[linux.git] / mm / slab_common.c

/*
 * Slab allocator functions that are independent of the allocator strategy
 *
 * (C) 2012 Christoph Lameter <[email protected]>
 */
#include <linux/slab.h>

#include <linux/mm.h>
#include <linux/poison.h>
#include <linux/interrupt.h>
#include <linux/memory.h>
#include <linux/compiler.h>
#include <linux/module.h>
#include <linux/cpu.h>
#include <linux/uaccess.h>
#include <linux/seq_file.h>
#include <linux/proc_fs.h>
#include <asm/cacheflush.h>
#include <asm/tlbflush.h>
#include <asm/page.h>
#include <linux/memcontrol.h>
#include <trace/events/kmem.h>

#include "slab.h"

enum slab_state slab_state;
LIST_HEAD(slab_caches);
DEFINE_MUTEX(slab_mutex);
struct kmem_cache *kmem_cache;

#ifdef CONFIG_DEBUG_VM
static int kmem_cache_sanity_check(const char *name, size_t size)
{
	struct kmem_cache *s = NULL;

	if (!name || in_interrupt() || size < sizeof(void *) ||
		size > KMALLOC_MAX_SIZE) {
		pr_err("kmem_cache_create(%s) integrity check failed\n", name);
		return -EINVAL;
	}

	list_for_each_entry(s, &slab_caches, list) {
		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.
		 */
		res = probe_kernel_address(s->name, tmp);
		if (res) {
			pr_err("Slab cache with size %d has lost its name\n",
			       s->object_size);
			continue;
		}

#if !defined(CONFIG_SLUB) || !defined(CONFIG_SLUB_DEBUG_ON)
		if (!strcmp(s->name, name)) {
			pr_err("%s (%s): Cache name already exists.\n",
			       __func__, name);
			dump_stack();
			s = NULL;
			return -EINVAL;
		}
#endif
	}

	WARN_ON(strchr(name, ' '));	/* It confuses parsers */
	return 0;
}
#else
static inline int kmem_cache_sanity_check(const char *name, size_t size)
{
	return 0;
}
#endif

#ifdef CONFIG_MEMCG_KMEM
int memcg_update_all_caches(int num_memcgs)
{
	struct kmem_cache *s;
	int ret = 0;
	mutex_lock(&slab_mutex);

	list_for_each_entry(s, &slab_caches, list) {
		if (!is_root_cache(s))
			continue;

		ret = memcg_update_cache_size(s, num_memcgs);
		/*
		 * See comment in memcontrol.c, memcg_update_cache_size:
		 * Instead of freeing the memory, we'll just leave the caches
		 * up to this point in an updated state.
		 */
		if (ret)
			goto out;
	}

	memcg_update_array_size(num_memcgs);
out:
	mutex_unlock(&slab_mutex);
	return ret;
}
#endif

/*
 * 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.
 */
unsigned long calculate_alignment(unsigned long flags,
		unsigned long align, unsigned long size)
{
	/*
	 * 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) {
		unsigned long 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 *));
}

static struct kmem_cache *
do_kmem_cache_create(char *name, size_t object_size, size_t size, size_t align,
		     unsigned long flags, void (*ctor)(void *),
		     struct mem_cgroup *memcg, struct kmem_cache *root_cache)
{
	struct kmem_cache *s;
	int err;

	err = -ENOMEM;
	s = kmem_cache_zalloc(kmem_cache, GFP_KERNEL);
	if (!s)
		goto out;

	s->name = name;
	s->object_size = object_size;
	s->size = size;
	s->align = align;
	s->ctor = ctor;

	err = memcg_alloc_cache_params(memcg, s, root_cache);
	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);
	memcg_register_cache(s);
out:
	if (err)
		return ERR_PTR(err);
	return s;

out_free_cache:
	memcg_free_cache_params(s);
	kfree(s);
	goto out;
}

/*
 * 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 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.
 */
struct kmem_cache *
kmem_cache_create(const char *name, size_t size, size_t align,
		  unsigned long flags, void (*ctor)(void *))
{
	struct kmem_cache *s;
	char *cache_name;
	int err;

	get_online_cpus();
	mutex_lock(&slab_mutex);

	err = kmem_cache_sanity_check(name, size);
	if (err)
		goto out_unlock;

	/*
	 * 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;

	s = __kmem_cache_alias(name, size, align, flags, ctor);
	if (s)
		goto out_unlock;

	cache_name = kstrdup(name, GFP_KERNEL);
	if (!cache_name) {
		err = -ENOMEM;
		goto out_unlock;
	}

	s = do_kmem_cache_create(cache_name, size, size,
				 calculate_alignment(flags, align, size),
				 flags, ctor, NULL, NULL);
	if (IS_ERR(s)) {
		err = PTR_ERR(s);
		kfree(cache_name);
	}

out_unlock:
	mutex_unlock(&slab_mutex);
	put_online_cpus();

	if (err) {
		if (flags & SLAB_PANIC)
			panic("kmem_cache_create: Failed to create slab '%s'. Error %d\n",
				name, err);
		else {
			printk(KERN_WARNING "kmem_cache_create(%s) failed with error %d",
				name, err);
			dump_stack();
		}
		return NULL;
	}
	return s;
}
EXPORT_SYMBOL(kmem_cache_create);

#ifdef CONFIG_MEMCG_KMEM
/*
 * kmem_cache_create_memcg - Create a cache for a memory cgroup.
 * @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.
 */
void kmem_cache_create_memcg(struct mem_cgroup *memcg, struct kmem_cache *root_cache)
{
	struct kmem_cache *s;
	char *cache_name;

	get_online_cpus();
	mutex_lock(&slab_mutex);

