[linux.git] / mm / page_cgroup.c

#include <linux/mm.h>
#include <linux/mmzone.h>
#include <linux/bootmem.h>
#include <linux/bit_spinlock.h>
#include <linux/page_cgroup.h>
#include <linux/hash.h>
#include <linux/slab.h>
#include <linux/memory.h>
#include <linux/vmalloc.h>
#include <linux/cgroup.h>
#include <linux/swapops.h>

static void __meminit
__init_page_cgroup(struct page_cgroup *pc, unsigned long pfn)
{
	pc->flags = 0;
	pc->mem_cgroup = NULL;
	pc->page = pfn_to_page(pfn);
	INIT_LIST_HEAD(&pc->lru);
}
static unsigned long total_usage;

#if !defined(CONFIG_SPARSEMEM)


void __meminit pgdat_page_cgroup_init(struct pglist_data *pgdat)
{
	pgdat->node_page_cgroup = NULL;
}

struct page_cgroup *lookup_page_cgroup(struct page *page)
{
	unsigned long pfn = page_to_pfn(page);
	unsigned long offset;
	struct page_cgroup *base;

	base = NODE_DATA(page_to_nid(page))->node_page_cgroup;
	if (unlikely(!base))
		return NULL;

	offset = pfn - NODE_DATA(page_to_nid(page))->node_start_pfn;
	return base + offset;
}

static int __init alloc_node_page_cgroup(int nid)
{
	struct page_cgroup *base, *pc;
	unsigned long table_size;
	unsigned long start_pfn, nr_pages, index;

	start_pfn = NODE_DATA(nid)->node_start_pfn;
	nr_pages = NODE_DATA(nid)->node_spanned_pages;

	if (!nr_pages)
		return 0;

	table_size = sizeof(struct page_cgroup) * nr_pages;

	base = __alloc_bootmem_node_nopanic(NODE_DATA(nid),
			table_size, PAGE_SIZE, __pa(MAX_DMA_ADDRESS));
	if (!base)
		return -ENOMEM;
	for (index = 0; index < nr_pages; index++) {
		pc = base + index;
		__init_page_cgroup(pc, start_pfn + index);
	}
	NODE_DATA(nid)->node_page_cgroup = base;
	total_usage += table_size;
	return 0;
}

void __init page_cgroup_init_flatmem(void)
{

	int nid, fail;

	if (mem_cgroup_disabled())
		return;

	for_each_online_node(nid)  {
		fail = alloc_node_page_cgroup(nid);
		if (fail)
			goto fail;
	}
	printk(KERN_INFO "allocated %ld bytes of page_cgroup\n", total_usage);
	printk(KERN_INFO "please try 'cgroup_disable=memory' option if you"
	" don't want memory cgroups\n");
	return;
fail:
	printk(KERN_CRIT "allocation of page_cgroup failed.\n");
	printk(KERN_CRIT "please try 'cgroup_disable=memory' boot option\n");
	panic("Out of memory");
}

#else /* CONFIG_FLAT_NODE_MEM_MAP */

struct page_cgroup *lookup_page_cgroup(struct page *page)
{
	unsigned long pfn = page_to_pfn(page);
	struct mem_section *section = __pfn_to_section(pfn);

	if (!section->page_cgroup)
		return NULL;
	return section->page_cgroup + pfn;
}

/* __alloc_bootmem...() is protected by !slab_available() */
static int __init_refok init_section_page_cgroup(unsigned long pfn)
{
	struct mem_section *section = __pfn_to_section(pfn);
	struct page_cgroup *base, *pc;
	unsigned long table_size;
	int nid, index;

	if (!section->page_cgroup) {
		nid = page_to_nid(pfn_to_page(pfn));
		table_size = sizeof(struct page_cgroup) * PAGES_PER_SECTION;
		VM_BUG_ON(!slab_is_available());
		if (node_state(nid, N_HIGH_MEMORY)) {
			base = kmalloc_node(table_size,
				GFP_KERNEL | __GFP_NOWARN, nid);
			if (!base)
				base = vmalloc_node(table_size, nid);
		} else {
			base = kmalloc(table_size, GFP_KERNEL | __GFP_NOWARN);
			if (!base)
				base = vmalloc(table_size);
		}
	} else {
		/*
 		 * We don't have to allocate page_cgroup again, but
		 * address of memmap may be changed. So, we have to initialize
		 * again.
		 */
		base = section->page_cgroup + pfn;
		table_size = 0;
		/* check address of memmap is changed or not. */
		if (base->page == pfn_to_page(pfn))
			return 0;
	}

