's avatar
Roman Zippel committed
28
#include <linux/math64.h>
Akinobu Mita's avatar
Akinobu Mita committed
29
#include <linux/fault-inject.h>
Christoph Lameter's avatar
Christoph Lameter committed
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75

/*
 * Lock order:
 *   1. slab_lock(page)
 *   2. slab->list_lock
 *
 *   The slab_lock protects operations on the object of a particular
 *   slab and its metadata in the page struct. If the slab lock
 *   has been taken then no allocations nor frees can be performed
 *   on the objects in the slab nor can the slab be added or removed
 *   from the partial or full lists since this would mean modifying
 *   the page_struct of the slab.
 *
 *   The list_lock protects the partial and full list on each node and
 *   the partial slab counter. If taken then no new slabs may be added or
 *   removed from the lists nor make the number of partial slabs be modified.
 *   (Note that the total number of slabs is an atomic value that may be
 *   modified without taking the list lock).
 *
 *   The list_lock is a centralized lock and thus we avoid taking it as
 *   much as possible. As long as SLUB does not have to handle partial
 *   slabs, operations can continue without any centralized lock. F.e.
 *   allocating a long series of objects that fill up slabs does not require
 *   the list lock.
 *
 *   The lock order is sometimes inverted when we are trying to get a slab
 *   off a list. We take the list_lock and then look for a page on the list
 *   to use. While we do that objects in the slabs may be freed. We can
 *   only operate on the slab if we have also taken the slab_lock. So we use
 *   a slab_trylock() on the slab. If trylock was successful then no frees
 *   can occur anymore and we can use the slab for allocations etc. If the
 *   slab_trylock() does not succeed then frees are in progress in the slab and
 *   we must stay away from it for a while since we may cause a bouncing
 *   cacheline if we try to acquire the lock. So go onto the next slab.
 *   If all pages are busy then we may allocate a new slab instead of reusing
 *   a partial slab. A new slab has noone operating on it and thus there is
 *   no danger of cacheline contention.
 *
 *   Interrupts are disabled during allocation and deallocation in order to
 *   make the slab allocator safe to use in the context of an irq. In addition
 *   interrupts are disabled to ensure that the processor does not change
 *   while handling per_cpu slabs, due to kernel preemption.
 *
 * SLUB assigns one slab for allocation to each processor.
 * Allocations only occur from these slabs called cpu slabs.
 *
Christoph Lameter's avatar
Christoph Lameter committed
76
77
 * Slabs with free elements are kept on a partial list and during regular
 * operations no list for full slabs is used. If an object in a full slab is
Christoph Lameter's avatar
Christoph Lameter committed
78
 * freed then the slab will show up again on the partial lists.
Christoph Lameter's avatar
Christoph Lameter committed
79
80
 * We track full slabs for debugging purposes though because otherwise we
 * cannot scan all objects.
Christoph Lameter's avatar
Christoph Lameter committed
81
82
83
84
85
86
87
 *
 * Slabs are freed when they become empty. Teardown and setup is
 * minimal so we rely on the page allocators per cpu caches for
 * fast frees and allocs.
 *
 * Overloading of page flags that are otherwise used for LRU management.
 *
88
89
90
91
92
93
94
95
96
97
98
99
 * PageActive 		The slab is frozen and exempt from list processing.
 * 			This means that the slab is dedicated to a purpose
 * 			such as satisfying allocations for a specific
 * 			processor. Objects may be freed in the slab while
 * 			it is frozen but slab_free will then skip the usual
 * 			list operations. It is up to the processor holding
 * 			the slab to integrate the slab into the slab lists
 * 			when the slab is no longer needed.
 *
 * 			One use of this flag is to mark slabs that are
 * 			used for allocations. Then such a slab becomes a cpu
 * 			slab. The cpu slab may be equipped with an additional
100
 * 			freelist that allows lockless access to
101
102
 * 			free objects in addition to the regular freelist
 * 			that requires the slab lock.
Christoph Lameter's avatar
Christoph Lameter committed
103
104
105
 *
 * PageError		Slab requires special handling due to debug
 * 			options set. This moves	slab handling out of
106
 * 			the fast path and disables lockless freelists.
Christoph Lameter's avatar
Christoph Lameter committed
107
108
 */

109
110
111
112
113
#define SLAB_DEBUG_FLAGS (SLAB_RED_ZONE | SLAB_POISON | SLAB_STORE_USER | \
		SLAB_TRACE | SLAB_DEBUG_FREE)

static inline int kmem_cache_debug(struct kmem_cache *s)
{
114
#ifdef CONFIG_SLUB_DEBUG
115
	return unlikely(s->flags & SLAB_DEBUG_FLAGS);
116
#else
117
	return 0;
118
#endif
119
}
120

Christoph Lameter's avatar
Christoph Lameter committed
121
122
123
124
125
126
127
128
129
130
131
/*
 * Issues still to be resolved:
 *
 * - Support PAGE_ALLOC_DEBUG. Should be easy to do.
 *
 * - Variable sizing of the per node arrays
 */

/* Enable to test recovery from slab corruption on boot */
#undef SLUB_RESILIENCY_TEST

