[linux-block.git] / drivers / md / dm-delay.c

/*
 * Copyright (C) 2005-2007 Red Hat GmbH
 *
 * A target that delays reads and/or writes and can send
 * them to different devices.
 *
 * This file is released under the GPL.
 */

#include <linux/module.h>
#include <linux/init.h>
#include <linux/blkdev.h>
#include <linux/bio.h>
#include <linux/slab.h>

#include "dm.h"
#include "dm-bio-list.h"

#define DM_MSG_PREFIX "delay"

struct delay_c {
	struct timer_list delay_timer;
	struct mutex timer_lock;
	struct work_struct flush_expired_bios;
	struct list_head delayed_bios;
	atomic_t may_delay;
	mempool_t *delayed_pool;

	struct dm_dev *dev_read;
	sector_t start_read;
	unsigned read_delay;
	unsigned reads;

	struct dm_dev *dev_write;
	sector_t start_write;
	unsigned write_delay;
	unsigned writes;
};

struct dm_delay_info {
	struct delay_c *context;
	struct list_head list;
	struct bio *bio;
	unsigned long expires;
};

static DEFINE_MUTEX(delayed_bios_lock);

static struct workqueue_struct *kdelayd_wq;
static struct kmem_cache *delayed_cache;

static void handle_delayed_timer(unsigned long data)
{
	struct delay_c *dc = (struct delay_c *)data;

	queue_work(kdelayd_wq, &dc->flush_expired_bios);
}

static void queue_timeout(struct delay_c *dc, unsigned long expires)
{
	mutex_lock(&dc->timer_lock);

	if (!timer_pending(&dc->delay_timer) || expires < dc->delay_timer.expires)
		mod_timer(&dc->delay_timer, expires);

	mutex_unlock(&dc->timer_lock);
}

static void flush_bios(struct bio *bio)
{
	struct bio *n;

	while (bio) {
		n = bio->bi_next;
		bio->bi_next = NULL;
		generic_make_request(bio);
		bio = n;
	}
}

static struct bio *flush_delayed_bios(struct delay_c *dc, int flush_all)
{
	struct dm_delay_info *delayed, *next;
	unsigned long next_expires = 0;
	int start_timer = 0;
	BIO_LIST(flush_bios);

	mutex_lock(&delayed_bios_lock);
	list_for_each_entry_safe(delayed, next, &dc->delayed_bios, list) {
		if (flush_all || time_after_eq(jiffies, delayed->expires)) {
			list_del(&delayed->list);
			bio_list_add(&flush_bios, delayed->bio);
			if ((bio_data_dir(delayed->bio) == WRITE))
				delayed->context->writes--;
			else
				delayed->context->reads--;
			mempool_free(delayed, dc->delayed_pool);
			continue;
		}

		if (!start_timer) {
			start_timer = 1;
			next_expires = delayed->expires;
		} else
			next_expires = min(next_expires, delayed->expires);
	}

	mutex_unlock(&delayed_bios_lock);

	if (start_timer)
		queue_timeout(dc, next_expires);

	return bio_list_get(&flush_bios);
}

static void flush_expired_bios(struct work_struct *work)
{
	struct delay_c *dc;

	dc = container_of(work, struct delay_c, flush_expired_bios);
	flush_bios(flush_delayed_bios(dc, 0));
}

/*
 * Mapping parameters:
 *    <device> <offset> <delay> [<write_device> <write_offset> <write_delay>]
 *
 * With separate write parameters, the first set is only used for reads.
 * Delays are specified in milliseconds.
 */
static int delay_ctr(struct dm_target *ti, unsigned int argc, char **argv)
{
	struct delay_c *dc;
	unsigned long long tmpll;

	if (argc != 3 && argc != 6) {
		ti->error = "requires exactly 3 or 6 arguments";
		return -EINVAL;
	}

	dc = kmalloc(sizeof(*dc), GFP_KERNEL);
	if (!dc) {
		ti->error = "Cannot allocate context";
		return -ENOMEM;
	}

	dc->reads = dc->writes = 0;

