[net-next-2.6.git] / block / blk-cgroup.c

/*
 * Common Block IO controller cgroup interface
 *
 * Based on ideas and code from CFQ, CFS and BFQ:
 * Copyright (C) 2003 Jens Axboe <axboe@kernel.dk>
 *
 * Copyright (C) 2008 Fabio Checconi <fabio@gandalf.sssup.it>
 *		      Paolo Valente <paolo.valente@unimore.it>
 *
 * Copyright (C) 2009 Vivek Goyal <vgoyal@redhat.com>
 * 	              Nauman Rafique <nauman@google.com>
 */
#include <linux/ioprio.h>
#include <linux/seq_file.h>
#include <linux/kdev_t.h>
#include <linux/module.h>
#include <linux/err.h>
#include "blk-cgroup.h"

static DEFINE_SPINLOCK(blkio_list_lock);
static LIST_HEAD(blkio_list);

struct blkio_cgroup blkio_root_cgroup = { .weight = 2*BLKIO_WEIGHT_DEFAULT };
EXPORT_SYMBOL_GPL(blkio_root_cgroup);

static struct cgroup_subsys_state *blkiocg_create(struct cgroup_subsys *,
						  struct cgroup *);
static int blkiocg_can_attach(struct cgroup_subsys *, struct cgroup *,
			      struct task_struct *, bool);
static void blkiocg_attach(struct cgroup_subsys *, struct cgroup *,
			   struct cgroup *, struct task_struct *, bool);
static void blkiocg_destroy(struct cgroup_subsys *, struct cgroup *);
static int blkiocg_populate(struct cgroup_subsys *, struct cgroup *);

struct cgroup_subsys blkio_subsys = {
	.name = "blkio",
	.create = blkiocg_create,
	.can_attach = blkiocg_can_attach,
	.attach = blkiocg_attach,
	.destroy = blkiocg_destroy,
	.populate = blkiocg_populate,
#ifdef CONFIG_BLK_CGROUP
	/* note: blkio_subsys_id is otherwise defined in blk-cgroup.h */
	.subsys_id = blkio_subsys_id,
#endif
	.use_id = 1,
	.module = THIS_MODULE,
};
EXPORT_SYMBOL_GPL(blkio_subsys);

struct blkio_cgroup *cgroup_to_blkio_cgroup(struct cgroup *cgroup)
{
	return container_of(cgroup_subsys_state(cgroup, blkio_subsys_id),
			    struct blkio_cgroup, css);
}
EXPORT_SYMBOL_GPL(cgroup_to_blkio_cgroup);

void blkiocg_update_blkio_group_stats(struct blkio_group *blkg,
			unsigned long time, unsigned long sectors)
{
	blkg->time += time;
	blkg->sectors += sectors;
}
EXPORT_SYMBOL_GPL(blkiocg_update_blkio_group_stats);

void blkiocg_add_blkio_group(struct blkio_cgroup *blkcg,
			struct blkio_group *blkg, void *key, dev_t dev)
{
	unsigned long flags;

	spin_lock_irqsave(&blkcg->lock, flags);
	rcu_assign_pointer(blkg->key, key);
	blkg->blkcg_id = css_id(&blkcg->css);
	hlist_add_head_rcu(&blkg->blkcg_node, &blkcg->blkg_list);
	spin_unlock_irqrestore(&blkcg->lock, flags);
#ifdef CONFIG_DEBUG_BLK_CGROUP
	/* Need to take css reference ? */
	cgroup_path(blkcg->css.cgroup, blkg->path, sizeof(blkg->path));
#endif
	blkg->dev = dev;
}
EXPORT_SYMBOL_GPL(blkiocg_add_blkio_group);

static void __blkiocg_del_blkio_group(struct blkio_group *blkg)
{
	hlist_del_init_rcu(&blkg->blkcg_node);
	blkg->blkcg_id = 0;
}

