[linux-2.6-block.git] / block / deadline-iosched.c

/*
 *  Deadline i/o scheduler.
 *
 *  Copyright (C) 2002 Jens Axboe <axboe@kernel.dk>
 */
#include <linux/kernel.h>
#include <linux/fs.h>
#include <linux/blkdev.h>
#include <linux/elevator.h>
#include <linux/bio.h>
#include <linux/module.h>
#include <linux/slab.h>
#include <linux/init.h>
#include <linux/compiler.h>
#include <linux/rbtree.h>

/*
 * See Documentation/block/deadline-iosched.txt
 */
static const int read_expire = HZ / 2;  /* max time before a read is submitted. */
static const int write_expire = 5 * HZ; /* ditto for writes, these limits are SOFT! */
static const int writes_starved = 2;    /* max times reads can starve a write */
static const int fifo_batch = 16;       /* # of sequential requests treated as one
				     by the above parameters. For throughput. */

struct deadline_data {
	/*
	 * run time data
	 */

	/*
	 * requests (deadline_rq s) are present on both sort_list and fifo_list
	 */
	struct rb_root sort_list[2];	
	struct list_head fifo_list[2];

	/*
	 * next in sort order. read, write or both are NULL
	 */
	struct request *next_rq[2];
	unsigned int batching;		/* number of sequential requests made */
	unsigned int starved;		/* times reads have starved writes */

	/*
	 * settings that change how the i/o scheduler behaves
	 */
	int fifo_expire[2];
	int fifo_batch;
	int writes_starved;
	int front_merges;
};

static inline struct rb_root *
deadline_rb_root(struct deadline_data *dd, struct request *rq)
{
	return &dd->sort_list[rq_data_dir(rq)];
}

/*
 * get the request after `rq' in sector-sorted order
 */
static inline struct request *
deadline_latter_request(struct request *rq)
{
	struct rb_node *node = rb_next(&rq->rb_node);

	if (node)
		return rb_entry_rq(node);

	return NULL;
}

static void
deadline_add_rq_rb(struct deadline_data *dd, struct request *rq)
{
	struct rb_root *root = deadline_rb_root(dd, rq);

	elv_rb_add(root, rq);
}

static inline void
deadline_del_rq_rb(struct deadline_data *dd, struct request *rq)
{
	const int data_dir = rq_data_dir(rq);

	if (dd->next_rq[data_dir] == rq)
		dd->next_rq[data_dir] = deadline_latter_request(rq);

	elv_rb_del(deadline_rb_root(dd, rq), rq);
}

/*
 * add rq to rbtree and fifo
 */
static void
deadline_add_request(struct request_queue *q, struct request *rq)
{
	struct deadline_data *dd = q->elevator->elevator_data;
	const int data_dir = rq_data_dir(rq);

	/*
	 * This may be a requeue of a write request that has locked its
	 * target zone. If it is the case, this releases the zone lock.
	 */
	blk_req_zone_write_unlock(rq);

	deadline_add_rq_rb(dd, rq);

	/*
	 * set expire time and add to fifo list
	 */
	rq->fifo_time = jiffies + dd->fifo_expire[data_dir];
	list_add_tail(&rq->queuelist, &dd->fifo_list[data_dir]);
}

/*
 * remove rq from rbtree and fifo.
 */
static void deadline_remove_request(struct request_queue *q, struct request *rq)
{
	struct deadline_data *dd = q->elevator->elevator_data;

	rq_fifo_clear(rq);
	deadline_del_rq_rb(dd, rq);
}

static enum elv_merge
deadline_merge(struct request_queue *q, struct request **req, struct bio *bio)
{
	struct deadline_data *dd = q->elevator->elevator_data;
	struct request *__rq;

	/*
	 * check for front merge
	 */
	if (dd->front_merges) {
		sector_t sector = bio_end_sector(bio);

		__rq = elv_rb_find(&dd->sort_list[bio_data_dir(bio)], sector);
		if (__rq) {
			BUG_ON(sector != blk_rq_pos(__rq));

			if (elv_bio_merge_ok(__rq, bio)) {
				*req = __rq;
				return ELEVATOR_FRONT_MERGE;
			}
		}
	}

	return ELEVATOR_NO_MERGE;
}

static void deadline_merged_request(struct request_queue *q,
				    struct request *req, enum elv_merge type)
{
	struct deadline_data *dd = q->elevator->elevator_data;

	/*
	 * if the merge was a front merge, we need to reposition request
	 */
	if (type == ELEVATOR_FRONT_MERGE) {
		elv_rb_del(deadline_rb_root(dd, req), req);
		deadline_add_rq_rb(dd, req);
	}
}

static void
deadline_merged_requests(struct request_queue *q, struct request *req,
			 struct request *next)
{
	/*
	 * if next expires before rq, assign its expire time to rq
	 * and move into next position (next will be deleted) in fifo
	 */
	if (!list_empty(&req->queuelist) && !list_empty(&next->queuelist)) {
		if (time_before((unsigned long)next->fifo_time,
				(unsigned long)req->fifo_time)) {
			list_move(&req->queuelist, &next->queuelist);
			req->fifo_time = next->fifo_time;
		}
	}

