[linux.git] / include / linux / blk-mq.h

#ifndef BLK_MQ_H
#define BLK_MQ_H

#include <linux/blkdev.h>

struct blk_mq_tags;
struct blk_flush_queue;

struct blk_mq_cpu_notifier {
	struct list_head list;
	void *data;
	int (*notify)(void *data, unsigned long action, unsigned int cpu);
};

struct blk_mq_ctxmap {
	unsigned int size;
	unsigned int bits_per_word;
	struct blk_align_bitmap *map;
};

struct blk_mq_hw_ctx {
	struct {
		spinlock_t		lock;
		struct list_head	dispatch;
	} ____cacheline_aligned_in_smp;

	unsigned long		state;		/* BLK_MQ_S_* flags */
	struct delayed_work	run_work;
	struct delayed_work	delay_work;
	cpumask_var_t		cpumask;
	int			next_cpu;
	int			next_cpu_batch;

	unsigned long		flags;		/* BLK_MQ_F_* flags */

	struct request_queue	*queue;
	struct blk_flush_queue	*fq;

	void			*driver_data;

	struct blk_mq_ctxmap	ctx_map;

	unsigned int		nr_ctx;
	struct blk_mq_ctx	**ctxs;

	atomic_t		wait_index;

	struct blk_mq_tags	*tags;

	unsigned long		queued;
	unsigned long		run;
#define BLK_MQ_MAX_DISPATCH_ORDER	10
	unsigned long		dispatched[BLK_MQ_MAX_DISPATCH_ORDER];

	unsigned int		numa_node;
	unsigned int		queue_num;

	atomic_t		nr_active;

	struct blk_mq_cpu_notifier	cpu_notifier;
	struct kobject		kobj;

	unsigned long		poll_invoked;
	unsigned long		poll_success;
};

struct blk_mq_tag_set {
	struct blk_mq_ops	*ops;
	unsigned int		nr_hw_queues;
	unsigned int		queue_depth;	/* max hw supported */
	unsigned int		reserved_tags;
	unsigned int		cmd_size;	/* per-request extra data */
	int			numa_node;
	unsigned int		timeout;
	unsigned int		flags;		/* BLK_MQ_F_* */
	void			*driver_data;

	struct blk_mq_tags	**tags;

	struct mutex		tag_list_lock;
	struct list_head	tag_list;
};

struct blk_mq_queue_data {
	struct request *rq;
	struct list_head *list;
	bool last;
};

typedef int (queue_rq_fn)(struct blk_mq_hw_ctx *, const struct blk_mq_queue_data *);
typedef struct blk_mq_hw_ctx *(map_queue_fn)(struct request_queue *, const int);
typedef enum blk_eh_timer_return (timeout_fn)(struct request *, bool);
typedef int (init_hctx_fn)(struct blk_mq_hw_ctx *, void *, unsigned int);
typedef void (exit_hctx_fn)(struct blk_mq_hw_ctx *, unsigned int);
typedef int (init_request_fn)(void *, struct request *, unsigned int,
		unsigned int, unsigned int);
typedef void (exit_request_fn)(void *, struct request *, unsigned int,
		unsigned int);

typedef void (busy_iter_fn)(struct blk_mq_hw_ctx *, struct request *, void *,
		bool);
typedef void (busy_tag_iter_fn)(struct request *, void *, bool);
typedef int (poll_fn)(struct blk_mq_hw_ctx *, unsigned int);


struct blk_mq_ops {
	/*
	 * Queue request
	 */
	queue_rq_fn		*queue_rq;

	/*
	 * Map to specific hardware queue
	 */
	map_queue_fn		*map_queue;

	/*
	 * Called on request timeout
	 */
	timeout_fn		*timeout;

	/*
	 * Called to poll for completion of a specific tag.
	 */
	poll_fn			*poll;

	softirq_done_fn		*complete;

	/*
	 * Called when the block layer side of a hardware queue has been
	 * set up, allowing the driver to allocate/init matching structures.
	 * Ditto for exit/teardown.
	 */
	init_hctx_fn		*init_hctx;
	exit_hctx_fn		*exit_hctx;

