slab_common.c 40.2 KB
Newer Older
1
// SPDX-License-Identifier: GPL-2.0
2 3 4 5 6 7 8 9 10 11 12
/*
 * Slab allocator functions that are independent of the allocator strategy
 *
 * (C) 2012 Christoph Lameter <cl@linux.com>
 */
#include <linux/slab.h>

#include <linux/mm.h>
#include <linux/poison.h>
#include <linux/interrupt.h>
#include <linux/memory.h>
13
#include <linux/cache.h>
14 15
#include <linux/compiler.h>
#include <linux/module.h>
16 17
#include <linux/cpu.h>
#include <linux/uaccess.h>
18 19
#include <linux/seq_file.h>
#include <linux/proc_fs.h>
20 21 22
#include <asm/cacheflush.h>
#include <asm/tlbflush.h>
#include <asm/page.h>
23
#include <linux/memcontrol.h>
24 25

#define CREATE_TRACE_POINTS
26
#include <trace/events/kmem.h>
27

28 29 30
#include "slab.h"

enum slab_state slab_state;
31 32
LIST_HEAD(slab_caches);
DEFINE_MUTEX(slab_mutex);
33
struct kmem_cache *kmem_cache;
34

35 36 37 38 39 40 41 42
#ifdef CONFIG_HARDENED_USERCOPY
bool usercopy_fallback __ro_after_init =
		IS_ENABLED(CONFIG_HARDENED_USERCOPY_FALLBACK);
module_param(usercopy_fallback, bool, 0400);
MODULE_PARM_DESC(usercopy_fallback,
		"WARN instead of reject usercopy whitelist violations");
#endif

43 44 45 46 47
static LIST_HEAD(slab_caches_to_rcu_destroy);
static void slab_caches_to_rcu_destroy_workfn(struct work_struct *work);
static DECLARE_WORK(slab_caches_to_rcu_destroy_work,
		    slab_caches_to_rcu_destroy_workfn);

48 49 50 51
/*
 * Set of flags that will prevent slab merging
 */
#define SLAB_NEVER_MERGE (SLAB_RED_ZONE | SLAB_POISON | SLAB_STORE_USER | \
52
		SLAB_TRACE | SLAB_TYPESAFE_BY_RCU | SLAB_NOLEAKTRACE | \
A
Alexander Potapenko 已提交
53
		SLAB_FAILSLAB | SLAB_KASAN)
54

V
Vladimir Davydov 已提交
55
#define SLAB_MERGE_SAME (SLAB_RECLAIM_ACCOUNT | SLAB_CACHE_DMA | \
56
			 SLAB_CACHE_DMA32 | SLAB_ACCOUNT)
57 58 59 60

/*
 * Merge control. If this is set then no merging of slab caches will occur.
 */
61
static bool slab_nomerge = !IS_ENABLED(CONFIG_SLAB_MERGE_DEFAULT);
62 63 64

static int __init setup_slab_nomerge(char *str)
{
65
	slab_nomerge = true;
66 67 68 69 70 71 72 73 74
	return 1;
}

#ifdef CONFIG_SLUB
__setup_param("slub_nomerge", slub_nomerge, setup_slab_nomerge, 0);
#endif

__setup("slab_nomerge", setup_slab_nomerge);

75 76 77 78 79 80 81 82 83
/*
 * Determine the size of a slab object
 */
unsigned int kmem_cache_size(struct kmem_cache *s)
{
	return s->object_size;
}
EXPORT_SYMBOL(kmem_cache_size);

84
#ifdef CONFIG_DEBUG_VM
85
static int kmem_cache_sanity_check(const char *name, unsigned int size)
86 87 88
{
	if (!name || in_interrupt() || size < sizeof(void *) ||
		size > KMALLOC_MAX_SIZE) {
89 90
		pr_err("kmem_cache_create(%s) integrity check failed\n", name);
		return -EINVAL;
91
	}
92

93
	WARN_ON(strchr(name, ' '));	/* It confuses parsers */
94 95 96
	return 0;
}
#else
97
static inline int kmem_cache_sanity_check(const char *name, unsigned int size)
98 99 100
{
	return 0;
}
101 102
#endif

103 104 105 106
void __kmem_cache_free_bulk(struct kmem_cache *s, size_t nr, void **p)
{
	size_t i;

107 108 109 110 111 112
	for (i = 0; i < nr; i++) {
		if (s)
			kmem_cache_free(s, p[i]);
		else
			kfree(p[i]);
	}
113 114
}

115
int __kmem_cache_alloc_bulk(struct kmem_cache *s, gfp_t flags, size_t nr,
116 117 118 119 120 121 122 123
								void **p)
{
	size_t i;

	for (i = 0; i < nr; i++) {
		void *x = p[i] = kmem_cache_alloc(s, flags);
		if (!x) {
			__kmem_cache_free_bulk(s, i, p);
124
			return 0;
125 126
		}
	}
127
	return i;
128 129
}

130
#ifdef CONFIG_MEMCG_KMEM
131 132 133

LIST_HEAD(slab_root_caches);

134
void slab_init_memcg_params(struct kmem_cache *s)
135
{
T
Tejun Heo 已提交
136
	s->memcg_params.root_cache = NULL;
137
	RCU_INIT_POINTER(s->memcg_params.memcg_caches, NULL);
T
Tejun Heo 已提交
138
	INIT_LIST_HEAD(&s->memcg_params.children);
139
	s->memcg_params.dying = false;
140 141 142
}

static int init_memcg_params(struct kmem_cache *s,
143
			     struct kmem_cache *root_cache)
144 145
{
	struct memcg_cache_array *arr;
146

T
Tejun Heo 已提交
147
	if (root_cache) {
148
		s->memcg_params.root_cache = root_cache;
T
Tejun Heo 已提交
149
		INIT_LIST_HEAD(&s->memcg_params.children_node);
150
		INIT_LIST_HEAD(&s->memcg_params.kmem_caches_node);
151
		return 0;
152
	}
153

154
	slab_init_memcg_params(s);
155

156 157
	if (!memcg_nr_cache_ids)
		return 0;
158

159 160 161
	arr = kvzalloc(sizeof(struct memcg_cache_array) +
		       memcg_nr_cache_ids * sizeof(void *),
		       GFP_KERNEL);
162 163
	if (!arr)
		return -ENOMEM;
164

165
	RCU_INIT_POINTER(s->memcg_params.memcg_caches, arr);
166 167 168
	return 0;
}

169
static void destroy_memcg_params(struct kmem_cache *s)
170
{
171
	if (is_root_cache(s))
172 173 174 175 176 177 178 179 180
		kvfree(rcu_access_pointer(s->memcg_params.memcg_caches));
}

static void free_memcg_params(struct rcu_head *rcu)
{
	struct memcg_cache_array *old;

	old = container_of(rcu, struct memcg_cache_array, rcu);
	kvfree(old);
181 182
}

183
static int update_memcg_params(struct kmem_cache *s, int new_array_size)
184
{
185
	struct memcg_cache_array *old, *new;
186

187 188
	new = kvzalloc(sizeof(struct memcg_cache_array) +
		       new_array_size * sizeof(void *), GFP_KERNEL);
189
	if (!new)
190 191
		return -ENOMEM;