/*
 * returns 0 if blkio_group was still on cgroup list. Otherwise returns 1
 * indicating that blk_group was unhashed by the time we got to it.
 */
int blkiocg_del_blkio_group(struct blkio_group *blkg)
{
	struct blkio_cgroup *blkcg;
	unsigned long flags;
	struct cgroup_subsys_state *css;
	int ret = 1;

	rcu_read_lock();
	css = css_lookup(&blkio_subsys, blkg->blkcg_id);
	if (!css)
		goto out;

	blkcg = container_of(css, struct blkio_cgroup, css);
	spin_lock_irqsave(&blkcg->lock, flags);
	if (!hlist_unhashed(&blkg->blkcg_node)) {
		__blkiocg_del_blkio_group(blkg);
		ret = 0;
	}
	spin_unlock_irqrestore(&blkcg->lock, flags);
out:
	rcu_read_unlock();
	return ret;
}
EXPORT_SYMBOL_GPL(blkiocg_del_blkio_group);

/* called under rcu_read_lock(). */
struct blkio_group *blkiocg_lookup_group(struct blkio_cgroup *blkcg, void *key)
{
	struct blkio_group *blkg;
	struct hlist_node *n;
	void *__key;

	hlist_for_each_entry_rcu(blkg, n, &blkcg->blkg_list, blkcg_node) {
		__key = blkg->key;
		if (__key == key)
			return blkg;
	}

	return NULL;
}
EXPORT_SYMBOL_GPL(blkiocg_lookup_group);

#define SHOW_FUNCTION(__VAR)						\
static u64 blkiocg_##__VAR##_read(struct cgroup *cgroup,		\
				       struct cftype *cftype)		\
{									\
	struct blkio_cgroup *blkcg;					\
									\
	blkcg = cgroup_to_blkio_cgroup(cgroup);				\
	return (u64)blkcg->__VAR;					\
}

SHOW_FUNCTION(weight);
#undef SHOW_FUNCTION

static int
blkiocg_weight_write(struct cgroup *cgroup, struct cftype *cftype, u64 val)
{
	struct blkio_cgroup *blkcg;
	struct blkio_group *blkg;
	struct hlist_node *n;
	struct blkio_policy_type *blkiop;

	if (val < BLKIO_WEIGHT_MIN || val > BLKIO_WEIGHT_MAX)
		return -EINVAL;

	blkcg = cgroup_to_blkio_cgroup(cgroup);
	spin_lock(&blkio_list_lock);
	spin_lock_irq(&blkcg->lock);
	blkcg->weight = (unsigned int)val;
	hlist_for_each_entry(blkg, n, &blkcg->blkg_list, blkcg_node) {
		list_for_each_entry(blkiop, &blkio_list, list)
			blkiop->ops.blkio_update_group_weight_fn(blkg,
					blkcg->weight);
	}
	spin_unlock_irq(&blkcg->lock);
	spin_unlock(&blkio_list_lock);
	return 0;
}

#define SHOW_FUNCTION_PER_GROUP(__VAR)					\
static int blkiocg_##__VAR##_read(struct cgroup *cgroup,		\
			struct cftype *cftype, struct seq_file *m)	\
{									\
	struct blkio_cgroup *blkcg;					\
	struct blkio_group *blkg;					\
	struct hlist_node *n;						\
									\
	if (!cgroup_lock_live_group(cgroup))				\
		return -ENODEV;						\
									\
	blkcg = cgroup_to_blkio_cgroup(cgroup);				\
	rcu_read_lock();						\
	hlist_for_each_entry_rcu(blkg, n, &blkcg->blkg_list, blkcg_node) {\
		if (blkg->dev)						\
			seq_printf(m, "%u:%u %lu\n", MAJOR(blkg->dev),	\
				 MINOR(blkg->dev), blkg->__VAR);	\
	}								\
	rcu_read_unlock();						\
	cgroup_unlock();						\
	return 0;							\
}

