slab.c 115.0 KB
Newer Older
L
Linus Torvalds 已提交
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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52
/*
 * 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
 * slabs and you must pass objects with the same intializations to
 * 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.
 *
A
Andrew Morton 已提交
53
 * The c_cpuarray may not be read with enabled local interrupts -
L
Linus Torvalds 已提交
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
L
Linus Torvalds 已提交
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
I
Ingo Molnar 已提交
71
 *	The global cache-chain is protected by the mutex 'cache_chain_mutex'.
L
Linus Torvalds 已提交
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.
L
Linus Torvalds 已提交
87 88 89 90
 */

#include	<linux/slab.h>
#include	<linux/mm.h>
91
#include	<linux/poison.h>
L
Linus Torvalds 已提交
92 93 94 95 96
#include	<linux/swap.h>
#include	<linux/cache.h>
#include	<linux/interrupt.h>
#include	<linux/init.h>
#include	<linux/compiler.h>
97
#include	<linux/cpuset.h>
L
Linus Torvalds 已提交
98 99 100 101 102 103 104
#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>
105
#include	<linux/string.h>
106
#include	<linux/uaccess.h>
107
#include	<linux/nodemask.h>
108
#include	<linux/mempolicy.h>
I
Ingo Molnar 已提交
109
#include	<linux/mutex.h>
110
#include	<linux/fault-inject.h>
I
Ingo Molnar 已提交
111
#include	<linux/rtmutex.h>
112
#include	<linux/reciprocal_div.h>
L
Linus Torvalds 已提交
113 114 115 116 117 118

#include	<asm/cacheflush.h>
#include	<asm/tlbflush.h>
#include	<asm/page.h>

/*
119
 * DEBUG	- 1 for kmem_cache_create() to honour; SLAB_RED_ZONE & SLAB_POISON.
L
Linus Torvalds 已提交
120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139
 *		  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 *)
D
David Woodhouse 已提交
140
#define	REDZONE_ALIGN		max(BYTES_PER_WORD, __alignof__(unsigned long long))
L
Linus Torvalds 已提交
141 142 143 144 145 146 147 148 149 150 151

#ifndef cache_line_size
#define cache_line_size()	L1_CACHE_BYTES
#endif

#ifndef ARCH_KMALLOC_MINALIGN
/*
 * Enforce a minimum alignment for the kmalloc caches.
 * Usually, the kmalloc caches are cache_line_size() aligned, except when
 * DEBUG and FORCED_DEBUG are enabled, then they are BYTES_PER_WORD aligned.
 * Some archs want to perform DMA into kmalloc caches and need a guaranteed
152 153 154
 * alignment larger than the alignment of a 64-bit integer.
 * ARCH_KMALLOC_MINALIGN allows that.
 * Note that increasing this value may disable some debug features.
L
Linus Torvalds 已提交
155
 */
156
#define ARCH_KMALLOC_MINALIGN __alignof__(unsigned long long)
L
Linus Torvalds 已提交
157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175
#endif

#ifndef ARCH_SLAB_MINALIGN
/*
 * Enforce a minimum alignment for all caches.
 * Intended for archs that get misalignment faults even for BYTES_PER_WORD
 * aligned buffers. Includes ARCH_KMALLOC_MINALIGN.
 * If possible: Do not enable this flag for CONFIG_DEBUG_SLAB, it disables
 * some debug features.
 */
#define ARCH_SLAB_MINALIGN 0
#endif

#ifndef ARCH_KMALLOC_FLAGS
#define ARCH_KMALLOC_FLAGS SLAB_HWCACHE_ALIGN
#endif

/* Legal flag mask for kmem_cache_create(). */
#if DEBUG
176
# define CREATE_MASK	(SLAB_RED_ZONE | \
L
Linus Torvalds 已提交
177
			 SLAB_POISON | SLAB_HWCACHE_ALIGN | \
178
			 SLAB_CACHE_DMA | \
179
			 SLAB_STORE_USER | \
L
Linus Torvalds 已提交
180
			 SLAB_RECLAIM_ACCOUNT | SLAB_PANIC | \
181
			 SLAB_DESTROY_BY_RCU | SLAB_MEM_SPREAD)
L
Linus Torvalds 已提交
182
#else
183
# define CREATE_MASK	(SLAB_HWCACHE_ALIGN | \
184
			 SLAB_CACHE_DMA | \
L
Linus Torvalds 已提交
185
			 SLAB_RECLAIM_ACCOUNT | SLAB_PANIC | \
186
			 SLAB_DESTROY_BY_RCU | SLAB_MEM_SPREAD)
L
Linus Torvalds 已提交
187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207
#endif

/*
 * 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.
 */

208
typedef unsigned int kmem_bufctl_t;
L
Linus Torvalds 已提交
209 210
#define BUFCTL_END	(((kmem_bufctl_t)(~0U))-0)
#define BUFCTL_FREE	(((kmem_bufctl_t)(~0U))-1)
211 212
#define	BUFCTL_ACTIVE	(((kmem_bufctl_t)(~0U))-2)
#define	SLAB_LIMIT	(((kmem_bufctl_t)(~0U))-3)
L
Linus Torvalds 已提交
213 214 215 216 217 218 219 220 221

/*
 * 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 {
P
Pekka Enberg 已提交
222 223 224 225 226 227
	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;
L
Linus Torvalds 已提交
228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246
};

/*
 * 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.
 *
 * We assume struct slab_rcu can overlay struct slab when destroying.
 */
struct slab_rcu {
P
Pekka Enberg 已提交
247
	struct rcu_head head;
248
	struct kmem_cache *cachep;
P
Pekka Enberg 已提交
249
	void *addr;
L
Linus Torvalds 已提交
250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268
};

/*
 * 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;
269
	spinlock_t lock;
A
Andrew Morton 已提交
270 271 272 273 274 275
	void *entry[0];	/*
			 * Must have this definition in here for the proper
			 * alignment of array_cache. Also simplifies accessing
			 * the entries.
			 * [0] is for gcc 2.95. It should really be [].
			 */
L
Linus Torvalds 已提交
276 277
};

A
Andrew Morton 已提交
278 279 280
/*
 * bootstrap: The caches do not work without cpuarrays anymore, but the
 * cpuarrays are allocated from the generic caches...
L
Linus Torvalds 已提交
281 282 283 284
 */
#define BOOT_CPUCACHE_ENTRIES	1
struct arraycache_init {
	struct array_cache cache;
P
Pekka Enberg 已提交
285
	void *entries[BOOT_CPUCACHE_ENTRIES];
L
Linus Torvalds 已提交
286 287 288
};

/*
289
 * The slab lists for all objects.
L
Linus Torvalds 已提交
290 291
 */
struct kmem_list3 {
P
Pekka Enberg 已提交
292 293 294 295 296
	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;
297
	unsigned int colour_next;	/* Per-node cache coloring */
P
Pekka Enberg 已提交
298 299 300
	spinlock_t list_lock;
	struct array_cache *shared;	/* shared per node */
	struct array_cache **alien;	/* on other nodes */
301 302
	unsigned long next_reap;	/* updated without locking */
	int free_touched;		/* updated without locking */
L
Linus Torvalds 已提交
303 304
};

305 306 307 308 309 310 311 312 313
/*
 * Need this for bootstrapping a per node allocator.
 */
#define NUM_INIT_LISTS (2 * MAX_NUMNODES + 1)
struct kmem_list3 __initdata initkmem_list3[NUM_INIT_LISTS];
#define	CACHE_CACHE 0
#define	SIZE_AC 1
#define	SIZE_L3 (1 + MAX_NUMNODES)

314 315 316 317
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);
318
static int enable_cpucache(struct kmem_cache *cachep);
319
static void cache_reap(struct work_struct *unused);
320

321
/*
A
Andrew Morton 已提交
322 323
 * 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.
324
 */
325
static __always_inline int index_of(const size_t size)
326
{
327 328
	extern void __bad_size(void);

329 330 331 332 333 334 335 336 337 338
	if (__builtin_constant_p(size)) {
		int i = 0;

#define CACHE(x) \
	if (size <=x) \
		return i; \
	else \
		i++;
#include "linux/kmalloc_sizes.h"
#undef CACHE
339
		__bad_size();
340
	} else
341
		__bad_size();
342 343 344
	return 0;
}

345 346
static int slab_early_init = 1;

347 348
#define INDEX_AC index_of(sizeof(struct arraycache_init))
#define INDEX_L3 index_of(sizeof(struct kmem_list3))
L
Linus Torvalds 已提交
349

P
Pekka Enberg 已提交
350
static void kmem_list3_init(struct kmem_list3 *parent)
351 352 353 354 355 356
{
	INIT_LIST_HEAD(&parent->slabs_full);
	INIT_LIST_HEAD(&parent->slabs_partial);
	INIT_LIST_HEAD(&parent->slabs_free);
	parent->shared = NULL;
	parent->alien = NULL;
357
	parent->colour_next = 0;
358 359 360 361 362
	spin_lock_init(&parent->list_lock);
	parent->free_objects = 0;
	parent->free_touched = 0;
}

A
Andrew Morton 已提交
363 364 365 366
#define MAKE_LIST(cachep, listp, slab, nodeid)				\
	do {								\
		INIT_LIST_HEAD(listp);					\
		list_splice(&(cachep->nodelists[nodeid]->slab), listp);	\
367 368
	} while (0)

A
Andrew Morton 已提交
369 370
#define	MAKE_ALL_LISTS(cachep, ptr, nodeid)				\
	do {								\
371 372 373 374
	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)
L
Linus Torvalds 已提交
375 376

/*
377
 * struct kmem_cache
L
Linus Torvalds 已提交
378 379 380
 *
 * manages a cache.
 */
P
Pekka Enberg 已提交
381

382
struct kmem_cache {
L
Linus Torvalds 已提交
383
/* 1) per-cpu data, touched during every alloc/free */
P
Pekka Enberg 已提交
384
	struct array_cache *array[NR_CPUS];
385
/* 2) Cache tunables. Protected by cache_chain_mutex */
P
Pekka Enberg 已提交
386 387 388
	unsigned int batchcount;
	unsigned int limit;
	unsigned int shared;
389

390
	unsigned int buffer_size;
391
	u32 reciprocal_buffer_size;
392 393
/* 3) touched by every alloc & free from the backend */

A
Andrew Morton 已提交
394 395
	unsigned int flags;		/* constant flags */
	unsigned int num;		/* # of objs per slab */
L
Linus Torvalds 已提交
396

397
/* 4) cache_grow/shrink */
L
Linus Torvalds 已提交
398
	/* order of pgs per slab (2^n) */
P
Pekka Enberg 已提交
399
	unsigned int gfporder;
L
Linus Torvalds 已提交
400 401

	/* force GFP flags, e.g. GFP_DMA */
P
Pekka Enberg 已提交
402
	gfp_t gfpflags;
L
Linus Torvalds 已提交
403

A
Andrew Morton 已提交
404
	size_t colour;			/* cache colouring range */
P
Pekka Enberg 已提交
405
	unsigned int colour_off;	/* colour offset */
406
	struct kmem_cache *slabp_cache;
P
Pekka Enberg 已提交
407
	unsigned int slab_size;
A
Andrew Morton 已提交
408
	unsigned int dflags;		/* dynamic flags */
L
Linus Torvalds 已提交
409 410

	/* constructor func */
411
	void (*ctor) (void *, struct kmem_cache *, unsigned long);
L
Linus Torvalds 已提交
412

413
/* 5) cache creation/removal */
P
Pekka Enberg 已提交
414 415
	const char *name;
	struct list_head next;
L
Linus Torvalds 已提交
416

417
/* 6) statistics */
L
Linus Torvalds 已提交
418
#if STATS
P
Pekka Enberg 已提交
419 420 421 422 423 424 425 426 427
	unsigned long num_active;
	unsigned long num_allocations;
	unsigned long high_mark;
	unsigned long grown;
	unsigned long reaped;
	unsigned long errors;
	unsigned long max_freeable;
	unsigned long node_allocs;
	unsigned long node_frees;
428
	unsigned long node_overflow;
P
Pekka Enberg 已提交
429 430 431 432
	atomic_t allochit;
	atomic_t allocmiss;
	atomic_t freehit;
	atomic_t freemiss;
L
Linus Torvalds 已提交
433 434
#endif
#if DEBUG
435 436 437 438 439 440 441 442
	/*
	 * If debugging is enabled, then the allocator can add additional
	 * fields and/or padding to every object. buffer_size contains the total
	 * object size including these internal fields, the following two
	 * variables contain the offset to the user object and its size.
	 */
	int obj_offset;
	int obj_size;
L
Linus Torvalds 已提交
443
#endif
E
Eric Dumazet 已提交
444 445 446 447 448 449 450 451 452 453 454
	/*
	 * We put nodelists[] at the end of kmem_cache, because we want to size
	 * this array to nr_node_ids slots instead of MAX_NUMNODES
	 * (see kmem_cache_init())
	 * We still use [MAX_NUMNODES] and not [1] or [0] because cache_cache
	 * is statically defined, so we reserve the max number of nodes.
	 */
	struct kmem_list3 *nodelists[MAX_NUMNODES];
	/*
	 * Do not add fields after nodelists[]
	 */
L
Linus Torvalds 已提交
455 456 457 458 459 460
};

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

#define BATCHREFILL_LIMIT	16
A
Andrew Morton 已提交
461 462 463
/*
 * Optimization question: fewer reaps means less probability for unnessary
 * cpucache drain/refill cycles.
L
Linus Torvalds 已提交
464
 *
A
Adrian Bunk 已提交
465
 * OTOH the cpuarrays can contain lots of objects,
L
Linus Torvalds 已提交
466 467 468 469 470 471 472 473 474 475
 * 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++)
476
#define	STATS_ADD_REAPED(x,y)	((x)->reaped += (y))
A
Andrew Morton 已提交
477 478 479 480 481
#define	STATS_SET_HIGH(x)						\
	do {								\
		if ((x)->num_active > (x)->high_mark)			\
			(x)->high_mark = (x)->num_active;		\
	} while (0)
L
Linus Torvalds 已提交
482 483
#define	STATS_INC_ERR(x)	((x)->errors++)
#define	STATS_INC_NODEALLOCS(x)	((x)->node_allocs++)
484
#define	STATS_INC_NODEFREES(x)	((x)->node_frees++)
485
#define STATS_INC_ACOVERFLOW(x)   ((x)->node_overflow++)
A
Andrew Morton 已提交
486 487 488 489 490
#define	STATS_SET_FREEABLE(x, i)					\
	do {								\
		if ((x)->max_freeable < i)				\
			(x)->max_freeable = i;				\
	} while (0)
L
Linus Torvalds 已提交
491 492 493 494 495 496 497 498 499
#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)
500
#define	STATS_ADD_REAPED(x,y)	do { } while (0)
L
Linus Torvalds 已提交
501 502 503
#define	STATS_SET_HIGH(x)	do { } while (0)
#define	STATS_INC_ERR(x)	do { } while (0)
#define	STATS_INC_NODEALLOCS(x)	do { } while (0)
504
#define	STATS_INC_NODEFREES(x)	do { } while (0)
505
#define STATS_INC_ACOVERFLOW(x)   do { } while (0)
A
Andrew Morton 已提交
506
#define	STATS_SET_FREEABLE(x, i) do { } while (0)
L
Linus Torvalds 已提交
507 508 509 510 511 512 513 514
#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

A
Andrew Morton 已提交
515 516
/*
 * memory layout of objects:
L
Linus Torvalds 已提交
517
 * 0		: objp
518
 * 0 .. cachep->obj_offset - BYTES_PER_WORD - 1: padding. This ensures that
L
Linus Torvalds 已提交
519 520
 * 		the end of an object is aligned with the end of the real
 * 		allocation. Catches writes behind the end of the allocation.
521
 * cachep->obj_offset - BYTES_PER_WORD .. cachep->obj_offset - 1:
L
Linus Torvalds 已提交
522
 * 		redzone word.
523 524
 * cachep->obj_offset: The real object.
 * cachep->buffer_size - 2* BYTES_PER_WORD: redzone word [BYTES_PER_WORD long]
A
Andrew Morton 已提交
525 526
 * cachep->buffer_size - 1* BYTES_PER_WORD: last caller address
 *					[BYTES_PER_WORD long]
L
Linus Torvalds 已提交
527
 */
528
static int obj_offset(struct kmem_cache *cachep)
L
Linus Torvalds 已提交
529
{
530
	return cachep->obj_offset;
L
Linus Torvalds 已提交
531 532
}

533
static int obj_size(struct kmem_cache *cachep)
L
Linus Torvalds 已提交
534
{
535
	return cachep->obj_size;
L
Linus Torvalds 已提交
536 537
}

538
static unsigned long long *dbg_redzone1(struct kmem_cache *cachep, void *objp)
L
Linus Torvalds 已提交
539 540
{
	BUG_ON(!(cachep->flags & SLAB_RED_ZONE));
541 542
	return (unsigned long long*) (objp + obj_offset(cachep) -
				      sizeof(unsigned long long));
L
Linus Torvalds 已提交
543 544
}

545
static unsigned long long *dbg_redzone2(struct kmem_cache *cachep, void *objp)
L
Linus Torvalds 已提交
546 547 548
{
	BUG_ON(!(cachep->flags & SLAB_RED_ZONE));
	if (cachep->flags & SLAB_STORE_USER)
549 550
		return (unsigned long long *)(objp + cachep->buffer_size -
					      sizeof(unsigned long long) -
D
David Woodhouse 已提交
551
					      REDZONE_ALIGN);
552 553
	return (unsigned long long *) (objp + cachep->buffer_size -
				       sizeof(unsigned long long));
L
Linus Torvalds 已提交
554 555
}

556
static void **dbg_userword(struct kmem_cache *cachep, void *objp)
L
Linus Torvalds 已提交
557 558
{
	BUG_ON(!(cachep->flags & SLAB_STORE_USER));
559
	return (void **)(objp + cachep->buffer_size - BYTES_PER_WORD);
L
Linus Torvalds 已提交
560 561 562 563
}

#else

564 565
#define obj_offset(x)			0
#define obj_size(cachep)		(cachep->buffer_size)
566 567
#define dbg_redzone1(cachep, objp)	({BUG(); (unsigned long long *)NULL;})
#define dbg_redzone2(cachep, objp)	({BUG(); (unsigned long long *)NULL;})
L
Linus Torvalds 已提交
568 569 570 571 572 573 574 575 576 577 578
#define dbg_userword(cachep, objp)	({BUG(); (void **)NULL;})

#endif

/*
 * Do not go above this order unless 0 objects fit into the slab.
 */
#define	BREAK_GFP_ORDER_HI	1
#define	BREAK_GFP_ORDER_LO	0
static int slab_break_gfp_order = BREAK_GFP_ORDER_LO;

A
Andrew Morton 已提交
579 580 581 582
/*
 * Functions for storing/retrieving the cachep and or slab from the page
 * allocator.  These are used to find the slab an obj belongs to.  With kfree(),
 * these are used to find the cache which an obj belongs to.
L
Linus Torvalds 已提交
583
 */
584 585 586 587 588 589 590
static inline void page_set_cache(struct page *page, struct kmem_cache *cache)
{
	page->lru.next = (struct list_head *)cache;
}

static inline struct kmem_cache *page_get_cache(struct page *page)
{
591
	page = compound_head(page);
592
	BUG_ON(!PageSlab(page));
593 594 595 596 597 598 599 600 601 602
	return (struct kmem_cache *)page->lru.next;
}

static inline void page_set_slab(struct page *page, struct slab *slab)
{
	page->lru.prev = (struct list_head *)slab;
}

static inline struct slab *page_get_slab(struct page *page)
{
603
	BUG_ON(!PageSlab(page));
604 605
	return (struct slab *)page->lru.prev;
}
L
Linus Torvalds 已提交
606

607 608
static inline struct kmem_cache *virt_to_cache(const void *obj)
{
609
	struct page *page = virt_to_head_page(obj);
610 611 612 613 614
	return page_get_cache(page);
}

static inline struct slab *virt_to_slab(const void *obj)
{
615
	struct page *page = virt_to_head_page(obj);
616 617 618
	return page_get_slab(page);
}

619 620 621 622 623 624
static inline void *index_to_obj(struct kmem_cache *cache, struct slab *slab,
				 unsigned int idx)
{
	return slab->s_mem + cache->buffer_size * idx;
}

625 626 627 628 629 630 631 632
/*
 * We want to avoid an expensive divide : (offset / cache->buffer_size)
 *   Using the fact that buffer_size is a constant for a particular cache,
 *   we can replace (offset / cache->buffer_size) by
 *   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)
633
{
634 635
	u32 offset = (obj - slab->s_mem);
	return reciprocal_divide(offset, cache->reciprocal_buffer_size);
636 637
}

A
Andrew Morton 已提交
638 639 640
/*
 * These are the default caches for kmalloc. Custom caches can have other sizes.
 */
L
Linus Torvalds 已提交
641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657
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>
P
Pekka Enberg 已提交
658
	{NULL,}
L
Linus Torvalds 已提交
659 660 661 662
#undef CACHE
};

static struct arraycache_init initarray_cache __initdata =
P
Pekka Enberg 已提交
663
    { {0, BOOT_CPUCACHE_ENTRIES, 1, 0} };
L
Linus Torvalds 已提交
664
static struct arraycache_init initarray_generic =
P
Pekka Enberg 已提交
665
    { {0, BOOT_CPUCACHE_ENTRIES, 1, 0} };
L
Linus Torvalds 已提交
666 667

/* internal cache of cache description objs */
668
static struct kmem_cache cache_cache = {
P
Pekka Enberg 已提交
669 670 671
	.batchcount = 1,
	.limit = BOOT_CPUCACHE_ENTRIES,
	.shared = 1,
672
	.buffer_size = sizeof(struct kmem_cache),
P
Pekka Enberg 已提交
673
	.name = "kmem_cache",
L
Linus Torvalds 已提交
674 675
};

676 677
#define BAD_ALIEN_MAGIC 0x01020304ul

678 679 680 681 682 683 684 685
#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.
686 687 688 689
 *
 * 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
690
 */
691 692 693 694
static struct lock_class_key on_slab_l3_key;
static struct lock_class_key on_slab_alc_key;

static inline void init_lock_keys(void)
695 696 697

{
	int q;
698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724
	struct cache_sizes *s = malloc_sizes;

	while (s->cs_size != ULONG_MAX) {
		for_each_node(q) {
			struct array_cache **alc;
			int r;
			struct kmem_list3 *l3 = s->cs_cachep->nodelists[q];
			if (!l3 || OFF_SLAB(s->cs_cachep))
				continue;
			lockdep_set_class(&l3->list_lock, &on_slab_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)
				continue;
			for_each_node(r) {
				if (alc[r])
					lockdep_set_class(&alc[r]->lock,
					     &on_slab_alc_key);
			}
		}
		s++;
725 726 727
	}
}
#else
728
static inline void init_lock_keys(void)
729 730 731 732
{
}
#endif