	/*
	 * kill knowledge of next, this one is a goner
	 */
	deadline_remove_request(q, next);
}

/*
 * move request from sort list to dispatch queue.
 */
static inline void
deadline_move_to_dispatch(struct deadline_data *dd, struct request *rq)
{
	struct request_queue *q = rq->q;

	/*
	 * For a zoned block device, write requests must write lock their
	 * target zone.
	 */
	blk_req_zone_write_lock(rq);

	deadline_remove_request(q, rq);
	elv_dispatch_add_tail(q, rq);
}

/*
 * move an entry to dispatch queue
 */
static void
deadline_move_request(struct deadline_data *dd, struct request *rq)
{
	const int data_dir = rq_data_dir(rq);

	dd->next_rq[READ] = NULL;
	dd->next_rq[WRITE] = NULL;
	dd->next_rq[data_dir] = deadline_latter_request(rq);

	/*
	 * take it off the sort and fifo list, move
	 * to dispatch queue
	 */
	deadline_move_to_dispatch(dd, rq);
}

/*
 * deadline_check_fifo returns 0 if there are no expired requests on the fifo,
 * 1 otherwise. Requires !list_empty(&dd->fifo_list[data_dir])
 */
static inline int deadline_check_fifo(struct deadline_data *dd, int ddir)
{
	struct request *rq = rq_entry_fifo(dd->fifo_list[ddir].next);

	/*
	 * rq is expired!
	 */
	if (time_after_eq(jiffies, (unsigned long)rq->fifo_time))
		return 1;

	return 0;
}

/*
 * For the specified data direction, return the next request to dispatch using
 * arrival ordered lists.
 */
static struct request *
deadline_fifo_request(struct deadline_data *dd, int data_dir)
{
	struct request *rq;

	if (WARN_ON_ONCE(data_dir != READ && data_dir != WRITE))
		return NULL;

	if (list_empty(&dd->fifo_list[data_dir]))
		return NULL;

	rq = rq_entry_fifo(dd->fifo_list[data_dir].next);
	if (data_dir == READ || !blk_queue_is_zoned(rq->q))
		return rq;

	/*
	 * Look for a write request that can be dispatched, that is one with
	 * an unlocked target zone.
	 */
	list_for_each_entry(rq, &dd->fifo_list[WRITE], queuelist) {
		if (blk_req_can_dispatch_to_zone(rq))
			return rq;
	}

	return NULL;
}

/*
 * For the specified data direction, return the next request to dispatch using
 * sector position sorted lists.
 */
static struct request *
deadline_next_request(struct deadline_data *dd, int data_dir)
{
	struct request *rq;

	if (WARN_ON_ONCE(data_dir != READ && data_dir != WRITE))
		return NULL;

	rq = dd->next_rq[data_dir];
	if (!rq)
		return NULL;

	if (data_dir == READ || !blk_queue_is_zoned(rq->q))
		return rq;

	/*
	 * Look for a write request that can be dispatched, that is one with
	 * an unlocked target zone.
	 */
	while (rq) {
		if (blk_req_can_dispatch_to_zone(rq))
			return rq;
		rq = deadline_latter_request(rq);
	}

	return NULL;
}

/*
 * deadline_dispatch_requests selects the best request according to
 * read/write expire, fifo_batch, etc
 */
static int deadline_dispatch_requests(struct request_queue *q, int force)
{
	struct deadline_data *dd = q->elevator->elevator_data;
	const int reads = !list_empty(&dd->fifo_list[READ]);
	const int writes = !list_empty(&dd->fifo_list[WRITE]);
	struct request *rq, *next_rq;
	int data_dir;

	/*
	 * batches are currently reads XOR writes
	 */
	rq = deadline_next_request(dd, WRITE);
	if (!rq)
		rq = deadline_next_request(dd, READ);

	if (rq && dd->batching < dd->fifo_batch)
		/* we have a next request are still entitled to batch */
		goto dispatch_request;

	/*
	 * at this point we are not running a batch. select the appropriate
	 * data direction (read / write)
	 */

	if (reads) {
		BUG_ON(RB_EMPTY_ROOT(&dd->sort_list[READ]));

		if (deadline_fifo_request(dd, WRITE) &&
		    (dd->starved++ >= dd->writes_starved))
			goto dispatch_writes;

		data_dir = READ;

		goto dispatch_find_request;
	}

	/*
	 * there are either no reads or writes have been starved
	 */

	if (writes) {
dispatch_writes:
		BUG_ON(RB_EMPTY_ROOT(&dd->sort_list[WRITE]));

		dd->starved = 0;

		data_dir = WRITE;

		goto dispatch_find_request;
	}

	return 0;

dispatch_find_request:
	/*
	 * we are not running a batch, find best request for selected data_dir
	 */
	next_rq = deadline_next_request(dd, data_dir);
	if (deadline_check_fifo(dd, data_dir) || !next_rq) {
		/*
		 * A deadline has expired, the last request was in the other
		 * direction, or we have run out of higher-sectored requests.
		 * Start again from the request with the earliest expiry time.
		 */
		rq = deadline_fifo_request(dd, data_dir);
	} else {
		/*
		 * The last req was the same dir and we have a next request in
		 * sort order. No expired requests so continue on from here.
		 */
		rq = next_rq;
	}

