[mirror_ubuntu-bionic-kernel.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 "blk-cgroup.h"
#include "cfq-iosched.h"

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

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

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

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;
}

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;
}

/* 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;
}

#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;

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

	blkcg = cgroup_to_blkio_cgroup(cgroup);
	spin_lock_irq(&blkcg->lock);
	blkcg->weight = (unsigned int)val;
	hlist_for_each_entry(blkg, n, &blkcg->blkg_list, blkcg_node)
		cfq_update_blkio_group_weight(blkg, blkcg->weight);
	spin_unlock_irq(&blkcg->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;
}
#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;

	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.
	 */
	cfq_unlink_blkio_group(key, blkg);
	goto remove_entry;
done:
	free_css_id(&blkio_subsys, &blkcg->css);
	rcu_read_unlock();
	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);
}

struct cgroup_subsys blkio_subsys = {
	.name = "blkio",
	.create = blkiocg_create,
	.can_attach = blkiocg_can_attach,
	.attach = blkiocg_attach,
	.destroy = blkiocg_destroy,
	.populate = blkiocg_populate,
	.subsys_id = blkio_subsys_id,
	.use_id = 1,
};
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>
31e4c28d	16	#include "blk-cgroup.h"
f2eecb91	17	#include "cfq-iosched.h"
b1c35769	18
31e4c28d VG	19	struct blkio_cgroup blkio_root_cgroup = { .weight = 2*BLKIO_WEIGHT_DEFAULT };
	20
	21	struct blkio_cgroup cgroup_to_blkio_cgroup(struct cgroup cgroup)
	22	{
	23	return container_of(cgroup_subsys_state(cgroup, blkio_subsys_id),
	24	struct blkio_cgroup, css);
	25	}
	26
22084190 VG	27	void blkiocg_update_blkio_group_stats(struct blkio_group *blkg,
	28	unsigned long time, unsigned long sectors)
	29	{
	30	blkg->time += time;
	31	blkg->sectors += sectors;
	32	}
	33
31e4c28d	34	void blkiocg_add_blkio_group(struct blkio_cgroup *blkcg,
22084190	35	struct blkio_group blkg, void key, dev_t dev)
31e4c28d VG	36	{
	37	unsigned long flags;
	38
	39	spin_lock_irqsave(&blkcg->lock, flags);
	40	rcu_assign_pointer(blkg->key, key);
b1c35769	41	blkg->blkcg_id = css_id(&blkcg->css);
31e4c28d VG	42	hlist_add_head_rcu(&blkg->blkcg_node, &blkcg->blkg_list);
31e4c28d VG	43	spin_unlock_irqrestore(&blkcg->lock, flags);
2868ef7b VG	44	#ifdef CONFIG_DEBUG_BLK_CGROUP
	45	/* Need to take css reference ? */
	46	cgroup_path(blkcg->css.cgroup, blkg->path, sizeof(blkg->path));
	47	#endif
22084190	48	blkg->dev = dev;
31e4c28d VG	49	}
31e4c28d VG	50
b1c35769 VG	51	static void __blkiocg_del_blkio_group(struct blkio_group *blkg)
	52	{
	53	hlist_del_init_rcu(&blkg->blkcg_node);
	54	blkg->blkcg_id = 0;
	55	}
	56
	57	/*
	58	* returns 0 if blkio_group was still on cgroup list. Otherwise returns 1
	59	* indicating that blk_group was unhashed by the time we got to it.
	60	*/
31e4c28d VG	61	int blkiocg_del_blkio_group(struct blkio_group *blkg)
31e4c28d VG	62	{
b1c35769 VG	63	struct blkio_cgroup *blkcg;
	64	unsigned long flags;
	65	struct cgroup_subsys_state *css;
	66	int ret = 1;
	67
	68	rcu_read_lock();
	69	css = css_lookup(&blkio_subsys, blkg->blkcg_id);
	70	if (!css)
	71	goto out;
	72
	73	blkcg = container_of(css, struct blkio_cgroup, css);
