slub.c 97.1 KB
Newer Older
Christoph Lameter's avatar
Christoph Lameter committed
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
/*
 * 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.
 *
 * (C) 2007 SGI, Christoph Lameter <clameter@sgi.com>
 */

#include <linux/mm.h>
#include <linux/module.h>
#include <linux/bit_spinlock.h>
#include <linux/interrupt.h>
#include <linux/bitops.h>
#include <linux/slab.h>
#include <linux/seq_file.h>
#include <linux/cpu.h>
#include <linux/cpuset.h>
#include <linux/mempolicy.h>
#include <linux/ctype.h>
#include <linux/kallsyms.h>
23
#include <linux/memory.h>
Christoph Lameter's avatar
Christoph Lameter committed
24
25
26
27
28
29
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

/*
 * 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
70
71
 * 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
72
 * freed then the slab will show up again on the partial lists.
Christoph Lameter's avatar
Christoph Lameter committed
73
74
 * We track full slabs for debugging purposes though because otherwise we
 * cannot scan all objects.
Christoph Lameter's avatar
Christoph Lameter committed
75
76
77
78
79
80
81
 *
 * 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.
 *
82
83
84
85
86
87
88
89
90
91
92
93
 * 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
94
 * 			freelist that allows lockless access to
95
96
 * 			free objects in addition to the regular freelist
 * 			that requires the slab lock.
Christoph Lameter's avatar
Christoph Lameter committed
97
98
99
 *
 * PageError		Slab requires special handling due to debug
 * 			options set. This moves	slab handling out of
100
 * 			the fast path and disables lockless freelists.
Christoph Lameter's avatar
Christoph Lameter committed
101
102
 */

103
104
105
106
107
108
109
110
#define FROZEN (1 << PG_active)

#ifdef CONFIG_SLUB_DEBUG
#define SLABDEBUG (1 << PG_error)
#else
#define SLABDEBUG 0
#endif

111
112
static inline int SlabFrozen(struct page *page)
{
113
	return page->flags & FROZEN;
114
115
116
117
}

static inline void SetSlabFrozen(struct page *page)
{
118
	page->flags |= FROZEN;
119
120
121
122
}

static inline void ClearSlabFrozen(struct page *page)
{
123
	page->flags &= ~FROZEN;
124
125
}

126
127
static inline int SlabDebug(struct page *page)
{
128
	return page->flags & SLABDEBUG;
129
130
131
132
}

static inline void SetSlabDebug(struct page *page)
{
133
	page->flags |= SLABDEBUG;
134
135
136
137
}

static inline void ClearSlabDebug(struct page *page)
{
138
	page->flags &= ~SLABDEBUG;
139
140
}

Christoph Lameter's avatar
Christoph Lameter committed
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
/*
 * 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

#if PAGE_SHIFT <= 12

/*
 * Small page size. Make sure that we do not fragment memory
 */
#define DEFAULT_MAX_ORDER 1
#define DEFAULT_MIN_OBJECTS 4

#else

/*
 * Large page machines are customarily able to handle larger
 * page orders.
 */
#define DEFAULT_MAX_ORDER 2
#define DEFAULT_MIN_OBJECTS 8

#endif

171
172
173
174
/*
 * Mininum number of partial slabs. These will be left on the partial
 * lists even if they are empty. kmem_cache_shrink may reclaim them.
 */
175
#define MIN_PARTIAL 5
Christoph Lameter's avatar
Christoph Lameter committed
176

177
178
179
180
181
182
183
/*
 * 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
184
185
#define DEBUG_DEFAULT_FLAGS (SLAB_DEBUG_FREE | SLAB_RED_ZONE | \
				SLAB_POISON | SLAB_STORE_USER)
Christoph Lameter's avatar
Christoph Lameter committed
186

Christoph Lameter's avatar
Christoph Lameter committed
187
188
189
190
191
192
193
194
195
196
/*
 * Set of flags that will prevent slab merging
 */
#define SLUB_NEVER_MERGE (SLAB_RED_ZONE | SLAB_POISON | SLAB_STORE_USER | \
		SLAB_TRACE | SLAB_DESTROY_BY_RCU)

#define SLUB_MERGE_SAME (SLAB_DEBUG_FREE | SLAB_RECLAIM_ACCOUNT | \
		SLAB_CACHE_DMA)

#ifndef ARCH_KMALLOC_MINALIGN
197
#define ARCH_KMALLOC_MINALIGN __alignof__(unsigned long long)
Christoph Lameter's avatar
Christoph Lameter committed
198
199
200
#endif

#ifndef ARCH_SLAB_MINALIGN
201
#define ARCH_SLAB_MINALIGN __alignof__(unsigned long long)
Christoph Lameter's avatar
Christoph Lameter committed
202
203
204
#endif

/* Internal SLUB flags */
205
206
#define __OBJECT_POISON		0x80000000 /* Poison object */
#define __SYSFS_ADD_DEFERRED	0x40000000 /* Not yet visible via sysfs */
Christoph Lameter's avatar
Christoph Lameter committed
207

208
209
210
211
212
/* Not all arches define cache_line_size */
#ifndef cache_line_size
#define cache_line_size()	L1_CACHE_BYTES
#endif

Christoph Lameter's avatar
Christoph Lameter committed
213
214
215
216
217
218
219
220
221
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
222
	UP,		/* Everything works but does not show up in sysfs */
Christoph Lameter's avatar
Christoph Lameter committed
223
224
225
226
227
	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
228
static LIST_HEAD(slab_caches);
Christoph Lameter's avatar
Christoph Lameter committed
229