	/*
	 * For a zoned block device, if we only have writes queued and none of
	 * them can be dispatched, rq will be NULL.
	 */
	if (!rq)
		return 0;

	dd->batching = 0;

dispatch_request:
	/*
	 * rq is the selected appropriate request.
	 */
	dd->batching++;
	deadline_move_request(dd, rq);

	return 1;
}

/*
 * For zoned block devices, write unlock the target zone of completed
 * write requests.
 */
static void
deadline_completed_request(struct request_queue *q, struct request *rq)
{
	blk_req_zone_write_unlock(rq);
}

static void deadline_exit_queue(struct elevator_queue *e)
{
	struct deadline_data *dd = e->elevator_data;

	BUG_ON(!list_empty(&dd->fifo_list[READ]));
	BUG_ON(!list_empty(&dd->fifo_list[WRITE]));

	kfree(dd);
}

/*
 * initialize elevator private data (deadline_data).
 */
static int deadline_init_queue(struct request_queue *q, struct elevator_type *e)
{
	struct deadline_data *dd;
	struct elevator_queue *eq;

	eq = elevator_alloc(q, e);
	if (!eq)
		return -ENOMEM;

	dd = kzalloc_node(sizeof(*dd), GFP_KERNEL, q->node);
	if (!dd) {
		kobject_put(&eq->kobj);
		return -ENOMEM;
	}
	eq->elevator_data = dd;

	INIT_LIST_HEAD(&dd->fifo_list[READ]);
	INIT_LIST_HEAD(&dd->fifo_list[WRITE]);
	dd->sort_list[READ] = RB_ROOT;
	dd->sort_list[WRITE] = RB_ROOT;
	dd->fifo_expire[READ] = read_expire;
	dd->fifo_expire[WRITE] = write_expire;
	dd->writes_starved = writes_starved;
	dd->front_merges = 1;
	dd->fifo_batch = fifo_batch;

	spin_lock_irq(q->queue_lock);
	q->elevator = eq;
	spin_unlock_irq(q->queue_lock);
	return 0;
}

/*
 * sysfs parts below
 */

static ssize_t
deadline_var_show(int var, char *page)
{
	return sprintf(page, "%d\n", var);
}

static void
deadline_var_store(int *var, const char *page)
{
	char *p = (char *) page;

	*var = simple_strtol(p, &p, 10);
}

#define SHOW_FUNCTION(__FUNC, __VAR, __CONV)				\
static ssize_t __FUNC(struct elevator_queue *e, char *page)		\
{									\
	struct deadline_data *dd = e->elevator_data;			\
	int __data = __VAR;						\
	if (__CONV)							\
		__data = jiffies_to_msecs(__data);			\
	return deadline_var_show(__data, (page));			\
}
SHOW_FUNCTION(deadline_read_expire_show, dd->fifo_expire[READ], 1);
SHOW_FUNCTION(deadline_write_expire_show, dd->fifo_expire[WRITE], 1);
SHOW_FUNCTION(deadline_writes_starved_show, dd->writes_starved, 0);
SHOW_FUNCTION(deadline_front_merges_show, dd->front_merges, 0);
SHOW_FUNCTION(deadline_fifo_batch_show, dd->fifo_batch, 0);
#undef SHOW_FUNCTION

#define STORE_FUNCTION(__FUNC, __PTR, MIN, MAX, __CONV)			\
static ssize_t __FUNC(struct elevator_queue *e, const char *page, size_t count)	\
{									\
	struct deadline_data *dd = e->elevator_data;			\
	int __data;							\
	deadline_var_store(&__data, (page));				\
	if (__data < (MIN))						\
		__data = (MIN);						\
	else if (__data > (MAX))					\
		__data = (MAX);						\
	if (__CONV)							\
		*(__PTR) = msecs_to_jiffies(__data);			\
	else								\
		*(__PTR) = __data;					\
	return count;							\
}
STORE_FUNCTION(deadline_read_expire_store, &dd->fifo_expire[READ], 0, INT_MAX, 1);
STORE_FUNCTION(deadline_write_expire_store, &dd->fifo_expire[WRITE], 0, INT_MAX, 1);
STORE_FUNCTION(deadline_writes_starved_store, &dd->writes_starved, INT_MIN, INT_MAX, 0);
STORE_FUNCTION(deadline_front_merges_store, &dd->front_merges, 0, 1, 0);
STORE_FUNCTION(deadline_fifo_batch_store, &dd->fifo_batch, 0, INT_MAX, 0);
#undef STORE_FUNCTION