	/*
	 * 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.
	 */
	if (cache_from_memcg_idx(root_cache, memcg_cache_id(memcg)))
		goto out_unlock;

	cache_name = memcg_create_cache_name(memcg, root_cache);
	if (!cache_name)
		goto out_unlock;

	s = do_kmem_cache_create(cache_name, root_cache->object_size,
				 root_cache->size, root_cache->align,
				 root_cache->flags, root_cache->ctor,
				 memcg, root_cache);
	if (IS_ERR(s))
		kfree(cache_name);

out_unlock:
	mutex_unlock(&slab_mutex);
	put_online_cpus();
}

static int kmem_cache_destroy_memcg_children(struct kmem_cache *s)
{
	int rc;

	if (!s->memcg_params ||
	    !s->memcg_params->is_root_cache)
		return 0;

	mutex_unlock(&slab_mutex);
	rc = __kmem_cache_destroy_memcg_children(s);
	mutex_lock(&slab_mutex);

	return rc;
}
#else
static int kmem_cache_destroy_memcg_children(struct kmem_cache *s)
{
	return 0;
}
#endif /* CONFIG_MEMCG_KMEM */

void slab_kmem_cache_release(struct kmem_cache *s)
{
	kfree(s->name);
	kmem_cache_free(kmem_cache, s);
}

void kmem_cache_destroy(struct kmem_cache *s)
{
	get_online_cpus();
	mutex_lock(&slab_mutex);

	s->refcount--;
	if (s->refcount)
		goto out_unlock;

	if (kmem_cache_destroy_memcg_children(s) != 0)
		goto out_unlock;

	list_del(&s->list);
	memcg_unregister_cache(s);

	if (__kmem_cache_shutdown(s) != 0) {
		list_add(&s->list, &slab_caches);
		memcg_register_cache(s);
		printk(KERN_ERR "kmem_cache_destroy %s: "
		       "Slab cache still has objects\n", s->name);
		dump_stack();
		goto out_unlock;
	}

	mutex_unlock(&slab_mutex);
	if (s->flags & SLAB_DESTROY_BY_RCU)
		rcu_barrier();

	memcg_free_cache_params(s);
#ifdef SLAB_SUPPORTS_SYSFS
	sysfs_slab_remove(s);
#else
	slab_kmem_cache_release(s);
#endif
	goto out_put_cpus;

out_unlock:
	mutex_unlock(&slab_mutex);
out_put_cpus:
	put_online_cpus();
}
EXPORT_SYMBOL(kmem_cache_destroy);

int slab_is_available(void)
{
	return slab_state >= UP;
}

#ifndef CONFIG_SLOB
/* Create a cache during boot when no slab services are available yet */
void __init create_boot_cache(struct kmem_cache *s, const char *name, size_t size,
		unsigned long flags)
{
	int err;

	s->name = name;
	s->size = s->object_size = size;
	s->align = calculate_alignment(flags, ARCH_KMALLOC_MINALIGN, size);
	err = __kmem_cache_create(s, flags);

	if (err)
		panic("Creation of kmalloc slab %s size=%zu failed. Reason %d\n",
					name, size, err);

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

struct kmem_cache *__init create_kmalloc_cache(const char *name, size_t size,
				unsigned long flags)
{
	struct kmem_cache *s = kmem_cache_zalloc(kmem_cache, GFP_NOWAIT);

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

	create_boot_cache(s, name, size, flags);
	list_add(&s->list, &slab_caches);
	s->refcount = 1;
	return s;
}

struct kmem_cache *kmalloc_caches[KMALLOC_SHIFT_HIGH + 1];
EXPORT_SYMBOL(kmalloc_caches);

#ifdef CONFIG_ZONE_DMA
struct kmem_cache *kmalloc_dma_caches[KMALLOC_SHIFT_HIGH + 1];
EXPORT_SYMBOL(kmalloc_dma_caches);
#endif

/*
 * 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.
 */
static s8 size_index[24] = {
	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 */
};

static inline int size_index_elem(size_t bytes)
{
	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)
{
	int index;

	if (unlikely(size > KMALLOC_MAX_SIZE)) {
		WARN_ON_ONCE(!(flags & __GFP_NOWARN));
		return NULL;
	}

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

		index = size_index[size_index_elem(size)];
	} else
		index = fls(size - 1);

#ifdef CONFIG_ZONE_DMA
	if (unlikely((flags & GFP_DMA)))
		return kmalloc_dma_caches[index];

#endif
	return kmalloc_caches[index];
}

/*
 * 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.
 */
void __init create_kmalloc_caches(unsigned long flags)
{
	int i;

	/*
	 * 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
	 */
	BUILD_BUG_ON(KMALLOC_MIN_SIZE > 256 ||
		(KMALLOC_MIN_SIZE & (KMALLOC_MIN_SIZE - 1)));

	for (i = 8; i < KMALLOC_MIN_SIZE; i += 8) {
		int elem = size_index_elem(i);

		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;
	}
	for (i = KMALLOC_SHIFT_LOW; i <= KMALLOC_SHIFT_HIGH; i++) {
		if (!kmalloc_caches[i]) {
			kmalloc_caches[i] = create_kmalloc_cache(NULL,
							1 << i, flags);
		}

		/*
		 * 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 && !kmalloc_caches[1] && i == 6)
			kmalloc_caches[1] = create_kmalloc_cache(NULL, 96, flags);

		if (KMALLOC_MIN_SIZE <= 64 && !kmalloc_caches[2] && i == 7)
			kmalloc_caches[2] = create_kmalloc_cache(NULL, 192, flags);
	}

