memcontrol.h 32.1 KB
Newer Older
1
/* SPDX-License-Identifier: GPL-2.0-or-later */
B
Balbir Singh 已提交
2 3 4 5 6
/* memcontrol.h - Memory Controller
 *
 * Copyright IBM Corporation, 2007
 * Author Balbir Singh <balbir@linux.vnet.ibm.com>
 *
7 8
 * Copyright 2007 OpenVZ SWsoft Inc
 * Author: Pavel Emelianov <xemul@openvz.org>
B
Balbir Singh 已提交
9 10 11 12
 */

#ifndef _LINUX_MEMCONTROL_H
#define _LINUX_MEMCONTROL_H
13
#include <linux/cgroup.h>
14
#include <linux/vm_event_item.h>
15
#include <linux/hardirq.h>
16
#include <linux/jump_label.h>
M
Michal Hocko 已提交
17 18 19
#include <linux/page_counter.h>
#include <linux/vmpressure.h>
#include <linux/eventfd.h>
20 21
#include <linux/mm.h>
#include <linux/vmstat.h>
M
Michal Hocko 已提交
22
#include <linux/writeback.h>
23
#include <linux/page-flags.h>
24

25
struct mem_cgroup;
26 27
struct page;
struct mm_struct;
28
struct kmem_cache;
29

30 31 32 33 34 35 36 37 38
/* 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,
39
	MEMCG_NR_STAT,
40 41
};

42 43
enum memcg_memory_event {
	MEMCG_LOW,
44 45 46
	MEMCG_HIGH,
	MEMCG_MAX,
	MEMCG_OOM,
R
Roman Gushchin 已提交
47
	MEMCG_OOM_KILL,
48 49
	MEMCG_SWAP_MAX,
	MEMCG_SWAP_FAIL,
50
	MEMCG_NR_MEMORY_EVENTS,
51 52
};

R
Roman Gushchin 已提交
53 54 55 56 57 58
enum mem_cgroup_protection {
	MEMCG_PROT_NONE,
	MEMCG_PROT_LOW,
	MEMCG_PROT_MIN,
};

59
struct mem_cgroup_reclaim_cookie {
60
	pg_data_t *pgdat;
61 62 63 64
	int priority;
	unsigned int generation;
};

65 66 67 68 69 70 71
#ifdef CONFIG_MEMCG

#define MEM_CGROUP_ID_SHIFT	16
#define MEM_CGROUP_ID_MAX	USHRT_MAX

struct mem_cgroup_id {
	int id;
72
	refcount_t ref;
73 74
};

M
Michal Hocko 已提交
75 76 77 78 79 80 81 82 83 84 85 86 87
/*
 * 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,
};

88 89
struct memcg_vmstats_percpu {
	long stat[MEMCG_NR_STAT];
90
	unsigned long events[NR_VM_EVENT_ITEMS];
M
Michal Hocko 已提交
91 92 93 94 95 96 97 98 99 100
	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;
};

101 102 103 104
struct lruvec_stat {
	long count[NR_VM_NODE_STAT_ITEMS];
};

105 106 107 108 109 110 111 112 113
/*
 * Bitmap of shrinker::id corresponding to memcg-aware shrinkers,
 * which have elements charged to this memcg.
 */
struct memcg_shrinker_map {
	struct rcu_head rcu;
	unsigned long map[0];
};

M
Michal Hocko 已提交
114 115 116
/*
 * per-zone information in memory controller.
 */
117
struct mem_cgroup_per_node {
M
Michal Hocko 已提交
118
	struct lruvec		lruvec;
119 120 121

	struct lruvec_stat __percpu *lruvec_stat_cpu;
	atomic_long_t		lruvec_stat[NR_VM_NODE_STAT_ITEMS];
122
	atomic_long_t		lruvec_stat_local[NR_VM_NODE_STAT_ITEMS];
123

124
	unsigned long		lru_zone_size[MAX_NR_ZONES][NR_LRU_LISTS];
M
Michal Hocko 已提交
125 126 127

	struct mem_cgroup_reclaim_iter	iter[DEF_PRIORITY + 1];

128 129 130
#ifdef CONFIG_MEMCG_KMEM
	struct memcg_shrinker_map __rcu	*shrinker_map;
#endif
M
Michal Hocko 已提交
131 132 133 134
	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;
135 136 137
	bool			congested;	/* memcg has many dirty pages */
						/* backed by a congested BDI */

M
Michal Hocko 已提交
138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167
	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;
};

168 169 170 171 172 173
enum memcg_kmem_state {
	KMEM_NONE,
	KMEM_ALLOCATED,
	KMEM_ONLINE,
};

174 175 176 177 178 179 180 181 182
#if defined(CONFIG_SMP)
struct memcg_padding {
	char x[0];
} ____cacheline_internodealigned_in_smp;
#define MEMCG_PADDING(name)      struct memcg_padding name;
#else
#define MEMCG_PADDING(name)
#endif

M
Michal Hocko 已提交
183 184 185 186 187 188 189 190 191
/*
 * 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;

