slub.c 110 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
#include <trace/events/kmem.h>

Christoph Lameter's avatar
Christoph Lameter committed
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
/*
 * 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
78
79
 * 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
80
 * freed then the slab will show up again on the partial lists.
Christoph Lameter's avatar
Christoph Lameter committed
81
82
 * We track full slabs for debugging purposes though because otherwise we
 * cannot scan all objects.
Christoph Lameter's avatar
Christoph Lameter committed
83
84
85
86
87
88
89
 *
 * 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.
 *
90
91
92
93
94
95
96
97
98
99
100
101
 * 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
102
 * 			freelist that allows lockless access to
103
104
 * 			free objects in addition to the regular freelist
 * 			that requires the slab lock.
Christoph Lameter's avatar
Christoph Lameter committed
105
106
107
 *
 * PageError		Slab requires special handling due to debug
 * 			options set. This moves	slab handling out of
108
 * 			the fast path and disables lockless freelists.
Christoph Lameter's avatar
Christoph Lameter committed
109
110
 */

111
112
113
114
115
#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)
{
116
#ifdef CONFIG_SLUB_DEBUG
117
	return unlikely(s->flags & SLAB_DEBUG_FLAGS);
118
#else
119
	return 0;
120
#endif
121
}
122

Christoph Lameter's avatar
Christoph Lameter committed
123
124
125
126
127
128
129
130
131
132
133
/*
 * 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

134
135
136
137
/*
 * Mininum number of partial slabs. These will be left on the partial
 * lists even if they are empty. kmem_cache_shrink may reclaim them.
 */
138
#define MIN_PARTIAL 5
Christoph Lameter's avatar
Christoph Lameter committed
139

140
141
142
143
144
145
146
/*
 * 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
147
148
#define DEBUG_DEFAULT_FLAGS (SLAB_DEBUG_FREE | SLAB_RED_ZONE | \
				SLAB_POISON | SLAB_STORE_USER)
Christoph Lameter's avatar
Christoph Lameter committed
149

150
/*
151
152
153
 * 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.
154
 */
155
#define DEBUG_METADATA_FLAGS (SLAB_RED_ZONE | SLAB_POISON | SLAB_STORE_USER)
156

Christoph Lameter's avatar
Christoph Lameter committed
157
158
159
160
/*
 * Set of flags that will prevent slab merging
 */
#define SLUB_NEVER_MERGE (SLAB_RED_ZONE | SLAB_POISON | SLAB_STORE_USER | \
161
162
		SLAB_TRACE | SLAB_DESTROY_BY_RCU | SLAB_NOLEAKTRACE | \
		SLAB_FAILSLAB)
Christoph Lameter's avatar
Christoph Lameter committed
163
164

#define SLUB_MERGE_SAME (SLAB_DEBUG_FREE | SLAB_RECLAIM_ACCOUNT | \
Vegard Nossum's avatar
Vegard Nossum committed
165
		SLAB_CACHE_DMA | SLAB_NOTRACK)
Christoph Lameter's avatar
Christoph Lameter committed
166

167
168
169
170
#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
171
/* Internal SLUB flags */
Christoph Lameter's avatar
Christoph Lameter committed
172
#define __OBJECT_POISON		0x80000000UL /* Poison object */
Christoph Lameter's avatar
Christoph Lameter committed
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 */
182
	PARTIAL,	/* Kmem_cache_node works */
Christoph Lameter's avatar
Christoph Lameter committed
183
	UP,		/* Everything works but does not show up in sysfs */
Christoph Lameter's avatar
Christoph Lameter committed
184
185
186
187
188
	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
189
static LIST_HEAD(slab_caches);
Christoph Lameter's avatar
Christoph Lameter committed
190

191
192
193
194
/*
 * Tracking user of a slab.
 */
struct track {
195
	unsigned long addr;	/* Called from address */
196
197
198
199
200
201
202
	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 };

203
#ifdef CONFIG_SYSFS
Christoph Lameter's avatar
Christoph Lameter committed
204
205
206
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 *);
207

Christoph Lameter's avatar
Christoph Lameter committed
208
#else
209
210
211
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; }
212
213
static inline void sysfs_slab_remove(struct kmem_cache *s)
{
214
	kfree(s->name);
215
216
	kfree(s);
}
217

Christoph Lameter's avatar
Christoph Lameter committed
218
219
#endif

220
static inline void stat(struct kmem_cache *s, enum stat_item si)
221
222
{
#ifdef CONFIG_SLUB_STATS
223
	__this_cpu_inc(s->cpu_slab->stat[si]);
224
225
226
#endif
}

