memcontrol.h 30.5 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
enum memcg_memory_event {
	MEMCG_LOW,
53 54 55
	MEMCG_HIGH,
	MEMCG_MAX,
	MEMCG_OOM,
R
Roman Gushchin 已提交
56
	MEMCG_OOM_KILL,
57 58
	MEMCG_SWAP_MAX,
	MEMCG_SWAP_FAIL,
59
	MEMCG_NR_MEMORY_EVENTS,
60 61
};

R
Roman Gushchin 已提交
62 63 64 65 66 67
enum mem_cgroup_protection {
	MEMCG_PROT_NONE,
	MEMCG_PROT_LOW,
	MEMCG_PROT_MIN,
};

68
struct mem_cgroup_reclaim_cookie {
69
	pg_data_t *pgdat;
70 71 72 73
	int priority;
	unsigned int generation;
};

74 75 76 77 78 79 80 81 82 83
#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
84 85 86 87 88 89 90 91 92 93 94 95 96 97
/*
 * 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 {
98
	long count[MEMCG_NR_STAT];
99
	unsigned long events[NR_VM_EVENT_ITEMS];
M
Michal Hocko 已提交
100 101 102 103 104 105 106 107 108 109
	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;
};

110 111 112 113
struct lruvec_stat {
	long count[NR_VM_NODE_STAT_ITEMS];
};

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;
119 120 121 122

	struct lruvec_stat __percpu *lruvec_stat_cpu;
	atomic_long_t		lruvec_stat[NR_VM_NODE_STAT_ITEMS];

123
	unsigned long		lru_zone_size[MAX_NR_ZONES][NR_LRU_LISTS];
M
Michal Hocko 已提交
124 125 126 127 128 129 130

	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;
131 132 133
	bool			congested;	/* memcg has many dirty pages */
						/* backed by a congested BDI */

M
Michal Hocko 已提交
134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163
	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;
};

164 165 166 167 168 169
enum memcg_kmem_state {
	KMEM_NONE,
	KMEM_ALLOCATED,
	KMEM_ONLINE,
};

170 171 172 173 174 175 176 177 178
#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
179 180 181 182 183 184 185 186 187
/*
 * 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;

188 189 190
	/* Private memcg ID. Used to ID objects that outlive the cgroup */
	struct mem_cgroup_id id;

M
Michal Hocko 已提交
191 192
	/* Accounted resources */
	struct page_counter memory;
193
	struct page_counter swap;
194 195

	/* Legacy consumer-oriented counters */
M
Michal Hocko 已提交
196 197
	struct page_counter memsw;
	struct page_counter kmem;
198
	struct page_counter tcpmem;
M
Michal Hocko 已提交
199

200
	/* Upper bound of normal memory consumption range */
M
Michal Hocko 已提交
201 202
	unsigned long high;

203 204 205
	/* Range enforcement for interrupt charges */
	struct work_struct high_work;

M
Michal Hocko 已提交
206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223
	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;

224
	/* memory.events */
225 226
	struct cgroup_file events_file;

227 228 229
	/* handle for "memory.swap.events" */
	struct cgroup_file swap_events_file;

M
Michal Hocko 已提交
230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246
	/* 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;
247 248 249 250 251 252
	/* taken only while moving_account > 0 */
	spinlock_t		move_lock;
	unsigned long		move_lock_flags;

	MEMCG_PADDING(_pad1_);

M
Michal Hocko 已提交
253 254 255 256 257
	/*
	 * set > 0 if pages under this cgroup are moving to other cgroup.
	 */
	atomic_t		moving_account;
	struct task_struct	*move_lock_task;
258

259
	/* memory.stat */
260
	struct mem_cgroup_stat_cpu __percpu *stat_cpu;
261 262 263

	MEMCG_PADDING(_pad2_);

264
	atomic_long_t		stat[MEMCG_NR_STAT];
265
	atomic_long_t		events[NR_VM_EVENT_ITEMS];
266
	atomic_long_t memory_events[MEMCG_NR_MEMORY_EVENTS];
M
Michal Hocko 已提交
267

268 269 270
	unsigned long		socket_pressure;