192 193 194
	/* Private memcg ID. Used to ID objects that outlive the cgroup */
	struct mem_cgroup_id id;

M
Michal Hocko 已提交
195 196
	/* Accounted resources */
	struct page_counter memory;
197
	struct page_counter swap;
198 199

	/* Legacy consumer-oriented counters */
M
Michal Hocko 已提交
200 201
	struct page_counter memsw;
	struct page_counter kmem;
202
	struct page_counter tcpmem;
M
Michal Hocko 已提交
203

204
	/* Upper bound of normal memory consumption range */
M
Michal Hocko 已提交
205 206
	unsigned long high;

207 208 209
	/* Range enforcement for interrupt charges */
	struct work_struct high_work;

M
Michal Hocko 已提交
210 211 212 213 214 215 216 217 218 219
	unsigned long soft_limit;

	/* vmpressure notifications */
	struct vmpressure vmpressure;

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

220 221 222 223 224
	/*
	 * Should the OOM killer kill all belonging tasks, had it kill one?
	 */
	bool oom_group;

M
Michal Hocko 已提交
225 226 227 228 229 230 231 232
	/* protected by memcg_oom_lock */
	bool		oom_lock;
	int		under_oom;

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

233
	/* memory.events */
234 235
	struct cgroup_file events_file;

236 237 238
	/* handle for "memory.swap.events" */
	struct cgroup_file swap_events_file;

M
Michal Hocko 已提交
239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255
	/* 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;
256 257 258 259 260 261
	/* taken only while moving_account > 0 */
	spinlock_t		move_lock;
	unsigned long		move_lock_flags;

	MEMCG_PADDING(_pad1_);

M
Michal Hocko 已提交
262 263 264 265 266
	/*
	 * set > 0 if pages under this cgroup are moving to other cgroup.
	 */
	atomic_t		moving_account;
	struct task_struct	*move_lock_task;
267

268
	/* memory.stat */
269
	struct memcg_vmstats_percpu __percpu *vmstats_percpu;
270 271 272

	MEMCG_PADDING(_pad2_);

273
	atomic_long_t		vmstats[MEMCG_NR_STAT];
274 275
	atomic_long_t		vmstats_local[MEMCG_NR_STAT];

276
	atomic_long_t		vmevents[NR_VM_EVENT_ITEMS];
277 278 279
	atomic_long_t		vmevents_local[NR_VM_EVENT_ITEMS];

	atomic_long_t		memory_events[MEMCG_NR_MEMORY_EVENTS];
M
Michal Hocko 已提交
280

281 282 283
	unsigned long		socket_pressure;

	/* Legacy tcp memory accounting */
284 285
	bool			tcpmem_active;
	int			tcpmem_pressure;
286

287
#ifdef CONFIG_MEMCG_KMEM
M
Michal Hocko 已提交
288 289
        /* Index in the kmem_cache->memcg_params.memcg_caches array */
	int kmemcg_id;
290
	enum memcg_kmem_state kmem_state;
291
	struct list_head kmem_caches;
M
Michal Hocko 已提交
292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312
#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 */
};
313

314 315 316 317 318 319
/*
 * size of first charge trial. "32" comes from vmscan.c's magic value.
 * TODO: maybe necessary to use big numbers in big irons.
 */
#define MEMCG_CHARGE_BATCH 32U

320
extern struct mem_cgroup *root_mem_cgroup;
T
Tejun Heo 已提交
321

322 323 324 325 326
static inline bool mem_cgroup_is_root(struct mem_cgroup *memcg)
{
	return (memcg == root_mem_cgroup);
}

327 328 329 330 331
static inline bool mem_cgroup_disabled(void)
{
	return !cgroup_subsys_enabled(memory_cgrp_subsys);
}

R
Roman Gushchin 已提交
332 333
enum mem_cgroup_protection mem_cgroup_protected(struct mem_cgroup *root,
						struct mem_cgroup *memcg);
334

335
int mem_cgroup_try_charge(struct page *page, struct mm_struct *mm,
336 337
			  gfp_t gfp_mask, struct mem_cgroup **memcgp,
			  bool compound);
338 339 340
int mem_cgroup_try_charge_delay(struct page *page, struct mm_struct *mm,
			  gfp_t gfp_mask, struct mem_cgroup **memcgp,
			  bool compound);
341
void mem_cgroup_commit_charge(struct page *page, struct mem_cgroup *memcg,
342 343 344
			      bool lrucare, bool compound);
void mem_cgroup_cancel_charge(struct page *page, struct mem_cgroup *memcg,
		bool compound);
345
void mem_cgroup_uncharge(struct page *page);
346
void mem_cgroup_uncharge_list(struct list_head *page_list);
347

348
void mem_cgroup_migrate(struct page *oldpage, struct page *newpage);
349

350 351
static struct mem_cgroup_per_node *
mem_cgroup_nodeinfo(struct mem_cgroup *memcg, int nid)
352
{
353
	return memcg->nodeinfo[nid];
354 355 356
}