733 734 735 736
/*
 * 1. Guard access to the cache-chain.
 * 2. Protect sanity of cpu_online_map against cpu hotplug events
 */
I
Ingo Molnar 已提交
737
static DEFINE_MUTEX(cache_chain_mutex);
L
Linus Torvalds 已提交
738 739 740 741 742 743 744 745
static struct list_head cache_chain;

/*
 * chicken and egg problem: delay the per-cpu array allocation
 * until the general caches are up.
 */
static enum {
	NONE,
746 747
	PARTIAL_AC,
	PARTIAL_L3,
L
Linus Torvalds 已提交
748 749 750
	FULL
} g_cpucache_up;

751 752 753 754 755 756 757 758
/*
 * used by boot code to determine if it can use slab based allocator
 */
int slab_is_available(void)
{
	return g_cpucache_up == FULL;
}

759
static DEFINE_PER_CPU(struct delayed_work, reap_work);
L
Linus Torvalds 已提交
760

761
static inline struct array_cache *cpu_cache_get(struct kmem_cache *cachep)
L
Linus Torvalds 已提交
762 763 764 765
{
	return cachep->array[smp_processor_id()];
}

A
Andrew Morton 已提交
766 767
static inline struct kmem_cache *__find_general_cachep(size_t size,
							gfp_t gfpflags)
L
Linus Torvalds 已提交
768 769 770 771 772
{
	struct cache_sizes *csizep = malloc_sizes;

#if DEBUG
	/* This happens if someone tries to call
P
Pekka Enberg 已提交
773 774 775
	 * kmem_cache_create(), or __kmalloc(), before
	 * the generic caches are initialized.
	 */
776
	BUG_ON(malloc_sizes[INDEX_AC].cs_cachep == NULL);
L
Linus Torvalds 已提交
777
#endif
778 779 780
	if (!size)
		return ZERO_SIZE_PTR;

L
Linus Torvalds 已提交
781 782 783 784
	while (size > csizep->cs_size)
		csizep++;

	/*
785
	 * Really subtle: The last entry with cs->cs_size==ULONG_MAX
L
Linus Torvalds 已提交
786 787 788
	 * has cs_{dma,}cachep==NULL. Thus no special case
	 * for large kmalloc calls required.
	 */
789
#ifdef CONFIG_ZONE_DMA
L
Linus Torvalds 已提交
790 791
	if (unlikely(gfpflags & GFP_DMA))
		return csizep->cs_dmacachep;
792
#endif
L
Linus Torvalds 已提交
793 794 795
	return csizep->cs_cachep;
}

A
Adrian Bunk 已提交
796
static struct kmem_cache *kmem_find_general_cachep(size_t size, gfp_t gfpflags)
797 798 799 800
{
	return __find_general_cachep(size, gfpflags);
}

801
static size_t slab_mgmt_size(size_t nr_objs, size_t align)
L
Linus Torvalds 已提交
802
{
803 804
	return ALIGN(sizeof(struct slab)+nr_objs*sizeof(kmem_bufctl_t), align);
}
L
Linus Torvalds 已提交
805

A
Andrew Morton 已提交
806 807 808
/*
 * Calculate the number of objects and left-over bytes for a given buffer size.
 */
809 810 811 812 813 814 815
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;
L
Linus Torvalds 已提交
816

817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864
	/*
	 * 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;
L
Linus Torvalds 已提交
865 866 867 868
}

#define slab_error(cachep, msg) __slab_error(__FUNCTION__, cachep, msg)

A
Andrew Morton 已提交
869 870
static void __slab_error(const char *function, struct kmem_cache *cachep,
			char *msg)
L
Linus Torvalds 已提交
871 872
{
	printk(KERN_ERR "slab error in %s(): cache `%s': %s\n",
P
Pekka Enberg 已提交
873
	       function, cachep->name, msg);
L
Linus Torvalds 已提交
874 875 876
	dump_stack();
}

877 878 879 880 881 882 883 884 885
/*
 * 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;
886
static int numa_platform __read_mostly = 1;
887 888 889 890 891 892 893
static int __init noaliencache_setup(char *s)
{
	use_alien_caches = 0;
	return 1;
}
__setup("noaliencache", noaliencache_setup);

894 895 896 897 898 899 900 901 902 903 904 905 906 907 908
#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.
 */
static DEFINE_PER_CPU(unsigned long, reap_node);

static void init_reap_node(int cpu)
{
	int node;

	node = next_node(cpu_to_node(cpu), node_online_map);
	if (node == MAX_NUMNODES)
909
		node = first_node(node_online_map);
910

911
	per_cpu(reap_node, cpu) = node;
912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928
}

static void next_reap_node(void)
{
	int node = __get_cpu_var(reap_node);

	node = next_node(node, node_online_map);
	if (unlikely(node >= MAX_NUMNODES))
		node = first_node(node_online_map);
	__get_cpu_var(reap_node) = node;
}

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

L
Linus Torvalds 已提交
929 930 931 932 933 934 935
/*
 * 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.
 */
936
static void __cpuinit start_cpu_timer(int cpu)
L
Linus Torvalds 已提交
937
{
938
	struct delayed_work *reap_work = &per_cpu(reap_work, cpu);
L
Linus Torvalds 已提交
939 940 941 942 943 944

	/*
	 * When this gets called from do_initcalls via cpucache_init(),
	 * init_workqueues() has already run, so keventd will be setup
	 * at that time.
	 */
945
	if (keventd_up() && reap_work->work.func == NULL) {
946
		init_reap_node(cpu);
947
		INIT_DELAYED_WORK(reap_work, cache_reap);
948 949
		schedule_delayed_work_on(cpu, reap_work,
					__round_jiffies_relative(HZ, cpu));
L
Linus Torvalds 已提交
950 951 952
	}
}

953
static struct array_cache *alloc_arraycache(int node, int entries,
P
Pekka Enberg 已提交
954
					    int batchcount)
L
Linus Torvalds 已提交
955
{
P
Pekka Enberg 已提交
956
	int memsize = sizeof(void *) * entries + sizeof(struct array_cache);
L
Linus Torvalds 已提交
957 958
	struct array_cache *nc = NULL;

959
	nc = kmalloc_node(memsize, GFP_KERNEL, node);
L
Linus Torvalds 已提交
960 961 962 963 964
	if (nc) {
		nc->avail = 0;
		nc->limit = entries;
		nc->batchcount = batchcount;
		nc->touched = 0;
965
		spin_lock_init(&nc->lock);
L
Linus Torvalds 已提交
966 967 968 969
	}
	return nc;
}

970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993
/*
 * Transfer objects in one arraycache to another.
 * Locking must be handled by the caller.
 *
 * Return the number of entries transferred.
 */
static int transfer_objects(struct array_cache *to,
		struct array_cache *from, unsigned int max)
{
	/* Figure out how many entries to transfer */
	int nr = min(min(from->avail, max), to->limit - to->avail);

	if (!nr)
		return 0;

	memcpy(to->entry + to->avail, from->entry + from->avail -nr,
			sizeof(void *) *nr);

	from->avail -= nr;
	to->avail += nr;
	to->touched = 1;
	return nr;
}

994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018
#ifndef CONFIG_NUMA

#define drain_alien_cache(cachep, alien) do { } while (0)
#define reap_alien(cachep, l3) do { } while (0)

static inline struct array_cache **alloc_alien_cache(int node, int limit)
{
	return (struct array_cache **)BAD_ALIEN_MAGIC;
}

static inline void free_alien_cache(struct array_cache **ac_ptr)
{
}

static inline int cache_free_alien(struct kmem_cache *cachep, void *objp)
{
	return 0;
}

static inline void *alternate_node_alloc(struct kmem_cache *cachep,
		gfp_t flags)
{
	return NULL;
}

1019
static inline void *____cache_alloc_node(struct kmem_cache *cachep,
1020 1021 1022 1023 1024 1025 1026
		 gfp_t flags, int nodeid)
{
	return NULL;
}

#else	/* CONFIG_NUMA */

1027
static void *____cache_alloc_node(struct kmem_cache *, gfp_t, int);
1028
static void *alternate_node_alloc(struct kmem_cache *, gfp_t);
1029

P
Pekka Enberg 已提交
1030
static struct array_cache **alloc_alien_cache(int node, int limit)
1031 1032
{
	struct array_cache **ac_ptr;
1033
	int memsize = sizeof(void *) * nr_node_ids;
1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046
	int i;

	if (limit > 1)
		limit = 12;
	ac_ptr = kmalloc_node(memsize, GFP_KERNEL, node);
	if (ac_ptr) {
		for_each_node(i) {
			if (i == node || !node_online(i)) {
				ac_ptr[i] = NULL;
				continue;
			}
			ac_ptr[i] = alloc_arraycache(node, limit, 0xbaadf00d);
			if (!ac_ptr[i]) {
P
Pekka Enberg 已提交
1047
				for (i--; i <= 0; i--)
1048 1049 1050 1051 1052 1053 1054 1055 1056
					kfree(ac_ptr[i]);
				kfree(ac_ptr);
				return NULL;
			}
		}
	}
	return ac_ptr;
}

P
Pekka Enberg 已提交
1057
static void free_alien_cache(struct array_cache **ac_ptr)
1058 1059 1060 1061 1062 1063
{
	int i;

	if (!ac_ptr)
		return;
	for_each_node(i)
P
Pekka Enberg 已提交
1064
	    kfree(ac_ptr[i]);
1065 1066 1067
	kfree(ac_ptr);
}

1068
static void __drain_alien_cache(struct kmem_cache *cachep,
P
Pekka Enberg 已提交
1069
				struct array_cache *ac, int node)
1070 1071 1072 1073 1074
{
	struct kmem_list3 *rl3 = cachep->nodelists[node];

	if (ac->avail) {
		spin_lock(&rl3->list_lock);
1075 1076 1077 1078 1079
		/*
		 * Stuff objects into the remote nodes shared array first.
		 * That way we could avoid the overhead of putting the objects
		 * into the free lists and getting them back later.
		 */
1080 1081
		if (rl3->shared)
			transfer_objects(rl3->shared, ac, ac->limit);
1082

1083
		free_block(cachep, ac->entry, ac->avail, node);
1084 1085 1086 1087 1088
		ac->avail = 0;
		spin_unlock(&rl3->list_lock);
	}
}

1089 1090 1091 1092 1093 1094 1095 1096 1097
/*
 * Called from cache_reap() to regularly drain alien caches round robin.
 */
static void reap_alien(struct kmem_cache *cachep, struct kmem_list3 *l3)
{
	int node = __get_cpu_var(reap_node);

	if (l3->alien) {
		struct array_cache *ac = l3->alien[node];
1098 1099

		if (ac && ac->avail && spin_trylock_irq(&ac->lock)) {
1100 1101 1102 1103 1104 1105
			__drain_alien_cache(cachep, ac, node);
			spin_unlock_irq(&ac->lock);
		}
	}
}

A
Andrew Morton 已提交
1106 1107
static void drain_alien_cache(struct kmem_cache *cachep,
				struct array_cache **alien)
1108
{
P
Pekka Enberg 已提交
1109
	int i = 0;
1110 1111 1112 1113
	struct array_cache *ac;
	unsigned long flags;

	for_each_online_node(i) {
1114
		ac = alien[i];
1115 1116 1117 1118 1119 1120 1121
		if (ac) {
			spin_lock_irqsave(&ac->lock, flags);
			__drain_alien_cache(cachep, ac, i);
			spin_unlock_irqrestore(&ac->lock, flags);
		}
	}
}
1122

1123
static inline int cache_free_alien(struct kmem_cache *cachep, void *objp)
1124 1125 1126 1127 1128
{
	struct slab *slabp = virt_to_slab(objp);
	int nodeid = slabp->nodeid;
	struct kmem_list3 *l3;
	struct array_cache *alien = NULL;
P
Pekka Enberg 已提交
1129 1130 1131
	int node;

	node = numa_node_id();
1132 1133 1134 1135 1136

	/*
	 * Make sure we are not freeing a object from another node to the array
	 * cache on this cpu.
	 */
1137
	if (likely(slabp->nodeid == node))
1138 1139
		return 0;

P
Pekka Enberg 已提交
1140
	l3 = cachep->nodelists[node];
1141 1142 1143
	STATS_INC_NODEFREES(cachep);
	if (l3->alien && l3->alien[nodeid]) {
		alien = l3->alien[nodeid];
1144
		spin_lock(&alien->lock);
1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157
		if (unlikely(alien->avail == alien->limit)) {
			STATS_INC_ACOVERFLOW(cachep);
			__drain_alien_cache(cachep, alien, nodeid);
		}
		alien->entry[alien->avail++] = objp;
		spin_unlock(&alien->lock);
	} else {
		spin_lock(&(cachep->nodelists[nodeid])->list_lock);
		free_block(cachep, &objp, 1, nodeid);
		spin_unlock(&(cachep->nodelists[nodeid])->list_lock);
	}
	return 1;
}
1158 1159
#endif

1160
static int __cpuinit cpuup_callback(struct notifier_block *nfb,
P
Pekka Enberg 已提交
1161
				    unsigned long action, void *hcpu)
L
Linus Torvalds 已提交
1162 1163
{
	long cpu = (long)hcpu;
1164
	struct kmem_cache *cachep;
1165 1166
	struct kmem_list3 *l3 = NULL;
	int node = cpu_to_node(cpu);
1167
	const int memsize = sizeof(struct kmem_list3);
L
Linus Torvalds 已提交
1168 1169

	switch (action) {
1170
	case CPU_LOCK_ACQUIRE:
I
Ingo Molnar 已提交
1171
		mutex_lock(&cache_chain_mutex);
1172 1173
		break;
	case CPU_UP_PREPARE:
1174
	case CPU_UP_PREPARE_FROZEN:
A
Andrew Morton 已提交
1175 1176
		/*
		 * We need to do this right in the beginning since
1177 1178 1179 1180 1181
		 * alloc_arraycache's are going to use this list.
		 * kmalloc_node allows us to add the slab to the right
		 * kmem_list3 and not this cpu's kmem_list3
		 */

L
Linus Torvalds 已提交
1182
		list_for_each_entry(cachep, &cache_chain, next) {
A
Andrew Morton 已提交
1183 1184
			/*
			 * Set up the size64 kmemlist for cpu before we can
1185 1186 1187 1188
			 * begin anything. Make sure some other cpu on this
			 * node has not already allocated this
			 */
			if (!cachep->nodelists[node]) {
A
Andrew Morton 已提交
1189 1190
				l3 = kmalloc_node(memsize, GFP_KERNEL, node);
				if (!l3)
1191 1192 1193
					goto bad;
				kmem_list3_init(l3);
				l3->next_reap = jiffies + REAPTIMEOUT_LIST3 +
P
Pekka Enberg 已提交
1194
				    ((unsigned long)cachep) % REAPTIMEOUT_LIST3;
1195

1196 1197 1198 1199 1200
				/*
				 * The l3s don't come and go as CPUs come and
				 * go.  cache_chain_mutex is sufficient
				 * protection here.
				 */
1201 1202
				cachep->nodelists[node] = l3;
			}
L
Linus Torvalds 已提交
1203

1204 1205
			spin_lock_irq(&cachep->nodelists[node]->list_lock);
			cachep->nodelists[node]->free_limit =
A
Andrew Morton 已提交
1206 1207
				(1 + nr_cpus_node(node)) *
				cachep->batchcount + cachep->num;
1208 1209 1210
			spin_unlock_irq(&cachep->nodelists[node]->list_lock);
		}

A
Andrew Morton 已提交
1211 1212 1213 1214
		/*
		 * Now we can go ahead with allocating the shared arrays and
		 * array caches
		 */
1215
		list_for_each_entry(cachep, &cache_chain, next) {
1216
			struct array_cache *nc;
1217
			struct array_cache *shared = NULL;
1218
			struct array_cache **alien = NULL;
1219

1220
			nc = alloc_arraycache(node, cachep->limit,
1221
						cachep->batchcount);
L
Linus Torvalds 已提交
1222 1223
			if (!nc)
				goto bad;
1224 1225
			if (cachep->shared) {
				shared = alloc_arraycache(node,
1226 1227
					cachep->shared * cachep->batchcount,
					0xbaadf00d);
1228 1229 1230
				if (!shared)
					goto bad;
			}
1231 1232 1233 1234 1235
			if (use_alien_caches) {
                                alien = alloc_alien_cache(node, cachep->limit);
                                if (!alien)
                                        goto bad;
                        }
L
Linus Torvalds 已提交
1236
			cachep->array[cpu] = nc;
1237 1238 1239
			l3 = cachep->nodelists[node];
			BUG_ON(!l3);

1240 1241 1242 1243 1244 1245 1246 1247
			spin_lock_irq(&l3->list_lock);
			if (!l3->shared) {
				/*
				 * We are serialised from CPU_DEAD or
				 * CPU_UP_CANCELLED by the cpucontrol lock
				 */
				l3->shared = shared;
				shared = NULL;
1248
			}
1249 1250 1251 1252 1253 1254 1255 1256 1257
#ifdef CONFIG_NUMA
			if (!l3->alien) {
				l3->alien = alien;
				alien = NULL;
			}
#endif
			spin_unlock_irq(&l3->list_lock);
			kfree(shared);
			free_alien_cache(alien);
L
Linus Torvalds 已提交
1258 1259 1260
		}
		break;
	case CPU_ONLINE:
1261
	case CPU_ONLINE_FROZEN:
L
Linus Torvalds 已提交
1262 1263 1264
		start_cpu_timer(cpu);
		break;
#ifdef CONFIG_HOTPLUG_CPU
1265
  	case CPU_DOWN_PREPARE:
1266
  	case CPU_DOWN_PREPARE_FROZEN:
1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277
		/*
		 * Shutdown cache reaper. Note that the cache_chain_mutex is
		 * held so that if cache_reap() is invoked it cannot do
		 * anything expensive but will only modify reap_work
		 * and reschedule the timer.
		*/
		cancel_rearming_delayed_work(&per_cpu(reap_work, cpu));
		/* Now the cache_reaper is guaranteed to be not running. */
		per_cpu(reap_work, cpu).work.func = NULL;
  		break;
  	case CPU_DOWN_FAILED:
1278
  	case CPU_DOWN_FAILED_FROZEN:
1279 1280
		start_cpu_timer(cpu);
  		break;
L
Linus Torvalds 已提交
1281
	case CPU_DEAD:
1282
	case CPU_DEAD_FROZEN:
1283 1284 1285 1286 1287 1288 1289 1290
		/*
		 * Even if all the cpus of a node are down, we don't free the
		 * kmem_list3 of any cache. This to avoid a race between
		 * cpu_down, and a kmalloc allocation from another cpu for
		 * memory from the node of the cpu going down.  The list3
		 * structure is usually allocated from kmem_cache_create() and
		 * gets destroyed at kmem_cache_destroy().
		 */
L
Linus Torvalds 已提交
1291
		/* fall thru */
1292
#endif
L
Linus Torvalds 已提交
1293
	case CPU_UP_CANCELED:
1294
	case CPU_UP_CANCELED_FROZEN:
L
Linus Torvalds 已提交
1295 1296
		list_for_each_entry(cachep, &cache_chain, next) {
			struct array_cache *nc;
1297 1298
			struct array_cache *shared;
			struct array_cache **alien;
1299
			cpumask_t mask;
L
Linus Torvalds 已提交
1300

1301
			mask = node_to_cpumask(node);
L
Linus Torvalds 已提交
1302 1303 1304
			/* cpu is dead; no one can alloc from it. */
			nc = cachep->array[cpu];
			cachep->array[cpu] = NULL;
1305 1306 1307
			l3 = cachep->nodelists[node];

			if (!l3)
1308
				goto free_array_cache;
1309

1310
			spin_lock_irq(&l3->list_lock);
1311 1312 1313 1314

			/* Free limit for this kmem_list3 */
			l3->free_limit -= cachep->batchcount;
			if (nc)
1315
				free_block(cachep, nc->entry, nc->avail, node);
1316 1317

			if (!cpus_empty(mask)) {
1318
				spin_unlock_irq(&l3->list_lock);
1319
				goto free_array_cache;
P
Pekka Enberg 已提交
1320
			}
1321

1322 1323
			shared = l3->shared;
			if (shared) {
1324 1325
				free_block(cachep, shared->entry,
					   shared->avail, node);
1326 1327 1328
				l3->shared = NULL;
			}

1329 1330 1331 1332 1333 1334 1335 1336 1337
			alien = l3->alien;
			l3->alien = NULL;

			spin_unlock_irq(&l3->list_lock);

			kfree(shared);
			if (alien) {
				drain_alien_cache(cachep, alien);
				free_alien_cache(alien);
1338
			}
1339
free_array_cache:
L
Linus Torvalds 已提交
1340 1341
			kfree(nc);
		}
1342 1343 1344 1345 1346 1347 1348 1349 1350
		/*
		 * In the previous loop, all the objects were freed to
		 * the respective cache's slabs,  now we can go ahead and
		 * shrink each nodelist to its limit.
		 */
		list_for_each_entry(cachep, &cache_chain, next) {
			l3 = cachep->nodelists[node];
			if (!l3)
				continue;
1351
			drain_freelist(cachep, l3, l3->free_objects);
1352
		}
1353 1354
		break;
	case CPU_LOCK_RELEASE:
I
Ingo Molnar 已提交
1355
		mutex_unlock(&cache_chain_mutex);
L
Linus Torvalds 已提交
1356 1357 1358
		break;
	}
	return NOTIFY_OK;
A
Andrew Morton 已提交
1359
bad:
L
Linus Torvalds 已提交
1360 1361 1362
	return NOTIFY_BAD;
}

1363 1364 1365
static struct notifier_block __cpuinitdata cpucache_notifier = {
	&cpuup_callback, NULL, 0
};
L
Linus Torvalds 已提交
1366

1367 1368 1369
/*
 * swap the static kmem_list3 with kmalloced memory
 */
A
Andrew Morton 已提交
1370 1371
static void init_list(struct kmem_cache *cachep, struct kmem_list3 *list,
			int nodeid)
1372 1373 1374 1375 1376 1377 1378 1379
{
	struct kmem_list3 *ptr;

	ptr = kmalloc_node(sizeof(struct kmem_list3), GFP_KERNEL, nodeid);
	BUG_ON(!ptr);

	local_irq_disable();
	memcpy(ptr, list, sizeof(struct kmem_list3));
1380 1381 1382 1383 1384
	/*
	 * Do not assume that spinlocks can be initialized via memcpy:
	 */
	spin_lock_init(&ptr->list_lock);

