[linux-2.6-block.git] / include / linux / blk-mq.h

/* SPDX-License-Identifier: GPL-2.0 */
#ifndef BLK_MQ_H
#define BLK_MQ_H

#include <linux/blkdev.h>
#include <linux/sbitmap.h>
#include <linux/srcu.h>

struct blk_mq_tags;
struct blk_flush_queue;

/**
 * struct blk_mq_hw_ctx - State for a hardware queue facing the hardware block device
 */
struct blk_mq_hw_ctx {
	struct {
		spinlock_t		lock;
		struct list_head	dispatch;
		unsigned long		state;		/* BLK_MQ_S_* flags */
	} ____cacheline_aligned_in_smp;

	struct delayed_work	run_work;
	cpumask_var_t		cpumask;
	int			next_cpu;
	int			next_cpu_batch;

	unsigned long		flags;		/* BLK_MQ_F_* flags */

	void			*sched_data;
	struct request_queue	*queue;
	struct blk_flush_queue	*fq;

	void			*driver_data;

	struct sbitmap		ctx_map;

	struct blk_mq_ctx	*dispatch_from;
	unsigned int		dispatch_busy;

	unsigned short		type;
	unsigned short		nr_ctx;
	struct blk_mq_ctx	**ctxs;

	spinlock_t		dispatch_wait_lock;
	wait_queue_entry_t	dispatch_wait;
	atomic_t		wait_index;

	struct blk_mq_tags	*tags;
	struct blk_mq_tags	*sched_tags;

	unsigned long		queued;
	unsigned long		run;
#define BLK_MQ_MAX_DISPATCH_ORDER	7
	unsigned long		dispatched[BLK_MQ_MAX_DISPATCH_ORDER];

	unsigned int		numa_node;
	unsigned int		queue_num;

	atomic_t		nr_active;
	unsigned int		nr_expired;

	struct hlist_node	cpuhp_dead;
	struct kobject		kobj;

	unsigned long		poll_considered;
	unsigned long		poll_invoked;
	unsigned long		poll_success;

#ifdef CONFIG_BLK_DEBUG_FS
	struct dentry		*debugfs_dir;
	struct dentry		*sched_debugfs_dir;
#endif

	/* Must be the last member - see also blk_mq_hw_ctx_size(). */
	struct srcu_struct	srcu[0];
};

struct blk_mq_queue_map {
	unsigned int *mq_map;
	unsigned int nr_queues;
	unsigned int queue_offset;
};

enum {
	HCTX_MAX_TYPES = 3,
};

struct blk_mq_tag_set {
	/*
	 * map[] holds ctx -> hctx mappings, one map exists for each type
	 * that the driver wishes to support. There are no restrictions
	 * on maps being of the same size, and it's perfectly legal to
	 * share maps between types.
	 */
	struct blk_mq_queue_map	map[HCTX_MAX_TYPES];
	unsigned int		nr_maps;	/* nr entries in map[] */
	const struct blk_mq_ops	*ops;
	unsigned int		nr_hw_queues;	/* nr hw queues across maps */
	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;
	bool last;
};

typedef blk_status_t (queue_rq_fn)(struct blk_mq_hw_ctx *,
		const struct blk_mq_queue_data *);
/* takes rq->cmd_flags as input, returns a hardware type index */
typedef int (rq_flags_to_type_fn)(struct request_queue *, unsigned int);
typedef bool (get_budget_fn)(struct blk_mq_hw_ctx *);
typedef void (put_budget_fn)(struct blk_mq_hw_ctx *);
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)(struct blk_mq_tag_set *set, struct request *,
		unsigned int, unsigned int);
typedef void (exit_request_fn)(struct blk_mq_tag_set *set, struct request *,
		unsigned int);

typedef bool (busy_iter_fn)(struct blk_mq_hw_ctx *, struct request *, void *,
		bool);
typedef bool (busy_tag_iter_fn)(struct request *, void *, bool);
typedef int (poll_fn)(struct blk_mq_hw_ctx *, unsigned int);
typedef int (map_queues_fn)(struct blk_mq_tag_set *set);
typedef bool (busy_fn)(struct request_queue *);
typedef void (complete_fn)(struct request *);


