memcontrol.h 28.2 KB
Newer Older
B
Balbir Singh 已提交
1 2 3 4 5
/* memcontrol.h - Memory Controller
 *
 * Copyright IBM Corporation, 2007
 * Author Balbir Singh <balbir@linux.vnet.ibm.com>
 *
6 7 8
 * Copyright 2007 OpenVZ SWsoft Inc
 * Author: Pavel Emelianov <xemul@openvz.org>
 *
B
Balbir Singh 已提交
9 10 11 12 13 14 15 16 17 18 19 20 21
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 */

#ifndef _LINUX_MEMCONTROL_H
#define _LINUX_MEMCONTROL_H
22
#include <linux/cgroup.h>
23
#include <linux/vm_event_item.h>
24
#include <linux/hardirq.h>
25
#include <linux/jump_label.h>
M
Michal Hocko 已提交
26 27 28
#include <linux/page_counter.h>
#include <linux/vmpressure.h>
#include <linux/eventfd.h>
29 30
#include <linux/mm.h>
#include <linux/vmstat.h>
M
Michal Hocko 已提交
31
#include <linux/writeback.h>
32
#include <linux/page-flags.h>
33

34
struct mem_cgroup;
35 36
struct page;
struct mm_struct;
37
struct kmem_cache;
38

39 40 41 42 43 44 45 46 47
/* Cgroup-specific page state, on top of universal node page state */
enum memcg_stat_item {
	MEMCG_CACHE = NR_VM_NODE_STAT_ITEMS,
	MEMCG_RSS,
	MEMCG_RSS_HUGE,
	MEMCG_SWAP,
	MEMCG_SOCK,
	/* XXX: why are these zone and not node counters? */
	MEMCG_KERNEL_STACK_KB,
48
	MEMCG_NR_STAT,
49 50
};

51 52 53 54 55 56 57 58 59
/* Cgroup-specific events, on top of universal VM events */
enum memcg_event_item {
	MEMCG_LOW = NR_VM_EVENT_ITEMS,
	MEMCG_HIGH,
	MEMCG_MAX,
	MEMCG_OOM,
	MEMCG_NR_EVENTS,
};

60
struct mem_cgroup_reclaim_cookie {
61
	pg_data_t *pgdat;
62 63 64 65
	int priority;
	unsigned int generation;
};

66 67 68 69 70 71 72 73 74 75
#ifdef CONFIG_MEMCG

#define MEM_CGROUP_ID_SHIFT	16
#define MEM_CGROUP_ID_MAX	USHRT_MAX

struct mem_cgroup_id {
	int id;
	atomic_t ref;
};

M
Michal Hocko 已提交
76 77 78 79 80 81 82 83 84 85 86 87 88 89
/*
 * Per memcg event counter is incremented at every pagein/pageout. With THP,
 * it will be incremated by the number of pages. This counter is used for
 * for trigger some periodic events. This is straightforward and better
 * than using jiffies etc. to handle periodic memcg event.
 */
enum mem_cgroup_events_target {
	MEM_CGROUP_TARGET_THRESH,
	MEM_CGROUP_TARGET_SOFTLIMIT,
	MEM_CGROUP_TARGET_NUMAINFO,
	MEM_CGROUP_NTARGETS,
};

struct mem_cgroup_stat_cpu {
90
	long count[MEMCG_NR_STAT];
M
Michal Hocko 已提交
91 92 93 94 95 96 97 98 99 100 101
	unsigned long events[MEMCG_NR_EVENTS];
	unsigned long nr_page_events;
	unsigned long targets[MEM_CGROUP_NTARGETS];
};

struct mem_cgroup_reclaim_iter {
	struct mem_cgroup *position;
	/* scan generation, increased every round-trip */
	unsigned int generation;
};

102 103 104 105
struct lruvec_stat {
	long count[NR_VM_NODE_STAT_ITEMS];
};

M
Michal Hocko 已提交
106 107 108
/*
 * per-zone information in memory controller.
 */
109
struct mem_cgroup_per_node {
M
Michal Hocko 已提交
110
	struct lruvec		lruvec;
111
	struct lruvec_stat __percpu *lruvec_stat;
112
	unsigned long		lru_zone_size[MAX_NR_ZONES][NR_LRU_LISTS];
M
Michal Hocko 已提交
113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149

	struct mem_cgroup_reclaim_iter	iter[DEF_PRIORITY + 1];

	struct rb_node		tree_node;	/* RB tree node */
	unsigned long		usage_in_excess;/* Set to the value by which */
						/* the soft limit is exceeded*/
	bool			on_tree;
	struct mem_cgroup	*memcg;		/* Back pointer, we cannot */
						/* use container_of	   */
};

struct mem_cgroup_threshold {
	struct eventfd_ctx *eventfd;
	unsigned long threshold;
};

/* For threshold */
struct mem_cgroup_threshold_ary {
	/* An array index points to threshold just below or equal to usage. */
	int current_threshold;
	/* Size of entries[] */
	unsigned int size;
	/* Array of thresholds */
	struct mem_cgroup_threshold entries[0];
};

struct mem_cgroup_thresholds {
	/* Primary thresholds array */
	struct mem_cgroup_threshold_ary *primary;
	/*
	 * Spare threshold array.
	 * This is needed to make mem_cgroup_unregister_event() "never fail".
	 * It must be able to store at least primary->size - 1 entries.
	 */
	struct mem_cgroup_threshold_ary *spare;
};

