cgroup.h 29.3 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14
#ifndef _LINUX_CGROUP_H
#define _LINUX_CGROUP_H
/*
 *  cgroup interface
 *
 *  Copyright (C) 2003 BULL SA
 *  Copyright (C) 2004-2006 Silicon Graphics, Inc.
 *
 */

#include <linux/sched.h>
#include <linux/cpumask.h>
#include <linux/nodemask.h>
#include <linux/rcupdate.h>
15
#include <linux/rculist.h>
B
Balbir Singh 已提交
16
#include <linux/cgroupstats.h>
17
#include <linux/prio_heap.h>
18
#include <linux/rwsem.h>
K
KAMEZAWA Hiroyuki 已提交
19
#include <linux/idr.h>
20
#include <linux/workqueue.h>
A
Aristeu Rozanski 已提交
21
#include <linux/xattr.h>
22
#include <linux/fs.h>
23
#include <linux/percpu-refcount.h>
24 25 26 27 28 29

#ifdef CONFIG_CGROUPS

struct cgroupfs_root;
struct cgroup_subsys;
struct inode;
30
struct cgroup;
K
KAMEZAWA Hiroyuki 已提交
31
struct css_id;
32
struct eventfd_ctx;
33 34 35

extern int cgroup_init_early(void);
extern int cgroup_init(void);
36
extern void cgroup_fork(struct task_struct *p);
37
extern void cgroup_post_fork(struct task_struct *p);
38
extern void cgroup_exit(struct task_struct *p, int run_callbacks);
B
Balbir Singh 已提交
39 40
extern int cgroupstats_build(struct cgroupstats *stats,
				struct dentry *dentry);
41
extern int cgroup_load_subsys(struct cgroup_subsys *ss);
B
Ben Blum 已提交
42
extern void cgroup_unload_subsys(struct cgroup_subsys *ss);
43

44
extern int proc_cgroup_show(struct seq_file *, void *);
45

46 47 48 49 50
/*
 * Define the enumeration of all cgroup subsystems.
 *
 * We define ids for builtin subsystems and then modular ones.
 */
51 52
#define SUBSYS(_x) _x ## _subsys_id,
enum cgroup_subsys_id {
53
#define IS_SUBSYS_ENABLED(option) IS_BUILTIN(option)
54
#include <linux/cgroup_subsys.h>
55 56 57 58 59 60
#undef IS_SUBSYS_ENABLED
	CGROUP_BUILTIN_SUBSYS_COUNT,

	__CGROUP_SUBSYS_TEMP_PLACEHOLDER = CGROUP_BUILTIN_SUBSYS_COUNT - 1,

#define IS_SUBSYS_ENABLED(option) IS_MODULE(option)
61
#include <linux/cgroup_subsys.h>
62
#undef IS_SUBSYS_ENABLED
63
	CGROUP_SUBSYS_COUNT,
64 65 66
};
#undef SUBSYS

67 68
/* Per-subsystem/per-cgroup state maintained by the system. */
struct cgroup_subsys_state {
69
	/* the cgroup that this css is attached to */
70 71
	struct cgroup *cgroup;

72 73 74
	/* the cgroup subsystem that this css is attached to */
	struct cgroup_subsys *ss;

75 76
	/* reference count - access via css_[try]get() and css_put() */
	struct percpu_ref refcnt;
77 78

	unsigned long flags;
K
KAMEZAWA Hiroyuki 已提交
79
	/* ID for this css, if possible */
A
Arnd Bergmann 已提交
80
	struct css_id __rcu *id;
81 82 83

	/* Used to put @cgroup->dentry on the last css_put() */
	struct work_struct dput_work;
84 85 86 87
};

/* bits in struct cgroup_subsys_state flags field */
enum {
88
	CSS_ROOT	= (1 << 0), /* this CSS is the root of the subsystem */
89
	CSS_ONLINE	= (1 << 1), /* between ->css_online() and ->css_offline() */
90 91
};

92 93 94 95 96
/**
 * css_get - obtain a reference on the specified css
 * @css: target css
 *
 * The caller must already have a reference.
97 98 99 100
 */
static inline void css_get(struct cgroup_subsys_state *css)
{
	/* We don't need to reference count the root state */
101
	if (!(css->flags & CSS_ROOT))
102
		percpu_ref_get(&css->refcnt);
103
}
P
Paul Menage 已提交
104

105 106 107 108 109 110 111 112 113 114
/**
 * css_tryget - try to obtain a reference on the specified css
 * @css: target css
 *
 * Obtain a reference on @css if it's alive.  The caller naturally needs to
 * ensure that @css is accessible but doesn't have to be holding a
 * reference on it - IOW, RCU protected access is good enough for this
 * function.  Returns %true if a reference count was successfully obtained;
 * %false otherwise.
 */
P
Paul Menage 已提交
115 116
static inline bool css_tryget(struct cgroup_subsys_state *css)
{
117
	if (css->flags & CSS_ROOT)
P
Paul Menage 已提交
118
		return true;
119
	return percpu_ref_tryget(&css->refcnt);
P
Paul Menage 已提交
120 121
}

