slab.c 116 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
/*
 * linux/mm/slab.c
 * Written by Mark Hemment, 1996/97.
 * (markhe@nextd.demon.co.uk)
 *
 * kmem_cache_destroy() + some cleanup - 1999 Andrea Arcangeli
 *
 * Major cleanup, different bufctl logic, per-cpu arrays
 *	(c) 2000 Manfred Spraul
 *
 * Cleanup, make the head arrays unconditional, preparation for NUMA
 * 	(c) 2002 Manfred Spraul
 *
 * An implementation of the Slab Allocator as described in outline in;
 *	UNIX Internals: The New Frontiers by Uresh Vahalia
 *	Pub: Prentice Hall	ISBN 0-13-101908-2
 * or with a little more detail in;
 *	The Slab Allocator: An Object-Caching Kernel Memory Allocator
 *	Jeff Bonwick (Sun Microsystems).
 *	Presented at: USENIX Summer 1994 Technical Conference
 *
 * The memory is organized in caches, one cache for each object type.
 * (e.g. inode_cache, dentry_cache, buffer_head, vm_area_struct)
 * Each cache consists out of many slabs (they are small (usually one
 * page long) and always contiguous), and each slab contains multiple
 * initialized objects.
 *
 * This means, that your constructor is used only for newly allocated
Simon Arlott's avatar
Simon Arlott committed
29
 * slabs and you must pass objects with the same initializations to
Linus Torvalds's avatar
Linus Torvalds committed
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
 * kmem_cache_free.
 *
 * Each cache can only support one memory type (GFP_DMA, GFP_HIGHMEM,
 * normal). If you need a special memory type, then must create a new
 * cache for that memory type.
 *
 * In order to reduce fragmentation, the slabs are sorted in 3 groups:
 *   full slabs with 0 free objects
 *   partial slabs
 *   empty slabs with no allocated objects
 *
 * If partial slabs exist, then new allocations come from these slabs,
 * otherwise from empty slabs or new slabs are allocated.
 *
 * kmem_cache_destroy() CAN CRASH if you try to allocate from the cache
 * during kmem_cache_destroy(). The caller must prevent concurrent allocs.
 *
 * Each cache has a short per-cpu head array, most allocs
 * and frees go into that array, and if that array overflows, then 1/2
 * of the entries in the array are given back into the global cache.
 * The head array is strictly LIFO and should improve the cache hit rates.
 * On SMP, it additionally reduces the spinlock operations.
 *
Andrew Morton's avatar
Andrew Morton committed
53
 * The c_cpuarray may not be read with enabled local interrupts -
Linus Torvalds's avatar
Linus Torvalds committed
54
55
56
57
 * it's changed with a smp_call_function().
 *
 * SMP synchronization:
 *  constructors and destructors are called without any locking.
58
 *  Several members in struct kmem_cache and struct slab never change, they
Linus Torvalds's avatar
Linus Torvalds committed
59
60
61
62
63
64
65
66
67
68
69
70
 *	are accessed without any locking.
 *  The per-cpu arrays are never accessed from the wrong cpu, no locking,
 *  	and local interrupts are disabled so slab code is preempt-safe.
 *  The non-constant members are protected with a per-cache irq spinlock.
 *
 * Many thanks to Mark Hemment, who wrote another per-cpu slab patch
 * in 2000 - many ideas in the current implementation are derived from
 * his patch.
 *
 * Further notes from the original documentation:
 *
 * 11 April '97.  Started multi-threading - markhe
71
 *	The global cache-chain is protected by the mutex 'slab_mutex'.
Linus Torvalds's avatar
Linus Torvalds committed
72
73
74
75
76
77
 *	The sem is only needed when accessing/extending the cache-chain, which
 *	can never happen inside an interrupt (kmem_cache_create(),
 *	kmem_cache_shrink() and kmem_cache_reap()).
 *
 *	At present, each engine can be growing a cache.  This should be blocked.
 *
78
79
80
81
82
83
84
85
86
 * 15 March 2005. NUMA slab allocator.
 *	Shai Fultheim <shai@scalex86.org>.
 *	Shobhit Dayal <shobhit@calsoftinc.com>
 *	Alok N Kataria <alokk@calsoftinc.com>
 *	Christoph Lameter <christoph@lameter.com>
 *
 *	Modified the slab allocator to be node aware on NUMA systems.
 *	Each node has its own list of partial, free and full slabs.
 *	All object allocations for a node occur from node specific slab lists.
Linus Torvalds's avatar
Linus Torvalds committed
87
88
89
 */

#include	<linux/slab.h>
90
#include	"slab.h"
Linus Torvalds's avatar
Linus Torvalds committed
91
#include	<linux/mm.h>
92
#include	<linux/poison.h>
Linus Torvalds's avatar
Linus Torvalds committed
93
94
95
96
97
#include	<linux/swap.h>
#include	<linux/cache.h>
#include	<linux/interrupt.h>
#include	<linux/init.h>
#include	<linux/compiler.h>
98
#include	<linux/cpuset.h>
99
#include	<linux/proc_fs.h>
Linus Torvalds's avatar
Linus Torvalds committed
100
101
102
103
104
105
106
#include	<linux/seq_file.h>
#include	<linux/notifier.h>
#include	<linux/kallsyms.h>
#include	<linux/cpu.h>
#include	<linux/sysctl.h>
#include	<linux/module.h>
#include	<linux/rcupdate.h>
107
#include	<linux/string.h>
108
#include	<linux/uaccess.h>
109
#include	<linux/nodemask.h>
110
#include	<linux/kmemleak.h>
111
#include	<linux/mempolicy.h>
Ingo Molnar's avatar
Ingo Molnar committed
112
#include	<linux/mutex.h>
113
#include	<linux/fault-inject.h>
Ingo Molnar's avatar
Ingo Molnar committed
114
#include	<linux/rtmutex.h>
115
#include	<linux/reciprocal_div.h>
116
#include	<linux/debugobjects.h>
Pekka Enberg's avatar
Pekka Enberg committed
117
#include	<linux/kmemcheck.h>
118
#include	<linux/memory.h>
119
#include	<linux/prefetch.h>
Linus Torvalds's avatar
Linus Torvalds committed
120