	/* Kmalloc array is now usable */
	slab_state = UP;

	for (i = 0; i <= KMALLOC_SHIFT_HIGH; i++) {
		struct kmem_cache *s = kmalloc_caches[i];
		char *n;

		if (s) {
			n = kasprintf(GFP_NOWAIT, "kmalloc-%d", kmalloc_size(i));

			BUG_ON(!n);
			s->name = n;
		}
	}

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

		if (s) {
			int size = kmalloc_size(i);
			char *n = kasprintf(GFP_NOWAIT,
				 "dma-kmalloc-%d", size);

			BUG_ON(!n);
			kmalloc_dma_caches[i] = create_kmalloc_cache(n,
				size, SLAB_CACHE_DMA | flags);
		}
	}
#endif
}
#endif /* !CONFIG_SLOB */

void *kmalloc_order(size_t size, gfp_t flags, unsigned int order)
{
	void *ret;
	struct page *page;

	flags |= __GFP_COMP;
	page = alloc_kmem_pages(flags, order);
	ret = page ? page_address(page) : NULL;
	kmemleak_alloc(ret, size, 1, flags);
	return ret;
}
EXPORT_SYMBOL(kmalloc_order);

#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

#ifdef CONFIG_SLABINFO

#ifdef CONFIG_SLAB
#define SLABINFO_RIGHTS (S_IWUSR | S_IRUSR)
#else
#define SLABINFO_RIGHTS S_IRUSR
#endif

void print_slabinfo_header(struct seq_file *m)
{
	/*
	 * 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
	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>");
#ifdef CONFIG_DEBUG_SLAB
	seq_puts(m, " : globalstat <listallocs> <maxobjs> <grown> <reaped> "
		 "<error> <maxfreeable> <nodeallocs> <remotefrees> <alienoverflow>");
	seq_puts(m, " : cpustat <allochit> <allocmiss> <freehit> <freemiss>");
#endif
	seq_putc(m, '\n');
}

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

	mutex_lock(&slab_mutex);
	if (!n)
		print_slabinfo_header(m);

	return seq_list_start(&slab_caches, *pos);
}

void *slab_next(struct seq_file *m, void *p, loff_t *pos)
{
	return seq_list_next(p, &slab_caches, pos);
}

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

static void
memcg_accumulate_slabinfo(struct kmem_cache *s, struct slabinfo *info)
{
	struct kmem_cache *c;
	struct slabinfo sinfo;
	int i;

	if (!is_root_cache(s))
		return;

	for_each_memcg_cache_index(i) {
		c = cache_from_memcg_idx(s, i);
		if (!c)
			continue;

		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;
	}
}

int cache_show(struct kmem_cache *s, struct seq_file *m)
{
	struct slabinfo sinfo;

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

	memcg_accumulate_slabinfo(s, &sinfo);

	seq_printf(m, "%-17s %6lu %6lu %6u %4u %4d",
		   cache_name(s), sinfo.active_objs, sinfo.num_objs, s->size,
		   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');
	return 0;
}

static int s_show(struct seq_file *m, void *p)
{
	struct kmem_cache *s = list_entry(p, struct kmem_cache, list);

	if (!is_root_cache(s))
		return 0;
	return cache_show(s, m);
}

/*
 * 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 = {
	.start = s_start,
	.next = slab_next,
	.stop = slab_stop,
	.show = s_show,
};