	/*
	 * Called for every command allocated by the block layer to allow
	 * the driver to set up driver specific data.
	 *
	 * Tag greater than or equal to queue_depth is for setting up
	 * flush request.
	 *
	 * Ditto for exit/teardown.
	 */
	init_request_fn		*init_request;
	exit_request_fn		*exit_request;
};

enum {
	BLK_MQ_RQ_QUEUE_OK	= 0,	/* queued fine */
	BLK_MQ_RQ_QUEUE_BUSY	= 1,	/* requeue IO for later */
	BLK_MQ_RQ_QUEUE_ERROR	= 2,	/* end IO with error */

	BLK_MQ_F_SHOULD_MERGE	= 1 << 0,
	BLK_MQ_F_TAG_SHARED	= 1 << 1,
	BLK_MQ_F_SG_MERGE	= 1 << 2,
	BLK_MQ_F_DEFER_ISSUE	= 1 << 4,
	BLK_MQ_F_ALLOC_POLICY_START_BIT = 8,
	BLK_MQ_F_ALLOC_POLICY_BITS = 1,

	BLK_MQ_S_STOPPED	= 0,
	BLK_MQ_S_TAG_ACTIVE	= 1,

	BLK_MQ_MAX_DEPTH	= 10240,

	BLK_MQ_CPU_WORK_BATCH	= 8,
};
#define BLK_MQ_FLAG_TO_ALLOC_POLICY(flags) \
	((flags >> BLK_MQ_F_ALLOC_POLICY_START_BIT) & \
		((1 << BLK_MQ_F_ALLOC_POLICY_BITS) - 1))
#define BLK_ALLOC_POLICY_TO_MQ_FLAG(policy) \
	((policy & ((1 << BLK_MQ_F_ALLOC_POLICY_BITS) - 1)) \
		<< BLK_MQ_F_ALLOC_POLICY_START_BIT)

struct request_queue *blk_mq_init_queue(struct blk_mq_tag_set *);
struct request_queue *blk_mq_init_allocated_queue(struct blk_mq_tag_set *set,
						  struct request_queue *q);
int blk_mq_register_disk(struct gendisk *);
void blk_mq_unregister_disk(struct gendisk *);

int blk_mq_alloc_tag_set(struct blk_mq_tag_set *set);
void blk_mq_free_tag_set(struct blk_mq_tag_set *set);

void blk_mq_flush_plug_list(struct blk_plug *plug, bool from_schedule);

void blk_mq_insert_request(struct request *, bool, bool, bool);
void blk_mq_free_request(struct request *rq);
void blk_mq_free_hctx_request(struct blk_mq_hw_ctx *, struct request *rq);
bool blk_mq_can_queue(struct blk_mq_hw_ctx *);
struct request *blk_mq_alloc_request(struct request_queue *q, int rw,
		gfp_t gfp, bool reserved);
struct request *blk_mq_tag_to_rq(struct blk_mq_tags *tags, unsigned int tag);
struct cpumask *blk_mq_tags_cpumask(struct blk_mq_tags *tags);

enum {
	BLK_MQ_UNIQUE_TAG_BITS = 16,
	BLK_MQ_UNIQUE_TAG_MASK = (1 << BLK_MQ_UNIQUE_TAG_BITS) - 1,
};

u32 blk_mq_unique_tag(struct request *rq);

static inline u16 blk_mq_unique_tag_to_hwq(u32 unique_tag)
{
	return unique_tag >> BLK_MQ_UNIQUE_TAG_BITS;
}

static inline u16 blk_mq_unique_tag_to_tag(u32 unique_tag)
{
	return unique_tag & BLK_MQ_UNIQUE_TAG_MASK;
}

struct blk_mq_hw_ctx *blk_mq_map_queue(struct request_queue *, const int ctx_index);
struct blk_mq_hw_ctx *blk_mq_alloc_single_hw_queue(struct blk_mq_tag_set *, unsigned int, int);

int blk_mq_request_started(struct request *rq);
void blk_mq_start_request(struct request *rq);
void blk_mq_end_request(struct request *rq, int error);
void __blk_mq_end_request(struct request *rq, int error);

void blk_mq_requeue_request(struct request *rq);
void blk_mq_add_to_requeue_list(struct request *rq, bool at_head);
void blk_mq_cancel_requeue_work(struct request_queue *q);
void blk_mq_kick_requeue_list(struct request_queue *q);
void blk_mq_abort_requeue_list(struct request_queue *q);
void blk_mq_complete_request(struct request *rq, int error);