	/* Legacy tcp memory accounting */
271 272
	bool			tcpmem_active;
	int			tcpmem_pressure;
273

274
#ifndef CONFIG_SLOB
M
Michal Hocko 已提交
275 276
        /* Index in the kmem_cache->memcg_params.memcg_caches array */
	int kmemcg_id;
277
	enum memcg_kmem_state kmem_state;
278
	struct list_head kmem_caches;
M
Michal Hocko 已提交
279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299
#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 */
};
300

301 302 303 304 305 306
/*
 * 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

307
extern struct mem_cgroup *root_mem_cgroup;
T
Tejun Heo 已提交
308

309 310 311 312 313
static inline bool mem_cgroup_disabled(void)
{
	return !cgroup_subsys_enabled(memory_cgrp_subsys);
}

R
Roman Gushchin 已提交
314 315
enum mem_cgroup_protection mem_cgroup_protected(struct mem_cgroup *root,
						struct mem_cgroup *memcg);
316

317
int mem_cgroup_try_charge(struct page *page, struct mm_struct *mm,
318 319
			  gfp_t gfp_mask, struct mem_cgroup **memcgp,
			  bool compound);
320 321 322
int mem_cgroup_try_charge_delay(struct page *page, struct mm_struct *mm,
			  gfp_t gfp_mask, struct mem_cgroup **memcgp,
			  bool compound);
323
void mem_cgroup_commit_charge(struct page *page, struct mem_cgroup *memcg,
324 325 326
			      bool lrucare, bool compound);
void mem_cgroup_cancel_charge(struct page *page, struct mem_cgroup *memcg,
		bool compound);
327
void mem_cgroup_uncharge(struct page *page);
328
void mem_cgroup_uncharge_list(struct list_head *page_list);
329

330
void mem_cgroup_migrate(struct page *oldpage, struct page *newpage);
331

332 333
static struct mem_cgroup_per_node *
mem_cgroup_nodeinfo(struct mem_cgroup *memcg, int nid)
334
{
335
	return memcg->nodeinfo[nid];
336 337 338
}

/**
339 340
 * mem_cgroup_lruvec - get the lru list vector for a node or a memcg zone
 * @node: node of the wanted lruvec
341 342
 * @memcg: memcg of the wanted lruvec
 *
343 344
 * 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
345 346
 * is disabled.
 */
347
static inline struct lruvec *mem_cgroup_lruvec(struct pglist_data *pgdat,
348
				struct mem_cgroup *memcg)
349
{
350
	struct mem_cgroup_per_node *mz;
351 352 353
	struct lruvec *lruvec;

	if (mem_cgroup_disabled()) {
354
		lruvec = node_lruvec(pgdat);
355 356 357
		goto out;
	}

358
	mz = mem_cgroup_nodeinfo(memcg, pgdat->node_id);
359 360 361 362
	lruvec = &mz->lruvec;
out:
	/*
	 * Since a node can be onlined after the mem_cgroup was created,
M
Mel Gorman 已提交
363
	 * we have to be prepared to initialize lruvec->pgdat here;
364 365
	 * and if offlined then reonlined, we need to reinitialize it.
	 */
366 367
	if (unlikely(lruvec->pgdat != pgdat))
		lruvec->pgdat = pgdat;
368 369 370
	return lruvec;
}

M
Mel Gorman 已提交
371
struct lruvec *mem_cgroup_page_lruvec(struct page *, struct pglist_data *);
372

373
bool task_in_mem_cgroup(struct task_struct *task, struct mem_cgroup *memcg);
374
struct mem_cgroup *mem_cgroup_from_task(struct task_struct *p);
375

376 377
struct mem_cgroup *get_mem_cgroup_from_mm(struct mm_struct *mm);

M
Michal Hocko 已提交
378 379 380 381 382
static inline
struct mem_cgroup *mem_cgroup_from_css(struct cgroup_subsys_state *css){
	return css ? container_of(css, struct mem_cgroup, css) : NULL;
}

383 384
static inline void mem_cgroup_put(struct mem_cgroup *memcg)
{
385 386
	if (memcg)
		css_put(&memcg->css);
387 388
}