#define DD_ATTR(name) \
	__ATTR(name, S_IRUGO|S_IWUSR, deadline_##name##_show, \
				      deadline_##name##_store)

static struct elv_fs_entry deadline_attrs[] = {
	DD_ATTR(read_expire),
	DD_ATTR(write_expire),
	DD_ATTR(writes_starved),
	DD_ATTR(front_merges),
	DD_ATTR(fifo_batch),
	__ATTR_NULL
};

static struct elevator_type iosched_deadline = {
	.ops.sq = {
		.elevator_merge_fn = 		deadline_merge,
		.elevator_merged_fn =		deadline_merged_request,
		.elevator_merge_req_fn =	deadline_merged_requests,
		.elevator_dispatch_fn =		deadline_dispatch_requests,
		.elevator_completed_req_fn =	deadline_completed_request,
		.elevator_add_req_fn =		deadline_add_request,
		.elevator_former_req_fn =	elv_rb_former_request,
		.elevator_latter_req_fn =	elv_rb_latter_request,
		.elevator_init_fn =		deadline_init_queue,
		.elevator_exit_fn =		deadline_exit_queue,
	},

	.elevator_attrs = deadline_attrs,
	.elevator_name = "deadline",
	.elevator_owner = THIS_MODULE,
};

static int __init deadline_init(void)
{
	return elv_register(&iosched_deadline);
}

static void __exit deadline_exit(void)
{
	elv_unregister(&iosched_deadline);
}

module_init(deadline_init);
module_exit(deadline_exit);