121
122
#include	<net/sock.h>

Linus Torvalds's avatar
Linus Torvalds committed
123
124
125
126
#include	<asm/cacheflush.h>
#include	<asm/tlbflush.h>
#include	<asm/page.h>

127
128
#include <trace/events/kmem.h>

129
130
#include	"internal.h"

Linus Torvalds's avatar
Linus Torvalds committed
131
/*
132
 * DEBUG	- 1 for kmem_cache_create() to honour; SLAB_RED_ZONE & SLAB_POISON.
Linus Torvalds's avatar
Linus Torvalds committed
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
 *		  0 for faster, smaller code (especially in the critical paths).
 *
 * STATS	- 1 to collect stats for /proc/slabinfo.
 *		  0 for faster, smaller code (especially in the critical paths).
 *
 * FORCED_DEBUG	- 1 enables SLAB_RED_ZONE and SLAB_POISON (if possible)
 */

#ifdef CONFIG_DEBUG_SLAB
#define	DEBUG		1
#define	STATS		1
#define	FORCED_DEBUG	1
#else
#define	DEBUG		0
#define	STATS		0
#define	FORCED_DEBUG	0
#endif

/* Shouldn't this be in a header file somewhere? */
#define	BYTES_PER_WORD		sizeof(void *)
David Woodhouse's avatar
David Woodhouse committed
153
#define	REDZONE_ALIGN		max(BYTES_PER_WORD, __alignof__(unsigned long long))
Linus Torvalds's avatar
Linus Torvalds committed
154
155
156
157
158

#ifndef ARCH_KMALLOC_FLAGS
#define ARCH_KMALLOC_FLAGS SLAB_HWCACHE_ALIGN
#endif

159
160
161
162
163
164
/*
 * true if a page was allocated from pfmemalloc reserves for network-based
 * swap
 */
static bool pfmemalloc_active __read_mostly;

Linus Torvalds's avatar
Linus Torvalds committed
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
/*
 * kmem_bufctl_t:
 *
 * Bufctl's are used for linking objs within a slab
 * linked offsets.
 *
 * This implementation relies on "struct page" for locating the cache &
 * slab an object belongs to.
 * This allows the bufctl structure to be small (one int), but limits
 * the number of objects a slab (not a cache) can contain when off-slab
 * bufctls are used. The limit is the size of the largest general cache
 * that does not use off-slab slabs.
 * For 32bit archs with 4 kB pages, is this 56.
 * This is not serious, as it is only for large objects, when it is unwise
 * to have too many per slab.
 * Note: This limit can be raised by introducing a general cache whose size
 * is less than 512 (PAGE_SIZE<<3), but greater than 256.
 */

184
typedef unsigned int kmem_bufctl_t;
Linus Torvalds's avatar
Linus Torvalds committed
185
186
#define BUFCTL_END	(((kmem_bufctl_t)(~0U))-0)
#define BUFCTL_FREE	(((kmem_bufctl_t)(~0U))-1)
187
188
#define	BUFCTL_ACTIVE	(((kmem_bufctl_t)(~0U))-2)
#define	SLAB_LIMIT	(((kmem_bufctl_t)(~0U))-3)
Linus Torvalds's avatar
Linus Torvalds committed
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204

/*
 * struct slab_rcu
 *
 * slab_destroy on a SLAB_DESTROY_BY_RCU cache uses this structure to
 * arrange for kmem_freepages to be called via RCU.  This is useful if
 * we need to approach a kernel structure obliquely, from its address
 * obtained without the usual locking.  We can lock the structure to
 * stabilize it and check it's still at the given address, only if we
 * can be sure that the memory has not been meanwhile reused for some
 * other kind of object (which our subsystem's lock might corrupt).
 *
 * rcu_read_lock before reading the address, then rcu_read_unlock after
 * taking the spinlock within the structure expected at that address.
 */
struct slab_rcu {
205
	struct rcu_head head;
206
	struct kmem_cache *cachep;
207
	void *addr;
Linus Torvalds's avatar
Linus Torvalds committed
208
209
};

210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
/*
 * struct slab
 *
 * Manages the objs in a slab. Placed either at the beginning of mem allocated
 * for a slab, or allocated from an general cache.
 * Slabs are chained into three list: fully used, partial, fully free slabs.
 */
struct slab {
	union {
		struct {
			struct list_head list;
			unsigned long colouroff;
			void *s_mem;		/* including colour offset */
			unsigned int inuse;	/* num of objs active in slab */
			kmem_bufctl_t free;
			unsigned short nodeid;
		};
		struct slab_rcu __slab_cover_slab_rcu;
	};
};

