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

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

#include "blk-stat.h"
#include "blk-mq-tag.h"

struct blk_mq_tag_set;

/**
 * struct blk_mq_ctx - State for a software queue facing the submitting CPUs
 */
struct blk_mq_ctx {
	struct {
		spinlock_t		lock;
		struct list_head	rq_list;
	}  ____cacheline_aligned_in_smp;

	unsigned int		cpu;
	unsigned int		index_hw;

	/* incremented at dispatch time */
	unsigned long		rq_dispatched[2];
	unsigned long		rq_merged;

	/* incremented at completion time */
	unsigned long		____cacheline_aligned_in_smp rq_completed[2];

	struct request_queue	*queue;
	struct kobject		kobj;
} ____cacheline_aligned_in_smp;

/*
 * Bits for request->gstate.  The lower two bits carry MQ_RQ_* state value
 * and the upper bits the generation number.
 */
enum mq_rq_state {
	MQ_RQ_IDLE		= 0,
	MQ_RQ_IN_FLIGHT		= 1,
	MQ_RQ_COMPLETE		= 2,

	MQ_RQ_STATE_BITS	= 2,
	MQ_RQ_STATE_MASK	= (1 << MQ_RQ_STATE_BITS) - 1,
	MQ_RQ_GEN_INC		= 1 << MQ_RQ_STATE_BITS,
};

void blk_mq_freeze_queue(struct request_queue *q);
void blk_mq_free_queue(struct request_queue *q);
int blk_mq_update_nr_requests(struct request_queue *q, unsigned int nr);
void blk_mq_wake_waiters(struct request_queue *q);
bool blk_mq_dispatch_rq_list(struct request_queue *, struct list_head *, bool);
void blk_mq_flush_busy_ctxs(struct blk_mq_hw_ctx *hctx, struct list_head *list);
bool blk_mq_get_driver_tag(struct request *rq, struct blk_mq_hw_ctx **hctx,
				bool wait);
struct request *blk_mq_dequeue_from_ctx(struct blk_mq_hw_ctx *hctx,
					struct blk_mq_ctx *start);

/*
 * Internal helpers for allocating/freeing the request map
 */
void blk_mq_free_rqs(struct blk_mq_tag_set *set, struct blk_mq_tags *tags,
		     unsigned int hctx_idx);
void blk_mq_free_rq_map(struct blk_mq_tags *tags);
struct blk_mq_tags *blk_mq_alloc_rq_map(struct blk_mq_tag_set *set,
					unsigned int hctx_idx,
					unsigned int nr_tags,
					unsigned int reserved_tags);
int blk_mq_alloc_rqs(struct blk_mq_tag_set *set, struct blk_mq_tags *tags,
		     unsigned int hctx_idx, unsigned int depth);

/*
 * Internal helpers for request insertion into sw queues
 */
void __blk_mq_insert_request(struct blk_mq_hw_ctx *hctx, struct request *rq,
				bool at_head);
void blk_mq_request_bypass_insert(struct request *rq, bool run_queue);
void blk_mq_insert_requests(struct blk_mq_hw_ctx *hctx, struct blk_mq_ctx *ctx,
				struct list_head *list);

/* Used by blk_insert_cloned_request() to issue request directly */
blk_status_t blk_mq_request_issue_directly(struct request *rq);

/*
 * CPU -> queue mappings
 */
extern int blk_mq_hw_queue_to_node(unsigned int *map, unsigned int);

static inline struct blk_mq_hw_ctx *blk_mq_map_queue(struct request_queue *q,
		int cpu)
{
	return q->queue_hw_ctx[q->mq_map[cpu]];
}

/*
 * sysfs helpers
 */
extern void blk_mq_sysfs_init(struct request_queue *q);
extern void blk_mq_sysfs_deinit(struct request_queue *q);
extern int __blk_mq_register_dev(struct device *dev, struct request_queue *q);
extern int blk_mq_sysfs_register(struct request_queue *q);
extern void blk_mq_sysfs_unregister(struct request_queue *q);
extern void blk_mq_hctx_kobj_init(struct blk_mq_hw_ctx *hctx);

void blk_mq_release(struct request_queue *q);