/**
357 358
 * mem_cgroup_lruvec - get the lru list vector for a node or a memcg zone
 * @node: node of the wanted lruvec
359 360
 * @memcg: memcg of the wanted lruvec
 *
361 362
 * 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
363 364
 * is disabled.
 */
365
static inline struct lruvec *mem_cgroup_lruvec(struct pglist_data *pgdat,
366
				struct mem_cgroup *memcg)
367
{
368
	struct mem_cgroup_per_node *mz;
369 370 371
	struct lruvec *lruvec;

	if (mem_cgroup_disabled()) {
372
		lruvec = node_lruvec(pgdat);
373 374 375
		goto out;
	}

376
	mz = mem_cgroup_nodeinfo(memcg, pgdat->node_id);
377 378 379 380
	lruvec = &mz->lruvec;
out:
	/*
	 * Since a node can be onlined after the mem_cgroup was created,
M
Mel Gorman 已提交
381
	 * we have to be prepared to initialize lruvec->pgdat here;
382 383
	 * and if offlined then reonlined, we need to reinitialize it.
	 */
384 385
	if (unlikely(lruvec->pgdat != pgdat))
		lruvec->pgdat = pgdat;
386 387 388
	return lruvec;
}

M
Mel Gorman 已提交
389
struct lruvec *mem_cgroup_page_lruvec(struct page *, struct pglist_data *);
390

391
bool task_in_mem_cgroup(struct task_struct *task, struct mem_cgroup *memcg);
392
struct mem_cgroup *mem_cgroup_from_task(struct task_struct *p);
393

394 395
struct mem_cgroup *get_mem_cgroup_from_mm(struct mm_struct *mm);

396 397
struct mem_cgroup *get_mem_cgroup_from_page(struct page *page);

M
Michal Hocko 已提交
398 399 400 401 402
static inline
struct mem_cgroup *mem_cgroup_from_css(struct cgroup_subsys_state *css){
	return css ? container_of(css, struct mem_cgroup, css) : NULL;
}

403 404
static inline void mem_cgroup_put(struct mem_cgroup *memcg)
{
405 406
	if (memcg)
		css_put(&memcg->css);
407 408
}

409 410 411
#define mem_cgroup_from_counter(counter, member)	\
	container_of(counter, struct mem_cgroup, member)

M
Michal Hocko 已提交
412 413 414 415
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 *);
416 417
int mem_cgroup_scan_tasks(struct mem_cgroup *,
			  int (*)(struct task_struct *, void *), void *);
M
Michal Hocko 已提交
418

419 420 421 422 423
static inline unsigned short mem_cgroup_id(struct mem_cgroup *memcg)
{
	if (mem_cgroup_disabled())
		return 0;

424
	return memcg->id.id;
425
}
426
struct mem_cgroup *mem_cgroup_from_id(unsigned short id);
427

428 429 430 431 432
static inline struct mem_cgroup *mem_cgroup_from_seq(struct seq_file *m)
{
	return mem_cgroup_from_css(seq_css(m));
}

433 434 435 436 437 438 439 440 441 442 443
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;
}

444 445 446 447 448 449 450 451 452 453 454 455 456 457
/**
 * 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
458 459 460 461 462 463 464 465 466
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
467

468 469
static inline bool mm_match_cgroup(struct mm_struct *mm,
				   struct mem_cgroup *memcg)
470
{
471
	struct mem_cgroup *task_memcg;
472
	bool match = false;
473

474
	rcu_read_lock();
475
	task_memcg = mem_cgroup_from_task(rcu_dereference(mm->owner));
476
	if (task_memcg)
477
		match = mem_cgroup_is_descendant(task_memcg, memcg);
478
	rcu_read_unlock();
479
	return match;
480
}
481

482
struct cgroup_subsys_state *mem_cgroup_css_from_page(struct page *page);
483
ino_t page_cgroup_ino(struct page *page);
484

485 486 487 488 489 490 491
static inline bool mem_cgroup_online(struct mem_cgroup *memcg)
{
	if (mem_cgroup_disabled())
		return true;
	return !!(memcg->css.flags & CSS_ONLINE);
}

492 493 494
/*
 * For memory reclaim.
 */
495
int mem_cgroup_select_victim_node(struct mem_cgroup *memcg);
M
Michal Hocko 已提交
496 497

void mem_cgroup_update_lru_size(struct lruvec *lruvec, enum lru_list lru,
498
		int zid, int nr_pages);
M
Michal Hocko 已提交
499

500 501 502 503 504 505 506 507
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
508 509
}

510 511
void mem_cgroup_handle_over_high(void);

512
unsigned long mem_cgroup_get_max(struct mem_cgroup *memcg);
513

514
void mem_cgroup_print_oom_context(struct mem_cgroup *memcg,
515
				struct task_struct *p);
516

517 518
void mem_cgroup_print_oom_meminfo(struct mem_cgroup *memcg);

519
static inline void mem_cgroup_enter_user_fault(void)
520
{
521 522
	WARN_ON(current->in_user_fault);
	current->in_user_fault = 1;
523 524
}