122 123 124 125 126 127
/**
 * css_put - put a css reference
 * @css: target css
 *
 * Put a reference obtained via css_get() and css_tryget().
 */
128 129
static inline void css_put(struct cgroup_subsys_state *css)
{
130
	if (!(css->flags & CSS_ROOT))
131
		percpu_ref_put(&css->refcnt);
132 133
}

134 135 136
/* bits in struct cgroup flags field */
enum {
	/* Control Group is dead */
137
	CGRP_DEAD,
P
Paul Menage 已提交
138 139 140 141
	/*
	 * Control Group has previously had a child cgroup or a task,
	 * but no longer (only if CGRP_NOTIFY_ON_RELEASE is set)
	 */
142 143 144
	CGRP_RELEASABLE,
	/* Control Group requires release notifications to userspace */
	CGRP_NOTIFY_ON_RELEASE,
145
	/*
146 147 148
	 * Clone the parent's configuration when creating a new child
	 * cpuset cgroup.  For historical reasons, this option can be
	 * specified at mount time and thus is implemented here.
149
	 */
150
	CGRP_CPUSET_CLONE_CHILDREN,
151 152
	/* see the comment above CGRP_ROOT_SANE_BEHAVIOR for details */
	CGRP_SANE_BEHAVIOR,
153 154
};

155 156 157
struct cgroup_name {
	struct rcu_head rcu_head;
	char name[];
158 159
};

160 161 162
struct cgroup {
	unsigned long flags;		/* "unsigned long" so bitops work */

163 164 165 166 167 168 169
	/*
	 * idr allocated in-hierarchy ID.
	 *
	 * The ID of the root cgroup is always 0, and a new cgroup
	 * will be assigned with a smallest available ID.
	 */
	int id;
T
Tejun Heo 已提交
170

171 172 173 174 175 176
	/*
	 * We link our 'sibling' struct into our parent's 'children'.
	 * Our children link their 'sibling' into our 'children'.
	 */
	struct list_head sibling;	/* my parent's children */
	struct list_head children;	/* my children */
T
Tejun Heo 已提交
177
	struct list_head files;		/* my files */
178

P
Paul Menage 已提交
179
	struct cgroup *parent;		/* my parent */
180
	struct dentry *dentry;		/* cgroup fs entry, RCU protected */
181

182 183 184 185 186 187 188 189
	/*
	 * Monotonically increasing unique serial number which defines a
	 * uniform order among all cgroups.  It's guaranteed that all
	 * ->children lists are in the ascending order of ->serial_nr.
	 * It's used to allow interrupting and resuming iterations.
	 */
	u64 serial_nr;

190 191 192 193 194 195 196 197 198 199 200 201 202
	/*
	 * This is a copy of dentry->d_name, and it's needed because
	 * we can't use dentry->d_name in cgroup_path().
	 *
	 * You must acquire rcu_read_lock() to access cgrp->name, and
	 * the only place that can change it is rename(), which is
	 * protected by parent dir's i_mutex.
	 *
	 * Normally you should use cgroup_name() wrapper rather than
	 * access it directly.
	 */
	struct cgroup_name __rcu *name;

203 204 205 206
	/* Private pointers for each registered subsystem */
	struct cgroup_subsys_state *subsys[CGROUP_SUBSYS_COUNT];

	struct cgroupfs_root *root;
207 208

	/*
209 210
	 * List of cgrp_cset_links pointing at css_sets with tasks in this
	 * cgroup.  Protected by css_set_lock.
211
	 */
212
	struct list_head cset_links;
213 214 215 216 217 218 219

	/*
	 * Linked list running through all cgroups that can
	 * potentially be reaped by the release agent. Protected by
	 * release_list_lock
	 */
	struct list_head release_list;
220

221 222 223 224 225 226
	/*
	 * list of pidlists, up to two for each namespace (one for procs, one
	 * for tasks); created on demand.
	 */
	struct list_head pidlists;
	struct mutex pidlist_mutex;
227

T
Tejun Heo 已提交
228 229 230
	/* dummy css with NULL ->ss, points back to this cgroup */
	struct cgroup_subsys_state dummy_css;

231
	/* For css percpu_ref killing and RCU-protected deletion */
232
	struct rcu_head rcu_head;
233
	struct work_struct destroy_work;
234
	atomic_t css_kill_cnt;
235

L
Lucas De Marchi 已提交
236
	/* List of events which userspace want to receive */
237 238
	struct list_head event_list;
	spinlock_t event_list_lock;
A
Aristeu Rozanski 已提交
239 240 241

	/* directory xattrs */
	struct simple_xattrs xattrs;
242 243
};

