[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 <linux/slab.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>
5a0e3ad6	18	#include <linux/slab.h>
31e4c28d	19	#include "blk-cgroup.h"
3e252066 VG	20
	21	static DEFINE_SPINLOCK(blkio_list_lock);
	22	static LIST_HEAD(blkio_list);
b1c35769	23
31e4c28d	24	struct blkio_cgroup blkio_root_cgroup = { .weight = 2*BLKIO_WEIGHT_DEFAULT };
9d6a986c VG	25	EXPORT_SYMBOL_GPL(blkio_root_cgroup);
9d6a986c VG	26
67523c48 BB	27	static struct cgroup_subsys_state blkiocg_create(struct cgroup_subsys ,
	28	struct cgroup *);
	29	static int blkiocg_can_attach(struct cgroup_subsys , struct cgroup ,
	30	struct task_struct *, bool);
	31	static void blkiocg_attach(struct cgroup_subsys , struct cgroup ,
	32	struct cgroup , struct task_struct , bool);
	33	static void blkiocg_destroy(struct cgroup_subsys , struct cgroup );
	34	static int blkiocg_populate(struct cgroup_subsys , struct cgroup );
	35
	36	struct cgroup_subsys blkio_subsys = {
	37	.name = "blkio",
	38	.create = blkiocg_create,
	39	.can_attach = blkiocg_can_attach,
	40	.attach = blkiocg_attach,
	41	.destroy = blkiocg_destroy,
	42	.populate = blkiocg_populate,
	43	#ifdef CONFIG_BLK_CGROUP
	44	/* note: blkio_subsys_id is otherwise defined in blk-cgroup.h */
	45	.subsys_id = blkio_subsys_id,
	46	#endif
	47	.use_id = 1,
	48	.module = THIS_MODULE,
	49	};
	50	EXPORT_SYMBOL_GPL(blkio_subsys);
	51
31e4c28d VG	52	struct blkio_cgroup cgroup_to_blkio_cgroup(struct cgroup cgroup)
	53	{
	54	return container_of(cgroup_subsys_state(cgroup, blkio_subsys_id),
	55	struct blkio_cgroup, css);
	56	}
9d6a986c	57	EXPORT_SYMBOL_GPL(cgroup_to_blkio_cgroup);
31e4c28d	58
22084190 VG	59	void blkiocg_update_blkio_group_stats(struct blkio_group *blkg,
	60	unsigned long time, unsigned long sectors)
	61	{
	62	blkg->time += time;
	63	blkg->sectors += sectors;
	64	}
9d6a986c	65	EXPORT_SYMBOL_GPL(blkiocg_update_blkio_group_stats);
22084190	66
31e4c28d	67	void blkiocg_add_blkio_group(struct blkio_cgroup *blkcg,
22084190	68	struct blkio_group blkg, void key, dev_t dev)
31e4c28d VG	69	{
	70	unsigned long flags;
	71
	72	spin_lock_irqsave(&blkcg->lock, flags);
	73	rcu_assign_pointer(blkg->key, key);
b1c35769	74	blkg->blkcg_id = css_id(&blkcg->css);
31e4c28d VG	75	hlist_add_head_rcu(&blkg->blkcg_node, &blkcg->blkg_list);
31e4c28d VG	76	spin_unlock_irqrestore(&blkcg->lock, flags);
2868ef7b VG	77	#ifdef CONFIG_DEBUG_BLK_CGROUP
	78	/* Need to take css reference ? */
	79	cgroup_path(blkcg->css.cgroup, blkg->path, sizeof(blkg->path));
	80	#endif
22084190	81	blkg->dev = dev;
31e4c28d	82	}
9d6a986c	83	EXPORT_SYMBOL_GPL(blkiocg_add_blkio_group);
31e4c28d	84
b1c35769 VG	85	static void __blkiocg_del_blkio_group(struct blkio_group *blkg)
	86	{
	87	hlist_del_init_rcu(&blkg->blkcg_node);
	88	blkg->blkcg_id = 0;
	89	}
	90
	91	/*
	92	* returns 0 if blkio_group was still on cgroup list. Otherwise returns 1
	93	* indicating that blk_group was unhashed by the time we got to it.