192 193 194 195 196
	old = rcu_dereference_protected(s->memcg_params.memcg_caches,
					lockdep_is_held(&slab_mutex));
	if (old)
		memcpy(new->entries, old->entries,
		       memcg_nr_cache_ids * sizeof(void *));
197

198 199
	rcu_assign_pointer(s->memcg_params.memcg_caches, new);
	if (old)
200
		call_rcu(&old->rcu, free_memcg_params);
201 202 203
	return 0;
}

204 205 206 207 208
int memcg_update_all_caches(int num_memcgs)
{
	struct kmem_cache *s;
	int ret = 0;

209
	mutex_lock(&slab_mutex);
210
	list_for_each_entry(s, &slab_root_caches, root_caches_node) {
211
		ret = update_memcg_params(s, num_memcgs);
212 213 214 215 216
		/*
		 * Instead of freeing the memory, we'll just leave the caches
		 * up to this point in an updated state.
		 */
		if (ret)
217
			break;
218 219 220 221
	}
	mutex_unlock(&slab_mutex);
	return ret;
}
222

223
void memcg_link_cache(struct kmem_cache *s, struct mem_cgroup *memcg)
224
{
225 226 227
	if (is_root_cache(s)) {
		list_add(&s->root_caches_node, &slab_root_caches);
	} else {
228
		s->memcg_params.memcg = memcg;
229 230 231 232 233 234 235 236 237 238 239 240 241 242 243
		list_add(&s->memcg_params.children_node,
			 &s->memcg_params.root_cache->memcg_params.children);
		list_add(&s->memcg_params.kmem_caches_node,
			 &s->memcg_params.memcg->kmem_caches);
	}
}

static void memcg_unlink_cache(struct kmem_cache *s)
{
	if (is_root_cache(s)) {
		list_del(&s->root_caches_node);
	} else {
		list_del(&s->memcg_params.children_node);
		list_del(&s->memcg_params.kmem_caches_node);
	}
244
}
245
#else
246
static inline int init_memcg_params(struct kmem_cache *s,
247
				    struct kmem_cache *root_cache)
248 249 250 251
{
	return 0;
}

252
static inline void destroy_memcg_params(struct kmem_cache *s)
253 254
{
}
255

256
static inline void memcg_unlink_cache(struct kmem_cache *s)
257 258
{
}
259
#endif /* CONFIG_MEMCG_KMEM */
260

261 262 263 264
/*
 * Figure out what the alignment of the objects will be given a set of
 * flags, a user specified alignment and the size of the objects.
 */
265 266
static unsigned int calculate_alignment(slab_flags_t flags,
		unsigned int align, unsigned int size)
267 268 269 270 271 272 273 274 275
{
	/*
	 * If the user wants hardware cache aligned objects then follow that
	 * suggestion if the object is sufficiently large.
	 *
	 * The hardware cache alignment cannot override the specified
	 * alignment though. If that is greater then use it.
	 */
	if (flags & SLAB_HWCACHE_ALIGN) {
276
		unsigned int ralign;
277 278 279 280 281 282 283 284 285 286 287 288 289

		ralign = cache_line_size();
		while (size <= ralign / 2)
			ralign /= 2;
		align = max(align, ralign);
	}

	if (align < ARCH_SLAB_MINALIGN)
		align = ARCH_SLAB_MINALIGN;

	return ALIGN(align, sizeof(void *));
}

290 291 292 293 294 295 296 297 298 299 300 301 302 303
/*
 * Find a mergeable slab cache
 */
int slab_unmergeable(struct kmem_cache *s)
{
	if (slab_nomerge || (s->flags & SLAB_NEVER_MERGE))
		return 1;

	if (!is_root_cache(s))
		return 1;

	if (s->ctor)
		return 1;

304 305 306
	if (s->usersize)
		return 1;

307 308 309 310 311 312 313 314 315
	/*
	 * We may have set a slab to be unmergeable during bootstrap.
	 */
	if (s->refcount < 0)
		return 1;

	return 0;
}

316
struct kmem_cache *find_mergeable(unsigned int size, unsigned int align,
317
		slab_flags_t flags, const char *name, void (*ctor)(void *))
318 319 320
{
	struct kmem_cache *s;

321
	if (slab_nomerge)
322 323 324 325 326 327 328 329 330 331
		return NULL;

	if (ctor)
		return NULL;

	size = ALIGN(size, sizeof(void *));
	align = calculate_alignment(flags, align, size);
	size = ALIGN(size, align);
	flags = kmem_cache_flags(size, flags, name, NULL);

332 333 334
	if (flags & SLAB_NEVER_MERGE)
		return NULL;

335
	list_for_each_entry_reverse(s, &slab_root_caches, root_caches_node) {
336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353
		if (slab_unmergeable(s))
			continue;

		if (size > s->size)
			continue;

		if ((flags & SLAB_MERGE_SAME) != (s->flags & SLAB_MERGE_SAME))
			continue;
		/*
		 * Check if alignment is compatible.
		 * Courtesy of Adrian Drzewiecki
		 */
		if ((s->size & ~(align - 1)) != s->size)
			continue;

		if (s->size - size >= sizeof(void *))
			continue;

354 355 356 357
		if (IS_ENABLED(CONFIG_SLAB) && align &&
			(align > s->align || s->align % align))
			continue;

358 359 360 361 362
		return s;
	}
	return NULL;
}

363
static struct kmem_cache *create_cache(const char *name,
364
		unsigned int object_size, unsigned int align,
365 366
		slab_flags_t flags, unsigned int useroffset,
		unsigned int usersize, void (*ctor)(void *),
367
		struct mem_cgroup *memcg, struct kmem_cache *root_cache)
368 369 370 371
{
	struct kmem_cache *s;
	int err;

372 373 374
	if (WARN_ON(useroffset + usersize > object_size))
		useroffset = usersize = 0;

375 376 377 378 379 380
	err = -ENOMEM;
	s = kmem_cache_zalloc(kmem_cache, GFP_KERNEL);
	if (!s)
		goto out;

	s->name = name;
381
	s->size = s->object_size = object_size;
382 383
	s->align = align;
	s->ctor = ctor;
384 385
	s->useroffset = useroffset;
	s->usersize = usersize;
386

387
	err = init_memcg_params(s, root_cache);
388 389 390 391 392 393 394 395 396
	if (err)
		goto out_free_cache;

	err = __kmem_cache_create(s, flags);
	if (err)
		goto out_free_cache;

	s->refcount = 1;
	list_add(&s->list, &slab_caches);
397
	memcg_link_cache(s, memcg);
398 399 400 401 402 403
out:
	if (err)
		return ERR_PTR(err);
	return s;

out_free_cache:
404
	destroy_memcg_params(s);
405
	kmem_cache_free(kmem_cache, s);
406 407
	goto out;
}
408