244 245 246 247
#define MAX_CGROUP_ROOT_NAMELEN 64

/* cgroupfs_root->flags */
enum {
248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271
	/*
	 * Unfortunately, cgroup core and various controllers are riddled
	 * with idiosyncrasies and pointless options.  The following flag,
	 * when set, will force sane behavior - some options are forced on,
	 * others are disallowed, and some controllers will change their
	 * hierarchical or other behaviors.
	 *
	 * The set of behaviors affected by this flag are still being
	 * determined and developed and the mount option for this flag is
	 * prefixed with __DEVEL__.  The prefix will be dropped once we
	 * reach the point where all behaviors are compatible with the
	 * planned unified hierarchy, which will automatically turn on this
	 * flag.
	 *
	 * The followings are the behaviors currently affected this flag.
	 *
	 * - Mount options "noprefix" and "clone_children" are disallowed.
	 *   Also, cgroupfs file cgroup.clone_children is not created.
	 *
	 * - When mounting an existing superblock, mount options should
	 *   match.
	 *
	 * - Remount is disallowed.
	 *
272 273
	 * - rename(2) is disallowed.
	 *
274 275
	 * - "tasks" is removed.  Everything should be at process
	 *   granularity.  Use "cgroup.procs" instead.
276
	 *
277 278
	 * - "release_agent" and "notify_on_release" are removed.
	 *   Replacement notification mechanism will be implemented.
279
	 *
280 281 282 283
	 * - cpuset: tasks will be kept in empty cpusets when hotplug happens
	 *   and take masks of ancestors with non-empty cpus/mems, instead of
	 *   being moved to an ancestor.
	 *
284 285
	 * - cpuset: a task can be moved into an empty cpuset, and again it
	 *   takes masks of ancestors.
286
	 *
287 288
	 * - memcg: use_hierarchy is on by default and the cgroup file for
	 *   the flag is not created.
289
	 *
290
	 * - blkcg: blk-throttle becomes properly hierarchical.
291 292 293
	 */
	CGRP_ROOT_SANE_BEHAVIOR	= (1 << 0),

294 295
	CGRP_ROOT_NOPREFIX	= (1 << 1), /* mounted subsystems have no named prefix */
	CGRP_ROOT_XATTR		= (1 << 2), /* supports extended attributes */
296 297 298 299 300

	/* mount options live below bit 16 */
	CGRP_ROOT_OPTION_MASK	= (1 << 16) - 1,

	CGRP_ROOT_SUBSYS_BOUND	= (1 << 16), /* subsystems finished binding */
301 302 303 304 305 306 307 308 309 310
};

/*
 * A cgroupfs_root represents the root of a cgroup hierarchy, and may be
 * associated with a superblock to form an active hierarchy.  This is
 * internal to cgroup core.  Don't access directly from controllers.
 */
struct cgroupfs_root {
	struct super_block *sb;

311
	/* The bitmask of subsystems attached to this hierarchy */
312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332
	unsigned long subsys_mask;

	/* Unique id for this hierarchy. */
	int hierarchy_id;

	/* A list running through the attached subsystems */
	struct list_head subsys_list;

	/* The root cgroup for this hierarchy */
	struct cgroup top_cgroup;

	/* Tracks how many cgroups are currently defined in hierarchy.*/
	int number_of_cgroups;

	/* A list running through the active hierarchies */
	struct list_head root_list;

	/* Hierarchy-specific flags */
	unsigned long flags;

	/* IDs for cgroups in this hierarchy */
333
	struct idr cgroup_idr;
334 335 336 337 338 339 340 341

	/* The path to use for release notifications. */
	char release_agent_path[PATH_MAX];

	/* The name for this hierarchy - may be empty */
	char name[MAX_CGROUP_ROOT_NAMELEN];
};

P
Paul Menage 已提交
342 343
/*
 * A css_set is a structure holding pointers to a set of
344 345
 * cgroup_subsys_state objects. This saves space in the task struct
 * object and speeds up fork()/exit(), since a single inc/dec and a
P
Paul Menage 已提交
346 347
 * list_add()/del() can bump the reference count on the entire cgroup
 * set for a task.
348 349 350 351 352
 */

struct css_set {

	/* Reference count */
353
	atomic_t refcount;
354

355 356 357 358 359 360
	/*
	 * List running through all cgroup groups in the same hash
	 * slot. Protected by css_set_lock
	 */
	struct hlist_node hlist;

361 362 363 364 365 366 367
	/*
	 * List running through all tasks using this cgroup
	 * group. Protected by css_set_lock
	 */
	struct list_head tasks;

	/*
368 369
	 * List of cgrp_cset_links pointing at cgroups referenced from this
	 * css_set.  Protected by css_set_lock.
370
	 */
371
	struct list_head cgrp_links;
372 373 374 375

	/*
	 * Set of subsystem states, one for each subsystem. This array
	 * is immutable after creation apart from the init_css_set
B
Ben Blum 已提交
376 377
	 * during subsystem registration (at boot time) and modular subsystem
	 * loading/unloading.
378 379
	 */
	struct cgroup_subsys_state *subsys[CGROUP_SUBSYS_COUNT];
380 381 382

	/* For RCU-protected deletion */
	struct rcu_head rcu_head;
383 384
};