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)
{
	proc_create("slabinfo", SLABINFO_RIGHTS, NULL,
						&proc_slabinfo_operations);
	return 0;
}
module_init(slab_proc_init);
#endif /* CONFIG_SLABINFO */
Commit	Line	Data
039363f3 CL	1	/*
	2	* Slab allocator functions that are independent of the allocator strategy
	3	*
	4	* (C) 2012 Christoph Lameter <[email protected]>
	5	*/
	6	#include <linux/slab.h>
	7
	8	#include <linux/mm.h>
	9	#include <linux/poison.h>
	10	#include <linux/interrupt.h>
	11	#include <linux/memory.h>
	12	#include <linux/compiler.h>
	13	#include <linux/module.h>
20cea968 CL	14	#include <linux/cpu.h>
20cea968 CL	15	#include <linux/uaccess.h>
b7454ad3 GC	16	#include <linux/seq_file.h>
b7454ad3 GC	17	#include <linux/proc_fs.h>
039363f3 CL	18	#include <asm/cacheflush.h>
	19	#include <asm/tlbflush.h>
	20	#include <asm/page.h>
2633d7a0	21	#include <linux/memcontrol.h>
f1b6eb6e	22	#include <trace/events/kmem.h>
039363f3	23
97d06609 CL	24	#include "slab.h"
	25
	26	enum slab_state slab_state;
18004c5d CL	27	LIST_HEAD(slab_caches);
18004c5d CL	28	DEFINE_MUTEX(slab_mutex);
9b030cb8	29	struct kmem_cache *kmem_cache;
97d06609	30
77be4b13	31	#ifdef CONFIG_DEBUG_VM
794b1248	32	static int kmem_cache_sanity_check(const char *name, size_t size)
039363f3 CL	33	{
	34	struct kmem_cache *s = NULL;
	35
039363f3 CL	36	if (!name \|\| in_interrupt() \|\| size < sizeof(void *) \|\|
039363f3 CL	37	size > KMALLOC_MAX_SIZE) {
77be4b13 SK	38	pr_err("kmem_cache_create(%s) integrity check failed\n", name);
77be4b13 SK	39	return -EINVAL;
039363f3	40	}
b920536a	41
20cea968 CL	42	list_for_each_entry(s, &slab_caches, list) {
	43	char tmp;
	44	int res;
	45
	46	/*
	47	* This happens when the module gets unloaded and doesn't
	48	* destroy its slab cache and no-one else reuses the vmalloc
	49	* area of the module. Print a warning.
	50	*/
	51	res = probe_kernel_address(s->name, tmp);
	52	if (res) {
77be4b13	53	pr_err("Slab cache with size %d has lost its name\n",
20cea968 CL	54	s->object_size);
	55	continue;
	56	}
	57
3e374919	58	#if !defined(CONFIG_SLUB) \|\| !defined(CONFIG_SLUB_DEBUG_ON)
794b1248	59	if (!strcmp(s->name, name)) {
77be4b13 SK	60	pr_err("%s (%s): Cache name already exists.\n",
77be4b13 SK	61	__func__, name);
20cea968 CL	62	dump_stack();
20cea968 CL	63	s = NULL;
77be4b13	64	return -EINVAL;
20cea968	65	}
3e374919	66	#endif
20cea968 CL	67	}
	68
	69	WARN_ON(strchr(name, ' ')); /* It confuses parsers */
77be4b13 SK	70	return 0;
	71	}
	72	#else
794b1248	73	static inline int kmem_cache_sanity_check(const char *name, size_t size)
77be4b13 SK	74	{
	75	return 0;
	76	}
20cea968 CL	77	#endif
20cea968 CL	78
55007d84 GC	79	#ifdef CONFIG_MEMCG_KMEM
	80	int memcg_update_all_caches(int num_memcgs)
	81	{
	82	struct kmem_cache *s;
	83	int ret = 0;
	84	mutex_lock(&slab_mutex);
	85
	86	list_for_each_entry(s, &slab_caches, list) {
	87	if (!is_root_cache(s))
	88	continue;
	89
	90	ret = memcg_update_cache_size(s, num_memcgs);
	91	/*
	92	* See comment in memcontrol.c, memcg_update_cache_size:
	93	* Instead of freeing the memory, we'll just leave the caches
	94	* up to this point in an updated state.
	95	*/
	96	if (ret)
	97	goto out;
	98	}
	99
	100	memcg_update_array_size(num_memcgs);
	101	out:
	102	mutex_unlock(&slab_mutex);
	103	return ret;
	104	}
	105	#endif
	106
45906855 CL	107	/*
	108	* Figure out what the alignment of the objects will be given a set of
	109	* flags, a user specified alignment and the size of the objects.
	110	*/
	111	unsigned long calculate_alignment(unsigned long flags,
	112	unsigned long align, unsigned long size)
	113	{
	114	/*
	115	* If the user wants hardware cache aligned objects then follow that
	116	* suggestion if the object is sufficiently large.
	117	*
	118	* The hardware cache alignment cannot override the specified
	119	* alignment though. If that is greater then use it.
	120	*/
	121	if (flags & SLAB_HWCACHE_ALIGN) {
	122	unsigned long ralign = cache_line_size();
	123	while (size <= ralign / 2)
	124	ralign /= 2;
	125	align = max(align, ralign);
	126	}
	127
	128	if (align < ARCH_SLAB_MINALIGN)
	129	align = ARCH_SLAB_MINALIGN;
	130
	131	return ALIGN(align, sizeof(void *));
	132	}
	133
794b1248 VD	134	static struct kmem_cache *
	135	do_kmem_cache_create(char *name, size_t object_size, size_t size, size_t align,
	136	unsigned long flags, void (ctor)(void ),
	137	struct mem_cgroup memcg, struct kmem_cache root_cache)
	138	{
	139	struct kmem_cache *s;
	140	int err;
	141