Christoph Lameter's avatar
Christoph Lameter committed
227
228
229
230
231
232
233
234
235
236
237
238
239
240
/********************************************************************
 * 			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)
{
	return s->node[node];
}

Christoph Lameter's avatar
Christoph Lameter committed
241
/* Verify that a pointer has an address that is valid within a slab page */
242
243
244
245
246
static inline int check_valid_pointer(struct kmem_cache *s,
				struct page *page, const void *object)
{
	void *base;

247
	if (!object)
248
249
		return 1;

250
	base = page_address(page);
251
	if (object < base || object >= base + page->objects * s->size ||
252
253
254
255
256
257
258
		(object - base) % s->size) {
		return 0;
	}

	return 1;
}

259
260
261
262
263
264
265
266
267
268
269
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 */
270
271
#define for_each_object(__p, __s, __addr, __objects) \
	for (__p = (__addr); __p < (__addr) + (__objects) * (__s)->size;\
272
273
274
275
			__p += (__s)->size)

/* Scan freelist */
#define for_each_free_object(__p, __s, __free) \
276
	for (__p = (__free); __p; __p = get_freepointer((__s), __p))
277
278
279
280
281
282
283

/* 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;
}

284
285
286
287
static inline struct kmem_cache_order_objects oo_make(int order,
						unsigned long size)
{
	struct kmem_cache_order_objects x = {
288
		(order << OO_SHIFT) + (PAGE_SIZE << order) / size
289
290
291
292
293
294
295
	};

	return x;
}

static inline int oo_order(struct kmem_cache_order_objects x)
{
296
	return x.x >> OO_SHIFT;
297
298
299
300
}

static inline int oo_objects(struct kmem_cache_order_objects x)
{
301
	return x.x & OO_MASK;
302
303
}

304
305
306
307
#ifdef CONFIG_SLUB_DEBUG
/*
 * Debug settings:
 */
308
309
310
#ifdef CONFIG_SLUB_DEBUG_ON
static int slub_debug = DEBUG_DEFAULT_FLAGS;
#else
311
static int slub_debug;
312
#endif
313
314

static char *slub_debug_slabs;
315
static int disable_higher_order_debug;
316

Christoph Lameter's avatar
Christoph Lameter committed
317
318
319
320
321
322
323
324
325
326
327
328
329
/*
 * 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) {
330
			printk(KERN_ERR "%8s 0x%p: ", text, addr + i);
Christoph Lameter's avatar
Christoph Lameter committed
331
332
			newline = 0;
		}
333
		printk(KERN_CONT " %02x", addr[i]);
Christoph Lameter's avatar
Christoph Lameter committed
334
335
336
		offset = i % 16;
		ascii[offset] = isgraph(addr[i]) ? addr[i] : '.';
		if (offset == 15) {
337
			printk(KERN_CONT " %s\n", ascii);
Christoph Lameter's avatar
Christoph Lameter committed
338
339
340
341
342
343
			newline = 1;
		}
	}
	if (!newline) {
		i %= 16;
		while (i < 16) {
344
			printk(KERN_CONT "   ");
Christoph Lameter's avatar
Christoph Lameter committed
345
346
347
			ascii[i] = ' ';
			i++;
		}
348
		printk(KERN_CONT " %s\n", ascii);
Christoph Lameter's avatar
Christoph Lameter committed
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
	}
}

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,
366
			enum track_item alloc, unsigned long addr)
Christoph Lameter's avatar
Christoph Lameter committed
367
{
Akinobu Mita's avatar
Akinobu Mita committed
368
	struct track *p = get_track(s, object, alloc);
Christoph Lameter's avatar
Christoph Lameter committed
369
370
371
372

	if (addr) {
		p->addr = addr;
		p->cpu = smp_processor_id();
373
		p->pid = current->pid;
Christoph Lameter's avatar
Christoph Lameter committed
374
375
376
377
378
379
380
		p->when = jiffies;
	} else
		memset(p, 0, sizeof(struct track));
}

static void init_tracking(struct kmem_cache *s, void *object)
{
381
382
383
	if (!(s->flags & SLAB_STORE_USER))
		return;

384
385
	set_track(s, object, TRACK_FREE, 0UL);
	set_track(s, object, TRACK_ALLOC, 0UL);
Christoph Lameter's avatar
Christoph Lameter committed
386
387
388
389
390
391
392
}