	94	*/
31e4c28d VG	95	int blkiocg_del_blkio_group(struct blkio_group *blkg)
31e4c28d VG	96	{
b1c35769 VG	97	struct blkio_cgroup *blkcg;
	98	unsigned long flags;
	99	struct cgroup_subsys_state *css;
	100	int ret = 1;
	101
	102	rcu_read_lock();
	103	css = css_lookup(&blkio_subsys, blkg->blkcg_id);
	104	if (!css)
	105	goto out;
	106
	107	blkcg = container_of(css, struct blkio_cgroup, css);
	108	spin_lock_irqsave(&blkcg->lock, flags);
	109	if (!hlist_unhashed(&blkg->blkcg_node)) {
	110	__blkiocg_del_blkio_group(blkg);
	111	ret = 0;
	112	}
	113	spin_unlock_irqrestore(&blkcg->lock, flags);
	114	out:
	115	rcu_read_unlock();
	116	return ret;
31e4c28d	117	}
9d6a986c	118	EXPORT_SYMBOL_GPL(blkiocg_del_blkio_group);
31e4c28d VG	119
	120	/* called under rcu_read_lock(). */
	121	struct blkio_group blkiocg_lookup_group(struct blkio_cgroup blkcg, void *key)
	122	{
	123	struct blkio_group *blkg;
	124	struct hlist_node *n;
	125	void *__key;
	126
	127	hlist_for_each_entry_rcu(blkg, n, &blkcg->blkg_list, blkcg_node) {
	128	__key = blkg->key;
	129	if (__key == key)
	130	return blkg;
	131	}
	132
	133	return NULL;
	134	}
9d6a986c	135	EXPORT_SYMBOL_GPL(blkiocg_lookup_group);
31e4c28d VG	136
	137	#define SHOW_FUNCTION(__VAR) \
	138	static u64 blkiocg_##__VAR##_read(struct cgroup *cgroup, \
	139	struct cftype *cftype) \
	140	{ \
	141	struct blkio_cgroup *blkcg; \
	142	\
	143	blkcg = cgroup_to_blkio_cgroup(cgroup); \
	144	return (u64)blkcg->__VAR; \
	145	}
	146
	147	SHOW_FUNCTION(weight);
	148	#undef SHOW_FUNCTION
	149
	150	static int
	151	blkiocg_weight_write(struct cgroup cgroup, struct cftype cftype, u64 val)
	152	{
	153	struct blkio_cgroup *blkcg;
f8d461d6 VG	154	struct blkio_group *blkg;
f8d461d6 VG	155	struct hlist_node *n;
3e252066	156	struct blkio_policy_type *blkiop;
31e4c28d VG	157
	158	if (val < BLKIO_WEIGHT_MIN \|\| val > BLKIO_WEIGHT_MAX)
	159	return -EINVAL;
	160
	161	blkcg = cgroup_to_blkio_cgroup(cgroup);
bcf4dd43	162	spin_lock(&blkio_list_lock);
f8d461d6	163	spin_lock_irq(&blkcg->lock);
31e4c28d	164	blkcg->weight = (unsigned int)val;
3e252066	165	hlist_for_each_entry(blkg, n, &blkcg->blkg_list, blkcg_node) {
3e252066 VG	166	list_for_each_entry(blkiop, &blkio_list, list)
	167	blkiop->ops.blkio_update_group_weight_fn(blkg,
	168	blkcg->weight);
3e252066	169	}
f8d461d6	170	spin_unlock_irq(&blkcg->lock);
bcf4dd43	171	spin_unlock(&blkio_list_lock);
31e4c28d VG	172	return 0;
	173	}
	174
22084190 VG	175	#define SHOW_FUNCTION_PER_GROUP(__VAR) \
	176	static int blkiocg_##__VAR##_read(struct cgroup *cgroup, \
	177	struct cftype cftype, struct seq_file m) \
	178	{ \
	179	struct blkio_cgroup *blkcg; \
	180	struct blkio_group *blkg; \
	181	struct hlist_node *n; \
	182	\