389 390 391
#define mem_cgroup_from_counter(counter, member)	\
	container_of(counter, struct mem_cgroup, member)

M
Michal Hocko 已提交
392 393 394 395
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 *);
396 397
int mem_cgroup_scan_tasks(struct mem_cgroup *,
			  int (*)(struct task_struct *, void *), void *);
M
Michal Hocko 已提交
398

399 400 401 402 403
static inline unsigned short mem_cgroup_id(struct mem_cgroup *memcg)
{
	if (mem_cgroup_disabled())
		return 0;

404
	return memcg->id.id;
405
}
406
struct mem_cgroup *mem_cgroup_from_id(unsigned short id);
407

408 409 410 411 412 413 414 415 416 417 418
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;
}

419 420 421 422 423 424 425 426 427 428 429 430 431 432
/**
 * 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
433 434 435 436 437 438 439 440 441
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
442

443 444
static inline bool mm_match_cgroup(struct mm_struct *mm,
				   struct mem_cgroup *memcg)
445
{
446
	struct mem_cgroup *task_memcg;
447
	bool match = false;
448

449
	rcu_read_lock();
450
	task_memcg = mem_cgroup_from_task(rcu_dereference(mm->owner));
451
	if (task_memcg)
452
		match = mem_cgroup_is_descendant(task_memcg, memcg);
453
	rcu_read_unlock();
454
	return match;
455
}
456

457
struct cgroup_subsys_state *mem_cgroup_css_from_page(struct page *page);
458
ino_t page_cgroup_ino(struct page *page);
459

460 461 462 463 464 465 466
static inline bool mem_cgroup_online(struct mem_cgroup *memcg)
{
	if (mem_cgroup_disabled())
		return true;
	return !!(memcg->css.flags & CSS_ONLINE);
}

467 468 469
/*
 * For memory reclaim.
 */
470
int mem_cgroup_select_victim_node(struct mem_cgroup *memcg);
M
Michal Hocko 已提交
471 472

void mem_cgroup_update_lru_size(struct lruvec *lruvec, enum lru_list lru,
473
		int zid, int nr_pages);
M
Michal Hocko 已提交
474

475 476 477
unsigned long mem_cgroup_node_nr_lru_pages(struct mem_cgroup *memcg,
					   int nid, unsigned int lru_mask);

M
Michal Hocko 已提交
478 479 480
static inline
unsigned long mem_cgroup_get_lru_size(struct lruvec *lruvec, enum lru_list lru)
{
481
	struct mem_cgroup_per_node *mz;
482 483
	unsigned long nr_pages = 0;
	int zid;
M
Michal Hocko 已提交
484

485
	mz = container_of(lruvec, struct mem_cgroup_per_node, lruvec);
486 487 488 489 490 491 492 493 494 495 496 497 498
	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
499 500
}

501 502
void mem_cgroup_handle_over_high(void);

503
unsigned long mem_cgroup_get_max(struct mem_cgroup *memcg);
504

505 506
void mem_cgroup_print_oom_info(struct mem_cgroup *memcg,
				struct task_struct *p);
507

508
static inline void mem_cgroup_oom_enable(void)
509
{
T
Tejun Heo 已提交
510 511
	WARN_ON(current->memcg_may_oom);
	current->memcg_may_oom = 1;
512 513
}

514
static inline void mem_cgroup_oom_disable(void)
515
{
T
Tejun Heo 已提交
516 517
	WARN_ON(!current->memcg_may_oom);
	current->memcg_may_oom = 0;
518 519
}

520 521
static inline bool task_in_memcg_oom(struct task_struct *p)
{
T
Tejun Heo 已提交
522
	return p->memcg_in_oom;
523 524
}

525
bool mem_cgroup_oom_synchronize(bool wait);
526

A
Andrew Morton 已提交
527
#ifdef CONFIG_MEMCG_SWAP
528 529
extern int do_swap_account;
#endif
530

531 532
struct mem_cgroup *lock_page_memcg(struct page *page);
void __unlock_page_memcg(struct mem_cgroup *memcg);
J
Johannes Weiner 已提交
533
void unlock_page_memcg(struct page *page);
534