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

	/*
	 * Return a queue map type for the given request/bio flags
	 */
	rq_flags_to_type_fn	*rq_flags_to_type;

	/*
	 * Reserve budget before queue request, once .queue_rq is
	 * run, it is driver's responsibility to release the
	 * reserved budget. Also we have to handle failure case
	 * of .get_budget for avoiding I/O deadlock.
	 */
	get_budget_fn		*get_budget;
	put_budget_fn		*put_budget;

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

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

	complete_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;
	/* Called from inside blk_get_request() */
	void (*initialize_rq_fn)(struct request *rq);

	/*
	 * If set, returns whether or not this queue currently is busy
	 */
	busy_fn			*busy;

	map_queues_fn		*map_queues;

#ifdef CONFIG_BLK_DEBUG_FS
	/*
	 * Used by the debugfs implementation to show driver-specific
	 * information about a request.
	 */
	void (*show_rq)(struct seq_file *m, struct request *rq);
#endif
};

enum {
	BLK_MQ_F_SHOULD_MERGE	= 1 << 0,
	BLK_MQ_F_TAG_SHARED	= 1 << 1,
	BLK_MQ_F_SG_MERGE	= 1 << 2,
	BLK_MQ_F_BLOCKING	= 1 << 5,
	BLK_MQ_F_NO_SCHED	= 1 << 6,
	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_S_SCHED_RESTART	= 2,

	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);
struct request_queue *blk_mq_init_sq_queue(struct blk_mq_tag_set *set,
						const struct blk_mq_ops *ops,
						unsigned int queue_depth,
						unsigned int set_flags);
int blk_mq_register_dev(struct device *, struct request_queue *);
void blk_mq_unregister_dev(struct device *, struct request_queue *);

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_free_request(struct request *rq);
bool blk_mq_can_queue(struct blk_mq_hw_ctx *);

enum {
	/* return when out of requests */
	BLK_MQ_REQ_NOWAIT	= (__force blk_mq_req_flags_t)(1 << 0),
	/* allocate from reserved pool */
	BLK_MQ_REQ_RESERVED	= (__force blk_mq_req_flags_t)(1 << 1),
	/* allocate internal/sched tag */
	BLK_MQ_REQ_INTERNAL	= (__force blk_mq_req_flags_t)(1 << 2),
	/* set RQF_PREEMPT */
	BLK_MQ_REQ_PREEMPT	= (__force blk_mq_req_flags_t)(1 << 3),
};

struct request *blk_mq_alloc_request(struct request_queue *q, unsigned int op,
		blk_mq_req_flags_t flags);
struct request *blk_mq_alloc_request_hctx(struct request_queue *q,
		unsigned int op, blk_mq_req_flags_t flags,
		unsigned int hctx_idx);
struct request *blk_mq_tag_to_rq(struct blk_mq_tags *tags, unsigned int tag);

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;
}


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

void blk_mq_requeue_request(struct request *rq, bool kick_requeue_list);
void blk_mq_add_to_requeue_list(struct request *rq, bool at_head,
				bool kick_requeue_list);
void blk_mq_kick_requeue_list(struct request_queue *q);
void blk_mq_delay_kick_requeue_list(struct request_queue *q, unsigned long msecs);
void blk_mq_complete_request(struct request *rq);
bool blk_mq_bio_list_merge(struct request_queue *q, struct list_head *list,
			   struct bio *bio);
bool blk_mq_queue_stopped(struct request_queue *q);
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_queue(struct blk_mq_hw_ctx *hctx, bool async);
void blk_mq_start_stopped_hw_queues(struct request_queue *q, bool async);
void blk_mq_quiesce_queue(struct request_queue *q);
void blk_mq_unquiesce_queue(struct request_queue *q);
void blk_mq_delay_run_hw_queue(struct blk_mq_hw_ctx *hctx, unsigned long msecs);
bool blk_mq_run_hw_queue(struct blk_mq_hw_ctx *hctx, bool async);
void blk_mq_run_hw_queues(struct request_queue *q, bool async);
void blk_mq_tagset_busy_iter(struct blk_mq_tag_set *tagset,
		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_freeze_queue_start(struct request_queue *q);
void blk_mq_freeze_queue_wait(struct request_queue *q);
int blk_mq_freeze_queue_wait_timeout(struct request_queue *q,
				     unsigned long timeout);

int blk_mq_map_queues(struct blk_mq_queue_map *qmap);
void blk_mq_update_nr_hw_queues(struct blk_mq_tag_set *set, int nr_hw_queues);

void blk_mq_quiesce_queue_nowait(struct request_queue *q);

unsigned int blk_mq_rq_cpu(struct request *rq);