Linus Torvalds's avatar
Linus Torvalds committed
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
/*
 * struct array_cache
 *
 * Purpose:
 * - LIFO ordering, to hand out cache-warm objects from _alloc
 * - reduce the number of linked list operations
 * - reduce spinlock operations
 *
 * The limit is stored in the per-cpu structure to reduce the data cache
 * footprint.
 *
 */
struct array_cache {
	unsigned int avail;
	unsigned int limit;
	unsigned int batchcount;
	unsigned int touched;
248
	spinlock_t lock;
249
	void *entry[];	/*
Andrew Morton's avatar
Andrew Morton committed
250
251
252
			 * Must have this definition in here for the proper
			 * alignment of array_cache. Also simplifies accessing
			 * the entries.
253
254
255
256
			 *
			 * Entries should not be directly dereferenced as
			 * entries belonging to slabs marked pfmemalloc will
			 * have the lower bits set SLAB_OBJ_PFMEMALLOC
Andrew Morton's avatar
Andrew Morton committed
257
			 */
Linus Torvalds's avatar
Linus Torvalds committed
258
259
};

260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
#define SLAB_OBJ_PFMEMALLOC	1
static inline bool is_obj_pfmemalloc(void *objp)
{
	return (unsigned long)objp & SLAB_OBJ_PFMEMALLOC;
}

static inline void set_obj_pfmemalloc(void **objp)
{
	*objp = (void *)((unsigned long)*objp | SLAB_OBJ_PFMEMALLOC);
	return;
}

static inline void clear_obj_pfmemalloc(void **objp)
{
	*objp = (void *)((unsigned long)*objp & ~SLAB_OBJ_PFMEMALLOC);
}

Andrew Morton's avatar
Andrew Morton committed
277
278
279
/*
 * bootstrap: The caches do not work without cpuarrays anymore, but the
 * cpuarrays are allocated from the generic caches...
Linus Torvalds's avatar
Linus Torvalds committed
280
281
282
283
 */
#define BOOT_CPUCACHE_ENTRIES	1
struct arraycache_init {
	struct array_cache cache;
284
	void *entries[BOOT_CPUCACHE_ENTRIES];
Linus Torvalds's avatar
Linus Torvalds committed
285
286
287
};

/*
288
 * The slab lists for all objects.
Linus Torvalds's avatar
Linus Torvalds committed
289
290
 */
struct kmem_list3 {
291
292
293
294
295
	struct list_head slabs_partial;	/* partial list first, better asm code */
	struct list_head slabs_full;
	struct list_head slabs_free;
	unsigned long free_objects;
	unsigned int free_limit;
296
	unsigned int colour_next;	/* Per-node cache coloring */
297
298
299
	spinlock_t list_lock;
	struct array_cache *shared;	/* shared per node */
	struct array_cache **alien;	/* on other nodes */
300
301
	unsigned long next_reap;	/* updated without locking */
	int free_touched;		/* updated without locking */
Linus Torvalds's avatar
Linus Torvalds committed
302
303
};

304
305
306
/*
 * Need this for bootstrapping a per node allocator.
 */
307
#define NUM_INIT_LISTS (3 * MAX_NUMNODES)
308
static struct kmem_list3 __initdata initkmem_list3[NUM_INIT_LISTS];
309
#define	CACHE_CACHE 0
310
311
#define	SIZE_AC MAX_NUMNODES
#define	SIZE_L3 (2 * MAX_NUMNODES)
312

313
314
315
316
static int drain_freelist(struct kmem_cache *cache,
			struct kmem_list3 *l3, int tofree);
static void free_block(struct kmem_cache *cachep, void **objpp, int len,
			int node);
317
static int enable_cpucache(struct kmem_cache *cachep, gfp_t gfp);
318
static void cache_reap(struct work_struct *unused);
319

320
/*
Andrew Morton's avatar
Andrew Morton committed
321
322
 * This function must be completely optimized away if a constant is passed to
 * it.  Mostly the same as what is in linux/slab.h except it returns an index.
323
 */
324
static __always_inline int index_of(const size_t size)
325
{
326
327
	extern void __bad_size(void);

328
329
330
331
332
333
334
335
	if (__builtin_constant_p(size)) {
		int i = 0;

#define CACHE(x) \
	if (size <=x) \
		return i; \
	else \
		i++;
336
#include <linux/kmalloc_sizes.h>
337
#undef CACHE
338
		__bad_size();
339
	} else
340
		__bad_size();
341
342
343
	return 0;
}

344
345
static int slab_early_init = 1;

346
347
#define INDEX_AC index_of(sizeof(struct arraycache_init))
#define INDEX_L3 index_of(sizeof(struct kmem_list3))
Linus Torvalds's avatar
Linus Torvalds committed
348

Pekka Enberg's avatar
Pekka Enberg committed
349
static void kmem_list3_init(struct kmem_list3 *parent)
350
351
352
353
354
355
{
	INIT_LIST_HEAD(&parent->slabs_full);
	INIT_LIST_HEAD(&parent->slabs_partial);
	INIT_LIST_HEAD(&parent->slabs_free);
	parent->shared = NULL;
	parent->alien = NULL;
356
	parent->colour_next = 0;
357
358
359
360
361
	spin_lock_init(&parent->list_lock);
	parent->free_objects = 0;
	parent->free_touched = 0;
}

Andrew Morton's avatar
Andrew Morton committed
362
363
364
365
#define MAKE_LIST(cachep, listp, slab, nodeid)				\
	do {								\
		INIT_LIST_HEAD(listp);					\
		list_splice(&(cachep->nodelists[nodeid]->slab), listp);	\
366
367
	} while (0)