/**
 * blk_mq_rq_state() - read the current MQ_RQ_* state of a request
 * @rq: target request.
 */
static inline int blk_mq_rq_state(struct request *rq)
{
	return READ_ONCE(rq->gstate) & MQ_RQ_STATE_MASK;
}

/**
 * blk_mq_rq_update_state() - set the current MQ_RQ_* state of a request
 * @rq: target request.
 * @state: new state to set.
 *
 * Set @rq's state to @state.  The caller is responsible for ensuring that
 * there are no other updaters.  A request can transition into IN_FLIGHT
 * only from IDLE and doing so increments the generation number.
 */
static inline void blk_mq_rq_update_state(struct request *rq,
					  enum mq_rq_state state)
{
	u64 old_val = READ_ONCE(rq->gstate);
	u64 new_val = (old_val & ~MQ_RQ_STATE_MASK) | state;

	if (state == MQ_RQ_IN_FLIGHT) {
		WARN_ON_ONCE((old_val & MQ_RQ_STATE_MASK) != MQ_RQ_IDLE);
		new_val += MQ_RQ_GEN_INC;
	}

	/* avoid exposing interim values */
	WRITE_ONCE(rq->gstate, new_val);
}

static inline struct blk_mq_ctx *__blk_mq_get_ctx(struct request_queue *q,
					   unsigned int cpu)
{
	return per_cpu_ptr(q->queue_ctx, cpu);
}

/*
 * This assumes per-cpu software queueing queues. They could be per-node
 * as well, for instance. For now this is hardcoded as-is. Note that we don't
 * care about preemption, since we know the ctx's are persistent. This does
 * mean that we can't rely on ctx always matching the currently running CPU.
 */
static inline struct blk_mq_ctx *blk_mq_get_ctx(struct request_queue *q)
{
	return __blk_mq_get_ctx(q, get_cpu());
}

static inline void blk_mq_put_ctx(struct blk_mq_ctx *ctx)
{
	put_cpu();
}

struct blk_mq_alloc_data {
	/* input parameter */
	struct request_queue *q;
	blk_mq_req_flags_t flags;
	unsigned int shallow_depth;

	/* input & output parameter */
	struct blk_mq_ctx *ctx;
	struct blk_mq_hw_ctx *hctx;
};

static inline struct blk_mq_tags *blk_mq_tags_from_data(struct blk_mq_alloc_data *data)
{
	if (data->flags & BLK_MQ_REQ_INTERNAL)
		return data->hctx->sched_tags;

	return data->hctx->tags;
}

static inline bool blk_mq_hctx_stopped(struct blk_mq_hw_ctx *hctx)
{
	return test_bit(BLK_MQ_S_STOPPED, &hctx->state);
}

static inline bool blk_mq_hw_queue_mapped(struct blk_mq_hw_ctx *hctx)
{
	return hctx->nr_ctx && hctx->tags;
}

void blk_mq_in_flight(struct request_queue *q, struct hd_struct *part,
			unsigned int inflight[2]);

static inline void blk_mq_put_dispatch_budget(struct blk_mq_hw_ctx *hctx)
{
	struct request_queue *q = hctx->queue;

	if (q->mq_ops->put_budget)
		q->mq_ops->put_budget(hctx);
}

static inline bool blk_mq_get_dispatch_budget(struct blk_mq_hw_ctx *hctx)
{
	struct request_queue *q = hctx->queue;

	if (q->mq_ops->get_budget)
		return q->mq_ops->get_budget(hctx);
	return true;
}

static inline void __blk_mq_put_driver_tag(struct blk_mq_hw_ctx *hctx,
					   struct request *rq)
{
	blk_mq_put_tag(hctx, hctx->tags, rq->mq_ctx, rq->tag);
	rq->tag = -1;

	if (rq->rq_flags & RQF_MQ_INFLIGHT) {
		rq->rq_flags &= ~RQF_MQ_INFLIGHT;
		atomic_dec(&hctx->nr_active);
	}
}

static inline void blk_mq_put_driver_tag_hctx(struct blk_mq_hw_ctx *hctx,
				       struct request *rq)
{
	if (rq->tag == -1 || rq->internal_tag == -1)
		return;