void blk_mq_stop_hw_queue(struct blk_mq_hw_ctx *hctx);
void blk_mq_start_hw_queue(struct blk_mq_hw_ctx *hctx);
void blk_mq_stop_hw_queues(struct request_queue *q);
void blk_mq_start_hw_queues(struct request_queue *q);
void blk_mq_start_stopped_hw_queues(struct request_queue *q, bool async);
void blk_mq_run_hw_queues(struct request_queue *q, bool async);
void blk_mq_delay_queue(struct blk_mq_hw_ctx *hctx, unsigned long msecs);
void blk_mq_all_tag_busy_iter(struct blk_mq_tags *tags, busy_tag_iter_fn *fn,
		void *priv);
void blk_mq_freeze_queue(struct request_queue *q);
void blk_mq_unfreeze_queue(struct request_queue *q);
void blk_mq_freeze_queue_start(struct request_queue *q);

/*
 * Driver command data is immediately after the request. So subtract request
 * size to get back to the original request, add request size to get the PDU.
 */
static inline struct request *blk_mq_rq_from_pdu(void *pdu)
{
	return pdu - sizeof(struct request);
}
static inline void *blk_mq_rq_to_pdu(struct request *rq)
{
	return rq + 1;
}

#define queue_for_each_hw_ctx(q, hctx, i)				\
	for ((i) = 0; (i) < (q)->nr_hw_queues &&			\
	     ({ hctx = (q)->queue_hw_ctx[i]; 1; }); (i)++)

#define queue_for_each_ctx(q, ctx, i)					\
	for ((i) = 0; (i) < (q)->nr_queues &&				\
	     ({ ctx = per_cpu_ptr((q)->queue_ctx, (i)); 1; }); (i)++)

#define hctx_for_each_ctx(hctx, ctx, i)					\
	for ((i) = 0; (i) < (hctx)->nr_ctx &&				\
	     ({ ctx = (hctx)->ctxs[(i)]; 1; }); (i)++)

#define blk_ctx_sum(q, sum)						\
({									\
	struct blk_mq_ctx *__x;						\
	unsigned int __ret = 0, __i;					\
									\
	queue_for_each_ctx((q), __x, __i)				\
		__ret += sum;						\
	__ret;								\
})