SHOW_FUNCTION_PER_GROUP(time);
SHOW_FUNCTION_PER_GROUP(sectors);
#ifdef CONFIG_DEBUG_BLK_CGROUP
SHOW_FUNCTION_PER_GROUP(dequeue);
#endif
#undef SHOW_FUNCTION_PER_GROUP

#ifdef CONFIG_DEBUG_BLK_CGROUP
void blkiocg_update_blkio_group_dequeue_stats(struct blkio_group *blkg,
			unsigned long dequeue)
{
	blkg->dequeue += dequeue;
}
EXPORT_SYMBOL_GPL(blkiocg_update_blkio_group_dequeue_stats);
#endif

struct cftype blkio_files[] = {
	{
		.name = "weight",
		.read_u64 = blkiocg_weight_read,
		.write_u64 = blkiocg_weight_write,
	},
	{
		.name = "time",
		.read_seq_string = blkiocg_time_read,
	},
	{
		.name = "sectors",
		.read_seq_string = blkiocg_sectors_read,
	},
#ifdef CONFIG_DEBUG_BLK_CGROUP
       {
		.name = "dequeue",
		.read_seq_string = blkiocg_dequeue_read,
       },
#endif
};

static int blkiocg_populate(struct cgroup_subsys *subsys, struct cgroup *cgroup)
{
	return cgroup_add_files(cgroup, subsys, blkio_files,
				ARRAY_SIZE(blkio_files));
}

static void blkiocg_destroy(struct cgroup_subsys *subsys, struct cgroup *cgroup)
{
	struct blkio_cgroup *blkcg = cgroup_to_blkio_cgroup(cgroup);
	unsigned long flags;
	struct blkio_group *blkg;
	void *key;
	struct blkio_policy_type *blkiop;

	rcu_read_lock();
remove_entry:
	spin_lock_irqsave(&blkcg->lock, flags);

	if (hlist_empty(&blkcg->blkg_list)) {
		spin_unlock_irqrestore(&blkcg->lock, flags);
		goto done;
	}

	blkg = hlist_entry(blkcg->blkg_list.first, struct blkio_group,
				blkcg_node);
	key = rcu_dereference(blkg->key);
	__blkiocg_del_blkio_group(blkg);

	spin_unlock_irqrestore(&blkcg->lock, flags);

	/*
	 * This blkio_group is being unlinked as associated cgroup is going
	 * away. Let all the IO controlling policies know about this event.
	 *
	 * Currently this is static call to one io controlling policy. Once
	 * we have more policies in place, we need some dynamic registration
	 * of callback function.
	 */
	spin_lock(&blkio_list_lock);
	list_for_each_entry(blkiop, &blkio_list, list)
		blkiop->ops.blkio_unlink_group_fn(key, blkg);
	spin_unlock(&blkio_list_lock);
	goto remove_entry;
done:
	free_css_id(&blkio_subsys, &blkcg->css);
	rcu_read_unlock();
	if (blkcg != &blkio_root_cgroup)
		kfree(blkcg);
}

static struct cgroup_subsys_state *
blkiocg_create(struct cgroup_subsys *subsys, struct cgroup *cgroup)
{
	struct blkio_cgroup *blkcg, *parent_blkcg;

	if (!cgroup->parent) {
		blkcg = &blkio_root_cgroup;
		goto done;
	}

	/* Currently we do not support hierarchy deeper than two level (0,1) */
	parent_blkcg = cgroup_to_blkio_cgroup(cgroup->parent);
	if (css_depth(&parent_blkcg->css) > 0)
		return ERR_PTR(-EINVAL);

	blkcg = kzalloc(sizeof(*blkcg), GFP_KERNEL);
	if (!blkcg)
		return ERR_PTR(-ENOMEM);