1385 1386 1387 1388 1389
	MAKE_ALL_LISTS(cachep, ptr, nodeid);
	cachep->nodelists[nodeid] = ptr;
	local_irq_enable();
}

A
Andrew Morton 已提交
1390 1391 1392
/*
 * Initialisation.  Called after the page allocator have been initialised and
 * before smp_init().
L
Linus Torvalds 已提交
1393 1394 1395 1396 1397 1398
 */
void __init kmem_cache_init(void)
{
	size_t left_over;
	struct cache_sizes *sizes;
	struct cache_names *names;
1399
	int i;
1400
	int order;
P
Pekka Enberg 已提交
1401
	int node;
1402

1403
	if (num_possible_nodes() == 1) {
1404
		use_alien_caches = 0;
1405 1406
		numa_platform = 0;
	}
1407

1408 1409 1410 1411 1412
	for (i = 0; i < NUM_INIT_LISTS; i++) {
		kmem_list3_init(&initkmem_list3[i]);
		if (i < MAX_NUMNODES)
			cache_cache.nodelists[i] = NULL;
	}
L
Linus Torvalds 已提交
1413 1414 1415 1416 1417 1418 1419 1420 1421 1422

	/*
	 * Fragmentation resistance on low memory - only use bigger
	 * page orders on machines with more than 32MB of memory.
	 */
	if (num_physpages > (32 << 20) >> PAGE_SHIFT)
		slab_break_gfp_order = BREAK_GFP_ORDER_HI;

	/* Bootstrap is tricky, because several objects are allocated
	 * from caches that do not exist yet:
A
Andrew Morton 已提交
1423 1424 1425
	 * 1) initialize the cache_cache cache: it contains the struct
	 *    kmem_cache structures of all caches, except cache_cache itself:
	 *    cache_cache is statically allocated.
1426 1427 1428
	 *    Initially an __init data area is used for the head array and the
	 *    kmem_list3 structures, it's replaced with a kmalloc allocated
	 *    array at the end of the bootstrap.
L
Linus Torvalds 已提交
1429
	 * 2) Create the first kmalloc cache.
1430
	 *    The struct kmem_cache for the new cache is allocated normally.
1431 1432 1433
	 *    An __init data area is used for the head array.
	 * 3) Create the remaining kmalloc caches, with minimally sized
	 *    head arrays.
L
Linus Torvalds 已提交
1434 1435
	 * 4) Replace the __init data head arrays for cache_cache and the first
	 *    kmalloc cache with kmalloc allocated arrays.
1436 1437 1438
	 * 5) Replace the __init data for kmem_list3 for cache_cache and
	 *    the other cache's with kmalloc allocated memory.
	 * 6) Resize the head arrays of the kmalloc caches to their final sizes.
L
Linus Torvalds 已提交
1439 1440
	 */

P
Pekka Enberg 已提交
1441 1442
	node = numa_node_id();

L
Linus Torvalds 已提交
1443 1444 1445 1446 1447
	/* 1) create the cache_cache */
	INIT_LIST_HEAD(&cache_chain);
	list_add(&cache_cache.next, &cache_chain);
	cache_cache.colour_off = cache_line_size();
	cache_cache.array[smp_processor_id()] = &initarray_cache.cache;
P
Pekka Enberg 已提交
1448
	cache_cache.nodelists[node] = &initkmem_list3[CACHE_CACHE];
L
Linus Torvalds 已提交
1449

E
Eric Dumazet 已提交
1450 1451 1452 1453 1454 1455 1456 1457 1458
	/*
	 * struct kmem_cache size depends on nr_node_ids, which
	 * can be less than MAX_NUMNODES.
	 */
	cache_cache.buffer_size = offsetof(struct kmem_cache, nodelists) +
				 nr_node_ids * sizeof(struct kmem_list3 *);
#if DEBUG
	cache_cache.obj_size = cache_cache.buffer_size;
#endif
A
Andrew Morton 已提交
1459 1460
	cache_cache.buffer_size = ALIGN(cache_cache.buffer_size,
					cache_line_size());
1461 1462
	cache_cache.reciprocal_buffer_size =
		reciprocal_value(cache_cache.buffer_size);
L
Linus Torvalds 已提交
1463

1464 1465 1466 1467 1468 1469
	for (order = 0; order < MAX_ORDER; order++) {
		cache_estimate(order, cache_cache.buffer_size,
			cache_line_size(), 0, &left_over, &cache_cache.num);
		if (cache_cache.num)
			break;
	}
1470
	BUG_ON(!cache_cache.num);
1471
	cache_cache.gfporder = order;
P
Pekka Enberg 已提交
1472 1473 1474
	cache_cache.colour = left_over / cache_cache.colour_off;
	cache_cache.slab_size = ALIGN(cache_cache.num * sizeof(kmem_bufctl_t) +
				      sizeof(struct slab), cache_line_size());
L
Linus Torvalds 已提交
1475 1476 1477 1478 1479

	/* 2+3) create the kmalloc caches */
	sizes = malloc_sizes;
	names = cache_names;

A
Andrew Morton 已提交
1480 1481 1482 1483
	/*
	 * Initialize the caches that provide memory for the array cache and the
	 * kmem_list3 structures first.  Without this, further allocations will
	 * bug.
1484 1485 1486
	 */

	sizes[INDEX_AC].cs_cachep = kmem_cache_create(names[INDEX_AC].name,
A
Andrew Morton 已提交
1487 1488 1489
					sizes[INDEX_AC].cs_size,
					ARCH_KMALLOC_MINALIGN,
					ARCH_KMALLOC_FLAGS|SLAB_PANIC,
1490
					NULL);
1491

A
Andrew Morton 已提交
1492
	if (INDEX_AC != INDEX_L3) {
1493
		sizes[INDEX_L3].cs_cachep =
A
Andrew Morton 已提交
1494 1495 1496 1497
			kmem_cache_create(names[INDEX_L3].name,
				sizes[INDEX_L3].cs_size,
				ARCH_KMALLOC_MINALIGN,
				ARCH_KMALLOC_FLAGS|SLAB_PANIC,
1498
				NULL);
A
Andrew Morton 已提交
1499
	}
1500

1501 1502
	slab_early_init = 0;

L
Linus Torvalds 已提交
1503
	while (sizes->cs_size != ULONG_MAX) {
1504 1505
		/*
		 * For performance, all the general caches are L1 aligned.
L
Linus Torvalds 已提交
1506 1507 1508
		 * This should be particularly beneficial on SMP boxes, as it
		 * eliminates "false sharing".
		 * Note for systems short on memory removing the alignment will
1509 1510
		 * allow tighter packing of the smaller caches.
		 */
A
Andrew Morton 已提交
1511
		if (!sizes->cs_cachep) {
1512
			sizes->cs_cachep = kmem_cache_create(names->name,
A
Andrew Morton 已提交
1513 1514 1515
					sizes->cs_size,
					ARCH_KMALLOC_MINALIGN,
					ARCH_KMALLOC_FLAGS|SLAB_PANIC,
1516
					NULL);
A
Andrew Morton 已提交
1517
		}
1518 1519 1520
#ifdef CONFIG_ZONE_DMA
		sizes->cs_dmacachep = kmem_cache_create(
					names->name_dma,
A
Andrew Morton 已提交
1521 1522 1523 1524
					sizes->cs_size,
					ARCH_KMALLOC_MINALIGN,
					ARCH_KMALLOC_FLAGS|SLAB_CACHE_DMA|
						SLAB_PANIC,
1525
					NULL);
1526
#endif
L
Linus Torvalds 已提交
1527 1528 1529 1530 1531
		sizes++;
		names++;
	}
	/* 4) Replace the bootstrap head arrays */
	{
1532
		struct array_cache *ptr;
1533

L
Linus Torvalds 已提交
1534
		ptr = kmalloc(sizeof(struct arraycache_init), GFP_KERNEL);
1535

L
Linus Torvalds 已提交
1536
		local_irq_disable();
1537 1538
		BUG_ON(cpu_cache_get(&cache_cache) != &initarray_cache.cache);
		memcpy(ptr, cpu_cache_get(&cache_cache),
P
Pekka Enberg 已提交
1539
		       sizeof(struct arraycache_init));
1540 1541 1542 1543 1544
		/*
		 * Do not assume that spinlocks can be initialized via memcpy:
		 */
		spin_lock_init(&ptr->lock);

L
Linus Torvalds 已提交
1545 1546
		cache_cache.array[smp_processor_id()] = ptr;
		local_irq_enable();
1547

L
Linus Torvalds 已提交
1548
		ptr = kmalloc(sizeof(struct arraycache_init), GFP_KERNEL);
1549

L
Linus Torvalds 已提交
1550
		local_irq_disable();
1551
		BUG_ON(cpu_cache_get(malloc_sizes[INDEX_AC].cs_cachep)
P
Pekka Enberg 已提交
1552
		       != &initarray_generic.cache);
1553
		memcpy(ptr, cpu_cache_get(malloc_sizes[INDEX_AC].cs_cachep),
P
Pekka Enberg 已提交
1554
		       sizeof(struct arraycache_init));
1555 1556 1557 1558 1559
		/*
		 * Do not assume that spinlocks can be initialized via memcpy:
		 */
		spin_lock_init(&ptr->lock);

1560
		malloc_sizes[INDEX_AC].cs_cachep->array[smp_processor_id()] =
P
Pekka Enberg 已提交
1561
		    ptr;
L
Linus Torvalds 已提交
1562 1563
		local_irq_enable();
	}
1564 1565
	/* 5) Replace the bootstrap kmem_list3's */
	{
P
Pekka Enberg 已提交
1566 1567
		int nid;

1568
		/* Replace the static kmem_list3 structures for the boot cpu */
P
Pekka Enberg 已提交
1569
		init_list(&cache_cache, &initkmem_list3[CACHE_CACHE], node);
1570

P
Pekka Enberg 已提交
1571
		for_each_online_node(nid) {
1572
			init_list(malloc_sizes[INDEX_AC].cs_cachep,
P
Pekka Enberg 已提交
1573
				  &initkmem_list3[SIZE_AC + nid], nid);
1574 1575 1576

			if (INDEX_AC != INDEX_L3) {
				init_list(malloc_sizes[INDEX_L3].cs_cachep,
P
Pekka Enberg 已提交
1577
					  &initkmem_list3[SIZE_L3 + nid], nid);
1578 1579 1580
			}
		}
	}
L
Linus Torvalds 已提交
1581

1582
	/* 6) resize the head arrays to their final sizes */
L
Linus Torvalds 已提交
1583
	{
1584
		struct kmem_cache *cachep;
I
Ingo Molnar 已提交
1585
		mutex_lock(&cache_chain_mutex);
L
Linus Torvalds 已提交
1586
		list_for_each_entry(cachep, &cache_chain, next)
1587 1588
			if (enable_cpucache(cachep))
				BUG();
I
Ingo Molnar 已提交
1589
		mutex_unlock(&cache_chain_mutex);
L
Linus Torvalds 已提交
1590 1591
	}

1592 1593 1594 1595
	/* Annotate slab for lockdep -- annotate the malloc caches */
	init_lock_keys();


L
Linus Torvalds 已提交
1596 1597 1598
	/* Done! */
	g_cpucache_up = FULL;

A
Andrew Morton 已提交
1599 1600 1601
	/*
	 * Register a cpu startup notifier callback that initializes
	 * cpu_cache_get for all new cpus
L
Linus Torvalds 已提交
1602 1603 1604
	 */
	register_cpu_notifier(&cpucache_notifier);

A
Andrew Morton 已提交
1605 1606 1607
	/*
	 * The reap timers are started later, with a module init call: That part
	 * of the kernel is not yet operational.
L
Linus Torvalds 已提交
1608 1609 1610 1611 1612 1613 1614
	 */
}

static int __init cpucache_init(void)
{
	int cpu;

A
Andrew Morton 已提交
1615 1616
	/*
	 * Register the timers that return unneeded pages to the page allocator
L
Linus Torvalds 已提交
1617
	 */
1618
	for_each_online_cpu(cpu)
A
Andrew Morton 已提交
1619
		start_cpu_timer(cpu);
L
Linus Torvalds 已提交
1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630
	return 0;
}
__initcall(cpucache_init);

/*
 * Interface to system's page allocator. No need to hold the cache-lock.
 *
 * If we requested dmaable memory, we will get it. Even if we
 * did not request dmaable memory, we might get it, but that
 * would be relatively rare and ignorable.
 */
1631
static void *kmem_getpages(struct kmem_cache *cachep, gfp_t flags, int nodeid)
L
Linus Torvalds 已提交
1632 1633
{
	struct page *page;
1634
	int nr_pages;
L
Linus Torvalds 已提交
1635 1636
	int i;

1637
#ifndef CONFIG_MMU
1638 1639 1640
	/*
	 * Nommu uses slab's for process anonymous memory allocations, and thus
	 * requires __GFP_COMP to properly refcount higher order allocations
1641
	 */
1642
	flags |= __GFP_COMP;
1643
#endif
1644

1645
	flags |= cachep->gfpflags;
1646 1647

	page = alloc_pages_node(nodeid, flags, cachep->gfporder);
L
Linus Torvalds 已提交
1648 1649 1650
	if (!page)
		return NULL;

1651
	nr_pages = (1 << cachep->gfporder);
L
Linus Torvalds 已提交
1652
	if (cachep->flags & SLAB_RECLAIM_ACCOUNT)
1653 1654 1655 1656 1657
		add_zone_page_state(page_zone(page),
			NR_SLAB_RECLAIMABLE, nr_pages);
	else
		add_zone_page_state(page_zone(page),
			NR_SLAB_UNRECLAIMABLE, nr_pages);
1658 1659 1660
	for (i = 0; i < nr_pages; i++)
		__SetPageSlab(page + i);
	return page_address(page);
L
Linus Torvalds 已提交
1661 1662 1663 1664 1665
}

/*
 * Interface to system's page release.
 */
1666
static void kmem_freepages(struct kmem_cache *cachep, void *addr)
L
Linus Torvalds 已提交
1667
{
P
Pekka Enberg 已提交
1668
	unsigned long i = (1 << cachep->gfporder);
L
Linus Torvalds 已提交
1669 1670 1671
	struct page *page = virt_to_page(addr);
	const unsigned long nr_freed = i;

1672 1673 1674 1675 1676 1677
	if (cachep->flags & SLAB_RECLAIM_ACCOUNT)
		sub_zone_page_state(page_zone(page),
				NR_SLAB_RECLAIMABLE, nr_freed);
	else
		sub_zone_page_state(page_zone(page),
				NR_SLAB_UNRECLAIMABLE, nr_freed);
L
Linus Torvalds 已提交
1678
	while (i--) {
N
Nick Piggin 已提交
1679 1680
		BUG_ON(!PageSlab(page));
		__ClearPageSlab(page);
L
Linus Torvalds 已提交
1681 1682 1683 1684 1685 1686 1687 1688 1689
		page++;
	}
	if (current->reclaim_state)
		current->reclaim_state->reclaimed_slab += nr_freed;
	free_pages((unsigned long)addr, cachep->gfporder);
}

static void kmem_rcu_free(struct rcu_head *head)
{
P
Pekka Enberg 已提交
1690
	struct slab_rcu *slab_rcu = (struct slab_rcu *)head;
1691
	struct kmem_cache *cachep = slab_rcu->cachep;
L
Linus Torvalds 已提交
1692 1693 1694 1695 1696 1697 1698 1699 1700

	kmem_freepages(cachep, slab_rcu->addr);
	if (OFF_SLAB(cachep))
		kmem_cache_free(cachep->slabp_cache, slab_rcu);
}

#if DEBUG

#ifdef CONFIG_DEBUG_PAGEALLOC
1701
static void store_stackinfo(struct kmem_cache *cachep, unsigned long *addr,
P
Pekka Enberg 已提交
1702
			    unsigned long caller)
L
Linus Torvalds 已提交
1703
{
1704
	int size = obj_size(cachep);
L
Linus Torvalds 已提交
1705

1706
	addr = (unsigned long *)&((char *)addr)[obj_offset(cachep)];
L
Linus Torvalds 已提交
1707

P
Pekka Enberg 已提交
1708
	if (size < 5 * sizeof(unsigned long))
L
Linus Torvalds 已提交
1709 1710
		return;

P
Pekka Enberg 已提交
1711 1712 1713 1714
	*addr++ = 0x12345678;
	*addr++ = caller;
	*addr++ = smp_processor_id();
	size -= 3 * sizeof(unsigned long);
L
Linus Torvalds 已提交
1715 1716 1717 1718 1719 1720 1721
	{
		unsigned long *sptr = &caller;
		unsigned long svalue;

		while (!kstack_end(sptr)) {
			svalue = *sptr++;
			if (kernel_text_address(svalue)) {
P
Pekka Enberg 已提交
1722
				*addr++ = svalue;
L
Linus Torvalds 已提交
1723 1724 1725 1726 1727 1728 1729
				size -= sizeof(unsigned long);
				if (size <= sizeof(unsigned long))
					break;
			}
		}

	}
P
Pekka Enberg 已提交
1730
	*addr++ = 0x87654321;
L
Linus Torvalds 已提交
1731 1732 1733
}
#endif

1734
static void poison_obj(struct kmem_cache *cachep, void *addr, unsigned char val)
L
Linus Torvalds 已提交
1735
{
1736 1737
	int size = obj_size(cachep);
	addr = &((char *)addr)[obj_offset(cachep)];
L
Linus Torvalds 已提交
1738 1739

	memset(addr, val, size);
P
Pekka Enberg 已提交
1740
	*(unsigned char *)(addr + size - 1) = POISON_END;
L
Linus Torvalds 已提交
1741 1742 1743 1744 1745
}

static void dump_line(char *data, int offset, int limit)
{
	int i;
D
Dave Jones 已提交
1746 1747 1748
	unsigned char error = 0;
	int bad_count = 0;

L
Linus Torvalds 已提交
1749
	printk(KERN_ERR "%03x:", offset);
D
Dave Jones 已提交
1750 1751 1752 1753 1754
	for (i = 0; i < limit; i++) {
		if (data[offset + i] != POISON_FREE) {
			error = data[offset + i];
			bad_count++;
		}
P
Pekka Enberg 已提交
1755
		printk(" %02x", (unsigned char)data[offset + i]);
D
Dave Jones 已提交
1756
	}
L
Linus Torvalds 已提交
1757
	printk("\n");
D
Dave Jones 已提交
1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771

	if (bad_count == 1) {
		error ^= POISON_FREE;
		if (!(error & (error - 1))) {
			printk(KERN_ERR "Single bit error detected. Probably "
					"bad RAM.\n");
#ifdef CONFIG_X86
			printk(KERN_ERR "Run memtest86+ or a similar memory "
					"test tool.\n");
#else
			printk(KERN_ERR "Run a memory test tool.\n");
#endif
		}
	}
L
Linus Torvalds 已提交
1772 1773 1774 1775 1776
}
#endif

#if DEBUG

1777
static void print_objinfo(struct kmem_cache *cachep, void *objp, int lines)
L
Linus Torvalds 已提交
1778 1779 1780 1781 1782
{
	int i, size;
	char *realobj;

	if (cachep->flags & SLAB_RED_ZONE) {
1783
		printk(KERN_ERR "Redzone: 0x%llx/0x%llx.\n",
A
Andrew Morton 已提交
1784 1785
			*dbg_redzone1(cachep, objp),
			*dbg_redzone2(cachep, objp));
L
Linus Torvalds 已提交
1786 1787 1788 1789
	}

	if (cachep->flags & SLAB_STORE_USER) {
		printk(KERN_ERR "Last user: [<%p>]",
A
Andrew Morton 已提交
1790
			*dbg_userword(cachep, objp));
L
Linus Torvalds 已提交
1791
		print_symbol("(%s)",
A
Andrew Morton 已提交
1792
				(unsigned long)*dbg_userword(cachep, objp));
L
Linus Torvalds 已提交
1793 1794
		printk("\n");
	}
1795 1796
	realobj = (char *)objp + obj_offset(cachep);
	size = obj_size(cachep);
P
Pekka Enberg 已提交
1797
	for (i = 0; i < size && lines; i += 16, lines--) {
L
Linus Torvalds 已提交
1798 1799
		int limit;
		limit = 16;
P
Pekka Enberg 已提交
1800 1801
		if (i + limit > size)
			limit = size - i;
L
Linus Torvalds 已提交
1802 1803 1804 1805
		dump_line(realobj, i, limit);
	}
}

1806
static void check_poison_obj(struct kmem_cache *cachep, void *objp)
L
Linus Torvalds 已提交
1807 1808 1809 1810 1811
{
	char *realobj;
	int size, i;
	int lines = 0;

1812 1813
	realobj = (char *)objp + obj_offset(cachep);
	size = obj_size(cachep);
L
Linus Torvalds 已提交
1814

P
Pekka Enberg 已提交
1815
	for (i = 0; i < size; i++) {
L
Linus Torvalds 已提交
1816
		char exp = POISON_FREE;
P
Pekka Enberg 已提交
1817
		if (i == size - 1)
L
Linus Torvalds 已提交
1818 1819 1820 1821 1822 1823
			exp = POISON_END;
		if (realobj[i] != exp) {
			int limit;
			/* Mismatch ! */
			/* Print header */
			if (lines == 0) {
P
Pekka Enberg 已提交
1824
				printk(KERN_ERR
1825 1826
					"Slab corruption: %s start=%p, len=%d\n",
					cachep->name, realobj, size);
L
Linus Torvalds 已提交
1827 1828 1829
				print_objinfo(cachep, objp, 0);
			}
			/* Hexdump the affected line */
P
Pekka Enberg 已提交
1830
			i = (i / 16) * 16;
L
Linus Torvalds 已提交
1831
			limit = 16;
P
Pekka Enberg 已提交
1832 1833
			if (i + limit > size)
				limit = size - i;
L
Linus Torvalds 已提交
1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845
			dump_line(realobj, i, limit);
			i += 16;
			lines++;
			/* Limit to 5 lines */
			if (lines > 5)
				break;
		}
	}
	if (lines != 0) {
		/* Print some data about the neighboring objects, if they
		 * exist:
		 */
1846
		struct slab *slabp = virt_to_slab(objp);
1847
		unsigned int objnr;
L
Linus Torvalds 已提交
1848