132
133
134
135
/*
 * Mininum number of partial slabs. These will be left on the partial
 * lists even if they are empty. kmem_cache_shrink may reclaim them.
 */
136
#define MIN_PARTIAL 5
Christoph Lameter's avatar
Christoph Lameter committed
137

138
139
140
141
142
143
144
/*
 * Maximum number of desirable partial slabs.
 * The existence of more partial slabs makes kmem_cache_shrink
 * sort the partial list by the number of objects in the.
 */
#define MAX_PARTIAL 10

Christoph Lameter's avatar
Christoph Lameter committed
145
146
#define DEBUG_DEFAULT_FLAGS (SLAB_DEBUG_FREE | SLAB_RED_ZONE | \
				SLAB_POISON | SLAB_STORE_USER)
Christoph Lameter's avatar
Christoph Lameter committed
147

148
/*
149
150
151
 * Debugging flags that require metadata to be stored in the slab.  These get
 * disabled when slub_debug=O is used and a cache's min order increases with
 * metadata.
152
 */
153
#define DEBUG_METADATA_FLAGS (SLAB_RED_ZONE | SLAB_POISON | SLAB_STORE_USER)
154

Christoph Lameter's avatar
Christoph Lameter committed
155
156
157
158
/*
 * Set of flags that will prevent slab merging
 */
#define SLUB_NEVER_MERGE (SLAB_RED_ZONE | SLAB_POISON | SLAB_STORE_USER | \
159
160
		SLAB_TRACE | SLAB_DESTROY_BY_RCU | SLAB_NOLEAKTRACE | \
		SLAB_FAILSLAB)
Christoph Lameter's avatar
Christoph Lameter committed
161
162

#define SLUB_MERGE_SAME (SLAB_DEBUG_FREE | SLAB_RECLAIM_ACCOUNT | \
Vegard Nossum's avatar
Vegard Nossum committed
163
		SLAB_CACHE_DMA | SLAB_NOTRACK)
Christoph Lameter's avatar
Christoph Lameter committed
164

165
166
167
168
#define OO_SHIFT	16
#define OO_MASK		((1 << OO_SHIFT) - 1)
#define MAX_OBJS_PER_PAGE	65535 /* since page.objects is u16 */

Christoph Lameter's avatar
Christoph Lameter committed
169
/* Internal SLUB flags */
Christoph Lameter's avatar
Christoph Lameter committed
170
171
#define __OBJECT_POISON		0x80000000UL /* Poison object */
#define __SYSFS_ADD_DEFERRED	0x40000000UL /* Not yet visible via sysfs */
Christoph Lameter's avatar
Christoph Lameter committed
172
173
174
175
176
177
178
179
180
181

static int kmem_size = sizeof(struct kmem_cache);

#ifdef CONFIG_SMP
static struct notifier_block slab_notifier;
#endif

static enum {
	DOWN,		/* No slab functionality available */
	PARTIAL,	/* kmem_cache_open() works but kmalloc does not */
Christoph Lameter's avatar
Christoph Lameter committed
182
	UP,		/* Everything works but does not show up in sysfs */
Christoph Lameter's avatar
Christoph Lameter committed
183
184
185
186
187
	SYSFS		/* Sysfs up */
} slab_state = DOWN;

/* A list of all slab caches on the system */
static DECLARE_RWSEM(slub_lock);
Adrian Bunk's avatar
Adrian Bunk committed
188
static LIST_HEAD(slab_caches);
Christoph Lameter's avatar
Christoph Lameter committed
189

190
191
192
193
/*
 * Tracking user of a slab.
 */
struct track {
194
	unsigned long addr;	/* Called from address */
195
196
197
198
199
200
201
	int cpu;		/* Was running on cpu */
	int pid;		/* Pid context */
	unsigned long when;	/* When did the operation occur */
};

enum track_item { TRACK_ALLOC, TRACK_FREE };

202
#ifdef CONFIG_SLUB_DEBUG
Christoph Lameter's avatar
Christoph Lameter committed
203
204
205
static int sysfs_slab_add(struct kmem_cache *);
static int sysfs_slab_alias(struct kmem_cache *, const char *);
static void sysfs_slab_remove(struct kmem_cache *);
206

Christoph Lameter's avatar
Christoph Lameter committed
207
#else
208
209
210
static inline int sysfs_slab_add(struct kmem_cache *s) { return 0; }
static inline int sysfs_slab_alias(struct kmem_cache *s, const char *p)
							{ return 0; }
211
212
213
214
static inline void sysfs_slab_remove(struct kmem_cache *s)
{
	kfree(s);
}
215

Christoph Lameter's avatar
Christoph Lameter committed
216
217
#endif

218
static inline void stat(struct kmem_cache *s, enum stat_item si)
219
220
{
#ifdef CONFIG_SLUB_STATS
221
	__this_cpu_inc(s->cpu_slab->stat[si]);
222
223
224
#endif
}