	142	err = -ENOMEM;
	143	s = kmem_cache_zalloc(kmem_cache, GFP_KERNEL);
	144	if (!s)
	145	goto out;
	146
	147	s->name = name;
	148	s->object_size = object_size;
	149	s->size = size;
	150	s->align = align;
	151	s->ctor = ctor;
	152
	153	err = memcg_alloc_cache_params(memcg, s, root_cache);
	154	if (err)
	155	goto out_free_cache;
	156
	157	err = __kmem_cache_create(s, flags);
	158	if (err)
	159	goto out_free_cache;
	160
	161	s->refcount = 1;
	162	list_add(&s->list, &slab_caches);
	163	memcg_register_cache(s);
	164	out:
	165	if (err)
	166	return ERR_PTR(err);
	167	return s;
	168
	169	out_free_cache:
	170	memcg_free_cache_params(s);
	171	kfree(s);
	172	goto out;
	173	}
45906855	174
77be4b13 SK	175	/*
	176	* kmem_cache_create - Create a cache.
	177	* @name: A string which is used in /proc/slabinfo to identify this cache.
	178	* @size: The size of objects to be created in this cache.
	179	* @align: The required alignment for the objects.
	180	* @flags: SLAB flags
	181	* @ctor: A constructor for the objects.
	182	*
	183	* Returns a ptr to the cache on success, NULL on failure.
	184	* Cannot be called within a interrupt, but can be interrupted.
	185	* The @ctor is run when new pages are allocated by the cache.
	186	*
	187	* The flags are
	188	*
	189	* %SLAB_POISON - Poison the slab with a known test pattern (a5a5a5a5)
	190	* to catch references to uninitialised memory.
	191	*
	192	* %SLAB_RED_ZONE - Insert `Red' zones around the allocated memory to check
	193	* for buffer overruns.
	194	*
	195	* %SLAB_HWCACHE_ALIGN - Align the objects in this cache to a hardware
	196	* cacheline. This can be beneficial if you're counting cycles as closely
	197	* as davem.
	198	*/
2633d7a0	199	struct kmem_cache *
794b1248 VD	200	kmem_cache_create(const char *name, size_t size, size_t align,
794b1248 VD	201	unsigned long flags, void (ctor)(void ))
77be4b13	202	{
794b1248 VD	203	struct kmem_cache *s;
794b1248 VD	204	char *cache_name;
3965fc36	205	int err;
039363f3	206
77be4b13 SK	207	get_online_cpus();
77be4b13 SK	208	mutex_lock(&slab_mutex);
686d550d	209
794b1248	210	err = kmem_cache_sanity_check(name, size);
3965fc36 VD	211	if (err)
3965fc36 VD	212	goto out_unlock;
686d550d	213
d8843922 GC	214	/*
	215	* Some allocators will constraint the set of valid flags to a subset
	216	* of all flags. We expect them to define CACHE_CREATE_MASK in this
	217	* case, and we'll just provide them with a sanitized version of the
	218	* passed flags.
	219	*/
	220	flags &= CACHE_CREATE_MASK;
686d550d	221
794b1248 VD	222	s = __kmem_cache_alias(name, size, align, flags, ctor);
794b1248 VD	223	if (s)
3965fc36	224	goto out_unlock;
2633d7a0	225
794b1248 VD	226	cache_name = kstrdup(name, GFP_KERNEL);
	227	if (!cache_name) {
	228	err = -ENOMEM;
	229	goto out_unlock;
	230	}
7c9adf5a	231
794b1248 VD	232	s = do_kmem_cache_create(cache_name, size, size,
	233	calculate_alignment(flags, align, size),
	234	flags, ctor, NULL, NULL);
	235	if (IS_ERR(s)) {
	236	err = PTR_ERR(s);
	237	kfree(cache_name);
	238	}
3965fc36 VD	239
3965fc36 VD	240	out_unlock:
20cea968 CL	241	mutex_unlock(&slab_mutex);
	242	put_online_cpus();
	243
ba3253c7	244	if (err) {
686d550d CL	245	if (flags & SLAB_PANIC)
	246	panic("kmem_cache_create: Failed to create slab '%s'. Error %d\n",
	247	name, err);
	248	else {
	249	printk(KERN_WARNING "kmem_cache_create(%s) failed with error %d",
	250	name, err);
	251	dump_stack();
	252	}
686d550d CL	253	return NULL;
686d550d CL	254	}
039363f3 CL	255	return s;
039363f3 CL	256	}
794b1248	257	EXPORT_SYMBOL(kmem_cache_create);
2633d7a0	258
794b1248 VD	259	#ifdef CONFIG_MEMCG_KMEM
	260	/*
	261	* kmem_cache_create_memcg - Create a cache for a memory cgroup.
	262	* @memcg: The memory cgroup the new cache is for.
	263	* @root_cache: The parent of the new cache.
	264	*
	265	* This function attempts to create a kmem cache that will serve allocation
	266	* requests going from @memcg to @root_cache. The new cache inherits properties
	267	* from its parent.
	268	*/
	269	void kmem_cache_create_memcg(struct mem_cgroup memcg, struct kmem_cache root_cache)
2633d7a0	270	{
794b1248 VD	271	struct kmem_cache *s;
	272	char *cache_name;
	273
	274	get_online_cpus();
	275	mutex_lock(&slab_mutex);
	276
	277	/*
	278	* Since per-memcg caches are created asynchronously on first
	279	* allocation (see memcg_kmem_get_cache()), several threads can try to
	280	* create the same cache, but only one of them may succeed.
	281	*/
	282	if (cache_from_memcg_idx(root_cache, memcg_cache_id(memcg)))