409 410 411
/**
 * kmem_cache_create_usercopy - Create a cache with a region suitable
 * for copying to userspace
412 413 414 415
 * @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
416 417
 * @useroffset: Usercopy region offset
 * @usersize: Usercopy region size
418 419 420 421 422 423 424 425 426 427
 * @ctor: A constructor for the objects.
 *
 * Cannot be called within a interrupt, but can be interrupted.
 * The @ctor is run when new pages are allocated by the cache.
 *
 * The flags are
 *
 * %SLAB_POISON - Poison the slab with a known test pattern (a5a5a5a5)
 * to catch references to uninitialised memory.
 *
428
 * %SLAB_RED_ZONE - Insert `Red` zones around the allocated memory to check
429 430 431 432 433
 * 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.
434 435
 *
 * Return: a pointer to the cache on success, NULL on failure.
436
 */
437
struct kmem_cache *
438 439
kmem_cache_create_usercopy(const char *name,
		  unsigned int size, unsigned int align,
440 441
		  slab_flags_t flags,
		  unsigned int useroffset, unsigned int usersize,
442
		  void (*ctor)(void *))
443
{
444
	struct kmem_cache *s = NULL;
445
	const char *cache_name;
446
	int err;
447

448
	get_online_cpus();
449
	get_online_mems();
450
	memcg_get_cache_ids();
451

452
	mutex_lock(&slab_mutex);
453

454
	err = kmem_cache_sanity_check(name, size);
A
Andrew Morton 已提交
455
	if (err) {
456
		goto out_unlock;
A
Andrew Morton 已提交
457
	}
458

459 460 461 462 463 464
	/* Refuse requests with allocator specific flags */
	if (flags & ~SLAB_FLAGS_PERMITTED) {
		err = -EINVAL;
		goto out_unlock;
	}

465 466 467 468 469 470 471
	/*
	 * Some allocators will constraint the set of valid flags to a subset
	 * of all flags. We expect them to define CACHE_CREATE_MASK in this
	 * case, and we'll just provide them with a sanitized version of the
	 * passed flags.
	 */
	flags &= CACHE_CREATE_MASK;
472

473 474 475 476 477 478 479
	/* Fail closed on bad usersize of useroffset values. */
	if (WARN_ON(!usersize && useroffset) ||
	    WARN_ON(size < usersize || size - usersize < useroffset))
		usersize = useroffset = 0;

	if (!usersize)
		s = __kmem_cache_alias(name, size, align, flags, ctor);
480
	if (s)
481
		goto out_unlock;
482

483
	cache_name = kstrdup_const(name, GFP_KERNEL);
484 485 486 487
	if (!cache_name) {
		err = -ENOMEM;
		goto out_unlock;
	}
488

489
	s = create_cache(cache_name, size,
490
			 calculate_alignment(flags, align, size),
491
			 flags, useroffset, usersize, ctor, NULL, NULL);
492 493
	if (IS_ERR(s)) {
		err = PTR_ERR(s);
494
		kfree_const(cache_name);
495
	}
496 497

out_unlock:
498
	mutex_unlock(&slab_mutex);
499

500
	memcg_put_cache_ids();
501
	put_online_mems();
502 503
	put_online_cpus();

504
	if (err) {
505 506 507 508
		if (flags & SLAB_PANIC)
			panic("kmem_cache_create: Failed to create slab '%s'. Error %d\n",
				name, err);
		else {
509
			pr_warn("kmem_cache_create(%s) failed with error %d\n",
510 511 512 513 514
				name, err);
			dump_stack();
		}
		return NULL;
	}
515 516
	return s;
}
517 518
EXPORT_SYMBOL(kmem_cache_create_usercopy);

519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543
/**
 * 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.
 *
 * Cannot be called within a interrupt, but can be interrupted.
 * The @ctor is run when new pages are allocated by the cache.
 *
 * 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.
 *
 * Return: a pointer to the cache on success, NULL on failure.
 */
544
struct kmem_cache *
545
kmem_cache_create(const char *name, unsigned int size, unsigned int align,
546 547
		slab_flags_t flags, void (*ctor)(void *))
{
548
	return kmem_cache_create_usercopy(name, size, align, flags, 0, 0,
549 550
					  ctor);
}
551
EXPORT_SYMBOL(kmem_cache_create);
552

553
static void slab_caches_to_rcu_destroy_workfn(struct work_struct *work)
554
{
555 556
	LIST_HEAD(to_destroy);
	struct kmem_cache *s, *s2;
557

558
	/*
559
	 * On destruction, SLAB_TYPESAFE_BY_RCU kmem_caches are put on the
560 561 562 563 564 565 566 567 568 569
	 * @slab_caches_to_rcu_destroy list.  The slab pages are freed
	 * through RCU and and the associated kmem_cache are dereferenced
	 * while freeing the pages, so the kmem_caches should be freed only
	 * after the pending RCU operations are finished.  As rcu_barrier()
	 * is a pretty slow operation, we batch all pending destructions
	 * asynchronously.
	 */
	mutex_lock(&slab_mutex);
	list_splice_init(&slab_caches_to_rcu_destroy, &to_destroy);
	mutex_unlock(&slab_mutex);
570

571 572 573 574 575 576 577 578 579 580 581 582
	if (list_empty(&to_destroy))
		return;

	rcu_barrier();

	list_for_each_entry_safe(s, s2, &to_destroy, list) {
#ifdef SLAB_SUPPORTS_SYSFS
		sysfs_slab_release(s);
#else
		slab_kmem_cache_release(s);
#endif
	}
583 584
}

585
static int shutdown_cache(struct kmem_cache *s)
586
{
587 588 589
	/* free asan quarantined objects */
	kasan_cache_shutdown(s);

590 591
	if (__kmem_cache_shutdown(s) != 0)
		return -EBUSY;
592

593
	memcg_unlink_cache(s);
594
	list_del(&s->list);
595

596
	if (s->flags & SLAB_TYPESAFE_BY_RCU) {
597 598 599
#ifdef SLAB_SUPPORTS_SYSFS
		sysfs_slab_unlink(s);
#endif
600 601 602
		list_add_tail(&s->list, &slab_caches_to_rcu_destroy);
		schedule_work(&slab_caches_to_rcu_destroy_work);
	} else {
603
#ifdef SLAB_SUPPORTS_SYSFS
604
		sysfs_slab_unlink(s);
605
		sysfs_slab_release(s);
606 607 608 609
#else
		slab_kmem_cache_release(s);
#endif
	}
610 611

	return 0;
612 613
}

614
#ifdef CONFIG_MEMCG_KMEM
615
/*
616
 * memcg_create_kmem_cache - Create a cache for a memory cgroup.
617 618 619 620 621 622 623
 * @memcg: The memory cgroup the new cache is for.
 * @root_cache: The parent of the new cache.
 *
 * This function attempts to create a kmem cache that will serve allocation
 * requests going from @memcg to @root_cache. The new cache inherits properties
 * from its parent.
 */
624 625
void memcg_create_kmem_cache(struct mem_cgroup *memcg,
			     struct kmem_cache *root_cache)
626
{
627
	static char memcg_name_buf[NAME_MAX + 1]; /* protected by slab_mutex */
M
Michal Hocko 已提交
628
	struct cgroup_subsys_state *css = &memcg->css;
629
	struct memcg_cache_array *arr;
630
	struct kmem_cache *s = NULL;
631
	char *cache_name;
632
	int idx;
633 634

	get_online_cpus();
635 636
	get_online_mems();

637 638
	mutex_lock(&slab_mutex);