525
static inline void mem_cgroup_exit_user_fault(void)
526
{
527 528
	WARN_ON(!current->in_user_fault);
	current->in_user_fault = 0;
529 530
}

531 532
static inline bool task_in_memcg_oom(struct task_struct *p)
{
T
Tejun Heo 已提交
533
	return p->memcg_in_oom;
534 535
}

536
bool mem_cgroup_oom_synchronize(bool wait);
537 538 539
struct mem_cgroup *mem_cgroup_get_oom_group(struct task_struct *victim,
					    struct mem_cgroup *oom_domain);
void mem_cgroup_print_oom_group(struct mem_cgroup *memcg);
540

A
Andrew Morton 已提交
541
#ifdef CONFIG_MEMCG_SWAP
542 543
extern int do_swap_account;
#endif
544

545 546
struct mem_cgroup *lock_page_memcg(struct page *page);
void __unlock_page_memcg(struct mem_cgroup *memcg);
J
Johannes Weiner 已提交
547
void unlock_page_memcg(struct page *page);
548

549 550 551 552 553 554 555 556 557 558 559 560 561 562
/*
 * idx can be of type enum memcg_stat_item or node_stat_item.
 * Keep in sync with memcg_exact_page_state().
 */
static inline unsigned long memcg_page_state(struct mem_cgroup *memcg, int idx)
{
	long x = atomic_long_read(&memcg->vmstats[idx]);
#ifdef CONFIG_SMP
	if (x < 0)
		x = 0;
#endif
	return x;
}

563 564 565 566
/*
 * idx can be of type enum memcg_stat_item or node_stat_item.
 * Keep in sync with memcg_exact_page_state().
 */
567 568
static inline unsigned long memcg_page_state_local(struct mem_cgroup *memcg,
						   int idx)
569
{
570
	long x = atomic_long_read(&memcg->vmstats_local[idx]);
571 572 573 574 575
#ifdef CONFIG_SMP
	if (x < 0)
		x = 0;
#endif
	return x;
576 577
}

578
void __mod_memcg_state(struct mem_cgroup *memcg, int idx, int val);
579

580
/* idx can be of type enum memcg_stat_item or node_stat_item */
581
static inline void mod_memcg_state(struct mem_cgroup *memcg,
582
				   int idx, int val)
583
{
584 585 586
	unsigned long flags;

	local_irq_save(flags);
587
	__mod_memcg_state(memcg, idx, val);
588
	local_irq_restore(flags);
589 590
}

M
Michal Hocko 已提交
591
/**
592
 * mod_memcg_page_state - update page state statistics
J
Johannes Weiner 已提交
593
 * @page: the page
M
Michal Hocko 已提交
594 595 596
 * @idx: page state item to account
 * @val: number of pages (positive or negative)
 *
597 598 599
 * 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:
600
 *
601
 *   lock_page(page) or lock_page_memcg(page)
602
 *   if (TestClearPageState(page))
603
 *     mod_memcg_page_state(page, state, -1);
604
 *   unlock_page(page) or unlock_page_memcg(page)
605 606
 *
 * Kernel pages are an exception to this, since they'll never move.
M
Michal Hocko 已提交
607
 */
608
static inline void __mod_memcg_page_state(struct page *page,
609
					  int idx, int val)
610 611 612 613 614
{
	if (page->mem_cgroup)
		__mod_memcg_state(page->mem_cgroup, idx, val);
}

615
static inline void mod_memcg_page_state(struct page *page,
616
					int idx, int val)
M
Michal Hocko 已提交
617
{
J
Johannes Weiner 已提交
618
	if (page->mem_cgroup)
619
		mod_memcg_state(page->mem_cgroup, idx, val);
M
Michal Hocko 已提交
620 621
}

622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639
static inline unsigned long lruvec_page_state(struct lruvec *lruvec,
					      enum node_stat_item idx)
{
	struct mem_cgroup_per_node *pn;
	long x;

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

	pn = container_of(lruvec, struct mem_cgroup_per_node, lruvec);
	x = atomic_long_read(&pn->lruvec_stat[idx]);
#ifdef CONFIG_SMP
	if (x < 0)
		x = 0;
#endif
	return x;
}

640 641
static inline unsigned long lruvec_page_state_local(struct lruvec *lruvec,
						    enum node_stat_item idx)
642
{
643
	struct mem_cgroup_per_node *pn;
644
	long x;
645 646 647 648 649

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

	pn = container_of(lruvec, struct mem_cgroup_per_node, lruvec);
650
	x = atomic_long_read(&pn->lruvec_stat_local[idx]);
651 652 653 654 655
#ifdef CONFIG_SMP
	if (x < 0)
		x = 0;
#endif
	return x;
656 657
}