/**
 * blk_mq_mark_complete() - Set request state to complete
 * @rq: request to set to complete state
 *
 * Returns true if request state was successfully set to complete. If
 * successful, the caller is responsibile for seeing this request is ended, as
 * blk_mq_complete_request will not work again.
 */
static inline bool blk_mq_mark_complete(struct request *rq)
{
	return cmpxchg(&rq->state, MQ_RQ_IN_FLIGHT, MQ_RQ_COMPLETE) ==
			MQ_RQ_IN_FLIGHT;
}

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

#endif
Commit	Line	Data
b2441318	1	/* SPDX-License-Identifier: GPL-2.0 */
320ae51f JA	2	#ifndef BLK_MQ_H
	3	#define BLK_MQ_H
	4
	5	#include <linux/blkdev.h>
88459642	6	#include <linux/sbitmap.h>
6a83e74d	7	#include <linux/srcu.h>
320ae51f JA	8
320ae51f JA	9	struct blk_mq_tags;
f70ced09	10	struct blk_flush_queue;
320ae51f	11
fe644072 LW	12	/**
	13	* struct blk_mq_hw_ctx - State for a hardware queue facing the hardware block device
	14	*/
320ae51f JA	15	struct blk_mq_hw_ctx {
	16	struct {
	17	spinlock_t lock;
	18	struct list_head dispatch;
8d354f13	19	unsigned long state; /* BLK_MQ_S_* flags */
320ae51f JA	20	} ____cacheline_aligned_in_smp;
320ae51f JA	21
9f993737	22	struct delayed_work run_work;
e4043dcf	23	cpumask_var_t cpumask;
506e931f JA	24	int next_cpu;
506e931f JA	25	int next_cpu_batch;
320ae51f JA	26
	27	unsigned long flags; /* BLK_MQ_F_* flags */
	28
bd166ef1	29	void *sched_data;
320ae51f	30	struct request_queue *queue;
f70ced09	31	struct blk_flush_queue *fq;
320ae51f JA	32
	33	void *driver_data;
	34
88459642	35	struct sbitmap ctx_map;
1429d7c9	36
b347689f	37	struct blk_mq_ctx *dispatch_from;
6e768717	38	unsigned int dispatch_busy;
b347689f	39
f31967f0 JA	40	unsigned short type;
f31967f0 JA	41	unsigned short nr_ctx;
6e768717	42	struct blk_mq_ctx **ctxs;
4bb659b1	43
5815839b	44	spinlock_t dispatch_wait_lock;
eb619fdb	45	wait_queue_entry_t dispatch_wait;
8537b120	46	atomic_t wait_index;
320ae51f	47
320ae51f	48	struct blk_mq_tags *tags;
bd166ef1	49	struct blk_mq_tags *sched_tags;
320ae51f JA	50
	51	unsigned long queued;
	52	unsigned long run;
8d354f13	53	#define BLK_MQ_MAX_DISPATCH_ORDER 7
320ae51f JA	54	unsigned long dispatched[BLK_MQ_MAX_DISPATCH_ORDER];
320ae51f JA	55
320ae51f	56	unsigned int numa_node;
17ded320	57	unsigned int queue_num;
320ae51f	58
0d2602ca	59	atomic_t nr_active;
1d9bd516	60	unsigned int nr_expired;
0d2602ca	61
9467f859	62	struct hlist_node cpuhp_dead;
320ae51f	63	struct kobject kobj;
05229bee	64
6e219353	65	unsigned long poll_considered;
05229bee JA	66	unsigned long poll_invoked;
05229bee JA	67	unsigned long poll_success;
9c1051aa OS	68
	69	#ifdef CONFIG_BLK_DEBUG_FS
	70	struct dentry *debugfs_dir;
d332ce09	71	struct dentry *sched_debugfs_dir;
9c1051aa	72	#endif
07319678 BVA	73
07319678 BVA	74	/* Must be the last member - see also blk_mq_hw_ctx_size(). */
05707b64	75	struct srcu_struct srcu[0];
320ae51f JA	76	};
320ae51f JA	77
ed76e329 JA	78	struct blk_mq_queue_map {
	79	unsigned int *mq_map;