	283	goto out_unlock;
	284
	285	cache_name = memcg_create_cache_name(memcg, root_cache);
	286	if (!cache_name)
	287	goto out_unlock;
	288
	289	s = do_kmem_cache_create(cache_name, root_cache->object_size,
	290	root_cache->size, root_cache->align,
	291	root_cache->flags, root_cache->ctor,
	292	memcg, root_cache);
5dfb4175	293	if (IS_ERR(s))
794b1248	294	kfree(cache_name);
794b1248 VD	295
	296	out_unlock:
	297	mutex_unlock(&slab_mutex);
	298	put_online_cpus();
2633d7a0	299	}
b8529907 VD	300
	301	static int kmem_cache_destroy_memcg_children(struct kmem_cache *s)
	302	{
	303	int rc;
	304
	305	if (!s->memcg_params \|\|
	306	!s->memcg_params->is_root_cache)
	307	return 0;
	308
	309	mutex_unlock(&slab_mutex);
	310	rc = __kmem_cache_destroy_memcg_children(s);
	311	mutex_lock(&slab_mutex);
	312
	313	return rc;
	314	}
	315	#else
	316	static int kmem_cache_destroy_memcg_children(struct kmem_cache *s)
	317	{
	318	return 0;
	319	}
794b1248	320	#endif /* CONFIG_MEMCG_KMEM */
97d06609	321
41a21285 CL	322	void slab_kmem_cache_release(struct kmem_cache *s)
	323	{
	324	kfree(s->name);
	325	kmem_cache_free(kmem_cache, s);
	326	}
	327
945cf2b6 CL	328	void kmem_cache_destroy(struct kmem_cache *s)
	329	{
	330	get_online_cpus();
	331	mutex_lock(&slab_mutex);
b8529907	332
945cf2b6	333	s->refcount--;
b8529907 VD	334	if (s->refcount)
	335	goto out_unlock;
	336
	337	if (kmem_cache_destroy_memcg_children(s) != 0)
	338	goto out_unlock;
	339
	340	list_del(&s->list);
	341	memcg_unregister_cache(s);
	342
	343	if (__kmem_cache_shutdown(s) != 0) {
	344	list_add(&s->list, &slab_caches);
	345	memcg_register_cache(s);
	346	printk(KERN_ERR "kmem_cache_destroy %s: "
	347	"Slab cache still has objects\n", s->name);
	348	dump_stack();
	349	goto out_unlock;
945cf2b6	350	}
b8529907 VD	351
	352	mutex_unlock(&slab_mutex);
	353	if (s->flags & SLAB_DESTROY_BY_RCU)
	354	rcu_barrier();
	355
	356	memcg_free_cache_params(s);
41a21285 CL	357	#ifdef SLAB_SUPPORTS_SYSFS
	358	sysfs_slab_remove(s);
	359	#else
	360	slab_kmem_cache_release(s);
	361	#endif
b8529907 VD	362	goto out_put_cpus;
	363
	364	out_unlock:
	365	mutex_unlock(&slab_mutex);
	366	out_put_cpus:
945cf2b6 CL	367	put_online_cpus();
	368	}
	369	EXPORT_SYMBOL(kmem_cache_destroy);
	370
97d06609 CL	371	int slab_is_available(void)
	372	{
	373	return slab_state >= UP;
	374	}
b7454ad3	375
45530c44 CL	376	#ifndef CONFIG_SLOB
	377	/* Create a cache during boot when no slab services are available yet */
	378	void __init create_boot_cache(struct kmem_cache s, const char name, size_t size,
	379	unsigned long flags)
	380	{
	381	int err;
	382
	383	s->name = name;
	384	s->size = s->object_size = size;
45906855	385	s->align = calculate_alignment(flags, ARCH_KMALLOC_MINALIGN, size);
45530c44 CL	386	err = __kmem_cache_create(s, flags);
	387
	388	if (err)
31ba7346	389	panic("Creation of kmalloc slab %s size=%zu failed. Reason %d\n",
45530c44 CL	390	name, size, err);
	391
	392	s->refcount = -1; /* Exempt from merging for now */
	393	}
	394
	395	struct kmem_cache __init create_kmalloc_cache(const char name, size_t size,
	396	unsigned long flags)
	397	{
	398	struct kmem_cache *s = kmem_cache_zalloc(kmem_cache, GFP_NOWAIT);
	399
	400	if (!s)
	401	panic("Out of memory when creating slab %s\n", name);
	402
	403	create_boot_cache(s, name, size, flags);
	404	list_add(&s->list, &slab_caches);
	405	s->refcount = 1;
	406	return s;
	407	}
	408
9425c58e CL	409	struct kmem_cache *kmalloc_caches[KMALLOC_SHIFT_HIGH + 1];
	410	EXPORT_SYMBOL(kmalloc_caches);
	411
	412	#ifdef CONFIG_ZONE_DMA
	413	struct kmem_cache *kmalloc_dma_caches[KMALLOC_SHIFT_HIGH + 1];
	414	EXPORT_SYMBOL(kmalloc_dma_caches);
	415	#endif
	416
2c59dd65 CL	417	/*
	418	* Conversion table for small slabs sizes / 8 to the index in the
	419	* kmalloc array. This is necessary for slabs < 192 since we have non power
	420	* of two cache sizes there. The size of larger slabs can be determined using
	421	* fls.
	422	*/
	423	static s8 size_index[24] = {
	424	3, /* 8 */
	425	4, /* 16 */
	426	5, /* 24 */
	427	5, /* 32 */
	428	6, /* 40 */
	429	6, /* 48 */
	430	6, /* 56 */
	431	6, /* 64 */
	432	1, /* 72 */
	433	1, /* 80 */
	434	1, /* 88 */
	435	1, /* 96 */
	436	7, /* 104 */
	437	7, /* 112 */
	438	7, /* 120 */
	439	7, /* 128 */
	440	2, /* 136 */
	441	2, /* 144 */
	442	2, /* 152 */
	443	2, /* 160 */
	444	2, /* 168 */
	445	2, /* 176 */
	446	2, /* 184 */
	447	2 /* 192 */
	448	};
	449
	450	static inline int size_index_elem(size_t bytes)
	451	{