230
231
232
233
234
235
236
237
238
239
240
241
/*
 * Tracking user of a slab.
 */
struct track {
	void *addr;		/* Called from address */
	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 };

242
#if defined(CONFIG_SYSFS) && defined(CONFIG_SLUB_DEBUG)
Christoph Lameter's avatar
Christoph Lameter committed
243
244
245
246
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 *);
#else
247
248
249
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; }
250
251
252
253
static inline void sysfs_slab_remove(struct kmem_cache *s)
{
	kfree(s);
}
Christoph Lameter's avatar
Christoph Lameter committed
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
#endif

/********************************************************************
 * 			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
}

274
275
static inline struct kmem_cache_cpu *get_cpu_slab(struct kmem_cache *s, int cpu)
{
276
277
278
279
280
#ifdef CONFIG_SMP
	return s->cpu_slab[cpu];
#else
	return &s->cpu_slab;
#endif
281
282
}

283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
static inline int check_valid_pointer(struct kmem_cache *s,
				struct page *page, const void *object)
{
	void *base;

	if (!object)
		return 1;

	base = page_address(page);
	if (object < base || object >= base + s->objects * s->size ||
		(object - base) % s->size) {
		return 0;
	}

	return 1;
}

300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
/*
 * Slow version of get and set free pointer.
 *
 * This version requires touching the cache lines of kmem_cache which
 * we avoid to do in the fast alloc free paths. There we obtain the offset
 * from the page struct.
 */
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 */
#define for_each_object(__p, __s, __addr) \
	for (__p = (__addr); __p < (__addr) + (__s)->objects * (__s)->size;\
			__p += (__s)->size)

/* Scan freelist */
#define for_each_free_object(__p, __s, __free) \
	for (__p = (__free); __p; __p = get_freepointer((__s), __p))

/* 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;
}

332
333
334
335
#ifdef CONFIG_SLUB_DEBUG
/*
 * Debug settings:
 */
336
337
338
#ifdef CONFIG_SLUB_DEBUG_ON
static int slub_debug = DEBUG_DEFAULT_FLAGS;
#else
339
static int slub_debug;
340
#endif
341
342
343

static char *slub_debug_slabs;

Christoph Lameter's avatar
Christoph Lameter committed
344
345
346
347
348
349
350
351
352
353
354
355
356
/*
 * 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) {
357
			printk(KERN_ERR "%8s 0x%p: ", text, addr + i);
Christoph Lameter's avatar
Christoph Lameter committed
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
			newline = 0;
		}
		printk(" %02x", addr[i]);
		offset = i % 16;
		ascii[offset] = isgraph(addr[i]) ? addr[i] : '.';
		if (offset == 15) {
			printk(" %s\n",ascii);
			newline = 1;
		}
	}
	if (!newline) {
		i %= 16;
		while (i < 16) {
			printk("   ");
			ascii[i] = ' ';
			i++;
		}
		printk(" %s\n", ascii);
	}
}

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,
				enum track_item alloc, void *addr)
{
	struct track *p;

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

	p += alloc;
	if (addr) {
		p->addr = addr;
		p->cpu = smp_processor_id();
		p->pid = current ? current->pid : -1;
		p->when = jiffies;
	} else
		memset(p, 0, sizeof(struct track));
}

static void init_tracking(struct kmem_cache *s, void *object)
{
414
415
416
417
418
	if (!(s->flags & SLAB_STORE_USER))
		return;

	set_track(s, object, TRACK_FREE, NULL);
	set_track(s, object, TRACK_ALLOC, NULL);
Christoph Lameter's avatar
Christoph Lameter committed
419
420
421
422
423
424
425
}

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

426
	printk(KERN_ERR "INFO: %s in ", s);
Christoph Lameter's avatar
Christoph Lameter committed
427
	__print_symbol("%s", (unsigned long)t->addr);
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
	printk(" age=%lu cpu=%u pid=%d\n", jiffies - t->when, t->cpu, t->pid);
}

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)
{
	printk(KERN_ERR "INFO: Slab 0x%p used=%u fp=0x%p flags=0x%04lx\n",
		page, page->inuse, page->freelist, page->flags);

}

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
460
461
}

462
463
464
465
466
467
468
469
470
471
472
473
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
474
475
{
	unsigned int off;	/* Offset of last byte */
476
477
478
479
480
481
482
483
484
485
486
487
488
	u8 *addr = page_address(page);

	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);

	print_section("Object", p, min(s->objsize, 128));
Christoph Lameter's avatar
Christoph Lameter committed
489
490
491
492
493
494
495
496
497
498

	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;

499
	if (s->flags & SLAB_STORE_USER)
Christoph Lameter's avatar
Christoph Lameter committed
500
501
502
503
		off += 2 * sizeof(struct track);

	if (off != s->size)
		/* Beginning of the filler is the free pointer */
504
505
506
		print_section("Padding", p + off, s->size - off);

	dump_stack();
Christoph Lameter's avatar
Christoph Lameter committed
507
508
509
510
511
}

static void object_err(struct kmem_cache *s, struct page *page,
			u8 *object, char *reason)
{
512
513
	slab_bug(s, reason);
	print_trailer(s, page, object);
Christoph Lameter's avatar
Christoph Lameter committed
514
515
}

516
static void slab_err(struct kmem_cache *s, struct page *page, char *fmt, ...)
Christoph Lameter's avatar
Christoph Lameter committed
517
518
519
520
{
	va_list args;
	char buf[100];

