[net-next-2.6.git] / kernel / cgroup_freezer.c

/*
 * cgroup_freezer.c -  control group freezer subsystem
 *
 * Copyright IBM Corporation, 2007
 *
 * Author : Cedric Le Goater <clg@fr.ibm.com>
 *
 * This program is free software; you can redistribute it and/or modify it
 * under the terms of version 2.1 of the GNU Lesser General Public License
 * as published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it would be useful, but
 * WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
 */

#include <linux/module.h>
#include <linux/cgroup.h>
#include <linux/fs.h>
#include <linux/uaccess.h>
#include <linux/freezer.h>
#include <linux/seq_file.h>

enum freezer_state {
	CGROUP_THAWED = 0,
	CGROUP_FREEZING,
	CGROUP_FROZEN,
};

struct freezer {
	struct cgroup_subsys_state css;
	enum freezer_state state;
	spinlock_t lock; /* protects _writes_ to state */
};

static inline struct freezer *cgroup_freezer(
		struct cgroup *cgroup)
{
	return container_of(
		cgroup_subsys_state(cgroup, freezer_subsys_id),
		struct freezer, css);
}

static inline struct freezer *task_freezer(struct task_struct *task)
{
	return container_of(task_subsys_state(task, freezer_subsys_id),
			    struct freezer, css);
}

int cgroup_frozen(struct task_struct *task)
{
	struct freezer *freezer;
	enum freezer_state state;

	task_lock(task);
	freezer = task_freezer(task);
	state = freezer->state;
	task_unlock(task);

	return state == CGROUP_FROZEN;
}

/*
 * cgroups_write_string() limits the size of freezer state strings to
 * CGROUP_LOCAL_BUFFER_SIZE
 */
static const char *freezer_state_strs[] = {
	"THAWED",
	"FREEZING",
	"FROZEN",
};

/*
 * State diagram
 * Transitions are caused by userspace writes to the freezer.state file.
 * The values in parenthesis are state labels. The rest are edge labels.
 *
 * (THAWED) --FROZEN--> (FREEZING) --FROZEN--> (FROZEN)
 *    ^ ^                    |                     |
 *    | \_______THAWED_______/                     |
 *    \__________________________THAWED____________/
 */

struct cgroup_subsys freezer_subsys;

/* Locks taken and their ordering
 * ------------------------------
 * css_set_lock
 * cgroup_mutex (AKA cgroup_lock)
 * task->alloc_lock (AKA task_lock)
 * freezer->lock
 * task->sighand->siglock
 *
 * cgroup code forces css_set_lock to be taken before task->alloc_lock
 *
 * freezer_create(), freezer_destroy():
 * cgroup_mutex [ by cgroup core ]
 *
 * can_attach():
 * cgroup_mutex
 *
 * cgroup_frozen():
 * task->alloc_lock (to get task's cgroup)
 *
 * freezer_fork() (preserving fork() performance means can't take cgroup_mutex):
 * task->alloc_lock (to get task's cgroup)
 * freezer->lock
 *  sighand->siglock (if the cgroup is freezing)
 *
 * freezer_read():
 * cgroup_mutex
 *  freezer->lock
 *   read_lock css_set_lock (cgroup iterator start)
 *
 * freezer_write() (freeze):
 * cgroup_mutex
 *  freezer->lock
 *   read_lock css_set_lock (cgroup iterator start)
 *    sighand->siglock
 *
 * freezer_write() (unfreeze):
 * cgroup_mutex
 *  freezer->lock
 *   read_lock css_set_lock (cgroup iterator start)
 *    task->alloc_lock (to prevent races with freeze_task())
 *     sighand->siglock
 */
static struct cgroup_subsys_state *freezer_create(struct cgroup_subsys *ss,
						  struct cgroup *cgroup)
{
	struct freezer *freezer;

	freezer = kzalloc(sizeof(struct freezer), GFP_KERNEL);
	if (!freezer)
		return ERR_PTR(-ENOMEM);

	spin_lock_init(&freezer->lock);
	freezer->state = CGROUP_THAWED;
	return &freezer->css;
}

static void freezer_destroy(struct cgroup_subsys *ss,
			    struct cgroup *cgroup)
{
	kfree(cgroup_freezer(cgroup));
}