MODULE_AUTHOR("Jens Axboe");
MODULE_LICENSE("GPL");
MODULE_DESCRIPTION("deadline IO scheduler");
Commit	Line	Data
1da177e4	1	/*
1da177e4 LT	2	* Deadline i/o scheduler.
1da177e4 LT	3	*
0fe23479	4	* Copyright (C) 2002 Jens Axboe <axboe@kernel.dk>
1da177e4 LT	5	*/
	6	#include <linux/kernel.h>
	7	#include <linux/fs.h>
	8	#include <linux/blkdev.h>
	9	#include <linux/elevator.h>
	10	#include <linux/bio.h>
1da177e4 LT	11	#include <linux/module.h>
	12	#include <linux/slab.h>
	13	#include <linux/init.h>
	14	#include <linux/compiler.h>
1da177e4 LT	15	#include <linux/rbtree.h>
	16
	17	/*
	18	* See Documentation/block/deadline-iosched.txt
	19	*/
64100099 AV	20	static const int read_expire = HZ / 2; /* max time before a read is submitted. */
	21	static const int write_expire = 5 * HZ; /* ditto for writes, these limits are SOFT! */
	22	static const int writes_starved = 2; /* max times reads can starve a write */
	23	static const int fifo_batch = 16; /* # of sequential requests treated as one
1da177e4 LT	24	by the above parameters. For throughput. */
1da177e4 LT	25
1da177e4 LT	26	struct deadline_data {
	27	/*
	28	* run time data
	29	*/
	30
	31	/*
	32	* requests (deadline_rq s) are present on both sort_list and fifo_list
	33	*/
	34	struct rb_root sort_list[2];
	35	struct list_head fifo_list[2];
4fb72f76	36
1da177e4 LT	37	/*
	38	* next in sort order. read, write or both are NULL
	39	*/
8840faa1	40	struct request *next_rq[2];
1da177e4	41	unsigned int batching; /* number of sequential requests made */
1da177e4 LT	42	unsigned int starved; /* times reads have starved writes */
	43
	44	/*
	45	* settings that change how the i/o scheduler behaves
	46	*/
	47	int fifo_expire[2];
	48	int fifo_batch;
	49	int writes_starved;
	50	int front_merges;
1da177e4 LT	51	};
1da177e4 LT	52
4fb72f76 AC	53	static inline struct rb_root *
	54	deadline_rb_root(struct deadline_data dd, struct request rq)
	55	{
	56	return &dd->sort_list[rq_data_dir(rq)];
	57	}
1da177e4	58
5d1a5366 AC	59	/*
	60	* get the request after `rq' in sector-sorted order
	61	*/
	62	static inline struct request *
	63	deadline_latter_request(struct request *rq)
	64	{
	65	struct rb_node *node = rb_next(&rq->rb_node);
	66
	67	if (node)
	68	return rb_entry_rq(node);
	69
	70	return NULL;
	71	}
	72
1da177e4	73	static void
8840faa1	74	deadline_add_rq_rb(struct deadline_data dd, struct request rq)
1da177e4	75	{
4fb72f76	76	struct rb_root *root = deadline_rb_root(dd, rq);
1da177e4	77
796d5116	78	elv_rb_add(root, rq);
1da177e4 LT	79	}
	80
	81	static inline void
8840faa1	82	deadline_del_rq_rb(struct deadline_data dd, struct request rq)
1da177e4	83	{
b8aca35a	84	const int data_dir = rq_data_dir(rq);
1da177e4	85
5d1a5366 AC	86	if (dd->next_rq[data_dir] == rq)
5d1a5366 AC	87	dd->next_rq[data_dir] = deadline_latter_request(rq);
1da177e4	88
4fb72f76	89	elv_rb_del(deadline_rb_root(dd, rq), rq);
1da177e4 LT	90	}
	91
	92	/*
8840faa1	93	* add rq to rbtree and fifo
1da177e4	94	*/
b4878f24	95	static void
1da177e4 LT	96	deadline_add_request(struct request_queue q, struct request rq)
	97	{
	98	struct deadline_data *dd = q->elevator->elevator_data;
8840faa1	99	const int data_dir = rq_data_dir(rq);
1da177e4	100
8dc8146f DLM	101	/*
	102	* This may be a requeue of a write request that has locked its
	103	* target zone. If it is the case, this releases the zone lock.
	104	*/
	105	blk_req_zone_write_unlock(rq);
	106
8840faa1	107	deadline_add_rq_rb(dd, rq);
9817064b	108
1da177e4	109	/*
4fb72f76	110	* set expire time and add to fifo list
1da177e4	111	*/
8b4922d3	112	rq->fifo_time = jiffies + dd->fifo_expire[data_dir];
8840faa1	113	list_add_tail(&rq->queuelist, &dd->fifo_list[data_dir]);
1da177e4 LT	114	}
	115
	116	/*
9817064b	117	* remove rq from rbtree and fifo.
1da177e4	118	*/
165125e1	119	static void deadline_remove_request(struct request_queue q, struct request rq)
1da177e4	120	{
b4878f24	121	struct deadline_data *dd = q->elevator->elevator_data;
1da177e4	122
8840faa1 JA	123	rq_fifo_clear(rq);
8840faa1 JA	124	deadline_del_rq_rb(dd, rq);
1da177e4 LT	125	}
1da177e4 LT	126
34fe7c05	127	static enum elv_merge
165125e1	128	deadline_merge(struct request_queue q, struct request req, struct bio bio)
1da177e4 LT	129	{
	130	struct deadline_data *dd = q->elevator->elevator_data;
	131	struct request *__rq;
1da177e4	132
1da177e4 LT	133	/*
	134	* check for front merge
	135	*/
	136	if (dd->front_merges) {
f73a1c7d	137	sector_t sector = bio_end_sector(bio);
1da177e4	138
b8aca35a	139	__rq = elv_rb_find(&dd->sort_list[bio_data_dir(bio)], sector);