521
522
	va_start(args, fmt);
	vsnprintf(buf, sizeof(buf), fmt, args);
Christoph Lameter's avatar
Christoph Lameter committed
523
	va_end(args);
524
525
	slab_bug(s, fmt);
	print_page_info(page);
Christoph Lameter's avatar
Christoph Lameter committed
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
	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);
		p[s->objsize -1] = POISON_END;
	}

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

544
static u8 *check_bytes(u8 *start, unsigned int value, unsigned int bytes)
Christoph Lameter's avatar
Christoph Lameter committed
545
546
547
{
	while (bytes) {
		if (*start != (u8)value)
548
			return start;
Christoph Lameter's avatar
Christoph Lameter committed
549
550
551
		start++;
		bytes--;
	}
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
	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,
			u8* start, unsigned int value, unsigned int bytes)
{
	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
584
585
586
587
588
589
590
591
592
}

/*
 * 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
593
 *
Christoph Lameter's avatar
Christoph Lameter committed
594
595
596
597
598
 * 	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
599
600
601
 * 	Padding is extended by another word if Redzoning is enabled and
 * 	objsize == inuse.
 *
Christoph Lameter's avatar
Christoph Lameter committed
602
603
604
605
 * 	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
606
607
 * 	Meta data starts here.
 *
Christoph Lameter's avatar
Christoph Lameter committed
608
609
 * 	A. Free pointer (if we cannot overwrite object on free)
 * 	B. Tracking data for SLAB_STORE_USER
Christoph Lameter's avatar
Christoph Lameter committed
610
611
612
613
614
 * 	C. Padding to reach required alignment boundary or at mininum
 * 		one word if debuggin is on to be able to detect writes
 * 		before the word boundary.
 *
 *	Padding is done using 0x5a (POISON_INUSE)
Christoph Lameter's avatar
Christoph Lameter committed
615
616
 *
 * object + s->size
Christoph Lameter's avatar
Christoph Lameter committed
617
 * 	Nothing is used beyond s->size.
Christoph Lameter's avatar
Christoph Lameter committed
618
 *
Christoph Lameter's avatar
Christoph Lameter committed
619
620
 * 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
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
 * 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;

639
640
	return check_bytes_and_report(s, page, p, "Object padding",
				p + off, POISON_INUSE, s->size - off);
Christoph Lameter's avatar
Christoph Lameter committed
641
642
643
644
}

static int slab_pad_check(struct kmem_cache *s, struct page *page)
{
645
646
647
648
649
	u8 *start;
	u8 *fault;
	u8 *end;
	int length;
	int remainder;
Christoph Lameter's avatar
Christoph Lameter committed
650
651
652
653

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

654
655
	start = page_address(page);
	end = start + (PAGE_SIZE << s->order);
Christoph Lameter's avatar
Christoph Lameter committed
656
	length = s->objects * s->size;
657
	remainder = end - (start + length);
Christoph Lameter's avatar
Christoph Lameter committed
658
659
660
	if (!remainder)
		return 1;

661
662
663
664
665
666
667
668
669
670
671
	fault = check_bytes(start + length, POISON_INUSE, remainder);
	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);
	print_section("Padding", start, length);

	restore_bytes(s, "slab padding", POISON_INUSE, start, end);
	return 0;
Christoph Lameter's avatar
Christoph Lameter committed
672
673
674
675
676
677
678
679
680
681
682
683
}

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;

684
685
		if (!check_bytes_and_report(s, page, object, "Redzone",
			endobject, red, s->inuse - s->objsize))
Christoph Lameter's avatar
Christoph Lameter committed
686
687
			return 0;
	} else {
688
689
690
		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
691
692
693
694
	}

	if (s->flags & SLAB_POISON) {
		if (!active && (s->flags & __OBJECT_POISON) &&
695
696
697
698
			(!check_bytes_and_report(s, page, p, "Poison", p,
					POISON_FREE, s->objsize - 1) ||
			 !check_bytes_and_report(s, page, p, "Poison",
			 	p + s->objsize -1, POISON_END, 1)))
Christoph Lameter's avatar
Christoph Lameter committed
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
			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");
		/*
		 * No choice but to zap it and thus loose the remainder
		 * of the free objects in this slab. May cause
Christoph Lameter's avatar
Christoph Lameter committed
719
		 * another error because the object count is now wrong.
Christoph Lameter's avatar
Christoph Lameter committed
720
721
722
723
724
725
726
727
728
729
730
731
		 */
		set_freepointer(s, p, NULL);
		return 0;
	}
	return 1;
}

static int check_slab(struct kmem_cache *s, struct page *page)
{
	VM_BUG_ON(!irqs_disabled());

	if (!PageSlab(page)) {
732
		slab_err(s, page, "Not a valid slab page");
Christoph Lameter's avatar
Christoph Lameter committed
733
734
735
		return 0;
	}
	if (page->inuse > s->objects) {
736
737
		slab_err(s, page, "inuse %u > max %u",
			s->name, page->inuse, s->objects);
Christoph Lameter's avatar
Christoph Lameter committed
738
739
740
741
742
743
744
745
		return 0;
	}
	/* Slab_pad_check fixes things up after itself */
	slab_pad_check(s, page);
	return 1;
}