/* Task is frozen or will freeze immediately when next it gets woken */
static bool is_task_frozen_enough(struct task_struct *task)
{
	return frozen(task) ||
		(task_is_stopped_or_traced(task) && freezing(task));
}

/*
 * The call to cgroup_lock() in the freezer.state write method prevents
 * a write to that file racing against an attach, and hence the
 * can_attach() result will remain valid until the attach completes.
 */
static int freezer_can_attach(struct cgroup_subsys *ss,
			      struct cgroup *new_cgroup,
			      struct task_struct *task)
{
	struct freezer *freezer;

	/*
	 * Anything frozen can't move or be moved to/from.
	 *
	 * Since orig_freezer->state == FROZEN means that @task has been
	 * frozen, so it's sufficient to check the latter condition.
	 */

	if (is_task_frozen_enough(task))
		return -EBUSY;

	freezer = cgroup_freezer(new_cgroup);
	if (freezer->state == CGROUP_FROZEN)
		return -EBUSY;

	return 0;
}

static void freezer_fork(struct cgroup_subsys *ss, struct task_struct *task)
{
	struct freezer *freezer;

	/*
	 * No lock is needed, since the task isn't on tasklist yet,
	 * so it can't be moved to another cgroup, which means the
	 * freezer won't be removed and will be valid during this
	 * function call.
	 */
	freezer = task_freezer(task);

	spin_lock_irq(&freezer->lock);
	BUG_ON(freezer->state == CGROUP_FROZEN);

	/* Locking avoids race with FREEZING -> THAWED transitions. */
	if (freezer->state == CGROUP_FREEZING)
		freeze_task(task, true);
	spin_unlock_irq(&freezer->lock);
}

/*
 * caller must hold freezer->lock
 */
static void update_freezer_state(struct cgroup *cgroup,
				 struct freezer *freezer)
{
	struct cgroup_iter it;
	struct task_struct *task;
	unsigned int nfrozen = 0, ntotal = 0;

	cgroup_iter_start(cgroup, &it);
	while ((task = cgroup_iter_next(cgroup, &it))) {
		ntotal++;
		if (is_task_frozen_enough(task))
			nfrozen++;
	}

	/*
	 * Transition to FROZEN when no new tasks can be added ensures
	 * that we never exist in the FROZEN state while there are unfrozen
	 * tasks.
	 */
	if (nfrozen == ntotal)
		freezer->state = CGROUP_FROZEN;
	else if (nfrozen > 0)
		freezer->state = CGROUP_FREEZING;
	else
		freezer->state = CGROUP_THAWED;
	cgroup_iter_end(cgroup, &it);
}

static int freezer_read(struct cgroup *cgroup, struct cftype *cft,
			struct seq_file *m)
{
	struct freezer *freezer;
	enum freezer_state state;

	if (!cgroup_lock_live_group(cgroup))
		return -ENODEV;

	freezer = cgroup_freezer(cgroup);
	spin_lock_irq(&freezer->lock);
	state = freezer->state;
	if (state == CGROUP_FREEZING) {
		/* We change from FREEZING to FROZEN lazily if the cgroup was
		 * only partially frozen when we exitted write. */
		update_freezer_state(cgroup, freezer);
		state = freezer->state;
	}
	spin_unlock_irq(&freezer->lock);
	cgroup_unlock();

	seq_puts(m, freezer_state_strs[state]);
	seq_putc(m, '\n');
	return 0;
}

static int try_to_freeze_cgroup(struct cgroup *cgroup, struct freezer *freezer)
{
	struct cgroup_iter it;
	struct task_struct *task;
	unsigned int num_cant_freeze_now = 0;

	freezer->state = CGROUP_FREEZING;
	cgroup_iter_start(cgroup, &it);
	while ((task = cgroup_iter_next(cgroup, &it))) {
		if (!freeze_task(task, true))
			continue;
		if (is_task_frozen_enough(task))
			continue;
		if (!freezing(task) && !freezer_should_skip(task))
			num_cant_freeze_now++;
	}
	cgroup_iter_end(cgroup, &it);

	return num_cant_freeze_now ? -EBUSY : 0;
}

static void unfreeze_cgroup(struct cgroup *cgroup, struct freezer *freezer)
{
	struct cgroup_iter it;
	struct task_struct *task;