535
/* idx can be of type enum memcg_stat_item or node_stat_item */
536
static inline unsigned long memcg_page_state(struct mem_cgroup *memcg,
537
					     int idx)
538
{
539 540 541 542 543 544
	long x = atomic_long_read(&memcg->stat[idx]);
#ifdef CONFIG_SMP
	if (x < 0)
		x = 0;
#endif
	return x;
545 546
}

547
/* idx can be of type enum memcg_stat_item or node_stat_item */
548
static inline void __mod_memcg_state(struct mem_cgroup *memcg,
549
				     int idx, int val)
550
{
551 552 553 554 555 556 557 558 559 560 561
	long x;

	if (mem_cgroup_disabled())
		return;

	x = val + __this_cpu_read(memcg->stat_cpu->count[idx]);
	if (unlikely(abs(x) > MEMCG_CHARGE_BATCH)) {
		atomic_long_add(x, &memcg->stat[idx]);
		x = 0;
	}
	__this_cpu_write(memcg->stat_cpu->count[idx], x);
562 563
}

564
/* idx can be of type enum memcg_stat_item or node_stat_item */
565
static inline void mod_memcg_state(struct mem_cgroup *memcg,
566
				   int idx, int val)
567
{
568 569 570
	unsigned long flags;

	local_irq_save(flags);
571
	__mod_memcg_state(memcg, idx, val);
572
	local_irq_restore(flags);
573 574
}

M
Michal Hocko 已提交
575
/**
576
 * mod_memcg_page_state - update page state statistics
J
Johannes Weiner 已提交
577
 * @page: the page
M
Michal Hocko 已提交
578 579 580
 * @idx: page state item to account
 * @val: number of pages (positive or negative)
 *
581 582 583
 * 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:
584
 *
585
 *   lock_page(page) or lock_page_memcg(page)
586
 *   if (TestClearPageState(page))
587
 *     mod_memcg_page_state(page, state, -1);
588
 *   unlock_page(page) or unlock_page_memcg(page)
589 590
 *
 * Kernel pages are an exception to this, since they'll never move.
M
Michal Hocko 已提交
591
 */
592
static inline void __mod_memcg_page_state(struct page *page,
593
					  int idx, int val)
594 595 596 597 598
{
	if (page->mem_cgroup)
		__mod_memcg_state(page->mem_cgroup, idx, val);
}

599
static inline void mod_memcg_page_state(struct page *page,
600
					int idx, int val)
M
Michal Hocko 已提交
601
{
J
Johannes Weiner 已提交
602
	if (page->mem_cgroup)
603
		mod_memcg_state(page->mem_cgroup, idx, val);
M
Michal Hocko 已提交
604 605
}

606 607
static inline unsigned long lruvec_page_state(struct lruvec *lruvec,
					      enum node_stat_item idx)
608
{
609
	struct mem_cgroup_per_node *pn;
610
	long x;
611 612 613 614 615

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

	pn = container_of(lruvec, struct mem_cgroup_per_node, lruvec);
616 617 618 619 620 621
	x = atomic_long_read(&pn->lruvec_stat[idx]);
#ifdef CONFIG_SMP
	if (x < 0)
		x = 0;
#endif
	return x;
622 623
}

624 625
static inline void __mod_lruvec_state(struct lruvec *lruvec,
				      enum node_stat_item idx, int val)
626
{
627
	struct mem_cgroup_per_node *pn;
628
	long x;
629

630
	/* Update node */
631
	__mod_node_page_state(lruvec_pgdat(lruvec), idx, val);
632

633 634
	if (mem_cgroup_disabled())
		return;
635

636
	pn = container_of(lruvec, struct mem_cgroup_per_node, lruvec);
637 638

	/* Update memcg */
639
	__mod_memcg_state(pn->memcg, idx, val);
640 641

	/* Update lruvec */
642 643 644 645 646 647
	x = val + __this_cpu_read(pn->lruvec_stat_cpu->count[idx]);
	if (unlikely(abs(x) > MEMCG_CHARGE_BATCH)) {
		atomic_long_add(x, &pn->lruvec_stat[idx]);
		x = 0;
	}
	__this_cpu_write(pn->lruvec_stat_cpu->count[idx], x);
648 649 650 651 652
}

static inline void mod_lruvec_state(struct lruvec *lruvec,
				    enum node_stat_item idx, int val)
{
653 654 655
	unsigned long flags;

	local_irq_save(flags);
656
	__mod_lruvec_state(lruvec, idx, val);
657
	local_irq_restore(flags);
658 659 660 661 662
}

static inline void __mod_lruvec_page_state(struct page *page,
					   enum node_stat_item idx, int val)
{
663 664
	pg_data_t *pgdat = page_pgdat(page);
	struct lruvec *lruvec;
665