Andrew Morton's avatar
Andrew Morton committed
368
369
#define	MAKE_ALL_LISTS(cachep, ptr, nodeid)				\
	do {								\
370
371
372
373
	MAKE_LIST((cachep), (&(ptr)->slabs_full), slabs_full, nodeid);	\
	MAKE_LIST((cachep), (&(ptr)->slabs_partial), slabs_partial, nodeid); \
	MAKE_LIST((cachep), (&(ptr)->slabs_free), slabs_free, nodeid);	\
	} while (0)
Linus Torvalds's avatar
Linus Torvalds committed
374
375
376
377
378

#define CFLGS_OFF_SLAB		(0x80000000UL)
#define	OFF_SLAB(x)	((x)->flags & CFLGS_OFF_SLAB)

#define BATCHREFILL_LIMIT	16
Andrew Morton's avatar
Andrew Morton committed
379
380
381
/*
 * Optimization question: fewer reaps means less probability for unnessary
 * cpucache drain/refill cycles.
Linus Torvalds's avatar
Linus Torvalds committed
382
 *
Adrian Bunk's avatar
Adrian Bunk committed
383
 * OTOH the cpuarrays can contain lots of objects,
Linus Torvalds's avatar
Linus Torvalds committed
384
385
386
387
388
389
390
391
392
393
 * which could lock up otherwise freeable slabs.
 */
#define REAPTIMEOUT_CPUC	(2*HZ)
#define REAPTIMEOUT_LIST3	(4*HZ)

#if STATS
#define	STATS_INC_ACTIVE(x)	((x)->num_active++)
#define	STATS_DEC_ACTIVE(x)	((x)->num_active--)
#define	STATS_INC_ALLOCED(x)	((x)->num_allocations++)
#define	STATS_INC_GROWN(x)	((x)->grown++)
394
#define	STATS_ADD_REAPED(x,y)	((x)->reaped += (y))
Andrew Morton's avatar
Andrew Morton committed
395
396
397
398
399
#define	STATS_SET_HIGH(x)						\
	do {								\
		if ((x)->num_active > (x)->high_mark)			\
			(x)->high_mark = (x)->num_active;		\
	} while (0)
Linus Torvalds's avatar
Linus Torvalds committed
400
401
#define	STATS_INC_ERR(x)	((x)->errors++)
#define	STATS_INC_NODEALLOCS(x)	((x)->node_allocs++)
402
#define	STATS_INC_NODEFREES(x)	((x)->node_frees++)
403
#define STATS_INC_ACOVERFLOW(x)   ((x)->node_overflow++)
Andrew Morton's avatar
Andrew Morton committed
404
405
406
407
408
#define	STATS_SET_FREEABLE(x, i)					\
	do {								\
		if ((x)->max_freeable < i)				\
			(x)->max_freeable = i;				\
	} while (0)
Linus Torvalds's avatar
Linus Torvalds committed
409
410
411
412
413
414
415
416
417
#define STATS_INC_ALLOCHIT(x)	atomic_inc(&(x)->allochit)
#define STATS_INC_ALLOCMISS(x)	atomic_inc(&(x)->allocmiss)
#define STATS_INC_FREEHIT(x)	atomic_inc(&(x)->freehit)
#define STATS_INC_FREEMISS(x)	atomic_inc(&(x)->freemiss)
#else
#define	STATS_INC_ACTIVE(x)	do { } while (0)
#define	STATS_DEC_ACTIVE(x)	do { } while (0)
#define	STATS_INC_ALLOCED(x)	do { } while (0)
#define	STATS_INC_GROWN(x)	do { } while (0)
418
#define	STATS_ADD_REAPED(x,y)	do { (void)(y); } while (0)
Linus Torvalds's avatar
Linus Torvalds committed
419
420
421
#define	STATS_SET_HIGH(x)	do { } while (0)
#define	STATS_INC_ERR(x)	do { } while (0)
#define	STATS_INC_NODEALLOCS(x)	do { } while (0)
422
#define	STATS_INC_NODEFREES(x)	do { } while (0)
423
#define STATS_INC_ACOVERFLOW(x)   do { } while (0)
Andrew Morton's avatar
Andrew Morton committed
424
#define	STATS_SET_FREEABLE(x, i) do { } while (0)
Linus Torvalds's avatar
Linus Torvalds committed
425
426
427
428
429
430
431
432
#define STATS_INC_ALLOCHIT(x)	do { } while (0)
#define STATS_INC_ALLOCMISS(x)	do { } while (0)
#define STATS_INC_FREEHIT(x)	do { } while (0)
#define STATS_INC_FREEMISS(x)	do { } while (0)
#endif

#if DEBUG

Andrew Morton's avatar
Andrew Morton committed
433
434
/*
 * memory layout of objects:
Linus Torvalds's avatar
Linus Torvalds committed
435
 * 0		: objp
436
 * 0 .. cachep->obj_offset - BYTES_PER_WORD - 1: padding. This ensures that
Linus Torvalds's avatar
Linus Torvalds committed
437
438
 * 		the end of an object is aligned with the end of the real
 * 		allocation. Catches writes behind the end of the allocation.
439
 * cachep->obj_offset - BYTES_PER_WORD .. cachep->obj_offset - 1:
Linus Torvalds's avatar
Linus Torvalds committed
440
 * 		redzone word.
441
 * cachep->obj_offset: The real object.
442
443
 * cachep->size - 2* BYTES_PER_WORD: redzone word [BYTES_PER_WORD long]
 * cachep->size - 1* BYTES_PER_WORD: last caller address
Andrew Morton's avatar
Andrew Morton committed
444
 *					[BYTES_PER_WORD long]
Linus Torvalds's avatar
Linus Torvalds committed
445
 */
446
static int obj_offset(struct kmem_cache *cachep)
Linus Torvalds's avatar
Linus Torvalds committed
447
{
448
	return cachep->obj_offset;
Linus Torvalds's avatar
Linus Torvalds committed
449
450
}