	452	return (bytes - 1) / 8;
	453	}
	454
	455	/*
	456	* Find the kmem_cache structure that serves a given size of
	457	* allocation
	458	*/
	459	struct kmem_cache *kmalloc_slab(size_t size, gfp_t flags)
	460	{
	461	int index;
	462
9de1bc87	463	if (unlikely(size > KMALLOC_MAX_SIZE)) {
907985f4	464	WARN_ON_ONCE(!(flags & __GFP_NOWARN));
6286ae97	465	return NULL;
907985f4	466	}
6286ae97	467
2c59dd65 CL	468	if (size <= 192) {
	469	if (!size)
	470	return ZERO_SIZE_PTR;
	471
	472	index = size_index[size_index_elem(size)];
	473	} else
	474	index = fls(size - 1);
	475
	476	#ifdef CONFIG_ZONE_DMA
b1e05416	477	if (unlikely((flags & GFP_DMA)))
2c59dd65 CL	478	return kmalloc_dma_caches[index];
	479
	480	#endif
	481	return kmalloc_caches[index];
	482	}
	483
f97d5f63 CL	484	/*
	485	* Create the kmalloc array. Some of the regular kmalloc arrays
	486	* may already have been created because they were needed to
	487	* enable allocations for slab creation.
	488	*/
	489	void __init create_kmalloc_caches(unsigned long flags)
	490	{
	491	int i;
	492
2c59dd65 CL	493	/*
	494	* Patch up the size_index table if we have strange large alignment
	495	* requirements for the kmalloc array. This is only the case for
	496	* MIPS it seems. The standard arches will not generate any code here.
	497	*
	498	* Largest permitted alignment is 256 bytes due to the way we
	499	* handle the index determination for the smaller caches.
	500	*
	501	* Make sure that nothing crazy happens if someone starts tinkering
	502	* around with ARCH_KMALLOC_MINALIGN
	503	*/
	504	BUILD_BUG_ON(KMALLOC_MIN_SIZE > 256 \|\|
	505	(KMALLOC_MIN_SIZE & (KMALLOC_MIN_SIZE - 1)));
	506
	507	for (i = 8; i < KMALLOC_MIN_SIZE; i += 8) {
	508	int elem = size_index_elem(i);
	509
	510	if (elem >= ARRAY_SIZE(size_index))
	511	break;
	512	size_index[elem] = KMALLOC_SHIFT_LOW;
	513	}
	514
	515	if (KMALLOC_MIN_SIZE >= 64) {
	516	/*
	517	* The 96 byte size cache is not used if the alignment
	518	* is 64 byte.
	519	*/
	520	for (i = 64 + 8; i <= 96; i += 8)
	521	size_index[size_index_elem(i)] = 7;
	522
	523	}
	524
	525	if (KMALLOC_MIN_SIZE >= 128) {
	526	/*
	527	* The 192 byte sized cache is not used if the alignment
	528	* is 128 byte. Redirect kmalloc to use the 256 byte cache
	529	* instead.
	530	*/
	531	for (i = 128 + 8; i <= 192; i += 8)
	532	size_index[size_index_elem(i)] = 8;
	533	}
8a965b3b CL	534	for (i = KMALLOC_SHIFT_LOW; i <= KMALLOC_SHIFT_HIGH; i++) {
8a965b3b CL	535	if (!kmalloc_caches[i]) {
f97d5f63 CL	536	kmalloc_caches[i] = create_kmalloc_cache(NULL,
f97d5f63 CL	537	1 << i, flags);
956e46ef	538	}
f97d5f63	539
956e46ef CM	540	/*
	541	* Caches that are not of the two-to-the-power-of size.
	542	* These have to be created immediately after the
	543	* earlier power of two caches
	544	*/
	545	if (KMALLOC_MIN_SIZE <= 32 && !kmalloc_caches[1] && i == 6)
	546	kmalloc_caches[1] = create_kmalloc_cache(NULL, 96, flags);
8a965b3b	547
956e46ef CM	548	if (KMALLOC_MIN_SIZE <= 64 && !kmalloc_caches[2] && i == 7)
956e46ef CM	549	kmalloc_caches[2] = create_kmalloc_cache(NULL, 192, flags);
8a965b3b CL	550	}
8a965b3b CL	551
f97d5f63 CL	552	/* Kmalloc array is now usable */
	553	slab_state = UP;
	554
	555	for (i = 0; i <= KMALLOC_SHIFT_HIGH; i++) {
	556	struct kmem_cache *s = kmalloc_caches[i];
	557	char *n;
	558
	559	if (s) {
	560	n = kasprintf(GFP_NOWAIT, "kmalloc-%d", kmalloc_size(i));
	561
	562	BUG_ON(!n);
	563	s->name = n;
	564	}
	565	}
	566
	567	#ifdef CONFIG_ZONE_DMA
	568	for (i = 0; i <= KMALLOC_SHIFT_HIGH; i++) {
	569	struct kmem_cache *s = kmalloc_caches[i];
	570
	571	if (s) {
	572	int size = kmalloc_size(i);
	573	char *n = kasprintf(GFP_NOWAIT,
	574	"dma-kmalloc-%d", size);
	575
	576	BUG_ON(!n);
	577	kmalloc_dma_caches[i] = create_kmalloc_cache(n,
	578	size, SLAB_CACHE_DMA \| flags);
	579	}
	580	}
	581	#endif
	582	}
45530c44 CL	583	#endif /* !CONFIG_SLOB */
45530c44 CL	584
52383431 VD	585	void *kmalloc_order(size_t size, gfp_t flags, unsigned int order)
	586	{
	587	void *ret;
	588	struct page *page;
	589
	590	flags \|= __GFP_COMP;
	591	page = alloc_kmem_pages(flags, order);
	592	ret = page ? page_address(page) : NULL;
	593	kmemleak_alloc(ret, size, 1, flags);
	594	return ret;
	595	}
	596	EXPORT_SYMBOL(kmalloc_order);
	597
f1b6eb6e CL	598	#ifdef CONFIG_TRACING
	599	void *kmalloc_order_trace(size_t size, gfp_t flags, unsigned int order)
	600	{
	601	void *ret = kmalloc_order(size, flags, order);
	602	trace_kmalloc(_RET_IP_, ret, size, PAGE_SIZE << order, flags);