	cgroup_iter_start(cgroup, &it);
	while ((task = cgroup_iter_next(cgroup, &it))) {
		thaw_process(task);
	}
	cgroup_iter_end(cgroup, &it);

	freezer->state = CGROUP_THAWED;
}

static int freezer_change_state(struct cgroup *cgroup,
				enum freezer_state goal_state)
{
	struct freezer *freezer;
	int retval = 0;

	freezer = cgroup_freezer(cgroup);

	spin_lock_irq(&freezer->lock);

	update_freezer_state(cgroup, freezer);
	if (goal_state == freezer->state)
		goto out;

	switch (goal_state) {
	case CGROUP_THAWED:
		unfreeze_cgroup(cgroup, freezer);
		break;
	case CGROUP_FROZEN:
		retval = try_to_freeze_cgroup(cgroup, freezer);
		break;
	default:
		BUG();
	}
out:
	spin_unlock_irq(&freezer->lock);

	return retval;
}

static int freezer_write(struct cgroup *cgroup,
			 struct cftype *cft,
			 const char *buffer)
{
	int retval;
	enum freezer_state goal_state;

	if (strcmp(buffer, freezer_state_strs[CGROUP_THAWED]) == 0)
		goal_state = CGROUP_THAWED;
	else if (strcmp(buffer, freezer_state_strs[CGROUP_FROZEN]) == 0)
		goal_state = CGROUP_FROZEN;
	else
		return -EIO;

	if (!cgroup_lock_live_group(cgroup))
		return -ENODEV;
	retval = freezer_change_state(cgroup, goal_state);
	cgroup_unlock();
	return retval;
}

static struct cftype files[] = {
	{
		.name = "state",
		.read_seq_string = freezer_read,
		.write_string = freezer_write,
	},
};

static int freezer_populate(struct cgroup_subsys *ss, struct cgroup *cgroup)
{
	return cgroup_add_files(cgroup, ss, files, ARRAY_SIZE(files));
}