385 386 387 388 389 390 391 392 393 394
/*
 * cgroup_map_cb is an abstract callback API for reporting map-valued
 * control files
 */

struct cgroup_map_cb {
	int (*fill)(struct cgroup_map_cb *cb, const char *key, u64 value);
	void *state;
};

P
Paul Menage 已提交
395 396
/*
 * struct cftype: handler definitions for cgroup control files
397 398
 *
 * When reading/writing to a file:
L
Li Zefan 已提交
399
 *	- the cgroup to use is file->f_dentry->d_parent->d_fsdata
400 401 402
 *	- the 'cftype' of the file is file->f_dentry->d_fsdata
 */

403
/* cftype->flags */
404
enum {
L
Li Zefan 已提交
405 406
	CFTYPE_ONLY_ON_ROOT	= (1 << 0),	/* only create on root cgrp */
	CFTYPE_NOT_ON_ROOT	= (1 << 1),	/* don't create on root cgrp */
407 408
	CFTYPE_INSANE		= (1 << 2),	/* don't create if sane_behavior */
};
409 410 411

#define MAX_CFTYPE_NAME		64

412
struct cftype {
P
Paul Menage 已提交
413 414
	/*
	 * By convention, the name should begin with the name of the
415 416
	 * subsystem, followed by a period.  Zero length string indicates
	 * end of cftype array.
P
Paul Menage 已提交
417
	 */
418 419
	char name[MAX_CFTYPE_NAME];
	int private;
L
Li Zefan 已提交
420 421 422 423
	/*
	 * If not 0, file mode is set to this value, otherwise it will
	 * be figured out automatically
	 */
A
Al Viro 已提交
424
	umode_t mode;
425 426 427 428 429 430 431

	/*
	 * If non-zero, defines the maximum length of string that can
	 * be passed to write_string; defaults to 64
	 */
	size_t max_write_len;

432 433 434
	/* CFTYPE_* flags */
	unsigned int flags;

435 436 437 438 439 440
	/*
	 * The subsys this file belongs to.  Initialized automatically
	 * during registration.  NULL for cgroup core files.
	 */
	struct cgroup_subsys *ss;

441
	int (*open)(struct inode *inode, struct file *file);
442
	ssize_t (*read)(struct cgroup_subsys_state *css, struct cftype *cft,
443 444
			struct file *file,
			char __user *buf, size_t nbytes, loff_t *ppos);
445
	/*
446
	 * read_u64() is a shortcut for the common case of returning a
447 448
	 * single integer. Use it in place of read()
	 */
449
	u64 (*read_u64)(struct cgroup_subsys_state *css, struct cftype *cft);
450 451 452
	/*
	 * read_s64() is a signed version of read_u64()
	 */
453
	s64 (*read_s64)(struct cgroup_subsys_state *css, struct cftype *cft);
454 455 456 457 458 459
	/*
	 * read_map() is used for defining a map of key/value
	 * pairs. It should call cb->fill(cb, key, value) for each
	 * entry. The key/value pairs (and their ordering) should not
	 * change between reboots.
	 */
460
	int (*read_map)(struct cgroup_subsys_state *css, struct cftype *cft,
461
			struct cgroup_map_cb *cb);
462 463 464 465
	/*
	 * read_seq_string() is used for outputting a simple sequence
	 * using seqfile.
	 */
466 467
	int (*read_seq_string)(struct cgroup_subsys_state *css,
			       struct cftype *cft, struct seq_file *m);
468

469
	ssize_t (*write)(struct cgroup_subsys_state *css, struct cftype *cft,
470 471
			 struct file *file,
			 const char __user *buf, size_t nbytes, loff_t *ppos);
472 473

	/*
474
	 * write_u64() is a shortcut for the common case of accepting
475 476 477
	 * a single integer (as parsed by simple_strtoull) from
	 * userspace. Use in place of write(); return 0 or error.
	 */
478 479
	int (*write_u64)(struct cgroup_subsys_state *css, struct cftype *cft,
			 u64 val);
480 481 482
	/*
	 * write_s64() is a signed version of write_u64()
	 */
483 484
	int (*write_s64)(struct cgroup_subsys_state *css, struct cftype *cft,
			 s64 val);
485

486 487 488 489 490
	/*
	 * write_string() is passed a nul-terminated kernelspace
	 * buffer of maximum length determined by max_write_len.
	 * Returns 0 or -ve error code.
	 */
491
	int (*write_string)(struct cgroup_subsys_state *css, struct cftype *cft,
492
			    const char *buffer);
493 494 495 496 497 498
	/*
	 * trigger() callback can be used to get some kick from the
	 * userspace, when the actual string written is not important
	 * at all. The private field can be used to determine the
	 * kick type for multiplexing.
	 */
499
	int (*trigger)(struct cgroup_subsys_state *css, unsigned int event);
500

501
	int (*release)(struct inode *inode, struct file *file);
502 503 504 505 506 507 508 509 510 511 512 513 514 515 516