658 659
void __mod_lruvec_state(struct lruvec *lruvec, enum node_stat_item idx,
			int val);
660 661 662 663

static inline void mod_lruvec_state(struct lruvec *lruvec,
				    enum node_stat_item idx, int val)
{
664 665 666
	unsigned long flags;

	local_irq_save(flags);
667
	__mod_lruvec_state(lruvec, idx, val);
668
	local_irq_restore(flags);
669 670 671 672 673
}

static inline void __mod_lruvec_page_state(struct page *page,
					   enum node_stat_item idx, int val)
{
674 675
	pg_data_t *pgdat = page_pgdat(page);
	struct lruvec *lruvec;
676

677 678 679
	/* Untracked pages have no memcg, no lruvec. Update only the node */
	if (!page->mem_cgroup) {
		__mod_node_page_state(pgdat, idx, val);
680
		return;
681 682 683 684
	}

	lruvec = mem_cgroup_lruvec(pgdat, page->mem_cgroup);
	__mod_lruvec_state(lruvec, idx, val);
685 686 687 688 689
}

static inline void mod_lruvec_page_state(struct page *page,
					 enum node_stat_item idx, int val)
{
690 691 692
	unsigned long flags;

	local_irq_save(flags);
693
	__mod_lruvec_page_state(page, idx, val);
694
	local_irq_restore(flags);
695 696
}

697
unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order,
698 699
						gfp_t gfp_mask,
						unsigned long *total_scanned);
D
David Rientjes 已提交
700

701 702
void __count_memcg_events(struct mem_cgroup *memcg, enum vm_event_item idx,
			  unsigned long count);
703

704
static inline void count_memcg_events(struct mem_cgroup *memcg,
705 706
				      enum vm_event_item idx,
				      unsigned long count)
707
{
708 709 710
	unsigned long flags;

	local_irq_save(flags);
711
	__count_memcg_events(memcg, idx, count);
712
	local_irq_restore(flags);
713 714 715
}

static inline void count_memcg_page_event(struct page *page,
716
					  enum vm_event_item idx)
717 718 719 720 721 722 723
{
	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)
724
{
M
Michal Hocko 已提交
725 726
	struct mem_cgroup *memcg;

727 728
	if (mem_cgroup_disabled())
		return;
M
Michal Hocko 已提交
729 730 731

	rcu_read_lock();
	memcg = mem_cgroup_from_task(rcu_dereference(mm->owner));
R
Roman Gushchin 已提交
732
	if (likely(memcg))
733
		count_memcg_events(memcg, idx, 1);
M
Michal Hocko 已提交
734
	rcu_read_unlock();
735
}
736

737 738
static inline void memcg_memory_event(struct mem_cgroup *memcg,
				      enum memcg_memory_event event)
739
{
740
	atomic_long_inc(&memcg->memory_events[event]);
741 742 743
	cgroup_file_notify(&memcg->events_file);
}

R
Roman Gushchin 已提交
744 745 746 747 748 749 750 751 752 753 754 755 756 757 758
static inline void memcg_memory_event_mm(struct mm_struct *mm,
					 enum memcg_memory_event event)
{
	struct mem_cgroup *memcg;

	if (mem_cgroup_disabled())
		return;

	rcu_read_lock();
	memcg = mem_cgroup_from_task(rcu_dereference(mm->owner));
	if (likely(memcg))
		memcg_memory_event(memcg, event);
	rcu_read_unlock();
}

759
#ifdef CONFIG_TRANSPARENT_HUGEPAGE
760
void mem_cgroup_split_huge_fixup(struct page *head);
761 762
#endif

A
Andrew Morton 已提交
763
#else /* CONFIG_MEMCG */
764 765 766 767

#define MEM_CGROUP_ID_SHIFT	0
#define MEM_CGROUP_ID_MAX	0

768 769
struct mem_cgroup;

770 771 772 773 774
static inline bool mem_cgroup_is_root(struct mem_cgroup *memcg)
{
	return true;
}

775 776 777 778 779
static inline bool mem_cgroup_disabled(void)
{
	return true;
}

780 781
static inline void memcg_memory_event(struct mem_cgroup *memcg,
				      enum memcg_memory_event event)
782 783 784
{
}

R
Roman Gushchin 已提交
785 786 787 788 789
static inline void memcg_memory_event_mm(struct mm_struct *mm,
					 enum memcg_memory_event event)
{
}

R
Roman Gushchin 已提交
790 791
static inline enum mem_cgroup_protection mem_cgroup_protected(
	struct mem_cgroup *root, struct mem_cgroup *memcg)
792
{
R
Roman Gushchin 已提交
793
	return MEMCG_PROT_NONE;
794 795
}

796 797
static inline int mem_cgroup_try_charge(struct page *page, struct mm_struct *mm,
					gfp_t gfp_mask,
798 799
					struct mem_cgroup **memcgp,
					bool compound)
800
{
801
	*memcgp = NULL;
802 803 804
	return 0;
}

805 806 807 808 809 810 811 812 813 814
static inline int mem_cgroup_try_charge_delay(struct page *page,
					      struct mm_struct *mm,
					      gfp_t gfp_mask,
					      struct mem_cgroup **memcgp,
					      bool compound)
{
	*memcgp = NULL;
	return 0;
}