666 667 668
	/* Untracked pages have no memcg, no lruvec. Update only the node */
	if (!page->mem_cgroup) {
		__mod_node_page_state(pgdat, idx, val);
669
		return;
670 671 672 673
	}

	lruvec = mem_cgroup_lruvec(pgdat, page->mem_cgroup);
	__mod_lruvec_state(lruvec, idx, val);
674 675 676 677 678
}

static inline void mod_lruvec_page_state(struct page *page,
					 enum node_stat_item idx, int val)
{
679 680 681
	unsigned long flags;

	local_irq_save(flags);
682
	__mod_lruvec_page_state(page, idx, val);
683
	local_irq_restore(flags);
684 685
}

686
unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order,
687 688
						gfp_t gfp_mask,
						unsigned long *total_scanned);
D
David Rientjes 已提交
689

690
static inline void __count_memcg_events(struct mem_cgroup *memcg,
691 692
					enum vm_event_item idx,
					unsigned long count)
693
{
694 695 696 697 698 699 700 701 702 703 704
	unsigned long x;

	if (mem_cgroup_disabled())
		return;

	x = count + __this_cpu_read(memcg->stat_cpu->events[idx]);
	if (unlikely(x > MEMCG_CHARGE_BATCH)) {
		atomic_long_add(x, &memcg->events[idx]);
		x = 0;
	}
	__this_cpu_write(memcg->stat_cpu->events[idx], x);
705 706
}

707
static inline void count_memcg_events(struct mem_cgroup *memcg,
708 709
				      enum vm_event_item idx,
				      unsigned long count)
710
{
711 712 713
	unsigned long flags;

	local_irq_save(flags);
714
	__count_memcg_events(memcg, idx, count);
715
	local_irq_restore(flags);
716 717 718
}

static inline void count_memcg_page_event(struct page *page,
719
					  enum vm_event_item idx)
720 721 722 723 724 725 726
{
	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)
727
{
M
Michal Hocko 已提交
728 729
	struct mem_cgroup *memcg;

730 731
	if (mem_cgroup_disabled())
		return;
M
Michal Hocko 已提交
732 733 734

	rcu_read_lock();
	memcg = mem_cgroup_from_task(rcu_dereference(mm->owner));
R
Roman Gushchin 已提交
735
	if (likely(memcg))
736
		count_memcg_events(memcg, idx, 1);
M
Michal Hocko 已提交
737
	rcu_read_unlock();
738
}
739

740 741
static inline void memcg_memory_event(struct mem_cgroup *memcg,
				      enum memcg_memory_event event)
742
{
743
	atomic_long_inc(&memcg->memory_events[event]);
744 745 746
	cgroup_file_notify(&memcg->events_file);
}

R
Roman Gushchin 已提交
747 748 749 750 751 752 753 754 755 756 757 758 759 760 761
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();
}

762
#ifdef CONFIG_TRANSPARENT_HUGEPAGE
763
void mem_cgroup_split_huge_fixup(struct page *head);
764 765
#endif

A
Andrew Morton 已提交
766
#else /* CONFIG_MEMCG */
767 768 769 770

#define MEM_CGROUP_ID_SHIFT	0
#define MEM_CGROUP_ID_MAX	0

771 772
struct mem_cgroup;

773 774 775 776 777
static inline bool mem_cgroup_disabled(void)
{
	return true;
}

778 779
static inline void memcg_memory_event(struct mem_cgroup *memcg,
				      enum memcg_memory_event event)
780 781 782
{
}

R
Roman Gushchin 已提交
783 784 785 786 787
static inline void memcg_memory_event_mm(struct mm_struct *mm,
					 enum memcg_memory_event event)
{
}