#endif
Commit	Line	Data
320ae51f JA	1	#ifndef BLK_MQ_H
	2	#define BLK_MQ_H
	3
	4	#include <linux/blkdev.h>
	5
	6	struct blk_mq_tags;
f70ced09	7	struct blk_flush_queue;
320ae51f JA	8
	9	struct blk_mq_cpu_notifier {
	10	struct list_head list;
	11	void *data;
e814e71b	12	int (notify)(void data, unsigned long action, unsigned int cpu);
320ae51f JA	13	};
320ae51f JA	14
1429d7c9	15	struct blk_mq_ctxmap {
569fd0ce	16	unsigned int size;
1429d7c9 JA	17	unsigned int bits_per_word;
	18	struct blk_align_bitmap *map;
	19	};
	20
320ae51f JA	21	struct blk_mq_hw_ctx {
	22	struct {
	23	spinlock_t lock;
	24	struct list_head dispatch;
	25	} ____cacheline_aligned_in_smp;
	26
	27	unsigned long state; /* BLK_MQ_S_* flags */
70f4db63 CH	28	struct delayed_work run_work;
70f4db63 CH	29	struct delayed_work delay_work;
e4043dcf	30	cpumask_var_t cpumask;
506e931f JA	31	int next_cpu;
506e931f JA	32	int next_cpu_batch;
320ae51f JA	33
	34	unsigned long flags; /* BLK_MQ_F_* flags */
	35
	36	struct request_queue *queue;
f70ced09	37	struct blk_flush_queue *fq;
320ae51f JA	38
	39	void *driver_data;
	40
1429d7c9 JA	41	struct blk_mq_ctxmap ctx_map;
1429d7c9 JA	42
4bb659b1 JA	43	unsigned int nr_ctx;
	44	struct blk_mq_ctx **ctxs;
	45
8537b120	46	atomic_t wait_index;
320ae51f	47
320ae51f JA	48	struct blk_mq_tags *tags;
	49
	50	unsigned long queued;
	51	unsigned long run;
	52	#define BLK_MQ_MAX_DISPATCH_ORDER 10
	53	unsigned long dispatched[BLK_MQ_MAX_DISPATCH_ORDER];
	54
320ae51f	55	unsigned int numa_node;
17ded320	56	unsigned int queue_num;
320ae51f	57
0d2602ca JA	58	atomic_t nr_active;
0d2602ca JA	59
320ae51f JA	60	struct blk_mq_cpu_notifier cpu_notifier;
320ae51f JA	61	struct kobject kobj;
05229bee JA	62
	63	unsigned long poll_invoked;
	64	unsigned long poll_success;
320ae51f JA	65	};
320ae51f JA	66
24d2f903	67	struct blk_mq_tag_set {
320ae51f JA	68	struct blk_mq_ops *ops;
320ae51f JA	69	unsigned int nr_hw_queues;
e3a2b3f9	70	unsigned int queue_depth; /* max hw supported */
320ae51f JA	71	unsigned int reserved_tags;
	72	unsigned int cmd_size; /* per-request extra data */
	73	int numa_node;
	74	unsigned int timeout;
	75	unsigned int flags; /* BLK_MQ_F_* */
24d2f903 CH	76	void *driver_data;
	77
	78	struct blk_mq_tags **tags;
0d2602ca JA	79
	80	struct mutex tag_list_lock;
	81	struct list_head tag_list;
320ae51f JA	82	};
320ae51f JA	83
74c45052 JA	84	struct blk_mq_queue_data {
	85	struct request *rq;
	86	struct list_head *list;
	87	bool last;
	88	};
	89
	90	typedef int (queue_rq_fn)(struct blk_mq_hw_ctx , const struct blk_mq_queue_data );
320ae51f	91	typedef struct blk_mq_hw_ctx (map_queue_fn)(struct request_queue , const int);
0152fb6b	92	typedef enum blk_eh_timer_return (timeout_fn)(struct request *, bool);
320ae51f JA	93	typedef int (init_hctx_fn)(struct blk_mq_hw_ctx , void , unsigned int);
320ae51f JA	94	typedef void (exit_hctx_fn)(struct blk_mq_hw_ctx *, unsigned int);
24d2f903 CH	95	typedef int (init_request_fn)(void , struct request , unsigned int,
	96	unsigned int, unsigned int);
	97	typedef void (exit_request_fn)(void , struct request , unsigned int,
	98	unsigned int);
320ae51f	99
81481eb4 CH	100	typedef void (busy_iter_fn)(struct blk_mq_hw_ctx , struct request , void *,
81481eb4 CH	101	bool);
f26cdc85	102	typedef void (busy_tag_iter_fn)(struct request , void , bool);
05229bee JA	103	typedef int (poll_fn)(struct blk_mq_hw_ctx *, unsigned int);
05229bee JA	104