	74	spin_lock_irqsave(&blkcg->lock, flags);
	75	if (!hlist_unhashed(&blkg->blkcg_node)) {
	76	__blkiocg_del_blkio_group(blkg);
	77	ret = 0;
	78	}
	79	spin_unlock_irqrestore(&blkcg->lock, flags);
	80	out:
	81	rcu_read_unlock();
	82	return ret;
31e4c28d VG	83	}
	84
	85	/* called under rcu_read_lock(). */
	86	struct blkio_group blkiocg_lookup_group(struct blkio_cgroup blkcg, void *key)
	87	{
	88	struct blkio_group *blkg;
	89	struct hlist_node *n;
	90	void *__key;
	91
	92	hlist_for_each_entry_rcu(blkg, n, &blkcg->blkg_list, blkcg_node) {
	93	__key = blkg->key;
	94	if (__key == key)
	95	return blkg;
	96	}
	97
	98	return NULL;
	99	}
	100
	101	#define SHOW_FUNCTION(__VAR) \
	102	static u64 blkiocg_##__VAR##_read(struct cgroup *cgroup, \
	103	struct cftype *cftype) \
	104	{ \
	105	struct blkio_cgroup *blkcg; \
	106	\
	107	blkcg = cgroup_to_blkio_cgroup(cgroup); \
	108	return (u64)blkcg->__VAR; \
	109	}
	110
	111	SHOW_FUNCTION(weight);
	112	#undef SHOW_FUNCTION
	113
	114	static int
	115	blkiocg_weight_write(struct cgroup cgroup, struct cftype cftype, u64 val)
	116	{
	117	struct blkio_cgroup *blkcg;
f8d461d6 VG	118	struct blkio_group *blkg;
f8d461d6 VG	119	struct hlist_node *n;
31e4c28d VG	120
	121	if (val < BLKIO_WEIGHT_MIN \|\| val > BLKIO_WEIGHT_MAX)
	122	return -EINVAL;
	123
	124	blkcg = cgroup_to_blkio_cgroup(cgroup);
f8d461d6	125	spin_lock_irq(&blkcg->lock);
31e4c28d	126	blkcg->weight = (unsigned int)val;
f8d461d6 VG	127	hlist_for_each_entry(blkg, n, &blkcg->blkg_list, blkcg_node)
	128	cfq_update_blkio_group_weight(blkg, blkcg->weight);
	129	spin_unlock_irq(&blkcg->lock);
31e4c28d VG	130	return 0;
	131	}
	132
22084190 VG	133	#define SHOW_FUNCTION_PER_GROUP(__VAR) \
	134	static int blkiocg_##__VAR##_read(struct cgroup *cgroup, \
	135	struct cftype cftype, struct seq_file m) \
	136	{ \
	137	struct blkio_cgroup *blkcg; \
	138	struct blkio_group *blkg; \
	139	struct hlist_node *n; \
	140	\
	141	if (!cgroup_lock_live_group(cgroup)) \
	142	return -ENODEV; \
	143	\
	144	blkcg = cgroup_to_blkio_cgroup(cgroup); \
	145	rcu_read_lock(); \
	146	hlist_for_each_entry_rcu(blkg, n, &blkcg->blkg_list, blkcg_node) {\
	147	if (blkg->dev) \
	148	seq_printf(m, "%u:%u %lu\n", MAJOR(blkg->dev), \
	149	MINOR(blkg->dev), blkg->__VAR); \
	150	} \
	151	rcu_read_unlock(); \
	152	cgroup_unlock(); \
	153	return 0; \
	154	}
	155
	156	SHOW_FUNCTION_PER_GROUP(time);
	157	SHOW_FUNCTION_PER_GROUP(sectors);
	158	#ifdef CONFIG_DEBUG_BLK_CGROUP
	159	SHOW_FUNCTION_PER_GROUP(dequeue);
	160	#endif
	161	#undef SHOW_FUNCTION_PER_GROUP
	162
	163	#ifdef CONFIG_DEBUG_BLK_CGROUP
	164	void blkiocg_update_blkio_group_dequeue_stats(struct blkio_group *blkg,
	165	unsigned long dequeue)
	166	{
	167	blkg->dequeue += dequeue;
	168	}
	169	#endif
	170
31e4c28d VG	171	struct cftype blkio_files[] = {
	172	{
	173	.name = "weight",
	174	.read_u64 = blkiocg_weight_read,
	175	.write_u64 = blkiocg_weight_write,
	176	},
22084190 VG	177	{
	178	.name = "time",
	179	.read_seq_string = blkiocg_time_read,
	180	},
	181	{
	182	.name = "sectors",
	183	.read_seq_string = blkiocg_sectors_read,
	184	},
	185	#ifdef CONFIG_DEBUG_BLK_CGROUP
	186	{
	187	.name = "dequeue",
	188	.read_seq_string = blkiocg_dequeue_read,
	189	},
	190	#endif
31e4c28d VG	191	};
	192
	193	static int blkiocg_populate(struct cgroup_subsys subsys, struct cgroup cgroup)
	194	{
	195	return cgroup_add_files(cgroup, subsys, blkio_files,