struct cgroup_subsys freezer_subsys = {
	.name		= "freezer",
	.create		= freezer_create,
	.destroy	= freezer_destroy,
	.populate	= freezer_populate,
	.subsys_id	= freezer_subsys_id,
	.can_attach	= freezer_can_attach,
	.attach		= NULL,
	.fork		= freezer_fork,
	.exit		= NULL,
};
Commit	Line	Data
dc52ddc0 MH	1	/*
	2	* cgroup_freezer.c - control group freezer subsystem
	3	*
	4	* Copyright IBM Corporation, 2007
	5	*
	6	* Author : Cedric Le Goater <clg@fr.ibm.com>
	7	*
	8	* This program is free software; you can redistribute it and/or modify it
	9	* under the terms of version 2.1 of the GNU Lesser General Public License
	10	* as published by the Free Software Foundation.
	11	*
	12	* This program is distributed in the hope that it would be useful, but
	13	* WITHOUT ANY WARRANTY; without even the implied warranty of
	14	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
	15	*/
	16
	17	#include <linux/module.h>
	18	#include <linux/cgroup.h>
	19	#include <linux/fs.h>
	20	#include <linux/uaccess.h>
	21	#include <linux/freezer.h>
	22	#include <linux/seq_file.h>
	23
	24	enum freezer_state {
81dcf33c MH	25	CGROUP_THAWED = 0,
	26	CGROUP_FREEZING,
	27	CGROUP_FROZEN,
dc52ddc0 MH	28	};
	29
	30	struct freezer {
	31	struct cgroup_subsys_state css;
	32	enum freezer_state state;
	33	spinlock_t lock; /* protects _writes_ to state */
	34	};
	35
	36	static inline struct freezer *cgroup_freezer(
	37	struct cgroup *cgroup)
	38	{
	39	return container_of(
	40	cgroup_subsys_state(cgroup, freezer_subsys_id),
	41	struct freezer, css);
	42	}
	43
	44	static inline struct freezer task_freezer(struct task_struct task)
	45	{
	46	return container_of(task_subsys_state(task, freezer_subsys_id),
	47	struct freezer, css);
	48	}
	49
	50	int cgroup_frozen(struct task_struct *task)
	51	{
	52	struct freezer *freezer;
	53	enum freezer_state state;
	54
	55	task_lock(task);
	56	freezer = task_freezer(task);
	57	state = freezer->state;
	58	task_unlock(task);
	59
81dcf33c	60	return state == CGROUP_FROZEN;
dc52ddc0 MH	61	}
	62
	63	/*
	64	* cgroups_write_string() limits the size of freezer state strings to
	65	* CGROUP_LOCAL_BUFFER_SIZE
	66	*/
	67	static const char *freezer_state_strs[] = {
81dcf33c	68	"THAWED",
dc52ddc0 MH	69	"FREEZING",
	70	"FROZEN",
	71	};
	72
	73	/*
	74	* State diagram
	75	* Transitions are caused by userspace writes to the freezer.state file.
	76	* The values in parenthesis are state labels. The rest are edge labels.
	77	*
81dcf33c MH	78	* (THAWED) --FROZEN--> (FREEZING) --FROZEN--> (FROZEN)
	79	* ^ ^ \| \|
	80	* \| \_______THAWED_______/ \|
	81	* \__________________________THAWED____________/
dc52ddc0 MH	82	*/
	83
	84	struct cgroup_subsys freezer_subsys;
	85
	86	/* Locks taken and their ordering
	87	* ------------------------------
	88	* css_set_lock
	89	* cgroup_mutex (AKA cgroup_lock)
	90	* task->alloc_lock (AKA task_lock)
	91	* freezer->lock
	92	* task->sighand->siglock
	93	*
	94	* cgroup code forces css_set_lock to be taken before task->alloc_lock
	95	*
	96	* freezer_create(), freezer_destroy():
	97	* cgroup_mutex [ by cgroup core ]
	98	*
	99	* can_attach():
	100	* cgroup_mutex
	101	*
	102	* cgroup_frozen():
	103	* task->alloc_lock (to get task's cgroup)
	104	*
	105	* freezer_fork() (preserving fork() performance means can't take cgroup_mutex):
	106	* task->alloc_lock (to get task's cgroup)
	107	* freezer->lock
	108	* sighand->siglock (if the cgroup is freezing)
	109	*
	110	* freezer_read():
	111	* cgroup_mutex
	112	* freezer->lock
	113	* read_lock css_set_lock (cgroup iterator start)
	114	*
	115	* freezer_write() (freeze):
	116	* cgroup_mutex
	117	* freezer->lock
	118	* read_lock css_set_lock (cgroup iterator start)
	119	* sighand->siglock
	120	*
	121	* freezer_write() (unfreeze):
	122	* cgroup_mutex