1849
		objnr = obj_to_index(cachep, slabp, objp);
L
Linus Torvalds 已提交
1850
		if (objnr) {
1851
			objp = index_to_obj(cachep, slabp, objnr - 1);
1852
			realobj = (char *)objp + obj_offset(cachep);
L
Linus Torvalds 已提交
1853
			printk(KERN_ERR "Prev obj: start=%p, len=%d\n",
P
Pekka Enberg 已提交
1854
			       realobj, size);
L
Linus Torvalds 已提交
1855 1856
			print_objinfo(cachep, objp, 2);
		}
P
Pekka Enberg 已提交
1857
		if (objnr + 1 < cachep->num) {
1858
			objp = index_to_obj(cachep, slabp, objnr + 1);
1859
			realobj = (char *)objp + obj_offset(cachep);
L
Linus Torvalds 已提交
1860
			printk(KERN_ERR "Next obj: start=%p, len=%d\n",
P
Pekka Enberg 已提交
1861
			       realobj, size);
L
Linus Torvalds 已提交
1862 1863 1864 1865 1866 1867
			print_objinfo(cachep, objp, 2);
		}
	}
}
#endif

1868 1869
#if DEBUG
/**
1870 1871 1872 1873 1874 1875
 * slab_destroy_objs - destroy a slab and its objects
 * @cachep: cache pointer being destroyed
 * @slabp: slab pointer being destroyed
 *
 * Call the registered destructor for each object in a slab that is being
 * destroyed.
L
Linus Torvalds 已提交
1876
 */
1877
static void slab_destroy_objs(struct kmem_cache *cachep, struct slab *slabp)
L
Linus Torvalds 已提交
1878 1879 1880
{
	int i;
	for (i = 0; i < cachep->num; i++) {
1881
		void *objp = index_to_obj(cachep, slabp, i);
L
Linus Torvalds 已提交
1882 1883 1884

		if (cachep->flags & SLAB_POISON) {
#ifdef CONFIG_DEBUG_PAGEALLOC
A
Andrew Morton 已提交
1885 1886
			if (cachep->buffer_size % PAGE_SIZE == 0 &&
					OFF_SLAB(cachep))
P
Pekka Enberg 已提交
1887
				kernel_map_pages(virt_to_page(objp),
A
Andrew Morton 已提交
1888
					cachep->buffer_size / PAGE_SIZE, 1);
L
Linus Torvalds 已提交
1889 1890 1891 1892 1893 1894 1895 1896 1897
			else
				check_poison_obj(cachep, objp);
#else
			check_poison_obj(cachep, objp);
#endif
		}
		if (cachep->flags & SLAB_RED_ZONE) {
			if (*dbg_redzone1(cachep, objp) != RED_INACTIVE)
				slab_error(cachep, "start of a freed object "
P
Pekka Enberg 已提交
1898
					   "was overwritten");
L
Linus Torvalds 已提交
1899 1900
			if (*dbg_redzone2(cachep, objp) != RED_INACTIVE)
				slab_error(cachep, "end of a freed object "
P
Pekka Enberg 已提交
1901
					   "was overwritten");
L
Linus Torvalds 已提交
1902 1903
		}
	}
1904
}
L
Linus Torvalds 已提交
1905
#else
1906
static void slab_destroy_objs(struct kmem_cache *cachep, struct slab *slabp)
1907 1908
{
}
L
Linus Torvalds 已提交
1909 1910
#endif

1911 1912 1913 1914 1915
/**
 * slab_destroy - destroy and release all objects in a slab
 * @cachep: cache pointer being destroyed
 * @slabp: slab pointer being destroyed
 *
1916
 * Destroy all the objs in a slab, and release the mem back to the system.
A
Andrew Morton 已提交
1917 1918
 * Before calling the slab must have been unlinked from the cache.  The
 * cache-lock is not held/needed.
1919
 */
1920
static void slab_destroy(struct kmem_cache *cachep, struct slab *slabp)
1921 1922 1923 1924
{
	void *addr = slabp->s_mem - slabp->colouroff;

	slab_destroy_objs(cachep, slabp);
L
Linus Torvalds 已提交
1925 1926 1927
	if (unlikely(cachep->flags & SLAB_DESTROY_BY_RCU)) {
		struct slab_rcu *slab_rcu;

P
Pekka Enberg 已提交
1928
		slab_rcu = (struct slab_rcu *)slabp;
L
Linus Torvalds 已提交
1929 1930 1931 1932 1933
		slab_rcu->cachep = cachep;
		slab_rcu->addr = addr;
		call_rcu(&slab_rcu->head, kmem_rcu_free);
	} else {
		kmem_freepages(cachep, addr);
1934 1935
		if (OFF_SLAB(cachep))
			kmem_cache_free(cachep->slabp_cache, slabp);
L
Linus Torvalds 已提交
1936 1937 1938
	}
}

A
Andrew Morton 已提交
1939 1940 1941 1942
/*
 * For setting up all the kmem_list3s for cache whose buffer_size is same as
 * size of kmem_list3.
 */
1943
static void __init set_up_list3s(struct kmem_cache *cachep, int index)
1944 1945 1946 1947
{
	int node;

	for_each_online_node(node) {
P
Pekka Enberg 已提交
1948
		cachep->nodelists[node] = &initkmem_list3[index + node];
1949
		cachep->nodelists[node]->next_reap = jiffies +
P
Pekka Enberg 已提交
1950 1951
		    REAPTIMEOUT_LIST3 +
		    ((unsigned long)cachep) % REAPTIMEOUT_LIST3;
1952 1953 1954
	}
}

1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
static void __kmem_cache_destroy(struct kmem_cache *cachep)
{
	int i;
	struct kmem_list3 *l3;

	for_each_online_cpu(i)
	    kfree(cachep->array[i]);

	/* NUMA: free the list3 structures */
	for_each_online_node(i) {
		l3 = cachep->nodelists[i];
		if (l3) {
			kfree(l3->shared);
			free_alien_cache(l3->alien);
			kfree(l3);
		}
	}
	kmem_cache_free(&cache_cache, cachep);
}


1976
/**
1977 1978 1979 1980 1981 1982 1983
 * calculate_slab_order - calculate size (page order) of slabs
 * @cachep: pointer to the cache that is being created
 * @size: size of objects to be created in this cache.
 * @align: required alignment for the objects.
 * @flags: slab allocation flags
 *
 * Also calculates the number of objects per slab.
1984 1985 1986 1987 1988
 *
 * This could be made much more intelligent.  For now, try to avoid using
 * high order pages for slabs.  When the gfp() functions are more friendly
 * towards high-order requests, this should be changed.
 */
A
Andrew Morton 已提交
1989
static size_t calculate_slab_order(struct kmem_cache *cachep,
R
Randy Dunlap 已提交
1990
			size_t size, size_t align, unsigned long flags)
1991
{
1992
	unsigned long offslab_limit;
1993
	size_t left_over = 0;
1994
	int gfporder;
1995

1996
	for (gfporder = 0; gfporder <= KMALLOC_MAX_ORDER; gfporder++) {
1997 1998 1999
		unsigned int num;
		size_t remainder;

2000
		cache_estimate(gfporder, size, align, flags, &remainder, &num);
2001 2002
		if (!num)
			continue;
2003

2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
		if (flags & CFLGS_OFF_SLAB) {
			/*
			 * Max number of objs-per-slab for caches which
			 * use off-slab slabs. Needed to avoid a possible
			 * looping condition in cache_grow().
			 */
			offslab_limit = size - sizeof(struct slab);
			offslab_limit /= sizeof(kmem_bufctl_t);

 			if (num > offslab_limit)
				break;
		}
2016

2017
		/* Found something acceptable - save it away */
2018
		cachep->num = num;
2019
		cachep->gfporder = gfporder;
2020 2021
		left_over = remainder;

2022 2023 2024 2025 2026 2027 2028 2029
		/*
		 * A VFS-reclaimable slab tends to have most allocations
		 * as GFP_NOFS and we really don't want to have to be allocating
		 * higher-order pages when we are unable to shrink dcache.
		 */
		if (flags & SLAB_RECLAIM_ACCOUNT)
			break;

2030 2031 2032 2033
		/*
		 * Large number of objects is good, but very large slabs are
		 * currently bad for the gfp()s.
		 */
2034
		if (gfporder >= slab_break_gfp_order)
2035 2036
			break;

2037 2038 2039
		/*
		 * Acceptable internal fragmentation?
		 */
A
Andrew Morton 已提交
2040
		if (left_over * 8 <= (PAGE_SIZE << gfporder))
2041 2042 2043 2044 2045
			break;
	}
	return left_over;
}

2046
static int __init_refok setup_cpu_cache(struct kmem_cache *cachep)
2047
{
2048 2049 2050
	if (g_cpucache_up == FULL)
		return enable_cpucache(cachep);

2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096
	if (g_cpucache_up == NONE) {
		/*
		 * Note: the first kmem_cache_create must create the cache
		 * that's used by kmalloc(24), otherwise the creation of
		 * further caches will BUG().
		 */
		cachep->array[smp_processor_id()] = &initarray_generic.cache;

		/*
		 * If the cache that's used by kmalloc(sizeof(kmem_list3)) is
		 * the first cache, then we need to set up all its list3s,
		 * otherwise the creation of further caches will BUG().
		 */
		set_up_list3s(cachep, SIZE_AC);
		if (INDEX_AC == INDEX_L3)
			g_cpucache_up = PARTIAL_L3;
		else
			g_cpucache_up = PARTIAL_AC;
	} else {
		cachep->array[smp_processor_id()] =
			kmalloc(sizeof(struct arraycache_init), GFP_KERNEL);

		if (g_cpucache_up == PARTIAL_AC) {
			set_up_list3s(cachep, SIZE_L3);
			g_cpucache_up = PARTIAL_L3;
		} else {
			int node;
			for_each_online_node(node) {
				cachep->nodelists[node] =
				    kmalloc_node(sizeof(struct kmem_list3),
						GFP_KERNEL, node);
				BUG_ON(!cachep->nodelists[node]);
				kmem_list3_init(cachep->nodelists[node]);
			}
		}
	}
	cachep->nodelists[numa_node_id()]->next_reap =
			jiffies + REAPTIMEOUT_LIST3 +
			((unsigned long)cachep) % REAPTIMEOUT_LIST3;

	cpu_cache_get(cachep)->avail = 0;
	cpu_cache_get(cachep)->limit = BOOT_CPUCACHE_ENTRIES;
	cpu_cache_get(cachep)->batchcount = 1;
	cpu_cache_get(cachep)->touched = 0;
	cachep->batchcount = 1;
	cachep->limit = BOOT_CPUCACHE_ENTRIES;
2097
	return 0;
2098 2099
}

L
Linus Torvalds 已提交
2100 2101 2102 2103 2104 2105 2106 2107 2108 2109
/**
 * kmem_cache_create - Create a cache.
 * @name: A string which is used in /proc/slabinfo to identify this cache.
 * @size: The size of objects to be created in this cache.
 * @align: The required alignment for the objects.
 * @flags: SLAB flags
 * @ctor: A constructor for the objects.
 *
 * Returns a ptr to the cache on success, NULL on failure.
 * Cannot be called within a int, but can be interrupted.
2110
 * The @ctor is run when new pages are allocated by the cache.
L
Linus Torvalds 已提交
2111 2112
 *
 * @name must be valid until the cache is destroyed. This implies that
A
Andrew Morton 已提交
2113 2114
 * the module calling this has to destroy the cache before getting unloaded.
 *
L
Linus Torvalds 已提交
2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126
 * The flags are
 *
 * %SLAB_POISON - Poison the slab with a known test pattern (a5a5a5a5)
 * to catch references to uninitialised memory.
 *
 * %SLAB_RED_ZONE - Insert `Red' zones around the allocated memory to check
 * for buffer overruns.
 *
 * %SLAB_HWCACHE_ALIGN - Align the objects in this cache to a hardware
 * cacheline.  This can be beneficial if you're counting cycles as closely
 * as davem.
 */
2127
struct kmem_cache *
L
Linus Torvalds 已提交
2128
kmem_cache_create (const char *name, size_t size, size_t align,
A
Andrew Morton 已提交
2129
	unsigned long flags,
2130
	void (*ctor)(void*, struct kmem_cache *, unsigned long))
L
Linus Torvalds 已提交
2131 2132
{
	size_t left_over, slab_size, ralign;
2133
	struct kmem_cache *cachep = NULL, *pc;
L
Linus Torvalds 已提交
2134 2135 2136 2137

	/*
	 * Sanity checks... these are all serious usage bugs.
	 */
A
Andrew Morton 已提交
2138
	if (!name || in_interrupt() || (size < BYTES_PER_WORD) ||
2139
	    size > KMALLOC_MAX_SIZE) {
A
Andrew Morton 已提交
2140 2141
		printk(KERN_ERR "%s: Early error in slab %s\n", __FUNCTION__,
				name);
P
Pekka Enberg 已提交
2142 2143
		BUG();
	}
L
Linus Torvalds 已提交
2144

2145
	/*
2146 2147
	 * We use cache_chain_mutex to ensure a consistent view of
	 * cpu_online_map as well.  Please see cpuup_callback
2148
	 */
I
Ingo Molnar 已提交
2149
	mutex_lock(&cache_chain_mutex);
2150

2151
	list_for_each_entry(pc, &cache_chain, next) {
2152 2153 2154 2155 2156 2157 2158 2159
		char tmp;
		int res;

		/*
		 * This happens when the module gets unloaded and doesn't
		 * destroy its slab cache and no-one else reuses the vmalloc
		 * area of the module.  Print a warning.
		 */
2160
		res = probe_kernel_address(pc->name, tmp);
2161
		if (res) {
2162 2163
			printk(KERN_ERR
			       "SLAB: cache with size %d has lost its name\n",
2164
			       pc->buffer_size);
2165 2166 2167
			continue;
		}

P
Pekka Enberg 已提交
2168
		if (!strcmp(pc->name, name)) {
2169 2170
			printk(KERN_ERR
			       "kmem_cache_create: duplicate cache %s\n", name);
2171 2172 2173 2174 2175
			dump_stack();
			goto oops;
		}
	}

L
Linus Torvalds 已提交
2176 2177 2178 2179 2180 2181 2182 2183 2184
#if DEBUG
	WARN_ON(strchr(name, ' '));	/* It confuses parsers */
#if FORCED_DEBUG
	/*
	 * Enable redzoning and last user accounting, except for caches with
	 * large objects, if the increased size would increase the object size
	 * above the next power of two: caches with object sizes just above a
	 * power of two have a significant amount of internal fragmentation.
	 */
D
David Woodhouse 已提交
2185 2186
	if (size < 4096 || fls(size - 1) == fls(size-1 + REDZONE_ALIGN +
						2 * sizeof(unsigned long long)))
P
Pekka Enberg 已提交
2187
		flags |= SLAB_RED_ZONE | SLAB_STORE_USER;
L
Linus Torvalds 已提交
2188 2189 2190 2191 2192 2193 2194
	if (!(flags & SLAB_DESTROY_BY_RCU))
		flags |= SLAB_POISON;
#endif
	if (flags & SLAB_DESTROY_BY_RCU)
		BUG_ON(flags & SLAB_POISON);
#endif
	/*
A
Andrew Morton 已提交
2195 2196
	 * Always checks flags, a caller might be expecting debug support which
	 * isn't available.
L
Linus Torvalds 已提交
2197
	 */
2198
	BUG_ON(flags & ~CREATE_MASK);
L
Linus Torvalds 已提交
2199

A
Andrew Morton 已提交
2200 2201
	/*
	 * Check that size is in terms of words.  This is needed to avoid
L
Linus Torvalds 已提交
2202 2203 2204
	 * unaligned accesses for some archs when redzoning is used, and makes
	 * sure any on-slab bufctl's are also correctly aligned.
	 */
P
Pekka Enberg 已提交
2205 2206 2207
	if (size & (BYTES_PER_WORD - 1)) {
		size += (BYTES_PER_WORD - 1);
		size &= ~(BYTES_PER_WORD - 1);
L
Linus Torvalds 已提交
2208 2209
	}

A
Andrew Morton 已提交
2210 2211
	/* calculate the final buffer alignment: */

L
Linus Torvalds 已提交
2212 2213
	/* 1) arch recommendation: can be overridden for debug */
	if (flags & SLAB_HWCACHE_ALIGN) {
A
Andrew Morton 已提交
2214 2215 2216 2217
		/*
		 * Default alignment: as specified by the arch code.  Except if
		 * an object is really small, then squeeze multiple objects into
		 * one cacheline.
L
Linus Torvalds 已提交
2218 2219
		 */
		ralign = cache_line_size();
P
Pekka Enberg 已提交
2220
		while (size <= ralign / 2)
L
Linus Torvalds 已提交
2221 2222 2223 2224
			ralign /= 2;
	} else {
		ralign = BYTES_PER_WORD;
	}
2225 2226

	/*
D
David Woodhouse 已提交
2227 2228 2229
	 * Redzoning and user store require word alignment or possibly larger.
	 * Note this will be overridden by architecture or caller mandated
	 * alignment if either is greater than BYTES_PER_WORD.
2230
	 */
D
David Woodhouse 已提交
2231 2232 2233 2234 2235 2236 2237 2238 2239 2240
	if (flags & SLAB_STORE_USER)
		ralign = BYTES_PER_WORD;

	if (flags & SLAB_RED_ZONE) {
		ralign = REDZONE_ALIGN;
		/* If redzoning, ensure that the second redzone is suitably
		 * aligned, by adjusting the object size accordingly. */
		size += REDZONE_ALIGN - 1;
		size &= ~(REDZONE_ALIGN - 1);
	}
2241

2242
	/* 2) arch mandated alignment */
L
Linus Torvalds 已提交
2243 2244 2245
	if (ralign < ARCH_SLAB_MINALIGN) {
		ralign = ARCH_SLAB_MINALIGN;
	}
2246
	/* 3) caller mandated alignment */
L
Linus Torvalds 已提交
2247 2248 2249
	if (ralign < align) {
		ralign = align;
	}
2250
	/* disable debug if necessary */
2251
	if (ralign > __alignof__(unsigned long long))
2252
		flags &= ~(SLAB_RED_ZONE | SLAB_STORE_USER);
A
Andrew Morton 已提交
2253
	/*
2254
	 * 4) Store it.
L
Linus Torvalds 已提交
2255 2256 2257 2258
	 */
	align = ralign;

	/* Get cache's description obj. */
2259
	cachep = kmem_cache_zalloc(&cache_cache, GFP_KERNEL);
L
Linus Torvalds 已提交
2260
	if (!cachep)
2261
		goto oops;
L
Linus Torvalds 已提交
2262 2263

#if DEBUG
2264
	cachep->obj_size = size;
L
Linus Torvalds 已提交
2265

2266 2267 2268 2269
	/*
	 * Both debugging options require word-alignment which is calculated
	 * into align above.
	 */
L
Linus Torvalds 已提交
2270 2271
	if (flags & SLAB_RED_ZONE) {
		/* add space for red zone words */
2272 2273
		cachep->obj_offset += sizeof(unsigned long long);
		size += 2 * sizeof(unsigned long long);
L
Linus Torvalds 已提交
2274 2275
	}
	if (flags & SLAB_STORE_USER) {
2276
		/* user store requires one word storage behind the end of
D
David Woodhouse 已提交
2277 2278
		 * the real object. But if the second red zone needs to be
		 * aligned to 64 bits, we must allow that much space.
L
Linus Torvalds 已提交
2279
		 */
D
David Woodhouse 已提交
2280 2281 2282 2283
		if (flags & SLAB_RED_ZONE)
			size += REDZONE_ALIGN;
		else
			size += BYTES_PER_WORD;
L
Linus Torvalds 已提交
2284 2285
	}
#if FORCED_DEBUG && defined(CONFIG_DEBUG_PAGEALLOC)
P
Pekka Enberg 已提交
2286
	if (size >= malloc_sizes[INDEX_L3 + 1].cs_size
2287 2288
	    && cachep->obj_size > cache_line_size() && size < PAGE_SIZE) {
		cachep->obj_offset += PAGE_SIZE - size;
L
Linus Torvalds 已提交
2289 2290 2291 2292 2293
		size = PAGE_SIZE;
	}
#endif
#endif

2294 2295 2296 2297 2298 2299
	/*
	 * Determine if the slab management is 'on' or 'off' slab.
	 * (bootstrapping cannot cope with offslab caches so don't do
	 * it too early on.)
	 */
	if ((size >= (PAGE_SIZE >> 3)) && !slab_early_init)
L
Linus Torvalds 已提交
2300 2301 2302 2303 2304 2305 2306 2307
		/*
		 * Size is large, assume best to place the slab management obj
		 * off-slab (should allow better packing of objs).
		 */
		flags |= CFLGS_OFF_SLAB;

	size = ALIGN(size, align);

2308
	left_over = calculate_slab_order(cachep, size, align, flags);
L
Linus Torvalds 已提交
2309 2310

	if (!cachep->num) {
2311 2312
		printk(KERN_ERR
		       "kmem_cache_create: couldn't create cache %s.\n", name);
L
Linus Torvalds 已提交
2313 2314
		kmem_cache_free(&cache_cache, cachep);
		cachep = NULL;
2315
		goto oops;
L
Linus Torvalds 已提交
2316
	}
P
Pekka Enberg 已提交
2317 2318
	slab_size = ALIGN(cachep->num * sizeof(kmem_bufctl_t)
			  + sizeof(struct slab), align);
L
Linus Torvalds 已提交
2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330

	/*
	 * If the slab has been placed off-slab, and we have enough space then
	 * move it on-slab. This is at the expense of any extra colouring.
	 */
	if (flags & CFLGS_OFF_SLAB && left_over >= slab_size) {
		flags &= ~CFLGS_OFF_SLAB;
		left_over -= slab_size;
	}

	if (flags & CFLGS_OFF_SLAB) {
		/* really off slab. No need for manual alignment */
P
Pekka Enberg 已提交
2331 2332
		slab_size =
		    cachep->num * sizeof(kmem_bufctl_t) + sizeof(struct slab);
L
Linus Torvalds 已提交
2333 2334 2335 2336 2337 2338
	}

	cachep->colour_off = cache_line_size();
	/* Offset must be a multiple of the alignment. */
	if (cachep->colour_off < align)
		cachep->colour_off = align;
P
Pekka Enberg 已提交
2339
	cachep->colour = left_over / cachep->colour_off;
L
Linus Torvalds 已提交
2340 2341 2342
	cachep->slab_size = slab_size;
	cachep->flags = flags;
	cachep->gfpflags = 0;
2343
	if (CONFIG_ZONE_DMA_FLAG && (flags & SLAB_CACHE_DMA))
L
Linus Torvalds 已提交
2344
		cachep->gfpflags |= GFP_DMA;
2345
	cachep->buffer_size = size;
2346
	cachep->reciprocal_buffer_size = reciprocal_value(size);
L
Linus Torvalds 已提交
2347