/*
Christoph Lameter's avatar
Christoph Lameter committed
746
747
 * 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
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
 */
static int on_freelist(struct kmem_cache *s, struct page *page, void *search)
{
	int nr = 0;
	void *fp = page->freelist;
	void *object = NULL;

	while (fp && nr <= s->objects) {
		if (fp == search)
			return 1;
		if (!check_valid_pointer(s, page, fp)) {
			if (object) {
				object_err(s, page, object,
					"Freechain corrupt");
				set_freepointer(s, object, NULL);
				break;
			} else {
765
				slab_err(s, page, "Freepointer corrupt");
Christoph Lameter's avatar
Christoph Lameter committed
766
767
				page->freelist = NULL;
				page->inuse = s->objects;
768
				slab_fix(s, "Freelist cleared");
Christoph Lameter's avatar
Christoph Lameter committed
769
770
771
772
773
774
775
776
777
778
				return 0;
			}
			break;
		}
		object = fp;
		fp = get_freepointer(s, object);
		nr++;
	}

	if (page->inuse != s->objects - nr) {
779
		slab_err(s, page, "Wrong object count. Counter is %d but "
780
			"counted were %d", page->inuse, s->objects - nr);
Christoph Lameter's avatar
Christoph Lameter committed
781
		page->inuse = s->objects - nr;
782
		slab_fix(s, "Object count adjusted.");
Christoph Lameter's avatar
Christoph Lameter committed
783
784
785
786
	}
	return search == NULL;
}

Christoph Lameter's avatar
Christoph Lameter committed
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
static void trace(struct kmem_cache *s, struct page *page, void *object, int alloc)
{
	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();
	}
}

803
/*
Christoph Lameter's avatar
Christoph Lameter committed
804
 * Tracking of fully allocated slabs for debugging purposes.
805
 */
Christoph Lameter's avatar
Christoph Lameter committed
806
static void add_full(struct kmem_cache_node *n, struct page *page)
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
{
	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);
}

Christoph Lameter's avatar
Christoph Lameter committed
827
828
829
830
831
832
833
834
835
836
837
838
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,
						void *object, void *addr)
Christoph Lameter's avatar
Christoph Lameter committed
839
840
841
842
843
{
	if (!check_slab(s, page))
		goto bad;

	if (object && !on_freelist(s, page, object)) {
844
		object_err(s, page, object, "Object already allocated");
845
		goto bad;
Christoph Lameter's avatar
Christoph Lameter committed
846
847
848
849
	}

	if (!check_valid_pointer(s, page, object)) {
		object_err(s, page, object, "Freelist Pointer check fails");
850
		goto bad;
Christoph Lameter's avatar
Christoph Lameter committed
851
852
	}

Christoph Lameter's avatar
Christoph Lameter committed
853
	if (object && !check_object(s, page, object, 0))
Christoph Lameter's avatar
Christoph Lameter committed
854
855
		goto bad;

Christoph Lameter's avatar
Christoph Lameter committed
856
857
858
859
860
	/* 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
861
	return 1;
Christoph Lameter's avatar
Christoph Lameter committed
862

Christoph Lameter's avatar
Christoph Lameter committed
863
864
865
866
867
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
868
		 * as used avoids touching the remaining objects.
Christoph Lameter's avatar
Christoph Lameter committed
869
		 */
870
		slab_fix(s, "Marking all objects used");
Christoph Lameter's avatar
Christoph Lameter committed
871
872
873
874
875
876
		page->inuse = s->objects;
		page->freelist = NULL;
	}
	return 0;
}

Christoph Lameter's avatar
Christoph Lameter committed
877
878
static int free_debug_processing(struct kmem_cache *s, struct page *page,
						void *object, void *addr)
Christoph Lameter's avatar
Christoph Lameter committed
879
880
881
882
883
{
	if (!check_slab(s, page))
		goto fail;

	if (!check_valid_pointer(s, page, object)) {
884
		slab_err(s, page, "Invalid object pointer 0x%p", object);
Christoph Lameter's avatar
Christoph Lameter committed
885
886
887
888
		goto fail;
	}

	if (on_freelist(s, page, object)) {
889
		object_err(s, page, object, "Object already free");
Christoph Lameter's avatar
Christoph Lameter committed
890
891
892
893
894
895
896
897
		goto fail;
	}

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

	if (unlikely(s != page->slab)) {
		if (!PageSlab(page))
898
899
			slab_err(s, page, "Attempt to free object(0x%p) "
				"outside of slab", object);
Christoph Lameter's avatar
Christoph Lameter committed
900
		else
901
		if (!page->slab) {
Christoph Lameter's avatar
Christoph Lameter committed
902
			printk(KERN_ERR
903
				"SLUB <none>: no slab for object 0x%p.\n",
Christoph Lameter's avatar
Christoph Lameter committed
904
						object);
905
906
			dump_stack();
		}
Christoph Lameter's avatar
Christoph Lameter committed
907
		else
908
909
			object_err(s, page, object,
					"page slab pointer corrupt.");
Christoph Lameter's avatar
Christoph Lameter committed
910
911
		goto fail;
	}
Christoph Lameter's avatar
Christoph Lameter committed
912
913
914
915
916
917
918
919

	/* Special debug activities for freeing objects */
	if (!SlabFrozen(page) && !page->freelist)
		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
920
	return 1;
Christoph Lameter's avatar
Christoph Lameter committed
921

Christoph Lameter's avatar
Christoph Lameter committed
922
fail:
923
	slab_fix(s, "Object at 0x%p not freed", object);
Christoph Lameter's avatar
Christoph Lameter committed
924
925
926
	return 0;
}

927
928
static int __init setup_slub_debug(char *str)
{
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
	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;

	slub_debug = 0;
	if (*str == '-')
		/*
		 * Switch off all debugging measures.
		 */
		goto out;