Christoph Lameter's avatar
Christoph Lameter committed
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
/********************************************************************
 * 			Core slab cache functions
 *******************************************************************/

int slab_is_available(void)
{
	return slab_state >= UP;
}

static inline struct kmem_cache_node *get_node(struct kmem_cache *s, int node)
{
#ifdef CONFIG_NUMA
	return s->node[node];
#else
	return &s->local_node;
#endif
}

Christoph Lameter's avatar
Christoph Lameter committed
243
/* Verify that a pointer has an address that is valid within a slab page */
244
245
246
247
248
static inline int check_valid_pointer(struct kmem_cache *s,
				struct page *page, const void *object)
{
	void *base;

249
	if (!object)
250
251
		return 1;

252
	base = page_address(page);
253
	if (object < base || object >= base + page->objects * s->size ||
254
255
256
257
258
259
260
		(object - base) % s->size) {
		return 0;
	}

	return 1;
}

261
262
263
264
265
266
267
268
269
270
271
static inline void *get_freepointer(struct kmem_cache *s, void *object)
{
	return *(void **)(object + s->offset);
}

static inline void set_freepointer(struct kmem_cache *s, void *object, void *fp)
{
	*(void **)(object + s->offset) = fp;
}

/* Loop over all objects in a slab */
272
273
#define for_each_object(__p, __s, __addr, __objects) \
	for (__p = (__addr); __p < (__addr) + (__objects) * (__s)->size;\
274
275
276
277
			__p += (__s)->size)

/* Scan freelist */
#define for_each_free_object(__p, __s, __free) \
278
	for (__p = (__free); __p; __p = get_freepointer((__s), __p))
279
280
281
282
283
284
285

/* Determine object index from a given position */
static inline int slab_index(void *p, struct kmem_cache *s, void *addr)
{
	return (p - addr) / s->size;
}

286
287
288
289
static inline struct kmem_cache_order_objects oo_make(int order,
						unsigned long size)
{
	struct kmem_cache_order_objects x = {
290
		(order << OO_SHIFT) + (PAGE_SIZE << order) / size
291
292
293
294
295
296
297
	};

	return x;
}

static inline int oo_order(struct kmem_cache_order_objects x)
{
298
	return x.x >> OO_SHIFT;
299
300
301
302
}

static inline int oo_objects(struct kmem_cache_order_objects x)
{
303
	return x.x & OO_MASK;
304
305
}

306
307
308
309
#ifdef CONFIG_SLUB_DEBUG
/*
 * Debug settings:
 */
310
311
312
#ifdef CONFIG_SLUB_DEBUG_ON
static int slub_debug = DEBUG_DEFAULT_FLAGS;
#else
313
static int slub_debug;
314
#endif
315
316

static char *slub_debug_slabs;
317
static int disable_higher_order_debug;
318

Christoph Lameter's avatar
Christoph Lameter committed
319
320
321
322
323
324
325
326
327
328
329
330
331
/*
 * Object debugging
 */
static void print_section(char *text, u8 *addr, unsigned int length)
{
	int i, offset;
	int newline = 1;
	char ascii[17];

	ascii[16] = 0;

	for (i = 0; i < length; i++) {
		if (newline) {
332
			printk(KERN_ERR "%8s 0x%p: ", text, addr + i);
Christoph Lameter's avatar
Christoph Lameter committed
333
334
			newline = 0;
		}
335
		printk(KERN_CONT " %02x", addr[i]);
Christoph Lameter's avatar
Christoph Lameter committed
336
337
338
		offset = i % 16;
		ascii[offset] = isgraph(addr[i]) ? addr[i] : '.';
		if (offset == 15) {
339
			printk(KERN_CONT " %s\n", ascii);
Christoph Lameter's avatar
Christoph Lameter committed
340
341
342
343
344
345
			newline = 1;
		}
	}
	if (!newline) {
		i %= 16;
		while (i < 16) {
346
			printk(KERN_CONT "   ");
Christoph Lameter's avatar
Christoph Lameter committed
347
348
349
			ascii[i] = ' ';
			i++;
		}
350
		printk(KERN_CONT " %s\n", ascii);
Christoph Lameter's avatar
Christoph Lameter committed
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
	}
}

static struct track *get_track(struct kmem_cache *s, void *object,
	enum track_item alloc)
{
	struct track *p;

	if (s->offset)
		p = object + s->offset + sizeof(void *);
	else
		p = object + s->inuse;

	return p + alloc;
}

static void set_track(struct kmem_cache *s, void *object,
368
			enum track_item alloc, unsigned long addr)
Christoph Lameter's avatar
Christoph Lameter committed
369
{
Akinobu Mita's avatar
Akinobu Mita committed
370
	struct track *p = get_track(s, object, alloc);
Christoph Lameter's avatar
Christoph Lameter committed
371
372
373
374

	if (addr) {
		p->addr = addr;
		p->cpu = smp_processor_id();
375
		p->pid = current->pid;
Christoph Lameter's avatar
Christoph Lameter committed
376
377
378
379
380
381
382
		p->when = jiffies;
	} else
		memset(p, 0, sizeof(struct track));
}

static void init_tracking(struct kmem_cache *s, void *object)
{
383
384
385
	if (!(s->flags & SLAB_STORE_USER))
		return;