639
	/*
640
	 * The memory cgroup could have been offlined while the cache
641 642
	 * creation work was pending.
	 */
643
	if (memcg->kmem_state != KMEM_ONLINE || root_cache->memcg_params.dying)
644 645
		goto out_unlock;

646 647 648 649
	idx = memcg_cache_id(memcg);
	arr = rcu_dereference_protected(root_cache->memcg_params.memcg_caches,
					lockdep_is_held(&slab_mutex));

650 651 652 653 654
	/*
	 * Since per-memcg caches are created asynchronously on first
	 * allocation (see memcg_kmem_get_cache()), several threads can try to
	 * create the same cache, but only one of them may succeed.
	 */
655
	if (arr->entries[idx])
656 657
		goto out_unlock;

658
	cgroup_name(css->cgroup, memcg_name_buf, sizeof(memcg_name_buf));
659 660
	cache_name = kasprintf(GFP_KERNEL, "%s(%llu:%s)", root_cache->name,
			       css->serial_nr, memcg_name_buf);
661 662 663
	if (!cache_name)
		goto out_unlock;

664
	s = create_cache(cache_name, root_cache->object_size,
665
			 root_cache->align,
666
			 root_cache->flags & CACHE_CREATE_MASK,
667
			 root_cache->useroffset, root_cache->usersize,
668
			 root_cache->ctor, memcg, root_cache);
669 670 671 672 673
	/*
	 * If we could not create a memcg cache, do not complain, because
	 * that's not critical at all as we can always proceed with the root
	 * cache.
	 */
674
	if (IS_ERR(s)) {
675
		kfree(cache_name);
676
		goto out_unlock;
677
	}
678

679 680 681 682 683 684
	/*
	 * Since readers won't lock (see cache_from_memcg_idx()), we need a
	 * barrier here to ensure nobody will see the kmem_cache partially
	 * initialized.
	 */
	smp_wmb();
685
	arr->entries[idx] = s;
686

687 688
out_unlock:
	mutex_unlock(&slab_mutex);
689 690

	put_online_mems();
691
	put_online_cpus();
692
}
693

694
static void kmemcg_workfn(struct work_struct *work)
695 696
{
	struct kmem_cache *s = container_of(work, struct kmem_cache,
697
					    memcg_params.work);
698 699 700 701 702 703

	get_online_cpus();
	get_online_mems();

	mutex_lock(&slab_mutex);

704
	s->memcg_params.work_fn(s);
705 706 707 708 709 710

	mutex_unlock(&slab_mutex);

	put_online_mems();
	put_online_cpus();

711
	/* done, put the ref from kmemcg_cache_deactivate() */
712 713 714
	css_put(&s->memcg_params.memcg->css);
}

715
static void kmemcg_rcufn(struct rcu_head *head)
716 717
{
	struct kmem_cache *s = container_of(head, struct kmem_cache,
718
					    memcg_params.rcu_head);
719 720

	/*
721
	 * We need to grab blocking locks.  Bounce to ->work.  The
722 723 724
	 * work item shares the space with the RCU head and can't be
	 * initialized eariler.
	 */
725 726
	INIT_WORK(&s->memcg_params.work, kmemcg_workfn);
	queue_work(memcg_kmem_cache_wq, &s->memcg_params.work);
727 728
}

729
static void kmemcg_cache_deactivate(struct kmem_cache *s)
730 731
{
	if (WARN_ON_ONCE(is_root_cache(s)) ||
732
	    WARN_ON_ONCE(s->memcg_params.work_fn))
733 734
		return;

735 736
	__kmemcg_cache_deactivate(s);

737 738 739
	if (s->memcg_params.root_cache->memcg_params.dying)
		return;

740 741 742
	/* pin memcg so that @s doesn't get destroyed in the middle */
	css_get(&s->memcg_params.memcg->css);

743
	s->memcg_params.work_fn = __kmemcg_cache_deactivate_after_rcu;
744
	call_rcu(&s->memcg_params.rcu_head, kmemcg_rcufn);
745 746
}

747 748 749 750
void memcg_deactivate_kmem_caches(struct mem_cgroup *memcg)
{
	int idx;
	struct memcg_cache_array *arr;
751
	struct kmem_cache *s, *c;
752 753 754

	idx = memcg_cache_id(memcg);

755 756 757
	get_online_cpus();
	get_online_mems();

758
	mutex_lock(&slab_mutex);
759
	list_for_each_entry(s, &slab_root_caches, root_caches_node) {
760 761
		arr = rcu_dereference_protected(s->memcg_params.memcg_caches,
						lockdep_is_held(&slab_mutex));
762 763 764 765
		c = arr->entries[idx];
		if (!c)
			continue;

766
		kmemcg_cache_deactivate(c);
767 768 769
		arr->entries[idx] = NULL;
	}
	mutex_unlock(&slab_mutex);
770 771 772

	put_online_mems();
	put_online_cpus();
773 774
}

775
void memcg_destroy_kmem_caches(struct mem_cgroup *memcg)
776
{
777
	struct kmem_cache *s, *s2;
778

779 780
	get_online_cpus();
	get_online_mems();
781 782

	mutex_lock(&slab_mutex);
783 784
	list_for_each_entry_safe(s, s2, &memcg->kmem_caches,
				 memcg_params.kmem_caches_node) {
785 786 787 788
		/*
		 * The cgroup is about to be freed and therefore has no charges
		 * left. Hence, all its caches must be empty by now.
		 */
789
		BUG_ON(shutdown_cache(s));
790 791
	}
	mutex_unlock(&slab_mutex);
792

793 794
	put_online_mems();
	put_online_cpus();
795
}
796

797
static int shutdown_memcg_caches(struct kmem_cache *s)
798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815
{
	struct memcg_cache_array *arr;
	struct kmem_cache *c, *c2;
	LIST_HEAD(busy);
	int i;

	BUG_ON(!is_root_cache(s));

	/*
	 * First, shutdown active caches, i.e. caches that belong to online
	 * memory cgroups.
	 */
	arr = rcu_dereference_protected(s->memcg_params.memcg_caches,
					lockdep_is_held(&slab_mutex));
	for_each_memcg_cache_index(i) {
		c = arr->entries[i];
		if (!c)
			continue;
816
		if (shutdown_cache(c))
817 818 819 820 821
			/*
			 * The cache still has objects. Move it to a temporary
			 * list so as not to try to destroy it for a second
			 * time while iterating over inactive caches below.
			 */
T
Tejun Heo 已提交
822
			list_move(&c->memcg_params.children_node, &busy);
823 824 825 826 827 828 829 830 831 832 833 834 835 836
		else
			/*
			 * The cache is empty and will be destroyed soon. Clear
			 * the pointer to it in the memcg_caches array so that
			 * it will never be accessed even if the root cache
			 * stays alive.
			 */
			arr->entries[i] = NULL;
	}

	/*
	 * Second, shutdown all caches left from memory cgroups that are now
	 * offline.
	 */
T
Tejun Heo 已提交
837 838
	list_for_each_entry_safe(c, c2, &s->memcg_params.children,
				 memcg_params.children_node)
839
		shutdown_cache(c);
840

T
Tejun Heo 已提交
841
	list_splice(&busy, &s->memcg_params.children);
842 843 844 845 846