	blkcg->weight = BLKIO_WEIGHT_DEFAULT;
done:
	spin_lock_init(&blkcg->lock);
	INIT_HLIST_HEAD(&blkcg->blkg_list);

	return &blkcg->css;
}

/*
 * We cannot support shared io contexts, as we have no mean to support
 * two tasks with the same ioc in two different groups without major rework
 * of the main cic data structures.  For now we allow a task to change
 * its cgroup only if it's the only owner of its ioc.
 */
static int blkiocg_can_attach(struct cgroup_subsys *subsys,
				struct cgroup *cgroup, struct task_struct *tsk,
				bool threadgroup)
{
	struct io_context *ioc;
	int ret = 0;

	/* task_lock() is needed to avoid races with exit_io_context() */
	task_lock(tsk);
	ioc = tsk->io_context;
	if (ioc && atomic_read(&ioc->nr_tasks) > 1)
		ret = -EINVAL;
	task_unlock(tsk);

	return ret;
}

static void blkiocg_attach(struct cgroup_subsys *subsys, struct cgroup *cgroup,
				struct cgroup *prev, struct task_struct *tsk,
				bool threadgroup)
{
	struct io_context *ioc;

	task_lock(tsk);
	ioc = tsk->io_context;
	if (ioc)
		ioc->cgroup_changed = 1;
	task_unlock(tsk);
}

void blkio_policy_register(struct blkio_policy_type *blkiop)
{
	spin_lock(&blkio_list_lock);
	list_add_tail(&blkiop->list, &blkio_list);
	spin_unlock(&blkio_list_lock);
}
EXPORT_SYMBOL_GPL(blkio_policy_register);

void blkio_policy_unregister(struct blkio_policy_type *blkiop)
{
	spin_lock(&blkio_list_lock);
	list_del_init(&blkiop->list);
	spin_unlock(&blkio_list_lock);
}
EXPORT_SYMBOL_GPL(blkio_policy_unregister);

static int __init init_cgroup_blkio(void)
{
	return cgroup_load_subsys(&blkio_subsys);
}

static void __exit exit_cgroup_blkio(void)
{
	cgroup_unload_subsys(&blkio_subsys);
}

