[deliverable/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_mq_cpu_notifier {
	struct list_head list;
	void *data;
	void (*notify)(void *data, unsigned long action, unsigned int cpu);
};

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;
	unsigned int		queue_num;

	void			*driver_data;

	unsigned int 		nr_ctx_map;
	unsigned long		*ctx_map;
	unsigned int		nr_ctx;
	struct blk_mq_ctx	**ctxs;

	unsigned int		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		cmd_size;	/* per-request extra data */

	struct blk_mq_cpu_notifier	cpu_notifier;
	struct kobject		kobj;
};

struct blk_mq_tag_set {
	struct blk_mq_ops	*ops;
	unsigned int		nr_hw_queues;
	unsigned int		queue_depth;
	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;
};

typedef int (queue_rq_fn)(struct blk_mq_hw_ctx *, struct request *);
typedef struct blk_mq_hw_ctx *(map_queue_fn)(struct request_queue *, const int);
typedef struct blk_mq_hw_ctx *(alloc_hctx_fn)(struct blk_mq_tag_set *,
		unsigned int);
typedef void (free_hctx_fn)(struct blk_mq_hw_ctx *, unsigned int);
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);

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
	 */
	rq_timed_out_fn		*timeout;

	softirq_done_fn		*complete;

	/*
	 * Override for hctx allocations (should probably go)
	 */
	alloc_hctx_fn		*alloc_hctx;
	free_hctx_fn		*free_hctx;

	/*
	 * 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.
	 * 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_SHOULD_SORT	= 1 << 1,

	BLK_MQ_S_STOPPED	= 0,

	BLK_MQ_MAX_DEPTH	= 2048,

	BLK_MQ_CPU_WORK_BATCH	= 8,
};

struct request_queue *blk_mq_init_queue(struct blk_mq_tag_set *);
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_run_queues(struct request_queue *q, bool async);
void blk_mq_free_request(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);
struct request *blk_mq_alloc_reserved_request(struct request_queue *q, int rw, gfp_t gfp);
struct request *blk_mq_tag_to_rq(struct blk_mq_tags *tags, unsigned int tag);

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);
void blk_mq_free_single_hw_queue(struct blk_mq_hw_ctx *, unsigned int);

void blk_mq_end_io(struct request *rq, int error);
void __blk_mq_end_io(struct request *rq, int error);

void blk_mq_requeue_request(struct request *rq);

void blk_mq_complete_request(struct request *rq);

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_delay_queue(struct blk_mq_hw_ctx *hctx, unsigned long msecs);

/*
 * Driver command data is immediately after the request. So subtract request
 * size to get back to the original request.
 */
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 (void *) rq + sizeof(*rq);
}