	603	return ret;
	604	}
	605	EXPORT_SYMBOL(kmalloc_order_trace);
	606	#endif
45530c44	607
b7454ad3	608	#ifdef CONFIG_SLABINFO
e9b4db2b WL	609
	610	#ifdef CONFIG_SLAB
	611	#define SLABINFO_RIGHTS (S_IWUSR \| S_IRUSR)
	612	#else
	613	#define SLABINFO_RIGHTS S_IRUSR
	614	#endif
	615
749c5415	616	void print_slabinfo_header(struct seq_file *m)
bcee6e2a GC	617	{
	618	/*
	619	* Output format version, so at least we can change it
	620	* without _too_ many complaints.
	621	*/
	622	#ifdef CONFIG_DEBUG_SLAB
	623	seq_puts(m, "slabinfo - version: 2.1 (statistics)\n");
	624	#else
	625	seq_puts(m, "slabinfo - version: 2.1\n");
	626	#endif
	627	seq_puts(m, "# name <active_objs> <num_objs> <objsize> "
	628	"<objperslab> <pagesperslab>");
	629	seq_puts(m, " : tunables <limit> <batchcount> <sharedfactor>");
	630	seq_puts(m, " : slabdata <active_slabs> <num_slabs> <sharedavail>");
	631	#ifdef CONFIG_DEBUG_SLAB
	632	seq_puts(m, " : globalstat <listallocs> <maxobjs> <grown> <reaped> "
	633	"<error> <maxfreeable> <nodeallocs> <remotefrees> <alienoverflow>");
	634	seq_puts(m, " : cpustat <allochit> <allocmiss> <freehit> <freemiss>");
	635	#endif
	636	seq_putc(m, '\n');
	637	}
	638
b7454ad3 GC	639	static void s_start(struct seq_file m, loff_t *pos)
	640	{
	641	loff_t n = *pos;
	642
	643	mutex_lock(&slab_mutex);
	644	if (!n)
	645	print_slabinfo_header(m);
	646
	647	return seq_list_start(&slab_caches, *pos);
	648	}
	649
276a2439	650	void slab_next(struct seq_file m, void p, loff_t pos)
b7454ad3 GC	651	{
	652	return seq_list_next(p, &slab_caches, pos);
	653	}
	654
276a2439	655	void slab_stop(struct seq_file m, void p)
b7454ad3 GC	656	{
	657	mutex_unlock(&slab_mutex);
	658	}
	659
749c5415 GC	660	static void
	661	memcg_accumulate_slabinfo(struct kmem_cache s, struct slabinfo info)
	662	{
	663	struct kmem_cache *c;
	664	struct slabinfo sinfo;
	665	int i;
	666
	667	if (!is_root_cache(s))
	668	return;
	669
	670	for_each_memcg_cache_index(i) {
2ade4de8	671	c = cache_from_memcg_idx(s, i);
749c5415 GC	672	if (!c)
	673	continue;
	674
	675	memset(&sinfo, 0, sizeof(sinfo));
	676	get_slabinfo(c, &sinfo);
	677
	678	info->active_slabs += sinfo.active_slabs;
	679	info->num_slabs += sinfo.num_slabs;
	680	info->shared_avail += sinfo.shared_avail;
	681	info->active_objs += sinfo.active_objs;
	682	info->num_objs += sinfo.num_objs;
	683	}
	684	}
	685
	686	int cache_show(struct kmem_cache s, struct seq_file m)
b7454ad3	687	{
0d7561c6 GC	688	struct slabinfo sinfo;
	689
	690	memset(&sinfo, 0, sizeof(sinfo));
	691	get_slabinfo(s, &sinfo);
	692
749c5415 GC	693	memcg_accumulate_slabinfo(s, &sinfo);
749c5415 GC	694
0d7561c6	695	seq_printf(m, "%-17s %6lu %6lu %6u %4u %4d",
749c5415	696	cache_name(s), sinfo.active_objs, sinfo.num_objs, s->size,
0d7561c6 GC	697	sinfo.objects_per_slab, (1 << sinfo.cache_order));
	698
	699	seq_printf(m, " : tunables %4u %4u %4u",
	700	sinfo.limit, sinfo.batchcount, sinfo.shared);
	701	seq_printf(m, " : slabdata %6lu %6lu %6lu",
	702	sinfo.active_slabs, sinfo.num_slabs, sinfo.shared_avail);
	703	slabinfo_show_stats(m, s);
	704	seq_putc(m, '\n');
	705	return 0;
b7454ad3 GC	706	}
b7454ad3 GC	707
749c5415 GC	708	static int s_show(struct seq_file m, void p)
	709	{
	710	struct kmem_cache *s = list_entry(p, struct kmem_cache, list);
	711
	712	if (!is_root_cache(s))
	713	return 0;
	714	return cache_show(s, m);
	715	}
	716
b7454ad3 GC	717	/*
	718	* slabinfo_op - iterator that generates /proc/slabinfo
	719	*
	720	* Output layout:
	721	* cache-name
	722	* num-active-objs
	723	* total-objs
	724	* object size
	725	* num-active-slabs
	726	* total-slabs
	727	* num-pages-per-slab
	728	* + further values on SMP and with statistics enabled
	729	*/
	730	static const struct seq_operations slabinfo_op = {
	731	.start = s_start,
276a2439 WL	732	.next = slab_next,
276a2439 WL	733	.stop = slab_stop,
b7454ad3 GC	734	.show = s_show,
	735	};
	736
	737	static int slabinfo_open(struct inode inode, struct file file)
	738	{
	739	return seq_open(file, &slabinfo_op);
	740	}
	741
	742	static const struct file_operations proc_slabinfo_operations = {
	743	.open = slabinfo_open,
	744	.read = seq_read,
	745	.write = slabinfo_write,
	746	.llseek = seq_lseek,
	747	.release = seq_release,
	748	};
	749
	750	static int __init slab_proc_init(void)
	751	{
e9b4db2b WL	752	proc_create("slabinfo", SLABINFO_RIGHTS, NULL,
e9b4db2b WL	753	&proc_slabinfo_operations);
b7454ad3 GC	754	return 0;
	755	}
	756	module_init(slab_proc_init);
	757	#endif /* CONFIG_SLABINFO */