	80	unsigned int nr_queues;
843477d4	81	unsigned int queue_offset;
ed76e329 JA	82	};
	83
	84	enum {
4b04cc6a	85	HCTX_MAX_TYPES = 3,
ed76e329 JA	86	};
ed76e329 JA	87
24d2f903	88	struct blk_mq_tag_set {
b3c661b1 JA	89	/*
	90	* map[] holds ctx -> hctx mappings, one map exists for each type
	91	* that the driver wishes to support. There are no restrictions
	92	* on maps being of the same size, and it's perfectly legal to
	93	* share maps between types.
	94	*/
ed76e329	95	struct blk_mq_queue_map map[HCTX_MAX_TYPES];
b3c661b1	96	unsigned int nr_maps; /* nr entries in map[] */
f8a5b122	97	const struct blk_mq_ops *ops;
ed76e329	98	unsigned int nr_hw_queues; /* nr hw queues across maps */
e3a2b3f9	99	unsigned int queue_depth; /* max hw supported */
320ae51f JA	100	unsigned int reserved_tags;
	101	unsigned int cmd_size; /* per-request extra data */
	102	int numa_node;
	103	unsigned int timeout;
	104	unsigned int flags; /* BLK_MQ_F_* */
24d2f903 CH	105	void *driver_data;
	106
	107	struct blk_mq_tags **tags;
0d2602ca JA	108
	109	struct mutex tag_list_lock;
	110	struct list_head tag_list;
320ae51f JA	111	};
320ae51f JA	112
74c45052 JA	113	struct blk_mq_queue_data {
74c45052 JA	114	struct request *rq;
74c45052 JA	115	bool last;
	116	};
	117
fc17b653 CH	118	typedef blk_status_t (queue_rq_fn)(struct blk_mq_hw_ctx *,
fc17b653 CH	119	const struct blk_mq_queue_data *);
b3c661b1 JA	120	/* takes rq->cmd_flags as input, returns a hardware type index */
b3c661b1 JA	121	typedef int (rq_flags_to_type_fn)(struct request_queue *, unsigned int);
88022d72	122	typedef bool (get_budget_fn)(struct blk_mq_hw_ctx *);
de148297	123	typedef void (put_budget_fn)(struct blk_mq_hw_ctx *);
0152fb6b	124	typedef enum blk_eh_timer_return (timeout_fn)(struct request *, bool);
320ae51f JA	125	typedef int (init_hctx_fn)(struct blk_mq_hw_ctx , void , unsigned int);
320ae51f JA	126	typedef void (exit_hctx_fn)(struct blk_mq_hw_ctx *, unsigned int);
d6296d39	127	typedef int (init_request_fn)(struct blk_mq_tag_set set, struct request ,
24d2f903	128	unsigned int, unsigned int);
d6296d39	129	typedef void (exit_request_fn)(struct blk_mq_tag_set set, struct request ,
24d2f903	130	unsigned int);
320ae51f	131
7baa8572	132	typedef bool (busy_iter_fn)(struct blk_mq_hw_ctx , struct request , void *,
81481eb4	133	bool);
7baa8572	134	typedef bool (busy_tag_iter_fn)(struct request , void , bool);
05229bee	135	typedef int (poll_fn)(struct blk_mq_hw_ctx *, unsigned int);
da695ba2	136	typedef int (map_queues_fn)(struct blk_mq_tag_set *set);
9ba20527	137	typedef bool (busy_fn)(struct request_queue *);
c7bb9ad1	138	typedef void (complete_fn)(struct request *);
05229bee	139
81481eb4	140
320ae51f JA	141	struct blk_mq_ops {
	142	/*
	143	* Queue request
	144	*/
	145	queue_rq_fn *queue_rq;
	146
b3c661b1 JA	147	/*
	148	* Return a queue map type for the given request/bio flags
	149	*/
	150	rq_flags_to_type_fn *rq_flags_to_type;
	151
de148297 ML	152	/*
	153	* Reserve budget before queue request, once .queue_rq is