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

393
	printk(KERN_ERR "INFO: %s in %pS age=%lu cpu=%u pid=%d\n",
394
		s, (void *)t->addr, jiffies - t->when, t->cpu, t->pid);
395
396
397
398
399
400
401
402
403
404
405
406
407
}

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)
{
408
409
	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);
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425

}

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
426
427
}

428
429
430
431
432
433
434
435
436
437
438
439
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
440
441
{
	unsigned int off;	/* Offset of last byte */
442
	u8 *addr = page_address(page);
443
444
445
446
447
448
449
450
451
452
453

	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);

454
	print_section("Object", p, min_t(unsigned long, s->objsize, PAGE_SIZE));
Christoph Lameter's avatar
Christoph Lameter committed
455
456
457
458
459
460
461
462
463
464

	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;

465
	if (s->flags & SLAB_STORE_USER)
Christoph Lameter's avatar
Christoph Lameter committed
466
467
468
469
		off += 2 * sizeof(struct track);

	if (off != s->size)
		/* Beginning of the filler is the free pointer */
470
471
472
		print_section("Padding", p + off, s->size - off);

	dump_stack();
Christoph Lameter's avatar
Christoph Lameter committed
473
474
475
476
477
}

static void object_err(struct kmem_cache *s, struct page *page,
			u8 *object, char *reason)
{
478
	slab_bug(s, "%s", reason);
479
	print_trailer(s, page, object);
Christoph Lameter's avatar
Christoph Lameter committed
480
481
}

482
static void slab_err(struct kmem_cache *s, struct page *page, char *fmt, ...)
Christoph Lameter's avatar
Christoph Lameter committed
483
484
485
486
{
	va_list args;
	char buf[100];

487
488
	va_start(args, fmt);
	vsnprintf(buf, sizeof(buf), fmt, args);
Christoph Lameter's avatar
Christoph Lameter committed
489
	va_end(args);
490
	slab_bug(s, "%s", buf);
491
	print_page_info(page);
Christoph Lameter's avatar
Christoph Lameter committed
492
493
494
	dump_stack();
}

495
static void init_object(struct kmem_cache *s, void *object, u8 val)
Christoph Lameter's avatar
Christoph Lameter committed
496
497
498
499
500
{
	u8 *p = object;

	if (s->flags & __OBJECT_POISON) {
		memset(p, POISON_FREE, s->objsize - 1);
501
		p[s->objsize - 1] = POISON_END;
Christoph Lameter's avatar
Christoph Lameter committed
502
503
504
	}

	if (s->flags & SLAB_RED_ZONE)
505
		memset(p + s->objsize, val, s->inuse - s->objsize);
Christoph Lameter's avatar
Christoph Lameter committed
506
507
}

508
static u8 *check_bytes(u8 *start, unsigned int value, unsigned int bytes)
Christoph Lameter's avatar
Christoph Lameter committed
509
510
511
{
	while (bytes) {
		if (*start != (u8)value)
512
			return start;
Christoph Lameter's avatar
Christoph Lameter committed
513
514
515
		start++;
		bytes--;
	}
516
517
518
519
520
521
522
523
524
525
526
527
	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,
528
			u8 *start, unsigned int value, unsigned int bytes)
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
{
	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
548
549
550
551
552
553
554
555
556
}

/*
 * 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
557
 *
Christoph Lameter's avatar
Christoph Lameter committed
558
559
560
561
562
 * 	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
563
564
565
 * 	Padding is extended by another word if Redzoning is enabled and
 * 	objsize == inuse.
 *
Christoph Lameter's avatar
Christoph Lameter committed
566
567
568
569
 * 	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
570
571
 * 	Meta data starts here.
 *
Christoph Lameter's avatar
Christoph Lameter committed
572
573
 * 	A. Free pointer (if we cannot overwrite object on free)
 * 	B. Tracking data for SLAB_STORE_USER
Christoph Lameter's avatar
Christoph Lameter committed
574
 * 	C. Padding to reach required alignment boundary or at mininum
Christoph Lameter's avatar
Christoph Lameter committed
575
 * 		one word if debugging is on to be able to detect writes
Christoph Lameter's avatar
Christoph Lameter committed
576
577
578
 * 		before the word boundary.
 *
 *	Padding is done using 0x5a (POISON_INUSE)
Christoph Lameter's avatar
Christoph Lameter committed
579
580
 *
 * object + s->size
Christoph Lameter's avatar
Christoph Lameter committed
581
 * 	Nothing is used beyond s->size.
Christoph Lameter's avatar
Christoph Lameter committed
582
 *
Christoph Lameter's avatar
Christoph Lameter committed
583
584
 * 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
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
 * 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;