	123	* freezer->lock
	124	* read_lock css_set_lock (cgroup iterator start)
	125	* task->alloc_lock (to prevent races with freeze_task())
	126	* sighand->siglock
	127	*/
	128	static struct cgroup_subsys_state freezer_create(struct cgroup_subsys ss,
	129	struct cgroup *cgroup)
	130	{
	131	struct freezer *freezer;
	132
	133	freezer = kzalloc(sizeof(struct freezer), GFP_KERNEL);
	134	if (!freezer)
	135	return ERR_PTR(-ENOMEM);
	136
	137	spin_lock_init(&freezer->lock);
81dcf33c	138	freezer->state = CGROUP_THAWED;
dc52ddc0 MH	139	return &freezer->css;
	140	}
	141
	142	static void freezer_destroy(struct cgroup_subsys *ss,
	143	struct cgroup *cgroup)
	144	{
	145	kfree(cgroup_freezer(cgroup));
	146	}
	147
957a4eea MH	148	/* Task is frozen or will freeze immediately when next it gets woken */
	149	static bool is_task_frozen_enough(struct task_struct *task)
	150	{
	151	return frozen(task) \|\|
	152	(task_is_stopped_or_traced(task) && freezing(task));
	153	}
dc52ddc0	154
957a4eea MH	155	/*
	156	* The call to cgroup_lock() in the freezer.state write method prevents
	157	* a write to that file racing against an attach, and hence the
	158	* can_attach() result will remain valid until the attach completes.
	159	*/
dc52ddc0 MH	160	static int freezer_can_attach(struct cgroup_subsys *ss,
	161	struct cgroup *new_cgroup,
	162	struct task_struct *task)
	163	{
	164	struct freezer *freezer;
957a4eea	165
80a6a2cf LZ	166	/*
	167	* Anything frozen can't move or be moved to/from.
	168	*
	169	* Since orig_freezer->state == FROZEN means that @task has been
	170	* frozen, so it's sufficient to check the latter condition.
	171	*/
957a4eea MH	172
	173	if (is_task_frozen_enough(task))
	174	return -EBUSY;
dc52ddc0	175
dc52ddc0	176	freezer = cgroup_freezer(new_cgroup);
81dcf33c	177	if (freezer->state == CGROUP_FROZEN)
957a4eea MH	178	return -EBUSY;
957a4eea MH	179
80a6a2cf	180	return 0;
dc52ddc0 MH	181	}
	182
	183	static void freezer_fork(struct cgroup_subsys ss, struct task_struct task)
	184	{
	185	struct freezer *freezer;
	186
68744676 LZ	187	/*
	188	* No lock is needed, since the task isn't on tasklist yet,
	189	* so it can't be moved to another cgroup, which means the
	190	* freezer won't be removed and will be valid during this
	191	* function call.
	192	*/
dc52ddc0	193	freezer = task_freezer(task);
dc52ddc0	194
dc52ddc0	195	spin_lock_irq(&freezer->lock);
7ccb9743 LZ	196	BUG_ON(freezer->state == CGROUP_FROZEN);
7ccb9743 LZ	197
81dcf33c MH	198	/* Locking avoids race with FREEZING -> THAWED transitions. */
81dcf33c MH	199	if (freezer->state == CGROUP_FREEZING)
dc52ddc0 MH	200	freeze_task(task, true);
	201	spin_unlock_irq(&freezer->lock);
	202	}
	203
	204	/*
	205	* caller must hold freezer->lock
	206	*/
1aece348 MH	207	static void update_freezer_state(struct cgroup *cgroup,
1aece348 MH	208	struct freezer *freezer)
dc52ddc0 MH	209	{
	210	struct cgroup_iter it;
	211	struct task_struct *task;
	212	unsigned int nfrozen = 0, ntotal = 0;
	213
	214	cgroup_iter_start(cgroup, &it);
	215	while ((task = cgroup_iter_next(cgroup, &it))) {
	216	ntotal++;
957a4eea	217	if (is_task_frozen_enough(task))
dc52ddc0 MH	218	nfrozen++;
	219	}
	220
	221	/*
	222	* Transition to FROZEN when no new tasks can be added ensures
	223	* that we never exist in the FROZEN state while there are unfrozen
	224	* tasks.
	225	*/
	226	if (nfrozen == ntotal)
81dcf33c	227	freezer->state = CGROUP_FROZEN;
1aece348 MH	228	else if (nfrozen > 0)
	229	freezer->state = CGROUP_FREEZING;
	230	else
	231	freezer->state = CGROUP_THAWED;
dc52ddc0 MH	232	cgroup_iter_end(cgroup, &it);
	233	}
	234
	235	static int freezer_read(struct cgroup cgroup, struct cftype cft,
	236	struct seq_file *m)
	237	{
	238	struct freezer *freezer;
	239	enum freezer_state state;
	240
