[mirror_ubuntu-jammy-kernel.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>
#include <linux/kmemleak.h>

static void __meminit init_page_cgroup(struct page_cgroup *pc, unsigned long id)
{
	pc->flags = 0;
	set_page_cgroup_array_id(pc, id);
	pc->mem_cgroup = NULL;
	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;
}

struct page *lookup_cgroup_page(struct page_cgroup *pc)
{
	unsigned long pfn;
	struct page *page;
	pg_data_t *pgdat;

	pgdat = NODE_DATA(page_cgroup_array_id(pc));
	pfn = pc - pgdat->node_page_cgroup + pgdat->node_start_pfn;
	page = pfn_to_page(pfn);
	VM_BUG_ON(pc != lookup_page_cgroup(page));
	return page;
}

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, nid);
	}
	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;
}

struct page *lookup_cgroup_page(struct page_cgroup *pc)
{
	struct mem_section *section;
	struct page *page;
	unsigned long nr;

	nr = page_cgroup_array_id(pc);
	section = __nr_to_section(nr);
	page = pfn_to_page(pc - section->page_cgroup);
	VM_BUG_ON(pc != lookup_page_cgroup(page));
	return page;
}

static void *__init_refok alloc_page_cgroup(size_t size, int nid)
{
	void *addr = NULL;

	addr = alloc_pages_exact(size, GFP_KERNEL | __GFP_NOWARN);
	if (addr)
		return addr;

	if (node_state(nid, N_HIGH_MEMORY))
		addr = vmalloc_node(size, nid);
	else
		addr = vmalloc(size);

	return addr;
}

#ifdef CONFIG_MEMORY_HOTPLUG
static void free_page_cgroup(void *addr)
{
	if (is_vmalloc_addr(addr)) {
		vfree(addr);
	} else {
		struct page *page = virt_to_page(addr);
		if (!PageReserved(page)) { /* Is bootmem ? */
			size_t table_size =
				sizeof(struct page_cgroup) * PAGES_PER_SECTION;
			free_pages_exact(addr, table_size);
		}
	}
}
#endif

static int __init_refok init_section_page_cgroup(unsigned long pfn)
{
	struct page_cgroup *base, *pc;
	struct mem_section *section;
	unsigned long table_size;
	unsigned long nr;
	int nid, index;

	nr = pfn_to_section_nr(pfn);
	section = __nr_to_section(nr);

	if (section->page_cgroup)
		return 0;

	nid = page_to_nid(pfn_to_page(pfn));
	table_size = sizeof(struct page_cgroup) * PAGES_PER_SECTION;
	base = alloc_page_cgroup(table_size, nid);

	/*
	 * The value stored in section->page_cgroup is (base - pfn)
	 * and it does not point to the memory block allocated above,
	 * causing kmemleak false positives.
	 */
	kmemleak_not_leak(base);

	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, nr);
	}

	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;
	free_page_cgroup(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;
	}