R
Roman Gushchin 已提交
788 789
static inline enum mem_cgroup_protection mem_cgroup_protected(
	struct mem_cgroup *root, struct mem_cgroup *memcg)
790
{
R
Roman Gushchin 已提交
791
	return MEMCG_PROT_NONE;
792 793
}

794 795
static inline int mem_cgroup_try_charge(struct page *page, struct mm_struct *mm,
					gfp_t gfp_mask,
796 797
					struct mem_cgroup **memcgp,
					bool compound)
798
{
799
	*memcgp = NULL;
800 801 802
	return 0;
}

803 804 805 806 807 808 809 810 811 812
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;
}

813 814
static inline void mem_cgroup_commit_charge(struct page *page,
					    struct mem_cgroup *memcg,
815
					    bool lrucare, bool compound)
816 817 818
{
}

819
static inline void mem_cgroup_cancel_charge(struct page *page,
820 821
					    struct mem_cgroup *memcg,
					    bool compound)
822 823 824
{
}

825
static inline void mem_cgroup_uncharge(struct page *page)
826 827 828
{
}

829
static inline void mem_cgroup_uncharge_list(struct list_head *page_list)
830 831 832
{
}

833
static inline void mem_cgroup_migrate(struct page *old, struct page *new)
834 835 836
{
}

837
static inline struct lruvec *mem_cgroup_lruvec(struct pglist_data *pgdat,
838
				struct mem_cgroup *memcg)
K
KAMEZAWA Hiroyuki 已提交
839
{
840
	return node_lruvec(pgdat);
K
KAMEZAWA Hiroyuki 已提交
841 842
}

843
static inline struct lruvec *mem_cgroup_page_lruvec(struct page *page,
M
Mel Gorman 已提交
844
						    struct pglist_data *pgdat)
845
{
M
Mel Gorman 已提交
846
	return &pgdat->lruvec;
847 848
}

849
static inline bool mm_match_cgroup(struct mm_struct *mm,
850
		struct mem_cgroup *memcg)
851
{
852
	return true;
853 854
}

855 856
static inline bool task_in_mem_cgroup(struct task_struct *task,
				      const struct mem_cgroup *memcg)
857
{
858
	return true;
859 860
}

861 862 863 864 865
static inline struct mem_cgroup *get_mem_cgroup_from_mm(struct mm_struct *mm)
{
	return NULL;
}

866 867 868 869
static inline void mem_cgroup_put(struct mem_cgroup *memcg)
{
}

870 871 872 873 874 875 876 877 878 879 880 881 882
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)
{
}

883 884 885 886 887 888
static inline int mem_cgroup_scan_tasks(struct mem_cgroup *memcg,
		int (*fn)(struct task_struct *, void *), void *arg)
{
	return 0;
}

889
static inline unsigned short mem_cgroup_id(struct mem_cgroup *memcg)
890
{
891 892 893 894 895 896 897 898
	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;
899
}
900

901 902 903 904 905
static inline struct mem_cgroup *lruvec_memcg(struct lruvec *lruvec)
{
	return NULL;
}

906
static inline bool mem_cgroup_online(struct mem_cgroup *memcg)
907
{
908
	return true;
909 910
}

911
static inline unsigned long
912
mem_cgroup_get_lru_size(struct lruvec *lruvec, enum lru_list lru)
913 914 915
{
	return 0;
}
916 917 918 919 920 921
static inline
unsigned long mem_cgroup_get_zone_lru_size(struct lruvec *lruvec,
		enum lru_list lru, int zone_idx)
{
	return 0;
}
922

923 924 925 926 927 928 929
static inline unsigned long
mem_cgroup_node_nr_lru_pages(struct mem_cgroup *memcg,
			     int nid, unsigned int lru_mask)
{
	return 0;
}

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

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

940 941 942 943 944 945
static inline struct mem_cgroup *lock_page_memcg(struct page *page)
{
	return NULL;
}

static inline void __unlock_page_memcg(struct mem_cgroup *memcg)
946 947 948
{
}

J
Johannes Weiner 已提交
949
static inline void unlock_page_memcg(struct page *page)
950 951 952
{
}