module_init(init_cgroup_blkio);
module_exit(exit_cgroup_blkio);
MODULE_LICENSE("GPL");
Commit	Line	Data
31e4c28d VG	1	/*
	2	* Common Block IO controller cgroup interface
	3	*
	4	* Based on ideas and code from CFQ, CFS and BFQ:
	5	* Copyright (C) 2003 Jens Axboe <axboe@kernel.dk>
	6	*
	7	* Copyright (C) 2008 Fabio Checconi <fabio@gandalf.sssup.it>
	8	* Paolo Valente <paolo.valente@unimore.it>
	9	*
	10	* Copyright (C) 2009 Vivek Goyal <vgoyal@redhat.com>
	11	* Nauman Rafique <nauman@google.com>
	12	*/
	13	#include <linux/ioprio.h>
22084190 VG	14	#include <linux/seq_file.h>
22084190 VG	15	#include <linux/kdev_t.h>
9d6a986c	16	#include <linux/module.h>
accee785	17	#include <linux/err.h>
31e4c28d	18	#include "blk-cgroup.h"
3e252066 VG	19
	20	static DEFINE_SPINLOCK(blkio_list_lock);
	21	static LIST_HEAD(blkio_list);
b1c35769	22
31e4c28d	23	struct blkio_cgroup blkio_root_cgroup = { .weight = 2*BLKIO_WEIGHT_DEFAULT };
9d6a986c VG	24	EXPORT_SYMBOL_GPL(blkio_root_cgroup);
9d6a986c VG	25
67523c48 BB	26	static struct cgroup_subsys_state blkiocg_create(struct cgroup_subsys ,
	27	struct cgroup *);
	28	static int blkiocg_can_attach(struct cgroup_subsys , struct cgroup ,
	29	struct task_struct *, bool);
	30	static void blkiocg_attach(struct cgroup_subsys , struct cgroup ,
	31	struct cgroup , struct task_struct , bool);
	32	static void blkiocg_destroy(struct cgroup_subsys , struct cgroup );
	33	static int blkiocg_populate(struct cgroup_subsys , struct cgroup );
	34
	35	struct cgroup_subsys blkio_subsys = {
	36	.name = "blkio",
	37	.create = blkiocg_create,
	38	.can_attach = blkiocg_can_attach,
	39	.attach = blkiocg_attach,
	40	.destroy = blkiocg_destroy,
	41	.populate = blkiocg_populate,
	42	#ifdef CONFIG_BLK_CGROUP
	43	/* note: blkio_subsys_id is otherwise defined in blk-cgroup.h */
	44	.subsys_id = blkio_subsys_id,
	45	#endif
	46	.use_id = 1,
	47	.module = THIS_MODULE,
	48	};
	49	EXPORT_SYMBOL_GPL(blkio_subsys);
	50
31e4c28d VG	51	struct blkio_cgroup cgroup_to_blkio_cgroup(struct cgroup cgroup)
	52	{
	53	return container_of(cgroup_subsys_state(cgroup, blkio_subsys_id),
	54	struct blkio_cgroup, css);
	55	}
9d6a986c	56	EXPORT_SYMBOL_GPL(cgroup_to_blkio_cgroup);
31e4c28d	57
22084190 VG	58	void blkiocg_update_blkio_group_stats(struct blkio_group *blkg,
	59	unsigned long time, unsigned long sectors)
	60	{
	61	blkg->time += time;
	62	blkg->sectors += sectors;
	63	}
9d6a986c	64	EXPORT_SYMBOL_GPL(blkiocg_update_blkio_group_stats);
22084190	65
31e4c28d	66	void blkiocg_add_blkio_group(struct blkio_cgroup *blkcg,
22084190	67	struct blkio_group blkg, void key, dev_t dev)
31e4c28d VG	68	{
	69	unsigned long flags;
	70
	71	spin_lock_irqsave(&blkcg->lock, flags);
	72	rcu_assign_pointer(blkg->key, key);
b1c35769	73	blkg->blkcg_id = css_id(&blkcg->css);
31e4c28d VG	74	hlist_add_head_rcu(&blkg->blkcg_node, &blkcg->blkg_list);
31e4c28d VG	75	spin_unlock_irqrestore(&blkcg->lock, flags);
2868ef7b VG	76	#ifdef CONFIG_DEBUG_BLK_CGROUP
	77	/* Need to take css reference ? */