	if (sscanf(argv[1], "%llu", &tmpll) != 1) {
		ti->error = "Invalid device sector";
		goto bad;
	}
	dc->start_read = tmpll;

	if (sscanf(argv[2], "%u", &dc->read_delay) != 1) {
		ti->error = "Invalid delay";
		goto bad;
	}

	if (dm_get_device(ti, argv[0], dc->start_read, ti->len,
			  dm_table_get_mode(ti->table), &dc->dev_read)) {
		ti->error = "Device lookup failed";
		goto bad;
	}

	if (argc == 3) {
		dc->dev_write = NULL;
		goto out;
	}

	if (sscanf(argv[4], "%llu", &tmpll) != 1) {
		ti->error = "Invalid write device sector";
		goto bad;
	}
	dc->start_write = tmpll;

	if (sscanf(argv[5], "%u", &dc->write_delay) != 1) {
		ti->error = "Invalid write delay";
		goto bad;
	}

	if (dm_get_device(ti, argv[3], dc->start_write, ti->len,
			  dm_table_get_mode(ti->table), &dc->dev_write)) {
		ti->error = "Write device lookup failed";
		dm_put_device(ti, dc->dev_read);
		goto bad;
	}

out:
	dc->delayed_pool = mempool_create_slab_pool(128, delayed_cache);
	if (!dc->delayed_pool) {
		DMERR("Couldn't create delayed bio pool.");
		goto bad;
	}

	setup_timer(&dc->delay_timer, handle_delayed_timer, (unsigned long)dc);

	INIT_WORK(&dc->flush_expired_bios, flush_expired_bios);
	INIT_LIST_HEAD(&dc->delayed_bios);
	mutex_init(&dc->timer_lock);
	atomic_set(&dc->may_delay, 1);

	ti->private = dc;
	return 0;

bad:
	kfree(dc);
	return -EINVAL;
}

static void delay_dtr(struct dm_target *ti)
{
	struct delay_c *dc = ti->private;

	flush_workqueue(kdelayd_wq);

	dm_put_device(ti, dc->dev_read);

	if (dc->dev_write)
		dm_put_device(ti, dc->dev_write);

	mempool_destroy(dc->delayed_pool);
	kfree(dc);
}

static int delay_bio(struct delay_c *dc, int delay, struct bio *bio)
{
	struct dm_delay_info *delayed;
	unsigned long expires = 0;

	if (!delay || !atomic_read(&dc->may_delay))
		return 1;

	delayed = mempool_alloc(dc->delayed_pool, GFP_NOIO);

	delayed->context = dc;
	delayed->bio = bio;
	delayed->expires = expires = jiffies + (delay * HZ / 1000);

	mutex_lock(&delayed_bios_lock);

	if (bio_data_dir(bio) == WRITE)
		dc->writes++;
	else
		dc->reads++;

	list_add_tail(&delayed->list, &dc->delayed_bios);

	mutex_unlock(&delayed_bios_lock);

	queue_timeout(dc, expires);

	return 0;
}

static void delay_presuspend(struct dm_target *ti)
{
	struct delay_c *dc = ti->private;

	atomic_set(&dc->may_delay, 0);
	del_timer_sync(&dc->delay_timer);
	flush_bios(flush_delayed_bios(dc, 1));
}

static void delay_resume(struct dm_target *ti)
{
	struct delay_c *dc = ti->private;

	atomic_set(&dc->may_delay, 1);
}

static int delay_map(struct dm_target *ti, struct bio *bio,
		     union map_info *map_context)
{
	struct delay_c *dc = ti->private;

	if ((bio_data_dir(bio) == WRITE) && (dc->dev_write)) {
		bio->bi_bdev = dc->dev_write->bdev;
		bio->bi_sector = dc->start_write +
				 (bio->bi_sector - ti->begin);

		return delay_bio(dc, dc->write_delay, bio);
	}

	bio->bi_bdev = dc->dev_read->bdev;
	bio->bi_sector = dc->start_read +
			 (bio->bi_sector - ti->begin);