	/*
	 * Determine which debug features should be switched on
	 */
	for ( ;*str && *str != ','; str++) {
		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;
		default:
			printk(KERN_ERR "slub_debug option '%c' "
				"unknown. skipped\n",*str);
		}
974
975
	}

976
check_slabs:
977
978
	if (*str == ',')
		slub_debug_slabs = str + 1;
979
out:
980
981
982
983
984
	return 1;
}

__setup("slub_debug", setup_slub_debug);

985
986
static unsigned long kmem_cache_flags(unsigned long objsize,
	unsigned long flags, const char *name,
987
	void (*ctor)(struct kmem_cache *, void *))
988
989
990
991
992
993
994
995
996
997
{
	/*
	 * The page->offset field is only 16 bit wide. This is an offset
	 * in units of words from the beginning of an object. If the slab
	 * size is bigger then we cannot move the free pointer behind the
	 * object anymore.
	 *
	 * On 32 bit platforms the limit is 256k. On 64bit platforms
	 * the limit is 512k.
	 *
998
	 * Debugging or ctor may create a need to move the free
999
1000
	 * pointer. Fail if this happens.
	 */
1001
1002
	if (objsize >= 65535 * sizeof(void *)) {
		BUG_ON(flags & (SLAB_RED_ZONE | SLAB_POISON |
1003
				SLAB_STORE_USER | SLAB_DESTROY_BY_RCU));
1004
1005
		BUG_ON(ctor);
	} else {
1006
1007
1008
1009
		/*
		 * Enable debugging if selected on the kernel commandline.
		 */
		if (slub_debug && (!slub_debug_slabs ||
1010
		    strncmp(slub_debug_slabs, name,
1011
		    	strlen(slub_debug_slabs)) == 0))
1012
1013
1014
1015
				flags |= slub_debug;
	}

	return flags;
1016
1017
}
#else
Christoph Lameter's avatar
Christoph Lameter committed
1018
1019
static inline void setup_object_debug(struct kmem_cache *s,
			struct page *page, void *object) {}
1020

Christoph Lameter's avatar
Christoph Lameter committed
1021
1022
static inline int alloc_debug_processing(struct kmem_cache *s,
	struct page *page, void *object, void *addr) { return 0; }
1023

Christoph Lameter's avatar
Christoph Lameter committed
1024
1025
static inline int free_debug_processing(struct kmem_cache *s,
	struct page *page, void *object, void *addr) { return 0; }
1026
1027
1028
1029
1030

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,
			void *object, int active) { return 1; }
Christoph Lameter's avatar
Christoph Lameter committed
1031
static inline void add_full(struct kmem_cache_node *n, struct page *page) {}
1032
1033
static inline unsigned long kmem_cache_flags(unsigned long objsize,
	unsigned long flags, const char *name,
1034
	void (*ctor)(struct kmem_cache *, void *))
1035
1036
1037
{
	return flags;
}
1038
1039
#define slub_debug 0
#endif
Christoph Lameter's avatar
Christoph Lameter committed
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
/*
 * Slab allocation and freeing
 */
static struct page *allocate_slab(struct kmem_cache *s, gfp_t flags, int node)
{
	struct page * page;
	int pages = 1 << s->order;

	if (s->order)
		flags |= __GFP_COMP;

	if (s->flags & SLAB_CACHE_DMA)
		flags |= SLUB_DMA;

1054
1055
1056
	if (s->flags & SLAB_RECLAIM_ACCOUNT)
		flags |= __GFP_RECLAIMABLE;

Christoph Lameter's avatar
Christoph Lameter committed
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
	if (node == -1)
		page = alloc_pages(flags, s->order);
	else
		page = alloc_pages_node(node, flags, s->order);

	if (!page)
		return NULL;

	mod_zone_page_state(page_zone(page),
		(s->flags & SLAB_RECLAIM_ACCOUNT) ?
		NR_SLAB_RECLAIMABLE : NR_SLAB_UNRECLAIMABLE,
		pages);

	return page;
}

static void setup_object(struct kmem_cache *s, struct page *page,
				void *object)
{
Christoph Lameter's avatar
Christoph Lameter committed
1076
	setup_object_debug(s, page, object);
1077
	if (unlikely(s->ctor))
1078
		s->ctor(s, object);
Christoph Lameter's avatar
Christoph Lameter committed
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
}

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

Christoph Lameter's avatar
Christoph Lameter committed
1089
	BUG_ON(flags & GFP_SLAB_BUG_MASK);
Christoph Lameter's avatar
Christoph Lameter committed
1090

Christoph Lameter's avatar
Christoph Lameter committed
1091
1092
	page = allocate_slab(s,
		flags & (GFP_RECLAIM_MASK | GFP_CONSTRAINT_MASK), node);
Christoph Lameter's avatar
Christoph Lameter committed
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
	if (!page)
		goto out;

	n = get_node(s, page_to_nid(page));
	if (n)
		atomic_long_inc(&n->nr_slabs);
	page->slab = s;
	page->flags |= 1 << PG_slab;
	if (s->flags & (SLAB_DEBUG_FREE | SLAB_RED_ZONE | SLAB_POISON |
			SLAB_STORE_USER | SLAB_TRACE))
1103
		SetSlabDebug(page);
Christoph Lameter's avatar
Christoph Lameter committed
1104
1105
1106
1107
1108
1109
1110

	start = page_address(page);

	if (unlikely(s->flags & SLAB_POISON))
		memset(start, POISON_INUSE, PAGE_SIZE << s->order);

	last = start;
1111
	for_each_object(p, s, start) {
Christoph Lameter's avatar
Christoph Lameter committed
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
		setup_object(s, page, last);
		set_freepointer(s, last, p);
		last = p;
	}
	setup_object(s, page, last);
	set_freepointer(s, last, NULL);