	__blk_mq_put_driver_tag(hctx, rq);
}

static inline void blk_mq_put_driver_tag(struct request *rq)
{
	struct blk_mq_hw_ctx *hctx;

	if (rq->tag == -1 || rq->internal_tag == -1)
		return;

	hctx = blk_mq_map_queue(rq->q, rq->mq_ctx->cpu);
	__blk_mq_put_driver_tag(hctx, rq);
}

#endif
Commit	Line	Data
b2441318	1	/* SPDX-License-Identifier: GPL-2.0 */
320ae51f JA	2	#ifndef INT_BLK_MQ_H
	3	#define INT_BLK_MQ_H
	4
cf43e6be	5	#include "blk-stat.h"
244c65a3	6	#include "blk-mq-tag.h"
cf43e6be	7
24d2f903 CH	8	struct blk_mq_tag_set;
24d2f903 CH	9
fe644072 LW	10	/**
	11	* struct blk_mq_ctx - State for a software queue facing the submitting CPUs
	12	*/
320ae51f JA	13	struct blk_mq_ctx {
	14	struct {
	15	spinlock_t lock;
	16	struct list_head rq_list;
	17	} ____cacheline_aligned_in_smp;
	18
	19	unsigned int cpu;
	20	unsigned int index_hw;
320ae51f JA	21
	22	/* incremented at dispatch time */
	23	unsigned long rq_dispatched[2];
	24	unsigned long rq_merged;
	25
	26	/* incremented at completion time */
	27	unsigned long ____cacheline_aligned_in_smp rq_completed[2];
	28
	29	struct request_queue *queue;
	30	struct kobject kobj;
4bb659b1	31	} ____cacheline_aligned_in_smp;
320ae51f	32
1d9bd516 TH	33	/*
	34	* Bits for request->gstate. The lower two bits carry MQ_RQ_* state value
	35	* and the upper bits the generation number.
	36	*/
	37	enum mq_rq_state {
	38	MQ_RQ_IDLE = 0,
	39	MQ_RQ_IN_FLIGHT = 1,
5a61c363	40	MQ_RQ_COMPLETE = 2,
1d9bd516 TH	41
	42	MQ_RQ_STATE_BITS = 2,
	43	MQ_RQ_STATE_MASK = (1 << MQ_RQ_STATE_BITS) - 1,
	44	MQ_RQ_GEN_INC = 1 << MQ_RQ_STATE_BITS,
	45	};
	46
780db207	47	void blk_mq_freeze_queue(struct request_queue *q);
3edcc0ce	48	void blk_mq_free_queue(struct request_queue *q);
e3a2b3f9	49	int blk_mq_update_nr_requests(struct request_queue *q, unsigned int nr);
aed3ea94	50	void blk_mq_wake_waiters(struct request_queue *q);
de148297	51	bool blk_mq_dispatch_rq_list(struct request_queue , struct list_head , bool);
2c3ad667	52	void blk_mq_flush_busy_ctxs(struct blk_mq_hw_ctx hctx, struct list_head list);
bd6737f1 JA	53	bool blk_mq_get_driver_tag(struct request rq, struct blk_mq_hw_ctx *hctx,
bd6737f1 JA	54	bool wait);
b347689f ML	55	struct request blk_mq_dequeue_from_ctx(struct blk_mq_hw_ctx hctx,
b347689f ML	56	struct blk_mq_ctx *start);
2c3ad667 JA	57
	58	/*
	59	* Internal helpers for allocating/freeing the request map
	60	*/
cc71a6f4 JA	61	void blk_mq_free_rqs(struct blk_mq_tag_set set, struct blk_mq_tags tags,
	62	unsigned int hctx_idx);
	63	void blk_mq_free_rq_map(struct blk_mq_tags *tags);
	64	struct blk_mq_tags blk_mq_alloc_rq_map(struct blk_mq_tag_set set,
	65	unsigned int hctx_idx,
	66	unsigned int nr_tags,
	67	unsigned int reserved_tags);
	68	int blk_mq_alloc_rqs(struct blk_mq_tag_set set, struct blk_mq_tags tags,
	69	unsigned int hctx_idx, unsigned int depth);
2c3ad667 JA	70
	71	/*
