cgroup.h 17.4 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13
#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>
14
#include <linux/rculist.h>
B
Balbir Singh 已提交
15
#include <linux/cgroupstats.h>
16
#include <linux/rwsem.h>
17
#include <linux/fs.h>
18
#include <linux/seq_file.h>
T
Tejun Heo 已提交
19
#include <linux/kernfs.h>
20

21
#include <linux/cgroup-defs.h>
22 23 24

#ifdef CONFIG_CGROUPS

25 26 27 28 29 30 31 32 33
/*
 * All weight knobs on the default hierarhcy should use the following min,
 * default and max values.  The default value is the logarithmic center of
 * MIN and MAX and allows 100x to be expressed in both directions.
 */
#define CGROUP_WEIGHT_MIN		1
#define CGROUP_WEIGHT_DFL		100
#define CGROUP_WEIGHT_MAX		10000

34 35 36 37 38 39
/* a css_task_iter should be treated as an opaque object */
struct css_task_iter {
	struct cgroup_subsys		*ss;

	struct list_head		*cset_pos;
	struct list_head		*cset_head;
40

41 42 43 44
	struct list_head		*task_pos;
	struct list_head		*tasks_head;
	struct list_head		*mg_tasks_head;
};
45

46 47
extern struct cgroup_root cgrp_dfl_root;
extern struct css_set init_css_set;
48

49
#define SUBSYS(_x) extern struct cgroup_subsys _x ## _cgrp_subsys;
50 51 52
#include <linux/cgroup_subsys.h>
#undef SUBSYS

53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79
bool css_has_online_children(struct cgroup_subsys_state *css);
struct cgroup_subsys_state *css_from_id(int id, struct cgroup_subsys *ss);
struct cgroup_subsys_state *cgroup_get_e_css(struct cgroup *cgroup,
					     struct cgroup_subsys *ss);
struct cgroup_subsys_state *css_tryget_online_from_dir(struct dentry *dentry,
						       struct cgroup_subsys *ss);

bool cgroup_is_descendant(struct cgroup *cgrp, struct cgroup *ancestor);
int cgroup_attach_task_all(struct task_struct *from, struct task_struct *);
int cgroup_transfer_tasks(struct cgroup *to, struct cgroup *from);

int cgroup_add_dfl_cftypes(struct cgroup_subsys *ss, struct cftype *cfts);
int cgroup_add_legacy_cftypes(struct cgroup_subsys *ss, struct cftype *cfts);
int cgroup_rm_cftypes(struct cftype *cfts);

char *task_cgroup_path(struct task_struct *task, char *buf, size_t buflen);
int cgroupstats_build(struct cgroupstats *stats, struct dentry *dentry);
int proc_cgroup_show(struct seq_file *m, struct pid_namespace *ns,
		     struct pid *pid, struct task_struct *tsk);

void cgroup_fork(struct task_struct *p);
void cgroup_post_fork(struct task_struct *p);
void cgroup_exit(struct task_struct *p);

int cgroup_init_early(void);
int cgroup_init(void);

T
Tejun Heo 已提交
80
/*
81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112
 * Iteration helpers and macros.
 */

struct cgroup_subsys_state *css_next_child(struct cgroup_subsys_state *pos,
					   struct cgroup_subsys_state *parent);
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);
struct cgroup_subsys_state *css_next_descendant_post(struct cgroup_subsys_state *pos,
						     struct cgroup_subsys_state *css);

struct task_struct *cgroup_taskset_first(struct cgroup_taskset *tset);
struct task_struct *cgroup_taskset_next(struct cgroup_taskset *tset);

void css_task_iter_start(struct cgroup_subsys_state *css,
			 struct css_task_iter *it);
struct task_struct *css_task_iter_next(struct css_task_iter *it);
void css_task_iter_end(struct css_task_iter *it);

/**
 * 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
 *
 * Walk @parent's children.  Must be called under rcu_read_lock().
 *
 * If a subsystem synchronizes ->css_online() and the start of iteration, a
 * css which finished ->css_online() is guaranteed to be visible in the
 * future iterations and will stay visible until the last reference is put.
 * A css which hasn't finished ->css_online() or already finished
 * ->css_offline() may show up during traversal.  It's each subsystem's
 * responsibility to synchronize against on/offlining.
T
Tejun Heo 已提交
113
 *
114 115 116
 * 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.
T
Tejun Heo 已提交
117
 */
118 119 120
#define css_for_each_child(pos, parent)					\
	for ((pos) = css_next_child(NULL, (parent)); (pos);		\
	     (pos) = css_next_child((pos), (parent)))
121