150 151 152 153 154 155
enum memcg_kmem_state {
	KMEM_NONE,
	KMEM_ALLOCATED,
	KMEM_ONLINE,
};

M
Michal Hocko 已提交
156 157 158 159 160 161 162 163 164
/*
 * The memory controller data structure. The memory controller controls both
 * page cache and RSS per cgroup. We would eventually like to provide
 * statistics based on the statistics developed by Rik Van Riel for clock-pro,
 * to help the administrator determine what knobs to tune.
 */
struct mem_cgroup {
	struct cgroup_subsys_state css;

165 166 167
	/* Private memcg ID. Used to ID objects that outlive the cgroup */
	struct mem_cgroup_id id;

M
Michal Hocko 已提交
168 169
	/* Accounted resources */
	struct page_counter memory;
170
	struct page_counter swap;
171 172

	/* Legacy consumer-oriented counters */
M
Michal Hocko 已提交
173 174
	struct page_counter memsw;
	struct page_counter kmem;
175
	struct page_counter tcpmem;
M
Michal Hocko 已提交
176 177 178 179 180

	/* Normal memory consumption range */
	unsigned long low;
	unsigned long high;

181 182 183
	/* Range enforcement for interrupt charges */
	struct work_struct high_work;

M
Michal Hocko 已提交
184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201
	unsigned long soft_limit;

	/* vmpressure notifications */
	struct vmpressure vmpressure;

	/*
	 * Should the accounting and control be hierarchical, per subtree?
	 */
	bool use_hierarchy;

	/* protected by memcg_oom_lock */
	bool		oom_lock;
	int		under_oom;

	int	swappiness;
	/* OOM-Killer disable */
	int		oom_kill_disable;

202 203 204
	/* handle for "memory.events" */
	struct cgroup_file events_file;

M
Michal Hocko 已提交
205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234
	/* protect arrays of thresholds */
	struct mutex thresholds_lock;

	/* thresholds for memory usage. RCU-protected */
	struct mem_cgroup_thresholds thresholds;

	/* thresholds for mem+swap usage. RCU-protected */
	struct mem_cgroup_thresholds memsw_thresholds;

	/* For oom notifier event fd */
	struct list_head oom_notify;

	/*
	 * Should we move charges of a task when a task is moved into this
	 * mem_cgroup ? And what type of charges should we move ?
	 */
	unsigned long move_charge_at_immigrate;
	/*
	 * set > 0 if pages under this cgroup are moving to other cgroup.
	 */
	atomic_t		moving_account;
	/* taken only while moving_account > 0 */
	spinlock_t		move_lock;
	struct task_struct	*move_lock_task;
	unsigned long		move_lock_flags;
	/*
	 * percpu counter.
	 */
	struct mem_cgroup_stat_cpu __percpu *stat;

235 236 237
	unsigned long		socket_pressure;

	/* Legacy tcp memory accounting */
238 239
	bool			tcpmem_active;
	int			tcpmem_pressure;
240

241
#ifndef CONFIG_SLOB
M
Michal Hocko 已提交
242 243
        /* Index in the kmem_cache->memcg_params.memcg_caches array */
	int kmemcg_id;
244
	enum memcg_kmem_state kmem_state;
245
	struct list_head kmem_caches;
M
Michal Hocko 已提交
246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266
#endif

	int last_scanned_node;
#if MAX_NUMNODES > 1
	nodemask_t	scan_nodes;
	atomic_t	numainfo_events;
	atomic_t	numainfo_updating;
#endif

#ifdef CONFIG_CGROUP_WRITEBACK
	struct list_head cgwb_list;
	struct wb_domain cgwb_domain;
#endif

	/* List of events which userspace want to receive */
	struct list_head event_list;
	spinlock_t event_list_lock;

	struct mem_cgroup_per_node *nodeinfo[0];
	/* WARNING: nodeinfo must be the last member here */
};
267 268

extern struct mem_cgroup *root_mem_cgroup;
T
Tejun Heo 已提交
269

270 271 272 273 274
static inline bool mem_cgroup_disabled(void)
{
	return !cgroup_subsys_enabled(memory_cgrp_subsys);
}

275 276
bool mem_cgroup_low(struct mem_cgroup *root, struct mem_cgroup *memcg);

277
int mem_cgroup_try_charge(struct page *page, struct mm_struct *mm,
278 279
			  gfp_t gfp_mask, struct mem_cgroup **memcgp,
			  bool compound);
280
void mem_cgroup_commit_charge(struct page *page, struct mem_cgroup *memcg,
281 282 283
			      bool lrucare, bool compound);
void mem_cgroup_cancel_charge(struct page *page, struct mem_cgroup *memcg,
		bool compound);
284
void mem_cgroup_uncharge(struct page *page);
285
void mem_cgroup_uncharge_list(struct list_head *page_list);
286

287
void mem_cgroup_migrate(struct page *oldpage, struct page *newpage);
288

289 290
static struct mem_cgroup_per_node *
mem_cgroup_nodeinfo(struct mem_cgroup *memcg, int nid)
291
{
292
	return memcg->nodeinfo[nid];
293 294 295
}