	return notifier_from_errno(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>
7952f988	12	#include <linux/kmemleak.h>
52d4b9ac	13
6b3ae58e	14	static void __meminit init_page_cgroup(struct page_cgroup *pc, unsigned long id)
52d4b9ac KH	15	{
52d4b9ac KH	16	pc->flags = 0;
6b3ae58e	17	set_page_cgroup_array_id(pc, id);
52d4b9ac	18	pc->mem_cgroup = NULL;
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
6b3ae58e JW	45	struct page lookup_cgroup_page(struct page_cgroup pc)
	46	{
	47	unsigned long pfn;
	48	struct page *page;
	49	pg_data_t *pgdat;
	50
	51	pgdat = NODE_DATA(page_cgroup_array_id(pc));
	52	pfn = pc - pgdat->node_page_cgroup + pgdat->node_start_pfn;
	53	page = pfn_to_page(pfn);
	54	VM_BUG_ON(pc != lookup_page_cgroup(page));
	55	return page;
	56	}
	57
52d4b9ac KH	58	static int __init alloc_node_page_cgroup(int nid)
	59	{
	60	struct page_cgroup base, pc;
	61	unsigned long table_size;
	62	unsigned long start_pfn, nr_pages, index;
	63
	64	start_pfn = NODE_DATA(nid)->node_start_pfn;
	65	nr_pages = NODE_DATA(nid)->node_spanned_pages;
	66
653d22c0 KH	67	if (!nr_pages)
	68	return 0;
	69
52d4b9ac	70	table_size = sizeof(struct page_cgroup) * nr_pages;
ca371c0d KH	71
	72	base = __alloc_bootmem_node_nopanic(NODE_DATA(nid),
	73	table_size, PAGE_SIZE, __pa(MAX_DMA_ADDRESS));
	74	if (!base)
52d4b9ac KH	75	return -ENOMEM;
	76	for (index = 0; index < nr_pages; index++) {
	77	pc = base + index;
6b3ae58e	78	init_page_cgroup(pc, nid);
52d4b9ac KH	79	}
	80	NODE_DATA(nid)->node_page_cgroup = base;
	81	total_usage += table_size;
	82	return 0;
	83	}
	84
ca371c0d	85	void __init page_cgroup_init_flatmem(void)
52d4b9ac KH	86	{
	87
	88	int nid, fail;
	89
f8d66542	90	if (mem_cgroup_disabled())
94b6da5a KH	91	return;
94b6da5a KH	92
52d4b9ac KH	93	for_each_online_node(nid) {
	94	fail = alloc_node_page_cgroup(nid);
	95	if (fail)
	96	goto fail;
	97	}
	98	printk(KERN_INFO "allocated %ld bytes of page_cgroup\n", total_usage);
8ca739e3 RD	99	printk(KERN_INFO "please try 'cgroup_disable=memory' option if you"
8ca739e3 RD	100	" don't want memory cgroups\n");
52d4b9ac KH	101	return;
52d4b9ac KH	102	fail:
8ca739e3 RD	103	printk(KERN_CRIT "allocation of page_cgroup failed.\n");
8ca739e3 RD	104	printk(KERN_CRIT "please try 'cgroup_disable=memory' boot option\n");
52d4b9ac KH	105	panic("Out of memory");
	106	}
	107
	108	#else /* CONFIG_FLAT_NODE_MEM_MAP */
	109
	110	struct page_cgroup lookup_page_cgroup(struct page page)
	111	{
	112	unsigned long pfn = page_to_pfn(page);
	113	struct mem_section *section = __pfn_to_section(pfn);
	114
d69b042f BS	115	if (!section->page_cgroup)
d69b042f BS	116	return NULL;
52d4b9ac KH	117	return section->page_cgroup + pfn;
	118	}
	119
6b3ae58e JW	120	struct page lookup_cgroup_page(struct page_cgroup pc)
	121	{
	122	struct mem_section *section;
	123	struct page *page;
	124	unsigned long nr;
	125
	126	nr = page_cgroup_array_id(pc);
	127	section = __nr_to_section(nr);
	128	page = pfn_to_page(pc - section->page_cgroup);
	129	VM_BUG_ON(pc != lookup_page_cgroup(page));
	130	return page;
	131	}
	132
dde79e00 MH	133	static void *__init_refok alloc_page_cgroup(size_t size, int nid)
	134	{
	135	void *addr = NULL;
	136
	137	addr = alloc_pages_exact(size, GFP_KERNEL \| __GFP_NOWARN);
	138	if (addr)
	139	return addr;
	140
	141	if (node_state(nid, N_HIGH_MEMORY))
	142	addr = vmalloc_node(size, nid);
	143	else
	144	addr = vmalloc(size);
	145
	146	return addr;
	147	}
	148
	149	#ifdef CONFIG_MEMORY_HOTPLUG
	150	static void free_page_cgroup(void *addr)
	151	{
	152	if (is_vmalloc_addr(addr)) {
	153	vfree(addr);