	/*
	 * register_event() callback will be used to add new userspace
	 * waiter for changes related to the cftype. Implement it if
	 * you want to provide this functionality. Use eventfd_signal()
	 * on eventfd to send notification to userspace.
	 */
	int (*register_event)(struct cgroup *cgrp, struct cftype *cft,
			struct eventfd_ctx *eventfd, const char *args);
	/*
	 * unregister_event() callback will be called when userspace
	 * closes the eventfd or on cgroup removing.
	 * This callback must be implemented, if you want provide
	 * notification functionality.
	 */
517
	void (*unregister_event)(struct cgroup *cgrp, struct cftype *cft,
518
			struct eventfd_ctx *eventfd);
519 520
};

521 522 523 524 525 526 527
/*
 * cftype_sets describe cftypes belonging to a subsystem and are chained at
 * cgroup_subsys->cftsets.  Each cftset points to an array of cftypes
 * terminated by zero length name.
 */
struct cftype_set {
	struct list_head		node;	/* chained at subsys->cftsets */
A
Aristeu Rozanski 已提交
528
	struct cftype			*cfts;
529 530
};

531 532 533 534 535 536 537 538 539
/*
 * See the comment above CGRP_ROOT_SANE_BEHAVIOR for details.  This
 * function can be called as long as @cgrp is accessible.
 */
static inline bool cgroup_sane_behavior(const struct cgroup *cgrp)
{
	return cgrp->root->flags & CGRP_ROOT_SANE_BEHAVIOR;
}

540 541 542 543 544 545
/* Caller should hold rcu_read_lock() */
static inline const char *cgroup_name(const struct cgroup *cgrp)
{
	return rcu_dereference(cgrp->name)->name;
}

A
Aristeu Rozanski 已提交
546
int cgroup_add_cftypes(struct cgroup_subsys *ss, struct cftype *cfts);
547
int cgroup_rm_cftypes(struct cftype *cfts);
548

549
bool cgroup_is_descendant(struct cgroup *cgrp, struct cgroup *ancestor);
550

L
Li Zefan 已提交
551
int cgroup_path(const struct cgroup *cgrp, char *buf, int buflen);
552
int task_cgroup_path(struct task_struct *task, char *buf, size_t buflen);
553

L
Li Zefan 已提交
554
int cgroup_task_count(const struct cgroup *cgrp);
555

556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577
/*
 * Control Group taskset, used to pass around set of tasks to cgroup_subsys
 * methods.
 */
struct cgroup_taskset;
struct task_struct *cgroup_taskset_first(struct cgroup_taskset *tset);
struct task_struct *cgroup_taskset_next(struct cgroup_taskset *tset);
struct cgroup *cgroup_taskset_cur_cgroup(struct cgroup_taskset *tset);
int cgroup_taskset_size(struct cgroup_taskset *tset);

/**
 * cgroup_taskset_for_each - iterate cgroup_taskset
 * @task: the loop cursor
 * @skip_cgrp: skip if task's cgroup matches this, %NULL to iterate through all
 * @tset: taskset to iterate
 */
#define cgroup_taskset_for_each(task, skip_cgrp, tset)			\
	for ((task) = cgroup_taskset_first((tset)); (task);		\
	     (task) = cgroup_taskset_next((tset)))			\
		if (!(skip_cgrp) ||					\
		    cgroup_taskset_cur_cgroup((tset)) != (skip_cgrp))

578 579 580 581
/*
 * Control Group subsystem type.
 * See Documentation/cgroups/cgroups.txt for details
 */
582 583

struct cgroup_subsys {
584 585 586 587 588 589 590 591 592 593 594
	struct cgroup_subsys_state *(*css_alloc)(struct cgroup_subsys_state *parent_css);
	int (*css_online)(struct cgroup_subsys_state *css);
	void (*css_offline)(struct cgroup_subsys_state *css);
	void (*css_free)(struct cgroup_subsys_state *css);

	int (*can_attach)(struct cgroup_subsys_state *css,
			  struct cgroup_taskset *tset);
	void (*cancel_attach)(struct cgroup_subsys_state *css,
			      struct cgroup_taskset *tset);
	void (*attach)(struct cgroup_subsys_state *css,
		       struct cgroup_taskset *tset);
595
	void (*fork)(struct task_struct *task);
596 597
	void (*exit)(struct cgroup_subsys_state *css,
		     struct cgroup_subsys_state *old_css,
598
		     struct task_struct *task);
599
	void (*bind)(struct cgroup_subsys_state *root_css);
600

601
	int subsys_id;
602
	int disabled;
603
	int early_init;
K
KAMEZAWA Hiroyuki 已提交
604 605 606 607 608
	/*
	 * True if this subsys uses ID. ID is not available before cgroup_init()
	 * (not available in early_init time.)
	 */
	bool use_id;
609