	/*
	 * A cache being destroyed must be empty. In particular, this means
	 * that all per memcg caches attached to it must be empty too.
	 */
T
Tejun Heo 已提交
847
	if (!list_empty(&s->memcg_params.children))
848 849 850
		return -EBUSY;
	return 0;
}
851 852 853 854 855 856 857 858

static void flush_memcg_workqueue(struct kmem_cache *s)
{
	mutex_lock(&slab_mutex);
	s->memcg_params.dying = true;
	mutex_unlock(&slab_mutex);

	/*
859
	 * SLAB and SLUB deactivate the kmem_caches through call_rcu. Make
860 861
	 * sure all registered rcu callbacks have been invoked.
	 */
862
	rcu_barrier();
863 864 865 866 867 868 869 870

	/*
	 * SLAB and SLUB create memcg kmem_caches through workqueue and SLUB
	 * deactivates the memcg kmem_caches through workqueue. Make sure all
	 * previous workitems on workqueue are processed.
	 */
	flush_workqueue(memcg_kmem_cache_wq);
}
871
#else
872
static inline int shutdown_memcg_caches(struct kmem_cache *s)
873 874 875
{
	return 0;
}
876 877 878 879

static inline void flush_memcg_workqueue(struct kmem_cache *s)
{
}
880
#endif /* CONFIG_MEMCG_KMEM */
881

882 883
void slab_kmem_cache_release(struct kmem_cache *s)
{
884
	__kmem_cache_release(s);
885
	destroy_memcg_params(s);
886
	kfree_const(s->name);
887 888 889
	kmem_cache_free(kmem_cache, s);
}

890 891
void kmem_cache_destroy(struct kmem_cache *s)
{
892
	int err;
893

894 895 896
	if (unlikely(!s))
		return;

897 898
	flush_memcg_workqueue(s);

899
	get_online_cpus();
900 901
	get_online_mems();

902
	mutex_lock(&slab_mutex);
903

904
	s->refcount--;
905 906 907
	if (s->refcount)
		goto out_unlock;

908
	err = shutdown_memcg_caches(s);
909
	if (!err)
910
		err = shutdown_cache(s);
911

912
	if (err) {
J
Joe Perches 已提交
913 914
		pr_err("kmem_cache_destroy %s: Slab cache still has objects\n",
		       s->name);
915 916
		dump_stack();
	}
917 918
out_unlock:
	mutex_unlock(&slab_mutex);
919

920
	put_online_mems();
921 922 923 924
	put_online_cpus();
}
EXPORT_SYMBOL(kmem_cache_destroy);

925 926 927 928 929 930
/**
 * 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.
931 932
 *
 * Return: %0 if all slabs were released, non-zero otherwise
933 934 935 936 937 938 939
 */
int kmem_cache_shrink(struct kmem_cache *cachep)
{
	int ret;

	get_online_cpus();
	get_online_mems();
940
	kasan_cache_shrink(cachep);
941
	ret = __kmem_cache_shrink(cachep);
942 943 944 945 946 947
	put_online_mems();
	put_online_cpus();
	return ret;
}
EXPORT_SYMBOL(kmem_cache_shrink);

948
bool slab_is_available(void)
949 950 951
{
	return slab_state >= UP;
}
952

953 954
#ifndef CONFIG_SLOB
/* Create a cache during boot when no slab services are available yet */
955 956 957
void __init create_boot_cache(struct kmem_cache *s, const char *name,
		unsigned int size, slab_flags_t flags,
		unsigned int useroffset, unsigned int usersize)
958 959 960 961 962
{
	int err;

	s->name = name;
	s->size = s->object_size = size;
963
	s->align = calculate_alignment(flags, ARCH_KMALLOC_MINALIGN, size);
964 965
	s->useroffset = useroffset;
	s->usersize = usersize;
966 967 968

	slab_init_memcg_params(s);

969 970 971
	err = __kmem_cache_create(s, flags);

	if (err)
972
		panic("Creation of kmalloc slab %s size=%u failed. Reason %d\n",
973 974 975 976 977
					name, size, err);

	s->refcount = -1;	/* Exempt from merging for now */
}

978 979 980
struct kmem_cache *__init create_kmalloc_cache(const char *name,
		unsigned int size, slab_flags_t flags,
		unsigned int useroffset, unsigned int usersize)
981 982 983 984 985 986
{
	struct kmem_cache *s = kmem_cache_zalloc(kmem_cache, GFP_NOWAIT);

	if (!s)
		panic("Out of memory when creating slab %s\n", name);

987
	create_boot_cache(s, name, size, flags, useroffset, usersize);
988
	list_add(&s->list, &slab_caches);
989
	memcg_link_cache(s, NULL);
990 991 992 993
	s->refcount = 1;
	return s;
}

994 995
struct kmem_cache *
kmalloc_caches[NR_KMALLOC_TYPES][KMALLOC_SHIFT_HIGH + 1] __ro_after_init;
996 997
EXPORT_SYMBOL(kmalloc_caches);

998 999 1000 1001 1002 1003
/*
 * Conversion table for small slabs sizes / 8 to the index in the
 * kmalloc array. This is necessary for slabs < 192 since we have non power
 * of two cache sizes there. The size of larger slabs can be determined using
 * fls.
 */
1004
static u8 size_index[24] __ro_after_init = {
1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030
	3,	/* 8 */
	4,	/* 16 */
	5,	/* 24 */
	5,	/* 32 */
	6,	/* 40 */
	6,	/* 48 */
	6,	/* 56 */
	6,	/* 64 */
	1,	/* 72 */
	1,	/* 80 */
	1,	/* 88 */
	1,	/* 96 */
	7,	/* 104 */
	7,	/* 112 */
	7,	/* 120 */
	7,	/* 128 */
	2,	/* 136 */
	2,	/* 144 */
	2,	/* 152 */
	2,	/* 160 */
	2,	/* 168 */
	2,	/* 176 */
	2,	/* 184 */
	2	/* 192 */
};

1031
static inline unsigned int size_index_elem(unsigned int bytes)
1032 1033 1034 1035 1036 1037 1038 1039 1040 1041
{
	return (bytes - 1) / 8;
}

/*
 * Find the kmem_cache structure that serves a given size of
 * allocation
 */
struct kmem_cache *kmalloc_slab(size_t size, gfp_t flags)
{
1042
	unsigned int index;
1043 1044 1045 1046 1047 1048

	if (size <= 192) {
		if (!size)
			return ZERO_SIZE_PTR;

		index = size_index[size_index_elem(size)];
1049
	} else {
1050
		if (WARN_ON_ONCE(size > KMALLOC_MAX_CACHE_SIZE))
1051
			return NULL;
1052
		index = fls(size - 1);
1053
	}
1054