	154	} else {
	155	struct page *page = virt_to_page(addr);
	156	if (!PageReserved(page)) { /* Is bootmem ? */
	157	size_t table_size =
	158	sizeof(struct page_cgroup) * PAGES_PER_SECTION;
	159	free_pages_exact(addr, table_size);
	160	}
	161	}
	162	}
	163	#endif
	164
feb16694	165	static int __init_refok init_section_page_cgroup(unsigned long pfn)
52d4b9ac	166	{
52d4b9ac	167	struct page_cgroup base, pc;
6b3ae58e	168	struct mem_section *section;
52d4b9ac	169	unsigned long table_size;
6b3ae58e	170	unsigned long nr;
52d4b9ac KH	171	int nid, index;
52d4b9ac KH	172
6b3ae58e JW	173	nr = pfn_to_section_nr(pfn);
	174	section = __nr_to_section(nr);
	175
	176	if (section->page_cgroup)
	177	return 0;
	178
	179	nid = page_to_nid(pfn_to_page(pfn));
	180	table_size = sizeof(struct page_cgroup) * PAGES_PER_SECTION;
dde79e00 MH	181	base = alloc_page_cgroup(table_size, nid);
dde79e00 MH	182
6b3ae58e JW	183	/*
	184	* The value stored in section->page_cgroup is (base - pfn)
	185	* and it does not point to the memory block allocated above,
	186	* causing kmemleak false positives.
	187	*/
	188	kmemleak_not_leak(base);
52d4b9ac KH	189
	190	if (!base) {
	191	printk(KERN_ERR "page cgroup allocation failure\n");
	192	return -ENOMEM;
	193	}
	194
	195	for (index = 0; index < PAGES_PER_SECTION; index++) {
	196	pc = base + index;
6b3ae58e	197	init_page_cgroup(pc, nr);
52d4b9ac KH	198	}
52d4b9ac KH	199
52d4b9ac KH	200	section->page_cgroup = base - pfn;
	201	total_usage += table_size;
	202	return 0;
	203	}
	204	#ifdef CONFIG_MEMORY_HOTPLUG
	205	void __free_page_cgroup(unsigned long pfn)
	206	{
	207	struct mem_section *ms;
	208	struct page_cgroup *base;
	209
	210	ms = __pfn_to_section(pfn);
	211	if (!ms \|\| !ms->page_cgroup)
	212	return;
	213	base = ms->page_cgroup + pfn;
dde79e00 MH	214	free_page_cgroup(base);
dde79e00 MH	215	ms->page_cgroup = NULL;
52d4b9ac KH	216	}
52d4b9ac KH	217
31168481	218	int __meminit online_page_cgroup(unsigned long start_pfn,
52d4b9ac KH	219	unsigned long nr_pages,
	220	int nid)
	221	{
	222	unsigned long start, end, pfn;
	223	int fail = 0;
	224
33c5d3d6	225	start = start_pfn & ~(PAGES_PER_SECTION - 1);
52d4b9ac KH	226	end = ALIGN(start_pfn + nr_pages, PAGES_PER_SECTION);
	227
	228	for (pfn = start; !fail && pfn < end; pfn += PAGES_PER_SECTION) {
	229	if (!pfn_present(pfn))
	230	continue;
	231	fail = init_section_page_cgroup(pfn);
	232	}
	233	if (!fail)
	234	return 0;
	235
	236	/* rollback */
	237	for (pfn = start; pfn < end; pfn += PAGES_PER_SECTION)
	238	__free_page_cgroup(pfn);
	239
	240	return -ENOMEM;
	241	}
	242
31168481	243	int __meminit offline_page_cgroup(unsigned long start_pfn,
52d4b9ac KH	244	unsigned long nr_pages, int nid)
	245	{
	246	unsigned long start, end, pfn;
	247
33c5d3d6	248	start = start_pfn & ~(PAGES_PER_SECTION - 1);
52d4b9ac KH	249	end = ALIGN(start_pfn + nr_pages, PAGES_PER_SECTION);
	250
	251	for (pfn = start; pfn < end; pfn += PAGES_PER_SECTION)
	252	__free_page_cgroup(pfn);
	253	return 0;
	254
	255	}
	256
31168481	257	static int __meminit page_cgroup_callback(struct notifier_block *self,
52d4b9ac KH	258	unsigned long action, void *arg)
	259	{
	260	struct memory_notify *mn = arg;
	261	int ret = 0;
	262	switch (action) {
	263	case MEM_GOING_ONLINE:
	264	ret = online_page_cgroup(mn->start_pfn,
	265	mn->nr_pages, mn->status_change_nid);
	266	break;
52d4b9ac KH	267	case MEM_OFFLINE:
	268	offline_page_cgroup(mn->start_pfn,
	269	mn->nr_pages, mn->status_change_nid);
	270	break;
dc19f9db	271	case MEM_CANCEL_ONLINE:
52d4b9ac KH	272	case MEM_GOING_OFFLINE:
	273	break;