	72	* Internal helpers for request insertion into sw queues
	73	*/
	74	void __blk_mq_insert_request(struct blk_mq_hw_ctx hctx, struct request rq,
	75	bool at_head);
b0850297	76	void blk_mq_request_bypass_insert(struct request *rq, bool run_queue);
bd166ef1 JA	77	void blk_mq_insert_requests(struct blk_mq_hw_ctx hctx, struct blk_mq_ctx ctx,
bd166ef1 JA	78	struct list_head *list);
320ae51f	79
396eaf21	80	/* Used by blk_insert_cloned_request() to issue request directly */
c77ff7fd	81	blk_status_t blk_mq_request_issue_directly(struct request *rq);
396eaf21	82
320ae51f JA	83	/*
	84	* CPU -> queue mappings
	85	*/
f14bbe77	86	extern int blk_mq_hw_queue_to_node(unsigned int *map, unsigned int);
320ae51f	87
7d7e0f90 CH	88	static inline struct blk_mq_hw_ctx blk_mq_map_queue(struct request_queue q,
	89	int cpu)
	90	{
	91	return q->queue_hw_ctx[q->mq_map[cpu]];
	92	}
	93
67aec14c JA	94	/*
	95	* sysfs helpers
	96	*/
737f98cf	97	extern void blk_mq_sysfs_init(struct request_queue *q);
7ea5fe31	98	extern void blk_mq_sysfs_deinit(struct request_queue *q);
2d0364c8	99	extern int __blk_mq_register_dev(struct device dev, struct request_queue q);
67aec14c JA	100	extern int blk_mq_sysfs_register(struct request_queue *q);
67aec14c JA	101	extern void blk_mq_sysfs_unregister(struct request_queue *q);
868f2f0b	102	extern void blk_mq_hctx_kobj_init(struct blk_mq_hw_ctx *hctx);
67aec14c	103
e09aae7e ML	104	void blk_mq_release(struct request_queue *q);
e09aae7e ML	105
1d9bd516 TH	106	/**
	107	* blk_mq_rq_state() - read the current MQ_RQ_* state of a request
	108	* @rq: target request.
	109	*/
	110	static inline int blk_mq_rq_state(struct request *rq)
	111	{
	112	return READ_ONCE(rq->gstate) & MQ_RQ_STATE_MASK;
	113	}
	114
	115	/**
	116	* blk_mq_rq_update_state() - set the current MQ_RQ_* state of a request
	117	* @rq: target request.
	118	* @state: new state to set.
	119	*
	120	* Set @rq's state to @state. The caller is responsible for ensuring that
	121	* there are no other updaters. A request can transition into IN_FLIGHT
	122	* only from IDLE and doing so increments the generation number.
	123	*/
	124	static inline void blk_mq_rq_update_state(struct request *rq,
	125	enum mq_rq_state state)
	126	{
	127	u64 old_val = READ_ONCE(rq->gstate);
	128	u64 new_val = (old_val & ~MQ_RQ_STATE_MASK) \| state;
	129
	130	if (state == MQ_RQ_IN_FLIGHT) {
	131	WARN_ON_ONCE((old_val & MQ_RQ_STATE_MASK) != MQ_RQ_IDLE);
	132	new_val += MQ_RQ_GEN_INC;
	133	}
	134
	135	/* avoid exposing interim values */
	136	WRITE_ONCE(rq->gstate, new_val);
	137	}
	138
1aecfe48 ML	139	static inline struct blk_mq_ctx __blk_mq_get_ctx(struct request_queue q,
	140	unsigned int cpu)
	141	{
	142	return per_cpu_ptr(q->queue_ctx, cpu);
	143	}
	144
	145	/*
	146	* This assumes per-cpu software queueing queues. They could be per-node