/**
296 297
 * mem_cgroup_lruvec - get the lru list vector for a node or a memcg zone
 * @node: node of the wanted lruvec
298 299
 * @memcg: memcg of the wanted lruvec
 *
300 301
 * Returns the lru list vector holding pages for a given @node or a given
 * @memcg and @zone. This can be the node lruvec, if the memory controller
302 303
 * is disabled.
 */
304
static inline struct lruvec *mem_cgroup_lruvec(struct pglist_data *pgdat,
305
				struct mem_cgroup *memcg)
306
{
307
	struct mem_cgroup_per_node *mz;
308 309 310
	struct lruvec *lruvec;

	if (mem_cgroup_disabled()) {
311
		lruvec = node_lruvec(pgdat);
312 313 314
		goto out;
	}

315
	mz = mem_cgroup_nodeinfo(memcg, pgdat->node_id);
316 317 318 319
	lruvec = &mz->lruvec;
out:
	/*
	 * Since a node can be onlined after the mem_cgroup was created,
M
Mel Gorman 已提交
320
	 * we have to be prepared to initialize lruvec->pgdat here;
321 322
	 * and if offlined then reonlined, we need to reinitialize it.
	 */
323 324
	if (unlikely(lruvec->pgdat != pgdat))
		lruvec->pgdat = pgdat;
325 326 327
	return lruvec;
}

M
Mel Gorman 已提交
328
struct lruvec *mem_cgroup_page_lruvec(struct page *, struct pglist_data *);
329

330
bool task_in_mem_cgroup(struct task_struct *task, struct mem_cgroup *memcg);
331
struct mem_cgroup *mem_cgroup_from_task(struct task_struct *p);
332

M
Michal Hocko 已提交
333 334 335 336 337
static inline
struct mem_cgroup *mem_cgroup_from_css(struct cgroup_subsys_state *css){
	return css ? container_of(css, struct mem_cgroup, css) : NULL;
}

338 339 340
#define mem_cgroup_from_counter(counter, member)	\
	container_of(counter, struct mem_cgroup, member)

M
Michal Hocko 已提交
341 342 343 344
struct mem_cgroup *mem_cgroup_iter(struct mem_cgroup *,
				   struct mem_cgroup *,
				   struct mem_cgroup_reclaim_cookie *);
void mem_cgroup_iter_break(struct mem_cgroup *, struct mem_cgroup *);
345 346
int mem_cgroup_scan_tasks(struct mem_cgroup *,
			  int (*)(struct task_struct *, void *), void *);
M
Michal Hocko 已提交
347

348 349 350 351 352
static inline unsigned short mem_cgroup_id(struct mem_cgroup *memcg)
{
	if (mem_cgroup_disabled())
		return 0;

353
	return memcg->id.id;
354
}
355
struct mem_cgroup *mem_cgroup_from_id(unsigned short id);
356

357 358 359 360 361 362 363 364 365 366 367
static inline struct mem_cgroup *lruvec_memcg(struct lruvec *lruvec)
{
	struct mem_cgroup_per_node *mz;

	if (mem_cgroup_disabled())
		return NULL;

	mz = container_of(lruvec, struct mem_cgroup_per_node, lruvec);
	return mz->memcg;
}

368 369 370 371 372 373 374 375 376 377 378 379 380 381
/**
 * parent_mem_cgroup - find the accounting parent of a memcg
 * @memcg: memcg whose parent to find
 *
 * Returns the parent memcg, or NULL if this is the root or the memory
 * controller is in legacy no-hierarchy mode.
 */
static inline struct mem_cgroup *parent_mem_cgroup(struct mem_cgroup *memcg)
{
	if (!memcg->memory.parent)
		return NULL;
	return mem_cgroup_from_counter(memcg->memory.parent, memory);
}

M
Michal Hocko 已提交
382 383 384 385 386 387 388 389 390
static inline bool mem_cgroup_is_descendant(struct mem_cgroup *memcg,
			      struct mem_cgroup *root)
{
	if (root == memcg)
		return true;
	if (!root->use_hierarchy)
		return false;
	return cgroup_is_descendant(memcg->css.cgroup, root->css.cgroup);
}
G
Glauber Costa 已提交
391

392 393
static inline bool mm_match_cgroup(struct mm_struct *mm,
				   struct mem_cgroup *memcg)
394
{
395
	struct mem_cgroup *task_memcg;
396
	bool match = false;
397

398
	rcu_read_lock();
399
	task_memcg = mem_cgroup_from_task(rcu_dereference(mm->owner));
400
	if (task_memcg)
401
		match = mem_cgroup_is_descendant(task_memcg, memcg);
402
	rcu_read_unlock();
403
	return match;
404
}
405

406
struct cgroup_subsys_state *mem_cgroup_css_from_page(struct page *page);
407
ino_t page_cgroup_ino(struct page *page);
408

409 410 411 412 413 414 415
static inline bool mem_cgroup_online(struct mem_cgroup *memcg)
{
	if (mem_cgroup_disabled())
		return true;
	return !!(memcg->css.flags & CSS_ONLINE);
}

416 417 418
/*
 * For memory reclaim.
 */
419
int mem_cgroup_select_victim_node(struct mem_cgroup *memcg);
M
Michal Hocko 已提交
420 421

void mem_cgroup_update_lru_size(struct lruvec *lruvec, enum lru_list lru,
422
		int zid, int nr_pages);
M
Michal Hocko 已提交
423