451
static unsigned long long *dbg_redzone1(struct kmem_cache *cachep, void *objp)
Linus Torvalds's avatar
Linus Torvalds committed
452
453
{
	BUG_ON(!(cachep->flags & SLAB_RED_ZONE));
454
455
	return (unsigned long long*) (objp + obj_offset(cachep) -
				      sizeof(unsigned long long));
Linus Torvalds's avatar
Linus Torvalds committed
456
457
}

458
static unsigned long long *dbg_redzone2(struct kmem_cache *cachep, void *objp)
Linus Torvalds's avatar
Linus Torvalds committed
459
460
461
{
	BUG_ON(!(cachep->flags & SLAB_RED_ZONE));
	if (cachep->flags & SLAB_STORE_USER)
462
		return (unsigned long long *)(objp + cachep->size -
463
					      sizeof(unsigned long long) -
David Woodhouse's avatar
David Woodhouse committed
464
					      REDZONE_ALIGN);
465
	return (unsigned long long *) (objp + cachep->size -
466
				       sizeof(unsigned long long));
Linus Torvalds's avatar
Linus Torvalds committed
467
468
}

469
static void **dbg_userword(struct kmem_cache *cachep, void *objp)
Linus Torvalds's avatar
Linus Torvalds committed
470
471
{
	BUG_ON(!(cachep->flags & SLAB_STORE_USER));
472
	return (void **)(objp + cachep->size - BYTES_PER_WORD);
Linus Torvalds's avatar
Linus Torvalds committed
473
474
475
476
}

#else

477
#define obj_offset(x)			0
478
479
#define dbg_redzone1(cachep, objp)	({BUG(); (unsigned long long *)NULL;})
#define dbg_redzone2(cachep, objp)	({BUG(); (unsigned long long *)NULL;})
Linus Torvalds's avatar
Linus Torvalds committed
480
481
482
483
484
#define dbg_userword(cachep, objp)	({BUG(); (void **)NULL;})

#endif

/*
485
486
 * Do not go above this order unless 0 objects fit into the slab or
 * overridden on the command line.
Linus Torvalds's avatar
Linus Torvalds committed
487
 */
488
489
490
#define	SLAB_MAX_ORDER_HI	1
#define	SLAB_MAX_ORDER_LO	0
static int slab_max_order = SLAB_MAX_ORDER_LO;
491
static bool slab_max_order_set __initdata;
Linus Torvalds's avatar
Linus Torvalds committed
492

493
494
static inline struct kmem_cache *virt_to_cache(const void *obj)
{
495
	struct page *page = virt_to_head_page(obj);
496
	return page->slab_cache;
497
498
499
500
}

static inline struct slab *virt_to_slab(const void *obj)
{
501
	struct page *page = virt_to_head_page(obj);
502
503
504

	VM_BUG_ON(!PageSlab(page));
	return page->slab_page;
505
506
}

507
508
509
static inline void *index_to_obj(struct kmem_cache *cache, struct slab *slab,
				 unsigned int idx)
{
510
	return slab->s_mem + cache->size * idx;
511
512
}

513
/*
514
515
516
 * We want to avoid an expensive divide : (offset / cache->size)
 *   Using the fact that size is a constant for a particular cache,
 *   we can replace (offset / cache->size) by
517
518
519
520
 *   reciprocal_divide(offset, cache->reciprocal_buffer_size)
 */
static inline unsigned int obj_to_index(const struct kmem_cache *cache,
					const struct slab *slab, void *obj)
521
{
522
523
	u32 offset = (obj - slab->s_mem);
	return reciprocal_divide(offset, cache->reciprocal_buffer_size);
524
525
}

Andrew Morton's avatar
Andrew Morton committed
526
527
528
/*
 * These are the default caches for kmalloc. Custom caches can have other sizes.
 */
Linus Torvalds's avatar
Linus Torvalds committed
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
struct cache_sizes malloc_sizes[] = {
#define CACHE(x) { .cs_size = (x) },
#include <linux/kmalloc_sizes.h>
	CACHE(ULONG_MAX)
#undef CACHE
};
EXPORT_SYMBOL(malloc_sizes);

/* Must match cache_sizes above. Out of line to keep cache footprint low. */
struct cache_names {
	char *name;
	char *name_dma;
};

static struct cache_names __initdata cache_names[] = {
#define CACHE(x) { .name = "size-" #x, .name_dma = "size-" #x "(DMA)" },
#include <linux/kmalloc_sizes.h>
546
	{NULL,}
Linus Torvalds's avatar
Linus Torvalds committed
547
548
549
550
#undef CACHE
};

static struct arraycache_init initarray_generic =
551
    { {0, BOOT_CPUCACHE_ENTRIES, 1, 0} };
Linus Torvalds's avatar
Linus Torvalds committed
552
553

/* internal cache of cache description objs */
554
static struct kmem_cache kmem_cache_boot = {
555
556
557
	.batchcount = 1,
	.limit = BOOT_CPUCACHE_ENTRIES,
	.shared = 1,
558
	.size = sizeof(struct kmem_cache),
559
	.name = "kmem_cache",
Linus Torvalds's avatar
Linus Torvalds committed
560
561
};