	if (!base) {
		printk(KERN_ERR "page cgroup allocation failure\n");
		return -ENOMEM;
	}

	for (index = 0; index < PAGES_PER_SECTION; index++) {
		pc = base + index;
		__init_page_cgroup(pc, pfn + index);
	}

	section->page_cgroup = base - pfn;
	total_usage += table_size;
	return 0;
}
#ifdef CONFIG_MEMORY_HOTPLUG
void __free_page_cgroup(unsigned long pfn)
{
	struct mem_section *ms;
	struct page_cgroup *base;

	ms = __pfn_to_section(pfn);
	if (!ms || !ms->page_cgroup)
		return;
	base = ms->page_cgroup + pfn;
	if (is_vmalloc_addr(base)) {
		vfree(base);
		ms->page_cgroup = NULL;
	} else {
		struct page *page = virt_to_page(base);
		if (!PageReserved(page)) { /* Is bootmem ? */
			kfree(base);
			ms->page_cgroup = NULL;
		}
	}
}

int __meminit online_page_cgroup(unsigned long start_pfn,
			unsigned long nr_pages,
			int nid)
{
	unsigned long start, end, pfn;
	int fail = 0;

	start = start_pfn & ~(PAGES_PER_SECTION - 1);
	end = ALIGN(start_pfn + nr_pages, PAGES_PER_SECTION);

	for (pfn = start; !fail && pfn < end; pfn += PAGES_PER_SECTION) {
		if (!pfn_present(pfn))
			continue;
		fail = init_section_page_cgroup(pfn);
	}
	if (!fail)
		return 0;

	/* rollback */
	for (pfn = start; pfn < end; pfn += PAGES_PER_SECTION)
		__free_page_cgroup(pfn);

	return -ENOMEM;
}

int __meminit offline_page_cgroup(unsigned long start_pfn,
		unsigned long nr_pages, int nid)
{
	unsigned long start, end, pfn;

	start = start_pfn & ~(PAGES_PER_SECTION - 1);
	end = ALIGN(start_pfn + nr_pages, PAGES_PER_SECTION);

	for (pfn = start; pfn < end; pfn += PAGES_PER_SECTION)
		__free_page_cgroup(pfn);
	return 0;

}

static int __meminit page_cgroup_callback(struct notifier_block *self,
			       unsigned long action, void *arg)
{
	struct memory_notify *mn = arg;
	int ret = 0;
	switch (action) {
	case MEM_GOING_ONLINE:
		ret = online_page_cgroup(mn->start_pfn,
				   mn->nr_pages, mn->status_change_nid);
		break;
	case MEM_OFFLINE:
		offline_page_cgroup(mn->start_pfn,
				mn->nr_pages, mn->status_change_nid);
		break;
	case MEM_CANCEL_ONLINE:
	case MEM_GOING_OFFLINE:
		break;
	case MEM_ONLINE:
	case MEM_CANCEL_OFFLINE:
		break;
	}

	if (ret)
		ret = notifier_from_errno(ret);
	else
		ret = NOTIFY_OK;

	return ret;
}

#endif

void __init page_cgroup_init(void)
{
	unsigned long pfn;
	int fail = 0;

	if (mem_cgroup_disabled())
		return;

	for (pfn = 0; !fail && pfn < max_pfn; pfn += PAGES_PER_SECTION) {
		if (!pfn_present(pfn))
			continue;
		fail = init_section_page_cgroup(pfn);
	}
	if (fail) {
		printk(KERN_CRIT "try 'cgroup_disable=memory' boot option\n");
		panic("Out of memory");
	} else {
		hotplug_memory_notifier(page_cgroup_callback, 0);
	}
	printk(KERN_INFO "allocated %ld bytes of page_cgroup\n", total_usage);
	printk(KERN_INFO "please try 'cgroup_disable=memory' option if you don't"
	" want memory cgroups\n");
}

void __meminit pgdat_page_cgroup_init(struct pglist_data *pgdat)
{
	return;
}

#endif


#ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP

static DEFINE_MUTEX(swap_cgroup_mutex);
struct swap_cgroup_ctrl {
	struct page **map;
	unsigned long length;
	spinlock_t	lock;
};

struct swap_cgroup_ctrl swap_cgroup_ctrl[MAX_SWAPFILES];

struct swap_cgroup {
	unsigned short		id;
};
#define SC_PER_PAGE	(PAGE_SIZE/sizeof(struct swap_cgroup))
#define SC_POS_MASK	(SC_PER_PAGE - 1)