	274	case MEM_ONLINE:
	275	case MEM_CANCEL_OFFLINE:
	276	break;
	277	}
dc19f9db	278
5fda1bd5	279	return notifier_from_errno(ret);
52d4b9ac KH	280	}
	281
	282	#endif
	283
	284	void __init page_cgroup_init(void)
	285	{
	286	unsigned long pfn;
	287	int fail = 0;
	288
f8d66542	289	if (mem_cgroup_disabled())
94b6da5a KH	290	return;
94b6da5a KH	291
52d4b9ac KH	292	for (pfn = 0; !fail && pfn < max_pfn; pfn += PAGES_PER_SECTION) {
	293	if (!pfn_present(pfn))
	294	continue;
	295	fail = init_section_page_cgroup(pfn);
	296	}
	297	if (fail) {
8ca739e3	298	printk(KERN_CRIT "try 'cgroup_disable=memory' boot option\n");
52d4b9ac KH	299	panic("Out of memory");
	300	} else {
	301	hotplug_memory_notifier(page_cgroup_callback, 0);
	302	}
	303	printk(KERN_INFO "allocated %ld bytes of page_cgroup\n", total_usage);
8ca739e3 RD	304	printk(KERN_INFO "please try 'cgroup_disable=memory' option if you don't"
8ca739e3 RD	305	" want memory cgroups\n");
52d4b9ac KH	306	}
52d4b9ac KH	307
31168481	308	void __meminit pgdat_page_cgroup_init(struct pglist_data *pgdat)
52d4b9ac KH	309	{
	310	return;
	311	}
	312
	313	#endif
27a7faa0 KH	314
	315
	316	#ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP
	317
	318	static DEFINE_MUTEX(swap_cgroup_mutex);
	319	struct swap_cgroup_ctrl {
	320	struct page **map;
	321	unsigned long length;
e9e58a4e	322	spinlock_t lock;
27a7faa0 KH	323	};
	324
	325	struct swap_cgroup_ctrl swap_cgroup_ctrl[MAX_SWAPFILES];
	326
27a7faa0	327	struct swap_cgroup {
a3b2d692	328	unsigned short id;
27a7faa0 KH	329	};
	330	#define SC_PER_PAGE (PAGE_SIZE/sizeof(struct swap_cgroup))
	331	#define SC_POS_MASK (SC_PER_PAGE - 1)
	332
	333	/*
	334	* SwapCgroup implements "lookup" and "exchange" operations.
	335	* In typical usage, this swap_cgroup is accessed via memcg's charge/uncharge
	336	* against SwapCache. At swap_free(), this is accessed directly from swap.
	337	*
	338	* This means,
	339	* - we have no race in "exchange" when we're accessed via SwapCache because
	340	* SwapCache(and its swp_entry) is under lock.
	341	* - When called via swap_free(), there is no user of this entry and no race.
	342	* Then, we don't need lock around "exchange".
	343	*
	344	* TODO: we can push these buffers out to HIGHMEM.
	345	*/
	346
	347	/*
	348	* allocate buffer for swap_cgroup.
	349	*/
	350	static int swap_cgroup_prepare(int type)
	351	{
	352	struct page *page;
	353	struct swap_cgroup_ctrl *ctrl;
	354	unsigned long idx, max;
	355
27a7faa0 KH	356	ctrl = &swap_cgroup_ctrl[type];
	357
	358	for (idx = 0; idx < ctrl->length; idx++) {
	359	page = alloc_page(GFP_KERNEL \| __GFP_ZERO);
	360	if (!page)
	361	goto not_enough_page;
	362	ctrl->map[idx] = page;
	363	}
	364	return 0;
	365	not_enough_page:
	366	max = idx;
	367	for (idx = 0; idx < max; idx++)
	368	__free_page(ctrl->map[idx]);
	369
	370	return -ENOMEM;
	371	}
	372
02491447 DN	373	/**
	374	* swap_cgroup_cmpxchg - cmpxchg mem_cgroup's id for this swp_entry.
	375	* @end: swap entry to be cmpxchged
	376	* @old: old id
	377	* @new: new id
	378	*
	379	* Returns old id at success, 0 at failure.
	380	* (There is no mem_cgroup useing 0 as its id)
	381	*/
	382	unsigned short swap_cgroup_cmpxchg(swp_entry_t ent,
	383	unsigned short old, unsigned short new)
	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;
e9e58a4e KH	392	unsigned long flags;
e9e58a4e KH	393	unsigned short retval;
02491447 DN	394
	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	retval = sc->id;
	402	if (retval == old)
	403	sc->id = new;
02491447	404	else
e9e58a4e KH	405	retval = 0;