1055
	return kmalloc_caches[kmalloc_type(flags)][index];
1056 1057
}

1058 1059 1060 1061 1062
/*
 * kmalloc_info[] is to make slub_debug=,kmalloc-xx option work at boot time.
 * kmalloc_index() supports up to 2^26=64MB, so the final entry of the table is
 * kmalloc-67108864.
 */
1063
const struct kmalloc_info_struct kmalloc_info[] __initconst = {
1064 1065 1066 1067 1068
	{NULL,                      0},		{"kmalloc-96",             96},
	{"kmalloc-192",           192},		{"kmalloc-8",               8},
	{"kmalloc-16",             16},		{"kmalloc-32",             32},
	{"kmalloc-64",             64},		{"kmalloc-128",           128},
	{"kmalloc-256",           256},		{"kmalloc-512",           512},
1069 1070 1071 1072 1073 1074 1075 1076 1077
	{"kmalloc-1k",           1024},		{"kmalloc-2k",           2048},
	{"kmalloc-4k",           4096},		{"kmalloc-8k",           8192},
	{"kmalloc-16k",         16384},		{"kmalloc-32k",         32768},
	{"kmalloc-64k",         65536},		{"kmalloc-128k",       131072},
	{"kmalloc-256k",       262144},		{"kmalloc-512k",       524288},
	{"kmalloc-1M",        1048576},		{"kmalloc-2M",        2097152},
	{"kmalloc-4M",        4194304},		{"kmalloc-8M",        8388608},
	{"kmalloc-16M",      16777216},		{"kmalloc-32M",      33554432},
	{"kmalloc-64M",      67108864}
1078 1079
};

1080
/*
1081 1082 1083 1084 1085 1086 1087 1088 1089
 * Patch up the size_index table if we have strange large alignment
 * requirements for the kmalloc array. This is only the case for
 * MIPS it seems. The standard arches will not generate any code here.
 *
 * Largest permitted alignment is 256 bytes due to the way we
 * handle the index determination for the smaller caches.
 *
 * Make sure that nothing crazy happens if someone starts tinkering
 * around with ARCH_KMALLOC_MINALIGN
1090
 */
1091
void __init setup_kmalloc_cache_index_table(void)
1092
{
1093
	unsigned int i;
1094

1095 1096 1097 1098
	BUILD_BUG_ON(KMALLOC_MIN_SIZE > 256 ||
		(KMALLOC_MIN_SIZE & (KMALLOC_MIN_SIZE - 1)));

	for (i = 8; i < KMALLOC_MIN_SIZE; i += 8) {
1099
		unsigned int elem = size_index_elem(i);
1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124

		if (elem >= ARRAY_SIZE(size_index))
			break;
		size_index[elem] = KMALLOC_SHIFT_LOW;
	}

	if (KMALLOC_MIN_SIZE >= 64) {
		/*
		 * The 96 byte size cache is not used if the alignment
		 * is 64 byte.
		 */
		for (i = 64 + 8; i <= 96; i += 8)
			size_index[size_index_elem(i)] = 7;

	}

	if (KMALLOC_MIN_SIZE >= 128) {
		/*
		 * The 192 byte sized cache is not used if the alignment
		 * is 128 byte. Redirect kmalloc to use the 256 byte cache
		 * instead.
		 */
		for (i = 128 + 8; i <= 192; i += 8)
			size_index[size_index_elem(i)] = 8;
	}
1125 1126
}

1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141
static const char *
kmalloc_cache_name(const char *prefix, unsigned int size)
{

	static const char units[3] = "\0kM";
	int idx = 0;

	while (size >= 1024 && (size % 1024 == 0)) {
		size /= 1024;
		idx++;
	}

	return kasprintf(GFP_NOWAIT, "%s-%u%c", prefix, size, units[idx]);
}

1142 1143
static void __init
new_kmalloc_cache(int idx, int type, slab_flags_t flags)
1144
{
1145 1146 1147 1148
	const char *name;

	if (type == KMALLOC_RECLAIM) {
		flags |= SLAB_RECLAIM_ACCOUNT;
1149
		name = kmalloc_cache_name("kmalloc-rcl",
1150 1151 1152 1153 1154 1155 1156
						kmalloc_info[idx].size);
		BUG_ON(!name);
	} else {
		name = kmalloc_info[idx].name;
	}

	kmalloc_caches[type][idx] = create_kmalloc_cache(name,
1157 1158
					kmalloc_info[idx].size, flags, 0,
					kmalloc_info[idx].size);
1159 1160
}

1161 1162 1163 1164 1165
/*
 * Create the kmalloc array. Some of the regular kmalloc arrays
 * may already have been created because they were needed to
 * enable allocations for slab creation.
 */
1166
void __init create_kmalloc_caches(slab_flags_t flags)
1167
{
1168
	int i, type;
1169

1170 1171 1172 1173
	for (type = KMALLOC_NORMAL; type <= KMALLOC_RECLAIM; type++) {
		for (i = KMALLOC_SHIFT_LOW; i <= KMALLOC_SHIFT_HIGH; i++) {
			if (!kmalloc_caches[type][i])
				new_kmalloc_cache(i, type, flags);
1174

1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186
			/*
			 * Caches that are not of the two-to-the-power-of size.
			 * These have to be created immediately after the
			 * earlier power of two caches
			 */
			if (KMALLOC_MIN_SIZE <= 32 && i == 6 &&
					!kmalloc_caches[type][1])
				new_kmalloc_cache(1, type, flags);
			if (KMALLOC_MIN_SIZE <= 64 && i == 7 &&
					!kmalloc_caches[type][2])
				new_kmalloc_cache(2, type, flags);
		}
1187 1188
	}

1189 1190 1191 1192 1193
	/* Kmalloc array is now usable */
	slab_state = UP;

#ifdef CONFIG_ZONE_DMA
	for (i = 0; i <= KMALLOC_SHIFT_HIGH; i++) {
1194
		struct kmem_cache *s = kmalloc_caches[KMALLOC_NORMAL][i];
1195 1196

		if (s) {
1197
			unsigned int size = kmalloc_size(i);
1198
			const char *n = kmalloc_cache_name("dma-kmalloc", size);
1199 1200

			BUG_ON(!n);
1201 1202
			kmalloc_caches[KMALLOC_DMA][i] = create_kmalloc_cache(
				n, size, SLAB_CACHE_DMA | flags, 0, 0);
1203 1204 1205 1206
		}
	}
#endif
}
1207 1208
#endif /* !CONFIG_SLOB */

V
Vladimir Davydov 已提交
1209 1210 1211 1212 1213
/*
 * To avoid unnecessary overhead, we pass through large allocation requests
 * directly to the page allocator. We use __GFP_COMP, because we will need to
 * know the allocation order to free the pages properly in kfree.
 */
V
Vladimir Davydov 已提交
1214 1215 1216 1217 1218 1219
void *kmalloc_order(size_t size, gfp_t flags, unsigned int order)
{
	void *ret;
	struct page *page;

	flags |= __GFP_COMP;
1220
	page = alloc_pages(flags, order);
V
Vladimir Davydov 已提交
1221
	ret = page ? page_address(page) : NULL;
1222
	ret = kasan_kmalloc_large(ret, size, flags);
1223
	/* As ret might get tagged, call kmemleak hook after KASAN. */
1224
	kmemleak_alloc(ret, size, 1, flags);
V
Vladimir Davydov 已提交
1225 1226 1227 1228
	return ret;
}
EXPORT_SYMBOL(kmalloc_order);

1229 1230 1231 1232 1233 1234 1235 1236 1237
#ifdef CONFIG_TRACING
void *kmalloc_order_trace(size_t size, gfp_t flags, unsigned int order)
{
	void *ret = kmalloc_order(size, flags, order);
	trace_kmalloc(_RET_IP_, ret, size, PAGE_SIZE << order, flags);
	return ret;
}
EXPORT_SYMBOL(kmalloc_order_trace);
#endif
1238