/*
 * SwapCgroup implements "lookup" and "exchange" operations.
 * In typical usage, this swap_cgroup is accessed via memcg's charge/uncharge
 * against SwapCache. At swap_free(), this is accessed directly from swap.
 *
 * This means,
 *  - we have no race in "exchange" when we're accessed via SwapCache because
 *    SwapCache(and its swp_entry) is under lock.
 *  - When called via swap_free(), there is no user of this entry and no race.
 * Then, we don't need lock around "exchange".
 *
 * TODO: we can push these buffers out to HIGHMEM.
 */

/*
 * allocate buffer for swap_cgroup.
 */
static int swap_cgroup_prepare(int type)
{
	struct page *page;
	struct swap_cgroup_ctrl *ctrl;
	unsigned long idx, max;

	ctrl = &swap_cgroup_ctrl[type];

	for (idx = 0; idx < ctrl->length; idx++) {
		page = alloc_page(GFP_KERNEL | __GFP_ZERO);
		if (!page)
			goto not_enough_page;
		ctrl->map[idx] = page;
	}
	return 0;
not_enough_page:
	max = idx;
	for (idx = 0; idx < max; idx++)
		__free_page(ctrl->map[idx]);

	return -ENOMEM;
}

/**
 * swap_cgroup_cmpxchg - cmpxchg mem_cgroup's id for this swp_entry.
 * @end: swap entry to be cmpxchged
 * @old: old id
 * @new: new id
 *
 * Returns old id at success, 0 at failure.
 * (There is no mem_cgroup useing 0 as its id)
 */
unsigned short swap_cgroup_cmpxchg(swp_entry_t ent,
					unsigned short old, unsigned short new)
{
	int type = swp_type(ent);
	unsigned long offset = swp_offset(ent);
	unsigned long idx = offset / SC_PER_PAGE;
	unsigned long pos = offset & SC_POS_MASK;
	struct swap_cgroup_ctrl *ctrl;
	struct page *mappage;
	struct swap_cgroup *sc;
	unsigned long flags;
	unsigned short retval;

	ctrl = &swap_cgroup_ctrl[type];

	mappage = ctrl->map[idx];
	sc = page_address(mappage);
	sc += pos;
	spin_lock_irqsave(&ctrl->lock, flags);
	retval = sc->id;
	if (retval == old)
		sc->id = new;
	else
		retval = 0;
	spin_unlock_irqrestore(&ctrl->lock, flags);
	return retval;
}

/**
 * swap_cgroup_record - record mem_cgroup for this swp_entry.
 * @ent: swap entry to be recorded into
 * @mem: mem_cgroup to be recorded
 *
 * Returns old value at success, 0 at failure.
 * (Of course, old value can be 0.)
 */
unsigned short swap_cgroup_record(swp_entry_t ent, unsigned short id)
{
	int type = swp_type(ent);
	unsigned long offset = swp_offset(ent);
	unsigned long idx = offset / SC_PER_PAGE;
	unsigned long pos = offset & SC_POS_MASK;
	struct swap_cgroup_ctrl *ctrl;
	struct page *mappage;
	struct swap_cgroup *sc;
	unsigned short old;
	unsigned long flags;

	ctrl = &swap_cgroup_ctrl[type];

	mappage = ctrl->map[idx];
	sc = page_address(mappage);
	sc += pos;
	spin_lock_irqsave(&ctrl->lock, flags);
	old = sc->id;
	sc->id = id;
	spin_unlock_irqrestore(&ctrl->lock, flags);

	return old;
}

/**
 * lookup_swap_cgroup - lookup mem_cgroup tied to swap entry
 * @ent: swap entry to be looked up.
 *
 * Returns CSS ID of mem_cgroup at success. 0 at failure. (0 is invalid ID)
 */
unsigned short lookup_swap_cgroup(swp_entry_t ent)
{
	int type = swp_type(ent);
	unsigned long offset = swp_offset(ent);
	unsigned long idx = offset / SC_PER_PAGE;
	unsigned long pos = offset & SC_POS_MASK;
	struct swap_cgroup_ctrl *ctrl;
	struct page *mappage;
	struct swap_cgroup *sc;
	unsigned short ret;

	ctrl = &swap_cgroup_ctrl[type];
	mappage = ctrl->map[idx];
	sc = page_address(mappage);
	sc += pos;
	ret = sc->id;
	return ret;
}