1da177e4	140	if (__rq) {
83096ebf	141	BUG_ON(sector != blk_rq_pos(__rq));
1da177e4	142
72ef799b	143	if (elv_bio_merge_ok(__rq, bio)) {
34fe7c05 CH	144	*req = __rq;
34fe7c05 CH	145	return ELEVATOR_FRONT_MERGE;
1da177e4 LT	146	}
	147	}
	148	}
	149
	150	return ELEVATOR_NO_MERGE;
1da177e4 LT	151	}
1da177e4 LT	152
165125e1	153	static void deadline_merged_request(struct request_queue *q,
34fe7c05	154	struct request *req, enum elv_merge type)
1da177e4 LT	155	{
1da177e4 LT	156	struct deadline_data *dd = q->elevator->elevator_data;
1da177e4	157
1da177e4 LT	158	/*
	159	* if the merge was a front merge, we need to reposition request
	160	*/
b8aca35a	161	if (type == ELEVATOR_FRONT_MERGE) {
4fb72f76	162	elv_rb_del(deadline_rb_root(dd, req), req);
8840faa1	163	deadline_add_rq_rb(dd, req);
1da177e4	164	}
1da177e4 LT	165	}
	166
	167	static void
165125e1	168	deadline_merged_requests(struct request_queue q, struct request req,
1da177e4 LT	169	struct request *next)
1da177e4 LT	170	{
1da177e4	171	/*
8840faa1 JA	172	* if next expires before rq, assign its expire time to rq
8840faa1 JA	173	* and move into next position (next will be deleted) in fifo
1da177e4	174	*/
8840faa1	175	if (!list_empty(&req->queuelist) && !list_empty(&next->queuelist)) {
9828c2c6 JK	176	if (time_before((unsigned long)next->fifo_time,
9828c2c6 JK	177	(unsigned long)req->fifo_time)) {
8840faa1	178	list_move(&req->queuelist, &next->queuelist);
8b4922d3	179	req->fifo_time = next->fifo_time;
1da177e4 LT	180	}
	181	}
	182
	183	/*
	184	* kill knowledge of next, this one is a goner
	185	*/
	186	deadline_remove_request(q, next);
	187	}
	188
	189	/*
	190	* move request from sort list to dispatch queue.
	191	*/
	192	static inline void
8840faa1	193	deadline_move_to_dispatch(struct deadline_data dd, struct request rq)
1da177e4	194	{
165125e1	195	struct request_queue *q = rq->q;
1da177e4	196
8dc8146f DLM	197	/*
	198	* For a zoned block device, write requests must write lock their
	199	* target zone.
	200	*/
	201	blk_req_zone_write_lock(rq);
	202
8840faa1 JA	203	deadline_remove_request(q, rq);
8840faa1 JA	204	elv_dispatch_add_tail(q, rq);
1da177e4 LT	205	}
	206
	207	/*
	208	* move an entry to dispatch queue
	209	*/
	210	static void
8840faa1	211	deadline_move_request(struct deadline_data dd, struct request rq)
1da177e4	212	{
b8aca35a	213	const int data_dir = rq_data_dir(rq);
1da177e4	214
8840faa1 JA	215	dd->next_rq[READ] = NULL;
8840faa1 JA	216	dd->next_rq[WRITE] = NULL;
5d1a5366	217	dd->next_rq[data_dir] = deadline_latter_request(rq);
1da177e4	218
1da177e4 LT	219	/*
	220	* take it off the sort and fifo list, move
	221	* to dispatch queue
	222	*/
8840faa1	223	deadline_move_to_dispatch(dd, rq);
1da177e4 LT	224	}
1da177e4 LT	225
1da177e4	226	/*
4fb72f76	227	* deadline_check_fifo returns 0 if there are no expired requests on the fifo,
1da177e4 LT	228	* 1 otherwise. Requires !list_empty(&dd->fifo_list[data_dir])
	229	*/
	230	static inline int deadline_check_fifo(struct deadline_data *dd, int ddir)
	231	{
8840faa1	232	struct request *rq = rq_entry_fifo(dd->fifo_list[ddir].next);
1da177e4 LT	233
1da177e4 LT	234	/*
8840faa1	235	* rq is expired!
1da177e4	236	*/
9828c2c6	237	if (time_after_eq(jiffies, (unsigned long)rq->fifo_time))
1da177e4 LT	238	return 1;
	239
	240	return 0;
	241	}
	242
c117bac7 DLM	243	/*
	244	* For the specified data direction, return the next request to dispatch using
	245	* arrival ordered lists.
	246	*/
	247	static struct request *
	248	deadline_fifo_request(struct deadline_data *dd, int data_dir)
	249	{
8dc8146f DLM	250	struct request *rq;
8dc8146f DLM	251
c117bac7 DLM	252	if (WARN_ON_ONCE(data_dir != READ && data_dir != WRITE))
	253	return NULL;
	254
	255	if (list_empty(&dd->fifo_list[data_dir]))
	256	return NULL;
	257
8dc8146f DLM	258	rq = rq_entry_fifo(dd->fifo_list[data_dir].next);
	259	if (data_dir == READ \|\| !blk_queue_is_zoned(rq->q))
	260	return rq;
	261
	262	/*
	263	* Look for a write request that can be dispatched, that is one with
	264	* an unlocked target zone.
	265	*/
	266	list_for_each_entry(rq, &dd->fifo_list[WRITE], queuelist) {
	267	if (blk_req_can_dispatch_to_zone(rq))
	268	return rq;
	269	}
	270
	271	return NULL;
c117bac7 DLM	272	}
	273
	274	/*
	275	* For the specified data direction, return the next request to dispatch using
	276	* sector position sorted lists.
	277	*/
	278	static struct request *
	279	deadline_next_request(struct deadline_data *dd, int data_dir)
	280	{
8dc8146f DLM	281	struct request *rq;
8dc8146f DLM	282