1239 1240 1241
#ifdef CONFIG_SLAB_FREELIST_RANDOM
/* Randomize a generic freelist */
static void freelist_randomize(struct rnd_state *state, unsigned int *list,
1242
			       unsigned int count)
1243 1244
{
	unsigned int rand;
1245
	unsigned int i;
1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285

	for (i = 0; i < count; i++)
		list[i] = i;

	/* Fisher-Yates shuffle */
	for (i = count - 1; i > 0; i--) {
		rand = prandom_u32_state(state);
		rand %= (i + 1);
		swap(list[i], list[rand]);
	}
}

/* Create a random sequence per cache */
int cache_random_seq_create(struct kmem_cache *cachep, unsigned int count,
				    gfp_t gfp)
{
	struct rnd_state state;

	if (count < 2 || cachep->random_seq)
		return 0;

	cachep->random_seq = kcalloc(count, sizeof(unsigned int), gfp);
	if (!cachep->random_seq)
		return -ENOMEM;

	/* Get best entropy at this stage of boot */
	prandom_seed_state(&state, get_random_long());

	freelist_randomize(&state, cachep->random_seq, count);
	return 0;
}

/* Destroy the per-cache random freelist sequence */
void cache_random_seq_destroy(struct kmem_cache *cachep)
{
	kfree(cachep->random_seq);
	cachep->random_seq = NULL;
}
#endif /* CONFIG_SLAB_FREELIST_RANDOM */

Y
Yang Shi 已提交
1286
#if defined(CONFIG_SLAB) || defined(CONFIG_SLUB_DEBUG)
1287
#ifdef CONFIG_SLAB
1288
#define SLABINFO_RIGHTS (0600)
1289
#else
1290
#define SLABINFO_RIGHTS (0400)
1291 1292
#endif

1293
static void print_slabinfo_header(struct seq_file *m)
1294 1295 1296 1297 1298 1299 1300 1301 1302 1303
{
	/*
	 * Output format version, so at least we can change it
	 * without _too_ many complaints.
	 */
#ifdef CONFIG_DEBUG_SLAB
	seq_puts(m, "slabinfo - version: 2.1 (statistics)\n");
#else
	seq_puts(m, "slabinfo - version: 2.1\n");
#endif
J
Joe Perches 已提交
1304
	seq_puts(m, "# name            <active_objs> <num_objs> <objsize> <objperslab> <pagesperslab>");
1305 1306 1307
	seq_puts(m, " : tunables <limit> <batchcount> <sharedfactor>");
	seq_puts(m, " : slabdata <active_slabs> <num_slabs> <sharedavail>");
#ifdef CONFIG_DEBUG_SLAB
J
Joe Perches 已提交
1308
	seq_puts(m, " : globalstat <listallocs> <maxobjs> <grown> <reaped> <error> <maxfreeable> <nodeallocs> <remotefrees> <alienoverflow>");
1309 1310 1311 1312 1313
	seq_puts(m, " : cpustat <allochit> <allocmiss> <freehit> <freemiss>");
#endif
	seq_putc(m, '\n');
}

1314
void *slab_start(struct seq_file *m, loff_t *pos)
1315 1316
{
	mutex_lock(&slab_mutex);
1317
	return seq_list_start(&slab_root_caches, *pos);
1318 1319
}

1320
void *slab_next(struct seq_file *m, void *p, loff_t *pos)
1321
{
1322
	return seq_list_next(p, &slab_root_caches, pos);
1323 1324
}

1325
void slab_stop(struct seq_file *m, void *p)
1326 1327 1328 1329
{
	mutex_unlock(&slab_mutex);
}

1330 1331 1332 1333 1334 1335 1336 1337 1338
static void
memcg_accumulate_slabinfo(struct kmem_cache *s, struct slabinfo *info)
{
	struct kmem_cache *c;
	struct slabinfo sinfo;

	if (!is_root_cache(s))
		return;

1339
	for_each_memcg_cache(c, s) {
1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350
		memset(&sinfo, 0, sizeof(sinfo));
		get_slabinfo(c, &sinfo);

		info->active_slabs += sinfo.active_slabs;
		info->num_slabs += sinfo.num_slabs;
		info->shared_avail += sinfo.shared_avail;
		info->active_objs += sinfo.active_objs;
		info->num_objs += sinfo.num_objs;
	}
}

1351
static void cache_show(struct kmem_cache *s, struct seq_file *m)
1352
{
1353 1354 1355 1356 1357
	struct slabinfo sinfo;

	memset(&sinfo, 0, sizeof(sinfo));
	get_slabinfo(s, &sinfo);

1358 1359
	memcg_accumulate_slabinfo(s, &sinfo);

1360
	seq_printf(m, "%-17s %6lu %6lu %6u %4u %4d",
1361
		   cache_name(s), sinfo.active_objs, sinfo.num_objs, s->size,
1362 1363 1364 1365 1366 1367 1368 1369
		   sinfo.objects_per_slab, (1 << sinfo.cache_order));

	seq_printf(m, " : tunables %4u %4u %4u",
		   sinfo.limit, sinfo.batchcount, sinfo.shared);
	seq_printf(m, " : slabdata %6lu %6lu %6lu",
		   sinfo.active_slabs, sinfo.num_slabs, sinfo.shared_avail);
	slabinfo_show_stats(m, s);
	seq_putc(m, '\n');
1370 1371
}

1372
static int slab_show(struct seq_file *m, void *p)
1373
{
1374
	struct kmem_cache *s = list_entry(p, struct kmem_cache, root_caches_node);
1375

1376
	if (p == slab_root_caches.next)
1377
		print_slabinfo_header(m);
1378
	cache_show(s, m);
1379 1380 1381
	return 0;
}

1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415
void dump_unreclaimable_slab(void)
{
	struct kmem_cache *s, *s2;
	struct slabinfo sinfo;

	/*
	 * Here acquiring slab_mutex is risky since we don't prefer to get
	 * sleep in oom path. But, without mutex hold, it may introduce a
	 * risk of crash.
	 * Use mutex_trylock to protect the list traverse, dump nothing
	 * without acquiring the mutex.
	 */
	if (!mutex_trylock(&slab_mutex)) {
		pr_warn("excessive unreclaimable slab but cannot dump stats\n");
		return;
	}

	pr_info("Unreclaimable slab info:\n");
	pr_info("Name                      Used          Total\n");

	list_for_each_entry_safe(s, s2, &slab_caches, list) {
		if (!is_root_cache(s) || (s->flags & SLAB_RECLAIM_ACCOUNT))
			continue;

		get_slabinfo(s, &sinfo);

		if (sinfo.num_objs > 0)
			pr_info("%-17s %10luKB %10luKB\n", cache_name(s),
				(sinfo.active_objs * s->size) / 1024,
				(sinfo.num_objs * s->size) / 1024);
	}
	mutex_unlock(&slab_mutex);
}