81481eb4	105
320ae51f JA	106	struct blk_mq_ops {
	107	/*
	108	* Queue request
	109	*/
	110	queue_rq_fn *queue_rq;
	111
	112	/*
	113	* Map to specific hardware queue
	114	*/
	115	map_queue_fn *map_queue;
	116
	117	/*
	118	* Called on request timeout
	119	*/
0152fb6b	120	timeout_fn *timeout;
320ae51f	121
05229bee JA	122	/*
	123	* Called to poll for completion of a specific tag.
	124	*/
	125	poll_fn *poll;
	126
30a91cb4 CH	127	softirq_done_fn *complete;
30a91cb4 CH	128
320ae51f JA	129	/*
	130	* Called when the block layer side of a hardware queue has been
	131	* set up, allowing the driver to allocate/init matching structures.
	132	* Ditto for exit/teardown.
	133	*/
	134	init_hctx_fn *init_hctx;
	135	exit_hctx_fn *exit_hctx;
e9b267d9 CH	136
	137	/*
	138	* Called for every command allocated by the block layer to allow
	139	* the driver to set up driver specific data.
f70ced09 ML	140	*
	141	* Tag greater than or equal to queue_depth is for setting up
	142	* flush request.
	143	*
e9b267d9 CH	144	* Ditto for exit/teardown.
	145	*/
	146	init_request_fn *init_request;
	147	exit_request_fn *exit_request;
320ae51f JA	148	};
	149
	150	enum {
	151	BLK_MQ_RQ_QUEUE_OK = 0, /* queued fine */
	152	BLK_MQ_RQ_QUEUE_BUSY = 1, /* requeue IO for later */
	153	BLK_MQ_RQ_QUEUE_ERROR = 2, /* end IO with error */
	154
	155	BLK_MQ_F_SHOULD_MERGE = 1 << 0,
8a58d1f1 JA	156	BLK_MQ_F_TAG_SHARED = 1 << 1,
8a58d1f1 JA	157	BLK_MQ_F_SG_MERGE = 1 << 2,
e167dfb5	158	BLK_MQ_F_DEFER_ISSUE = 1 << 4,
24391c0d SL	159	BLK_MQ_F_ALLOC_POLICY_START_BIT = 8,
24391c0d SL	160	BLK_MQ_F_ALLOC_POLICY_BITS = 1,
320ae51f	161
5d12f905	162	BLK_MQ_S_STOPPED = 0,
0d2602ca	163	BLK_MQ_S_TAG_ACTIVE = 1,
320ae51f	164
a4391c64	165	BLK_MQ_MAX_DEPTH = 10240,
506e931f JA	166
506e931f JA	167	BLK_MQ_CPU_WORK_BATCH = 8,
320ae51f	168	};
24391c0d SL	169	#define BLK_MQ_FLAG_TO_ALLOC_POLICY(flags) \
	170	((flags >> BLK_MQ_F_ALLOC_POLICY_START_BIT) & \
	171	((1 << BLK_MQ_F_ALLOC_POLICY_BITS) - 1))
	172	#define BLK_ALLOC_POLICY_TO_MQ_FLAG(policy) \
	173	((policy & ((1 << BLK_MQ_F_ALLOC_POLICY_BITS) - 1)) \
	174	<< BLK_MQ_F_ALLOC_POLICY_START_BIT)
320ae51f	175
24d2f903	176	struct request_queue blk_mq_init_queue(struct blk_mq_tag_set );
b62c21b7 MS	177	struct request_queue blk_mq_init_allocated_queue(struct blk_mq_tag_set set,
b62c21b7 MS	178	struct request_queue *q);
320ae51f JA	179	int blk_mq_register_disk(struct gendisk *);
320ae51f JA	180	void blk_mq_unregister_disk(struct gendisk *);
320ae51f	181
24d2f903 CH	182	int blk_mq_alloc_tag_set(struct blk_mq_tag_set *set);
	183	void blk_mq_free_tag_set(struct blk_mq_tag_set *set);
	184
320ae51f JA	185	void blk_mq_flush_plug_list(struct blk_plug *plug, bool from_schedule);
320ae51f JA	186
feb71dae	187	void blk_mq_insert_request(struct request *, bool, bool, bool);
320ae51f	188	void blk_mq_free_request(struct request *rq);
7c7f2f2b	189	void blk_mq_free_hctx_request(struct blk_mq_hw_ctx , struct request rq);
320ae51f	190	bool blk_mq_can_queue(struct blk_mq_hw_ctx *);
4ce01dd1 CH	191	struct request blk_mq_alloc_request(struct request_queue q, int rw,
4ce01dd1 CH	192	gfp_t gfp, bool reserved);
0e62f51f	193	struct request blk_mq_tag_to_rq(struct blk_mq_tags tags, unsigned int tag);
f26cdc85	194	struct cpumask blk_mq_tags_cpumask(struct blk_mq_tags tags);
320ae51f	195
205fb5f5 BVA	196	enum {
	197	BLK_MQ_UNIQUE_TAG_BITS = 16,
	198	BLK_MQ_UNIQUE_TAG_MASK = (1 << BLK_MQ_UNIQUE_TAG_BITS) - 1,