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

static void __free_slab(struct kmem_cache *s, struct page *page)
{
	int pages = 1 << s->order;

1129
	if (unlikely(SlabDebug(page))) {
Christoph Lameter's avatar
Christoph Lameter committed
1130
1131
1132
		void *p;

		slab_pad_check(s, page);
1133
		for_each_object(p, s, page_address(page))
Christoph Lameter's avatar
Christoph Lameter committed
1134
			check_object(s, page, p, 0);
1135
		ClearSlabDebug(page);
Christoph Lameter's avatar
Christoph Lameter committed
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
	}

	mod_zone_page_state(page_zone(page),
		(s->flags & SLAB_RECLAIM_ACCOUNT) ?
		NR_SLAB_RECLAIMABLE : NR_SLAB_UNRECLAIMABLE,
		- pages);

	__free_pages(page, s->order);
}

static void rcu_free_slab(struct rcu_head *h)
{
	struct page *page;

	page = container_of((struct list_head *)h, struct page, lru);
	__free_slab(page->slab, page);
}

static void free_slab(struct kmem_cache *s, struct page *page)
{
	if (unlikely(s->flags & SLAB_DESTROY_BY_RCU)) {
		/*
		 * RCU free overloads the RCU head over the LRU
		 */
		struct rcu_head *head = (void *)&page->lru;

		call_rcu(head, rcu_free_slab);
	} else
		__free_slab(s, page);
}

static void discard_slab(struct kmem_cache *s, struct page *page)
{
	struct kmem_cache_node *n = get_node(s, page_to_nid(page));

	atomic_long_dec(&n->nr_slabs);
	reset_page_mapcount(page);
1173
	__ClearPageSlab(page);
Christoph Lameter's avatar
Christoph Lameter committed
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
	free_slab(s, page);
}

/*
 * Per slab locking using the pagelock
 */
static __always_inline void slab_lock(struct page *page)
{
	bit_spin_lock(PG_locked, &page->flags);
}

static __always_inline void slab_unlock(struct page *page)
{
	bit_spin_unlock(PG_locked, &page->flags);
}

static __always_inline int slab_trylock(struct page *page)
{
	int rc = 1;

	rc = bit_spin_trylock(PG_locked, &page->flags);
	return rc;
}

/*
 * Management of partially allocated slabs
 */
1201
1202
static void add_partial(struct kmem_cache_node *n,
				struct page *page, int tail)
Christoph Lameter's avatar
Christoph Lameter committed
1203
{
Christoph Lameter's avatar
Christoph Lameter committed
1204
1205
	spin_lock(&n->list_lock);
	n->nr_partial++;
1206
1207
1208
1209
	if (tail)
		list_add_tail(&page->lru, &n->partial);
	else
		list_add(&page->lru, &n->partial);
Christoph Lameter's avatar
Christoph Lameter committed
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
	spin_unlock(&n->list_lock);
}

static void remove_partial(struct kmem_cache *s,
						struct page *page)
{
	struct kmem_cache_node *n = get_node(s, page_to_nid(page));

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

/*
Christoph Lameter's avatar
Christoph Lameter committed
1225
 * Lock slab and remove from the partial list.
Christoph Lameter's avatar
Christoph Lameter committed
1226
 *
Christoph Lameter's avatar
Christoph Lameter committed
1227
 * Must hold list_lock.
Christoph Lameter's avatar
Christoph Lameter committed
1228
 */
1229
static inline int lock_and_freeze_slab(struct kmem_cache_node *n, struct page *page)
Christoph Lameter's avatar
Christoph Lameter committed
1230
1231
1232
1233
{
	if (slab_trylock(page)) {
		list_del(&page->lru);
		n->nr_partial--;
1234
		SetSlabFrozen(page);
Christoph Lameter's avatar
Christoph Lameter committed
1235
1236
1237
1238
1239
1240
		return 1;
	}
	return 0;
}

/*
Christoph Lameter's avatar
Christoph Lameter committed
1241
 * Try to allocate a partial slab from a specific node.
Christoph Lameter's avatar
Christoph Lameter committed
1242
1243
1244
1245
1246
1247
1248
1249
 */
static struct page *get_partial_node(struct kmem_cache_node *n)
{
	struct page *page;

	/*
	 * Racy check. If we mistakenly see no partial slabs then we
	 * just allocate an empty slab. If we mistakenly try to get a
Christoph Lameter's avatar
Christoph Lameter committed
1250
1251
	 * partial slab and there is none available then get_partials()
	 * will return NULL.
Christoph Lameter's avatar
Christoph Lameter committed
1252
1253
1254
1255
1256
1257
	 */
	if (!n || !n->nr_partial)
		return NULL;

	spin_lock(&n->list_lock);
	list_for_each_entry(page, &n->partial, lru)
1258
		if (lock_and_freeze_slab(n, page))
Christoph Lameter's avatar
Christoph Lameter committed
1259
1260
1261
1262
1263
1264
1265
1266
			goto out;
	page = NULL;
out:
	spin_unlock(&n->list_lock);
	return page;
}

/*
Christoph Lameter's avatar
Christoph Lameter committed
1267
 * Get a page from somewhere. Search in increasing NUMA distances.
Christoph Lameter's avatar
Christoph Lameter committed
1268
1269
1270
1271
1272
1273
1274
1275
1276
 */
static struct page *get_any_partial(struct kmem_cache *s, gfp_t flags)
{
#ifdef CONFIG_NUMA
	struct zonelist *zonelist;
	struct zone **z;
	struct page *page;