	return delay_bio(dc, dc->read_delay, bio);
}

static int delay_status(struct dm_target *ti, status_type_t type,
			char *result, unsigned maxlen)
{
	struct delay_c *dc = ti->private;
	int sz = 0;

	switch (type) {
	case STATUSTYPE_INFO:
		DMEMIT("%u %u", dc->reads, dc->writes);
		break;

	case STATUSTYPE_TABLE:
		DMEMIT("%s %llu %u", dc->dev_read->name,
		       (unsigned long long) dc->start_read,
		       dc->read_delay);
		if (dc->dev_write)
			DMEMIT("%s %llu %u", dc->dev_write->name,
			       (unsigned long long) dc->start_write,
			       dc->write_delay);
		break;
	}

	return 0;
}

static struct target_type delay_target = {
	.name	     = "delay",
	.version     = {1, 0, 2},
	.module      = THIS_MODULE,
	.ctr	     = delay_ctr,
	.dtr	     = delay_dtr,
	.map	     = delay_map,
	.presuspend  = delay_presuspend,
	.resume	     = delay_resume,
	.status	     = delay_status,
};

static int __init dm_delay_init(void)
{
	int r = -ENOMEM;

	kdelayd_wq = create_workqueue("kdelayd");
	if (!kdelayd_wq) {
		DMERR("Couldn't start kdelayd");
		goto bad_queue;
	}

	delayed_cache = KMEM_CACHE(dm_delay_info, 0);
	if (!delayed_cache) {
		DMERR("Couldn't create delayed bio cache.");
		goto bad_memcache;
	}

	r = dm_register_target(&delay_target);
	if (r < 0) {
		DMERR("register failed %d", r);
		goto bad_register;
	}

	return 0;

bad_register:
	kmem_cache_destroy(delayed_cache);
bad_memcache:
	destroy_workqueue(kdelayd_wq);
bad_queue:
	return r;
}

static void __exit dm_delay_exit(void)
{
	int r = dm_unregister_target(&delay_target);

	if (r < 0)
		DMERR("unregister failed %d", r);

	kmem_cache_destroy(delayed_cache);
	destroy_workqueue(kdelayd_wq);
}

/* Module hooks */
module_init(dm_delay_init);
module_exit(dm_delay_exit);