	78	cgroup_path(blkcg->css.cgroup, blkg->path, sizeof(blkg->path));
	79	#endif
22084190	80	blkg->dev = dev;
31e4c28d	81	}
9d6a986c	82	EXPORT_SYMBOL_GPL(blkiocg_add_blkio_group);
31e4c28d	83
b1c35769 VG	84	static void __blkiocg_del_blkio_group(struct blkio_group *blkg)
	85	{
	86	hlist_del_init_rcu(&blkg->blkcg_node);
	87	blkg->blkcg_id = 0;
	88	}
	89
	90	/*
	91	* returns 0 if blkio_group was still on cgroup list. Otherwise returns 1
	92	* indicating that blk_group was unhashed by the time we got to it.
	93	*/
31e4c28d VG	94	int blkiocg_del_blkio_group(struct blkio_group *blkg)
31e4c28d VG	95	{
b1c35769 VG	96	struct blkio_cgroup *blkcg;
	97	unsigned long flags;
	98	struct cgroup_subsys_state *css;
	99	int ret = 1;
	100
	101	rcu_read_lock();
	102	css = css_lookup(&blkio_subsys, blkg->blkcg_id);
	103	if (!css)
	104	goto out;
	105
	106	blkcg = container_of(css, struct blkio_cgroup, css);
	107	spin_lock_irqsave(&blkcg->lock, flags);
	108	if (!hlist_unhashed(&blkg->blkcg_node)) {
	109	__blkiocg_del_blkio_group(blkg);
	110	ret = 0;
	111	}
	112	spin_unlock_irqrestore(&blkcg->lock, flags);
	113	out:
	114	rcu_read_unlock();
	115	return ret;
31e4c28d	116	}
9d6a986c	117	EXPORT_SYMBOL_GPL(blkiocg_del_blkio_group);
31e4c28d VG	118
	119	/* called under rcu_read_lock(). */
	120	struct blkio_group blkiocg_lookup_group(struct blkio_cgroup blkcg, void *key)
	121	{
	122	struct blkio_group *blkg;
	123	struct hlist_node *n;
	124	void *__key;
	125
	126	hlist_for_each_entry_rcu(blkg, n, &blkcg->blkg_list, blkcg_node) {
	127	__key = blkg->key;
	128	if (__key == key)
	129	return blkg;
	130	}
	131
	132	return NULL;
	133	}
9d6a986c	134	EXPORT_SYMBOL_GPL(blkiocg_lookup_group);
31e4c28d VG	135
	136	#define SHOW_FUNCTION(__VAR) \
	137	static u64 blkiocg_##__VAR##_read(struct cgroup *cgroup, \
	138	struct cftype *cftype) \
	139	{ \
	140	struct blkio_cgroup *blkcg; \
	141	\
	142	blkcg = cgroup_to_blkio_cgroup(cgroup); \
	143	return (u64)blkcg->__VAR; \
	144	}
	145
	146	SHOW_FUNCTION(weight);
	147	#undef SHOW_FUNCTION
	148
	149	static int
	150	blkiocg_weight_write(struct cgroup cgroup, struct cftype cftype, u64 val)
	151	{
	152	struct blkio_cgroup *blkcg;
f8d461d6 VG	153	struct blkio_group *blkg;
f8d461d6 VG	154	struct hlist_node *n;
3e252066	155	struct blkio_policy_type *blkiop;
31e4c28d VG	156
	157	if (val < BLKIO_WEIGHT_MIN \|\| val > BLKIO_WEIGHT_MAX)
	158	return -EINVAL;
	159
	160	blkcg = cgroup_to_blkio_cgroup(cgroup);
bcf4dd43	161	spin_lock(&blkio_list_lock);
f8d461d6	162	spin_lock_irq(&blkcg->lock);
31e4c28d	163	blkcg->weight = (unsigned int)val;
3e252066	164	hlist_for_each_entry(blkg, n, &blkcg->blkg_list, blkcg_node) {
3e252066 VG	165	list_for_each_entry(blkiop, &blkio_list, list)
	166	blkiop->ops.blkio_update_group_weight_fn(blkg,
	167	blkcg->weight);
3e252066	168	}
f8d461d6	169	spin_unlock_irq(&blkcg->lock);
bcf4dd43	170	spin_unlock(&blkio_list_lock);
31e4c28d VG	171	return 0;
	172	}
	173
22084190 VG	174	#define SHOW_FUNCTION_PER_GROUP(__VAR) \