386
387
	set_track(s, object, TRACK_FREE, 0UL);
	set_track(s, object, TRACK_ALLOC, 0UL);
Christoph Lameter's avatar
Christoph Lameter committed
388
389
390
391
392
393
394
}

static void print_track(const char *s, struct track *t)
{
	if (!t->addr)
		return;

395
	printk(KERN_ERR "INFO: %s in %pS age=%lu cpu=%u pid=%d\n",
396
		s, (void *)t->addr, jiffies - t->when, t->cpu, t->pid);
397
398
399
400
401
402
403
404
405
406
407
408
409
}

static void print_tracking(struct kmem_cache *s, void *object)
{
	if (!(s->flags & SLAB_STORE_USER))
		return;

	print_track("Allocated", get_track(s, object, TRACK_ALLOC));
	print_track("Freed", get_track(s, object, TRACK_FREE));
}

static void print_page_info(struct page *page)
{
410
411
	printk(KERN_ERR "INFO: Slab 0x%p objects=%u used=%u fp=0x%p flags=0x%04lx\n",
		page, page->objects, page->inuse, page->freelist, page->flags);
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427

}

static void slab_bug(struct kmem_cache *s, char *fmt, ...)
{
	va_list args;
	char buf[100];

	va_start(args, fmt);
	vsnprintf(buf, sizeof(buf), fmt, args);
	va_end(args);
	printk(KERN_ERR "========================================"
			"=====================================\n");
	printk(KERN_ERR "BUG %s: %s\n", s->name, buf);
	printk(KERN_ERR "----------------------------------------"
			"-------------------------------------\n\n");
Christoph Lameter's avatar
Christoph Lameter committed
428
429
}

430
431
432
433
434
435
436
437
438
439
440
441
static void slab_fix(struct kmem_cache *s, char *fmt, ...)
{
	va_list args;
	char buf[100];

	va_start(args, fmt);
	vsnprintf(buf, sizeof(buf), fmt, args);
	va_end(args);
	printk(KERN_ERR "FIX %s: %s\n", s->name, buf);
}

static void print_trailer(struct kmem_cache *s, struct page *page, u8 *p)
Christoph Lameter's avatar
Christoph Lameter committed
442
443
{
	unsigned int off;	/* Offset of last byte */
444
	u8 *addr = page_address(page);
445
446
447
448
449
450
451
452
453
454
455

	print_tracking(s, p);

	print_page_info(page);

	printk(KERN_ERR "INFO: Object 0x%p @offset=%tu fp=0x%p\n\n",
			p, p - addr, get_freepointer(s, p));

	if (p > addr + 16)
		print_section("Bytes b4", p - 16, 16);

456
	print_section("Object", p, min_t(unsigned long, s->objsize, PAGE_SIZE));
Christoph Lameter's avatar
Christoph Lameter committed
457
458
459
460
461
462
463
464
465
466

	if (s->flags & SLAB_RED_ZONE)
		print_section("Redzone", p + s->objsize,
			s->inuse - s->objsize);

	if (s->offset)
		off = s->offset + sizeof(void *);
	else
		off = s->inuse;

467
	if (s->flags & SLAB_STORE_USER)
Christoph Lameter's avatar
Christoph Lameter committed
468
469
470
471
		off += 2 * sizeof(struct track);

	if (off != s->size)
		/* Beginning of the filler is the free pointer */
472
473
474
		print_section("Padding", p + off, s->size - off);

	dump_stack();
Christoph Lameter's avatar
Christoph Lameter committed
475
476
477
478
479
}

static void object_err(struct kmem_cache *s, struct page *page,
			u8 *object, char *reason)
{
480
	slab_bug(s, "%s", reason);
481
	print_trailer(s, page, object);
Christoph Lameter's avatar
Christoph Lameter committed
482
483
}

484
static void slab_err(struct kmem_cache *s, struct page *page, char *fmt, ...)
Christoph Lameter's avatar
Christoph Lameter committed
485
486
487
488
{
	va_list args;
	char buf[100];

489
490
	va_start(args, fmt);
	vsnprintf(buf, sizeof(buf), fmt, args);
Christoph Lameter's avatar
Christoph Lameter committed
491
	va_end(args);
492
	slab_bug(s, "%s", buf);
493
	print_page_info(page);
Christoph Lameter's avatar
Christoph Lameter committed
494
495
496
497
498
499
500
501
502
	dump_stack();
}

static void init_object(struct kmem_cache *s, void *object, int active)
{
	u8 *p = object;

	if (s->flags & __OBJECT_POISON) {
		memset(p, POISON_FREE, s->objsize - 1);
503
		p[s->objsize - 1] = POISON_END;
Christoph Lameter's avatar
Christoph Lameter committed
504
505
506
507
508
509
510
511
	}

	if (s->flags & SLAB_RED_ZONE)
		memset(p + s->objsize,
			active ? SLUB_RED_ACTIVE : SLUB_RED_INACTIVE,
			s->inuse - s->objsize);
}