603
604
	return check_bytes_and_report(s, page, p, "Object padding",
				p + off, POISON_INUSE, s->size - off);
Christoph Lameter's avatar
Christoph Lameter committed
605
606
}

607
/* Check the pad bytes at the end of a slab page */
Christoph Lameter's avatar
Christoph Lameter committed
608
609
static int slab_pad_check(struct kmem_cache *s, struct page *page)
{
610
611
612
613
614
	u8 *start;
	u8 *fault;
	u8 *end;
	int length;
	int remainder;
Christoph Lameter's avatar
Christoph Lameter committed
615
616
617
618

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

619
	start = page_address(page);
620
	length = (PAGE_SIZE << compound_order(page));
621
622
	end = start + length;
	remainder = length % s->size;
Christoph Lameter's avatar
Christoph Lameter committed
623
624
625
	if (!remainder)
		return 1;

626
	fault = check_bytes(end - remainder, POISON_INUSE, remainder);
627
628
629
630
631
632
	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);
633
	print_section("Padding", end - remainder, remainder);
634

Eric Dumazet's avatar
Eric Dumazet committed
635
	restore_bytes(s, "slab padding", POISON_INUSE, end - remainder, end);
636
	return 0;
Christoph Lameter's avatar
Christoph Lameter committed
637
638
639
}

static int check_object(struct kmem_cache *s, struct page *page,
640
					void *object, u8 val)
Christoph Lameter's avatar
Christoph Lameter committed
641
642
643
644
645
{
	u8 *p = object;
	u8 *endobject = object + s->objsize;

	if (s->flags & SLAB_RED_ZONE) {
646
		if (!check_bytes_and_report(s, page, object, "Redzone",
647
			endobject, val, s->inuse - s->objsize))
Christoph Lameter's avatar
Christoph Lameter committed
648
649
			return 0;
	} else {
Ingo Molnar's avatar
Ingo Molnar committed
650
651
652
653
		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
654
655
656
	}

	if (s->flags & SLAB_POISON) {
657
		if (val != SLUB_RED_ACTIVE && (s->flags & __OBJECT_POISON) &&
658
659
660
			(!check_bytes_and_report(s, page, p, "Poison", p,
					POISON_FREE, s->objsize - 1) ||
			 !check_bytes_and_report(s, page, p, "Poison",
661
				p + s->objsize - 1, POISON_END, 1)))
Christoph Lameter's avatar
Christoph Lameter committed
662
663
664
665
666
667
668
			return 0;
		/*
		 * check_pad_bytes cleans up on its own.
		 */
		check_pad_bytes(s, page, p);
	}

669
	if (!s->offset && val == SLUB_RED_ACTIVE)
Christoph Lameter's avatar
Christoph Lameter committed
670
671
672
673
674
675
676
677
678
679
		/*
		 * 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
680
		 * No choice but to zap it and thus lose the remainder
Christoph Lameter's avatar
Christoph Lameter committed
681
		 * of the free objects in this slab. May cause
Christoph Lameter's avatar
Christoph Lameter committed
682
		 * another error because the object count is now wrong.
Christoph Lameter's avatar
Christoph Lameter committed
683
		 */
684
		set_freepointer(s, p, NULL);
Christoph Lameter's avatar
Christoph Lameter committed
685
686
687
688
689
690
691
		return 0;
	}
	return 1;
}

static int check_slab(struct kmem_cache *s, struct page *page)
{
692
693
	int maxobj;

Christoph Lameter's avatar
Christoph Lameter committed
694
695
696
	VM_BUG_ON(!irqs_disabled());

	if (!PageSlab(page)) {
697
		slab_err(s, page, "Not a valid slab page");
Christoph Lameter's avatar
Christoph Lameter committed
698
699
		return 0;
	}
700
701
702
703
704
705
706
707

	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) {
708
		slab_err(s, page, "inuse %u > max %u",
709
			s->name, page->inuse, page->objects);
Christoph Lameter's avatar
Christoph Lameter committed
710
711
712
713
714
715
716
717
		return 0;
	}
	/* Slab_pad_check fixes things up after itself */
	slab_pad_check(s, page);
	return 1;
}

/*
Christoph Lameter's avatar
Christoph Lameter committed
718
719
 * 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
720
721
722
723
724
725
 */
static int on_freelist(struct kmem_cache *s, struct page *page, void *search)
{
	int nr = 0;
	void *fp = page->freelist;
	void *object = NULL;
726
	unsigned long max_objects;
Christoph Lameter's avatar
Christoph Lameter committed
727