2348
	if (flags & CFLGS_OFF_SLAB) {
2349
		cachep->slabp_cache = kmem_find_general_cachep(slab_size, 0u);
2350 2351 2352 2353 2354 2355 2356
		/*
		 * This is a possibility for one of the malloc_sizes caches.
		 * But since we go off slab only for object size greater than
		 * PAGE_SIZE/8, and malloc_sizes gets created in ascending order,
		 * this should not happen at all.
		 * But leave a BUG_ON for some lucky dude.
		 */
2357
		BUG_ON(ZERO_OR_NULL_PTR(cachep->slabp_cache));
2358
	}
L
Linus Torvalds 已提交
2359 2360 2361
	cachep->ctor = ctor;
	cachep->name = name;

2362 2363 2364 2365 2366
	if (setup_cpu_cache(cachep)) {
		__kmem_cache_destroy(cachep);
		cachep = NULL;
		goto oops;
	}
L
Linus Torvalds 已提交
2367 2368 2369

	/* cache setup completed, link it into the list */
	list_add(&cachep->next, &cache_chain);
A
Andrew Morton 已提交
2370
oops:
L
Linus Torvalds 已提交
2371 2372
	if (!cachep && (flags & SLAB_PANIC))
		panic("kmem_cache_create(): failed to create slab `%s'\n",
P
Pekka Enberg 已提交
2373
		      name);
I
Ingo Molnar 已提交
2374
	mutex_unlock(&cache_chain_mutex);
L
Linus Torvalds 已提交
2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389
	return cachep;
}
EXPORT_SYMBOL(kmem_cache_create);

#if DEBUG
static void check_irq_off(void)
{
	BUG_ON(!irqs_disabled());
}

static void check_irq_on(void)
{
	BUG_ON(irqs_disabled());
}

2390
static void check_spinlock_acquired(struct kmem_cache *cachep)
L
Linus Torvalds 已提交
2391 2392 2393
{
#ifdef CONFIG_SMP
	check_irq_off();
2394
	assert_spin_locked(&cachep->nodelists[numa_node_id()]->list_lock);
L
Linus Torvalds 已提交
2395 2396
#endif
}
2397

2398
static void check_spinlock_acquired_node(struct kmem_cache *cachep, int node)
2399 2400 2401 2402 2403 2404 2405
{
#ifdef CONFIG_SMP
	check_irq_off();
	assert_spin_locked(&cachep->nodelists[node]->list_lock);
#endif
}

L
Linus Torvalds 已提交
2406 2407 2408 2409
#else
#define check_irq_off()	do { } while(0)
#define check_irq_on()	do { } while(0)
#define check_spinlock_acquired(x) do { } while(0)
2410
#define check_spinlock_acquired_node(x, y) do { } while(0)
L
Linus Torvalds 已提交
2411 2412
#endif

2413 2414 2415 2416
static void drain_array(struct kmem_cache *cachep, struct kmem_list3 *l3,
			struct array_cache *ac,
			int force, int node);

L
Linus Torvalds 已提交
2417 2418
static void do_drain(void *arg)
{
A
Andrew Morton 已提交
2419
	struct kmem_cache *cachep = arg;
L
Linus Torvalds 已提交
2420
	struct array_cache *ac;
2421
	int node = numa_node_id();
L
Linus Torvalds 已提交
2422 2423

	check_irq_off();
2424
	ac = cpu_cache_get(cachep);
2425 2426 2427
	spin_lock(&cachep->nodelists[node]->list_lock);
	free_block(cachep, ac->entry, ac->avail, node);
	spin_unlock(&cachep->nodelists[node]->list_lock);
L
Linus Torvalds 已提交
2428 2429 2430
	ac->avail = 0;
}

2431
static void drain_cpu_caches(struct kmem_cache *cachep)
L
Linus Torvalds 已提交
2432
{
2433 2434 2435
	struct kmem_list3 *l3;
	int node;

A
Andrew Morton 已提交
2436
	on_each_cpu(do_drain, cachep, 1, 1);
L
Linus Torvalds 已提交
2437
	check_irq_on();
P
Pekka Enberg 已提交
2438
	for_each_online_node(node) {
2439
		l3 = cachep->nodelists[node];
2440 2441 2442 2443 2444 2445 2446
		if (l3 && l3->alien)
			drain_alien_cache(cachep, l3->alien);
	}

	for_each_online_node(node) {
		l3 = cachep->nodelists[node];
		if (l3)
2447
			drain_array(cachep, l3, l3->shared, 1, node);
2448
	}
L
Linus Torvalds 已提交
2449 2450
}

2451 2452 2453 2454 2455 2456 2457 2458
/*
 * Remove slabs from the list of free slabs.
 * Specify the number of slabs to drain in tofree.
 *
 * Returns the actual number of slabs released.
 */
static int drain_freelist(struct kmem_cache *cache,
			struct kmem_list3 *l3, int tofree)
L
Linus Torvalds 已提交
2459
{
2460 2461
	struct list_head *p;
	int nr_freed;
L
Linus Torvalds 已提交
2462 2463
	struct slab *slabp;

2464 2465
	nr_freed = 0;
	while (nr_freed < tofree && !list_empty(&l3->slabs_free)) {
L
Linus Torvalds 已提交
2466

2467
		spin_lock_irq(&l3->list_lock);
2468
		p = l3->slabs_free.prev;
2469 2470 2471 2472
		if (p == &l3->slabs_free) {
			spin_unlock_irq(&l3->list_lock);
			goto out;
		}
L
Linus Torvalds 已提交
2473

2474
		slabp = list_entry(p, struct slab, list);
L
Linus Torvalds 已提交
2475
#if DEBUG
2476
		BUG_ON(slabp->inuse);
L
Linus Torvalds 已提交
2477 2478
#endif
		list_del(&slabp->list);
2479 2480 2481 2482 2483
		/*
		 * Safe to drop the lock. The slab is no longer linked
		 * to the cache.
		 */
		l3->free_objects -= cache->num;
2484
		spin_unlock_irq(&l3->list_lock);
2485 2486
		slab_destroy(cache, slabp);
		nr_freed++;
L
Linus Torvalds 已提交
2487
	}
2488 2489
out:
	return nr_freed;
L
Linus Torvalds 已提交
2490 2491
}

2492
/* Called with cache_chain_mutex held to protect against cpu hotplug */
2493
static int __cache_shrink(struct kmem_cache *cachep)
2494 2495 2496 2497 2498 2499 2500 2501 2502
{
	int ret = 0, i = 0;
	struct kmem_list3 *l3;

	drain_cpu_caches(cachep);

	check_irq_on();
	for_each_online_node(i) {
		l3 = cachep->nodelists[i];
2503 2504 2505 2506 2507 2508 2509
		if (!l3)
			continue;

		drain_freelist(cachep, l3, l3->free_objects);

		ret += !list_empty(&l3->slabs_full) ||
			!list_empty(&l3->slabs_partial);
2510 2511 2512 2513
	}
	return (ret ? 1 : 0);
}

L
Linus Torvalds 已提交
2514 2515 2516 2517 2518 2519 2520
/**
 * kmem_cache_shrink - Shrink a cache.
 * @cachep: The cache to shrink.
 *
 * Releases as many slabs as possible for a cache.
 * To help debugging, a zero exit status indicates all slabs were released.
 */
2521
int kmem_cache_shrink(struct kmem_cache *cachep)
L
Linus Torvalds 已提交
2522
{
2523
	int ret;
2524
	BUG_ON(!cachep || in_interrupt());
L
Linus Torvalds 已提交
2525

2526 2527 2528 2529
	mutex_lock(&cache_chain_mutex);
	ret = __cache_shrink(cachep);
	mutex_unlock(&cache_chain_mutex);
	return ret;
L
Linus Torvalds 已提交
2530 2531 2532 2533 2534 2535 2536
}
EXPORT_SYMBOL(kmem_cache_shrink);

/**
 * kmem_cache_destroy - delete a cache
 * @cachep: the cache to destroy
 *
2537
 * Remove a &struct kmem_cache object from the slab cache.
L
Linus Torvalds 已提交
2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548
 *
 * It is expected this function will be called by a module when it is
 * unloaded.  This will remove the cache completely, and avoid a duplicate
 * cache being allocated each time a module is loaded and unloaded, if the
 * module doesn't have persistent in-kernel storage across loads and unloads.
 *
 * The cache must be empty before calling this function.
 *
 * The caller must guarantee that noone will allocate memory from the cache
 * during the kmem_cache_destroy().
 */
2549
void kmem_cache_destroy(struct kmem_cache *cachep)
L
Linus Torvalds 已提交
2550
{
2551
	BUG_ON(!cachep || in_interrupt());
L
Linus Torvalds 已提交
2552 2553

	/* Find the cache in the chain of caches. */
I
Ingo Molnar 已提交
2554
	mutex_lock(&cache_chain_mutex);
L
Linus Torvalds 已提交
2555 2556 2557 2558 2559 2560
	/*
	 * the chain is never empty, cache_cache is never destroyed
	 */
	list_del(&cachep->next);
	if (__cache_shrink(cachep)) {
		slab_error(cachep, "Can't free all objects");
P
Pekka Enberg 已提交
2561
		list_add(&cachep->next, &cache_chain);
I
Ingo Molnar 已提交
2562
		mutex_unlock(&cache_chain_mutex);
2563
		return;
L
Linus Torvalds 已提交
2564 2565 2566
	}

	if (unlikely(cachep->flags & SLAB_DESTROY_BY_RCU))
2567
		synchronize_rcu();
L
Linus Torvalds 已提交
2568

2569
	__kmem_cache_destroy(cachep);
2570
	mutex_unlock(&cache_chain_mutex);
L
Linus Torvalds 已提交
2571 2572 2573
}
EXPORT_SYMBOL(kmem_cache_destroy);

2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584
/*
 * Get the memory for a slab management obj.
 * For a slab cache when the slab descriptor is off-slab, slab descriptors
 * always come from malloc_sizes caches.  The slab descriptor cannot
 * come from the same cache which is getting created because,
 * when we are searching for an appropriate cache for these
 * descriptors in kmem_cache_create, we search through the malloc_sizes array.
 * If we are creating a malloc_sizes cache here it would not be visible to
 * kmem_find_general_cachep till the initialization is complete.
 * Hence we cannot have slabp_cache same as the original cache.
 */
2585
static struct slab *alloc_slabmgmt(struct kmem_cache *cachep, void *objp,
2586 2587
				   int colour_off, gfp_t local_flags,
				   int nodeid)
L
Linus Torvalds 已提交
2588 2589
{
	struct slab *slabp;
P
Pekka Enberg 已提交
2590

L
Linus Torvalds 已提交
2591 2592
	if (OFF_SLAB(cachep)) {
		/* Slab management obj is off-slab. */
2593
		slabp = kmem_cache_alloc_node(cachep->slabp_cache,
2594
					      local_flags & ~GFP_THISNODE, nodeid);
L
Linus Torvalds 已提交
2595 2596 2597
		if (!slabp)
			return NULL;
	} else {
P
Pekka Enberg 已提交
2598
		slabp = objp + colour_off;
L
Linus Torvalds 已提交
2599 2600 2601 2602
		colour_off += cachep->slab_size;
	}
	slabp->inuse = 0;
	slabp->colouroff = colour_off;
P
Pekka Enberg 已提交
2603
	slabp->s_mem = objp + colour_off;
2604
	slabp->nodeid = nodeid;
L
Linus Torvalds 已提交
2605 2606 2607 2608 2609
	return slabp;
}

static inline kmem_bufctl_t *slab_bufctl(struct slab *slabp)
{
P
Pekka Enberg 已提交
2610
	return (kmem_bufctl_t *) (slabp + 1);
L
Linus Torvalds 已提交
2611 2612
}

2613
static void cache_init_objs(struct kmem_cache *cachep,
C
Christoph Lameter 已提交
2614
			    struct slab *slabp)
L
Linus Torvalds 已提交
2615 2616 2617 2618
{
	int i;

	for (i = 0; i < cachep->num; i++) {
2619
		void *objp = index_to_obj(cachep, slabp, i);
L
Linus Torvalds 已提交
2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631
#if DEBUG
		/* need to poison the objs? */
		if (cachep->flags & SLAB_POISON)
			poison_obj(cachep, objp, POISON_FREE);
		if (cachep->flags & SLAB_STORE_USER)
			*dbg_userword(cachep, objp) = NULL;

		if (cachep->flags & SLAB_RED_ZONE) {
			*dbg_redzone1(cachep, objp) = RED_INACTIVE;
			*dbg_redzone2(cachep, objp) = RED_INACTIVE;
		}
		/*
A
Andrew Morton 已提交
2632 2633 2634
		 * Constructors are not allowed to allocate memory from the same
		 * cache which they are a constructor for.  Otherwise, deadlock.
		 * They must also be threaded.
L
Linus Torvalds 已提交
2635 2636
		 */
		if (cachep->ctor && !(cachep->flags & SLAB_POISON))
2637
			cachep->ctor(objp + obj_offset(cachep), cachep,
C
Christoph Lameter 已提交
2638
				     0);
L
Linus Torvalds 已提交
2639 2640 2641 2642

		if (cachep->flags & SLAB_RED_ZONE) {
			if (*dbg_redzone2(cachep, objp) != RED_INACTIVE)
				slab_error(cachep, "constructor overwrote the"
P
Pekka Enberg 已提交
2643
					   " end of an object");
L
Linus Torvalds 已提交
2644 2645
			if (*dbg_redzone1(cachep, objp) != RED_INACTIVE)
				slab_error(cachep, "constructor overwrote the"
P
Pekka Enberg 已提交
2646
					   " start of an object");
L
Linus Torvalds 已提交
2647
		}
A
Andrew Morton 已提交
2648 2649
		if ((cachep->buffer_size % PAGE_SIZE) == 0 &&
			    OFF_SLAB(cachep) && cachep->flags & SLAB_POISON)
P
Pekka Enberg 已提交
2650
			kernel_map_pages(virt_to_page(objp),
2651
					 cachep->buffer_size / PAGE_SIZE, 0);
L
Linus Torvalds 已提交
2652 2653
#else
		if (cachep->ctor)
C
Christoph Lameter 已提交
2654
			cachep->ctor(objp, cachep, 0);
L
Linus Torvalds 已提交
2655
#endif
P
Pekka Enberg 已提交
2656
		slab_bufctl(slabp)[i] = i + 1;
L
Linus Torvalds 已提交
2657
	}
P
Pekka Enberg 已提交
2658
	slab_bufctl(slabp)[i - 1] = BUFCTL_END;
L
Linus Torvalds 已提交
2659 2660 2661
	slabp->free = 0;
}

2662
static void kmem_flagcheck(struct kmem_cache *cachep, gfp_t flags)
L
Linus Torvalds 已提交
2663
{
2664 2665 2666 2667 2668 2669
	if (CONFIG_ZONE_DMA_FLAG) {
		if (flags & GFP_DMA)
			BUG_ON(!(cachep->gfpflags & GFP_DMA));
		else
			BUG_ON(cachep->gfpflags & GFP_DMA);
	}
L
Linus Torvalds 已提交
2670 2671
}

A
Andrew Morton 已提交
2672 2673
static void *slab_get_obj(struct kmem_cache *cachep, struct slab *slabp,
				int nodeid)
2674
{
2675
	void *objp = index_to_obj(cachep, slabp, slabp->free);
2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688
	kmem_bufctl_t next;

	slabp->inuse++;
	next = slab_bufctl(slabp)[slabp->free];
#if DEBUG
	slab_bufctl(slabp)[slabp->free] = BUFCTL_FREE;
	WARN_ON(slabp->nodeid != nodeid);
#endif
	slabp->free = next;

	return objp;
}

A
Andrew Morton 已提交
2689 2690
static void slab_put_obj(struct kmem_cache *cachep, struct slab *slabp,
				void *objp, int nodeid)
2691
{
2692
	unsigned int objnr = obj_to_index(cachep, slabp, objp);
2693 2694 2695 2696 2697

#if DEBUG
	/* Verify that the slab belongs to the intended node */
	WARN_ON(slabp->nodeid != nodeid);

2698
	if (slab_bufctl(slabp)[objnr] + 1 <= SLAB_LIMIT + 1) {
2699
		printk(KERN_ERR "slab: double free detected in cache "
A
Andrew Morton 已提交
2700
				"'%s', objp %p\n", cachep->name, objp);
2701 2702 2703 2704 2705 2706 2707 2708
		BUG();
	}
#endif
	slab_bufctl(slabp)[objnr] = slabp->free;
	slabp->free = objnr;
	slabp->inuse--;
}

2709 2710 2711 2712 2713 2714 2715
/*
 * Map pages beginning at addr to the given cache and slab. This is required
 * for the slab allocator to be able to lookup the cache and slab of a
 * virtual address for kfree, ksize, kmem_ptr_validate, and slab debugging.
 */
static void slab_map_pages(struct kmem_cache *cache, struct slab *slab,
			   void *addr)
L
Linus Torvalds 已提交
2716
{
2717
	int nr_pages;
L
Linus Torvalds 已提交
2718 2719
	struct page *page;

2720
	page = virt_to_page(addr);
2721

2722
	nr_pages = 1;
2723
	if (likely(!PageCompound(page)))
2724 2725
		nr_pages <<= cache->gfporder;

L
Linus Torvalds 已提交
2726
	do {
2727 2728
		page_set_cache(page, cache);
		page_set_slab(page, slab);
L
Linus Torvalds 已提交
2729
		page++;
2730
	} while (--nr_pages);
L
Linus Torvalds 已提交
2731 2732 2733 2734 2735 2736
}

/*
 * Grow (by 1) the number of slabs within a cache.  This is called by
 * kmem_cache_alloc() when there are no active objs left in a cache.
 */
2737 2738
static int cache_grow(struct kmem_cache *cachep,
		gfp_t flags, int nodeid, void *objp)
L
Linus Torvalds 已提交
2739
{
P
Pekka Enberg 已提交
2740 2741 2742
	struct slab *slabp;
	size_t offset;
	gfp_t local_flags;
2743
	struct kmem_list3 *l3;
L
Linus Torvalds 已提交
2744

A
Andrew Morton 已提交
2745 2746 2747
	/*
	 * Be lazy and only check for valid flags here,  keeping it out of the
	 * critical path in kmem_cache_alloc().
L
Linus Torvalds 已提交
2748
	 */
2749
	BUG_ON(flags & ~(GFP_DMA | __GFP_ZERO | GFP_LEVEL_MASK));
L
Linus Torvalds 已提交
2750

2751
	local_flags = (flags & GFP_LEVEL_MASK);
2752
	/* Take the l3 list lock to change the colour_next on this node */
L
Linus Torvalds 已提交
2753
	check_irq_off();
2754 2755
	l3 = cachep->nodelists[nodeid];
	spin_lock(&l3->list_lock);
L
Linus Torvalds 已提交
2756 2757

	/* Get colour for the slab, and cal the next value. */
2758 2759 2760 2761 2762
	offset = l3->colour_next;
	l3->colour_next++;
	if (l3->colour_next >= cachep->colour)
		l3->colour_next = 0;
	spin_unlock(&l3->list_lock);
L
Linus Torvalds 已提交
2763

2764
	offset *= cachep->colour_off;
L
Linus Torvalds 已提交
2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776

	if (local_flags & __GFP_WAIT)
		local_irq_enable();

	/*
	 * The test for missing atomic flag is performed here, rather than
	 * the more obvious place, simply to reduce the critical path length
	 * in kmem_cache_alloc(). If a caller is seriously mis-behaving they
	 * will eventually be caught here (where it matters).
	 */
	kmem_flagcheck(cachep, flags);

A
Andrew Morton 已提交
2777 2778 2779
	/*
	 * Get mem for the objs.  Attempt to allocate a physical page from
	 * 'nodeid'.
2780
	 */
2781
	if (!objp)
2782
		objp = kmem_getpages(cachep, local_flags, nodeid);
A
Andrew Morton 已提交
2783
	if (!objp)
L
Linus Torvalds 已提交
2784 2785 2786
		goto failed;

	/* Get slab management. */
2787 2788
	slabp = alloc_slabmgmt(cachep, objp, offset,
			local_flags & ~GFP_THISNODE, nodeid);
A
Andrew Morton 已提交
2789
	if (!slabp)
L
Linus Torvalds 已提交
2790 2791
		goto opps1;

2792
	slabp->nodeid = nodeid;
2793
	slab_map_pages(cachep, slabp, objp);
L
Linus Torvalds 已提交
2794

C
Christoph Lameter 已提交
2795
	cache_init_objs(cachep, slabp);
L
Linus Torvalds 已提交
2796 2797 2798 2799

	if (local_flags & __GFP_WAIT)
		local_irq_disable();
	check_irq_off();
2800
	spin_lock(&l3->list_lock);
L
Linus Torvalds 已提交
2801 2802

	/* Make slab active. */
2803
	list_add_tail(&slabp->list, &(l3->slabs_free));
L
Linus Torvalds 已提交
2804
	STATS_INC_GROWN(cachep);
2805 2806
	l3->free_objects += cachep->num;
	spin_unlock(&l3->list_lock);
L
Linus Torvalds 已提交
2807
	return 1;
A
Andrew Morton 已提交
2808
opps1:
L
Linus Torvalds 已提交
2809
	kmem_freepages(cachep, objp);
A
Andrew Morton 已提交
2810
failed:
L
Linus Torvalds 已提交
2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826
	if (local_flags & __GFP_WAIT)
		local_irq_disable();
	return 0;
}

#if DEBUG

/*
 * Perform extra freeing checks:
 * - detect bad pointers.
 * - POISON/RED_ZONE checking
 */
static void kfree_debugcheck(const void *objp)
{
	if (!virt_addr_valid(objp)) {
		printk(KERN_ERR "kfree_debugcheck: out of range ptr %lxh.\n",
P
Pekka Enberg 已提交
2827 2828
		       (unsigned long)objp);
		BUG();
L
Linus Torvalds 已提交
2829 2830 2831
	}
}

2832 2833
static inline void verify_redzone_free(struct kmem_cache *cache, void *obj)
{
2834
	unsigned long long redzone1, redzone2;
2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849

	redzone1 = *dbg_redzone1(cache, obj);
	redzone2 = *dbg_redzone2(cache, obj);

	/*
	 * Redzone is ok.
	 */
	if (redzone1 == RED_ACTIVE && redzone2 == RED_ACTIVE)
		return;

	if (redzone1 == RED_INACTIVE && redzone2 == RED_INACTIVE)
		slab_error(cache, "double free detected");
	else
		slab_error(cache, "memory outside object was overwritten");

2850
	printk(KERN_ERR "%p: redzone 1:0x%llx, redzone 2:0x%llx.\n",
2851 2852 2853
			obj, redzone1, redzone2);
}

2854
static void *cache_free_debugcheck(struct kmem_cache *cachep, void *objp,
P
Pekka Enberg 已提交
2855
				   void *caller)
L
Linus Torvalds 已提交
2856 2857 2858 2859 2860
{
	struct page *page;
	unsigned int objnr;
	struct slab *slabp;