	154	* run, it is driver's responsibility to release the
	155	* reserved budget. Also we have to handle failure case
	156	* of .get_budget for avoiding I/O deadlock.
	157	*/
	158	get_budget_fn *get_budget;
	159	put_budget_fn *put_budget;
	160
320ae51f JA	161	/*
	162	* Called on request timeout
	163	*/
0152fb6b	164	timeout_fn *timeout;
320ae51f	165
05229bee JA	166	/*
	167	* Called to poll for completion of a specific tag.
	168	*/
	169	poll_fn *poll;
	170
c7bb9ad1	171	complete_fn *complete;
30a91cb4	172
320ae51f JA	173	/*
	174	* Called when the block layer side of a hardware queue has been
	175	* set up, allowing the driver to allocate/init matching structures.
	176	* Ditto for exit/teardown.
	177	*/
	178	init_hctx_fn *init_hctx;
	179	exit_hctx_fn *exit_hctx;
e9b267d9 CH	180
	181	/*
	182	* Called for every command allocated by the block layer to allow
	183	* the driver to set up driver specific data.
f70ced09 ML	184	*
	185	* Tag greater than or equal to queue_depth is for setting up
	186	* flush request.
	187	*
e9b267d9 CH	188	* Ditto for exit/teardown.
	189	*/
	190	init_request_fn *init_request;
	191	exit_request_fn *exit_request;
d280bab3 BVA	192	/* Called from inside blk_get_request() */
d280bab3 BVA	193	void (initialize_rq_fn)(struct request rq);
da695ba2	194
9ba20527 JA	195	/*
	196	* If set, returns whether or not this queue currently is busy
	197	*/
	198	busy_fn *busy;
	199
da695ba2	200	map_queues_fn *map_queues;
2836ee4b BVA	201
	202	#ifdef CONFIG_BLK_DEBUG_FS
	203	/*
	204	* Used by the debugfs implementation to show driver-specific
	205	* information about a request.
	206	*/
	207	void (show_rq)(struct seq_file m, struct request *rq);
	208	#endif
320ae51f JA	209	};
	210
	211	enum {
320ae51f	212	BLK_MQ_F_SHOULD_MERGE = 1 << 0,
8a58d1f1 JA	213	BLK_MQ_F_TAG_SHARED = 1 << 1,
8a58d1f1 JA	214	BLK_MQ_F_SG_MERGE = 1 << 2,
1b792f2f	215	BLK_MQ_F_BLOCKING = 1 << 5,
d3484991	216	BLK_MQ_F_NO_SCHED = 1 << 6,
24391c0d SL	217	BLK_MQ_F_ALLOC_POLICY_START_BIT = 8,
24391c0d SL	218	BLK_MQ_F_ALLOC_POLICY_BITS = 1,
320ae51f	219
5d12f905	220	BLK_MQ_S_STOPPED = 0,
0d2602ca	221	BLK_MQ_S_TAG_ACTIVE = 1,
bd166ef1	222	BLK_MQ_S_SCHED_RESTART = 2,
320ae51f	223
a4391c64	224	BLK_MQ_MAX_DEPTH = 10240,
506e931f JA	225
506e931f JA	226	BLK_MQ_CPU_WORK_BATCH = 8,
320ae51f	227	};
24391c0d SL	228	#define BLK_MQ_FLAG_TO_ALLOC_POLICY(flags) \
	229	((flags >> BLK_MQ_F_ALLOC_POLICY_START_BIT) & \
	230	((1 << BLK_MQ_F_ALLOC_POLICY_BITS) - 1))
	231	#define BLK_ALLOC_POLICY_TO_MQ_FLAG(policy) \
	232	((policy & ((1 << BLK_MQ_F_ALLOC_POLICY_BITS) - 1)) \
	233	<< BLK_MQ_F_ALLOC_POLICY_START_BIT)
320ae51f	234
24d2f903	235	struct request_queue blk_mq_init_queue(struct blk_mq_tag_set );
b62c21b7 MS	236	struct request_queue blk_mq_init_allocated_queue(struct blk_mq_tag_set set,
b62c21b7 MS	237	struct request_queue *q);
9316a9ed JA	238	struct request_queue blk_mq_init_sq_queue(struct blk_mq_tag_set set,
	239	const struct blk_mq_ops *ops,