610 611 612 613 614 615 616 617 618 619 620 621 622 623 624
	/*
	 * If %false, this subsystem is properly hierarchical -
	 * configuration, resource accounting and restriction on a parent
	 * cgroup cover those of its children.  If %true, hierarchy support
	 * is broken in some ways - some subsystems ignore hierarchy
	 * completely while others are only implemented half-way.
	 *
	 * It's now disallowed to create nested cgroups if the subsystem is
	 * broken and cgroup core will emit a warning message on such
	 * cases.  Eventually, all subsystems will be made properly
	 * hierarchical and this will go away.
	 */
	bool broken_hierarchy;
	bool warned_broken_hierarchy;

625 626 627
#define MAX_CGROUP_TYPE_NAMELEN 32
	const char *name;

628 629
	/*
	 * Link to parent, and list entry in parent's children.
L
Li Zefan 已提交
630
	 * Protected by cgroup_lock()
631 632
	 */
	struct cgroupfs_root *root;
633
	struct list_head sibling;
K
KAMEZAWA Hiroyuki 已提交
634 635
	/* used when use_id == true */
	struct idr idr;
636
	spinlock_t id_lock;
637

638 639 640 641 642 643 644
	/* list of cftype_sets */
	struct list_head cftsets;

	/* base cftypes, automatically [de]registered with subsys itself */
	struct cftype *base_cftypes;
	struct cftype_set base_cftset;

645 646
	/* should be defined only by modular subsystems */
	struct module *module;
647 648 649
};

#define SUBSYS(_x) extern struct cgroup_subsys _x ## _subsys;
650
#define IS_SUBSYS_ENABLED(option) IS_BUILTIN(option)
651
#include <linux/cgroup_subsys.h>
652
#undef IS_SUBSYS_ENABLED
653 654
#undef SUBSYS

T
Tejun Heo 已提交
655 656 657 658 659 660 661 662 663 664 665 666
/**
 * css_parent - find the parent css
 * @css: the target cgroup_subsys_state
 *
 * Return the parent css of @css.  This function is guaranteed to return
 * non-NULL parent as long as @css isn't the root.
 */
static inline
struct cgroup_subsys_state *css_parent(struct cgroup_subsys_state *css)
{
	struct cgroup *parent_cgrp = css->cgroup->parent;

T
Tejun Heo 已提交
667 668 669 670 671 672 673
	if (!parent_cgrp)
		return NULL;

	if (css->ss)
		return parent_cgrp->subsys[css->ss->subsys_id];
	else
		return &parent_cgrp->dummy_css;
T
Tejun Heo 已提交
674 675
}

676 677 678 679 680 681 682 683 684
/**
 * cgroup_css - obtain a cgroup's css for the specified subsystem
 * @cgrp: the cgroup of interest
 * @subsys_id: the subsystem of interest
 *
 * Return @cgrp's css (cgroup_subsys_state) associated with @subsys_id.
 */
static inline struct cgroup_subsys_state *cgroup_css(struct cgroup *cgrp,
						     int subsys_id)
685
{
L
Li Zefan 已提交
686
	return cgrp->subsys[subsys_id];
687 688
}

689 690 691 692 693 694 695 696 697 698 699 700
/**
 * task_css_set_check - obtain a task's css_set with extra access conditions
 * @task: the task to obtain css_set for
 * @__c: extra condition expression to be passed to rcu_dereference_check()
 *
 * A task's css_set is RCU protected, initialized and exited while holding
 * task_lock(), and can only be modified while holding both cgroup_mutex
 * and task_lock() while the task is alive.  This macro verifies that the
 * caller is inside proper critical section and returns @task's css_set.
 *
 * The caller can also specify additional allowed conditions via @__c, such
 * as locks used during the cgroup_subsys::attach() methods.
701
 */
T
Tejun Heo 已提交
702 703
#ifdef CONFIG_PROVE_RCU
extern struct mutex cgroup_mutex;
704 705 706 707
#define task_css_set_check(task, __c)					\
	rcu_dereference_check((task)->cgroups,				\
		lockdep_is_held(&(task)->alloc_lock) ||			\
		lockdep_is_held(&cgroup_mutex) || (__c))
T
Tejun Heo 已提交
708
#else
709 710
#define task_css_set_check(task, __c)					\
	rcu_dereference((task)->cgroups)
T
Tejun Heo 已提交
711
#endif
712

713
/**
714
 * task_css_check - obtain css for (task, subsys) w/ extra access conds
715 716 717 718 719 720 721
 * @task: the target task
 * @subsys_id: the target subsystem ID
 * @__c: extra condition expression to be passed to rcu_dereference_check()
 *
 * Return the cgroup_subsys_state for the (@task, @subsys_id) pair.  The
 * synchronization rules are the same as task_css_set_check().
 */
722
#define task_css_check(task, subsys_id, __c)				\
723 724 725 726 727 728 729 730 731 732 733 734 735 736
	task_css_set_check((task), (__c))->subsys[(subsys_id)]

/**
 * task_css_set - obtain a task's css_set
 * @task: the task to obtain css_set for
 *
 * See task_css_set_check().
 */
static inline struct css_set *task_css_set(struct task_struct *task)
{
	return task_css_set_check(task, false);
}