728
	while (fp && nr <= page->objects) {
Christoph Lameter's avatar
Christoph Lameter committed
729
730
731
732
733
734
		if (fp == search)
			return 1;
		if (!check_valid_pointer(s, page, fp)) {
			if (object) {
				object_err(s, page, object,
					"Freechain corrupt");
735
				set_freepointer(s, object, NULL);
Christoph Lameter's avatar
Christoph Lameter committed
736
737
				break;
			} else {
738
				slab_err(s, page, "Freepointer corrupt");
739
				page->freelist = NULL;
740
				page->inuse = page->objects;
741
				slab_fix(s, "Freelist cleared");
Christoph Lameter's avatar
Christoph Lameter committed
742
743
744
745
746
747
748
749
750
				return 0;
			}
			break;
		}
		object = fp;
		fp = get_freepointer(s, object);
		nr++;
	}

751
	max_objects = (PAGE_SIZE << compound_order(page)) / s->size;
752
753
	if (max_objects > MAX_OBJS_PER_PAGE)
		max_objects = MAX_OBJS_PER_PAGE;
754
755
756
757
758
759
760

	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.");
	}
761
	if (page->inuse != page->objects - nr) {
762
		slab_err(s, page, "Wrong object count. Counter is %d but "
763
764
			"counted were %d", page->inuse, page->objects - nr);
		page->inuse = page->objects - nr;
765
		slab_fix(s, "Object count adjusted.");
Christoph Lameter's avatar
Christoph Lameter committed
766
767
768
769
	}
	return search == NULL;
}

770
771
static void trace(struct kmem_cache *s, struct page *page, void *object,
								int alloc)
Christoph Lameter's avatar
Christoph Lameter committed
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
{
	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();
	}
}

787
788
789
790
791
792
/*
 * Hooks for other subsystems that check memory allocations. In a typical
 * production configuration these hooks all should produce no code at all.
 */
static inline int slab_pre_alloc_hook(struct kmem_cache *s, gfp_t flags)
{
793
	flags &= gfp_allowed_mask;
794
795
796
797
798
799
800
801
	lockdep_trace_alloc(flags);
	might_sleep_if(flags & __GFP_WAIT);

	return should_failslab(s->objsize, flags, s->flags);
}

static inline void slab_post_alloc_hook(struct kmem_cache *s, gfp_t flags, void *object)
{
802
	flags &= gfp_allowed_mask;
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
	kmemcheck_slab_alloc(s, flags, object, s->objsize);
	kmemleak_alloc_recursive(object, s->objsize, 1, s->flags, flags);
}

static inline void slab_free_hook(struct kmem_cache *s, void *x)
{
	kmemleak_free_recursive(x, s->flags);
}

static inline void slab_free_hook_irq(struct kmem_cache *s, void *object)
{
	kmemcheck_slab_free(s, object, s->objsize);
	debug_check_no_locks_freed(object, s->objsize);
	if (!(s->flags & SLAB_DEBUG_OBJECTS))
		debug_check_no_obj_freed(object, s->objsize);
}

820
/*
Christoph Lameter's avatar
Christoph Lameter committed
821
 * Tracking of fully allocated slabs for debugging purposes.
822
 */
Christoph Lameter's avatar
Christoph Lameter committed
823
static void add_full(struct kmem_cache_node *n, struct page *page)
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
{
	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);
}

844
845
846
847
848
849
850
851
/* 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);
}

852
853
854
855
856
static inline unsigned long node_nr_slabs(struct kmem_cache_node *n)
{
	return atomic_long_read(&n->nr_slabs);
}

857
static inline void inc_slabs_node(struct kmem_cache *s, int node, int objects)
858
859
860
861
862
863
864
865
866
{
	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).
	 */
867
	if (n) {
868
		atomic_long_inc(&n->nr_slabs);
869
870
		atomic_long_add(objects, &n->total_objects);
	}
871
}
872
static inline void dec_slabs_node(struct kmem_cache *s, int node, int objects)
873
874
875
876
{
	struct kmem_cache_node *n = get_node(s, node);

	atomic_long_dec(&n->nr_slabs);
877
	atomic_long_sub(objects, &n->total_objects);
878
879
880
}

/* Object debug checks for alloc/free paths */
Christoph Lameter's avatar
Christoph Lameter committed
881
882
883
884
885
886
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;

