[mirror_ubuntu-artful-kernel.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	delayed_work;
	cpumask_var_t		cpumask;

	unsigned long		flags;		/* BLK_MQ_F_* flags */

	struct request_queue	*queue;
	unsigned int		queue_num;

	void			*driver_data;

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

	struct request		**rqs;
	struct list_head	page_list;
	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		queue_depth;
	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_reg {
	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_* */
};

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_reg *,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 blk_mq_hw_ctx *,
		struct request *, unsigned int);
typedef void (exit_request_fn)(void *, struct blk_mq_hw_ctx *,
		struct request *, 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_F_SHOULD_IPI	= 1 << 2,

	BLK_MQ_S_STOPPED	= 0,

	BLK_MQ_MAX_DEPTH	= 2048,
};

struct request_queue *blk_mq_init_queue(struct blk_mq_reg *, void *);
int blk_mq_register_disk(struct gendisk *);
void blk_mq_unregister_disk(struct gendisk *);

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_rq_from_tag(struct request_queue *q, 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_reg *, unsigned int);
void blk_mq_free_single_hw_queue(struct blk_mq_hw_ctx *, unsigned int);

bool blk_mq_end_io_partial(struct request *rq, int error,
		unsigned int nr_bytes);
static inline void blk_mq_end_io(struct request *rq, int error)
{
	bool done = !blk_mq_end_io_partial(rq, error, blk_rq_bytes(rq));
	BUG_ON(!done);
}

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_stopped_hw_queues(struct request_queue *q);

/*
 * 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);
}

static inline struct request *blk_mq_tag_to_rq(struct blk_mq_hw_ctx *hctx,
					       unsigned int tag)
{
	return hctx->rqs[tag];
}