/**
737
 * task_css - obtain css for (task, subsys)
738 739 740
 * @task: the target task
 * @subsys_id: the target subsystem ID
 *
741
 * See task_css_check().
742
 */
743 744
static inline struct cgroup_subsys_state *task_css(struct task_struct *task,
						   int subsys_id)
745
{
746
	return task_css_check(task, subsys_id, false);
747 748
}

749 750
static inline struct cgroup *task_cgroup(struct task_struct *task,
					 int subsys_id)
751
{
752
	return task_css(task, subsys_id)->cgroup;
753 754
}

L
Li Zefan 已提交
755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772
/**
 * cgroup_from_id - lookup cgroup by id
 * @ss: cgroup subsys to be looked into
 * @id: the cgroup id
 *
 * Returns the cgroup if there's valid one with @id, otherwise returns NULL.
 * Should be called under rcu_read_lock().
 */
static inline struct cgroup *cgroup_from_id(struct cgroup_subsys *ss, int id)
{
#ifdef CONFIG_PROVE_RCU
	rcu_lockdep_assert(rcu_read_lock_held() ||
			   lockdep_is_held(&cgroup_mutex),
			   "cgroup_from_id() needs proper protection");
#endif
	return idr_find(&ss->root->cgroup_idr, id);
}

773 774
struct cgroup_subsys_state *css_next_child(struct cgroup_subsys_state *pos,
					   struct cgroup_subsys_state *parent);
775

776
/**
777 778 779
 * css_for_each_child - iterate through children of a css
 * @pos: the css * to use as the loop cursor
 * @parent: css whose children to walk
780
 *
781 782
 * Walk @parent's children.  Must be called under rcu_read_lock().  A child
 * css which hasn't finished ->css_online() or already has finished
783
 * ->css_offline() may show up during traversal and it's each subsystem's
784 785
 * responsibility to verify that each @pos is alive.
 *
786
 * If a subsystem synchronizes against the parent in its ->css_online() and
787
 * before starting iterating, a css which finished ->css_online() is
788
 * guaranteed to be visible in the future iterations.
789 790 791 792
 *
 * It is allowed to temporarily drop RCU read lock during iteration.  The
 * caller is responsible for ensuring that @pos remains accessible until
 * the start of the next iteration by, for example, bumping the css refcnt.
793
 */
794 795 796
#define css_for_each_child(pos, parent)					\
	for ((pos) = css_next_child(NULL, (parent)); (pos);		\
	     (pos) = css_next_child((pos), (parent)))
797

798 799 800 801 802 803
struct cgroup_subsys_state *
css_next_descendant_pre(struct cgroup_subsys_state *pos,
			struct cgroup_subsys_state *css);

struct cgroup_subsys_state *
css_rightmost_descendant(struct cgroup_subsys_state *pos);
804 805

/**
806 807 808
 * css_for_each_descendant_pre - pre-order walk of a css's descendants
 * @pos: the css * to use as the loop cursor
 * @root: css whose descendants to walk
809
 *
810 811
 * Walk @root's descendants.  Must be called under rcu_read_lock().  A
 * descendant css which hasn't finished ->css_online() or already has
812
 * finished ->css_offline() may show up during traversal and it's each
813 814
 * subsystem's responsibility to verify that each @pos is alive.
 *
815 816
 * If a subsystem synchronizes against the parent in its ->css_online() and
 * before starting iterating, and synchronizes against @pos on each
817
 * iteration, any descendant css which finished ->css_online() is
818 819 820 821 822
 * guaranteed to be visible in the future iterations.
 *
 * In other words, the following guarantees that a descendant can't escape
 * state updates of its ancestors.
 *
823
 * my_online(@css)
824
 * {
825 826
 *	Lock @css's parent and @css;
 *	Inherit state from the parent;
827 828 829
 *	Unlock both.
 * }
 *
830
 * my_update_state(@css)
831
 * {
832 833 834
 *	Lock @css;
 *	Update @css's state;
 *	Unlock @css;
835
 *
836
 *	css_for_each_descendant_pre(@pos, @css) {
837
 *		Lock @pos;
838
 *		Verify @pos is alive and inherit state from @pos's parent;
839 840 841 842 843 844 845 846 847 848
 *		Unlock @pos;
 *	}
 * }
 *
 * As long as the inheriting step, including checking the parent state, is
 * enclosed inside @pos locking, double-locking the parent isn't necessary
 * while inheriting.  The state update to the parent is guaranteed to be
 * visible by walking order and, as long as inheriting operations to the
 * same @pos are atomic to each other, multiple updates racing each other
 * still result in the correct state.  It's guaranateed that at least one
849
 * inheritance happens for any css after the latest update to its parent.
850 851 852 853 854
 *
 * If checking parent's state requires locking the parent, each inheriting
 * iteration should lock and unlock both @pos->parent and @pos.
 *
 * Alternatively, a subsystem may choose to use a single global lock to
855
 * synchronize ->css_online() and ->css_offline() against tree-walking
856
 * operations.
857 858 859 860
 *
 * It is allowed to temporarily drop RCU read lock during iteration.  The
 * caller is responsible for ensuring that @pos remains accessible until
 * the start of the next iteration by, for example, bumping the css refcnt.
861
 */
862 863 864
#define css_for_each_descendant_pre(pos, css)				\
	for ((pos) = css_next_descendant_pre(NULL, (css)); (pos);	\
	     (pos) = css_next_descendant_pre((pos), (css)))
865