424 425 426
unsigned long mem_cgroup_node_nr_lru_pages(struct mem_cgroup *memcg,
					   int nid, unsigned int lru_mask);

M
Michal Hocko 已提交
427 428 429
static inline
unsigned long mem_cgroup_get_lru_size(struct lruvec *lruvec, enum lru_list lru)
{
430
	struct mem_cgroup_per_node *mz;
431 432
	unsigned long nr_pages = 0;
	int zid;
M
Michal Hocko 已提交
433

434
	mz = container_of(lruvec, struct mem_cgroup_per_node, lruvec);
435 436 437 438 439 440 441 442 443 444 445 446 447
	for (zid = 0; zid < MAX_NR_ZONES; zid++)
		nr_pages += mz->lru_zone_size[zid][lru];
	return nr_pages;
}

static inline
unsigned long mem_cgroup_get_zone_lru_size(struct lruvec *lruvec,
		enum lru_list lru, int zone_idx)
{
	struct mem_cgroup_per_node *mz;

	mz = container_of(lruvec, struct mem_cgroup_per_node, lruvec);
	return mz->lru_zone_size[zone_idx][lru];
M
Michal Hocko 已提交
448 449
}

450 451
void mem_cgroup_handle_over_high(void);

452 453
unsigned long mem_cgroup_get_limit(struct mem_cgroup *memcg);

454 455
void mem_cgroup_print_oom_info(struct mem_cgroup *memcg,
				struct task_struct *p);
456

457
static inline void mem_cgroup_oom_enable(void)
458
{
T
Tejun Heo 已提交
459 460
	WARN_ON(current->memcg_may_oom);
	current->memcg_may_oom = 1;
461 462
}

463
static inline void mem_cgroup_oom_disable(void)
464
{
T
Tejun Heo 已提交
465 466
	WARN_ON(!current->memcg_may_oom);
	current->memcg_may_oom = 0;
467 468
}

469 470
static inline bool task_in_memcg_oom(struct task_struct *p)
{
T
Tejun Heo 已提交
471
	return p->memcg_in_oom;
472 473
}

474
bool mem_cgroup_oom_synchronize(bool wait);
475

A
Andrew Morton 已提交
476
#ifdef CONFIG_MEMCG_SWAP
477 478
extern int do_swap_account;
#endif
479

480 481
struct mem_cgroup *lock_page_memcg(struct page *page);
void __unlock_page_memcg(struct mem_cgroup *memcg);
J
Johannes Weiner 已提交
482
void unlock_page_memcg(struct page *page);
483

484
/* idx can be of type enum memcg_stat_item or node_stat_item */
485
static inline unsigned long memcg_page_state(struct mem_cgroup *memcg,
486
					     int idx)
487 488 489 490 491 492 493 494 495 496 497 498 499
{
	long val = 0;
	int cpu;

	for_each_possible_cpu(cpu)
		val += per_cpu(memcg->stat->count[idx], cpu);

	if (val < 0)
		val = 0;

	return val;
}

500
/* idx can be of type enum memcg_stat_item or node_stat_item */
501
static inline void __mod_memcg_state(struct mem_cgroup *memcg,
502
				     int idx, int val)
503 504
{
	if (!mem_cgroup_disabled())
505
		__this_cpu_add(memcg->stat->count[idx], val);
506 507
}

508
/* idx can be of type enum memcg_stat_item or node_stat_item */
509
static inline void mod_memcg_state(struct mem_cgroup *memcg,
510
				   int idx, int val)
511
{
512 513
	if (!mem_cgroup_disabled())
		this_cpu_add(memcg->stat->count[idx], val);
514 515
}

M
Michal Hocko 已提交
516
/**
517
 * mod_memcg_page_state - update page state statistics
J
Johannes Weiner 已提交
518
 * @page: the page
M
Michal Hocko 已提交
519 520 521
 * @idx: page state item to account
 * @val: number of pages (positive or negative)
 *
522 523 524
 * The @page must be locked or the caller must use lock_page_memcg()
 * to prevent double accounting when the page is concurrently being
 * moved to another memcg:
525
 *
526
 *   lock_page(page) or lock_page_memcg(page)
527
 *   if (TestClearPageState(page))
528
 *     mod_memcg_page_state(page, state, -1);
529
 *   unlock_page(page) or unlock_page_memcg(page)
530 531
 *
 * Kernel pages are an exception to this, since they'll never move.
M
Michal Hocko 已提交
532
 */
533
static inline void __mod_memcg_page_state(struct page *page,
534
					  int idx, int val)
535 536 537 538 539
{
	if (page->mem_cgroup)
		__mod_memcg_state(page->mem_cgroup, idx, val);
}

540
static inline void mod_memcg_page_state(struct page *page,
541
					int idx, int val)
M
Michal Hocko 已提交
542
{
J
Johannes Weiner 已提交
543
	if (page->mem_cgroup)
544
		mod_memcg_state(page->mem_cgroup, idx, val);
M
Michal Hocko 已提交
545 546
}