512
static u8 *check_bytes(u8 *start, unsigned int value, unsigned int bytes)
Christoph Lameter's avatar
Christoph Lameter committed
513
514
515
{
	while (bytes) {
		if (*start != (u8)value)
516
			return start;
Christoph Lameter's avatar
Christoph Lameter committed
517
518
519
		start++;
		bytes--;
	}
520
521
522
523
524
525
526
527
528
529
530
531
	return NULL;
}

static void restore_bytes(struct kmem_cache *s, char *message, u8 data,
						void *from, void *to)
{
	slab_fix(s, "Restoring 0x%p-0x%p=0x%x\n", from, to - 1, data);
	memset(from, data, to - from);
}

static int check_bytes_and_report(struct kmem_cache *s, struct page *page,
			u8 *object, char *what,
532
			u8 *start, unsigned int value, unsigned int bytes)
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
{
	u8 *fault;
	u8 *end;

	fault = check_bytes(start, value, bytes);
	if (!fault)
		return 1;

	end = start + bytes;
	while (end > fault && end[-1] == value)
		end--;

	slab_bug(s, "%s overwritten", what);
	printk(KERN_ERR "INFO: 0x%p-0x%p. First byte 0x%x instead of 0x%x\n",
					fault, end - 1, fault[0], value);
	print_trailer(s, page, object);

	restore_bytes(s, what, value, fault, end);
	return 0;
Christoph Lameter's avatar
Christoph Lameter committed
552
553
554
555
556
557
558
559
560
}

/*
 * Object layout:
 *
 * object address
 * 	Bytes of the object to be managed.
 * 	If the freepointer may overlay the object then the free
 * 	pointer is the first word of the object.
Christoph Lameter's avatar
Christoph Lameter committed
561
 *
Christoph Lameter's avatar
Christoph Lameter committed
562
563
564
565
566
 * 	Poisoning uses 0x6b (POISON_FREE) and the last byte is
 * 	0xa5 (POISON_END)
 *
 * object + s->objsize
 * 	Padding to reach word boundary. This is also used for Redzoning.
Christoph Lameter's avatar
Christoph Lameter committed
567
568
569
 * 	Padding is extended by another word if Redzoning is enabled and
 * 	objsize == inuse.
 *
Christoph Lameter's avatar
Christoph Lameter committed
570
571
572
573
 * 	We fill with 0xbb (RED_INACTIVE) for inactive objects and with
 * 	0xcc (RED_ACTIVE) for objects in use.
 *
 * object + s->inuse
Christoph Lameter's avatar
Christoph Lameter committed
574
575
 * 	Meta data starts here.
 *
Christoph Lameter's avatar
Christoph Lameter committed
576
577
 * 	A. Free pointer (if we cannot overwrite object on free)
 * 	B. Tracking data for SLAB_STORE_USER
Christoph Lameter's avatar
Christoph Lameter committed
578
 * 	C. Padding to reach required alignment boundary or at mininum
Christoph Lameter's avatar
Christoph Lameter committed
579
 * 		one word if debugging is on to be able to detect writes
Christoph Lameter's avatar
Christoph Lameter committed
580
581
582
 * 		before the word boundary.
 *
 *	Padding is done using 0x5a (POISON_INUSE)
Christoph Lameter's avatar
Christoph Lameter committed
583
584
 *
 * object + s->size
Christoph Lameter's avatar
Christoph Lameter committed
585
 * 	Nothing is used beyond s->size.
Christoph Lameter's avatar
Christoph Lameter committed
586
 *
Christoph Lameter's avatar
Christoph Lameter committed
587
588
 * If slabcaches are merged then the objsize and inuse boundaries are mostly
 * ignored. And therefore no slab options that rely on these boundaries
Christoph Lameter's avatar
Christoph Lameter committed
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
 * may be used with merged slabcaches.
 */

static int check_pad_bytes(struct kmem_cache *s, struct page *page, u8 *p)
{
	unsigned long off = s->inuse;	/* The end of info */

	if (s->offset)
		/* Freepointer is placed after the object. */
		off += sizeof(void *);

	if (s->flags & SLAB_STORE_USER)
		/* We also have user information there */
		off += 2 * sizeof(struct track);

	if (s->size == off)
		return 1;

607
608
	return check_bytes_and_report(s, page, p, "Object padding",
				p + off, POISON_INUSE, s->size - off);
Christoph Lameter's avatar
Christoph Lameter committed
609
610
}

611
/* Check the pad bytes at the end of a slab page */
Christoph Lameter's avatar
Christoph Lameter committed
612
613
static int slab_pad_check(struct kmem_cache *s, struct page *page)
{
614
615
616
617
618
	u8 *start;
	u8 *fault;
	u8 *end;
	int length;
	int remainder;
Christoph Lameter's avatar
Christoph Lameter committed
619
620
621
622

	if (!(s->flags & SLAB_POISON))
		return 1;

623
	start = page_address(page);
624
	length = (PAGE_SIZE << compound_order(page));
625
626
	end = start + length;
	remainder = length % s->size;
Christoph Lameter's avatar
Christoph Lameter committed
627
628
629
	if (!remainder)
		return 1;