866 867 868
struct cgroup_subsys_state *
css_next_descendant_post(struct cgroup_subsys_state *pos,
			 struct cgroup_subsys_state *css);
869 870

/**
871 872 873
 * css_for_each_descendant_post - post-order walk of a css's descendants
 * @pos: the css * to use as the loop cursor
 * @css: css whose descendants to walk
874
 *
875
 * Similar to css_for_each_descendant_pre() but performs post-order
876 877 878
 * traversal instead.  Note that the walk visibility guarantee described in
 * pre-order walk doesn't apply the same to post-order walks.
 */
879 880 881
#define css_for_each_descendant_post(pos, css)				\
	for ((pos) = css_next_descendant_post(NULL, (css)); (pos);	\
	     (pos) = css_next_descendant_post((pos), (css)))
882

883 884
/* A cgroup_task_iter should be treated as an opaque object */
struct cgroup_task_iter {
885
	struct cgroup			*origin_cgrp;
886 887
	struct list_head		*cset_link;
	struct list_head		*task;
888 889
};

890
void cgroup_task_iter_start(struct cgroup *cgrp, struct cgroup_task_iter *it);
891 892
struct task_struct *cgroup_task_iter_next(struct cgroup_task_iter *it);
void cgroup_task_iter_end(struct cgroup_task_iter *it);
T
Tejun Heo 已提交
893 894 895 896 897 898

int cgroup_scan_tasks(struct cgroup *cgrp,
		      bool (*test)(struct task_struct *, void *),
		      void (*process)(struct task_struct *, void *),
		      void *data, struct ptr_heap *heap);

M
Michael S. Tsirkin 已提交
899
int cgroup_attach_task_all(struct task_struct *from, struct task_struct *);
900
int cgroup_transfer_tasks(struct cgroup *to, struct cgroup *from);
M
Michael S. Tsirkin 已提交
901

K
KAMEZAWA Hiroyuki 已提交
902 903 904 905 906 907 908 909
/*
 * CSS ID is ID for cgroup_subsys_state structs under subsys. This only works
 * if cgroup_subsys.use_id == true. It can be used for looking up and scanning.
 * CSS ID is assigned at cgroup allocation (create) automatically
 * and removed when subsys calls free_css_id() function. This is because
 * the lifetime of cgroup_subsys_state is subsys's matter.
 *
 * Looking up and scanning function should be called under rcu_read_lock().
L
Li Zefan 已提交
910
 * Taking cgroup_mutex is not necessary for following calls.
K
KAMEZAWA Hiroyuki 已提交
911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926
 * But the css returned by this routine can be "not populated yet" or "being
 * destroyed". The caller should check css and cgroup's status.
 */

/*
 * Typically Called at ->destroy(), or somewhere the subsys frees
 * cgroup_subsys_state.
 */
void free_css_id(struct cgroup_subsys *ss, struct cgroup_subsys_state *css);

/* Find a cgroup_subsys_state which has given ID */

struct cgroup_subsys_state *css_lookup(struct cgroup_subsys *ss, int id);

/* Returns true if root is ancestor of cg */
bool css_is_ancestor(struct cgroup_subsys_state *cg,
927
		     const struct cgroup_subsys_state *root);
K
KAMEZAWA Hiroyuki 已提交
928 929 930

/* Get id and depth of css */
unsigned short css_id(struct cgroup_subsys_state *css);
S
Stephane Eranian 已提交
931
struct cgroup_subsys_state *cgroup_css_from_dir(struct file *f, int id);
K
KAMEZAWA Hiroyuki 已提交
932

933 934 935 936
#else /* !CONFIG_CGROUPS */

static inline int cgroup_init_early(void) { return 0; }
static inline int cgroup_init(void) { return 0; }
937
static inline void cgroup_fork(struct task_struct *p) {}
938
static inline void cgroup_post_fork(struct task_struct *p) {}
939
static inline void cgroup_exit(struct task_struct *p, int callbacks) {}
940

B
Balbir Singh 已提交
941 942 943 944 945
static inline int cgroupstats_build(struct cgroupstats *stats,
					struct dentry *dentry)
{
	return -EINVAL;
}
946

947
/* No cgroups - nothing to do */
M
Michael S. Tsirkin 已提交
948 949 950 951 952
static inline int cgroup_attach_task_all(struct task_struct *from,
					 struct task_struct *t)
{
	return 0;
}
953

954 955 956
#endif /* !CONFIG_CGROUPS */

#endif /* _LINUX_CGROUP_H */