	240	unsigned int queue_depth,
	241	unsigned int set_flags);
b21d5b30 MB	242	int blk_mq_register_dev(struct device , struct request_queue );
b21d5b30 MB	243	void blk_mq_unregister_dev(struct device , struct request_queue );
320ae51f	244
24d2f903 CH	245	int blk_mq_alloc_tag_set(struct blk_mq_tag_set *set);
	246	void blk_mq_free_tag_set(struct blk_mq_tag_set *set);
	247
320ae51f JA	248	void blk_mq_flush_plug_list(struct blk_plug *plug, bool from_schedule);
320ae51f JA	249
320ae51f JA	250	void blk_mq_free_request(struct request *rq);
320ae51f JA	251	bool blk_mq_can_queue(struct blk_mq_hw_ctx *);
6f3b0e8b CH	252
6f3b0e8b CH	253	enum {
9a95e4ef BVA	254	/* return when out of requests */
	255	BLK_MQ_REQ_NOWAIT = (__force blk_mq_req_flags_t)(1 << 0),
	256	/* allocate from reserved pool */
	257	BLK_MQ_REQ_RESERVED = (__force blk_mq_req_flags_t)(1 << 1),
	258	/* allocate internal/sched tag */
	259	BLK_MQ_REQ_INTERNAL = (__force blk_mq_req_flags_t)(1 << 2),
	260	/* set RQF_PREEMPT */
	261	BLK_MQ_REQ_PREEMPT = (__force blk_mq_req_flags_t)(1 << 3),
6f3b0e8b CH	262	};
6f3b0e8b CH	263
cd6ce148	264	struct request blk_mq_alloc_request(struct request_queue q, unsigned int op,
9a95e4ef	265	blk_mq_req_flags_t flags);
cd6ce148	266	struct request blk_mq_alloc_request_hctx(struct request_queue q,
9a95e4ef BVA	267	unsigned int op, blk_mq_req_flags_t flags,
9a95e4ef BVA	268	unsigned int hctx_idx);
0e62f51f	269	struct request blk_mq_tag_to_rq(struct blk_mq_tags tags, unsigned int tag);
320ae51f	270
205fb5f5 BVA	271	enum {
	272	BLK_MQ_UNIQUE_TAG_BITS = 16,
	273	BLK_MQ_UNIQUE_TAG_MASK = (1 << BLK_MQ_UNIQUE_TAG_BITS) - 1,
	274	};
	275
	276	u32 blk_mq_unique_tag(struct request *rq);
	277
	278	static inline u16 blk_mq_unique_tag_to_hwq(u32 unique_tag)
	279	{
	280	return unique_tag >> BLK_MQ_UNIQUE_TAG_BITS;
	281	}
	282
	283	static inline u16 blk_mq_unique_tag_to_tag(u32 unique_tag)
	284	{
	285	return unique_tag & BLK_MQ_UNIQUE_TAG_MASK;
	286	}
	287
320ae51f	288
973c0191	289	int blk_mq_request_started(struct request *rq);
e2490073	290	void blk_mq_start_request(struct request *rq);
2a842aca CH	291	void blk_mq_end_request(struct request *rq, blk_status_t error);
2a842aca CH	292	void __blk_mq_end_request(struct request *rq, blk_status_t error);
320ae51f	293
2b053aca BVA	294	void blk_mq_requeue_request(struct request *rq, bool kick_requeue_list);
	295	void blk_mq_add_to_requeue_list(struct request *rq, bool at_head,
	296	bool kick_requeue_list);
6fca6a61	297	void blk_mq_kick_requeue_list(struct request_queue *q);
2849450a	298	void blk_mq_delay_kick_requeue_list(struct request_queue *q, unsigned long msecs);
08e0029a	299	void blk_mq_complete_request(struct request *rq);
9c558734 JA	300	bool blk_mq_bio_list_merge(struct request_queue q, struct list_head list,
9c558734 JA	301	struct bio *bio);
fd001443	302	bool blk_mq_queue_stopped(struct request_queue *q);
320ae51f JA	303	void blk_mq_stop_hw_queue(struct blk_mq_hw_ctx *hctx);
320ae51f JA	304	void blk_mq_start_hw_queue(struct blk_mq_hw_ctx *hctx);