815 816
static inline void mem_cgroup_commit_charge(struct page *page,
					    struct mem_cgroup *memcg,
817
					    bool lrucare, bool compound)
818 819 820
{
}

821
static inline void mem_cgroup_cancel_charge(struct page *page,
822 823
					    struct mem_cgroup *memcg,
					    bool compound)
824 825 826
{
}

827
static inline void mem_cgroup_uncharge(struct page *page)
828 829 830
{
}

831
static inline void mem_cgroup_uncharge_list(struct list_head *page_list)
832 833 834
{
}

835
static inline void mem_cgroup_migrate(struct page *old, struct page *new)
836 837 838
{
}

839
static inline struct lruvec *mem_cgroup_lruvec(struct pglist_data *pgdat,
840
				struct mem_cgroup *memcg)
K
KAMEZAWA Hiroyuki 已提交
841
{
842
	return node_lruvec(pgdat);
K
KAMEZAWA Hiroyuki 已提交
843 844
}

845
static inline struct lruvec *mem_cgroup_page_lruvec(struct page *page,
M
Mel Gorman 已提交
846
						    struct pglist_data *pgdat)
847
{
M
Mel Gorman 已提交
848
	return &pgdat->lruvec;
849 850
}

851
static inline bool mm_match_cgroup(struct mm_struct *mm,
852
		struct mem_cgroup *memcg)
853
{
854
	return true;
855 856
}

857 858
static inline bool task_in_mem_cgroup(struct task_struct *task,
				      const struct mem_cgroup *memcg)
859
{
860
	return true;
861 862
}

863 864 865 866 867
static inline struct mem_cgroup *get_mem_cgroup_from_mm(struct mm_struct *mm)
{
	return NULL;
}

868 869 870 871 872
static inline struct mem_cgroup *get_mem_cgroup_from_page(struct page *page)
{
	return NULL;
}

873 874 875 876
static inline void mem_cgroup_put(struct mem_cgroup *memcg)
{
}

877 878 879 880 881 882 883 884 885 886 887 888 889
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)
{
}

890 891 892 893 894 895
static inline int mem_cgroup_scan_tasks(struct mem_cgroup *memcg,
		int (*fn)(struct task_struct *, void *), void *arg)
{
	return 0;
}

896
static inline unsigned short mem_cgroup_id(struct mem_cgroup *memcg)
897
{
898 899 900 901 902 903 904 905
	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;
906
}
907

908 909 910 911 912
static inline struct mem_cgroup *mem_cgroup_from_seq(struct seq_file *m)
{
	return NULL;
}

913 914 915 916 917
static inline struct mem_cgroup *lruvec_memcg(struct lruvec *lruvec)
{
	return NULL;
}

918
static inline bool mem_cgroup_online(struct mem_cgroup *memcg)
919
{
920
	return true;
921 922
}

923 924 925 926 927 928
static inline
unsigned long mem_cgroup_get_zone_lru_size(struct lruvec *lruvec,
		enum lru_list lru, int zone_idx)
{
	return 0;
}
929

930
static inline unsigned long mem_cgroup_get_max(struct mem_cgroup *memcg)
931 932 933 934
{
	return 0;
}

935
static inline void
936 937 938 939 940 941
mem_cgroup_print_oom_context(struct mem_cgroup *memcg, struct task_struct *p)
{
}

static inline void
mem_cgroup_print_oom_meminfo(struct mem_cgroup *memcg)
942 943 944
{
}

945 946 947 948 949 950
static inline struct mem_cgroup *lock_page_memcg(struct page *page)
{
	return NULL;
}

static inline void __unlock_page_memcg(struct mem_cgroup *memcg)
951 952 953
{
}

J
Johannes Weiner 已提交
954
static inline void unlock_page_memcg(struct page *page)
955 956 957
{
}

958 959 960 961
static inline void mem_cgroup_handle_over_high(void)
{
}

962
static inline void mem_cgroup_enter_user_fault(void)
963 964 965
{
}

966
static inline void mem_cgroup_exit_user_fault(void)
967 968 969
{
}

970 971 972 973 974
static inline bool task_in_memcg_oom(struct task_struct *p)
{
	return false;
}

975
static inline bool mem_cgroup_oom_synchronize(bool wait)
976 977 978 979
{
	return false;
}

980 981 982 983 984 985 986 987 988 989
static inline struct mem_cgroup *mem_cgroup_get_oom_group(
	struct task_struct *victim, struct mem_cgroup *oom_domain)
{
	return NULL;
}

static inline void mem_cgroup_print_oom_group(struct mem_cgroup *memcg)
{
}

990 991 992 993 994
static inline unsigned long memcg_page_state(struct mem_cgroup *memcg, int idx)
{
	return 0;
}

995 996
static inline unsigned long memcg_page_state_local(struct mem_cgroup *memcg,
						   int idx)
997 998 999 1000
{
	return 0;
}

1001
static inline void __mod_memcg_state(struct mem_cgroup *memcg,
1002
				     int idx,
1003
				     int nr)
1004 1005 1006
{
}