int swap_cgroup_swapon(int type, unsigned long max_pages)
{
	void *array;
	unsigned long array_size;
	unsigned long length;
	struct swap_cgroup_ctrl *ctrl;

	if (!do_swap_account)
		return 0;

	length = ((max_pages/SC_PER_PAGE) + 1);
	array_size = length * sizeof(void *);

	array = vmalloc(array_size);
	if (!array)
		goto nomem;

	memset(array, 0, array_size);
	ctrl = &swap_cgroup_ctrl[type];
	mutex_lock(&swap_cgroup_mutex);
	ctrl->length = length;
	ctrl->map = array;
	spin_lock_init(&ctrl->lock);
	if (swap_cgroup_prepare(type)) {
		/* memory shortage */
		ctrl->map = NULL;
		ctrl->length = 0;
		vfree(array);
		mutex_unlock(&swap_cgroup_mutex);
		goto nomem;
	}
	mutex_unlock(&swap_cgroup_mutex);

	return 0;
nomem:
	printk(KERN_INFO "couldn't allocate enough memory for swap_cgroup.\n");
	printk(KERN_INFO
		"swap_cgroup can be disabled by noswapaccount boot option\n");
	return -ENOMEM;
}

void swap_cgroup_swapoff(int type)
{
	int i;
	struct swap_cgroup_ctrl *ctrl;

	if (!do_swap_account)
		return;

	mutex_lock(&swap_cgroup_mutex);
	ctrl = &swap_cgroup_ctrl[type];
	if (ctrl->map) {
		for (i = 0; i < ctrl->length; i++) {
			struct page *page = ctrl->map[i];
			if (page)
				__free_page(page);
		}
		vfree(ctrl->map);
		ctrl->map = NULL;
		ctrl->length = 0;
	}
	mutex_unlock(&swap_cgroup_mutex);
}