630
	fault = check_bytes(end - remainder, POISON_INUSE, remainder);
631
632
633
634
635
636
	if (!fault)
		return 1;
	while (end > fault && end[-1] == POISON_INUSE)
		end--;

	slab_err(s, page, "Padding overwritten. 0x%p-0x%p", fault, end - 1);
637
	print_section("Padding", end - remainder, remainder);
638

Eric Dumazet's avatar
Eric Dumazet committed
639
	restore_bytes(s, "slab padding", POISON_INUSE, end - remainder, end);
640
	return 0;
Christoph Lameter's avatar
Christoph Lameter committed
641
642
643
644
645
646
647
648
649
650
651
652
}

static int check_object(struct kmem_cache *s, struct page *page,
					void *object, int active)
{
	u8 *p = object;
	u8 *endobject = object + s->objsize;

	if (s->flags & SLAB_RED_ZONE) {
		unsigned int red =
			active ? SLUB_RED_ACTIVE : SLUB_RED_INACTIVE;

653
654
		if (!check_bytes_and_report(s, page, object, "Redzone",
			endobject, red, s->inuse - s->objsize))
Christoph Lameter's avatar
Christoph Lameter committed
655
656
			return 0;
	} else {
Ingo Molnar's avatar
Ingo Molnar committed
657
658
659
660
		if ((s->flags & SLAB_POISON) && s->objsize < s->inuse) {
			check_bytes_and_report(s, page, p, "Alignment padding",
				endobject, POISON_INUSE, s->inuse - s->objsize);
		}
Christoph Lameter's avatar
Christoph Lameter committed
661
662
663
664
	}

	if (s->flags & SLAB_POISON) {
		if (!active && (s->flags & __OBJECT_POISON) &&
665
666
667
			(!check_bytes_and_report(s, page, p, "Poison", p,
					POISON_FREE, s->objsize - 1) ||
			 !check_bytes_and_report(s, page, p, "Poison",
668
				p + s->objsize - 1, POISON_END, 1)))
Christoph Lameter's avatar
Christoph Lameter committed
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
			return 0;
		/*
		 * check_pad_bytes cleans up on its own.
		 */
		check_pad_bytes(s, page, p);
	}

	if (!s->offset && active)
		/*
		 * Object and freepointer overlap. Cannot check
		 * freepointer while object is allocated.
		 */
		return 1;

	/* Check free pointer validity */
	if (!check_valid_pointer(s, page, get_freepointer(s, p))) {
		object_err(s, page, p, "Freepointer corrupt");
		/*
Nick Andrew's avatar
Nick Andrew committed
687
		 * No choice but to zap it and thus lose the remainder
Christoph Lameter's avatar
Christoph Lameter committed
688
		 * of the free objects in this slab. May cause
Christoph Lameter's avatar
Christoph Lameter committed
689
		 * another error because the object count is now wrong.
Christoph Lameter's avatar
Christoph Lameter committed
690
		 */
691
		set_freepointer(s, p, NULL);
Christoph Lameter's avatar
Christoph Lameter committed
692
693
694
695
696
697
698
		return 0;
	}
	return 1;
}

static int check_slab(struct kmem_cache *s, struct page *page)
{
699
700
	int maxobj;

Christoph Lameter's avatar
Christoph Lameter committed
701
702
703
	VM_BUG_ON(!irqs_disabled());

	if (!PageSlab(page)) {
704
		slab_err(s, page, "Not a valid slab page");
Christoph Lameter's avatar
Christoph Lameter committed
705
706
		return 0;
	}
707
708
709
710
711
712
713
714

	maxobj = (PAGE_SIZE << compound_order(page)) / s->size;
	if (page->objects > maxobj) {
		slab_err(s, page, "objects %u > max %u",
			s->name, page->objects, maxobj);
		return 0;
	}
	if (page->inuse > page->objects) {
715
		slab_err(s, page, "inuse %u > max %u",
716
			s->name, page->inuse, page->objects);
Christoph Lameter's avatar
Christoph Lameter committed
717
718
719
720
721
722
723
724
		return 0;
	}
	/* Slab_pad_check fixes things up after itself */
	slab_pad_check(s, page);
	return 1;
}

/*
Christoph Lameter's avatar
Christoph Lameter committed
725
726
 * Determine if a certain object on a page is on the freelist. Must hold the
 * slab lock to guarantee that the chains are in a consistent state.
Christoph Lameter's avatar
Christoph Lameter committed
727
728
729
730
731
732
 */
static int on_freelist(struct kmem_cache *s, struct page *page, void *search)
{
	int nr = 0;
	void *fp = page->freelist;
	void *object = NULL;
733
	unsigned long max_objects;
Christoph Lameter's avatar
Christoph Lameter committed
734

735
	while (fp && nr <= page->objects) {
Christoph Lameter's avatar
Christoph Lameter committed
736
737
738
739
740
741
		if (fp == search)
			return 1;
		if (!check_valid_pointer(s, page, fp)) {
			if (object) {
				object_err(s, page, object,
					"Freechain corrupt");
742
				set_freepointer(s, object, NULL);
Christoph Lameter's avatar
Christoph Lameter committed
743
744
				break;
			} else {
745
				slab_err(s, page, "Freepointer corrupt");
746
				page->freelist = NULL;
747
				page->inuse = page->objects;
748
				slab_fix(s, "Freelist cleared");
Christoph Lameter's avatar
Christoph Lameter committed
749
750
751
752
753
754
755
756
757
				return 0;
			}
			break;
		}
		object = fp;
		fp = get_freepointer(s, object);
		nr++;
	}