1007
static inline void mod_memcg_state(struct mem_cgroup *memcg,
1008
				   int idx,
1009
				   int nr)
1010 1011 1012
{
}

1013
static inline void __mod_memcg_page_state(struct page *page,
1014
					  int idx,
1015
					  int nr)
1016 1017 1018
{
}

1019
static inline void mod_memcg_page_state(struct page *page,
1020
					int idx,
1021
					int nr)
1022 1023 1024
{
}

1025 1026 1027 1028 1029 1030
static inline unsigned long lruvec_page_state(struct lruvec *lruvec,
					      enum node_stat_item idx)
{
	return node_page_state(lruvec_pgdat(lruvec), idx);
}

1031 1032
static inline unsigned long lruvec_page_state_local(struct lruvec *lruvec,
						    enum node_stat_item idx)
1033
{
1034
	return node_page_state(lruvec_pgdat(lruvec), idx);
1035 1036
}

1037 1038
static inline void __mod_lruvec_state(struct lruvec *lruvec,
				      enum node_stat_item idx, int val)
1039
{
1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058
	__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);
1059 1060
}

1061
static inline
1062
unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order,
1063 1064
					    gfp_t gfp_mask,
					    unsigned long *total_scanned)
1065
{
1066
	return 0;
1067 1068
}

1069
static inline void mem_cgroup_split_huge_fixup(struct page *head)
1070 1071 1072
{
}

1073 1074 1075 1076 1077 1078
static inline void count_memcg_events(struct mem_cgroup *memcg,
				      enum vm_event_item idx,
				      unsigned long count)
{
}

1079 1080 1081 1082 1083 1084
static inline void __count_memcg_events(struct mem_cgroup *memcg,
					enum vm_event_item idx,
					unsigned long count)
{
}

1085
static inline void count_memcg_page_event(struct page *page,
1086
					  int idx)
1087 1088 1089
{
}

1090
static inline
1091
void count_memcg_event_mm(struct mm_struct *mm, enum vm_event_item idx)
1092 1093
{
}
A
Andrew Morton 已提交
1094
#endif /* CONFIG_MEMCG */
1095

1096
/* idx can be of type enum memcg_stat_item or node_stat_item */
1097
static inline void __inc_memcg_state(struct mem_cgroup *memcg,
1098
				     int idx)
1099 1100 1101 1102
{
	__mod_memcg_state(memcg, idx, 1);
}

1103
/* idx can be of type enum memcg_stat_item or node_stat_item */
1104
static inline void __dec_memcg_state(struct mem_cgroup *memcg,
1105
				     int idx)
1106 1107 1108 1109
{
	__mod_memcg_state(memcg, idx, -1);
}

1110
/* idx can be of type enum memcg_stat_item or node_stat_item */
1111
static inline void __inc_memcg_page_state(struct page *page,
1112
					  int idx)
1113 1114 1115 1116
{
	__mod_memcg_page_state(page, idx, 1);
}

1117
/* idx can be of type enum memcg_stat_item or node_stat_item */
1118
static inline void __dec_memcg_page_state(struct page *page,
1119
					  int idx)
1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147
{
	__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);
}

1148
/* idx can be of type enum memcg_stat_item or node_stat_item */
1149
static inline void inc_memcg_state(struct mem_cgroup *memcg,
1150
				   int idx)
1151 1152 1153 1154
{
	mod_memcg_state(memcg, idx, 1);
}

1155
/* idx can be of type enum memcg_stat_item or node_stat_item */
1156
static inline void dec_memcg_state(struct mem_cgroup *memcg,
1157
				   int idx)
1158 1159 1160 1161
{
	mod_memcg_state(memcg, idx, -1);
}

1162
/* idx can be of type enum memcg_stat_item or node_stat_item */
1163
static inline void inc_memcg_page_state(struct page *page,
1164
					int idx)
1165 1166 1167 1168
{
	mod_memcg_page_state(page, idx, 1);
}

1169
/* idx can be of type enum memcg_stat_item or node_stat_item */
1170
static inline void dec_memcg_page_state(struct page *page,
1171
					int idx)
1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199
{
	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);
}

1200
#ifdef CONFIG_CGROUP_WRITEBACK
T
Tejun Heo 已提交
1201 1202

struct wb_domain *mem_cgroup_wb_domain(struct bdi_writeback *wb);
1203 1204 1205
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
1206 1207 1208 1209 1210 1211 1212 1213

#else	/* CONFIG_CGROUP_WRITEBACK */

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

1214
static inline void mem_cgroup_wb_stats(struct bdi_writeback *wb,
1215 1216
				       unsigned long *pfilepages,
				       unsigned long *pheadroom,
1217 1218 1219 1220 1221
				       unsigned long *pdirty,
				       unsigned long *pwriteback)
{
}