953 954 955 956
static inline void mem_cgroup_handle_over_high(void)
{
}

957
static inline void mem_cgroup_oom_enable(void)
958 959 960
{
}

961
static inline void mem_cgroup_oom_disable(void)
962 963 964
{
}

965 966 967 968 969
static inline bool task_in_memcg_oom(struct task_struct *p)
{
	return false;
}

970
static inline bool mem_cgroup_oom_synchronize(bool wait)
971 972 973 974
{
	return false;
}

975
static inline unsigned long memcg_page_state(struct mem_cgroup *memcg,
976
					     int idx)
977 978 979 980
{
	return 0;
}

981
static inline void __mod_memcg_state(struct mem_cgroup *memcg,
982
				     int idx,
983
				     int nr)
984 985 986
{
}

987
static inline void mod_memcg_state(struct mem_cgroup *memcg,
988
				   int idx,
989
				   int nr)
990 991 992
{
}

993
static inline void __mod_memcg_page_state(struct page *page,
994
					  int idx,
995
					  int nr)
996 997 998
{
}

999
static inline void mod_memcg_page_state(struct page *page,
1000
					int idx,
1001
					int nr)
1002 1003 1004
{
}

1005 1006
static inline unsigned long lruvec_page_state(struct lruvec *lruvec,
					      enum node_stat_item idx)
1007
{
1008
	return node_page_state(lruvec_pgdat(lruvec), idx);
1009 1010
}

1011 1012
static inline void __mod_lruvec_state(struct lruvec *lruvec,
				      enum node_stat_item idx, int val)
1013
{
1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032
	__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);
1033 1034
}

1035
static inline
1036
unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order,
1037 1038
					    gfp_t gfp_mask,
					    unsigned long *total_scanned)
1039
{
1040
	return 0;
1041 1042
}

1043
static inline void mem_cgroup_split_huge_fixup(struct page *head)
1044 1045 1046
{
}

1047 1048 1049 1050 1051 1052 1053
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,
1054
					  int idx)
1055 1056 1057
{
}

1058
static inline
1059
void count_memcg_event_mm(struct mm_struct *mm, enum vm_event_item idx)
1060 1061
{
}
A
Andrew Morton 已提交
1062
#endif /* CONFIG_MEMCG */
1063

1064
/* idx can be of type enum memcg_stat_item or node_stat_item */
1065
static inline void __inc_memcg_state(struct mem_cgroup *memcg,
1066
				     int idx)
1067 1068 1069 1070
{
	__mod_memcg_state(memcg, idx, 1);
}

1071
/* idx can be of type enum memcg_stat_item or node_stat_item */
1072
static inline void __dec_memcg_state(struct mem_cgroup *memcg,
1073
				     int idx)
1074 1075 1076 1077
{
	__mod_memcg_state(memcg, idx, -1);
}

1078
/* idx can be of type enum memcg_stat_item or node_stat_item */
1079
static inline void __inc_memcg_page_state(struct page *page,
1080
					  int idx)
1081 1082 1083 1084
{
	__mod_memcg_page_state(page, idx, 1);
}

1085
/* idx can be of type enum memcg_stat_item or node_stat_item */
1086
static inline void __dec_memcg_page_state(struct page *page,
1087
					  int idx)
1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115
{
	__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);
}

1116
/* idx can be of type enum memcg_stat_item or node_stat_item */
1117
static inline void inc_memcg_state(struct mem_cgroup *memcg,
1118
				   int idx)
1119 1120 1121 1122
{
	mod_memcg_state(memcg, idx, 1);
}

1123
/* idx can be of type enum memcg_stat_item or node_stat_item */
1124
static inline void dec_memcg_state(struct mem_cgroup *memcg,
1125
				   int idx)
1126 1127 1128 1129
{
	mod_memcg_state(memcg, idx, -1);
}

1130
/* idx can be of type enum memcg_stat_item or node_stat_item */
1131
static inline void inc_memcg_page_state(struct page *page,
1132
					int idx)
1133 1134 1135 1136
{
	mod_memcg_page_state(page, idx, 1);
}