#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 */
	21	struct delayed_work delayed_work;
e4043dcf	22	cpumask_var_t cpumask;
320ae51f JA	23
	24	unsigned long flags; /* BLK_MQ_F_* flags */
	25
	26	struct request_queue *queue;
	27	unsigned int queue_num;
	28
	29	void *driver_data;
	30
	31	unsigned int nr_ctx;
	32	struct blk_mq_ctx **ctxs;
	33	unsigned int nr_ctx_map;
	34	unsigned long *ctx_map;
	35
	36	struct request **rqs;
	37	struct list_head page_list;
	38	struct blk_mq_tags *tags;
	39
	40	unsigned long queued;
	41	unsigned long run;
	42	#define BLK_MQ_MAX_DISPATCH_ORDER 10
	43	unsigned long dispatched[BLK_MQ_MAX_DISPATCH_ORDER];
	44
	45	unsigned int queue_depth;
	46	unsigned int numa_node;
	47	unsigned int cmd_size; /* per-request extra data */
	48
	49	struct blk_mq_cpu_notifier cpu_notifier;
	50	struct kobject kobj;
	51	};
	52
	53	struct blk_mq_reg {
	54	struct blk_mq_ops *ops;
	55	unsigned int nr_hw_queues;
	56	unsigned int queue_depth;
	57	unsigned int reserved_tags;
	58	unsigned int cmd_size; /* per-request extra data */
	59	int numa_node;
	60	unsigned int timeout;
	61	unsigned int flags; /* BLK_MQ_F_* */
	62	};
	63
	64	typedef int (queue_rq_fn)(struct blk_mq_hw_ctx , struct request );
	65	typedef struct blk_mq_hw_ctx (map_queue_fn)(struct request_queue , const int);
	66	typedef struct blk_mq_hw_ctx (alloc_hctx_fn)(struct blk_mq_reg ,unsigned int);
	67	typedef void (free_hctx_fn)(struct blk_mq_hw_ctx *, unsigned int);
	68	typedef int (init_hctx_fn)(struct blk_mq_hw_ctx , void , unsigned int);
	69	typedef void (exit_hctx_fn)(struct blk_mq_hw_ctx *, unsigned int);
e9b267d9 CH	70	typedef int (init_request_fn)(void , struct blk_mq_hw_ctx ,
	71	struct request *, unsigned int);
	72	typedef void (exit_request_fn)(void , struct blk_mq_hw_ctx ,
	73	struct request *, unsigned int);
320ae51f JA	74
	75	struct blk_mq_ops {
	76	/*
	77	* Queue request
	78	*/
	79	queue_rq_fn *queue_rq;
	80
	81	/*
	82	* Map to specific hardware queue
	83	*/
	84	map_queue_fn *map_queue;
	85
	86	/*
	87	* Called on request timeout
	88	*/
	89	rq_timed_out_fn *timeout;
	90
30a91cb4 CH	91	softirq_done_fn *complete;
30a91cb4 CH	92
320ae51f JA	93	/*
	94	* Override for hctx allocations (should probably go)
	95	*/
	96	alloc_hctx_fn *alloc_hctx;
	97	free_hctx_fn *free_hctx;
	98
	99	/*
	100	* Called when the block layer side of a hardware queue has been
	101	* set up, allowing the driver to allocate/init matching structures.
	102	* Ditto for exit/teardown.
	103	*/
	104	init_hctx_fn *init_hctx;
	105	exit_hctx_fn *exit_hctx;
e9b267d9 CH	106
	107	/*
	108	* Called for every command allocated by the block layer to allow
	109	* the driver to set up driver specific data.
	110	* Ditto for exit/teardown.
	111	*/
	112	init_request_fn *init_request;
	113	exit_request_fn *exit_request;
320ae51f JA	114	};
	115
	116	enum {
	117	BLK_MQ_RQ_QUEUE_OK = 0, /* queued fine */
	118	BLK_MQ_RQ_QUEUE_BUSY = 1, /* requeue IO for later */
	119	BLK_MQ_RQ_QUEUE_ERROR = 2, /* end IO with error */
	120
	121	BLK_MQ_F_SHOULD_MERGE = 1 << 0,
	122	BLK_MQ_F_SHOULD_SORT = 1 << 1,
	123	BLK_MQ_F_SHOULD_IPI = 1 << 2,
	124
5d12f905	125	BLK_MQ_S_STOPPED = 0,
320ae51f JA	126
	127	BLK_MQ_MAX_DEPTH = 2048,
	128	};
	129
	130	struct request_queue blk_mq_init_queue(struct blk_mq_reg , void *);
320ae51f JA	131	int blk_mq_register_disk(struct gendisk *);
320ae51f JA	132	void blk_mq_unregister_disk(struct gendisk *);
320ae51f JA	133
	134	void blk_mq_flush_plug_list(struct blk_plug *plug, bool from_schedule);
	135
feb71dae	136	void blk_mq_insert_request(struct request *, bool, bool, bool);
320ae51f JA	137	void blk_mq_run_queues(struct request_queue *q, bool async);
	138	void blk_mq_free_request(struct request *rq);
	139	bool blk_mq_can_queue(struct blk_mq_hw_ctx *);
18741986	140	struct request blk_mq_alloc_request(struct request_queue q, int rw, gfp_t gfp);
320ae51f JA	141	struct request blk_mq_alloc_reserved_request(struct request_queue q, int rw, gfp_t gfp);
	142	struct request blk_mq_rq_from_tag(struct request_queue q, unsigned int tag);
	143
	144	struct blk_mq_hw_ctx blk_mq_map_queue(struct request_queue , const int ctx_index);
	145	struct blk_mq_hw_ctx blk_mq_alloc_single_hw_queue(struct blk_mq_reg , unsigned int);
	146	void blk_mq_free_single_hw_queue(struct blk_mq_hw_ctx *, unsigned int);
	147
d6a25b31 CH	148	bool blk_mq_end_io_partial(struct request *rq, int error,
	149	unsigned int nr_bytes);
	150	static inline void blk_mq_end_io(struct request *rq, int error)
	151	{
	152	bool done = !blk_mq_end_io_partial(rq, error, blk_rq_bytes(rq));
	153	BUG_ON(!done);
	154	}
320ae51f	155
30a91cb4 CH	156	void blk_mq_complete_request(struct request *rq);
30a91cb4 CH	157
320ae51f JA	158	void blk_mq_stop_hw_queue(struct blk_mq_hw_ctx *hctx);
320ae51f JA	159	void blk_mq_start_hw_queue(struct blk_mq_hw_ctx *hctx);
280d45f6	160	void blk_mq_stop_hw_queues(struct request_queue *q);
320ae51f JA	161	void blk_mq_start_stopped_hw_queues(struct request_queue *q);
	162
	163	/*
	164	* Driver command data is immediately after the request. So subtract request
	165	* size to get back to the original request.
	166	*/
	167	static inline struct request blk_mq_rq_from_pdu(void pdu)
	168	{
	169	return pdu - sizeof(struct request);
	170	}
	171	static inline void blk_mq_rq_to_pdu(struct request rq)
	172	{
	173	return (void ) rq + sizeof(rq);
	174	}
	175
	176	static inline struct request blk_mq_tag_to_rq(struct blk_mq_hw_ctx hctx,
	177	unsigned int tag)
	178	{
	179	return hctx->rqs[tag];
	180	}
	181
	182	#define queue_for_each_hw_ctx(q, hctx, i) \
0d0b7d42 JA	183	for ((i) = 0; (i) < (q)->nr_hw_queues && \
0d0b7d42 JA	184	({ hctx = (q)->queue_hw_ctx[i]; 1; }); (i)++)
320ae51f JA	185
320ae51f JA	186	#define queue_for_each_ctx(q, ctx, i) \
0d0b7d42 JA	187	for ((i) = 0; (i) < (q)->nr_queues && \
0d0b7d42 JA	188	({ ctx = per_cpu_ptr((q)->queue_ctx, (i)); 1; }); (i)++)
320ae51f JA	189
320ae51f JA	190	#define hctx_for_each_ctx(hctx, ctx, i) \
0d0b7d42 JA	191	for ((i) = 0; (i) < (hctx)->nr_ctx && \
0d0b7d42 JA	192	({ ctx = (hctx)->ctxs[(i)]; 1; }); (i)++)
320ae51f JA	193
	194	#define blk_ctx_sum(q, sum) \
	195	({ \
	196	struct blk_mq_ctx *__x; \
	197	unsigned int __ret = 0, __i; \
	198	\
	199	queue_for_each_ctx((q), __x, __i) \
	200	__ret += sum; \
	201	__ret; \
	202	})
	203
	204	#endif