#endif
Commit	Line	Data
52d4b9ac KH	1	#include <linux/mm.h>
	2	#include <linux/mmzone.h>
	3	#include <linux/bootmem.h>
	4	#include <linux/bit_spinlock.h>
	5	#include <linux/page_cgroup.h>
	6	#include <linux/hash.h>
94b6da5a	7	#include <linux/slab.h>
52d4b9ac	8	#include <linux/memory.h>
4c821042	9	#include <linux/vmalloc.h>
94b6da5a	10	#include <linux/cgroup.h>
27a7faa0	11	#include <linux/swapops.h>
52d4b9ac KH	12
	13	static void __meminit
	14	__init_page_cgroup(struct page_cgroup *pc, unsigned long pfn)
	15	{
	16	pc->flags = 0;
	17	pc->mem_cgroup = NULL;
	18	pc->page = pfn_to_page(pfn);
08e552c6	19	INIT_LIST_HEAD(&pc->lru);
52d4b9ac KH	20	}
	21	static unsigned long total_usage;
	22
	23	#if !defined(CONFIG_SPARSEMEM)
	24
	25
31168481	26	void __meminit pgdat_page_cgroup_init(struct pglist_data *pgdat)
52d4b9ac KH	27	{
	28	pgdat->node_page_cgroup = NULL;
	29	}
	30
	31	struct page_cgroup lookup_page_cgroup(struct page page)
	32	{
	33	unsigned long pfn = page_to_pfn(page);
	34	unsigned long offset;
	35	struct page_cgroup *base;
	36
	37	base = NODE_DATA(page_to_nid(page))->node_page_cgroup;
	38	if (unlikely(!base))
	39	return NULL;
	40
	41	offset = pfn - NODE_DATA(page_to_nid(page))->node_start_pfn;
	42	return base + offset;
	43	}
	44
	45	static int __init alloc_node_page_cgroup(int nid)
	46	{
	47	struct page_cgroup base, pc;
	48	unsigned long table_size;
	49	unsigned long start_pfn, nr_pages, index;
	50
	51	start_pfn = NODE_DATA(nid)->node_start_pfn;
	52	nr_pages = NODE_DATA(nid)->node_spanned_pages;
	53
653d22c0 KH	54	if (!nr_pages)
	55	return 0;
	56
52d4b9ac	57	table_size = sizeof(struct page_cgroup) * nr_pages;
ca371c0d KH	58
	59	base = __alloc_bootmem_node_nopanic(NODE_DATA(nid),
	60	table_size, PAGE_SIZE, __pa(MAX_DMA_ADDRESS));
	61	if (!base)
52d4b9ac KH	62	return -ENOMEM;
	63	for (index = 0; index < nr_pages; index++) {
	64	pc = base + index;
	65	__init_page_cgroup(pc, start_pfn + index);
	66	}
	67	NODE_DATA(nid)->node_page_cgroup = base;
	68	total_usage += table_size;
	69	return 0;
	70	}
	71
ca371c0d	72	void __init page_cgroup_init_flatmem(void)
52d4b9ac KH	73	{
	74
	75	int nid, fail;
	76
f8d66542	77	if (mem_cgroup_disabled())
94b6da5a KH	78	return;
94b6da5a KH	79
52d4b9ac KH	80	for_each_online_node(nid) {
	81	fail = alloc_node_page_cgroup(nid);
	82	if (fail)
	83	goto fail;
	84	}
	85	printk(KERN_INFO "allocated %ld bytes of page_cgroup\n", total_usage);
8ca739e3 RD	86	printk(KERN_INFO "please try 'cgroup_disable=memory' option if you"
8ca739e3 RD	87	" don't want memory cgroups\n");
52d4b9ac KH	88	return;
52d4b9ac KH	89	fail:
8ca739e3 RD	90	printk(KERN_CRIT "allocation of page_cgroup failed.\n");
8ca739e3 RD	91	printk(KERN_CRIT "please try 'cgroup_disable=memory' boot option\n");
52d4b9ac KH	92	panic("Out of memory");
	93	}
	94
	95	#else /* CONFIG_FLAT_NODE_MEM_MAP */
	96
	97	struct page_cgroup lookup_page_cgroup(struct page page)
	98	{
	99	unsigned long pfn = page_to_pfn(page);
	100	struct mem_section *section = __pfn_to_section(pfn);
	101
d69b042f BS	102	if (!section->page_cgroup)
d69b042f BS	103	return NULL;
52d4b9ac KH	104	return section->page_cgroup + pfn;
	105	}
	106
31168481	107	/* __alloc_bootmem...() is protected by !slab_available() */
feb16694	108	static int __init_refok init_section_page_cgroup(unsigned long pfn)
52d4b9ac	109	{
0753b0ef	110	struct mem_section *section = __pfn_to_section(pfn);
52d4b9ac KH	111	struct page_cgroup base, pc;
	112	unsigned long table_size;
	113	int nid, index;
	114
dc19f9db KH	115	if (!section->page_cgroup) {
	116	nid = page_to_nid(pfn_to_page(pfn));
	117	table_size = sizeof(struct page_cgroup) * PAGES_PER_SECTION;