	/*
Christoph Lameter's avatar
Christoph Lameter committed
1277
1278
1279
1280
	 * The defrag ratio allows a configuration of the tradeoffs between
	 * inter node defragmentation and node local allocations. A lower
	 * defrag_ratio increases the tendency to do local allocations
	 * instead of attempting to obtain partial slabs from other nodes.
Christoph Lameter's avatar
Christoph Lameter committed
1281
	 *
Christoph Lameter's avatar
Christoph Lameter committed
1282
1283
1284
1285
	 * If the defrag_ratio is set to 0 then kmalloc() always
	 * returns node local objects. If the ratio is higher then kmalloc()
	 * may return off node objects because partial slabs are obtained
	 * from other nodes and filled up.
Christoph Lameter's avatar
Christoph Lameter committed
1286
1287
	 *
	 * If /sys/slab/xx/defrag_ratio is set to 100 (which makes
Christoph Lameter's avatar
Christoph Lameter committed
1288
1289
1290
1291
1292
	 * defrag_ratio = 1000) then every (well almost) allocation will
	 * first attempt to defrag slab caches on other nodes. This means
	 * scanning over all nodes to look for partial slabs which may be
	 * expensive if we do it every time we are trying to find a slab
	 * with available objects.
Christoph Lameter's avatar
Christoph Lameter committed
1293
	 */
1294
1295
	if (!s->remote_node_defrag_ratio ||
			get_cycles() % 1024 > s->remote_node_defrag_ratio)
Christoph Lameter's avatar
Christoph Lameter committed
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
		return NULL;

	zonelist = &NODE_DATA(slab_node(current->mempolicy))
					->node_zonelists[gfp_zone(flags)];
	for (z = zonelist->zones; *z; z++) {
		struct kmem_cache_node *n;

		n = get_node(s, zone_to_nid(*z));

		if (n && cpuset_zone_allowed_hardwall(*z, flags) &&
Christoph Lameter's avatar
Christoph Lameter committed
1306
				n->nr_partial > MIN_PARTIAL) {
Christoph Lameter's avatar
Christoph Lameter committed
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
			page = get_partial_node(n);
			if (page)
				return page;
		}
	}
#endif
	return NULL;
}

/*
 * Get a partial page, lock it and return it.
 */
static struct page *get_partial(struct kmem_cache *s, gfp_t flags, int node)
{
	struct page *page;
	int searchnode = (node == -1) ? numa_node_id() : node;

	page = get_partial_node(get_node(s, searchnode));
	if (page || (flags & __GFP_THISNODE))
		return page;

	return get_any_partial(s, flags);
}

/*
 * Move a page back to the lists.
 *
 * Must be called with the slab lock held.
 *
 * On exit the slab lock will have been dropped.
 */
1338
static void unfreeze_slab(struct kmem_cache *s, struct page *page, int tail)
Christoph Lameter's avatar
Christoph Lameter committed
1339
{
Christoph Lameter's avatar
Christoph Lameter committed
1340
1341
	struct kmem_cache_node *n = get_node(s, page_to_nid(page));

1342
	ClearSlabFrozen(page);
Christoph Lameter's avatar
Christoph Lameter committed
1343
	if (page->inuse) {
Christoph Lameter's avatar
Christoph Lameter committed
1344

Christoph Lameter's avatar
Christoph Lameter committed
1345
		if (page->freelist)
1346
			add_partial(n, page, tail);
1347
		else if (SlabDebug(page) && (s->flags & SLAB_STORE_USER))
Christoph Lameter's avatar
Christoph Lameter committed
1348
			add_full(n, page);
Christoph Lameter's avatar
Christoph Lameter committed
1349
		slab_unlock(page);
Christoph Lameter's avatar
Christoph Lameter committed
1350

Christoph Lameter's avatar
Christoph Lameter committed
1351
	} else {
Christoph Lameter's avatar
Christoph Lameter committed
1352
1353
		if (n->nr_partial < MIN_PARTIAL) {
			/*
Christoph Lameter's avatar
Christoph Lameter committed
1354
1355
1356
1357
1358
1359
			 * Adding an empty slab to the partial slabs in order
			 * to avoid page allocator overhead. This slab needs
			 * to come after the other slabs with objects in
			 * order to fill them up. That way the size of the
			 * partial list stays small. kmem_cache_shrink can
			 * reclaim empty slabs from the partial list.
Christoph Lameter's avatar
Christoph Lameter committed
1360
			 */
1361
			add_partial(n, page, 1);
Christoph Lameter's avatar
Christoph Lameter committed
1362
1363
1364
1365
1366
			slab_unlock(page);
		} else {
			slab_unlock(page);
			discard_slab(s, page);
		}
Christoph Lameter's avatar
Christoph Lameter committed
1367
1368
1369
1370
1371
1372
	}
}

/*
 * Remove the cpu slab
 */
1373
static void deactivate_slab(struct kmem_cache *s, struct kmem_cache_cpu *c)
Christoph Lameter's avatar
Christoph Lameter committed
1374
{
1375
	struct page *page = c->page;
1376
	int tail = 1;
1377
1378
1379
1380
1381
	/*
	 * Merge cpu freelist into freelist. Typically we get here
	 * because both freelists are empty. So this is unlikely
	 * to occur.
	 */
1382
	while (unlikely(c->freelist)) {
1383
1384
		void **object;