2861
	objp -= obj_offset(cachep);
L
Linus Torvalds 已提交
2862
	kfree_debugcheck(objp);
2863
	page = virt_to_head_page(objp);
L
Linus Torvalds 已提交
2864

2865
	slabp = page_get_slab(page);
L
Linus Torvalds 已提交
2866 2867

	if (cachep->flags & SLAB_RED_ZONE) {
2868
		verify_redzone_free(cachep, objp);
L
Linus Torvalds 已提交
2869 2870 2871 2872 2873 2874
		*dbg_redzone1(cachep, objp) = RED_INACTIVE;
		*dbg_redzone2(cachep, objp) = RED_INACTIVE;
	}
	if (cachep->flags & SLAB_STORE_USER)
		*dbg_userword(cachep, objp) = caller;

2875
	objnr = obj_to_index(cachep, slabp, objp);
L
Linus Torvalds 已提交
2876 2877

	BUG_ON(objnr >= cachep->num);
2878
	BUG_ON(objp != index_to_obj(cachep, slabp, objnr));
L
Linus Torvalds 已提交
2879

2880 2881 2882
#ifdef CONFIG_DEBUG_SLAB_LEAK
	slab_bufctl(slabp)[objnr] = BUFCTL_FREE;
#endif
L
Linus Torvalds 已提交
2883 2884
	if (cachep->flags & SLAB_POISON) {
#ifdef CONFIG_DEBUG_PAGEALLOC
A
Andrew Morton 已提交
2885
		if ((cachep->buffer_size % PAGE_SIZE)==0 && OFF_SLAB(cachep)) {
L
Linus Torvalds 已提交
2886
			store_stackinfo(cachep, objp, (unsigned long)caller);
P
Pekka Enberg 已提交
2887
			kernel_map_pages(virt_to_page(objp),
2888
					 cachep->buffer_size / PAGE_SIZE, 0);
L
Linus Torvalds 已提交
2889 2890 2891 2892 2893 2894 2895 2896 2897 2898
		} else {
			poison_obj(cachep, objp, POISON_FREE);
		}
#else
		poison_obj(cachep, objp, POISON_FREE);
#endif
	}
	return objp;
}

2899
static void check_slabp(struct kmem_cache *cachep, struct slab *slabp)
L
Linus Torvalds 已提交
2900 2901 2902
{
	kmem_bufctl_t i;
	int entries = 0;
P
Pekka Enberg 已提交
2903

L
Linus Torvalds 已提交
2904 2905 2906 2907 2908 2909 2910
	/* Check slab's freelist to see if this obj is there. */
	for (i = slabp->free; i != BUFCTL_END; i = slab_bufctl(slabp)[i]) {
		entries++;
		if (entries > cachep->num || i >= cachep->num)
			goto bad;
	}
	if (entries != cachep->num - slabp->inuse) {
A
Andrew Morton 已提交
2911 2912 2913 2914
bad:
		printk(KERN_ERR "slab: Internal list corruption detected in "
				"cache '%s'(%d), slabp %p(%d). Hexdump:\n",
			cachep->name, cachep->num, slabp, slabp->inuse);
P
Pekka Enberg 已提交
2915
		for (i = 0;
2916
		     i < sizeof(*slabp) + cachep->num * sizeof(kmem_bufctl_t);
P
Pekka Enberg 已提交
2917
		     i++) {
A
Andrew Morton 已提交
2918
			if (i % 16 == 0)
L
Linus Torvalds 已提交
2919
				printk("\n%03x:", i);
P
Pekka Enberg 已提交
2920
			printk(" %02x", ((unsigned char *)slabp)[i]);
L
Linus Torvalds 已提交
2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931
		}
		printk("\n");
		BUG();
	}
}
#else
#define kfree_debugcheck(x) do { } while(0)
#define cache_free_debugcheck(x,objp,z) (objp)
#define check_slabp(x,y) do { } while(0)
#endif

2932
static void *cache_alloc_refill(struct kmem_cache *cachep, gfp_t flags)
L
Linus Torvalds 已提交
2933 2934 2935 2936
{
	int batchcount;
	struct kmem_list3 *l3;
	struct array_cache *ac;
P
Pekka Enberg 已提交
2937 2938 2939
	int node;

	node = numa_node_id();
L
Linus Torvalds 已提交
2940 2941

	check_irq_off();
2942
	ac = cpu_cache_get(cachep);
A
Andrew Morton 已提交
2943
retry:
L
Linus Torvalds 已提交
2944 2945
	batchcount = ac->batchcount;
	if (!ac->touched && batchcount > BATCHREFILL_LIMIT) {
A
Andrew Morton 已提交
2946 2947 2948 2949
		/*
		 * If there was little recent activity on this cache, then
		 * perform only a partial refill.  Otherwise we could generate
		 * refill bouncing.
L
Linus Torvalds 已提交
2950 2951 2952
		 */
		batchcount = BATCHREFILL_LIMIT;
	}
P
Pekka Enberg 已提交
2953
	l3 = cachep->nodelists[node];
2954 2955 2956

	BUG_ON(ac->avail > 0 || !l3);
	spin_lock(&l3->list_lock);
L
Linus Torvalds 已提交
2957

2958 2959 2960 2961
	/* See if we can refill from the shared array */
	if (l3->shared && transfer_objects(ac, l3->shared, batchcount))
		goto alloc_done;

L
Linus Torvalds 已提交
2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976
	while (batchcount > 0) {
		struct list_head *entry;
		struct slab *slabp;
		/* Get slab alloc is to come from. */
		entry = l3->slabs_partial.next;
		if (entry == &l3->slabs_partial) {
			l3->free_touched = 1;
			entry = l3->slabs_free.next;
			if (entry == &l3->slabs_free)
				goto must_grow;
		}

		slabp = list_entry(entry, struct slab, list);
		check_slabp(cachep, slabp);
		check_spinlock_acquired(cachep);
2977 2978 2979 2980 2981 2982 2983 2984

		/*
		 * The slab was either on partial or free list so
		 * there must be at least one object available for
		 * allocation.
		 */
		BUG_ON(slabp->inuse < 0 || slabp->inuse >= cachep->num);

L
Linus Torvalds 已提交
2985 2986 2987 2988 2989
		while (slabp->inuse < cachep->num && batchcount--) {
			STATS_INC_ALLOCED(cachep);
			STATS_INC_ACTIVE(cachep);
			STATS_SET_HIGH(cachep);

2990
			ac->entry[ac->avail++] = slab_get_obj(cachep, slabp,
P
Pekka Enberg 已提交
2991
							    node);
L
Linus Torvalds 已提交
2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002
		}
		check_slabp(cachep, slabp);

		/* move slabp to correct slabp list: */
		list_del(&slabp->list);
		if (slabp->free == BUFCTL_END)
			list_add(&slabp->list, &l3->slabs_full);
		else
			list_add(&slabp->list, &l3->slabs_partial);
	}

A
Andrew Morton 已提交
3003
must_grow:
L
Linus Torvalds 已提交
3004
	l3->free_objects -= ac->avail;
A
Andrew Morton 已提交
3005
alloc_done:
3006
	spin_unlock(&l3->list_lock);
L
Linus Torvalds 已提交
3007 3008 3009

	if (unlikely(!ac->avail)) {
		int x;
3010
		x = cache_grow(cachep, flags | GFP_THISNODE, node, NULL);
3011

A
Andrew Morton 已提交
3012
		/* cache_grow can reenable interrupts, then ac could change. */
3013
		ac = cpu_cache_get(cachep);
A
Andrew Morton 已提交
3014
		if (!x && ac->avail == 0)	/* no objects in sight? abort */
L
Linus Torvalds 已提交
3015 3016
			return NULL;

A
Andrew Morton 已提交
3017
		if (!ac->avail)		/* objects refilled by interrupt? */
L
Linus Torvalds 已提交
3018 3019 3020
			goto retry;
	}
	ac->touched = 1;
3021
	return ac->entry[--ac->avail];
L
Linus Torvalds 已提交
3022 3023
}

A
Andrew Morton 已提交
3024 3025
static inline void cache_alloc_debugcheck_before(struct kmem_cache *cachep,
						gfp_t flags)
L
Linus Torvalds 已提交
3026 3027 3028 3029 3030 3031 3032 3033
{
	might_sleep_if(flags & __GFP_WAIT);
#if DEBUG
	kmem_flagcheck(cachep, flags);
#endif
}

#if DEBUG
A
Andrew Morton 已提交
3034 3035
static void *cache_alloc_debugcheck_after(struct kmem_cache *cachep,
				gfp_t flags, void *objp, void *caller)
L
Linus Torvalds 已提交
3036
{
P
Pekka Enberg 已提交
3037
	if (!objp)
L
Linus Torvalds 已提交
3038
		return objp;
P
Pekka Enberg 已提交
3039
	if (cachep->flags & SLAB_POISON) {
L
Linus Torvalds 已提交
3040
#ifdef CONFIG_DEBUG_PAGEALLOC
3041
		if ((cachep->buffer_size % PAGE_SIZE) == 0 && OFF_SLAB(cachep))
P
Pekka Enberg 已提交
3042
			kernel_map_pages(virt_to_page(objp),
3043
					 cachep->buffer_size / PAGE_SIZE, 1);
L
Linus Torvalds 已提交
3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054
		else
			check_poison_obj(cachep, objp);
#else
		check_poison_obj(cachep, objp);
#endif
		poison_obj(cachep, objp, POISON_INUSE);
	}
	if (cachep->flags & SLAB_STORE_USER)
		*dbg_userword(cachep, objp) = caller;

	if (cachep->flags & SLAB_RED_ZONE) {
A
Andrew Morton 已提交
3055 3056 3057 3058
		if (*dbg_redzone1(cachep, objp) != RED_INACTIVE ||
				*dbg_redzone2(cachep, objp) != RED_INACTIVE) {
			slab_error(cachep, "double free, or memory outside"
						" object was overwritten");
P
Pekka Enberg 已提交
3059
			printk(KERN_ERR
3060
				"%p: redzone 1:0x%llx, redzone 2:0x%llx\n",
A
Andrew Morton 已提交
3061 3062
				objp, *dbg_redzone1(cachep, objp),
				*dbg_redzone2(cachep, objp));
L
Linus Torvalds 已提交
3063 3064 3065 3066
		}
		*dbg_redzone1(cachep, objp) = RED_ACTIVE;
		*dbg_redzone2(cachep, objp) = RED_ACTIVE;
	}
3067 3068 3069 3070 3071
#ifdef CONFIG_DEBUG_SLAB_LEAK
	{
		struct slab *slabp;
		unsigned objnr;

3072
		slabp = page_get_slab(virt_to_head_page(objp));
3073 3074 3075 3076
		objnr = (unsigned)(objp - slabp->s_mem) / cachep->buffer_size;
		slab_bufctl(slabp)[objnr] = BUFCTL_ACTIVE;
	}
#endif
3077
	objp += obj_offset(cachep);
3078
	if (cachep->ctor && cachep->flags & SLAB_POISON)
C
Christoph Lameter 已提交
3079
		cachep->ctor(objp, cachep, 0);
3080 3081 3082 3083 3084 3085
#if ARCH_SLAB_MINALIGN
	if ((u32)objp & (ARCH_SLAB_MINALIGN-1)) {
		printk(KERN_ERR "0x%p: not aligned to ARCH_SLAB_MINALIGN=%d\n",
		       objp, ARCH_SLAB_MINALIGN);
	}
#endif
L
Linus Torvalds 已提交
3086 3087 3088 3089 3090 3091
	return objp;
}
#else
#define cache_alloc_debugcheck_after(a,b,objp,d) (objp)
#endif

3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104
#ifdef CONFIG_FAILSLAB

static struct failslab_attr {

	struct fault_attr attr;

	u32 ignore_gfp_wait;
#ifdef CONFIG_FAULT_INJECTION_DEBUG_FS
	struct dentry *ignore_gfp_wait_file;
#endif

} failslab = {
	.attr = FAULT_ATTR_INITIALIZER,
3105
	.ignore_gfp_wait = 1,
3106 3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133
};

static int __init setup_failslab(char *str)
{
	return setup_fault_attr(&failslab.attr, str);
}
__setup("failslab=", setup_failslab);

static int should_failslab(struct kmem_cache *cachep, gfp_t flags)
{
	if (cachep == &cache_cache)
		return 0;
	if (flags & __GFP_NOFAIL)
		return 0;
	if (failslab.ignore_gfp_wait && (flags & __GFP_WAIT))
		return 0;

	return should_fail(&failslab.attr, obj_size(cachep));
}

#ifdef CONFIG_FAULT_INJECTION_DEBUG_FS

static int __init failslab_debugfs(void)
{
	mode_t mode = S_IFREG | S_IRUSR | S_IWUSR;
	struct dentry *dir;
	int err;

3134
	err = init_fault_attr_dentries(&failslab.attr, "failslab");
3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164
	if (err)
		return err;
	dir = failslab.attr.dentries.dir;

	failslab.ignore_gfp_wait_file =
		debugfs_create_bool("ignore-gfp-wait", mode, dir,
				      &failslab.ignore_gfp_wait);

	if (!failslab.ignore_gfp_wait_file) {
		err = -ENOMEM;
		debugfs_remove(failslab.ignore_gfp_wait_file);
		cleanup_fault_attr_dentries(&failslab.attr);
	}

	return err;
}

late_initcall(failslab_debugfs);

#endif /* CONFIG_FAULT_INJECTION_DEBUG_FS */

#else /* CONFIG_FAILSLAB */

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

#endif /* CONFIG_FAILSLAB */

3165
static inline void *____cache_alloc(struct kmem_cache *cachep, gfp_t flags)
L
Linus Torvalds 已提交
3166
{
P
Pekka Enberg 已提交
3167
	void *objp;
L
Linus Torvalds 已提交
3168 3169
	struct array_cache *ac;

3170
	check_irq_off();
3171

3172
	ac = cpu_cache_get(cachep);
L
Linus Torvalds 已提交
3173 3174 3175
	if (likely(ac->avail)) {
		STATS_INC_ALLOCHIT(cachep);
		ac->touched = 1;
3176
		objp = ac->entry[--ac->avail];
L
Linus Torvalds 已提交
3177 3178 3179 3180
	} else {
		STATS_INC_ALLOCMISS(cachep);
		objp = cache_alloc_refill(cachep, flags);
	}
3181 3182 3183
	return objp;
}

3184
#ifdef CONFIG_NUMA
3185
/*
3186
 * Try allocating on another node if PF_SPREAD_SLAB|PF_MEMPOLICY.
3187 3188 3189 3190 3191 3192 3193 3194
 *
 * If we are in_interrupt, then process context, including cpusets and
 * mempolicy, may not apply and should not be used for allocation policy.
 */
static void *alternate_node_alloc(struct kmem_cache *cachep, gfp_t flags)
{
	int nid_alloc, nid_here;

3195
	if (in_interrupt() || (flags & __GFP_THISNODE))
3196 3197 3198 3199 3200 3201 3202
		return NULL;
	nid_alloc = nid_here = numa_node_id();
	if (cpuset_do_slab_mem_spread() && (cachep->flags & SLAB_MEM_SPREAD))
		nid_alloc = cpuset_mem_spread_node();
	else if (current->mempolicy)
		nid_alloc = slab_node(current->mempolicy);
	if (nid_alloc != nid_here)
3203
		return ____cache_alloc_node(cachep, flags, nid_alloc);
3204 3205 3206
	return NULL;
}

3207 3208
/*
 * Fallback function if there was no memory available and no objects on a
3209 3210 3211 3212 3213
 * certain node and fall back is permitted. First we scan all the
 * available nodelists for available objects. If that fails then we
 * perform an allocation without specifying a node. This allows the page
 * allocator to do its reclaim / fallback magic. We then insert the
 * slab into the proper nodelist and then allocate from it.
3214
 */
3215
static void *fallback_alloc(struct kmem_cache *cache, gfp_t flags)
3216
{
3217 3218
	struct zonelist *zonelist;
	gfp_t local_flags;
3219 3220
	struct zone **z;
	void *obj = NULL;
3221
	int nid;
3222 3223 3224 3225 3226 3227 3228

	if (flags & __GFP_THISNODE)
		return NULL;

	zonelist = &NODE_DATA(slab_node(current->mempolicy))
			->node_zonelists[gfp_zone(flags)];
	local_flags = (flags & GFP_LEVEL_MASK);
3229

3230 3231 3232 3233 3234
retry:
	/*
	 * Look through allowed nodes for objects available
	 * from existing per node queues.
	 */
3235
	for (z = zonelist->zones; *z && !obj; z++) {
3236
		nid = zone_to_nid(*z);
3237

3238
		if (cpuset_zone_allowed_hardwall(*z, flags) &&
3239 3240 3241 3242 3243 3244
			cache->nodelists[nid] &&
			cache->nodelists[nid]->free_objects)
				obj = ____cache_alloc_node(cache,
					flags | GFP_THISNODE, nid);
	}

3245
	if (!obj) {
3246 3247 3248 3249 3250 3251
		/*
		 * This allocation will be performed within the constraints
		 * of the current cpuset / memory policy requirements.
		 * We may trigger various forms of reclaim on the allowed
		 * set and go into memory reserves if necessary.
		 */
3252 3253 3254
		if (local_flags & __GFP_WAIT)
			local_irq_enable();
		kmem_flagcheck(cache, flags);
3255
		obj = kmem_getpages(cache, flags, -1);
3256 3257
		if (local_flags & __GFP_WAIT)
			local_irq_disable();
3258 3259 3260 3261 3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273
		if (obj) {
			/*
			 * Insert into the appropriate per node queues
			 */
			nid = page_to_nid(virt_to_page(obj));
			if (cache_grow(cache, flags, nid, obj)) {
				obj = ____cache_alloc_node(cache,
					flags | GFP_THISNODE, nid);
				if (!obj)
					/*
					 * Another processor may allocate the
					 * objects in the slab since we are
					 * not holding any locks.
					 */
					goto retry;
			} else {
3274
				/* cache_grow already freed obj */
3275 3276 3277
				obj = NULL;
			}
		}
3278
	}
3279 3280 3281
	return obj;
}

3282 3283
/*
 * A interface to enable slab creation on nodeid
L
Linus Torvalds 已提交
3284
 */
3285
static void *____cache_alloc_node(struct kmem_cache *cachep, gfp_t flags,
A
Andrew Morton 已提交
3286
				int nodeid)
3287 3288
{
	struct list_head *entry;
P
Pekka Enberg 已提交
3289 3290 3291 3292 3293 3294 3295 3296
	struct slab *slabp;
	struct kmem_list3 *l3;
	void *obj;
	int x;

	l3 = cachep->nodelists[nodeid];
	BUG_ON(!l3);

A
Andrew Morton 已提交
3297
retry:
3298
	check_irq_off();
P
Pekka Enberg 已提交
3299 3300 3301 3302 3303 3304 3305 3306 3307 3308 3309 3310 3311 3312 3313 3314 3315 3316 3317
	spin_lock(&l3->list_lock);
	entry = l3->slabs_partial.next;
	if (entry == &l3->slabs_partial) {
		l3->free_touched = 1;
		entry = l3->slabs_free.next;
		if (entry == &l3->slabs_free)
			goto must_grow;
	}

	slabp = list_entry(entry, struct slab, list);
	check_spinlock_acquired_node(cachep, nodeid);
	check_slabp(cachep, slabp);

	STATS_INC_NODEALLOCS(cachep);
	STATS_INC_ACTIVE(cachep);
	STATS_SET_HIGH(cachep);

	BUG_ON(slabp->inuse == cachep->num);

3318
	obj = slab_get_obj(cachep, slabp, nodeid);
P
Pekka Enberg 已提交
3319 3320 3321 3322 3323
	check_slabp(cachep, slabp);
	l3->free_objects--;
	/* move slabp to correct slabp list: */
	list_del(&slabp->list);

A
Andrew Morton 已提交
3324
	if (slabp->free == BUFCTL_END)
P
Pekka Enberg 已提交
3325
		list_add(&slabp->list, &l3->slabs_full);
A
Andrew Morton 已提交
3326
	else
P
Pekka Enberg 已提交
3327
		list_add(&slabp->list, &l3->slabs_partial);
3328

P
Pekka Enberg 已提交
3329 3330
	spin_unlock(&l3->list_lock);
	goto done;
3331

A
Andrew Morton 已提交
3332
must_grow:
P
Pekka Enberg 已提交
3333
	spin_unlock(&l3->list_lock);
3334
	x = cache_grow(cachep, flags | GFP_THISNODE, nodeid, NULL);
3335 3336
	if (x)
		goto retry;
L
Linus Torvalds 已提交
3337

3338
	return fallback_alloc(cachep, flags);
3339

A
Andrew Morton 已提交
3340
done:
P
Pekka Enberg 已提交
3341
	return obj;
3342
}
3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362

/**
 * kmem_cache_alloc_node - Allocate an object on the specified node
 * @cachep: The cache to allocate from.
 * @flags: See kmalloc().
 * @nodeid: node number of the target node.
 * @caller: return address of caller, used for debug information
 *
 * Identical to kmem_cache_alloc but it will allocate memory on the given
 * node, which can improve the performance for cpu bound structures.
 *
 * Fallback to other node is possible if __GFP_THISNODE is not set.
 */
static __always_inline void *
__cache_alloc_node(struct kmem_cache *cachep, gfp_t flags, int nodeid,
		   void *caller)
{
	unsigned long save_flags;
	void *ptr;

3363 3364 3365
	if (should_failslab(cachep, flags))
		return NULL;

3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377 3378 3379 3380 3381 3382 3383 3384 3385 3386 3387 3388 3389 3390 3391 3392 3393 3394
	cache_alloc_debugcheck_before(cachep, flags);
	local_irq_save(save_flags);

	if (unlikely(nodeid == -1))
		nodeid = numa_node_id();

	if (unlikely(!cachep->nodelists[nodeid])) {
		/* Node not bootstrapped yet */
		ptr = fallback_alloc(cachep, flags);
		goto out;
	}

	if (nodeid == numa_node_id()) {
		/*
		 * Use the locally cached objects if possible.
		 * However ____cache_alloc does not allow fallback
		 * to other nodes. It may fail while we still have
		 * objects on other nodes available.
		 */
		ptr = ____cache_alloc(cachep, flags);
		if (ptr)
			goto out;
	}
	/* ___cache_alloc_node can fall back to other nodes */
	ptr = ____cache_alloc_node(cachep, flags, nodeid);
  out:
	local_irq_restore(save_flags);
	ptr = cache_alloc_debugcheck_after(cachep, flags, ptr, caller);

3395 3396 3397
	if (unlikely((flags & __GFP_ZERO) && ptr))
		memset(ptr, 0, obj_size(cachep));