ca371c0d	118	VM_BUG_ON(!slab_is_available());
f52407ce SL	119	if (node_state(nid, N_HIGH_MEMORY)) {
f52407ce SL	120	base = kmalloc_node(table_size,
ca371c0d	121	GFP_KERNEL \| __GFP_NOWARN, nid);
f52407ce SL	122	if (!base)
	123	base = vmalloc_node(table_size, nid);
	124	} else {
	125	base = kmalloc(table_size, GFP_KERNEL \| __GFP_NOWARN);
	126	if (!base)
	127	base = vmalloc(table_size);
	128	}
dc19f9db KH	129	} else {
	130	/*
	131	* We don't have to allocate page_cgroup again, but
	132	* address of memmap may be changed. So, we have to initialize
	133	* again.
	134	*/
	135	base = section->page_cgroup + pfn;
	136	table_size = 0;
	137	/* check address of memmap is changed or not. */
	138	if (base->page == pfn_to_page(pfn))
	139	return 0;
94b6da5a	140	}
52d4b9ac KH	141
	142	if (!base) {
	143	printk(KERN_ERR "page cgroup allocation failure\n");
	144	return -ENOMEM;
	145	}
	146
	147	for (index = 0; index < PAGES_PER_SECTION; index++) {
	148	pc = base + index;
	149	__init_page_cgroup(pc, pfn + index);
	150	}
	151
52d4b9ac KH	152	section->page_cgroup = base - pfn;
	153	total_usage += table_size;
	154	return 0;
	155	}
	156	#ifdef CONFIG_MEMORY_HOTPLUG
	157	void __free_page_cgroup(unsigned long pfn)
	158	{
	159	struct mem_section *ms;
	160	struct page_cgroup *base;
	161
	162	ms = __pfn_to_section(pfn);
	163	if (!ms \|\| !ms->page_cgroup)
	164	return;
	165	base = ms->page_cgroup + pfn;
94b6da5a	166	if (is_vmalloc_addr(base)) {
52d4b9ac	167	vfree(base);
94b6da5a KH	168	ms->page_cgroup = NULL;
	169	} else {
	170	struct page *page = virt_to_page(base);
	171	if (!PageReserved(page)) { /* Is bootmem ? */
	172	kfree(base);
	173	ms->page_cgroup = NULL;
	174	}
	175	}
52d4b9ac KH	176	}
52d4b9ac KH	177
31168481	178	int __meminit online_page_cgroup(unsigned long start_pfn,
52d4b9ac KH	179	unsigned long nr_pages,
	180	int nid)
	181	{
	182	unsigned long start, end, pfn;
	183	int fail = 0;
	184
33c5d3d6	185	start = start_pfn & ~(PAGES_PER_SECTION - 1);
52d4b9ac KH	186	end = ALIGN(start_pfn + nr_pages, PAGES_PER_SECTION);
	187
	188	for (pfn = start; !fail && pfn < end; pfn += PAGES_PER_SECTION) {
	189	if (!pfn_present(pfn))
	190	continue;
	191	fail = init_section_page_cgroup(pfn);
	192	}
	193	if (!fail)
	194	return 0;
	195
	196	/* rollback */
	197	for (pfn = start; pfn < end; pfn += PAGES_PER_SECTION)
	198	__free_page_cgroup(pfn);
	199
	200	return -ENOMEM;
	201	}
	202
31168481	203	int __meminit offline_page_cgroup(unsigned long start_pfn,
52d4b9ac KH	204	unsigned long nr_pages, int nid)
	205	{
	206	unsigned long start, end, pfn;
	207
33c5d3d6	208	start = start_pfn & ~(PAGES_PER_SECTION - 1);
52d4b9ac KH	209	end = ALIGN(start_pfn + nr_pages, PAGES_PER_SECTION);
	210
	211	for (pfn = start; pfn < end; pfn += PAGES_PER_SECTION)
	212	__free_page_cgroup(pfn);
	213	return 0;
	214
	215	}
	216
31168481	217	static int __meminit page_cgroup_callback(struct notifier_block *self,
52d4b9ac KH	218	unsigned long action, void *arg)
	219	{
	220	struct memory_notify *mn = arg;
	221	int ret = 0;
	222	switch (action) {
	223	case MEM_GOING_ONLINE:
	224	ret = online_page_cgroup(mn->start_pfn,
	225	mn->nr_pages, mn->status_change_nid);
	226	break;
52d4b9ac KH	227	case MEM_OFFLINE:
	228	offline_page_cgroup(mn->start_pfn,
	229	mn->nr_pages, mn->status_change_nid);
	230	break;
dc19f9db	231	case MEM_CANCEL_ONLINE:
52d4b9ac KH	232	case MEM_GOING_OFFLINE:
	233	break;
	234	case MEM_ONLINE:
	235	case MEM_CANCEL_OFFLINE:
	236	break;
	237	}
dc19f9db KH	238
	239	if (ret)
	240	ret = notifier_from_errno(ret);
	241	else
	242	ret = NOTIFY_OK;
	243
52d4b9ac KH	244	return ret;
	245	}
	246
	247	#endif
	248
	249	void __init page_cgroup_init(void)