122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212
/**
 * 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
 *
 * Walk @root's descendants.  @root is included in the iteration and the
 * first node to be visited.  Must be called under rcu_read_lock().
 *
 * If a subsystem synchronizes ->css_online() and the start of iteration, a
 * css which finished ->css_online() is guaranteed to be visible in the
 * future iterations and will stay visible until the last reference is put.
 * A css which hasn't finished ->css_online() or already finished
 * ->css_offline() may show up during traversal.  It's each subsystem's
 * responsibility to synchronize against on/offlining.
 *
 * For example, the following guarantees that a descendant can't escape
 * state updates of its ancestors.
 *
 * my_online(@css)
 * {
 *	Lock @css's parent and @css;
 *	Inherit state from the parent;
 *	Unlock both.
 * }
 *
 * my_update_state(@css)
 * {
 *	css_for_each_descendant_pre(@pos, @css) {
 *		Lock @pos;
 *		if (@pos == @css)
 *			Update @css's state;
 *		else
 *			Verify @pos is alive and inherit state from its parent;
 *		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
 * inheritance happens for any css after the latest update to its parent.
 *
 * 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
 * synchronize ->css_online() and ->css_offline() against tree-walking
 * operations.
 *
 * 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.
 */
#define css_for_each_descendant_pre(pos, css)				\
	for ((pos) = css_next_descendant_pre(NULL, (css)); (pos);	\
	     (pos) = css_next_descendant_pre((pos), (css)))

/**
 * 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
 *
 * Similar to css_for_each_descendant_pre() but performs post-order
 * traversal instead.  @root is included in the iteration and the last
 * node to be visited.
 *
 * If a subsystem synchronizes ->css_online() and the start of iteration, a
 * css which finished ->css_online() is guaranteed to be visible in the
 * future iterations and will stay visible until the last reference is put.
 * A css which hasn't finished ->css_online() or already finished
 * ->css_offline() may show up during traversal.  It's each subsystem's
 * responsibility to synchronize against on/offlining.
 *
 * Note that the walk visibility guarantee example described in pre-order
 * walk doesn't apply the same to post-order walks.
 */
#define css_for_each_descendant_post(pos, css)				\
	for ((pos) = css_next_descendant_post(NULL, (css)); (pos);	\
	     (pos) = css_next_descendant_post((pos), (css)))

/**
 * cgroup_taskset_for_each - iterate cgroup_taskset
 * @task: the loop cursor
 * @tset: taskset to iterate
 */
#define cgroup_taskset_for_each(task, tset)				\
	for ((task) = cgroup_taskset_first((tset)); (task);		\
	     (task) = cgroup_taskset_next((tset)))
213

214 215 216
/*
 * Inline functions.
 */
217

218 219 220 221 222
/**
 * css_get - obtain a reference on the specified css
 * @css: target css
 *
 * The caller must already have a reference.
223 224 225
 */
static inline void css_get(struct cgroup_subsys_state *css)
{
226 227
	if (!(css->flags & CSS_NO_REF))
		percpu_ref_get(&css->refcnt);
228
}
P
Paul Menage 已提交
229

230 231 232 233 234 235 236 237 238 239 240 241 242
/**
 * css_get_many - obtain references on the specified css
 * @css: target css
 * @n: number of references to get
 *
 * The caller must already have a reference.
 */
static inline void css_get_many(struct cgroup_subsys_state *css, unsigned int n)
{
	if (!(css->flags & CSS_NO_REF))
		percpu_ref_get_many(&css->refcnt, n);
}

T
Tejun Heo 已提交
243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260
/**
 * css_tryget - try to obtain a reference on the specified css
 * @css: target css
 *
 * Obtain a reference on @css unless it already has reached zero and is
 * being released.  This function doesn't care whether @css is on or
 * offline.  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.
 */
static inline bool css_tryget(struct cgroup_subsys_state *css)
{
	if (!(css->flags & CSS_NO_REF))
		return percpu_ref_tryget(&css->refcnt);
	return true;
}

261
/**
262
 * css_tryget_online - try to obtain a reference on the specified css if online
263 264
 * @css: target css
 *
265 266
 * Obtain a reference on @css if it's online.  The caller naturally needs
 * to ensure that @css is accessible but doesn't have to be holding a
267 268 269 270
 * reference on it - IOW, RCU protected access is good enough for this
 * function.  Returns %true if a reference count was successfully obtained;
 * %false otherwise.
 */
271
static inline bool css_tryget_online(struct cgroup_subsys_state *css)
P
Paul Menage 已提交
272
{
273 274 275
	if (!(css->flags & CSS_NO_REF))
		return percpu_ref_tryget_live(&css->refcnt);
	return true;
P
Paul Menage 已提交
276
}
277 278