	175	static int blkiocg_##__VAR##_read(struct cgroup *cgroup, \
	176	struct cftype cftype, struct seq_file m) \
	177	{ \
	178	struct blkio_cgroup *blkcg; \
	179	struct blkio_group *blkg; \
	180	struct hlist_node *n; \
	181	\
	182	if (!cgroup_lock_live_group(cgroup)) \
	183	return -ENODEV; \
	184	\
	185	blkcg = cgroup_to_blkio_cgroup(cgroup); \
	186	rcu_read_lock(); \
	187	hlist_for_each_entry_rcu(blkg, n, &blkcg->blkg_list, blkcg_node) {\
	188	if (blkg->dev) \
	189	seq_printf(m, "%u:%u %lu\n", MAJOR(blkg->dev), \
	190	MINOR(blkg->dev), blkg->__VAR); \
	191	} \
	192	rcu_read_unlock(); \
	193	cgroup_unlock(); \
	194	return 0; \
	195	}
	196
	197	SHOW_FUNCTION_PER_GROUP(time);
	198	SHOW_FUNCTION_PER_GROUP(sectors);
	199	#ifdef CONFIG_DEBUG_BLK_CGROUP
	200	SHOW_FUNCTION_PER_GROUP(dequeue);
	201	#endif
	202	#undef SHOW_FUNCTION_PER_GROUP
	203
	204	#ifdef CONFIG_DEBUG_BLK_CGROUP
	205	void blkiocg_update_blkio_group_dequeue_stats(struct blkio_group *blkg,
	206	unsigned long dequeue)
	207	{
	208	blkg->dequeue += dequeue;
	209	}
9d6a986c	210	EXPORT_SYMBOL_GPL(blkiocg_update_blkio_group_dequeue_stats);
22084190 VG	211	#endif
22084190 VG	212
31e4c28d VG	213	struct cftype blkio_files[] = {
	214	{
	215	.name = "weight",
	216	.read_u64 = blkiocg_weight_read,
	217	.write_u64 = blkiocg_weight_write,
	218	},
22084190 VG	219	{
	220	.name = "time",
	221	.read_seq_string = blkiocg_time_read,
	222	},
	223	{
	224	.name = "sectors",
	225	.read_seq_string = blkiocg_sectors_read,
	226	},
	227	#ifdef CONFIG_DEBUG_BLK_CGROUP
	228	{
	229	.name = "dequeue",
	230	.read_seq_string = blkiocg_dequeue_read,
	231	},
	232	#endif
31e4c28d VG	233	};
	234
	235	static int blkiocg_populate(struct cgroup_subsys subsys, struct cgroup cgroup)
	236	{
	237	return cgroup_add_files(cgroup, subsys, blkio_files,
	238	ARRAY_SIZE(blkio_files));
	239	}
	240
	241	static void blkiocg_destroy(struct cgroup_subsys subsys, struct cgroup cgroup)
	242	{
	243	struct blkio_cgroup *blkcg = cgroup_to_blkio_cgroup(cgroup);
b1c35769 VG	244	unsigned long flags;
	245	struct blkio_group *blkg;
	246	void *key;
3e252066	247	struct blkio_policy_type *blkiop;
b1c35769 VG	248
	249	rcu_read_lock();
	250	remove_entry:
	251	spin_lock_irqsave(&blkcg->lock, flags);
	252
	253	if (hlist_empty(&blkcg->blkg_list)) {
	254	spin_unlock_irqrestore(&blkcg->lock, flags);
	255	goto done;
	256	}
	257
	258	blkg = hlist_entry(blkcg->blkg_list.first, struct blkio_group,
	259	blkcg_node);
	260	key = rcu_dereference(blkg->key);
	261	__blkiocg_del_blkio_group(blkg);
31e4c28d	262
b1c35769 VG	263	spin_unlock_irqrestore(&blkcg->lock, flags);
	264
	265	/*
	266	* This blkio_group is being unlinked as associated cgroup is going
	267	* away. Let all the IO controlling policies know about this event.
	268	*
	269	* Currently this is static call to one io controlling policy. Once
	270	* we have more policies in place, we need some dynamic registration
	271	* of callback function.
	272	*/
3e252066 VG	273	spin_lock(&blkio_list_lock);
	274	list_for_each_entry(blkiop, &blkio_list, list)