758
	max_objects = (PAGE_SIZE << compound_order(page)) / s->size;
759
760
	if (max_objects > MAX_OBJS_PER_PAGE)
		max_objects = MAX_OBJS_PER_PAGE;
761
762
763
764
765
766
767

	if (page->objects != max_objects) {
		slab_err(s, page, "Wrong number of objects. Found %d but "
			"should be %d", page->objects, max_objects);
		page->objects = max_objects;
		slab_fix(s, "Number of objects adjusted.");
	}
768
	if (page->inuse != page->objects - nr) {
769
		slab_err(s, page, "Wrong object count. Counter is %d but "
770
771
			"counted were %d", page->inuse, page->objects - nr);
		page->inuse = page->objects - nr;
772
		slab_fix(s, "Object count adjusted.");
Christoph Lameter's avatar
Christoph Lameter committed
773
774
775
776
	}
	return search == NULL;
}

777
778
static void trace(struct kmem_cache *s, struct page *page, void *object,
								int alloc)
Christoph Lameter's avatar
Christoph Lameter committed
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
{
	if (s->flags & SLAB_TRACE) {
		printk(KERN_INFO "TRACE %s %s 0x%p inuse=%d fp=0x%p\n",
			s->name,
			alloc ? "alloc" : "free",
			object, page->inuse,
			page->freelist);

		if (!alloc)
			print_section("Object", (void *)object, s->objsize);

		dump_stack();
	}
}

794
/*
Christoph Lameter's avatar
Christoph Lameter committed
795
 * Tracking of fully allocated slabs for debugging purposes.
796
 */
Christoph Lameter's avatar
Christoph Lameter committed
797
static void add_full(struct kmem_cache_node *n, struct page *page)
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
{
	spin_lock(&n->list_lock);
	list_add(&page->lru, &n->full);
	spin_unlock(&n->list_lock);
}

static void remove_full(struct kmem_cache *s, struct page *page)
{
	struct kmem_cache_node *n;

	if (!(s->flags & SLAB_STORE_USER))
		return;

	n = get_node(s, page_to_nid(page));

	spin_lock(&n->list_lock);
	list_del(&page->lru);
	spin_unlock(&n->list_lock);
}

818
819
820
821
822
823
824
825
/* Tracking of the number of slabs for debugging purposes */
static inline unsigned long slabs_node(struct kmem_cache *s, int node)
{
	struct kmem_cache_node *n = get_node(s, node);

	return atomic_long_read(&n->nr_slabs);
}

826
827
828
829
830
static inline unsigned long node_nr_slabs(struct kmem_cache_node *n)
{
	return atomic_long_read(&n->nr_slabs);
}

831
static inline void inc_slabs_node(struct kmem_cache *s, int node, int objects)
832
833
834
835
836
837
838
839
840
{
	struct kmem_cache_node *n = get_node(s, node);

	/*
	 * May be called early in order to allocate a slab for the
	 * kmem_cache_node structure. Solve the chicken-egg
	 * dilemma by deferring the increment of the count during
	 * bootstrap (see early_kmem_cache_node_alloc).
	 */
841
	if (!NUMA_BUILD || n) {
842
		atomic_long_inc(&n->nr_slabs);
843
844
		atomic_long_add(objects, &n->total_objects);
	}
845
}
846
static inline void dec_slabs_node(struct kmem_cache *s, int node, int objects)
847
848
849
850
{
	struct kmem_cache_node *n = get_node(s, node);

	atomic_long_dec(&n->nr_slabs);
851
	atomic_long_sub(objects, &n->total_objects);
852
853
854
}

/* Object debug checks for alloc/free paths */
Christoph Lameter's avatar
Christoph Lameter committed
855
856
857
858
859
860
861
862
863
864
865
static void setup_object_debug(struct kmem_cache *s, struct page *page,
								void *object)
{
	if (!(s->flags & (SLAB_STORE_USER|SLAB_RED_ZONE|__OBJECT_POISON)))
		return;

	init_object(s, object, 0);
	init_tracking(s, object);
}

static int alloc_debug_processing(struct kmem_cache *s, struct page *page,
866
					void *object, unsigned long addr)
Christoph Lameter's avatar
Christoph Lameter committed
867
868
869
870
{
	if (!check_slab(s, page))
		goto bad;

871
	if (!on_freelist(s, page, object)) {
872
		object_err(s, page, object, "Object already allocated");
873
		goto bad;
Christoph Lameter's avatar
Christoph Lameter committed
874
875
876
877
	}

	if (!check_valid_pointer(s, page, object)) {
		object_err(s, page, object, "Freelist Pointer check fails");
878
		goto bad;
Christoph Lameter's avatar
Christoph Lameter committed
879
880
	}