562
563
#define BAD_ALIEN_MAGIC 0x01020304ul

564
565
566
567
568
569
570
571
#ifdef CONFIG_LOCKDEP

/*
 * Slab sometimes uses the kmalloc slabs to store the slab headers
 * for other slabs "off slab".
 * The locking for this is tricky in that it nests within the locks
 * of all other slabs in a few places; to deal with this special
 * locking we put on-slab caches into a separate lock-class.
572
573
574
575
 *
 * We set lock class for alien array caches which are up during init.
 * The lock annotation will be lost if all cpus of a node goes down and
 * then comes back up during hotplug
576
 */
577
578
579
static struct lock_class_key on_slab_l3_key;
static struct lock_class_key on_slab_alc_key;

580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
static struct lock_class_key debugobj_l3_key;
static struct lock_class_key debugobj_alc_key;

static void slab_set_lock_classes(struct kmem_cache *cachep,
		struct lock_class_key *l3_key, struct lock_class_key *alc_key,
		int q)
{
	struct array_cache **alc;
	struct kmem_list3 *l3;
	int r;

	l3 = cachep->nodelists[q];
	if (!l3)
		return;

	lockdep_set_class(&l3->list_lock, l3_key);
	alc = l3->alien;
	/*
	 * FIXME: This check for BAD_ALIEN_MAGIC
	 * should go away when common slab code is taught to
	 * work even without alien caches.
	 * Currently, non NUMA code returns BAD_ALIEN_MAGIC
	 * for alloc_alien_cache,
	 */
	if (!alc || (unsigned long)alc == BAD_ALIEN_MAGIC)
		return;
	for_each_node(r) {
		if (alc[r])
			lockdep_set_class(&alc[r]->lock, alc_key);
	}
}

static void slab_set_debugobj_lock_classes_node(struct kmem_cache *cachep, int node)
{
	slab_set_lock_classes(cachep, &debugobj_l3_key, &debugobj_alc_key, node);
}

static void slab_set_debugobj_lock_classes(struct kmem_cache *cachep)
{
	int node;

	for_each_online_node(node)
		slab_set_debugobj_lock_classes_node(cachep, node);
}

625
static void init_node_lock_keys(int q)
626
{
627
628
	struct cache_sizes *s = malloc_sizes;

629
	if (slab_state < UP)
630
631
632
633
634
635
636
		return;

	for (s = malloc_sizes; s->cs_size != ULONG_MAX; s++) {
		struct kmem_list3 *l3;

		l3 = s->cs_cachep->nodelists[q];
		if (!l3 || OFF_SLAB(s->cs_cachep))
637
			continue;
638
639
640

		slab_set_lock_classes(s->cs_cachep, &on_slab_l3_key,
				&on_slab_alc_key, q);
641
642
	}
}
643
644
645
646
647
648
649
650

static inline void init_lock_keys(void)
{
	int node;

	for_each_node(node)
		init_node_lock_keys(node);
}
651
#else
652
653
654
655
static void init_node_lock_keys(int q)
{
}

656
static inline void init_lock_keys(void)
657
658
{
}
659
660
661
662
663
664
665
666

static void slab_set_debugobj_lock_classes_node(struct kmem_cache *cachep, int node)
{
}

static void slab_set_debugobj_lock_classes(struct kmem_cache *cachep)
{
}
667
668
#endif

669
static DEFINE_PER_CPU(struct delayed_work, slab_reap_work);
Linus Torvalds's avatar
Linus Torvalds committed
670

671
static inline struct array_cache *cpu_cache_get(struct kmem_cache *cachep)
Linus Torvalds's avatar
Linus Torvalds committed
672
673
674
675
{
	return cachep->array[smp_processor_id()];
}

Andrew Morton's avatar
Andrew Morton committed
676
677
static inline struct kmem_cache *__find_general_cachep(size_t size,
							gfp_t gfpflags)
Linus Torvalds's avatar
Linus Torvalds committed
678
679
680
681
682
{
	struct cache_sizes *csizep = malloc_sizes;

#if DEBUG
	/* This happens if someone tries to call
683
684
685
	 * kmem_cache_create(), or __kmalloc(), before
	 * the generic caches are initialized.
	 */
686
	BUG_ON(malloc_sizes[INDEX_AC].cs_cachep == NULL);
Linus Torvalds's avatar
Linus Torvalds committed
687
#endif
688
689
690
	if (!size)
		return ZERO_SIZE_PTR;

Linus Torvalds's avatar
Linus Torvalds committed
691
692
693
694
	while (size > csizep->cs_size)
		csizep++;

	/*
695
	 * Really subtle: The last entry with cs->cs_size==ULONG_MAX
Linus Torvalds's avatar
Linus Torvalds committed
696
697
698
	 * has cs_{dma,}cachep==NULL. Thus no special case
	 * for large kmalloc calls required.
	 */
699
#ifdef CONFIG_ZONE_DMA
Linus Torvalds's avatar
Linus Torvalds committed
700
701
	if (unlikely(gfpflags & GFP_DMA))
		return csizep->cs_dmacachep;
702
#endif
Linus Torvalds's avatar
Linus Torvalds committed
703
704
705
	return csizep->cs_cachep;
}

706
static struct kmem_cache *kmem_find_general_cachep(size_t size, gfp_t gfpflags)
707
708
709
710
{
	return __find_general_cachep(size, gfpflags);
}

711
static size_t slab_mgmt_size(size_t nr_objs, size_t align)
Linus Torvalds's avatar
Linus Torvalds committed
712
{
713
714
	return ALIGN(sizeof(struct slab)+nr_objs*sizeof(kmem_bufctl_t), align);
}
Linus Torvalds's avatar
Linus Torvalds committed
715