	196	ARRAY_SIZE(blkio_files));
	197	}
	198
	199	static void blkiocg_destroy(struct cgroup_subsys subsys, struct cgroup cgroup)
	200	{
	201	struct blkio_cgroup *blkcg = cgroup_to_blkio_cgroup(cgroup);
b1c35769 VG	202	unsigned long flags;
	203	struct blkio_group *blkg;
	204	void *key;
	205
	206	rcu_read_lock();
	207	remove_entry:
	208	spin_lock_irqsave(&blkcg->lock, flags);
	209
	210	if (hlist_empty(&blkcg->blkg_list)) {
	211	spin_unlock_irqrestore(&blkcg->lock, flags);
	212	goto done;
	213	}
	214
	215	blkg = hlist_entry(blkcg->blkg_list.first, struct blkio_group,
	216	blkcg_node);
	217	key = rcu_dereference(blkg->key);
	218	__blkiocg_del_blkio_group(blkg);
31e4c28d	219
b1c35769 VG	220	spin_unlock_irqrestore(&blkcg->lock, flags);
	221
	222	/*
	223	* This blkio_group is being unlinked as associated cgroup is going
	224	* away. Let all the IO controlling policies know about this event.
	225	*
	226	* Currently this is static call to one io controlling policy. Once
	227	* we have more policies in place, we need some dynamic registration
	228	* of callback function.
	229	*/
	230	cfq_unlink_blkio_group(key, blkg);
	231	goto remove_entry;
	232	done:
31e4c28d	233	free_css_id(&blkio_subsys, &blkcg->css);
b1c35769	234	rcu_read_unlock();
31e4c28d VG	235	kfree(blkcg);
	236	}
	237
	238	static struct cgroup_subsys_state *
	239	blkiocg_create(struct cgroup_subsys subsys, struct cgroup cgroup)
	240	{
	241	struct blkio_cgroup blkcg, parent_blkcg;
	242
	243	if (!cgroup->parent) {
	244	blkcg = &blkio_root_cgroup;
	245	goto done;
	246	}
	247
	248	/* Currently we do not support hierarchy deeper than two level (0,1) */
	249	parent_blkcg = cgroup_to_blkio_cgroup(cgroup->parent);
	250	if (css_depth(&parent_blkcg->css) > 0)
	251	return ERR_PTR(-EINVAL);
	252
	253	blkcg = kzalloc(sizeof(*blkcg), GFP_KERNEL);
	254	if (!blkcg)
	255	return ERR_PTR(-ENOMEM);
	256
	257	blkcg->weight = BLKIO_WEIGHT_DEFAULT;
	258	done:
	259	spin_lock_init(&blkcg->lock);
	260	INIT_HLIST_HEAD(&blkcg->blkg_list);
	261
	262	return &blkcg->css;
	263	}
	264
	265	/*
	266	* We cannot support shared io contexts, as we have no mean to support
	267	* two tasks with the same ioc in two different groups without major rework
	268	* of the main cic data structures. For now we allow a task to change
	269	* its cgroup only if it's the only owner of its ioc.
	270	*/
	271	static int blkiocg_can_attach(struct cgroup_subsys *subsys,
	272	struct cgroup cgroup, struct task_struct tsk,
	273	bool threadgroup)
	274	{
	275	struct io_context *ioc;
	276	int ret = 0;
	277
	278	/* task_lock() is needed to avoid races with exit_io_context() */
	279	task_lock(tsk);
	280	ioc = tsk->io_context;
	281	if (ioc && atomic_read(&ioc->nr_tasks) > 1)
	282	ret = -EINVAL;
	283	task_unlock(tsk);
	284
	285	return ret;
	286	}
	287
	288	static void blkiocg_attach(struct cgroup_subsys subsys, struct cgroup cgroup,
	289	struct cgroup prev, struct task_struct tsk,
	290	bool threadgroup)
	291	{
	292	struct io_context *ioc;
	293
	294	task_lock(tsk);
	295	ioc = tsk->io_context;
	296	if (ioc)
	297	ioc->cgroup_changed = 1;
	298	task_unlock(tsk);
299	}
300
301	struct cgroup_subsys blkio_subsys = {
302	.name = "blkio",
303	.create = blkiocg_create,
304	.can_attach = blkiocg_can_attach,
305	.attach = blkiocg_attach,
306	.destroy = blkiocg_destroy,
307	.populate = blkiocg_populate,
308	.subsys_id = blkio_subsys_id,
309	.use_id = 1,
310	};