881
	if (!check_object(s, page, object, 0))
Christoph Lameter's avatar
Christoph Lameter committed
882
883
		goto bad;

Christoph Lameter's avatar
Christoph Lameter committed
884
885
886
887
888
	/* Success perform special debug activities for allocs */
	if (s->flags & SLAB_STORE_USER)
		set_track(s, object, TRACK_ALLOC, addr);
	trace(s, page, object, 1);
	init_object(s, object, 1);
Christoph Lameter's avatar
Christoph Lameter committed
889
	return 1;
Christoph Lameter's avatar
Christoph Lameter committed
890

Christoph Lameter's avatar
Christoph Lameter committed
891
892
893
894
895
bad:
	if (PageSlab(page)) {
		/*
		 * If this is a slab page then lets do the best we can
		 * to avoid issues in the future. Marking all objects
Christoph Lameter's avatar
Christoph Lameter committed
896
		 * as used avoids touching the remaining objects.
Christoph Lameter's avatar
Christoph Lameter committed
897
		 */
898
		slab_fix(s, "Marking all objects used");
899
		page->inuse = page->objects;
900
		page->freelist = NULL;
Christoph Lameter's avatar
Christoph Lameter committed
901
902
903
904
	}
	return 0;
}

Christoph Lameter's avatar
Christoph Lameter committed
905
static int free_debug_processing(struct kmem_cache *s, struct page *page,
906
					void *object, unsigned long addr)
Christoph Lameter's avatar
Christoph Lameter committed
907
908
909
910
911
{
	if (!check_slab(s, page))
		goto fail;

	if (!check_valid_pointer(s, page, object)) {
912
		slab_err(s, page, "Invalid object pointer 0x%p", object);
Christoph Lameter's avatar
Christoph Lameter committed
913
914
915
916
		goto fail;
	}

	if (on_freelist(s, page, object)) {
917
		object_err(s, page, object, "Object already free");
Christoph Lameter's avatar
Christoph Lameter committed
918
919
920
921
922
923
924
		goto fail;
	}

	if (!check_object(s, page, object, 1))
		return 0;

	if (unlikely(s != page->slab)) {
Ingo Molnar's avatar
Ingo Molnar committed
925
		if (!PageSlab(page)) {
926
927
			slab_err(s, page, "Attempt to free object(0x%p) "
				"outside of slab", object);
Ingo Molnar's avatar
Ingo Molnar committed
928
		} else if (!page->slab) {
Christoph Lameter's avatar
Christoph Lameter committed
929
			printk(KERN_ERR
930
				"SLUB <none>: no slab for object 0x%p.\n",
Christoph Lameter's avatar
Christoph Lameter committed
931
						object);
932
			dump_stack();
933
		} else
934
935
			object_err(s, page, object,
					"page slab pointer corrupt.");
Christoph Lameter's avatar
Christoph Lameter committed
936
937
		goto fail;
	}
Christoph Lameter's avatar
Christoph Lameter committed
938
939

	/* Special debug activities for freeing objects */
940
	if (!PageSlubFrozen(page) && !page->freelist)
Christoph Lameter's avatar
Christoph Lameter committed
941
942
943
944
945
		remove_full(s, page);
	if (s->flags & SLAB_STORE_USER)
		set_track(s, object, TRACK_FREE, addr);
	trace(s, page, object, 0);
	init_object(s, object, 0);
Christoph Lameter's avatar
Christoph Lameter committed
946
	return 1;
Christoph Lameter's avatar
Christoph Lameter committed
947

Christoph Lameter's avatar
Christoph Lameter committed
948
fail:
949
	slab_fix(s, "Object at 0x%p not freed", object);
Christoph Lameter's avatar
Christoph Lameter committed
950
951
952
	return 0;
}

953
954
static int __init setup_slub_debug(char *str)
{
955
956
957
958
959
960
961
962
963
964
965
966
967
968
	slub_debug = DEBUG_DEFAULT_FLAGS;
	if (*str++ != '=' || !*str)
		/*
		 * No options specified. Switch on full debugging.
		 */
		goto out;

	if (*str == ',')
		/*
		 * No options but restriction on slabs. This means full
		 * debugging for slabs matching a pattern.
		 */
		goto check_slabs;

969
970
971
972
973
974
975
976
977
	if (tolower(*str) == 'o') {
		/*
		 * Avoid enabling debugging on caches if its minimum order
		 * would increase as a result.
		 */
		disable_higher_order_debug = 1;
		goto out;
	}

978
979
980
981
982
983
984
985
986
987
	slub_debug = 0;
	if (*str == '-')
		/*
		 * Switch off all debugging measures.
		 */
		goto out;

	/*
	 * Determine which debug features should be switched on
	 */
988
	for (; *str && *str != ','; str++) {
989
990
991
992
993
994
995
996
997
998
999
1000
		switch (tolower(*str)) {
		case 'f':
			slub_debug |= SLAB_DEBUG_FREE;
			break;
		case 'z':
			slub_debug |= SLAB_RED_ZONE;
			break;
		case 'p':
			slub_debug |= SLAB_POISON;
			break;
		case 'u':
			slub_debug |= SLAB_STORE_USER;