#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;
	7
	8	struct blk_mq_cpu_notifier {
	9	struct list_head list;
	10	void *data;
	11	void (notify)(void data, unsigned long action, unsigned int cpu);
	12	};
	13
	14	struct blk_mq_hw_ctx {
	15	struct {
	16	spinlock_t lock;
	17	struct list_head dispatch;
	18	} ____cacheline_aligned_in_smp;
	19
	20	unsigned long state; /* BLK_MQ_S_* flags */
70f4db63 CH	21	struct delayed_work run_work;
70f4db63 CH	22	struct delayed_work delay_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
	29	struct request_queue *queue;
	30	unsigned int queue_num;
	31
	32	void *driver_data;
	33
320ae51f JA	34	unsigned int nr_ctx_map;
320ae51f JA	35	unsigned long *ctx_map;
4bb659b1 JA	36	unsigned int nr_ctx;
	37	struct blk_mq_ctx **ctxs;
	38
	39	unsigned int wait_index;
320ae51f	40
320ae51f JA	41	struct blk_mq_tags *tags;
	42
	43	unsigned long queued;
	44	unsigned long run;
	45	#define BLK_MQ_MAX_DISPATCH_ORDER 10
	46	unsigned long dispatched[BLK_MQ_MAX_DISPATCH_ORDER];
	47
320ae51f JA	48	unsigned int numa_node;
	49	unsigned int cmd_size; /* per-request extra data */
	50
	51	struct blk_mq_cpu_notifier cpu_notifier;
	52	struct kobject kobj;
	53	};
	54
24d2f903	55	struct blk_mq_tag_set {
320ae51f JA	56	struct blk_mq_ops *ops;
	57	unsigned int nr_hw_queues;
	58	unsigned int queue_depth;
	59	unsigned int reserved_tags;
	60	unsigned int cmd_size; /* per-request extra data */
	61	int numa_node;
	62	unsigned int timeout;
	63	unsigned int flags; /* BLK_MQ_F_* */
24d2f903 CH	64	void *driver_data;
	65
	66	struct blk_mq_tags **tags;
320ae51f JA	67	};
	68
	69	typedef int (queue_rq_fn)(struct blk_mq_hw_ctx , struct request );
	70	typedef struct blk_mq_hw_ctx (map_queue_fn)(struct request_queue , const int);
24d2f903 CH	71	typedef struct blk_mq_hw_ctx (alloc_hctx_fn)(struct blk_mq_tag_set ,
24d2f903 CH	72	unsigned int);
320ae51f JA	73	typedef void (free_hctx_fn)(struct blk_mq_hw_ctx *, unsigned int);
	74	typedef int (init_hctx_fn)(struct blk_mq_hw_ctx , void , unsigned int);
	75	typedef void (exit_hctx_fn)(struct blk_mq_hw_ctx *, unsigned int);
24d2f903 CH	76	typedef int (init_request_fn)(void , struct request , unsigned int,
	77	unsigned int, unsigned int);
	78	typedef void (exit_request_fn)(void , struct request , unsigned int,
	79	unsigned int);
320ae51f JA	80
	81	struct blk_mq_ops {
	82	/*
	83	* Queue request
	84	*/
	85	queue_rq_fn *queue_rq;
	86
	87	/*
	88	* Map to specific hardware queue
	89	*/
	90	map_queue_fn *map_queue;
	91
	92	/*
	93	* Called on request timeout
	94	*/
	95	rq_timed_out_fn *timeout;
	96
30a91cb4 CH	97	softirq_done_fn *complete;
30a91cb4 CH	98
320ae51f JA	99	/*
	100	* Override for hctx allocations (should probably go)
	101	*/
	102	alloc_hctx_fn *alloc_hctx;
	103	free_hctx_fn *free_hctx;
	104
	105	/*
	106	* Called when the block layer side of a hardware queue has been
	107	* set up, allowing the driver to allocate/init matching structures.
	108	* Ditto for exit/teardown.
	109	*/
	110	init_hctx_fn *init_hctx;
	111	exit_hctx_fn *exit_hctx;
e9b267d9 CH	112
	113	/*
	114	* Called for every command allocated by the block layer to allow
	115	* the driver to set up driver specific data.
	116	* Ditto for exit/teardown.
	117	*/
	118	init_request_fn *init_request;
	119	exit_request_fn *exit_request;
320ae51f JA	120	};
	121
	122	enum {
	123	BLK_MQ_RQ_QUEUE_OK = 0, /* queued fine */
	124	BLK_MQ_RQ_QUEUE_BUSY = 1, /* requeue IO for later */
	125	BLK_MQ_RQ_QUEUE_ERROR = 2, /* end IO with error */
	126
	127	BLK_MQ_F_SHOULD_MERGE = 1 << 0,