	199	};
	200
	201	u32 blk_mq_unique_tag(struct request *rq);
	202
	203	static inline u16 blk_mq_unique_tag_to_hwq(u32 unique_tag)
	204	{
	205	return unique_tag >> BLK_MQ_UNIQUE_TAG_BITS;
	206	}
	207
	208	static inline u16 blk_mq_unique_tag_to_tag(u32 unique_tag)
	209	{
	210	return unique_tag & BLK_MQ_UNIQUE_TAG_MASK;
	211	}
	212
320ae51f	213	struct blk_mq_hw_ctx blk_mq_map_queue(struct request_queue , const int ctx_index);
f14bbe77	214	struct blk_mq_hw_ctx blk_mq_alloc_single_hw_queue(struct blk_mq_tag_set , unsigned int, int);
320ae51f	215
973c0191	216	int blk_mq_request_started(struct request *rq);
e2490073	217	void blk_mq_start_request(struct request *rq);
c8a446ad CH	218	void blk_mq_end_request(struct request *rq, int error);
c8a446ad CH	219	void __blk_mq_end_request(struct request *rq, int error);
320ae51f	220
ed0791b2	221	void blk_mq_requeue_request(struct request *rq);
6fca6a61	222	void blk_mq_add_to_requeue_list(struct request *rq, bool at_head);
c68ed59f	223	void blk_mq_cancel_requeue_work(struct request_queue *q);
6fca6a61	224	void blk_mq_kick_requeue_list(struct request_queue *q);
1885b24d	225	void blk_mq_abort_requeue_list(struct request_queue *q);
f4829a9b	226	void blk_mq_complete_request(struct request *rq, int error);
30a91cb4	227
320ae51f JA	228	void blk_mq_stop_hw_queue(struct blk_mq_hw_ctx *hctx);
320ae51f JA	229	void blk_mq_start_hw_queue(struct blk_mq_hw_ctx *hctx);
280d45f6	230	void blk_mq_stop_hw_queues(struct request_queue *q);
2f268556	231	void blk_mq_start_hw_queues(struct request_queue *q);
1b4a3258	232	void blk_mq_start_stopped_hw_queues(struct request_queue *q, bool async);
b94ec296	233	void blk_mq_run_hw_queues(struct request_queue *q, bool async);
70f4db63	234	void blk_mq_delay_queue(struct blk_mq_hw_ctx *hctx, unsigned long msecs);
f26cdc85 KB	235	void blk_mq_all_tag_busy_iter(struct blk_mq_tags tags, busy_tag_iter_fn fn,
f26cdc85 KB	236	void *priv);
c761d96b	237	void blk_mq_freeze_queue(struct request_queue *q);
b4c6a028 KB	238	void blk_mq_unfreeze_queue(struct request_queue *q);
b4c6a028 KB	239	void blk_mq_freeze_queue_start(struct request_queue *q);
320ae51f JA	240
	241	/*
	242	* Driver command data is immediately after the request. So subtract request
2963e3f7	243	* size to get back to the original request, add request size to get the PDU.
320ae51f JA	244	*/
	245	static inline struct request blk_mq_rq_from_pdu(void pdu)
	246	{
	247	return pdu - sizeof(struct request);
	248	}
	249	static inline void blk_mq_rq_to_pdu(struct request rq)
	250	{
2963e3f7	251	return rq + 1;
320ae51f JA	252	}
320ae51f JA	253
320ae51f	254	#define queue_for_each_hw_ctx(q, hctx, i) \
0d0b7d42 JA	255	for ((i) = 0; (i) < (q)->nr_hw_queues && \
0d0b7d42 JA	256	({ hctx = (q)->queue_hw_ctx[i]; 1; }); (i)++)
320ae51f JA	257
320ae51f JA	258	#define queue_for_each_ctx(q, ctx, i) \
0d0b7d42 JA	259	for ((i) = 0; (i) < (q)->nr_queues && \
0d0b7d42 JA	260	({ ctx = per_cpu_ptr((q)->queue_ctx, (i)); 1; }); (i)++)
320ae51f JA	261
320ae51f JA	262	#define hctx_for_each_ctx(hctx, ctx, i) \
0d0b7d42 JA	263	for ((i) = 0; (i) < (hctx)->nr_ctx && \
0d0b7d42 JA	264	({ ctx = (hctx)->ctxs[(i)]; 1; }); (i)++)
320ae51f JA	265
	266	#define blk_ctx_sum(q, sum) \
	267	({ \
	268	struct blk_mq_ctx *__x; \
	269	unsigned int __ret = 0, __i; \
	270	\
	271	queue_for_each_ctx((q), __x, __i) \
	272	__ret += sum; \
	273	__ret; \
	274	})
	275
	276	#endif