Andrew Morton's avatar
Andrew Morton committed
716
717
718
/*
 * Calculate the number of objects and left-over bytes for a given buffer size.
 */
719
720
721
722
723
724
725
static void cache_estimate(unsigned long gfporder, size_t buffer_size,
			   size_t align, int flags, size_t *left_over,
			   unsigned int *num)
{
	int nr_objs;
	size_t mgmt_size;
	size_t slab_size = PAGE_SIZE << gfporder;
Linus Torvalds's avatar
Linus Torvalds committed
726

727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
	/*
	 * The slab management structure can be either off the slab or
	 * on it. For the latter case, the memory allocated for a
	 * slab is used for:
	 *
	 * - The struct slab
	 * - One kmem_bufctl_t for each object
	 * - Padding to respect alignment of @align
	 * - @buffer_size bytes for each object
	 *
	 * If the slab management structure is off the slab, then the
	 * alignment will already be calculated into the size. Because
	 * the slabs are all pages aligned, the objects will be at the
	 * correct alignment when allocated.
	 */
	if (flags & CFLGS_OFF_SLAB) {
		mgmt_size = 0;
		nr_objs = slab_size / buffer_size;

		if (nr_objs > SLAB_LIMIT)
			nr_objs = SLAB_LIMIT;
	} else {
		/*
		 * Ignore padding for the initial guess. The padding
		 * is at most @align-1 bytes, and @buffer_size is at
		 * least @align. In the worst case, this result will
		 * be one greater than the number of objects that fit
		 * into the memory allocation when taking the padding
		 * into account.
		 */
		nr_objs = (slab_size - sizeof(struct slab)) /
			  (buffer_size + sizeof(kmem_bufctl_t));

		/*
		 * This calculated number will be either the right
		 * amount, or one greater than what we want.
		 */
		if (slab_mgmt_size(nr_objs, align) + nr_objs*buffer_size
		       > slab_size)
			nr_objs--;

		if (nr_objs > SLAB_LIMIT)
			nr_objs = SLAB_LIMIT;

		mgmt_size = slab_mgmt_size(nr_objs, align);
	}
	*num = nr_objs;
	*left_over = slab_size - nr_objs*buffer_size - mgmt_size;
Linus Torvalds's avatar
Linus Torvalds committed
775
776
}

777
#if DEBUG
778
#define slab_error(cachep, msg) __slab_error(__func__, cachep, msg)
Linus Torvalds's avatar
Linus Torvalds committed
779

Andrew Morton's avatar
Andrew Morton committed
780
781
static void __slab_error(const char *function, struct kmem_cache *cachep,
			char *msg)
Linus Torvalds's avatar
Linus Torvalds committed
782
783
{
	printk(KERN_ERR "slab error in %s(): cache `%s': %s\n",
784
	       function, cachep->name, msg);
Linus Torvalds's avatar
Linus Torvalds committed
785
	dump_stack();
786
	add_taint(TAINT_BAD_PAGE);
Linus Torvalds's avatar
Linus Torvalds committed
787
}
788
#endif
Linus Torvalds's avatar
Linus Torvalds committed
789

790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
/*
 * By default on NUMA we use alien caches to stage the freeing of
 * objects allocated from other nodes. This causes massive memory
 * inefficiencies when using fake NUMA setup to split memory into a
 * large number of small nodes, so it can be disabled on the command
 * line
  */

static int use_alien_caches __read_mostly = 1;
static int __init noaliencache_setup(char *s)
{
	use_alien_caches = 0;
	return 1;
}
__setup("noaliencache", noaliencache_setup);

806
807
808
809
810
811
812
813
814
815
816
static int __init slab_max_order_setup(char *str)
{
	get_option(&str, &slab_max_order);
	slab_max_order = slab_max_order < 0 ? 0 :
				min(slab_max_order, MAX_ORDER - 1);
	slab_max_order_set = true;

	return 1;
}
__setup("slab_max_order=", slab_max_order_setup);

817
818
819
820
821
822
823
#ifdef CONFIG_NUMA
/*
 * Special reaping functions for NUMA systems called from cache_reap().
 * These take care of doing round robin flushing of alien caches (containing
 * objects freed on different nodes from which they were allocated) and the
 * flushing of remote pcps by calling drain_node_pages.
 */
824
static DEFINE_PER_CPU(unsigned long, slab_reap_node);
825
826
827
828
829

static void init_reap_node(int cpu)
{
	int node;

830
	node = next_node(cpu_to_mem(cpu), node_online_map);
831
	if (node == MAX_NUMNODES)
832
		node = first_node(node_online_map);
833

834
	per_cpu(slab_reap_node, cpu) = node;
835
836
837
838
}

static void next_reap_node(void)
{
839
	int node = __this_cpu_read(slab_reap_node);
840
841
842
843

	node = next_node(node, node_online_map);
	if (unlikely(node >= MAX_NUMNODES))
		node = first_node(node_online_map);
844
	__this_cpu_write(slab_reap_node, node);
845
846
847
848
849
850
851
}

#else
#define init_reap_node(cpu) do { } while (0)
#define next_reap_node(void) do { } while (0)
#endif

Linus Torvalds's avatar
Linus Torvalds committed
852
853
854
855
856
857
858
/*
 * Initiate the reap timer running on the target CPU.  We run at around 1 to 2Hz
 * via the workqueue/eventd.
 * Add the CPU number into the expiration time to minimize the possibility of
 * the CPUs getting into lockstep and contending for the global cache chain
 * lock.
 */
859
static void __cpuinit start_cpu_timer(int cpu)
Linus Torvalds's avatar
Linus Torvalds committed
860
{
861
	struct delayed_work *reap_work = &per_cpu(slab_reap_work, cpu);
Linus Torvalds's avatar
Linus Torvalds committed
862
863
864
865
866
867