	250	{
	251	unsigned long pfn;
	252	int fail = 0;
	253
f8d66542	254	if (mem_cgroup_disabled())
94b6da5a KH	255	return;
94b6da5a KH	256
52d4b9ac KH	257	for (pfn = 0; !fail && pfn < max_pfn; pfn += PAGES_PER_SECTION) {
	258	if (!pfn_present(pfn))
	259	continue;
	260	fail = init_section_page_cgroup(pfn);
	261	}
	262	if (fail) {
8ca739e3	263	printk(KERN_CRIT "try 'cgroup_disable=memory' boot option\n");
52d4b9ac KH	264	panic("Out of memory");
	265	} else {
	266	hotplug_memory_notifier(page_cgroup_callback, 0);
	267	}
	268	printk(KERN_INFO "allocated %ld bytes of page_cgroup\n", total_usage);
8ca739e3 RD	269	printk(KERN_INFO "please try 'cgroup_disable=memory' option if you don't"
8ca739e3 RD	270	" want memory cgroups\n");
52d4b9ac KH	271	}
52d4b9ac KH	272
31168481	273	void __meminit pgdat_page_cgroup_init(struct pglist_data *pgdat)
52d4b9ac KH	274	{
	275	return;
	276	}
	277
	278	#endif
27a7faa0 KH	279
	280
	281	#ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP
	282
	283	static DEFINE_MUTEX(swap_cgroup_mutex);
	284	struct swap_cgroup_ctrl {
	285	struct page **map;
	286	unsigned long length;
e9e58a4e	287	spinlock_t lock;
27a7faa0 KH	288	};
	289
	290	struct swap_cgroup_ctrl swap_cgroup_ctrl[MAX_SWAPFILES];
	291
27a7faa0	292	struct swap_cgroup {
a3b2d692	293	unsigned short id;
27a7faa0 KH	294	};
	295	#define SC_PER_PAGE (PAGE_SIZE/sizeof(struct swap_cgroup))
	296	#define SC_POS_MASK (SC_PER_PAGE - 1)
	297
	298	/*
	299	* SwapCgroup implements "lookup" and "exchange" operations.
	300	* In typical usage, this swap_cgroup is accessed via memcg's charge/uncharge
	301	* against SwapCache. At swap_free(), this is accessed directly from swap.
	302	*
	303	* This means,
	304	* - we have no race in "exchange" when we're accessed via SwapCache because
	305	* SwapCache(and its swp_entry) is under lock.
	306	* - When called via swap_free(), there is no user of this entry and no race.
	307	* Then, we don't need lock around "exchange".
	308	*
	309	* TODO: we can push these buffers out to HIGHMEM.
	310	*/
	311
	312	/*
	313	* allocate buffer for swap_cgroup.
	314	*/
	315	static int swap_cgroup_prepare(int type)
	316	{
	317	struct page *page;
	318	struct swap_cgroup_ctrl *ctrl;
	319	unsigned long idx, max;
	320
27a7faa0 KH	321	ctrl = &swap_cgroup_ctrl[type];
	322
	323	for (idx = 0; idx < ctrl->length; idx++) {
	324	page = alloc_page(GFP_KERNEL \| __GFP_ZERO);
	325	if (!page)
	326	goto not_enough_page;
	327	ctrl->map[idx] = page;
	328	}
	329	return 0;
	330	not_enough_page:
	331	max = idx;
	332	for (idx = 0; idx < max; idx++)
	333	__free_page(ctrl->map[idx]);
	334
	335	return -ENOMEM;
	336	}
	337
02491447 DN	338	/**
	339	* swap_cgroup_cmpxchg - cmpxchg mem_cgroup's id for this swp_entry.
	340	* @end: swap entry to be cmpxchged
	341	* @old: old id
	342	* @new: new id
	343	*
	344	* Returns old id at success, 0 at failure.
	345	* (There is no mem_cgroup useing 0 as its id)
	346	*/
	347	unsigned short swap_cgroup_cmpxchg(swp_entry_t ent,
	348	unsigned short old, unsigned short new)
	349	{
	350	int type = swp_type(ent);
	351	unsigned long offset = swp_offset(ent);
	352	unsigned long idx = offset / SC_PER_PAGE;
	353	unsigned long pos = offset & SC_POS_MASK;
	354	struct swap_cgroup_ctrl *ctrl;
	355	struct page *mappage;
	356	struct swap_cgroup *sc;
e9e58a4e KH	357	unsigned long flags;
e9e58a4e KH	358	unsigned short retval;
02491447 DN	359
	360	ctrl = &swap_cgroup_ctrl[type];
	361
	362	mappage = ctrl->map[idx];
	363	sc = page_address(mappage);
	364	sc += pos;
e9e58a4e KH	365	spin_lock_irqsave(&ctrl->lock, flags);
	366	retval = sc->id;
	367	if (retval == old)
	368	sc->id = new;