Y
Yang Shi 已提交
1416
#if defined(CONFIG_MEMCG)
1417 1418
void *memcg_slab_start(struct seq_file *m, loff_t *pos)
{
1419
	struct mem_cgroup *memcg = mem_cgroup_from_seq(m);
1420 1421 1422 1423 1424 1425 1426

	mutex_lock(&slab_mutex);
	return seq_list_start(&memcg->kmem_caches, *pos);
}

void *memcg_slab_next(struct seq_file *m, void *p, loff_t *pos)
{
1427
	struct mem_cgroup *memcg = mem_cgroup_from_seq(m);
1428 1429 1430 1431 1432 1433 1434 1435 1436

	return seq_list_next(p, &memcg->kmem_caches, pos);
}

void memcg_slab_stop(struct seq_file *m, void *p)
{
	mutex_unlock(&slab_mutex);
}

1437 1438
int memcg_slab_show(struct seq_file *m, void *p)
{
1439 1440
	struct kmem_cache *s = list_entry(p, struct kmem_cache,
					  memcg_params.kmem_caches_node);
1441
	struct mem_cgroup *memcg = mem_cgroup_from_seq(m);
1442

1443
	if (p == memcg->kmem_caches.next)
1444
		print_slabinfo_header(m);
1445
	cache_show(s, m);
1446
	return 0;
1447
}
1448
#endif
1449

1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463
/*
 * 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
 */
static const struct seq_operations slabinfo_op = {
1464
	.start = slab_start,
1465 1466
	.next = slab_next,
	.stop = slab_stop,
1467
	.show = slab_show,
1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484
};

static int slabinfo_open(struct inode *inode, struct file *file)
{
	return seq_open(file, &slabinfo_op);
}

static const struct file_operations proc_slabinfo_operations = {
	.open		= slabinfo_open,
	.read		= seq_read,
	.write          = slabinfo_write,
	.llseek		= seq_lseek,
	.release	= seq_release,
};

static int __init slab_proc_init(void)
{
1485 1486
	proc_create("slabinfo", SLABINFO_RIGHTS, NULL,
						&proc_slabinfo_operations);
1487 1488 1489
	return 0;
}
module_init(slab_proc_init);
Y
Yang Shi 已提交
1490
#endif /* CONFIG_SLAB || CONFIG_SLUB_DEBUG */
1491 1492 1493 1494 1495 1496 1497 1498 1499 1500

static __always_inline void *__do_krealloc(const void *p, size_t new_size,
					   gfp_t flags)
{
	void *ret;
	size_t ks = 0;

	if (p)
		ks = ksize(p);

1501
	if (ks >= new_size) {
1502
		p = kasan_krealloc((void *)p, new_size, flags);
1503
		return (void *)p;
1504
	}
1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521

	ret = kmalloc_track_caller(new_size, flags);
	if (ret && p)
		memcpy(ret, p, ks);

	return ret;
}

/**
 * __krealloc - like krealloc() but don't free @p.
 * @p: object to reallocate memory for.
 * @new_size: how many bytes of memory are required.
 * @flags: the type of memory to allocate.
 *
 * This function is like krealloc() except it never frees the originally
 * allocated buffer. Use this if you don't want to free the buffer immediately
 * like, for example, with RCU.
1522 1523
 *
 * Return: pointer to the allocated memory or %NULL in case of error
1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544
 */
void *__krealloc(const void *p, size_t new_size, gfp_t flags)
{
	if (unlikely(!new_size))
		return ZERO_SIZE_PTR;

	return __do_krealloc(p, new_size, flags);

}
EXPORT_SYMBOL(__krealloc);

/**
 * krealloc - reallocate memory. The contents will remain unchanged.
 * @p: object to reallocate memory for.
 * @new_size: how many bytes of memory are required.
 * @flags: the type of memory to allocate.
 *
 * The contents of the object pointed to are preserved up to the
 * lesser of the new and old sizes.  If @p is %NULL, krealloc()
 * behaves exactly like kmalloc().  If @new_size is 0 and @p is not a
 * %NULL pointer, the object pointed to is freed.
1545 1546
 *
 * Return: pointer to the allocated memory or %NULL in case of error
1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557
 */
void *krealloc(const void *p, size_t new_size, gfp_t flags)
{
	void *ret;

	if (unlikely(!new_size)) {
		kfree(p);
		return ZERO_SIZE_PTR;
	}

	ret = __do_krealloc(p, new_size, flags);
1558
	if (ret && kasan_reset_tag(p) != kasan_reset_tag(ret))
1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588
		kfree(p);

	return ret;
}
EXPORT_SYMBOL(krealloc);

/**
 * kzfree - like kfree but zero memory
 * @p: object to free memory of
 *
 * The memory of the object @p points to is zeroed before freed.
 * If @p is %NULL, kzfree() does nothing.
 *
 * Note: this function zeroes the whole allocated buffer which can be a good
 * deal bigger than the requested buffer size passed to kmalloc(). So be
 * careful when using this function in performance sensitive code.
 */
void kzfree(const void *p)
{
	size_t ks;
	void *mem = (void *)p;

	if (unlikely(ZERO_OR_NULL_PTR(mem)))
		return;
	ks = ksize(mem);
	memset(mem, 0, ks);
	kfree(mem);
}
EXPORT_SYMBOL(kzfree);

1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604
/**
 * 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.
 *
 * Return: size of the actual memory used by @objp in bytes
 */
size_t ksize(const void *objp)
{
1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625
	size_t size;

	if (WARN_ON_ONCE(!objp))
		return 0;
	/*
	 * We need to check that the pointed to object is valid, and only then
	 * unpoison the shadow memory below. We use __kasan_check_read(), to
	 * generate a more useful report at the time ksize() is called (rather
	 * than later where behaviour is undefined due to potential
	 * use-after-free or double-free).
	 *
	 * If the pointed to memory is invalid we return 0, to avoid users of
	 * ksize() writing to and potentially corrupting the memory region.
	 *
	 * We want to perform the check before __ksize(), to avoid potentially
	 * crashing in __ksize() due to accessing invalid metadata.
	 */
	if (unlikely(objp == ZERO_SIZE_PTR) || !__kasan_check_read(objp, 1))
		return 0;

	size = __ksize(objp);
1626 1627 1628 1629 1630 1631 1632 1633 1634
	/*
	 * We assume that ksize callers could use whole allocated area,
	 * so we need to unpoison this area.
	 */
	kasan_unpoison_shadow(objp, size);
	return size;
}
EXPORT_SYMBOL(ksize);

1635 1636 1637 1638 1639 1640 1641
/* Tracepoints definitions. */
EXPORT_TRACEPOINT_SYMBOL(kmalloc);
EXPORT_TRACEPOINT_SYMBOL(kmem_cache_alloc);
EXPORT_TRACEPOINT_SYMBOL(kmalloc_node);
EXPORT_TRACEPOINT_SYMBOL(kmem_cache_alloc_node);
EXPORT_TRACEPOINT_SYMBOL(kfree);
EXPORT_TRACEPOINT_SYMBOL(kmem_cache_free);
1642 1643 1644 1645 1646 1647 1648 1649

int should_failslab(struct kmem_cache *s, gfp_t gfpflags)
{
	if (__should_failslab(s, gfpflags))
		return -ENOMEM;
	return 0;
}
ALLOW_ERROR_INJECTION(should_failslab, ERRNO);