1385
1386
		tail = 0;	/* Hot objects. Put the slab first */

1387
		/* Retrieve object from cpu_freelist */
1388
		object = c->freelist;
1389
		c->freelist = c->freelist[c->offset];
1390
1391

		/* And put onto the regular freelist */
1392
		object[c->offset] = page->freelist;
1393
1394
1395
		page->freelist = object;
		page->inuse--;
	}
1396
	c->page = NULL;
1397
	unfreeze_slab(s, page, tail);
Christoph Lameter's avatar
Christoph Lameter committed
1398
1399
}

1400
static inline void flush_slab(struct kmem_cache *s, struct kmem_cache_cpu *c)
Christoph Lameter's avatar
Christoph Lameter committed
1401
{
1402
1403
	slab_lock(c->page);
	deactivate_slab(s, c);
Christoph Lameter's avatar
Christoph Lameter committed
1404
1405
1406
1407
1408
1409
}

/*
 * Flush cpu slab.
 * Called from IPI handler with interrupts disabled.
 */
1410
static inline void __flush_cpu_slab(struct kmem_cache *s, int cpu)
Christoph Lameter's avatar
Christoph Lameter committed
1411
{
1412
	struct kmem_cache_cpu *c = get_cpu_slab(s, cpu);
Christoph Lameter's avatar
Christoph Lameter committed
1413

1414
1415
	if (likely(c && c->page))
		flush_slab(s, c);
Christoph Lameter's avatar
Christoph Lameter committed
1416
1417
1418
1419
1420
1421
}

static void flush_cpu_slab(void *d)
{
	struct kmem_cache *s = d;

1422
	__flush_cpu_slab(s, smp_processor_id());
Christoph Lameter's avatar
Christoph Lameter committed
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
}

static void flush_all(struct kmem_cache *s)
{
#ifdef CONFIG_SMP
	on_each_cpu(flush_cpu_slab, s, 1, 1);
#else
	unsigned long flags;

	local_irq_save(flags);
	flush_cpu_slab(s);
	local_irq_restore(flags);
#endif
}

1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
/*
 * Check if the objects in a per cpu structure fit numa
 * locality expectations.
 */
static inline int node_match(struct kmem_cache_cpu *c, int node)
{
#ifdef CONFIG_NUMA
	if (node != -1 && c->node != node)
		return 0;
#endif
	return 1;
}

Christoph Lameter's avatar
Christoph Lameter committed
1451
/*
1452
1453
1454
1455
 * Slow path. The lockless freelist is empty or we need to perform
 * debugging duties.
 *
 * Interrupts are disabled.
Christoph Lameter's avatar
Christoph Lameter committed
1456
 *
1457
1458
1459
 * Processing is still very fast if new objects have been freed to the
 * regular freelist. In that case we simply take over the regular freelist
 * as the lockless freelist and zap the regular freelist.
Christoph Lameter's avatar
Christoph Lameter committed
1460
 *
1461
1462
1463
 * If that is not working then we fall back to the partial lists. We take the
 * first element of the freelist as the object to allocate now and move the
 * rest of the freelist to the lockless freelist.
Christoph Lameter's avatar
Christoph Lameter committed
1464
 *
1465
1466
 * And if we were unable to get a new slab from the partial slab lists then
 * we need to allocate a new slab. This is slowest path since we may sleep.
Christoph Lameter's avatar
Christoph Lameter committed
1467
 */
1468
static void *__slab_alloc(struct kmem_cache *s,
1469
		gfp_t gfpflags, int node, void *addr, struct kmem_cache_cpu *c)
Christoph Lameter's avatar
Christoph Lameter committed
1470
1471
{
	void **object;
1472
	struct page *new;
Christoph Lameter's avatar
Christoph Lameter committed
1473

1474
	if (!c->page)
Christoph Lameter's avatar
Christoph Lameter committed
1475
1476
		goto new_slab;

1477
1478
	slab_lock(c->page);
	if (unlikely(!node_match(c, node)))
Christoph Lameter's avatar
Christoph Lameter committed
1479
		goto another_slab;
1480
load_freelist:
1481
	object = c->page->freelist;
Christoph Lameter's avatar
Christoph Lameter committed
1482
1483
	if (unlikely(!object))
		goto another_slab;
1484
	if (unlikely(SlabDebug(c->page)))
Christoph Lameter's avatar
Christoph Lameter committed
1485
1486
		goto debug;

1487
	object = c->page->freelist;
1488
	c->freelist = object[c->offset];
1489
1490
1491
1492
	c->page->inuse = s->objects;
	c->page->freelist = NULL;
	c->node = page_to_nid(c->page);
	slab_unlock(c->page);
Christoph Lameter's avatar
Christoph Lameter committed
1493
1494
1495
	return object;

another_slab:
1496
	deactivate_slab(s, c);
Christoph Lameter's avatar
Christoph Lameter committed
1497
1498

new_slab:
1499
1500
1501
	new = get_partial(s, gfpflags, node);
	if (new) {
		c->page = new;
1502
		goto load_freelist;
Christoph Lameter's avatar
Christoph Lameter committed
1503
1504
	}

1505
1506
1507
	if (gfpflags & __GFP_WAIT)
		local_irq_enable();

1508
	new = new_slab(s, gfpflags, node);
1509
1510
1511
1512

	if (gfpflags & __GFP_WAIT)
		local_irq_disable();

1513
1514
	if (new) {
		c = get_cpu_slab(s, smp_processor_id());
1515
		if (c->page)