	241	if (!cgroup_lock_live_group(cgroup))
	242	return -ENODEV;
	243
	244	freezer = cgroup_freezer(cgroup);
	245	spin_lock_irq(&freezer->lock);
	246	state = freezer->state;
81dcf33c	247	if (state == CGROUP_FREEZING) {
dc52ddc0 MH	248	/* We change from FREEZING to FROZEN lazily if the cgroup was
dc52ddc0 MH	249	* only partially frozen when we exitted write. */
1aece348	250	update_freezer_state(cgroup, freezer);
dc52ddc0 MH	251	state = freezer->state;
	252	}
	253	spin_unlock_irq(&freezer->lock);
	254	cgroup_unlock();
	255
	256	seq_puts(m, freezer_state_strs[state]);
	257	seq_putc(m, '\n');
	258	return 0;
	259	}
	260
	261	static int try_to_freeze_cgroup(struct cgroup cgroup, struct freezer freezer)
	262	{
	263	struct cgroup_iter it;
	264	struct task_struct *task;
	265	unsigned int num_cant_freeze_now = 0;
	266
81dcf33c	267	freezer->state = CGROUP_FREEZING;
dc52ddc0 MH	268	cgroup_iter_start(cgroup, &it);
	269	while ((task = cgroup_iter_next(cgroup, &it))) {
	270	if (!freeze_task(task, true))
	271	continue;
957a4eea	272	if (is_task_frozen_enough(task))
dc52ddc0 MH	273	continue;
	274	if (!freezing(task) && !freezer_should_skip(task))
	275	num_cant_freeze_now++;
	276	}
	277	cgroup_iter_end(cgroup, &it);
	278
	279	return num_cant_freeze_now ? -EBUSY : 0;
	280	}
	281
00c2e63c	282	static void unfreeze_cgroup(struct cgroup cgroup, struct freezer freezer)
dc52ddc0 MH	283	{
	284	struct cgroup_iter it;
	285	struct task_struct *task;
	286
	287	cgroup_iter_start(cgroup, &it);
	288	while ((task = cgroup_iter_next(cgroup, &it))) {
00c2e63c	289	thaw_process(task);
dc52ddc0 MH	290	}
dc52ddc0 MH	291	cgroup_iter_end(cgroup, &it);
dc52ddc0	292
00c2e63c	293	freezer->state = CGROUP_THAWED;
dc52ddc0 MH	294	}
	295
	296	static int freezer_change_state(struct cgroup *cgroup,
	297	enum freezer_state goal_state)
	298	{
	299	struct freezer *freezer;
	300	int retval = 0;
	301
	302	freezer = cgroup_freezer(cgroup);
51308ee5	303
dc52ddc0	304	spin_lock_irq(&freezer->lock);
51308ee5	305
1aece348	306	update_freezer_state(cgroup, freezer);
dc52ddc0 MH	307	if (goal_state == freezer->state)
dc52ddc0 MH	308	goto out;
51308ee5 LZ	309
51308ee5 LZ	310	switch (goal_state) {
81dcf33c	311	case CGROUP_THAWED:
51308ee5	312	unfreeze_cgroup(cgroup, freezer);
dc52ddc0	313	break;
81dcf33c	314	case CGROUP_FROZEN:
51308ee5	315	retval = try_to_freeze_cgroup(cgroup, freezer);
dc52ddc0 MH	316	break;
dc52ddc0 MH	317	default:
51308ee5	318	BUG();
dc52ddc0 MH	319	}
	320	out:
	321	spin_unlock_irq(&freezer->lock);
	322
	323	return retval;
	324	}
	325
	326	static int freezer_write(struct cgroup *cgroup,
	327	struct cftype *cft,
	328	const char *buffer)
	329	{
	330	int retval;
	331	enum freezer_state goal_state;
	332
81dcf33c MH	333	if (strcmp(buffer, freezer_state_strs[CGROUP_THAWED]) == 0)
	334	goal_state = CGROUP_THAWED;
	335	else if (strcmp(buffer, freezer_state_strs[CGROUP_FROZEN]) == 0)
	336	goal_state = CGROUP_FROZEN;
dc52ddc0 MH	337	else
	338	return -EIO;
	339
	340	if (!cgroup_lock_live_group(cgroup))
	341	return -ENODEV;
	342	retval = freezer_change_state(cgroup, goal_state);
	343	cgroup_unlock();
	344	return retval;
	345	}
	346
	347	static struct cftype files[] = {
	348	{
	349	.name = "state",
	350	.read_seq_string = freezer_read,
	351	.write_string = freezer_write,
	352	},
	353	};
	354
	355	static int freezer_populate(struct cgroup_subsys ss, struct cgroup cgroup)
	356	{
	357	return cgroup_add_files(cgroup, ss, files, ARRAY_SIZE(files));
	358	}
	359
	360	struct cgroup_subsys freezer_subsys = {
	361	.name = "freezer",
	362	.create = freezer_create,
	363	.destroy = freezer_destroy,
	364	.populate = freezer_populate,
	365	.subsys_id = freezer_subsys_id,
	366	.can_attach = freezer_can_attach,
	367	.attach = NULL,
	368	.fork = freezer_fork,
	369	.exit = NULL,
	370	};