/**
279 280 281
 * css_put - put a css reference
 * @css: target css
 *
282
 * Put a reference obtained via css_get() and css_tryget_online().
283
 */
284 285
static inline void css_put(struct cgroup_subsys_state *css)
{
286 287
	if (!(css->flags & CSS_NO_REF))
		percpu_ref_put(&css->refcnt);
288
}
289

290 291 292 293 294 295
/**
 * css_put_many - put css references
 * @css: target css
 * @n: number of references to put
 *
 * Put references obtained via css_get() and css_tryget_online().
296
 */
297 298 299 300 301
static inline void css_put_many(struct cgroup_subsys_state *css, unsigned int n)
{
	if (!(css->flags & CSS_NO_REF))
		percpu_ref_put_many(&css->refcnt, n);
}
302

303 304 305 306 307 308 309 310 311 312 313 314
/**
 * 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.
315
 */
T
Tejun Heo 已提交
316 317
#ifdef CONFIG_PROVE_RCU
extern struct mutex cgroup_mutex;
318
extern struct rw_semaphore css_set_rwsem;
319 320
#define task_css_set_check(task, __c)					\
	rcu_dereference_check((task)->cgroups,				\
321 322 323
		lockdep_is_held(&cgroup_mutex) ||			\
		lockdep_is_held(&css_set_rwsem) ||			\
		((task)->flags & PF_EXITING) || (__c))
T
Tejun Heo 已提交
324
#else
325 326
#define task_css_set_check(task, __c)					\
	rcu_dereference((task)->cgroups)
T
Tejun Heo 已提交
327
#endif
328

329
/**
330
 * task_css_check - obtain css for (task, subsys) w/ extra access conds
331 332 333 334 335 336 337
 * @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().
 */
338
#define task_css_check(task, subsys_id, __c)				\
339 340 341 342 343 344 345 346 347 348 349 350 351 352
	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);
}

/**
353
 * task_css - obtain css for (task, subsys)
354 355 356
 * @task: the target task
 * @subsys_id: the target subsystem ID
 *
357
 * See task_css_check().
358
 */
359 360
static inline struct cgroup_subsys_state *task_css(struct task_struct *task,
						   int subsys_id)
361
{
362
	return task_css_check(task, subsys_id, false);
363 364
}

365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389
/**
 * task_get_css - find and get the css for (task, subsys)
 * @task: the target task
 * @subsys_id: the target subsystem ID
 *
 * Find the css for the (@task, @subsys_id) combination, increment a
 * reference on and return it.  This function is guaranteed to return a
 * valid css.
 */
static inline struct cgroup_subsys_state *
task_get_css(struct task_struct *task, int subsys_id)
{
	struct cgroup_subsys_state *css;

	rcu_read_lock();
	while (true) {
		css = task_css(task, subsys_id);
		if (likely(css_tryget_online(css)))
			break;
		cpu_relax();
	}
	rcu_read_unlock();
	return css;
}

T
Tejun Heo 已提交
390 391 392 393 394 395 396 397 398 399 400 401 402 403
/**
 * task_css_is_root - test whether a task belongs to the root css
 * @task: the target task
 * @subsys_id: the target subsystem ID
 *
 * Test whether @task belongs to the root css on the specified subsystem.
 * May be invoked in any context.
 */
static inline bool task_css_is_root(struct task_struct *task, int subsys_id)
{
	return task_css_check(task, subsys_id, true) ==
		init_css_set.subsys[subsys_id];
}

404 405
static inline struct cgroup *task_cgroup(struct task_struct *task,
					 int subsys_id)
406
{
407
	return task_css(task, subsys_id)->cgroup;
408 409
}

410
/**
411 412
 * cgroup_on_dfl - test whether a cgroup is on the default hierarchy
 * @cgrp: the cgroup of interest
413
 *
414 415 416 417
 * The default hierarchy is the v2 interface of cgroup and this function
 * can be used to test whether a cgroup is on the default hierarchy for
 * cases where a subsystem should behave differnetly depending on the
 * interface version.
418
 *
419 420
 * The set of behaviors which change on the default hierarchy are still
 * being determined and the mount option is prefixed with __DEVEL__.
421
 *
422
 * List of changed behaviors:
423
 *
424 425
 * - Mount options "noprefix", "xattr", "clone_children", "release_agent"
 *   and "name" are disallowed.
426
 *
427
 * - When mounting an existing superblock, mount options should match.
428
 *
429
 * - Remount is disallowed.
430
 *
431
 * - rename(2) is disallowed.
432
 *
433 434
 * - "tasks" is removed.  Everything should be at process granularity.  Use
 *   "cgroup.procs" instead.
435
 *
436 437
 * - "cgroup.procs" is not sorted.  pids will be unique unless they got
 *   recycled inbetween reads.
438
 *
439 440
 * - "release_agent" and "notify_on_release" are removed.  Replacement
 *   notification mechanism will be implemented.
441
 *
442
 * - "cgroup.clone_children" is removed.
443
 *
444 445 446 447
 * - "cgroup.subtree_populated" is available.  Its value is 0 if the cgroup
 *   and its descendants contain no task; otherwise, 1.  The file also
 *   generates kernfs notification which can be monitored through poll and
 *   [di]notify when the value of the file changes.
448
 *
449 450 451
 * - 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.
452
 *
453 454
 * - cpuset: a task can be moved into an empty cpuset, and again it takes
 *   masks of ancestors.
455
 *
456 457 458 459 460 461
 * - memcg: use_hierarchy is on by default and the cgroup file for the flag
 *   is not created.
 *
 * - blkcg: blk-throttle becomes properly hierarchical.
 *
 * - debug: disallowed on the default hierarchy.
462
 */
T
Tejun Heo 已提交
463 464 465 466
static inline bool cgroup_on_dfl(const struct cgroup *cgrp)
{
	return cgrp->root == &cgrp_dfl_root;
}
467