	275	blkiop->ops.blkio_unlink_group_fn(key, blkg);
	276	spin_unlock(&blkio_list_lock);
b1c35769 VG	277	goto remove_entry;
b1c35769 VG	278	done:
31e4c28d	279	free_css_id(&blkio_subsys, &blkcg->css);
b1c35769	280	rcu_read_unlock();
67523c48 BB	281	if (blkcg != &blkio_root_cgroup)
67523c48 BB	282	kfree(blkcg);
31e4c28d VG	283	}
	284
	285	static struct cgroup_subsys_state *
	286	blkiocg_create(struct cgroup_subsys subsys, struct cgroup cgroup)
	287	{
	288	struct blkio_cgroup blkcg, parent_blkcg;
	289
	290	if (!cgroup->parent) {
	291	blkcg = &blkio_root_cgroup;
	292	goto done;
	293	}
	294
	295	/* Currently we do not support hierarchy deeper than two level (0,1) */
	296	parent_blkcg = cgroup_to_blkio_cgroup(cgroup->parent);
	297	if (css_depth(&parent_blkcg->css) > 0)
	298	return ERR_PTR(-EINVAL);
	299
	300	blkcg = kzalloc(sizeof(*blkcg), GFP_KERNEL);
	301	if (!blkcg)
	302	return ERR_PTR(-ENOMEM);
	303
	304	blkcg->weight = BLKIO_WEIGHT_DEFAULT;
	305	done:
	306	spin_lock_init(&blkcg->lock);
	307	INIT_HLIST_HEAD(&blkcg->blkg_list);
	308
	309	return &blkcg->css;
	310	}
	311
	312	/*
	313	* We cannot support shared io contexts, as we have no mean to support
	314	* two tasks with the same ioc in two different groups without major rework
	315	* of the main cic data structures. For now we allow a task to change
	316	* its cgroup only if it's the only owner of its ioc.
	317	*/
	318	static int blkiocg_can_attach(struct cgroup_subsys *subsys,
	319	struct cgroup cgroup, struct task_struct tsk,
	320	bool threadgroup)
	321	{
	322	struct io_context *ioc;
	323	int ret = 0;
	324
	325	/* task_lock() is needed to avoid races with exit_io_context() */
	326	task_lock(tsk);
	327	ioc = tsk->io_context;
	328	if (ioc && atomic_read(&ioc->nr_tasks) > 1)
	329	ret = -EINVAL;
	330	task_unlock(tsk);
	331
	332	return ret;
	333	}
	334
	335	static void blkiocg_attach(struct cgroup_subsys subsys, struct cgroup cgroup,
	336	struct cgroup prev, struct task_struct tsk,
	337	bool threadgroup)
	338	{
	339	struct io_context *ioc;
	340
	341	task_lock(tsk);
	342	ioc = tsk->io_context;
	343	if (ioc)
	344	ioc->cgroup_changed = 1;
	345	task_unlock(tsk);
	346	}
347
3e252066 VG	348	void blkio_policy_register(struct blkio_policy_type *blkiop)
	349	{
	350	spin_lock(&blkio_list_lock);
	351	list_add_tail(&blkiop->list, &blkio_list);
	352	spin_unlock(&blkio_list_lock);
	353	}
	354	EXPORT_SYMBOL_GPL(blkio_policy_register);
	355
	356	void blkio_policy_unregister(struct blkio_policy_type *blkiop)
	357	{
	358	spin_lock(&blkio_list_lock);
	359	list_del_init(&blkiop->list);
	360	spin_unlock(&blkio_list_lock);
	361	}
	362	EXPORT_SYMBOL_GPL(blkio_policy_unregister);
67523c48 BB	363
	364	static int __init init_cgroup_blkio(void)
	365	{
	366	return cgroup_load_subsys(&blkio_subsys);
	367	}
	368
	369	static void __exit exit_cgroup_blkio(void)
	370	{
	371	cgroup_unload_subsys(&blkio_subsys);
	372	}
	373
	374	module_init(init_cgroup_blkio);
	375	module_exit(exit_cgroup_blkio);
	376	MODULE_LICENSE("GPL");