1137
/* idx can be of type enum memcg_stat_item or node_stat_item */
1138
static inline void dec_memcg_page_state(struct page *page,
1139
					int idx)
1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167
{
	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);
}

1168
#ifdef CONFIG_CGROUP_WRITEBACK
T
Tejun Heo 已提交
1169 1170

struct wb_domain *mem_cgroup_wb_domain(struct bdi_writeback *wb);
1171 1172 1173
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
1174 1175 1176 1177 1178 1179 1180 1181

#else	/* CONFIG_CGROUP_WRITEBACK */

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

1182
static inline void mem_cgroup_wb_stats(struct bdi_writeback *wb,
1183 1184
				       unsigned long *pfilepages,
				       unsigned long *pheadroom,
1185 1186 1187 1188 1189
				       unsigned long *pdirty,
				       unsigned long *pwriteback)
{
}

T
Tejun Heo 已提交
1190
#endif	/* CONFIG_CGROUP_WRITEBACK */
1191

G
Glauber Costa 已提交
1192
struct sock;
1193 1194
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);
1195
#ifdef CONFIG_MEMCG
1196 1197
extern struct static_key_false memcg_sockets_enabled_key;
#define mem_cgroup_sockets_enabled static_branch_unlikely(&memcg_sockets_enabled_key)
1198 1199
void mem_cgroup_sk_alloc(struct sock *sk);
void mem_cgroup_sk_free(struct sock *sk);
1200
static inline bool mem_cgroup_under_socket_pressure(struct mem_cgroup *memcg)
1201
{
1202
	if (!cgroup_subsys_on_dfl(memory_cgrp_subsys) && memcg->tcpmem_pressure)
1203 1204 1205 1206 1207 1208
		return true;
	do {
		if (time_before(jiffies, memcg->socket_pressure))
			return true;
	} while ((memcg = parent_mem_cgroup(memcg)));
	return false;
1209 1210
}
#else
1211
#define mem_cgroup_sockets_enabled 0
1212 1213
static inline void mem_cgroup_sk_alloc(struct sock *sk) { };
static inline void mem_cgroup_sk_free(struct sock *sk) { };
1214
static inline bool mem_cgroup_under_socket_pressure(struct mem_cgroup *memcg)
1215 1216 1217 1218
{
	return false;
}
#endif
1219

1220 1221 1222 1223 1224 1225 1226
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);

1227
#if defined(CONFIG_MEMCG) && !defined(CONFIG_SLOB)
1228
extern struct static_key_false memcg_kmem_enabled_key;
1229
extern struct workqueue_struct *memcg_kmem_cache_wq;
1230

1231
extern int memcg_nr_cache_ids;
1232 1233
void memcg_get_cache_ids(void);
void memcg_put_cache_ids(void);
1234 1235 1236 1237 1238 1239

/*
 * 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
 */
1240
#define for_each_memcg_cache_index(_idx)	\
1241
	for ((_idx) = 0; (_idx) < memcg_nr_cache_ids; (_idx)++)
1242

1243 1244
static inline bool memcg_kmem_enabled(void)
{
1245
	return static_branch_unlikely(&memcg_kmem_enabled_key);
1246 1247
}

M
Michal Hocko 已提交
1248
/*
J
Jesper Dangaard Brouer 已提交
1249
 * helper for accessing a memcg's index. It will be used as an index in the
M
Michal Hocko 已提交
1250 1251 1252 1253 1254 1255 1256
 * 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;
}
1257

1258
#else
1259 1260 1261
#define for_each_memcg_cache_index(_idx)	\
	for (; NULL; )

1262 1263 1264 1265 1266
static inline bool memcg_kmem_enabled(void)
{
	return false;
}

1267 1268 1269 1270 1271
static inline int memcg_cache_id(struct mem_cgroup *memcg)
{
	return -1;
}

1272 1273 1274 1275 1276 1277 1278 1279
static inline void memcg_get_cache_ids(void)
{
}

static inline void memcg_put_cache_ids(void)
{
}

1280 1281
#endif /* CONFIG_MEMCG && !CONFIG_SLOB */

B
Balbir Singh 已提交
1282
#endif /* _LINUX_MEMCONTROL_H */