3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438
	return ptr;
}

static __always_inline void *
__do_cache_alloc(struct kmem_cache *cache, gfp_t flags)
{
	void *objp;

	if (unlikely(current->flags & (PF_SPREAD_SLAB | PF_MEMPOLICY))) {
		objp = alternate_node_alloc(cache, flags);
		if (objp)
			goto out;
	}
	objp = ____cache_alloc(cache, flags);

	/*
	 * We may just have run out of memory on the local node.
	 * ____cache_alloc_node() knows how to locate memory on other nodes
	 */
 	if (!objp)
 		objp = ____cache_alloc_node(cache, flags, numa_node_id());

  out:
	return objp;
}
#else

static __always_inline void *
__do_cache_alloc(struct kmem_cache *cachep, gfp_t flags)
{
	return ____cache_alloc(cachep, flags);
}

#endif /* CONFIG_NUMA */

static __always_inline void *
__cache_alloc(struct kmem_cache *cachep, gfp_t flags, void *caller)
{
	unsigned long save_flags;
	void *objp;

3439 3440 3441
	if (should_failslab(cachep, flags))
		return NULL;

3442 3443 3444 3445 3446 3447 3448
	cache_alloc_debugcheck_before(cachep, flags);
	local_irq_save(save_flags);
	objp = __do_cache_alloc(cachep, flags);
	local_irq_restore(save_flags);
	objp = cache_alloc_debugcheck_after(cachep, flags, objp, caller);
	prefetchw(objp);

3449 3450 3451
	if (unlikely((flags & __GFP_ZERO) && objp))
		memset(objp, 0, obj_size(cachep));

3452 3453
	return objp;
}
3454 3455 3456 3457

/*
 * Caller needs to acquire correct kmem_list's list_lock
 */
3458
static void free_block(struct kmem_cache *cachep, void **objpp, int nr_objects,
P
Pekka Enberg 已提交
3459
		       int node)
L
Linus Torvalds 已提交
3460 3461
{
	int i;
3462
	struct kmem_list3 *l3;
L
Linus Torvalds 已提交
3463 3464 3465 3466 3467

	for (i = 0; i < nr_objects; i++) {
		void *objp = objpp[i];
		struct slab *slabp;

3468
		slabp = virt_to_slab(objp);
3469
		l3 = cachep->nodelists[node];
L
Linus Torvalds 已提交
3470
		list_del(&slabp->list);
3471
		check_spinlock_acquired_node(cachep, node);
L
Linus Torvalds 已提交
3472
		check_slabp(cachep, slabp);
3473
		slab_put_obj(cachep, slabp, objp, node);
L
Linus Torvalds 已提交
3474
		STATS_DEC_ACTIVE(cachep);
3475
		l3->free_objects++;
L
Linus Torvalds 已提交
3476 3477 3478 3479
		check_slabp(cachep, slabp);

		/* fixup slab chains */
		if (slabp->inuse == 0) {
3480 3481
			if (l3->free_objects > l3->free_limit) {
				l3->free_objects -= cachep->num;
3482 3483 3484 3485 3486 3487
				/* No need to drop any previously held
				 * lock here, even if we have a off-slab slab
				 * descriptor it is guaranteed to come from
				 * a different cache, refer to comments before
				 * alloc_slabmgmt.
				 */
L
Linus Torvalds 已提交
3488 3489
				slab_destroy(cachep, slabp);
			} else {
3490
				list_add(&slabp->list, &l3->slabs_free);
L
Linus Torvalds 已提交
3491 3492 3493 3494 3495 3496
			}
		} else {
			/* Unconditionally move a slab to the end of the
			 * partial list on free - maximum time for the
			 * other objects to be freed, too.
			 */
3497
			list_add_tail(&slabp->list, &l3->slabs_partial);
L
Linus Torvalds 已提交
3498 3499 3500 3501
		}
	}
}

3502
static void cache_flusharray(struct kmem_cache *cachep, struct array_cache *ac)
L
Linus Torvalds 已提交
3503 3504
{
	int batchcount;
3505
	struct kmem_list3 *l3;
3506
	int node = numa_node_id();
L
Linus Torvalds 已提交
3507 3508 3509 3510 3511 3512

	batchcount = ac->batchcount;
#if DEBUG
	BUG_ON(!batchcount || batchcount > ac->avail);
#endif
	check_irq_off();
3513
	l3 = cachep->nodelists[node];
3514
	spin_lock(&l3->list_lock);
3515 3516
	if (l3->shared) {
		struct array_cache *shared_array = l3->shared;
P
Pekka Enberg 已提交
3517
		int max = shared_array->limit - shared_array->avail;
L
Linus Torvalds 已提交
3518 3519 3520
		if (max) {
			if (batchcount > max)
				batchcount = max;
3521
			memcpy(&(shared_array->entry[shared_array->avail]),
P
Pekka Enberg 已提交
3522
			       ac->entry, sizeof(void *) * batchcount);
L
Linus Torvalds 已提交
3523 3524 3525 3526 3527
			shared_array->avail += batchcount;
			goto free_done;
		}
	}

3528
	free_block(cachep, ac->entry, batchcount, node);
A
Andrew Morton 已提交
3529
free_done:
L
Linus Torvalds 已提交
3530 3531 3532 3533 3534
#if STATS
	{
		int i = 0;
		struct list_head *p;

3535 3536
		p = l3->slabs_free.next;
		while (p != &(l3->slabs_free)) {
L
Linus Torvalds 已提交
3537 3538 3539 3540 3541 3542 3543 3544 3545 3546 3547
			struct slab *slabp;

			slabp = list_entry(p, struct slab, list);
			BUG_ON(slabp->inuse);

			i++;
			p = p->next;
		}
		STATS_SET_FREEABLE(cachep, i);
	}
#endif
3548
	spin_unlock(&l3->list_lock);
L
Linus Torvalds 已提交
3549
	ac->avail -= batchcount;
A
Andrew Morton 已提交
3550
	memmove(ac->entry, &(ac->entry[batchcount]), sizeof(void *)*ac->avail);
L
Linus Torvalds 已提交
3551 3552 3553
}

/*
A
Andrew Morton 已提交
3554 3555
 * Release an obj back to its cache. If the obj has a constructed state, it must
 * be in this state _before_ it is released.  Called with disabled ints.
L
Linus Torvalds 已提交
3556
 */
3557
static inline void __cache_free(struct kmem_cache *cachep, void *objp)
L
Linus Torvalds 已提交
3558
{
3559
	struct array_cache *ac = cpu_cache_get(cachep);
L
Linus Torvalds 已提交
3560 3561 3562 3563

	check_irq_off();
	objp = cache_free_debugcheck(cachep, objp, __builtin_return_address(0));

3564 3565 3566 3567 3568 3569 3570 3571
	/*
	 * Skip calling cache_free_alien() when the platform is not numa.
	 * This will avoid cache misses that happen while accessing slabp (which
	 * is per page memory  reference) to get nodeid. Instead use a global
	 * variable to skip the call, which is mostly likely to be present in
	 * the cache.
	 */
	if (numa_platform && cache_free_alien(cachep, objp))
3572 3573
		return;

L
Linus Torvalds 已提交
3574 3575
	if (likely(ac->avail < ac->limit)) {
		STATS_INC_FREEHIT(cachep);
3576
		ac->entry[ac->avail++] = objp;
L
Linus Torvalds 已提交
3577 3578 3579 3580
		return;
	} else {
		STATS_INC_FREEMISS(cachep);
		cache_flusharray(cachep, ac);
3581
		ac->entry[ac->avail++] = objp;
L
Linus Torvalds 已提交
3582 3583 3584 3585 3586 3587 3588 3589 3590 3591 3592
	}
}

/**
 * kmem_cache_alloc - Allocate an object
 * @cachep: The cache to allocate from.
 * @flags: See kmalloc().
 *
 * Allocate an object from this cache.  The flags are only relevant
 * if the cache has no available objects.
 */
3593
void *kmem_cache_alloc(struct kmem_cache *cachep, gfp_t flags)
L
Linus Torvalds 已提交
3594
{
3595
	return __cache_alloc(cachep, flags, __builtin_return_address(0));
L
Linus Torvalds 已提交
3596 3597 3598 3599 3600 3601 3602 3603 3604 3605 3606 3607 3608 3609 3610 3611 3612
}
EXPORT_SYMBOL(kmem_cache_alloc);

/**
 * kmem_ptr_validate - check if an untrusted pointer might
 *	be a slab entry.
 * @cachep: the cache we're checking against
 * @ptr: pointer to validate
 *
 * This verifies that the untrusted pointer looks sane:
 * it is _not_ a guarantee that the pointer is actually
 * part of the slab cache in question, but it at least
 * validates that the pointer can be dereferenced and
 * looks half-way sane.
 *
 * Currently only used for dentry validation.
 */
3613
int kmem_ptr_validate(struct kmem_cache *cachep, const void *ptr)
L
Linus Torvalds 已提交
3614
{
P
Pekka Enberg 已提交
3615
	unsigned long addr = (unsigned long)ptr;
L
Linus Torvalds 已提交
3616
	unsigned long min_addr = PAGE_OFFSET;
P
Pekka Enberg 已提交
3617
	unsigned long align_mask = BYTES_PER_WORD - 1;
3618
	unsigned long size = cachep->buffer_size;
L
Linus Torvalds 已提交
3619 3620 3621 3622 3623 3624 3625 3626 3627 3628 3629 3630 3631 3632 3633
	struct page *page;

	if (unlikely(addr < min_addr))
		goto out;
	if (unlikely(addr > (unsigned long)high_memory - size))
		goto out;
	if (unlikely(addr & align_mask))
		goto out;
	if (unlikely(!kern_addr_valid(addr)))
		goto out;
	if (unlikely(!kern_addr_valid(addr + size - 1)))
		goto out;
	page = virt_to_page(ptr);
	if (unlikely(!PageSlab(page)))
		goto out;
3634
	if (unlikely(page_get_cache(page) != cachep))
L
Linus Torvalds 已提交
3635 3636
		goto out;
	return 1;
A
Andrew Morton 已提交
3637
out:
L
Linus Torvalds 已提交
3638 3639 3640 3641
	return 0;
}

#ifdef CONFIG_NUMA
3642 3643 3644 3645 3646
void *kmem_cache_alloc_node(struct kmem_cache *cachep, gfp_t flags, int nodeid)
{
	return __cache_alloc_node(cachep, flags, nodeid,
			__builtin_return_address(0));
}
L
Linus Torvalds 已提交
3647 3648
EXPORT_SYMBOL(kmem_cache_alloc_node);

3649 3650
static __always_inline void *
__do_kmalloc_node(size_t size, gfp_t flags, int node, void *caller)
3651
{
3652
	struct kmem_cache *cachep;
3653 3654

	cachep = kmem_find_general_cachep(size, flags);
3655 3656
	if (unlikely(ZERO_OR_NULL_PTR(cachep)))
		return cachep;
3657 3658
	return kmem_cache_alloc_node(cachep, flags, node);
}
3659 3660 3661 3662 3663 3664 3665

#ifdef CONFIG_DEBUG_SLAB
void *__kmalloc_node(size_t size, gfp_t flags, int node)
{
	return __do_kmalloc_node(size, flags, node,
			__builtin_return_address(0));
}
3666
EXPORT_SYMBOL(__kmalloc_node);
3667 3668 3669 3670 3671 3672 3673 3674 3675 3676 3677 3678 3679 3680 3681

void *__kmalloc_node_track_caller(size_t size, gfp_t flags,
		int node, void *caller)
{
	return __do_kmalloc_node(size, flags, node, caller);
}
EXPORT_SYMBOL(__kmalloc_node_track_caller);
#else
void *__kmalloc_node(size_t size, gfp_t flags, int node)
{
	return __do_kmalloc_node(size, flags, node, NULL);
}
EXPORT_SYMBOL(__kmalloc_node);
#endif /* CONFIG_DEBUG_SLAB */
#endif /* CONFIG_NUMA */
L
Linus Torvalds 已提交
3682 3683

/**
3684
 * __do_kmalloc - allocate memory
L
Linus Torvalds 已提交
3685
 * @size: how many bytes of memory are required.
3686
 * @flags: the type of memory to allocate (see kmalloc).
3687
 * @caller: function caller for debug tracking of the caller
L
Linus Torvalds 已提交
3688
 */
3689 3690
static __always_inline void *__do_kmalloc(size_t size, gfp_t flags,
					  void *caller)
L
Linus Torvalds 已提交
3691
{
3692
	struct kmem_cache *cachep;
L
Linus Torvalds 已提交
3693

3694 3695 3696 3697 3698 3699
	/* If you want to save a few bytes .text space: replace
	 * __ with kmem_.
	 * Then kmalloc uses the uninlined functions instead of the inline
	 * functions.
	 */
	cachep = __find_general_cachep(size, flags);
3700 3701
	if (unlikely(ZERO_OR_NULL_PTR(cachep)))
		return cachep;
3702 3703 3704 3705
	return __cache_alloc(cachep, flags, caller);
}


3706
#ifdef CONFIG_DEBUG_SLAB
3707 3708
void *__kmalloc(size_t size, gfp_t flags)
{
3709
	return __do_kmalloc(size, flags, __builtin_return_address(0));
L
Linus Torvalds 已提交
3710 3711 3712
}
EXPORT_SYMBOL(__kmalloc);

3713 3714 3715 3716 3717
void *__kmalloc_track_caller(size_t size, gfp_t flags, void *caller)
{
	return __do_kmalloc(size, flags, caller);
}
EXPORT_SYMBOL(__kmalloc_track_caller);
3718 3719 3720 3721 3722 3723 3724

#else
void *__kmalloc(size_t size, gfp_t flags)
{
	return __do_kmalloc(size, flags, NULL);
}
EXPORT_SYMBOL(__kmalloc);
3725 3726
#endif

L
Linus Torvalds 已提交
3727 3728 3729 3730 3731 3732 3733 3734
/**
 * kmem_cache_free - Deallocate an object
 * @cachep: The cache the allocation was from.
 * @objp: The previously allocated object.
 *
 * Free an object which was previously allocated from this
 * cache.
 */
3735
void kmem_cache_free(struct kmem_cache *cachep, void *objp)
L
Linus Torvalds 已提交
3736 3737 3738
{
	unsigned long flags;

3739 3740
	BUG_ON(virt_to_cache(objp) != cachep);

L
Linus Torvalds 已提交
3741
	local_irq_save(flags);
3742
	debug_check_no_locks_freed(objp, obj_size(cachep));
3743
	__cache_free(cachep, objp);
L
Linus Torvalds 已提交
3744 3745 3746 3747 3748 3749 3750 3751
	local_irq_restore(flags);
}
EXPORT_SYMBOL(kmem_cache_free);

/**
 * kfree - free previously allocated memory
 * @objp: pointer returned by kmalloc.
 *
3752 3753
 * If @objp is NULL, no operation is performed.
 *
L
Linus Torvalds 已提交
3754 3755 3756 3757 3758
 * Don't free memory not originally allocated by kmalloc()
 * or you will run into trouble.
 */
void kfree(const void *objp)
{
3759
	struct kmem_cache *c;
L
Linus Torvalds 已提交
3760 3761
	unsigned long flags;

3762
	if (unlikely(ZERO_OR_NULL_PTR(objp)))
L
Linus Torvalds 已提交
3763 3764 3765
		return;
	local_irq_save(flags);
	kfree_debugcheck(objp);
3766
	c = virt_to_cache(objp);
3767
	debug_check_no_locks_freed(objp, obj_size(c));
3768
	__cache_free(c, (void *)objp);
L
Linus Torvalds 已提交
3769 3770 3771 3772
	local_irq_restore(flags);
}
EXPORT_SYMBOL(kfree);

3773
unsigned int kmem_cache_size(struct kmem_cache *cachep)
L
Linus Torvalds 已提交
3774
{
3775
	return obj_size(cachep);
L
Linus Torvalds 已提交
3776 3777 3778
}
EXPORT_SYMBOL(kmem_cache_size);

3779
const char *kmem_cache_name(struct kmem_cache *cachep)
3780 3781 3782 3783 3784
{
	return cachep->name;
}
EXPORT_SYMBOL_GPL(kmem_cache_name);

3785
/*
3786
 * This initializes kmem_list3 or resizes varioius caches for all nodes.
3787
 */
3788
static int alloc_kmemlist(struct kmem_cache *cachep)
3789 3790 3791
{
	int node;
	struct kmem_list3 *l3;
3792
	struct array_cache *new_shared;
3793
	struct array_cache **new_alien = NULL;
3794 3795

	for_each_online_node(node) {
3796

3797 3798 3799 3800 3801
                if (use_alien_caches) {
                        new_alien = alloc_alien_cache(node, cachep->limit);
                        if (!new_alien)
                                goto fail;
                }
3802

3803 3804 3805
		new_shared = NULL;
		if (cachep->shared) {
			new_shared = alloc_arraycache(node,
3806
				cachep->shared*cachep->batchcount,
A
Andrew Morton 已提交
3807
					0xbaadf00d);
3808 3809 3810 3811
			if (!new_shared) {
				free_alien_cache(new_alien);
				goto fail;
			}
3812
		}
3813

A
Andrew Morton 已提交
3814 3815
		l3 = cachep->nodelists[node];
		if (l3) {
3816 3817
			struct array_cache *shared = l3->shared;

3818 3819
			spin_lock_irq(&l3->list_lock);

3820
			if (shared)
3821 3822
				free_block(cachep, shared->entry,
						shared->avail, node);
3823

3824 3825
			l3->shared = new_shared;
			if (!l3->alien) {
3826 3827 3828
				l3->alien = new_alien;
				new_alien = NULL;
			}
P
Pekka Enberg 已提交
3829
			l3->free_limit = (1 + nr_cpus_node(node)) *
A
Andrew Morton 已提交
3830
					cachep->batchcount + cachep->num;
3831
			spin_unlock_irq(&l3->list_lock);
3832
			kfree(shared);
3833 3834 3835
			free_alien_cache(new_alien);
			continue;
		}
A
Andrew Morton 已提交
3836
		l3 = kmalloc_node(sizeof(struct kmem_list3), GFP_KERNEL, node);
3837 3838 3839
		if (!l3) {
			free_alien_cache(new_alien);
			kfree(new_shared);
3840
			goto fail;
3841
		}
3842 3843 3844

		kmem_list3_init(l3);
		l3->next_reap = jiffies + REAPTIMEOUT_LIST3 +
A
Andrew Morton 已提交
3845
				((unsigned long)cachep) % REAPTIMEOUT_LIST3;
3846
		l3->shared = new_shared;
3847
		l3->alien = new_alien;
P
Pekka Enberg 已提交
3848
		l3->free_limit = (1 + nr_cpus_node(node)) *
A
Andrew Morton 已提交
3849
					cachep->batchcount + cachep->num;
3850 3851
		cachep->nodelists[node] = l3;
	}
3852
	return 0;
3853

A
Andrew Morton 已提交
3854
fail:
3855 3856 3857 3858 3859 3860 3861 3862 3863 3864 3865 3866 3867 3868 3869
	if (!cachep->next.next) {
		/* Cache is not active yet. Roll back what we did */
		node--;
		while (node >= 0) {
			if (cachep->nodelists[node]) {
				l3 = cachep->nodelists[node];

				kfree(l3->shared);
				free_alien_cache(l3->alien);
				kfree(l3);
				cachep->nodelists[node] = NULL;
			}
			node--;
		}
	}
3870
	return -ENOMEM;
3871 3872
}

L
Linus Torvalds 已提交
3873
struct ccupdate_struct {
3874
	struct kmem_cache *cachep;
L
Linus Torvalds 已提交
3875 3876 3877 3878 3879
	struct array_cache *new[NR_CPUS];
};

static void do_ccupdate_local(void *info)
{
A
Andrew Morton 已提交
3880
	struct ccupdate_struct *new = info;
L
Linus Torvalds 已提交
3881 3882 3883
	struct array_cache *old;

	check_irq_off();
3884
	old = cpu_cache_get(new->cachep);
3885

L
Linus Torvalds 已提交
3886 3887 3888 3889
	new->cachep->array[smp_processor_id()] = new->new[smp_processor_id()];
	new->new[smp_processor_id()] = old;
}

3890
/* Always called with the cache_chain_mutex held */
A
Andrew Morton 已提交
3891 3892
static int do_tune_cpucache(struct kmem_cache *cachep, int limit,
				int batchcount, int shared)
L
Linus Torvalds 已提交
3893
{
3894
	struct ccupdate_struct *new;
3895
	int i;
L
Linus Torvalds 已提交
3896

3897 3898 3899 3900
	new = kzalloc(sizeof(*new), GFP_KERNEL);
	if (!new)
		return -ENOMEM;

3901
	for_each_online_cpu(i) {
3902
		new->new[i] = alloc_arraycache(cpu_to_node(i), limit,
A
Andrew Morton 已提交
3903
						batchcount);
3904
		if (!new->new[i]) {
P
Pekka Enberg 已提交
3905
			for (i--; i >= 0; i--)
3906 3907
				kfree(new->new[i]);
			kfree(new);
3908
			return -ENOMEM;
L
Linus Torvalds 已提交
3909 3910
		}
	}
3911
	new->cachep = cachep;
L
Linus Torvalds 已提交
3912

3913
	on_each_cpu(do_ccupdate_local, (void *)new, 1, 1);
3914

L
Linus Torvalds 已提交
3915 3916 3917
	check_irq_on();
	cachep->batchcount = batchcount;
	cachep->limit = limit;
3918
	cachep->shared = shared;
L
Linus Torvalds 已提交
3919

3920
	for_each_online_cpu(i) {
3921
		struct array_cache *ccold = new->new[i];
L
Linus Torvalds 已提交
3922 3923
		if (!ccold)
			continue;
3924
		spin_lock_irq(&cachep->nodelists[cpu_to_node(i)]->list_lock);
3925
		free_block(cachep, ccold->entry, ccold->avail, cpu_to_node(i));
3926
		spin_unlock_irq(&cachep->nodelists[cpu_to_node(i)]->list_lock);
L
Linus Torvalds 已提交
3927 3928
		kfree(ccold);
	}
3929
	kfree(new);
3930
	return alloc_kmemlist(cachep);
L
Linus Torvalds 已提交
3931 3932
}

3933
/* Called with cache_chain_mutex held always */
3934
static int enable_cpucache(struct kmem_cache *cachep)
L
Linus Torvalds 已提交
3935 3936 3937 3938
{
	int err;
	int limit, shared;