280d45f6	305	void blk_mq_stop_hw_queues(struct request_queue *q);
2f268556	306	void blk_mq_start_hw_queues(struct request_queue *q);
ae911c5e	307	void blk_mq_start_stopped_hw_queue(struct blk_mq_hw_ctx *hctx, bool async);
1b4a3258	308	void blk_mq_start_stopped_hw_queues(struct request_queue *q, bool async);
97e01209	309	void blk_mq_quiesce_queue(struct request_queue *q);
e4e73913	310	void blk_mq_unquiesce_queue(struct request_queue *q);
7587a5ae	311	void blk_mq_delay_run_hw_queue(struct blk_mq_hw_ctx *hctx, unsigned long msecs);
79f720a7	312	bool blk_mq_run_hw_queue(struct blk_mq_hw_ctx *hctx, bool async);
b94ec296	313	void blk_mq_run_hw_queues(struct request_queue *q, bool async);
e0489487 SG	314	void blk_mq_tagset_busy_iter(struct blk_mq_tag_set *tagset,
e0489487 SG	315	busy_tag_iter_fn fn, void priv);
c761d96b	316	void blk_mq_freeze_queue(struct request_queue *q);
b4c6a028	317	void blk_mq_unfreeze_queue(struct request_queue *q);
1671d522	318	void blk_freeze_queue_start(struct request_queue *q);
6bae363e	319	void blk_mq_freeze_queue_wait(struct request_queue *q);
f91328c4 KB	320	int blk_mq_freeze_queue_wait_timeout(struct request_queue *q,
f91328c4 KB	321	unsigned long timeout);
320ae51f	322
ed76e329	323	int blk_mq_map_queues(struct blk_mq_queue_map *qmap);
868f2f0b KB	324	void blk_mq_update_nr_hw_queues(struct blk_mq_tag_set *set, int nr_hw_queues);
868f2f0b KB	325
852ec809	326	void blk_mq_quiesce_queue_nowait(struct request_queue *q);
4f084b41	327
9cf2bab6 JA	328	unsigned int blk_mq_rq_cpu(struct request *rq);
9cf2bab6 JA	329
0fc09f92 KB	330	/**
	331	* blk_mq_mark_complete() - Set request state to complete
	332	* @rq: request to set to complete state
	333	*
	334	* Returns true if request state was successfully set to complete. If
	335	* successful, the caller is responsibile for seeing this request is ended, as
	336	* blk_mq_complete_request will not work again.
	337	*/
	338	static inline bool blk_mq_mark_complete(struct request *rq)
	339	{
	340	return cmpxchg(&rq->state, MQ_RQ_IN_FLIGHT, MQ_RQ_COMPLETE) ==
	341	MQ_RQ_IN_FLIGHT;
	342	}
	343
320ae51f JA	344	/*
320ae51f JA	345	* Driver command data is immediately after the request. So subtract request
2963e3f7	346	* size to get back to the original request, add request size to get the PDU.
320ae51f JA	347	*/
	348	static inline struct request blk_mq_rq_from_pdu(void pdu)
	349	{
	350	return pdu - sizeof(struct request);
	351	}
	352	static inline void blk_mq_rq_to_pdu(struct request rq)
	353	{
2963e3f7	354	return rq + 1;
320ae51f JA	355	}
320ae51f JA	356
320ae51f	357	#define queue_for_each_hw_ctx(q, hctx, i) \
0d0b7d42 JA	358	for ((i) = 0; (i) < (q)->nr_hw_queues && \
0d0b7d42 JA	359	({ hctx = (q)->queue_hw_ctx[i]; 1; }); (i)++)
320ae51f	360
320ae51f	361	#define hctx_for_each_ctx(hctx, ctx, i) \
0d0b7d42 JA	362	for ((i) = 0; (i) < (hctx)->nr_ctx && \
0d0b7d42 JA	363	({ ctx = (hctx)->ctxs[(i)]; 1; }); (i)++)
320ae51f	364
320ae51f	365	#endif