468 469 470 471 472
/* no synchronization, the result can only be used as a hint */
static inline bool cgroup_has_tasks(struct cgroup *cgrp)
{
	return !list_empty(&cgrp->cset_links);
}
473

474
/* returns ino associated with a cgroup */
T
Tejun Heo 已提交
475 476
static inline ino_t cgroup_ino(struct cgroup *cgrp)
{
477
	return cgrp->kn->ino;
T
Tejun Heo 已提交
478
}
479

T
Tejun Heo 已提交
480 481
/* cft/css accessors for cftype->write() operation */
static inline struct cftype *of_cft(struct kernfs_open_file *of)
482
{
T
Tejun Heo 已提交
483
	return of->kn->priv;
484
}
T
Tejun Heo 已提交
485

T
Tejun Heo 已提交
486
struct cgroup_subsys_state *of_css(struct kernfs_open_file *of);
487

T
Tejun Heo 已提交
488 489 490 491 492
/* cft/css accessors for cftype->seq_*() operations */
static inline struct cftype *seq_cft(struct seq_file *seq)
{
	return of_cft(seq->private);
}
T
Tejun Heo 已提交
493

T
Tejun Heo 已提交
494 495 496 497
static inline struct cgroup_subsys_state *seq_css(struct seq_file *seq)
{
	return of_css(seq->private);
}
M
Michael S. Tsirkin 已提交
498

T
Tejun Heo 已提交
499 500 501 502
/*
 * Name / path handling functions.  All are thin wrappers around the kernfs
 * counterparts and can be called under any context.
 */
K
KAMEZAWA Hiroyuki 已提交
503

T
Tejun Heo 已提交
504 505
static inline int cgroup_name(struct cgroup *cgrp, char *buf, size_t buflen)
{
506
	return kernfs_name(cgrp->kn, buf, buflen);
T
Tejun Heo 已提交
507
}
508

T
Tejun Heo 已提交
509 510 511
static inline char * __must_check cgroup_path(struct cgroup *cgrp, char *buf,
					      size_t buflen)
{
512
	return kernfs_path(cgrp->kn, buf, buflen);
T
Tejun Heo 已提交
513
}
514

T
Tejun Heo 已提交
515 516
static inline void pr_cont_cgroup_name(struct cgroup *cgrp)
{
517
	pr_cont_kernfs_name(cgrp->kn);
T
Tejun Heo 已提交
518
}
519

T
Tejun Heo 已提交
520
static inline void pr_cont_cgroup_path(struct cgroup *cgrp)
B
Balbir Singh 已提交
521
{
522
	pr_cont_kernfs_path(cgrp->kn);
B
Balbir Singh 已提交
523
}
524 525 526

#else /* !CONFIG_CGROUPS */

527 528
struct cgroup_subsys_state;

529
static inline void css_put(struct cgroup_subsys_state *css) {}
M
Michael S. Tsirkin 已提交
530
static inline int cgroup_attach_task_all(struct task_struct *from,
531 532 533 534
					 struct task_struct *t) { return 0; }
static inline int cgroupstats_build(struct cgroupstats *stats,
				    struct dentry *dentry) { return -EINVAL; }

535
static inline void cgroup_fork(struct task_struct *p) {}
536
static inline void cgroup_post_fork(struct task_struct *p) {}
537
static inline void cgroup_exit(struct task_struct *p) {}
538

539 540
static inline int cgroup_init_early(void) { return 0; }
static inline int cgroup_init(void) { return 0; }
541

542 543 544
#endif /* !CONFIG_CGROUPS */

#endif /* _LINUX_CGROUP_H */