547 548
static inline unsigned long lruvec_page_state(struct lruvec *lruvec,
					      enum node_stat_item idx)
549
{
550 551 552 553 554 555 556 557 558 559 560 561 562 563 564
	struct mem_cgroup_per_node *pn;
	long val = 0;
	int cpu;

	if (mem_cgroup_disabled())
		return node_page_state(lruvec_pgdat(lruvec), idx);

	pn = container_of(lruvec, struct mem_cgroup_per_node, lruvec);
	for_each_possible_cpu(cpu)
		val += per_cpu(pn->lruvec_stat->count[idx], cpu);

	if (val < 0)
		val = 0;

	return val;
565 566
}

567 568
static inline void __mod_lruvec_state(struct lruvec *lruvec,
				      enum node_stat_item idx, int val)
569
{
570 571
	struct mem_cgroup_per_node *pn;

572
	/* Update node */
573
	__mod_node_page_state(lruvec_pgdat(lruvec), idx, val);
574

575 576
	if (mem_cgroup_disabled())
		return;
577

578
	pn = container_of(lruvec, struct mem_cgroup_per_node, lruvec);
579 580

	/* Update memcg */
581
	__mod_memcg_state(pn->memcg, idx, val);
582 583

	/* Update lruvec */
584 585 586 587 588 589
	__this_cpu_add(pn->lruvec_stat->count[idx], val);
}

static inline void mod_lruvec_state(struct lruvec *lruvec,
				    enum node_stat_item idx, int val)
{
590 591 592
	preempt_disable();
	__mod_lruvec_state(lruvec, idx, val);
	preempt_enable();
593 594 595 596 597
}

static inline void __mod_lruvec_page_state(struct page *page,
					   enum node_stat_item idx, int val)
{
598 599
	pg_data_t *pgdat = page_pgdat(page);
	struct lruvec *lruvec;
600

601 602 603
	/* Untracked pages have no memcg, no lruvec. Update only the node */
	if (!page->mem_cgroup) {
		__mod_node_page_state(pgdat, idx, val);
604
		return;
605 606 607 608
	}

	lruvec = mem_cgroup_lruvec(pgdat, page->mem_cgroup);
	__mod_lruvec_state(lruvec, idx, val);
609 610 611 612 613
}

static inline void mod_lruvec_page_state(struct page *page,
					 enum node_stat_item idx, int val)
{
614 615 616
	preempt_disable();
	__mod_lruvec_page_state(page, idx, val);
	preempt_enable();
617 618
}

619
unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order,
620 621
						gfp_t gfp_mask,
						unsigned long *total_scanned);
D
David Rientjes 已提交
622

623 624 625 626 627 628 629 630 631
/* idx can be of type enum memcg_event_item or vm_event_item */
static inline void __count_memcg_events(struct mem_cgroup *memcg,
					int idx, unsigned long count)
{
	if (!mem_cgroup_disabled())
		__this_cpu_add(memcg->stat->events[idx], count);
}

/* idx can be of type enum memcg_event_item or vm_event_item */
632
static inline void count_memcg_events(struct mem_cgroup *memcg,
633
				      int idx, unsigned long count)
634 635 636 637 638
{
	if (!mem_cgroup_disabled())
		this_cpu_add(memcg->stat->events[idx], count);
}

639
/* idx can be of type enum memcg_event_item or vm_event_item */
640
static inline void count_memcg_page_event(struct page *page,
641
					  int idx)
642 643 644 645 646 647 648
{
	if (page->mem_cgroup)
		count_memcg_events(page->mem_cgroup, idx, 1);
}

static inline void count_memcg_event_mm(struct mm_struct *mm,
					enum vm_event_item idx)
649
{
M
Michal Hocko 已提交
650 651
	struct mem_cgroup *memcg;

652 653
	if (mem_cgroup_disabled())
		return;
M
Michal Hocko 已提交
654 655 656

	rcu_read_lock();
	memcg = mem_cgroup_from_task(rcu_dereference(mm->owner));
657
	if (likely(memcg)) {
658
		count_memcg_events(memcg, idx, 1);
659 660 661
		if (idx == OOM_KILL)
			cgroup_file_notify(&memcg->events_file);
	}
M
Michal Hocko 已提交
662
	rcu_read_unlock();
663
}
664 665 666 667 668 669 670 671

static inline void mem_cgroup_event(struct mem_cgroup *memcg,
				    enum memcg_event_item event)
{
	count_memcg_events(memcg, event, 1);
	cgroup_file_notify(&memcg->events_file);
}

672
#ifdef CONFIG_TRANSPARENT_HUGEPAGE
673
void mem_cgroup_split_huge_fixup(struct page *head);
674 675
#endif

A
Andrew Morton 已提交
676
#else /* CONFIG_MEMCG */
677 678 679 680

#define MEM_CGROUP_ID_SHIFT	0
#define MEM_CGROUP_ID_MAX	0

681 682
struct mem_cgroup;

683 684 685 686 687
static inline bool mem_cgroup_disabled(void)
{
	return true;
}

688
static inline void mem_cgroup_event(struct mem_cgroup *memcg,
689
				    enum memcg_event_item event)
690 691 692 693 694 695 696 697 698
{
}

static inline bool mem_cgroup_low(struct mem_cgroup *root,
				  struct mem_cgroup *memcg)
{
	return false;
}

699 700
static inline int mem_cgroup_try_charge(struct page *page, struct mm_struct *mm,
					gfp_t gfp_mask,
701 702
					struct mem_cgroup **memcgp,
					bool compound)
703
{
704
	*memcgp = NULL;
705 706 707
	return 0;
}