c117bac7 DLM	283	if (WARN_ON_ONCE(data_dir != READ && data_dir != WRITE))
	284	return NULL;
	285
8dc8146f DLM	286	rq = dd->next_rq[data_dir];
	287	if (!rq)
	288	return NULL;
	289
	290	if (data_dir == READ \|\| !blk_queue_is_zoned(rq->q))
	291	return rq;
	292
	293	/*
	294	* Look for a write request that can be dispatched, that is one with
	295	* an unlocked target zone.
	296	*/
	297	while (rq) {
	298	if (blk_req_can_dispatch_to_zone(rq))
	299	return rq;
	300	rq = deadline_latter_request(rq);
	301	}
	302
	303	return NULL;
c117bac7 DLM	304	}
c117bac7 DLM	305
1da177e4 LT	306	/*
	307	* deadline_dispatch_requests selects the best request according to
	308	* read/write expire, fifo_batch, etc
	309	*/
165125e1	310	static int deadline_dispatch_requests(struct request_queue *q, int force)
1da177e4	311	{
b4878f24	312	struct deadline_data *dd = q->elevator->elevator_data;
1da177e4 LT	313	const int reads = !list_empty(&dd->fifo_list[READ]);
1da177e4 LT	314	const int writes = !list_empty(&dd->fifo_list[WRITE]);
c117bac7	315	struct request rq, next_rq;
4b0dc07e	316	int data_dir;
1da177e4 LT	317
	318	/*
	319	* batches are currently reads XOR writes
	320	*/
c117bac7 DLM	321	rq = deadline_next_request(dd, WRITE);
	322	if (!rq)
	323	rq = deadline_next_request(dd, READ);
1da177e4	324
63de428b AC	325	if (rq && dd->batching < dd->fifo_batch)
	326	/* we have a next request are still entitled to batch */
	327	goto dispatch_request;
1da177e4 LT	328
	329	/*
	330	* at this point we are not running a batch. select the appropriate
	331	* data direction (read / write)
	332	*/
	333
	334	if (reads) {
dd67d051	335	BUG_ON(RB_EMPTY_ROOT(&dd->sort_list[READ]));
1da177e4	336
8dc8146f DLM	337	if (deadline_fifo_request(dd, WRITE) &&
8dc8146f DLM	338	(dd->starved++ >= dd->writes_starved))
1da177e4 LT	339	goto dispatch_writes;
	340
	341	data_dir = READ;
1da177e4 LT	342
	343	goto dispatch_find_request;
	344	}
	345
	346	/*
	347	* there are either no reads or writes have been starved
	348	*/
	349
	350	if (writes) {
	351	dispatch_writes:
dd67d051	352	BUG_ON(RB_EMPTY_ROOT(&dd->sort_list[WRITE]));
1da177e4 LT	353
	354	dd->starved = 0;
	355
	356	data_dir = WRITE;
1da177e4 LT	357
	358	goto dispatch_find_request;
	359	}
	360
	361	return 0;
	362
	363	dispatch_find_request:
	364	/*
	365	* we are not running a batch, find best request for selected data_dir
	366	*/
c117bac7 DLM	367	next_rq = deadline_next_request(dd, data_dir);
c117bac7 DLM	368	if (deadline_check_fifo(dd, data_dir) \|\| !next_rq) {
6f5d8aa6 AC	369	/*
	370	* A deadline has expired, the last request was in the other
	371	* direction, or we have run out of higher-sectored requests.
	372	* Start again from the request with the earliest expiry time.
	373	*/
c117bac7	374	rq = deadline_fifo_request(dd, data_dir);
6f5d8aa6	375	} else {
1da177e4 LT	376	/*
	377	* The last req was the same dir and we have a next request in
	378	* sort order. No expired requests so continue on from here.
	379	*/
c117bac7	380	rq = next_rq;
1da177e4 LT	381	}
1da177e4 LT	382
8dc8146f DLM	383	/*
	384	* For a zoned block device, if we only have writes queued and none of
	385	* them can be dispatched, rq will be NULL.
	386	*/
	387	if (!rq)
	388	return 0;
	389
dfb3d72a AC	390	dd->batching = 0;
dfb3d72a AC	391
1da177e4 LT	392	dispatch_request:
1da177e4 LT	393	/*
8840faa1	394	* rq is the selected appropriate request.
1da177e4 LT	395	*/
1da177e4 LT	396	dd->batching++;
8840faa1	397	deadline_move_request(dd, rq);
1da177e4 LT	398
	399	return 1;
	400	}
	401
8dc8146f DLM	402	/*
	403	* For zoned block devices, write unlock the target zone of completed
	404	* write requests.
	405	*/
	406	static void
	407	deadline_completed_request(struct request_queue q, struct request rq)
	408	{
	409	blk_req_zone_write_unlock(rq);
	410	}
	411
b374d18a	412	static void deadline_exit_queue(struct elevator_queue *e)
1da177e4 LT	413	{
	414	struct deadline_data *dd = e->elevator_data;
	415
	416	BUG_ON(!list_empty(&dd->fifo_list[READ]));
	417	BUG_ON(!list_empty(&dd->fifo_list[WRITE]));
	418
1da177e4 LT	419	kfree(dd);
	420	}
	421
	422	/*
8840faa1	423	* initialize elevator private data (deadline_data).
1da177e4	424	*/
d50235b7	425	static int deadline_init_queue(struct request_queue q, struct elevator_type e)
1da177e4 LT	426	{
1da177e4 LT	427	struct deadline_data *dd;
d50235b7 JM	428	struct elevator_queue *eq;
	429
	430	eq = elevator_alloc(q, e);
	431	if (!eq)
	432	return -ENOMEM;
1da177e4	433
c1b511eb	434	dd = kzalloc_node(sizeof(*dd), GFP_KERNEL, q->node);
d50235b7 JM	435	if (!dd) {
d50235b7 JM	436	kobject_put(&eq->kobj);