887
	init_object(s, object, SLUB_RED_INACTIVE);
Christoph Lameter's avatar
Christoph Lameter committed
888
889
890
	init_tracking(s, object);
}

891
static noinline int alloc_debug_processing(struct kmem_cache *s, struct page *page,
892
					void *object, unsigned long addr)
Christoph Lameter's avatar
Christoph Lameter committed
893
894
895
896
{
	if (!check_slab(s, page))
		goto bad;

897
	if (!on_freelist(s, page, object)) {
898
		object_err(s, page, object, "Object already allocated");
899
		goto bad;
Christoph Lameter's avatar
Christoph Lameter committed
900
901
902
903
	}

	if (!check_valid_pointer(s, page, object)) {
		object_err(s, page, object, "Freelist Pointer check fails");
904
		goto bad;
Christoph Lameter's avatar
Christoph Lameter committed
905
906
	}

907
	if (!check_object(s, page, object, SLUB_RED_INACTIVE))
Christoph Lameter's avatar
Christoph Lameter committed
908
909
		goto bad;

Christoph Lameter's avatar
Christoph Lameter committed
910
911
912
913
	/* 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);
914
	init_object(s, object, SLUB_RED_ACTIVE);
Christoph Lameter's avatar
Christoph Lameter committed
915
	return 1;
Christoph Lameter's avatar
Christoph Lameter committed
916

Christoph Lameter's avatar
Christoph Lameter committed
917
918
919
920
921
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
922
		 * as used avoids touching the remaining objects.
Christoph Lameter's avatar
Christoph Lameter committed
923
		 */
924
		slab_fix(s, "Marking all objects used");
925
		page->inuse = page->objects;
926
		page->freelist = NULL;
Christoph Lameter's avatar
Christoph Lameter committed
927
928
929
930
	}
	return 0;
}

931
932
static noinline int free_debug_processing(struct kmem_cache *s,
		 struct page *page, void *object, unsigned long addr)
Christoph Lameter's avatar
Christoph Lameter committed
933
934
935
936
937
{
	if (!check_slab(s, page))
		goto fail;

	if (!check_valid_pointer(s, page, object)) {
938
		slab_err(s, page, "Invalid object pointer 0x%p", object);
Christoph Lameter's avatar
Christoph Lameter committed
939
940
941
942
		goto fail;
	}

	if (on_freelist(s, page, object)) {
943
		object_err(s, page, object, "Object already free");
Christoph Lameter's avatar
Christoph Lameter committed
944
945
946
		goto fail;
	}

947
	if (!check_object(s, page, object, SLUB_RED_ACTIVE))
Christoph Lameter's avatar
Christoph Lameter committed
948
949
950
		return 0;

	if (unlikely(s != page->slab)) {
Ingo Molnar's avatar
Ingo Molnar committed
951
		if (!PageSlab(page)) {
952
953
			slab_err(s, page, "Attempt to free object(0x%p) "
				"outside of slab", object);
Ingo Molnar's avatar
Ingo Molnar committed
954
		} else if (!page->slab) {
Christoph Lameter's avatar
Christoph Lameter committed
955
			printk(KERN_ERR
956
				"SLUB <none>: no slab for object 0x%p.\n",
Christoph Lameter's avatar
Christoph Lameter committed
957
						object);
958
			dump_stack();
959
		} else
960
961
			object_err(s, page, object,
					"page slab pointer corrupt.");
Christoph Lameter's avatar
Christoph Lameter committed
962
963
		goto fail;
	}
Christoph Lameter's avatar
Christoph Lameter committed
964
965

	/* Special debug activities for freeing objects */
966
	if (!PageSlubFrozen(page) && !page->freelist)
Christoph Lameter's avatar
Christoph Lameter committed
967
968
969
970
		remove_full(s, page);
	if (s->flags & SLAB_STORE_USER)
		set_track(s, object, TRACK_FREE, addr);
	trace(s, page, object, 0);
971
	init_object(s, object, SLUB_RED_INACTIVE);
Christoph Lameter's avatar
Christoph Lameter committed
972
	return 1;
Christoph Lameter's avatar
Christoph Lameter committed
973

Christoph Lameter's avatar
Christoph Lameter committed
974
fail:
975
	slab_fix(s, "Object at 0x%p not freed", object);
Christoph Lameter's avatar
Christoph Lameter committed
976
977
978
	return 0;
}

979
980
static int __init setup_slub_debug(char *str)
{
981
982
983
984
985
986
987
988
989
990
991
992
993
994
	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;