	183	if (!cgroup_lock_live_group(cgroup)) \
	184	return -ENODEV; \
	185	\
	186	blkcg = cgroup_to_blkio_cgroup(cgroup); \
	187	rcu_read_lock(); \
	188	hlist_for_each_entry_rcu(blkg, n, &blkcg->blkg_list, blkcg_node) {\
	189	if (blkg->dev) \
	190	seq_printf(m, "%u:%u %lu\n", MAJOR(blkg->dev), \
	191	MINOR(blkg->dev), blkg->__VAR); \
	192	} \
	193	rcu_read_unlock(); \
	194	cgroup_unlock(); \
	195	return 0; \
	196	}
	197
	198	SHOW_FUNCTION_PER_GROUP(time);
	199	SHOW_FUNCTION_PER_GROUP(sectors);
	200	#ifdef CONFIG_DEBUG_BLK_CGROUP
	201	SHOW_FUNCTION_PER_GROUP(dequeue);
	202	#endif
	203	#undef SHOW_FUNCTION_PER_GROUP
	204
	205	#ifdef CONFIG_DEBUG_BLK_CGROUP
	206	void blkiocg_update_blkio_group_dequeue_stats(struct blkio_group *blkg,
	207	unsigned long dequeue)
	208	{
	209	blkg->dequeue += dequeue;
	210	}
9d6a986c	211	EXPORT_SYMBOL_GPL(blkiocg_update_blkio_group_dequeue_stats);
22084190 VG	212	#endif
22084190 VG	213
31e4c28d VG	214	struct cftype blkio_files[] = {
	215	{
	216	.name = "weight",
	217	.read_u64 = blkiocg_weight_read,
	218	.write_u64 = blkiocg_weight_write,
	219	},
22084190 VG	220	{
	221	.name = "time",
	222	.read_seq_string = blkiocg_time_read,
	223	},
	224	{
	225	.name = "sectors",
	226	.read_seq_string = blkiocg_sectors_read,
	227	},
	228	#ifdef CONFIG_DEBUG_BLK_CGROUP
	229	{
	230	.name = "dequeue",
	231	.read_seq_string = blkiocg_dequeue_read,
	232	},
	233	#endif
31e4c28d VG	234	};
	235
	236	static int blkiocg_populate(struct cgroup_subsys subsys, struct cgroup cgroup)
	237	{
	238	return cgroup_add_files(cgroup, subsys, blkio_files,
	239	ARRAY_SIZE(blkio_files));
	240	}
	241
	242	static void blkiocg_destroy(struct cgroup_subsys subsys, struct cgroup cgroup)
	243	{
	244	struct blkio_cgroup *blkcg = cgroup_to_blkio_cgroup(cgroup);
b1c35769 VG	245	unsigned long flags;
	246	struct blkio_group *blkg;
	247	void *key;
3e252066	248	struct blkio_policy_type *blkiop;
b1c35769 VG	249
	250	rcu_read_lock();
	251	remove_entry:
	252	spin_lock_irqsave(&blkcg->lock, flags);
	253
	254	if (hlist_empty(&blkcg->blkg_list)) {
	255	spin_unlock_irqrestore(&blkcg->lock, flags);
	256	goto done;
	257	}
	258
	259	blkg = hlist_entry(blkcg->blkg_list.first, struct blkio_group,
	260	blkcg_node);
	261	key = rcu_dereference(blkg->key);
	262	__blkiocg_del_blkio_group(blkg);
31e4c28d	263
b1c35769 VG	264	spin_unlock_irqrestore(&blkcg->lock, flags);
	265
	266	/*
	267	* This blkio_group is being unlinked as associated cgroup is going
	268	* away. Let all the IO controlling policies know about this event.
	269	*
	270	* Currently this is static call to one io controlling policy. Once
	271	* we have more policies in place, we need some dynamic registration
	272	* of callback function.