708 709
static inline void mem_cgroup_commit_charge(struct page *page,
					    struct mem_cgroup *memcg,
710
					    bool lrucare, bool compound)
711 712 713
{
}

714
static inline void mem_cgroup_cancel_charge(struct page *page,
715 716
					    struct mem_cgroup *memcg,
					    bool compound)
717 718 719
{
}

720
static inline void mem_cgroup_uncharge(struct page *page)
721 722 723
{
}

724
static inline void mem_cgroup_uncharge_list(struct list_head *page_list)
725 726 727
{
}

728
static inline void mem_cgroup_migrate(struct page *old, struct page *new)
729 730 731
{
}

732
static inline struct lruvec *mem_cgroup_lruvec(struct pglist_data *pgdat,
733
				struct mem_cgroup *memcg)
K
KAMEZAWA Hiroyuki 已提交
734
{
735
	return node_lruvec(pgdat);
K
KAMEZAWA Hiroyuki 已提交
736 737
}

738
static inline struct lruvec *mem_cgroup_page_lruvec(struct page *page,
M
Mel Gorman 已提交
739
						    struct pglist_data *pgdat)
740
{
M
Mel Gorman 已提交
741
	return &pgdat->lruvec;
742 743
}

744
static inline bool mm_match_cgroup(struct mm_struct *mm,
745
		struct mem_cgroup *memcg)
746
{
747
	return true;
748 749
}

750 751
static inline bool task_in_mem_cgroup(struct task_struct *task,
				      const struct mem_cgroup *memcg)
752
{
753
	return true;
754 755
}

756 757 758 759 760 761 762 763 764 765 766 767 768
static inline struct mem_cgroup *
mem_cgroup_iter(struct mem_cgroup *root,
		struct mem_cgroup *prev,
		struct mem_cgroup_reclaim_cookie *reclaim)
{
	return NULL;
}

static inline void mem_cgroup_iter_break(struct mem_cgroup *root,
					 struct mem_cgroup *prev)
{
}

769 770 771 772 773 774
static inline int mem_cgroup_scan_tasks(struct mem_cgroup *memcg,
		int (*fn)(struct task_struct *, void *), void *arg)
{
	return 0;
}

775
static inline unsigned short mem_cgroup_id(struct mem_cgroup *memcg)
776
{
777 778 779 780 781 782 783 784
	return 0;
}

static inline struct mem_cgroup *mem_cgroup_from_id(unsigned short id)
{
	WARN_ON_ONCE(id);
	/* XXX: This should always return root_mem_cgroup */
	return NULL;
785
}
786

787 788 789 790 791
static inline struct mem_cgroup *lruvec_memcg(struct lruvec *lruvec)
{
	return NULL;
}

792
static inline bool mem_cgroup_online(struct mem_cgroup *memcg)
793
{
794
	return true;
795 796
}

797
static inline unsigned long
798
mem_cgroup_get_lru_size(struct lruvec *lruvec, enum lru_list lru)
799 800 801
{
	return 0;
}
802 803 804 805 806 807
static inline
unsigned long mem_cgroup_get_zone_lru_size(struct lruvec *lruvec,
		enum lru_list lru, int zone_idx)
{
	return 0;
}
808

809 810 811 812 813 814 815
static inline unsigned long
mem_cgroup_node_nr_lru_pages(struct mem_cgroup *memcg,
			     int nid, unsigned int lru_mask)
{
	return 0;
}

816 817 818 819 820
static inline unsigned long mem_cgroup_get_limit(struct mem_cgroup *memcg)
{
	return 0;
}

821 822 823 824 825
static inline void
mem_cgroup_print_oom_info(struct mem_cgroup *memcg, struct task_struct *p)
{
}

826 827 828 829 830 831
static inline struct mem_cgroup *lock_page_memcg(struct page *page)
{
	return NULL;
}

static inline void __unlock_page_memcg(struct mem_cgroup *memcg)
832 833 834
{
}

J
Johannes Weiner 已提交
835
static inline void unlock_page_memcg(struct page *page)
836 837 838
{
}

839 840 841 842
static inline void mem_cgroup_handle_over_high(void)
{
}

843
static inline void mem_cgroup_oom_enable(void)
844 845 846
{
}

847
static inline void mem_cgroup_oom_disable(void)
848 849 850
{
}

851 852 853 854 855
static inline bool task_in_memcg_oom(struct task_struct *p)
{
	return false;
}

856
static inline bool mem_cgroup_oom_synchronize(bool wait)
857 858 859 860
{
	return false;
}

861
static inline unsigned long memcg_page_state(struct mem_cgroup *memcg,
862
					     int idx)
863 864 865 866
{
	return 0;
}

867
static inline void __mod_memcg_state(struct mem_cgroup *memcg,
868
				     int idx,
869
				     int nr)
870 871 872
{
}

873
static inline void mod_memcg_state(struct mem_cgroup *memcg,
874
				   int idx,
875
				   int nr)
876 877 878
{
}

879
static inline void __mod_memcg_page_state(struct page *page,
880
					  int idx,
881
					  int nr)
882 883 884
{
}

885
static inline void mod_memcg_page_state(struct page *page,
886
					int idx,
887
					int nr)
888 889 890
{
}