T
Tejun Heo 已提交
1222
#endif	/* CONFIG_CGROUP_WRITEBACK */
1223

G
Glauber Costa 已提交
1224
struct sock;
1225 1226
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);
1227
#ifdef CONFIG_MEMCG
1228 1229
extern struct static_key_false memcg_sockets_enabled_key;
#define mem_cgroup_sockets_enabled static_branch_unlikely(&memcg_sockets_enabled_key)
1230 1231
void mem_cgroup_sk_alloc(struct sock *sk);
void mem_cgroup_sk_free(struct sock *sk);
1232
static inline bool mem_cgroup_under_socket_pressure(struct mem_cgroup *memcg)
1233
{
1234
	if (!cgroup_subsys_on_dfl(memory_cgrp_subsys) && memcg->tcpmem_pressure)
1235 1236 1237 1238 1239 1240
		return true;
	do {
		if (time_before(jiffies, memcg->socket_pressure))
			return true;
	} while ((memcg = parent_mem_cgroup(memcg)));
	return false;
1241 1242
}
#else
1243
#define mem_cgroup_sockets_enabled 0
1244 1245
static inline void mem_cgroup_sk_alloc(struct sock *sk) { };
static inline void mem_cgroup_sk_free(struct sock *sk) { };
1246
static inline bool mem_cgroup_under_socket_pressure(struct mem_cgroup *memcg)
1247 1248 1249 1250
{
	return false;
}
#endif
1251

1252 1253
struct kmem_cache *memcg_kmem_get_cache(struct kmem_cache *cachep);
void memcg_kmem_put_cache(struct kmem_cache *cachep);
1254 1255

#ifdef CONFIG_MEMCG_KMEM
1256 1257 1258 1259
int __memcg_kmem_charge(struct page *page, gfp_t gfp, int order);
void __memcg_kmem_uncharge(struct page *page, int order);
int __memcg_kmem_charge_memcg(struct page *page, gfp_t gfp, int order,
			      struct mem_cgroup *memcg);
1260

1261
extern struct static_key_false memcg_kmem_enabled_key;
1262
extern struct workqueue_struct *memcg_kmem_cache_wq;
1263

1264
extern int memcg_nr_cache_ids;
1265 1266
void memcg_get_cache_ids(void);
void memcg_put_cache_ids(void);
1267 1268 1269 1270 1271 1272

/*
 * 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
 */
1273
#define for_each_memcg_cache_index(_idx)	\
1274
	for ((_idx) = 0; (_idx) < memcg_nr_cache_ids; (_idx)++)
1275

1276 1277
static inline bool memcg_kmem_enabled(void)
{
1278
	return static_branch_unlikely(&memcg_kmem_enabled_key);
1279 1280
}

1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300
static inline int memcg_kmem_charge(struct page *page, gfp_t gfp, int order)
{
	if (memcg_kmem_enabled())
		return __memcg_kmem_charge(page, gfp, order);
	return 0;
}

static inline void memcg_kmem_uncharge(struct page *page, int order)
{
	if (memcg_kmem_enabled())
		__memcg_kmem_uncharge(page, order);
}

static inline int memcg_kmem_charge_memcg(struct page *page, gfp_t gfp,
					  int order, struct mem_cgroup *memcg)
{
	if (memcg_kmem_enabled())
		return __memcg_kmem_charge_memcg(page, gfp, order, memcg);
	return 0;
}
M
Michal Hocko 已提交
1301
/*
J
Jesper Dangaard Brouer 已提交
1302
 * helper for accessing a memcg's index. It will be used as an index in the
M
Michal Hocko 已提交
1303 1304 1305 1306 1307 1308 1309
 * 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;
}
1310

1311 1312
extern int memcg_expand_shrinker_maps(int new_id);

1313 1314
extern void memcg_set_shrinker_bit(struct mem_cgroup *memcg,
				   int nid, int shrinker_id);
1315
#else
1316 1317 1318 1319 1320 1321 1322 1323 1324 1325

static inline int memcg_kmem_charge(struct page *page, gfp_t gfp, int order)
{
	return 0;
}

static inline void memcg_kmem_uncharge(struct page *page, int order)
{
}

1326 1327 1328 1329 1330 1331 1332 1333 1334
static inline int __memcg_kmem_charge(struct page *page, gfp_t gfp, int order)
{
	return 0;
}

static inline void __memcg_kmem_uncharge(struct page *page, int order)
{
}

1335 1336 1337
#define for_each_memcg_cache_index(_idx)	\
	for (; NULL; )

1338 1339 1340 1341 1342
static inline bool memcg_kmem_enabled(void)
{
	return false;
}

1343 1344 1345 1346 1347
static inline int memcg_cache_id(struct mem_cgroup *memcg)
{
	return -1;
}

1348 1349 1350 1351 1352 1353 1354 1355
static inline void memcg_get_cache_ids(void)
{
}

static inline void memcg_put_cache_ids(void)
{
}

1356 1357
static inline void memcg_set_shrinker_bit(struct mem_cgroup *memcg,
					  int nid, int shrinker_id) { }
1358
#endif /* CONFIG_MEMCG_KMEM */
1359

B
Balbir Singh 已提交
1360
#endif /* _LINUX_MEMCONTROL_H */