	406	spin_unlock_irqrestore(&ctrl->lock, flags);
	407	return retval;
02491447 DN	408	}
02491447 DN	409
27a7faa0 KH	410	/**
	411	* swap_cgroup_record - record mem_cgroup for this swp_entry.
	412	* @ent: swap entry to be recorded into
	413	* @mem: mem_cgroup to be recorded
	414	*
a3b2d692 KH	415	* Returns old value at success, 0 at failure.
a3b2d692 KH	416	* (Of course, old value can be 0.)
27a7faa0	417	*/
a3b2d692	418	unsigned short swap_cgroup_record(swp_entry_t ent, unsigned short id)
27a7faa0 KH	419	{
	420	int type = swp_type(ent);
	421	unsigned long offset = swp_offset(ent);
	422	unsigned long idx = offset / SC_PER_PAGE;
	423	unsigned long pos = offset & SC_POS_MASK;
	424	struct swap_cgroup_ctrl *ctrl;
	425	struct page *mappage;
	426	struct swap_cgroup *sc;
a3b2d692	427	unsigned short old;
e9e58a4e	428	unsigned long flags;
27a7faa0	429
27a7faa0 KH	430	ctrl = &swap_cgroup_ctrl[type];
	431
	432	mappage = ctrl->map[idx];
	433	sc = page_address(mappage);
	434	sc += pos;
e9e58a4e KH	435	spin_lock_irqsave(&ctrl->lock, flags);
	436	old = sc->id;
	437	sc->id = id;
	438	spin_unlock_irqrestore(&ctrl->lock, flags);
27a7faa0 KH	439
	440	return old;
	441	}
	442
	443	/**
	444	* lookup_swap_cgroup - lookup mem_cgroup tied to swap entry
	445	* @ent: swap entry to be looked up.
	446	*
a3b2d692	447	* Returns CSS ID of mem_cgroup at success. 0 at failure. (0 is invalid ID)
27a7faa0	448	*/
a3b2d692	449	unsigned short lookup_swap_cgroup(swp_entry_t ent)
27a7faa0 KH	450	{
	451	int type = swp_type(ent);
	452	unsigned long offset = swp_offset(ent);
	453	unsigned long idx = offset / SC_PER_PAGE;
	454	unsigned long pos = offset & SC_POS_MASK;
	455	struct swap_cgroup_ctrl *ctrl;
	456	struct page *mappage;
	457	struct swap_cgroup *sc;
a3b2d692	458	unsigned short ret;
27a7faa0	459
27a7faa0 KH	460	ctrl = &swap_cgroup_ctrl[type];
	461	mappage = ctrl->map[idx];
	462	sc = page_address(mappage);
	463	sc += pos;
a3b2d692	464	ret = sc->id;
27a7faa0 KH	465	return ret;
	466	}
	467
	468	int swap_cgroup_swapon(int type, unsigned long max_pages)
	469	{
	470	void *array;
	471	unsigned long array_size;
	472	unsigned long length;
	473	struct swap_cgroup_ctrl *ctrl;
	474
	475	if (!do_swap_account)
	476	return 0;
	477
	478	length = ((max_pages/SC_PER_PAGE) + 1);
	479	array_size = length * sizeof(void *);
	480
	481	array = vmalloc(array_size);
	482	if (!array)
	483	goto nomem;
	484
	485	memset(array, 0, array_size);
	486	ctrl = &swap_cgroup_ctrl[type];
	487	mutex_lock(&swap_cgroup_mutex);
	488	ctrl->length = length;
	489	ctrl->map = array;
e9e58a4e	490	spin_lock_init(&ctrl->lock);
27a7faa0 KH	491	if (swap_cgroup_prepare(type)) {
	492	/* memory shortage */
	493	ctrl->map = NULL;
	494	ctrl->length = 0;
	495	vfree(array);
	496	mutex_unlock(&swap_cgroup_mutex);
	497	goto nomem;
	498	}
	499	mutex_unlock(&swap_cgroup_mutex);
	500
27a7faa0 KH	501	return 0;
	502	nomem:
	503	printk(KERN_INFO "couldn't allocate enough memory for swap_cgroup.\n");
	504	printk(KERN_INFO
	505	"swap_cgroup can be disabled by noswapaccount boot option\n");
	506	return -ENOMEM;
	507	}
	508
	509	void swap_cgroup_swapoff(int type)
	510	{
	511	int i;
	512	struct swap_cgroup_ctrl *ctrl;
	513
	514	if (!do_swap_account)
	515	return;
	516
	517	mutex_lock(&swap_cgroup_mutex);
	518	ctrl = &swap_cgroup_ctrl[type];
	519	if (ctrl->map) {
	520	for (i = 0; i < ctrl->length; i++) {
	521	struct page *page = ctrl->map[i];
	522	if (page)
	523	__free_page(page);
	524	}
	525	vfree(ctrl->map);
	526	ctrl->map = NULL;
	527	ctrl->length = 0;
	528	}
	529	mutex_unlock(&swap_cgroup_mutex);
	530	}
	531
	532	#endif