995
996
997
998
999
1000
1001
1002
1003
	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;
	}

1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
	slub_debug = 0;
	if (*str == '-')
		/*
		 * Switch off all debugging measures.
		 */
		goto out;

	/*
	 * Determine which debug features should be switched on
	 */
1014
	for (; *str && *str != ','; str++) {
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
		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;
			break;
		case 't':
			slub_debug |= SLAB_TRACE;
			break;
1031
1032
1033
		case 'a':
			slub_debug |= SLAB_FAILSLAB;
			break;
1034
1035
		default:
			printk(KERN_ERR "slub_debug option '%c' "
1036
				"unknown. skipped\n", *str);
1037
		}
1038
1039
	}

1040
check_slabs:
1041
1042
	if (*str == ',')
		slub_debug_slabs = str + 1;
1043
out:
1044
1045
1046
1047
1048
	return 1;
}

__setup("slub_debug", setup_slub_debug);

1049
1050
static unsigned long kmem_cache_flags(unsigned long objsize,
	unsigned long flags, const char *name,
1051
	void (*ctor)(void *))
1052
1053
{
	/*
1054
	 * Enable debugging if selected on the kernel commandline.
1055
	 */
1056
	if (slub_debug && (!slub_debug_slabs ||
1057
1058
		!strncmp(slub_debug_slabs, name, strlen(slub_debug_slabs))))
		flags |= slub_debug;
1059
1060

	return flags;
1061
1062
}
#else
Christoph Lameter's avatar
Christoph Lameter committed
1063
1064
static inline void setup_object_debug(struct kmem_cache *s,
			struct page *page, void *object) {}
1065

Christoph Lameter's avatar
Christoph Lameter committed
1066
static inline int alloc_debug_processing(struct kmem_cache *s,
1067
	struct page *page, void *object, unsigned long addr) { return 0; }
1068

Christoph Lameter's avatar
Christoph Lameter committed
1069
static inline int free_debug_processing(struct kmem_cache *s,
1070
	struct page *page, void *object, unsigned long addr) { return 0; }
1071
1072
1073
1074

static inline int slab_pad_check(struct kmem_cache *s, struct page *page)
			{ return 1; }
static inline int check_object(struct kmem_cache *s, struct page *page,
1075
			void *object, u8 val) { return 1; }
Christoph Lameter's avatar
Christoph Lameter committed
1076
static inline void add_full(struct kmem_cache_node *n, struct page *page) {}
1077
1078
static inline unsigned long kmem_cache_flags(unsigned long objsize,
	unsigned long flags, const char *name,
1079
	void (*ctor)(void *))
1080
1081
1082
{
	return flags;
}
1083
#define slub_debug 0
1084

1085
1086
#define disable_higher_order_debug 0

1087
1088
static inline unsigned long slabs_node(struct kmem_cache *s, int node)
							{ return 0; }
1089
1090
static inline unsigned long node_nr_slabs(struct kmem_cache_node *n)
							{ return 0; }
1091
1092
1093
1094
static inline void inc_slabs_node(struct kmem_cache *s, int node,
							int objects) {}
static inline void dec_slabs_node(struct kmem_cache *s, int node,
							int objects) {}
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106

static inline int slab_pre_alloc_hook(struct kmem_cache *s, gfp_t flags)
							{ return 0; }

static inline void slab_post_alloc_hook(struct kmem_cache *s, gfp_t flags,
		void *object) {}

static inline void slab_free_hook(struct kmem_cache *s, void *x) {}

static inline void slab_free_hook_irq(struct kmem_cache *s,
		void *object) {}

1107
#endif /* CONFIG_SLUB_DEBUG */
1108

Christoph Lameter's avatar
Christoph Lameter committed
1109
1110
1111
/*
 * Slab allocation and freeing
 */
1112
1113
1114
1115
1116
static inline struct page *alloc_slab_page(gfp_t flags, int node,
					struct kmem_cache_order_objects oo)
{
	int order = oo_order(oo);

1117
1118
	flags |= __GFP_NOTRACK;

1119
	if (node == NUMA_NO_NODE)
1120
1121
		return alloc_pages(flags, order);
	else
1122
		return alloc_pages_exact_node(node, flags, order);
1123
1124
}

Christoph Lameter's avatar
Christoph Lameter committed
1125
1126
static struct page *allocate_slab(struct kmem_cache *s, gfp_t flags, int node)
{
1127
	struct page *page;
1128
	struct kmem_cache_order_objects oo = s->oo;
1129
	gfp_t alloc_gfp;
Christoph Lameter's avatar
Christoph Lameter committed
1130