	128	BLK_MQ_F_SHOULD_SORT = 1 << 1,
320ae51f	129
5d12f905	130	BLK_MQ_S_STOPPED = 0,
320ae51f JA	131
320ae51f JA	132	BLK_MQ_MAX_DEPTH = 2048,
506e931f JA	133
506e931f JA	134	BLK_MQ_CPU_WORK_BATCH = 8,
320ae51f JA	135	};
320ae51f JA	136
24d2f903	137	struct request_queue blk_mq_init_queue(struct blk_mq_tag_set );
320ae51f JA	138	int blk_mq_register_disk(struct gendisk *);
320ae51f JA	139	void blk_mq_unregister_disk(struct gendisk *);
320ae51f	140
24d2f903 CH	141	int blk_mq_alloc_tag_set(struct blk_mq_tag_set *set);
	142	void blk_mq_free_tag_set(struct blk_mq_tag_set *set);
	143
320ae51f JA	144	void blk_mq_flush_plug_list(struct blk_plug *plug, bool from_schedule);
320ae51f JA	145
feb71dae	146	void blk_mq_insert_request(struct request *, bool, bool, bool);
320ae51f JA	147	void blk_mq_run_queues(struct request_queue *q, bool async);
	148	void blk_mq_free_request(struct request *rq);
	149	bool blk_mq_can_queue(struct blk_mq_hw_ctx *);
18741986	150	struct request blk_mq_alloc_request(struct request_queue q, int rw, gfp_t gfp);
320ae51f	151	struct request blk_mq_alloc_reserved_request(struct request_queue q, int rw, gfp_t gfp);
24d2f903	152	struct request blk_mq_tag_to_rq(struct blk_mq_tags tags, unsigned int tag);
320ae51f JA	153
320ae51f JA	154	struct blk_mq_hw_ctx blk_mq_map_queue(struct request_queue , const int ctx_index);
24d2f903	155	struct blk_mq_hw_ctx blk_mq_alloc_single_hw_queue(struct blk_mq_tag_set , unsigned int);
320ae51f JA	156	void blk_mq_free_single_hw_queue(struct blk_mq_hw_ctx *, unsigned int);
320ae51f JA	157
63151a44 CH	158	void blk_mq_end_io(struct request *rq, int error);
63151a44 CH	159	void __blk_mq_end_io(struct request *rq, int error);
320ae51f	160
ed0791b2 CH	161	void blk_mq_requeue_request(struct request *rq);
ed0791b2 CH	162
30a91cb4 CH	163	void blk_mq_complete_request(struct request *rq);
30a91cb4 CH	164
320ae51f JA	165	void blk_mq_stop_hw_queue(struct blk_mq_hw_ctx *hctx);
320ae51f JA	166	void blk_mq_start_hw_queue(struct blk_mq_hw_ctx *hctx);
280d45f6	167	void blk_mq_stop_hw_queues(struct request_queue *q);
2f268556	168	void blk_mq_start_hw_queues(struct request_queue *q);
1b4a3258	169	void blk_mq_start_stopped_hw_queues(struct request_queue *q, bool async);
70f4db63	170	void blk_mq_delay_queue(struct blk_mq_hw_ctx *hctx, unsigned long msecs);
320ae51f JA	171
	172	/*
	173	* Driver command data is immediately after the request. So subtract request
	174	* size to get back to the original request.
	175	*/
	176	static inline struct request blk_mq_rq_from_pdu(void pdu)
	177	{
	178	return pdu - sizeof(struct request);
	179	}
	180	static inline void blk_mq_rq_to_pdu(struct request rq)
	181	{
	182	return (void ) rq + sizeof(rq);
	183	}
	184
320ae51f	185	#define queue_for_each_hw_ctx(q, hctx, i) \
0d0b7d42 JA	186	for ((i) = 0; (i) < (q)->nr_hw_queues && \
0d0b7d42 JA	187	({ hctx = (q)->queue_hw_ctx[i]; 1; }); (i)++)
320ae51f JA	188
320ae51f JA	189	#define queue_for_each_ctx(q, ctx, i) \
0d0b7d42 JA	190	for ((i) = 0; (i) < (q)->nr_queues && \
0d0b7d42 JA	191	({ ctx = per_cpu_ptr((q)->queue_ctx, (i)); 1; }); (i)++)
320ae51f JA	192
320ae51f JA	193	#define hctx_for_each_ctx(hctx, ctx, i) \
0d0b7d42 JA	194	for ((i) = 0; (i) < (hctx)->nr_ctx && \
0d0b7d42 JA	195	({ ctx = (hctx)->ctxs[(i)]; 1; }); (i)++)
320ae51f JA	196
	197	#define blk_ctx_sum(q, sum) \
	198	({ \
	199	struct blk_mq_ctx *__x; \
	200	unsigned int __ret = 0, __i; \
	201	\
	202	queue_for_each_ctx((q), __x, __i) \
	203	__ret += sum; \
	204	__ret; \
	205	})
	206
	207	#endif