b2fab5ac	437	return -ENOMEM;
d50235b7 JM	438	}
d50235b7 JM	439	eq->elevator_data = dd;
1da177e4	440
1da177e4 LT	441	INIT_LIST_HEAD(&dd->fifo_list[READ]);
	442	INIT_LIST_HEAD(&dd->fifo_list[WRITE]);
	443	dd->sort_list[READ] = RB_ROOT;
	444	dd->sort_list[WRITE] = RB_ROOT;
1da177e4 LT	445	dd->fifo_expire[READ] = read_expire;
	446	dd->fifo_expire[WRITE] = write_expire;
	447	dd->writes_starved = writes_starved;
	448	dd->front_merges = 1;
	449	dd->fifo_batch = fifo_batch;
b2fab5ac	450
d50235b7 JM	451	spin_lock_irq(q->queue_lock);
	452	q->elevator = eq;
	453	spin_unlock_irq(q->queue_lock);
b2fab5ac	454	return 0;
1da177e4 LT	455	}
1da177e4 LT	456
1da177e4 LT	457	/*
	458	* sysfs parts below
	459	*/
1da177e4 LT	460
	461	static ssize_t
	462	deadline_var_show(int var, char *page)
	463	{
	464	return sprintf(page, "%d\n", var);
	465	}
	466
235f8da1	467	static void
235f8da1	468	deadline_var_store(int var, const char page)
1da177e4 LT	469	{
	470	char p = (char ) page;
	471
	472	*var = simple_strtol(p, &p, 10);
1da177e4 LT	473	}
	474
	475	#define SHOW_FUNCTION(__FUNC, __VAR, __CONV) \
b374d18a	476	static ssize_t __FUNC(struct elevator_queue e, char page) \
1da177e4	477	{ \
3d1ab40f AV	478	struct deadline_data *dd = e->elevator_data; \
3d1ab40f AV	479	int __data = __VAR; \
1da177e4 LT	480	if (__CONV) \
	481	__data = jiffies_to_msecs(__data); \
	482	return deadline_var_show(__data, (page)); \
	483	}
e572ec7e AV	484	SHOW_FUNCTION(deadline_read_expire_show, dd->fifo_expire[READ], 1);
	485	SHOW_FUNCTION(deadline_write_expire_show, dd->fifo_expire[WRITE], 1);
	486	SHOW_FUNCTION(deadline_writes_starved_show, dd->writes_starved, 0);
	487	SHOW_FUNCTION(deadline_front_merges_show, dd->front_merges, 0);
	488	SHOW_FUNCTION(deadline_fifo_batch_show, dd->fifo_batch, 0);
1da177e4 LT	489	#undef SHOW_FUNCTION
	490
	491	#define STORE_FUNCTION(__FUNC, __PTR, MIN, MAX, __CONV) \
b374d18a	492	static ssize_t __FUNC(struct elevator_queue e, const char page, size_t count) \
1da177e4	493	{ \
3d1ab40f	494	struct deadline_data *dd = e->elevator_data; \
1da177e4	495	int __data; \
235f8da1	496	deadline_var_store(&__data, (page)); \
1da177e4 LT	497	if (__data < (MIN)) \
	498	__data = (MIN); \
	499	else if (__data > (MAX)) \
	500	__data = (MAX); \
	501	if (__CONV) \
	502	*(__PTR) = msecs_to_jiffies(__data); \
	503	else \
	504	*(__PTR) = __data; \
235f8da1	505	return count; \
1da177e4	506	}
e572ec7e AV	507	STORE_FUNCTION(deadline_read_expire_store, &dd->fifo_expire[READ], 0, INT_MAX, 1);
	508	STORE_FUNCTION(deadline_write_expire_store, &dd->fifo_expire[WRITE], 0, INT_MAX, 1);
	509	STORE_FUNCTION(deadline_writes_starved_store, &dd->writes_starved, INT_MIN, INT_MAX, 0);
	510	STORE_FUNCTION(deadline_front_merges_store, &dd->front_merges, 0, 1, 0);
	511	STORE_FUNCTION(deadline_fifo_batch_store, &dd->fifo_batch, 0, INT_MAX, 0);
1da177e4 LT	512	#undef STORE_FUNCTION
1da177e4 LT	513
e572ec7e AV	514	#define DD_ATTR(name) \
	515	__ATTR(name, S_IRUGO\|S_IWUSR, deadline_##name##_show, \
	516	deadline_##name##_store)
	517
	518	static struct elv_fs_entry deadline_attrs[] = {
	519	DD_ATTR(read_expire),
	520	DD_ATTR(write_expire),
	521	DD_ATTR(writes_starved),
	522	DD_ATTR(front_merges),
	523	DD_ATTR(fifo_batch),
	524	__ATTR_NULL
1da177e4 LT	525	};
1da177e4 LT	526
1da177e4	527	static struct elevator_type iosched_deadline = {
c51ca6cf	528	.ops.sq = {
1da177e4 LT	529	.elevator_merge_fn = deadline_merge,
	530	.elevator_merged_fn = deadline_merged_request,
	531	.elevator_merge_req_fn = deadline_merged_requests,
b4878f24	532	.elevator_dispatch_fn = deadline_dispatch_requests,
8dc8146f	533	.elevator_completed_req_fn = deadline_completed_request,
b4878f24	534	.elevator_add_req_fn = deadline_add_request,
b8aca35a JA	535	.elevator_former_req_fn = elv_rb_former_request,
b8aca35a JA	536	.elevator_latter_req_fn = elv_rb_latter_request,
1da177e4 LT	537	.elevator_init_fn = deadline_init_queue,
	538	.elevator_exit_fn = deadline_exit_queue,
	539	},
	540
3d1ab40f	541	.elevator_attrs = deadline_attrs,
1da177e4 LT	542	.elevator_name = "deadline",
	543	.elevator_owner = THIS_MODULE,
	544	};
	545
	546	static int __init deadline_init(void)
	547	{
3d3c2379	548	return elv_register(&iosched_deadline);
1da177e4 LT	549	}
	550
	551	static void __exit deadline_exit(void)
	552	{
1da177e4 LT	553	elv_unregister(&iosched_deadline);
	554	}
	555
	556	module_init(deadline_init);
	557	module_exit(deadline_exit);
	558
	559	MODULE_AUTHOR("Jens Axboe");
	560	MODULE_LICENSE("GPL");
	561	MODULE_DESCRIPTION("deadline IO scheduler");