891 892
static inline unsigned long lruvec_page_state(struct lruvec *lruvec,
					      enum node_stat_item idx)
893
{
894
	return node_page_state(lruvec_pgdat(lruvec), idx);
895 896
}

897 898
static inline void __mod_lruvec_state(struct lruvec *lruvec,
				      enum node_stat_item idx, int val)
899
{
900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918
	__mod_node_page_state(lruvec_pgdat(lruvec), idx, val);
}

static inline void mod_lruvec_state(struct lruvec *lruvec,
				    enum node_stat_item idx, int val)
{
	mod_node_page_state(lruvec_pgdat(lruvec), idx, val);
}

static inline void __mod_lruvec_page_state(struct page *page,
					   enum node_stat_item idx, int val)
{
	__mod_node_page_state(page_pgdat(page), idx, val);
}

static inline void mod_lruvec_page_state(struct page *page,
					 enum node_stat_item idx, int val)
{
	mod_node_page_state(page_pgdat(page), idx, val);
919 920
}

921
static inline
922
unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order,
923 924
					    gfp_t gfp_mask,
					    unsigned long *total_scanned)
925
{
926
	return 0;
927 928
}

929
static inline void mem_cgroup_split_huge_fixup(struct page *head)
930 931 932
{
}

933 934 935 936 937 938 939
static inline void count_memcg_events(struct mem_cgroup *memcg,
				      enum vm_event_item idx,
				      unsigned long count)
{
}

static inline void count_memcg_page_event(struct page *page,
940
					  int idx)
941 942 943
{
}

944
static inline
945
void count_memcg_event_mm(struct mm_struct *mm, enum vm_event_item idx)
946 947
{
}
A
Andrew Morton 已提交
948
#endif /* CONFIG_MEMCG */
949

950
/* idx can be of type enum memcg_stat_item or node_stat_item */
951
static inline void __inc_memcg_state(struct mem_cgroup *memcg,
952
				     int idx)
953 954 955 956
{
	__mod_memcg_state(memcg, idx, 1);
}

957
/* idx can be of type enum memcg_stat_item or node_stat_item */
958
static inline void __dec_memcg_state(struct mem_cgroup *memcg,
959
				     int idx)
960 961 962 963
{
	__mod_memcg_state(memcg, idx, -1);
}

964
/* idx can be of type enum memcg_stat_item or node_stat_item */
965
static inline void __inc_memcg_page_state(struct page *page,
966
					  int idx)
967 968 969 970
{
	__mod_memcg_page_state(page, idx, 1);
}

971
/* idx can be of type enum memcg_stat_item or node_stat_item */
972
static inline void __dec_memcg_page_state(struct page *page,
973
					  int idx)
974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001
{
	__mod_memcg_page_state(page, idx, -1);
}

static inline void __inc_lruvec_state(struct lruvec *lruvec,
				      enum node_stat_item idx)
{
	__mod_lruvec_state(lruvec, idx, 1);
}

static inline void __dec_lruvec_state(struct lruvec *lruvec,
				      enum node_stat_item idx)
{
	__mod_lruvec_state(lruvec, idx, -1);
}

static inline void __inc_lruvec_page_state(struct page *page,
					   enum node_stat_item idx)
{
	__mod_lruvec_page_state(page, idx, 1);
}

static inline void __dec_lruvec_page_state(struct page *page,
					   enum node_stat_item idx)
{
	__mod_lruvec_page_state(page, idx, -1);
}

1002
/* idx can be of type enum memcg_stat_item or node_stat_item */
1003
static inline void inc_memcg_state(struct mem_cgroup *memcg,
1004
				   int idx)
1005 1006 1007 1008
{
	mod_memcg_state(memcg, idx, 1);
}

1009
/* idx can be of type enum memcg_stat_item or node_stat_item */
1010
static inline void dec_memcg_state(struct mem_cgroup *memcg,
1011
				   int idx)
1012 1013 1014 1015
{
	mod_memcg_state(memcg, idx, -1);
}

1016
/* idx can be of type enum memcg_stat_item or node_stat_item */
1017
static inline void inc_memcg_page_state(struct page *page,
1018
					int idx)
1019 1020 1021 1022
{
	mod_memcg_page_state(page, idx, 1);
}

1023
/* idx can be of type enum memcg_stat_item or node_stat_item */
1024
static inline void dec_memcg_page_state(struct page *page,
1025
					int idx)
1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053
{
	mod_memcg_page_state(page, idx, -1);
}

static inline void inc_lruvec_state(struct lruvec *lruvec,
				    enum node_stat_item idx)
{
	mod_lruvec_state(lruvec, idx, 1);
}

static inline void dec_lruvec_state(struct lruvec *lruvec,
				    enum node_stat_item idx)
{
	mod_lruvec_state(lruvec, idx, -1);
}

static inline void inc_lruvec_page_state(struct page *page,
					 enum node_stat_item idx)
{
	mod_lruvec_page_state(page, idx, 1);
}

static inline void dec_lruvec_page_state(struct page *page,
					 enum node_stat_item idx)
{
	mod_lruvec_page_state(page, idx, -1);
}