A
Andrew Morton 已提交
3939 3940
	/*
	 * The head array serves three purposes:
L
Linus Torvalds 已提交
3941 3942
	 * - create a LIFO ordering, i.e. return objects that are cache-warm
	 * - reduce the number of spinlock operations.
A
Andrew Morton 已提交
3943
	 * - reduce the number of linked list operations on the slab and
L
Linus Torvalds 已提交
3944 3945 3946 3947
	 *   bufctl chains: array operations are cheaper.
	 * The numbers are guessed, we should auto-tune as described by
	 * Bonwick.
	 */
3948
	if (cachep->buffer_size > 131072)
L
Linus Torvalds 已提交
3949
		limit = 1;
3950
	else if (cachep->buffer_size > PAGE_SIZE)
L
Linus Torvalds 已提交
3951
		limit = 8;
3952
	else if (cachep->buffer_size > 1024)
L
Linus Torvalds 已提交
3953
		limit = 24;
3954
	else if (cachep->buffer_size > 256)
L
Linus Torvalds 已提交
3955 3956 3957 3958
		limit = 54;
	else
		limit = 120;

A
Andrew Morton 已提交
3959 3960
	/*
	 * CPU bound tasks (e.g. network routing) can exhibit cpu bound
L
Linus Torvalds 已提交
3961 3962 3963 3964 3965 3966 3967 3968
	 * allocation behaviour: Most allocs on one cpu, most free operations
	 * on another cpu. For these cases, an efficient object passing between
	 * cpus is necessary. This is provided by a shared array. The array
	 * replaces Bonwick's magazine layer.
	 * On uniprocessor, it's functionally equivalent (but less efficient)
	 * to a larger limit. Thus disabled by default.
	 */
	shared = 0;
3969
	if (cachep->buffer_size <= PAGE_SIZE && num_possible_cpus() > 1)
L
Linus Torvalds 已提交
3970 3971 3972
		shared = 8;

#if DEBUG
A
Andrew Morton 已提交
3973 3974 3975
	/*
	 * With debugging enabled, large batchcount lead to excessively long
	 * periods with disabled local interrupts. Limit the batchcount
L
Linus Torvalds 已提交
3976 3977 3978 3979
	 */
	if (limit > 32)
		limit = 32;
#endif
P
Pekka Enberg 已提交
3980
	err = do_tune_cpucache(cachep, limit, (limit + 1) / 2, shared);
L
Linus Torvalds 已提交
3981 3982
	if (err)
		printk(KERN_ERR "enable_cpucache failed for %s, error %d.\n",
P
Pekka Enberg 已提交
3983
		       cachep->name, -err);
3984
	return err;
L
Linus Torvalds 已提交
3985 3986
}

3987 3988
/*
 * Drain an array if it contains any elements taking the l3 lock only if
3989 3990
 * necessary. Note that the l3 listlock also protects the array_cache
 * if drain_array() is used on the shared array.
3991 3992 3993
 */
void drain_array(struct kmem_cache *cachep, struct kmem_list3 *l3,
			 struct array_cache *ac, int force, int node)
L
Linus Torvalds 已提交
3994 3995 3996
{
	int tofree;

3997 3998
	if (!ac || !ac->avail)
		return;
L
Linus Torvalds 已提交
3999 4000
	if (ac->touched && !force) {
		ac->touched = 0;
4001
	} else {
4002
		spin_lock_irq(&l3->list_lock);
4003 4004 4005 4006 4007 4008 4009 4010 4011
		if (ac->avail) {
			tofree = force ? ac->avail : (ac->limit + 4) / 5;
			if (tofree > ac->avail)
				tofree = (ac->avail + 1) / 2;
			free_block(cachep, ac->entry, tofree, node);
			ac->avail -= tofree;
			memmove(ac->entry, &(ac->entry[tofree]),
				sizeof(void *) * ac->avail);
		}
4012
		spin_unlock_irq(&l3->list_lock);
L
Linus Torvalds 已提交
4013 4014 4015 4016 4017
	}
}

/**
 * cache_reap - Reclaim memory from caches.
4018
 * @w: work descriptor
L
Linus Torvalds 已提交
4019 4020 4021 4022 4023 4024
 *
 * Called from workqueue/eventd every few seconds.
 * Purpose:
 * - clear the per-cpu caches for this CPU.
 * - return freeable pages to the main free memory pool.
 *
A
Andrew Morton 已提交
4025 4026
 * If we cannot acquire the cache chain mutex then just give up - we'll try
 * again on the next iteration.
L
Linus Torvalds 已提交
4027
 */
4028
static void cache_reap(struct work_struct *w)
L
Linus Torvalds 已提交
4029
{
4030
	struct kmem_cache *searchp;
4031
	struct kmem_list3 *l3;
4032
	int node = numa_node_id();
4033 4034
	struct delayed_work *work =
		container_of(w, struct delayed_work, work);
L
Linus Torvalds 已提交
4035

4036
	if (!mutex_trylock(&cache_chain_mutex))
L
Linus Torvalds 已提交
4037
		/* Give up. Setup the next iteration. */
4038
		goto out;
L
Linus Torvalds 已提交
4039

4040
	list_for_each_entry(searchp, &cache_chain, next) {
L
Linus Torvalds 已提交
4041 4042
		check_irq_on();

4043 4044 4045 4046 4047
		/*
		 * We only take the l3 lock if absolutely necessary and we
		 * have established with reasonable certainty that
		 * we can do some work if the lock was obtained.
		 */
4048
		l3 = searchp->nodelists[node];
4049

4050
		reap_alien(searchp, l3);
L
Linus Torvalds 已提交
4051

4052
		drain_array(searchp, l3, cpu_cache_get(searchp), 0, node);
L
Linus Torvalds 已提交
4053

4054 4055 4056 4057
		/*
		 * These are racy checks but it does not matter
		 * if we skip one check or scan twice.
		 */
4058
		if (time_after(l3->next_reap, jiffies))
4059
			goto next;
L
Linus Torvalds 已提交
4060

4061
		l3->next_reap = jiffies + REAPTIMEOUT_LIST3;
L
Linus Torvalds 已提交
4062

4063
		drain_array(searchp, l3, l3->shared, 0, node);
L
Linus Torvalds 已提交
4064

4065
		if (l3->free_touched)
4066
			l3->free_touched = 0;
4067 4068
		else {
			int freed;
L
Linus Torvalds 已提交
4069

4070 4071 4072 4073
			freed = drain_freelist(searchp, l3, (l3->free_limit +
				5 * searchp->num - 1) / (5 * searchp->num));
			STATS_ADD_REAPED(searchp, freed);
		}
4074
next:
L
Linus Torvalds 已提交
4075 4076 4077
		cond_resched();
	}
	check_irq_on();
I
Ingo Molnar 已提交
4078
	mutex_unlock(&cache_chain_mutex);
4079
	next_reap_node();
4080
out:
A
Andrew Morton 已提交
4081
	/* Set up the next iteration */
4082
	schedule_delayed_work(work, round_jiffies_relative(REAPTIMEOUT_CPUC));
L
Linus Torvalds 已提交
4083 4084 4085 4086
}

#ifdef CONFIG_PROC_FS

4087
static void print_slabinfo_header(struct seq_file *m)
L
Linus Torvalds 已提交
4088
{
4089 4090 4091 4092
	/*
	 * Output format version, so at least we can change it
	 * without _too_ many complaints.
	 */
L
Linus Torvalds 已提交
4093
#if STATS
4094
	seq_puts(m, "slabinfo - version: 2.1 (statistics)\n");
L
Linus Torvalds 已提交
4095
#else
4096
	seq_puts(m, "slabinfo - version: 2.1\n");
L
Linus Torvalds 已提交
4097
#endif
4098 4099 4100 4101
	seq_puts(m, "# name            <active_objs> <num_objs> <objsize> "
		 "<objperslab> <pagesperslab>");
	seq_puts(m, " : tunables <limit> <batchcount> <sharedfactor>");
	seq_puts(m, " : slabdata <active_slabs> <num_slabs> <sharedavail>");
L
Linus Torvalds 已提交
4102
#if STATS
4103
	seq_puts(m, " : globalstat <listallocs> <maxobjs> <grown> <reaped> "
4104
		 "<error> <maxfreeable> <nodeallocs> <remotefrees> <alienoverflow>");
4105
	seq_puts(m, " : cpustat <allochit> <allocmiss> <freehit> <freemiss>");
L
Linus Torvalds 已提交
4106
#endif
4107 4108 4109 4110 4111 4112 4113
	seq_putc(m, '\n');
}

static void *s_start(struct seq_file *m, loff_t *pos)
{
	loff_t n = *pos;

I
Ingo Molnar 已提交
4114
	mutex_lock(&cache_chain_mutex);
4115 4116
	if (!n)
		print_slabinfo_header(m);
4117 4118

	return seq_list_start(&cache_chain, *pos);
L
Linus Torvalds 已提交
4119 4120 4121 4122
}

static void *s_next(struct seq_file *m, void *p, loff_t *pos)
{
4123
	return seq_list_next(p, &cache_chain, pos);
L
Linus Torvalds 已提交
4124 4125 4126 4127
}

static void s_stop(struct seq_file *m, void *p)
{
I
Ingo Molnar 已提交
4128
	mutex_unlock(&cache_chain_mutex);
L
Linus Torvalds 已提交
4129 4130 4131 4132
}

static int s_show(struct seq_file *m, void *p)
{
4133
	struct kmem_cache *cachep = list_entry(p, struct kmem_cache, next);
P
Pekka Enberg 已提交
4134 4135 4136 4137 4138
	struct slab *slabp;
	unsigned long active_objs;
	unsigned long num_objs;
	unsigned long active_slabs = 0;
	unsigned long num_slabs, free_objects = 0, shared_avail = 0;
4139
	const char *name;
L
Linus Torvalds 已提交
4140
	char *error = NULL;
4141 4142
	int node;
	struct kmem_list3 *l3;
L
Linus Torvalds 已提交
4143 4144 4145

	active_objs = 0;
	num_slabs = 0;
4146 4147 4148 4149 4150
	for_each_online_node(node) {
		l3 = cachep->nodelists[node];
		if (!l3)
			continue;

4151 4152
		check_irq_on();
		spin_lock_irq(&l3->list_lock);
4153

4154
		list_for_each_entry(slabp, &l3->slabs_full, list) {
4155 4156 4157 4158 4159
			if (slabp->inuse != cachep->num && !error)
				error = "slabs_full accounting error";
			active_objs += cachep->num;
			active_slabs++;
		}
4160
		list_for_each_entry(slabp, &l3->slabs_partial, list) {
4161 4162 4163 4164 4165 4166 4167
			if (slabp->inuse == cachep->num && !error)
				error = "slabs_partial inuse accounting error";
			if (!slabp->inuse && !error)
				error = "slabs_partial/inuse accounting error";
			active_objs += slabp->inuse;
			active_slabs++;
		}
4168
		list_for_each_entry(slabp, &l3->slabs_free, list) {
4169 4170 4171 4172 4173
			if (slabp->inuse && !error)
				error = "slabs_free/inuse accounting error";
			num_slabs++;
		}
		free_objects += l3->free_objects;
4174 4175
		if (l3->shared)
			shared_avail += l3->shared->avail;
4176

4177
		spin_unlock_irq(&l3->list_lock);
L
Linus Torvalds 已提交
4178
	}
P
Pekka Enberg 已提交
4179 4180
	num_slabs += active_slabs;
	num_objs = num_slabs * cachep->num;
4181
	if (num_objs - active_objs != free_objects && !error)
L
Linus Torvalds 已提交
4182 4183
		error = "free_objects accounting error";

P
Pekka Enberg 已提交
4184
	name = cachep->name;
L
Linus Torvalds 已提交
4185 4186 4187 4188
	if (error)
		printk(KERN_ERR "slab: cache %s error: %s\n", name, error);

	seq_printf(m, "%-17s %6lu %6lu %6u %4u %4d",
4189
		   name, active_objs, num_objs, cachep->buffer_size,
P
Pekka Enberg 已提交
4190
		   cachep->num, (1 << cachep->gfporder));
L
Linus Torvalds 已提交
4191
	seq_printf(m, " : tunables %4u %4u %4u",
P
Pekka Enberg 已提交
4192
		   cachep->limit, cachep->batchcount, cachep->shared);
4193
	seq_printf(m, " : slabdata %6lu %6lu %6lu",
P
Pekka Enberg 已提交
4194
		   active_slabs, num_slabs, shared_avail);
L
Linus Torvalds 已提交
4195
#if STATS
P
Pekka Enberg 已提交
4196
	{			/* list3 stats */
L
Linus Torvalds 已提交
4197 4198 4199 4200 4201 4202 4203
		unsigned long high = cachep->high_mark;
		unsigned long allocs = cachep->num_allocations;
		unsigned long grown = cachep->grown;
		unsigned long reaped = cachep->reaped;
		unsigned long errors = cachep->errors;
		unsigned long max_freeable = cachep->max_freeable;
		unsigned long node_allocs = cachep->node_allocs;
4204
		unsigned long node_frees = cachep->node_frees;
4205
		unsigned long overflows = cachep->node_overflow;
L
Linus Torvalds 已提交
4206

4207
		seq_printf(m, " : globalstat %7lu %6lu %5lu %4lu \
4208
				%4lu %4lu %4lu %4lu %4lu", allocs, high, grown,
A
Andrew Morton 已提交
4209
				reaped, errors, max_freeable, node_allocs,
4210
				node_frees, overflows);
L
Linus Torvalds 已提交
4211 4212 4213 4214 4215 4216 4217 4218 4219
	}
	/* cpu stats */
	{
		unsigned long allochit = atomic_read(&cachep->allochit);
		unsigned long allocmiss = atomic_read(&cachep->allocmiss);
		unsigned long freehit = atomic_read(&cachep->freehit);
		unsigned long freemiss = atomic_read(&cachep->freemiss);

		seq_printf(m, " : cpustat %6lu %6lu %6lu %6lu",
P
Pekka Enberg 已提交
4220
			   allochit, allocmiss, freehit, freemiss);
L
Linus Torvalds 已提交
4221 4222 4223 4224 4225 4226 4227 4228 4229 4230 4231 4232 4233 4234 4235 4236 4237 4238 4239 4240
	}
#endif
	seq_putc(m, '\n');
	return 0;
}

/*
 * slabinfo_op - iterator that generates /proc/slabinfo
 *
 * Output layout:
 * cache-name
 * num-active-objs
 * total-objs
 * object size
 * num-active-slabs
 * total-slabs
 * num-pages-per-slab
 * + further values on SMP and with statistics enabled
 */

4241
const struct seq_operations slabinfo_op = {
P
Pekka Enberg 已提交
4242 4243 4244 4245
	.start = s_start,
	.next = s_next,
	.stop = s_stop,
	.show = s_show,
L
Linus Torvalds 已提交
4246 4247 4248 4249 4250 4251 4252 4253 4254 4255
};

#define MAX_SLABINFO_WRITE 128
/**
 * slabinfo_write - Tuning for the slab allocator
 * @file: unused
 * @buffer: user buffer
 * @count: data length
 * @ppos: unused
 */
P
Pekka Enberg 已提交
4256 4257
ssize_t slabinfo_write(struct file *file, const char __user * buffer,
		       size_t count, loff_t *ppos)
L
Linus Torvalds 已提交
4258
{
P
Pekka Enberg 已提交
4259
	char kbuf[MAX_SLABINFO_WRITE + 1], *tmp;
L
Linus Torvalds 已提交
4260
	int limit, batchcount, shared, res;
4261
	struct kmem_cache *cachep;
P
Pekka Enberg 已提交
4262

L
Linus Torvalds 已提交
4263 4264 4265 4266
	if (count > MAX_SLABINFO_WRITE)
		return -EINVAL;
	if (copy_from_user(&kbuf, buffer, count))
		return -EFAULT;
P
Pekka Enberg 已提交
4267
	kbuf[MAX_SLABINFO_WRITE] = '\0';
L
Linus Torvalds 已提交
4268 4269 4270 4271 4272 4273 4274 4275 4276 4277

	tmp = strchr(kbuf, ' ');
	if (!tmp)
		return -EINVAL;
	*tmp = '\0';
	tmp++;
	if (sscanf(tmp, " %d %d %d", &limit, &batchcount, &shared) != 3)
		return -EINVAL;

	/* Find the cache in the chain of caches. */
I
Ingo Molnar 已提交
4278
	mutex_lock(&cache_chain_mutex);
L
Linus Torvalds 已提交
4279
	res = -EINVAL;
4280
	list_for_each_entry(cachep, &cache_chain, next) {
L
Linus Torvalds 已提交
4281
		if (!strcmp(cachep->name, kbuf)) {
A
Andrew Morton 已提交
4282 4283
			if (limit < 1 || batchcount < 1 ||
					batchcount > limit || shared < 0) {
4284
				res = 0;
L
Linus Torvalds 已提交
4285
			} else {
4286
				res = do_tune_cpucache(cachep, limit,
P
Pekka Enberg 已提交
4287
						       batchcount, shared);
L
Linus Torvalds 已提交
4288 4289 4290 4291
			}
			break;
		}
	}
I
Ingo Molnar 已提交
4292
	mutex_unlock(&cache_chain_mutex);
L
Linus Torvalds 已提交
4293 4294 4295 4296
	if (res >= 0)
		res = count;
	return res;
}
4297 4298 4299 4300 4301 4302

#ifdef CONFIG_DEBUG_SLAB_LEAK

static void *leaks_start(struct seq_file *m, loff_t *pos)
{
	mutex_lock(&cache_chain_mutex);
4303
	return seq_list_start(&cache_chain, *pos);
4304 4305 4306 4307 4308 4309 4310 4311 4312 4313 4314 4315 4316 4317 4318 4319 4320 4321 4322 4323 4324 4325 4326 4327 4328 4329 4330 4331 4332 4333 4334 4335 4336 4337 4338 4339 4340 4341 4342 4343 4344 4345 4346 4347 4348 4349 4350 4351 4352 4353
}

static inline int add_caller(unsigned long *n, unsigned long v)
{
	unsigned long *p;
	int l;
	if (!v)
		return 1;
	l = n[1];
	p = n + 2;
	while (l) {
		int i = l/2;
		unsigned long *q = p + 2 * i;
		if (*q == v) {
			q[1]++;
			return 1;
		}
		if (*q > v) {
			l = i;
		} else {
			p = q + 2;
			l -= i + 1;
		}
	}
	if (++n[1] == n[0])
		return 0;
	memmove(p + 2, p, n[1] * 2 * sizeof(unsigned long) - ((void *)p - (void *)n));
	p[0] = v;
	p[1] = 1;
	return 1;
}

static void handle_slab(unsigned long *n, struct kmem_cache *c, struct slab *s)
{
	void *p;
	int i;
	if (n[0] == n[1])
		return;
	for (i = 0, p = s->s_mem; i < c->num; i++, p += c->buffer_size) {
		if (slab_bufctl(s)[i] != BUFCTL_ACTIVE)
			continue;
		if (!add_caller(n, (unsigned long)*dbg_userword(c, p)))
			return;
	}
}

static void show_symbol(struct seq_file *m, unsigned long address)
{
#ifdef CONFIG_KALLSYMS
	unsigned long offset, size;
4354
	char modname[MODULE_NAME_LEN], name[KSYM_NAME_LEN];
4355

4356
	if (lookup_symbol_attrs(address, &size, &offset, modname, name) == 0) {
4357
		seq_printf(m, "%s+%#lx/%#lx", name, offset, size);
4358
		if (modname[0])
4359 4360 4361 4362 4363 4364 4365 4366 4367
			seq_printf(m, " [%s]", modname);
		return;
	}
#endif
	seq_printf(m, "%p", (void *)address);
}

static int leaks_show(struct seq_file *m, void *p)
{
4368
	struct kmem_cache *cachep = list_entry(p, struct kmem_cache, next);
4369 4370 4371 4372 4373 4374 4375 4376 4377 4378 4379 4380 4381 4382 4383 4384 4385 4386 4387 4388 4389 4390 4391 4392
	struct slab *slabp;
	struct kmem_list3 *l3;
	const char *name;
	unsigned long *n = m->private;
	int node;
	int i;

	if (!(cachep->flags & SLAB_STORE_USER))
		return 0;
	if (!(cachep->flags & SLAB_RED_ZONE))
		return 0;

	/* OK, we can do it */

	n[1] = 0;

	for_each_online_node(node) {
		l3 = cachep->nodelists[node];
		if (!l3)
			continue;

		check_irq_on();
		spin_lock_irq(&l3->list_lock);

4393
		list_for_each_entry(slabp, &l3->slabs_full, list)
4394
			handle_slab(n, cachep, slabp);
4395
		list_for_each_entry(slabp, &l3->slabs_partial, list)
4396 4397 4398 4399 4400 4401 4402 4403 4404 4405 4406 4407 4408 4409 4410 4411 4412 4413 4414 4415 4416 4417 4418 4419 4420 4421
			handle_slab(n, cachep, slabp);
		spin_unlock_irq(&l3->list_lock);
	}
	name = cachep->name;
	if (n[0] == n[1]) {
		/* Increase the buffer size */
		mutex_unlock(&cache_chain_mutex);
		m->private = kzalloc(n[0] * 4 * sizeof(unsigned long), GFP_KERNEL);
		if (!m->private) {
			/* Too bad, we are really out */
			m->private = n;
			mutex_lock(&cache_chain_mutex);
			return -ENOMEM;
		}
		*(unsigned long *)m->private = n[0] * 2;
		kfree(n);
		mutex_lock(&cache_chain_mutex);
		/* Now make sure this entry will be retried */
		m->count = m->size;
		return 0;
	}
	for (i = 0; i < n[1]; i++) {
		seq_printf(m, "%s: %lu ", name, n[2*i+3]);
		show_symbol(m, n[2*i+2]);
		seq_putc(m, '\n');
	}
4422

4423 4424 4425
	return 0;
}

4426
const struct seq_operations slabstats_op = {
4427 4428 4429 4430 4431 4432
	.start = leaks_start,
	.next = s_next,
	.stop = s_stop,
	.show = leaks_show,
};
#endif
L
Linus Torvalds 已提交
4433 4434
#endif

4435 4436 4437 4438 4439 4440 4441 4442 4443 4444 4445 4446
/**
 * ksize - get the actual amount of memory allocated for a given object
 * @objp: Pointer to the object
 *
 * kmalloc may internally round up allocations and return more memory
 * than requested. ksize() can be used to determine the actual amount of
 * memory allocated. The caller may use this additional memory, even though
 * a smaller amount of memory was initially specified with the kmalloc call.
 * The caller must guarantee that objp points to a valid object previously
 * allocated with either kmalloc() or kmem_cache_alloc(). The object
 * must not be freed during the duration of the call.
 */
P
Pekka Enberg 已提交
4447
size_t ksize(const void *objp)
L
Linus Torvalds 已提交
4448
{
4449
	if (unlikely(ZERO_OR_NULL_PTR(objp)))
4450
		return 0;
L
Linus Torvalds 已提交
4451

4452
	return obj_size(virt_to_cache(objp));
L
Linus Torvalds 已提交
4453
}