1131
	flags |= s->allocflags;
1132

1133
1134
1135
1136
1137
1138
1139
	/*
	 * Let the initial higher-order allocation fail under memory pressure
	 * so we fall-back to the minimum order allocation.
	 */
	alloc_gfp = (flags | __GFP_NOWARN | __GFP_NORETRY) & ~__GFP_NOFAIL;

	page = alloc_slab_page(alloc_gfp, node, oo);
1140
1141
1142
1143
1144
1145
1146
1147
1148
	if (unlikely(!page)) {
		oo = s->min;
		/*
		 * Allocation may have failed due to fragmentation.
		 * Try a lower order alloc if possible
		 */
		page = alloc_slab_page(flags, node, oo);
		if (!page)
			return NULL;
Christoph Lameter's avatar
Christoph Lameter committed
1149

1150
		stat(s, ORDER_FALLBACK);
1151
	}
Vegard Nossum's avatar
Vegard Nossum committed
1152
1153

	if (kmemcheck_enabled
1154
		&& !(s->flags & (SLAB_NOTRACK | DEBUG_DEFAULT_FLAGS))) {
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
		int pages = 1 << oo_order(oo);

		kmemcheck_alloc_shadow(page, oo_order(oo), flags, node);

		/*
		 * Objects from caches that have a constructor don't get
		 * cleared when they're allocated, so we need to do it here.
		 */
		if (s->ctor)
			kmemcheck_mark_uninitialized_pages(page, pages);
		else
			kmemcheck_mark_unallocated_pages(page, pages);
Vegard Nossum's avatar
Vegard Nossum committed
1167
1168
	}

1169
	page->objects = oo_objects(oo);
Christoph Lameter's avatar
Christoph Lameter committed
1170
1171
1172
	mod_zone_page_state(page_zone(page),
		(s->flags & SLAB_RECLAIM_ACCOUNT) ?
		NR_SLAB_RECLAIMABLE : NR_SLAB_UNRECLAIMABLE,
1173
		1 << oo_order(oo));
Christoph Lameter's avatar
Christoph Lameter committed
1174
1175
1176
1177
1178
1179
1180

	return page;
}

static void setup_object(struct kmem_cache *s, struct page *page,
				void *object)
{
Christoph Lameter's avatar
Christoph Lameter committed
1181
	setup_object_debug(s, page, object);
1182
	if (unlikely(s->ctor))
1183
		s->ctor(object);
Christoph Lameter's avatar
Christoph Lameter committed
1184
1185
1186
1187
1188
1189
1190
1191
1192
}

static struct page *new_slab(struct kmem_cache *s, gfp_t flags, int node)
{
	struct page *page;
	void *start;
	void *last;
	void *p;

Christoph Lameter's avatar
Christoph Lameter committed
1193
	BUG_ON(flags & GFP_SLAB_BUG_MASK);
Christoph Lameter's avatar
Christoph Lameter committed
1194

Christoph Lameter's avatar
Christoph Lameter committed
1195
1196
	page = allocate_slab(s,
		flags & (GFP_RECLAIM_MASK | GFP_CONSTRAINT_MASK), node);
Christoph Lameter's avatar
Christoph Lameter committed
1197
1198
1199
	if (!page)
		goto out;

1200
	inc_slabs_node(s, page_to_nid(page), page->objects);
Christoph Lameter's avatar
Christoph Lameter committed
1201
1202
1203
1204
1205
1206
	page->slab = s;
	page->flags |= 1 << PG_slab;

	start = page_address(page);

	if (unlikely(s->flags & SLAB_POISON))
1207
		memset(start, POISON_INUSE, PAGE_SIZE << compound_order(page));
Christoph Lameter's avatar
Christoph Lameter committed
1208
1209

	last = start;
1210
	for_each_object(p, s, start, page->objects) {
Christoph Lameter's avatar
Christoph Lameter committed
1211
1212
1213
1214
1215
		setup_object(s, page, last);
		set_freepointer(s, last, p);
		last = p;
	}
	setup_object(s, page, last);
1216
	set_freepointer(s, last, NULL);
Christoph Lameter's avatar
Christoph Lameter committed
1217
1218
1219
1220
1221
1222
1223
1224
1225

	page->freelist = start;
	page->inuse = 0;
out:
	return page;
}

static void __free_slab(struct kmem_cache *s, struct page *page)
{