	/*
	 * When this gets called from do_initcalls via cpucache_init(),
	 * init_workqueues() has already run, so keventd will be setup
	 * at that time.
	 */
868
	if (keventd_up() && reap_work->work.func == NULL) {
869
		init_reap_node(cpu);
870
		INIT_DEFERRABLE_WORK(reap_work, cache_reap);
871
872
		schedule_delayed_work_on(cpu, reap_work,
					__round_jiffies_relative(HZ, cpu));
Linus Torvalds's avatar
Linus Torvalds committed
873
874
875
	}
}

876
static struct array_cache *alloc_arraycache(int node, int entries,
877
					    int batchcount, gfp_t gfp)
Linus Torvalds's avatar
Linus Torvalds committed
878
{
879
	int memsize = sizeof(void *) * entries + sizeof(struct array_cache);
Linus Torvalds's avatar
Linus Torvalds committed
880
881
	struct array_cache *nc = NULL;

882
	nc = kmalloc_node(memsize, gfp, node);
883
884
	/*
	 * The array_cache structures contain pointers to free object.
Lucas De Marchi's avatar
Lucas De Marchi committed
885
	 * However, when such objects are allocated or transferred to another
886
887
888
889
890
	 * cache the pointers are not cleared and they could be counted as
	 * valid references during a kmemleak scan. Therefore, kmemleak must
	 * not scan such objects.
	 */
	kmemleak_no_scan(nc);
Linus Torvalds's avatar
Linus Torvalds committed
891
892
893
894
895
	if (nc) {
		nc->avail = 0;
		nc->limit = entries;
		nc->batchcount = batchcount;
		nc->touched = 0;
896
		spin_lock_init(&nc->lock);
Linus Torvalds's avatar
Linus Torvalds committed
897
898
899
900
	}
	return nc;
}

901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
static inline bool is_slab_pfmemalloc(struct slab *slabp)
{
	struct page *page = virt_to_page(slabp->s_mem);

	return PageSlabPfmemalloc(page);
}

/* Clears pfmemalloc_active if no slabs have pfmalloc set */
static void recheck_pfmemalloc_active(struct kmem_cache *cachep,
						struct array_cache *ac)
{
	struct kmem_list3 *l3 = cachep->nodelists[numa_mem_id()];
	struct slab *slabp;
	unsigned long flags;

	if (!pfmemalloc_active)
		return;

	spin_lock_irqsave(&l3->list_lock, flags);
	list_for_each_entry(slabp, &l3->slabs_full, list)
		if (is_slab_pfmemalloc(slabp))
			goto out;

	list_for_each_entry(slabp, &l3->slabs_partial, list)
		if (is_slab_pfmemalloc(slabp))
			goto out;

	list_for_each_entry(slabp, &l3->slabs_free, list)
		if (is_slab_pfmemalloc(slabp))
			goto out;

	pfmemalloc_active = false;
out:
	spin_unlock_irqrestore(&l3->list_lock, flags);
}

937
static void *__ac_get_obj(struct kmem_cache *cachep, struct array_cache *ac,
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
						gfp_t flags, bool force_refill)
{
	int i;
	void *objp = ac->entry[--ac->avail];

	/* Ensure the caller is allowed to use objects from PFMEMALLOC slab */
	if (unlikely(is_obj_pfmemalloc(objp))) {
		struct kmem_list3 *l3;

		if (gfp_pfmemalloc_allowed(flags)) {
			clear_obj_pfmemalloc(&objp);
			return objp;
		}

		/* The caller cannot use PFMEMALLOC objects, find another one */
953
		for (i = 0; i < ac->avail; i++) {
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
			/* If a !PFMEMALLOC object is found, swap them */
			if (!is_obj_pfmemalloc(ac->entry[i])) {
				objp = ac->entry[i];
				ac->entry[i] = ac->entry[ac->avail];
				ac->entry[ac->avail] = objp;
				return objp;
			}
		}

		/*
		 * If there are empty slabs on the slabs_free list and we are
		 * being forced to refill the cache, mark this one !pfmemalloc.
		 */
		l3 = cachep->nodelists[numa_mem_id()];
		if (!list_empty(&l3->slabs_free) && force_refill) {
			struct slab *slabp = virt_to_slab(objp);
970
			ClearPageSlabPfmemalloc(virt_to_head_page(slabp->s_mem));
971
972
973
974
975
976
977
978
979
980
981
982
983
			clear_obj_pfmemalloc(&objp);
			recheck_pfmemalloc_active(cachep, ac);
			return objp;
		}

		/* No !PFMEMALLOC objects available */
		ac->avail++;
		objp = NULL;
	}

	return objp;
}

984
985
986
987
988
989
990
991
992
993
994
995
996
997
static inline void *ac_get_obj(struct kmem_cache *cachep,
			struct array_cache *ac, gfp_t flags, bool force_refill)
{
	void *objp;

	if (unlikely(sk_memalloc_socks()))
		objp = __ac_get_obj(cachep, ac, flags, force_refill);
	else
		objp = ac->entry[--ac->avail];

	return objp;
}

static void *__ac_put_obj(struct kmem_cache *cachep, struct array_cache *ac,
998
999
1000
								void *objp)
{
	if (unlikely(pfmemalloc_active)) {