02491447	369	else
e9e58a4e KH	370	retval = 0;
	371	spin_unlock_irqrestore(&ctrl->lock, flags);
	372	return retval;
02491447 DN	373	}
02491447 DN	374
27a7faa0 KH	375	/**
	376	* swap_cgroup_record - record mem_cgroup for this swp_entry.
	377	* @ent: swap entry to be recorded into
	378	* @mem: mem_cgroup to be recorded
	379	*
a3b2d692 KH	380	* Returns old value at success, 0 at failure.
a3b2d692 KH	381	* (Of course, old value can be 0.)
27a7faa0	382	*/
a3b2d692	383	unsigned short swap_cgroup_record(swp_entry_t ent, unsigned short id)
27a7faa0 KH	384	{
	385	int type = swp_type(ent);
	386	unsigned long offset = swp_offset(ent);
	387	unsigned long idx = offset / SC_PER_PAGE;
	388	unsigned long pos = offset & SC_POS_MASK;
	389	struct swap_cgroup_ctrl *ctrl;
	390	struct page *mappage;
	391	struct swap_cgroup *sc;
a3b2d692	392	unsigned short old;
e9e58a4e	393	unsigned long flags;
27a7faa0	394
27a7faa0 KH	395	ctrl = &swap_cgroup_ctrl[type];
	396
	397	mappage = ctrl->map[idx];
	398	sc = page_address(mappage);
	399	sc += pos;
e9e58a4e KH	400	spin_lock_irqsave(&ctrl->lock, flags);
	401	old = sc->id;
	402	sc->id = id;
	403	spin_unlock_irqrestore(&ctrl->lock, flags);
27a7faa0 KH	404
	405	return old;
	406	}
	407
	408	/**
	409	* lookup_swap_cgroup - lookup mem_cgroup tied to swap entry
	410	* @ent: swap entry to be looked up.
	411	*
a3b2d692	412	* Returns CSS ID of mem_cgroup at success. 0 at failure. (0 is invalid ID)
27a7faa0	413	*/
a3b2d692	414	unsigned short lookup_swap_cgroup(swp_entry_t ent)
27a7faa0 KH	415	{
	416	int type = swp_type(ent);
	417	unsigned long offset = swp_offset(ent);
	418	unsigned long idx = offset / SC_PER_PAGE;
	419	unsigned long pos = offset & SC_POS_MASK;
	420	struct swap_cgroup_ctrl *ctrl;
	421	struct page *mappage;
	422	struct swap_cgroup *sc;
a3b2d692	423	unsigned short ret;
27a7faa0	424
27a7faa0 KH	425	ctrl = &swap_cgroup_ctrl[type];
	426	mappage = ctrl->map[idx];
	427	sc = page_address(mappage);
	428	sc += pos;
a3b2d692	429	ret = sc->id;
27a7faa0 KH	430	return ret;
	431	}
	432
	433	int swap_cgroup_swapon(int type, unsigned long max_pages)
	434	{
	435	void *array;
	436	unsigned long array_size;
	437	unsigned long length;
	438	struct swap_cgroup_ctrl *ctrl;
	439
	440	if (!do_swap_account)
	441	return 0;
	442
	443	length = ((max_pages/SC_PER_PAGE) + 1);
	444	array_size = length * sizeof(void *);
	445
	446	array = vmalloc(array_size);
	447	if (!array)
	448	goto nomem;
	449
	450	memset(array, 0, array_size);
	451	ctrl = &swap_cgroup_ctrl[type];
	452	mutex_lock(&swap_cgroup_mutex);
	453	ctrl->length = length;
	454	ctrl->map = array;
e9e58a4e	455	spin_lock_init(&ctrl->lock);
27a7faa0 KH	456	if (swap_cgroup_prepare(type)) {
	457	/* memory shortage */
	458	ctrl->map = NULL;
	459	ctrl->length = 0;
	460	vfree(array);
	461	mutex_unlock(&swap_cgroup_mutex);
	462	goto nomem;
	463	}
	464	mutex_unlock(&swap_cgroup_mutex);
	465
27a7faa0 KH	466	return 0;
	467	nomem:
	468	printk(KERN_INFO "couldn't allocate enough memory for swap_cgroup.\n");
	469	printk(KERN_INFO
	470	"swap_cgroup can be disabled by noswapaccount boot option\n");
	471	return -ENOMEM;
	472	}
	473
	474	void swap_cgroup_swapoff(int type)
	475	{
	476	int i;
	477	struct swap_cgroup_ctrl *ctrl;
	478
	479	if (!do_swap_account)
	480	return;
	481
	482	mutex_lock(&swap_cgroup_mutex);
	483	ctrl = &swap_cgroup_ctrl[type];
	484	if (ctrl->map) {
	485	for (i = 0; i < ctrl->length; i++) {
	486	struct page *page = ctrl->map[i];
	487	if (page)
	488	__free_page(page);
	489	}
	490	vfree(ctrl->map);
	491	ctrl->map = NULL;
	492	ctrl->length = 0;
	493	}
	494	mutex_unlock(&swap_cgroup_mutex);
	495	}
	496
	497	#endif