	273	*/
3e252066 VG	274	spin_lock(&blkio_list_lock);
	275	list_for_each_entry(blkiop, &blkio_list, list)
	276	blkiop->ops.blkio_unlink_group_fn(key, blkg);
	277	spin_unlock(&blkio_list_lock);
b1c35769 VG	278	goto remove_entry;
b1c35769 VG	279	done:
31e4c28d	280	free_css_id(&blkio_subsys, &blkcg->css);
b1c35769	281	rcu_read_unlock();
67523c48 BB	282	if (blkcg != &blkio_root_cgroup)
67523c48 BB	283	kfree(blkcg);
31e4c28d VG	284	}
	285
	286	static struct cgroup_subsys_state *
	287	blkiocg_create(struct cgroup_subsys subsys, struct cgroup cgroup)
	288	{
	289	struct blkio_cgroup blkcg, parent_blkcg;
	290
	291	if (!cgroup->parent) {
	292	blkcg = &blkio_root_cgroup;
	293	goto done;
	294	}
	295
	296	/* Currently we do not support hierarchy deeper than two level (0,1) */
	297	parent_blkcg = cgroup_to_blkio_cgroup(cgroup->parent);
	298	if (css_depth(&parent_blkcg->css) > 0)
	299	return ERR_PTR(-EINVAL);
	300
	301	blkcg = kzalloc(sizeof(*blkcg), GFP_KERNEL);
	302	if (!blkcg)
	303	return ERR_PTR(-ENOMEM);
	304
	305	blkcg->weight = BLKIO_WEIGHT_DEFAULT;
	306	done:
	307	spin_lock_init(&blkcg->lock);
	308	INIT_HLIST_HEAD(&blkcg->blkg_list);
	309
	310	return &blkcg->css;
	311	}
	312
	313	/*
	314	* We cannot support shared io contexts, as we have no mean to support
	315	* two tasks with the same ioc in two different groups without major rework
	316	* of the main cic data structures. For now we allow a task to change
	317	* its cgroup only if it's the only owner of its ioc.
	318	*/
	319	static int blkiocg_can_attach(struct cgroup_subsys *subsys,
	320	struct cgroup cgroup, struct task_struct tsk,
	321	bool threadgroup)
	322	{
	323	struct io_context *ioc;
	324	int ret = 0;
	325
	326	/* task_lock() is needed to avoid races with exit_io_context() */
	327	task_lock(tsk);
	328	ioc = tsk->io_context;
	329	if (ioc && atomic_read(&ioc->nr_tasks) > 1)
	330	ret = -EINVAL;
	331	task_unlock(tsk);
	332
	333	return ret;
	334	}
	335
	336	static void blkiocg_attach(struct cgroup_subsys subsys, struct cgroup cgroup,
	337	struct cgroup prev, struct task_struct tsk,
	338	bool threadgroup)
	339	{
	340	struct io_context *ioc;
	341
	342	task_lock(tsk);
	343	ioc = tsk->io_context;
	344	if (ioc)
	345	ioc->cgroup_changed = 1;
	346	task_unlock(tsk);
	347	}
348
3e252066 VG	349	void blkio_policy_register(struct blkio_policy_type *blkiop)
	350	{
	351	spin_lock(&blkio_list_lock);
	352	list_add_tail(&blkiop->list, &blkio_list);
	353	spin_unlock(&blkio_list_lock);
	354	}
	355	EXPORT_SYMBOL_GPL(blkio_policy_register);
	356
	357	void blkio_policy_unregister(struct blkio_policy_type *blkiop)
	358	{
	359	spin_lock(&blkio_list_lock);
	360	list_del_init(&blkiop->list);
	361	spin_unlock(&blkio_list_lock);
	362	}
	363	EXPORT_SYMBOL_GPL(blkio_policy_unregister);
67523c48 BB	364
	365	static int __init init_cgroup_blkio(void)
	366	{
	367	return cgroup_load_subsys(&blkio_subsys);
	368	}
	369
	370	static void __exit exit_cgroup_blkio(void)
	371	{
	372	cgroup_unload_subsys(&blkio_subsys);
	373	}
	374
	375	module_init(init_cgroup_blkio);
	376	module_exit(exit_cgroup_blkio);
	377	MODULE_LICENSE("GPL");