1054
#ifdef CONFIG_CGROUP_WRITEBACK
T
Tejun Heo 已提交
1055

1056
struct list_head *mem_cgroup_cgwb_list(struct mem_cgroup *memcg);
T
Tejun Heo 已提交
1057
struct wb_domain *mem_cgroup_wb_domain(struct bdi_writeback *wb);
1058 1059 1060
void mem_cgroup_wb_stats(struct bdi_writeback *wb, unsigned long *pfilepages,
			 unsigned long *pheadroom, unsigned long *pdirty,
			 unsigned long *pwriteback);
T
Tejun Heo 已提交
1061 1062 1063 1064 1065 1066 1067 1068

#else	/* CONFIG_CGROUP_WRITEBACK */

static inline struct wb_domain *mem_cgroup_wb_domain(struct bdi_writeback *wb)
{
	return NULL;
}

1069
static inline void mem_cgroup_wb_stats(struct bdi_writeback *wb,
1070 1071
				       unsigned long *pfilepages,
				       unsigned long *pheadroom,
1072 1073 1074 1075 1076
				       unsigned long *pdirty,
				       unsigned long *pwriteback)
{
}

T
Tejun Heo 已提交
1077
#endif	/* CONFIG_CGROUP_WRITEBACK */
1078

G
Glauber Costa 已提交
1079
struct sock;
1080 1081
bool mem_cgroup_charge_skmem(struct mem_cgroup *memcg, unsigned int nr_pages);
void mem_cgroup_uncharge_skmem(struct mem_cgroup *memcg, unsigned int nr_pages);
1082
#ifdef CONFIG_MEMCG
1083 1084
extern struct static_key_false memcg_sockets_enabled_key;
#define mem_cgroup_sockets_enabled static_branch_unlikely(&memcg_sockets_enabled_key)
1085 1086
void mem_cgroup_sk_alloc(struct sock *sk);
void mem_cgroup_sk_free(struct sock *sk);
1087
static inline bool mem_cgroup_under_socket_pressure(struct mem_cgroup *memcg)
1088
{
1089
	if (!cgroup_subsys_on_dfl(memory_cgrp_subsys) && memcg->tcpmem_pressure)
1090 1091 1092 1093 1094 1095
		return true;
	do {
		if (time_before(jiffies, memcg->socket_pressure))
			return true;
	} while ((memcg = parent_mem_cgroup(memcg)));
	return false;
1096 1097
}
#else
1098
#define mem_cgroup_sockets_enabled 0
1099 1100
static inline void mem_cgroup_sk_alloc(struct sock *sk) { };
static inline void mem_cgroup_sk_free(struct sock *sk) { };
1101
static inline bool mem_cgroup_under_socket_pressure(struct mem_cgroup *memcg)
1102 1103 1104 1105
{
	return false;
}
#endif
1106

1107 1108 1109 1110 1111 1112 1113
struct kmem_cache *memcg_kmem_get_cache(struct kmem_cache *cachep);
void memcg_kmem_put_cache(struct kmem_cache *cachep);
int memcg_kmem_charge_memcg(struct page *page, gfp_t gfp, int order,
			    struct mem_cgroup *memcg);
int memcg_kmem_charge(struct page *page, gfp_t gfp, int order);
void memcg_kmem_uncharge(struct page *page, int order);

1114
#if defined(CONFIG_MEMCG) && !defined(CONFIG_SLOB)
1115
extern struct static_key_false memcg_kmem_enabled_key;
1116
extern struct workqueue_struct *memcg_kmem_cache_wq;
1117

1118
extern int memcg_nr_cache_ids;
1119 1120
void memcg_get_cache_ids(void);
void memcg_put_cache_ids(void);
1121 1122 1123 1124 1125 1126

/*
 * Helper macro to loop through all memcg-specific caches. Callers must still
 * check if the cache is valid (it is either valid or NULL).
 * the slab_mutex must be held when looping through those caches
 */
1127
#define for_each_memcg_cache_index(_idx)	\
1128
	for ((_idx) = 0; (_idx) < memcg_nr_cache_ids; (_idx)++)
1129

1130 1131
static inline bool memcg_kmem_enabled(void)
{
1132
	return static_branch_unlikely(&memcg_kmem_enabled_key);
1133 1134
}

M
Michal Hocko 已提交
1135
/*
J
Jesper Dangaard Brouer 已提交
1136
 * helper for accessing a memcg's index. It will be used as an index in the
M
Michal Hocko 已提交
1137 1138 1139 1140 1141 1142 1143
 * child cache array in kmem_cache, and also to derive its name. This function
 * will return -1 when this is not a kmem-limited memcg.
 */
static inline int memcg_cache_id(struct mem_cgroup *memcg)
{
	return memcg ? memcg->kmemcg_id : -1;
}
1144

1145
#else
1146 1147 1148
#define for_each_memcg_cache_index(_idx)	\
	for (; NULL; )

1149 1150 1151 1152 1153
static inline bool memcg_kmem_enabled(void)
{
	return false;
}

1154 1155 1156 1157 1158
static inline int memcg_cache_id(struct mem_cgroup *memcg)
{
	return -1;
}

1159 1160 1161 1162 1163 1164 1165 1166
static inline void memcg_get_cache_ids(void)
{
}

static inline void memcg_put_cache_ids(void)
{
}

1167 1168
#endif /* CONFIG_MEMCG && !CONFIG_SLOB */

B
Balbir Singh 已提交
1169
#endif /* _LINUX_MEMCONTROL_H */