	147	* as well, for instance. For now this is hardcoded as-is. Note that we don't
	148	* care about preemption, since we know the ctx's are persistent. This does
	149	* mean that we can't rely on ctx always matching the currently running CPU.
	150	*/
	151	static inline struct blk_mq_ctx blk_mq_get_ctx(struct request_queue q)
	152	{
	153	return __blk_mq_get_ctx(q, get_cpu());
	154	}
	155
	156	static inline void blk_mq_put_ctx(struct blk_mq_ctx *ctx)
	157	{
	158	put_cpu();
	159	}
	160
cb96a42c ML	161	struct blk_mq_alloc_data {
	162	/* input parameter */
	163	struct request_queue *q;
9a95e4ef	164	blk_mq_req_flags_t flags;
229a9287	165	unsigned int shallow_depth;
cb96a42c ML	166
	167	/* input & output parameter */
	168	struct blk_mq_ctx *ctx;
	169	struct blk_mq_hw_ctx *hctx;
	170	};
	171
4941115b JA	172	static inline struct blk_mq_tags blk_mq_tags_from_data(struct blk_mq_alloc_data data)
4941115b JA	173	{
bd166ef1 JA	174	if (data->flags & BLK_MQ_REQ_INTERNAL)
	175	return data->hctx->sched_tags;
	176
4941115b JA	177	return data->hctx->tags;
	178	}
	179
5d1b25c1 BVA	180	static inline bool blk_mq_hctx_stopped(struct blk_mq_hw_ctx *hctx)
	181	{
	182	return test_bit(BLK_MQ_S_STOPPED, &hctx->state);
	183	}
	184
19c66e59 ML	185	static inline bool blk_mq_hw_queue_mapped(struct blk_mq_hw_ctx *hctx)
	186	{
	187	return hctx->nr_ctx && hctx->tags;
	188	}
	189
f299b7c7 JA	190	void blk_mq_in_flight(struct request_queue q, struct hd_struct part,
	191	unsigned int inflight[2]);
	192
de148297 ML	193	static inline void blk_mq_put_dispatch_budget(struct blk_mq_hw_ctx *hctx)
	194	{
	195	struct request_queue *q = hctx->queue;
	196
	197	if (q->mq_ops->put_budget)
	198	q->mq_ops->put_budget(hctx);
	199	}
	200
88022d72	201	static inline bool blk_mq_get_dispatch_budget(struct blk_mq_hw_ctx *hctx)
de148297 ML	202	{
	203	struct request_queue *q = hctx->queue;
	204
	205	if (q->mq_ops->get_budget)
	206	return q->mq_ops->get_budget(hctx);
88022d72	207	return true;
de148297 ML	208	}
de148297 ML	209
244c65a3 ML	210	static inline void __blk_mq_put_driver_tag(struct blk_mq_hw_ctx *hctx,
	211	struct request *rq)
	212	{
	213	blk_mq_put_tag(hctx, hctx->tags, rq->mq_ctx, rq->tag);
	214	rq->tag = -1;
	215
	216	if (rq->rq_flags & RQF_MQ_INFLIGHT) {
	217	rq->rq_flags &= ~RQF_MQ_INFLIGHT;
	218	atomic_dec(&hctx->nr_active);
	219	}
	220	}
	221
	222	static inline void blk_mq_put_driver_tag_hctx(struct blk_mq_hw_ctx *hctx,
	223	struct request *rq)
	224	{
	225	if (rq->tag == -1 \|\| rq->internal_tag == -1)
	226	return;
	227
	228	__blk_mq_put_driver_tag(hctx, rq);
	229	}
	230
	231	static inline void blk_mq_put_driver_tag(struct request *rq)
	232	{
	233	struct blk_mq_hw_ctx *hctx;
	234
	235	if (rq->tag == -1 \|\| rq->internal_tag == -1)
	236	return;
	237
	238	hctx = blk_mq_map_queue(rq->q, rq->mq_ctx->cpu);
	239	__blk_mq_put_driver_tag(hctx, rq);
	240	}
	241
320ae51f	242	#endif