MODULE_DESCRIPTION(DM_NAME " delay target");
MODULE_AUTHOR("Heinz Mauelshagen <mauelshagen@redhat.com>");
MODULE_LICENSE("GPL");
Commit	Line	Data
26b9f228 HM	1	/*
	2	* Copyright (C) 2005-2007 Red Hat GmbH
	3	*
	4	* A target that delays reads and/or writes and can send
	5	* them to different devices.
	6	*
	7	* This file is released under the GPL.
	8	*/
	9
	10	#include <linux/module.h>
	11	#include <linux/init.h>
	12	#include <linux/blkdev.h>
	13	#include <linux/bio.h>
	14	#include <linux/slab.h>
	15
	16	#include "dm.h"
	17	#include "dm-bio-list.h"
	18
	19	#define DM_MSG_PREFIX "delay"
	20
	21	struct delay_c {
	22	struct timer_list delay_timer;
ac818646	23	struct mutex timer_lock;
26b9f228 HM	24	struct work_struct flush_expired_bios;
	25	struct list_head delayed_bios;
	26	atomic_t may_delay;
	27	mempool_t *delayed_pool;
	28
	29	struct dm_dev *dev_read;
	30	sector_t start_read;
	31	unsigned read_delay;
	32	unsigned reads;
	33
	34	struct dm_dev *dev_write;
	35	sector_t start_write;
	36	unsigned write_delay;
	37	unsigned writes;
	38	};
	39
028867ac	40	struct dm_delay_info {
26b9f228 HM	41	struct delay_c *context;
	42	struct list_head list;
	43	struct bio *bio;
	44	unsigned long expires;
	45	};
	46
	47	static DEFINE_MUTEX(delayed_bios_lock);
	48
	49	static struct workqueue_struct *kdelayd_wq;
	50	static struct kmem_cache *delayed_cache;
	51
	52	static void handle_delayed_timer(unsigned long data)
	53	{
	54	struct delay_c dc = (struct delay_c )data;
	55
	56	queue_work(kdelayd_wq, &dc->flush_expired_bios);
	57	}
	58
	59	static void queue_timeout(struct delay_c *dc, unsigned long expires)
	60	{
ac818646	61	mutex_lock(&dc->timer_lock);
26b9f228 HM	62
	63	if (!timer_pending(&dc->delay_timer) \|\| expires < dc->delay_timer.expires)
	64	mod_timer(&dc->delay_timer, expires);
	65
ac818646	66	mutex_unlock(&dc->timer_lock);
26b9f228 HM	67	}
	68
	69	static void flush_bios(struct bio *bio)
	70	{
	71	struct bio *n;
	72
	73	while (bio) {
	74	n = bio->bi_next;
	75	bio->bi_next = NULL;
	76	generic_make_request(bio);
	77	bio = n;
	78	}
	79	}
	80
	81	static struct bio flush_delayed_bios(struct delay_c dc, int flush_all)
	82	{
028867ac	83	struct dm_delay_info delayed, next;
26b9f228 HM	84	unsigned long next_expires = 0;
	85	int start_timer = 0;
	86	BIO_LIST(flush_bios);
	87
	88	mutex_lock(&delayed_bios_lock);
	89	list_for_each_entry_safe(delayed, next, &dc->delayed_bios, list) {
	90	if (flush_all \|\| time_after_eq(jiffies, delayed->expires)) {
	91	list_del(&delayed->list);
	92	bio_list_add(&flush_bios, delayed->bio);
	93	if ((bio_data_dir(delayed->bio) == WRITE))
	94	delayed->context->writes--;
	95	else
	96	delayed->context->reads--;
	97	mempool_free(delayed, dc->delayed_pool);
	98	continue;
	99	}
	100
	101	if (!start_timer) {
	102	start_timer = 1;
	103	next_expires = delayed->expires;
	104	} else
	105	next_expires = min(next_expires, delayed->expires);
	106	}
	107
	108	mutex_unlock(&delayed_bios_lock);
	109
	110	if (start_timer)
	111	queue_timeout(dc, next_expires);
	112
	113	return bio_list_get(&flush_bios);
	114	}
	115
	116	static void flush_expired_bios(struct work_struct *work)
	117	{
	118	struct delay_c *dc;
	119
	120	dc = container_of(work, struct delay_c, flush_expired_bios);
	121	flush_bios(flush_delayed_bios(dc, 0));
	122	}
	123
	124	/*
	125	* Mapping parameters:
	126	* <device> <offset> <delay> [<write_device> <write_offset> <write_delay>]
	127	*
	128	* With separate write parameters, the first set is only used for reads.
	129	* Delays are specified in milliseconds.
	130	*/
	131	static int delay_ctr(struct dm_target ti, unsigned int argc, char *argv)
	132	{
	133	struct delay_c *dc;
	134	unsigned long long tmpll;
	135
	136	if (argc != 3 && argc != 6) {
	137	ti->error = "requires exactly 3 or 6 arguments";
	138	return -EINVAL;
	139	}
	140
	141	dc = kmalloc(sizeof(*dc), GFP_KERNEL);
	142	if (!dc) {
	143	ti->error = "Cannot allocate context";
	144	return -ENOMEM;
	145	}
	146
	147	dc->reads = dc->writes = 0;
148
149	if (sscanf(argv[1], "%llu", &tmpll) != 1) {
150	ti->error = "Invalid device sector";
151	goto bad;
152	}
153	dc->start_read = tmpll;
154
155	if (sscanf(argv[2], "%u", &dc->read_delay) != 1) {
156	ti->error = "Invalid delay";
157	goto bad;
158	}
159
160	if (dm_get_device(ti, argv[0], dc->start_read, ti->len,
161	dm_table_get_mode(ti->table), &dc->dev_read)) {
162	ti->error = "Device lookup failed";
163	goto bad;
164	}
165
166	if (argc == 3) {
167	dc->dev_write = NULL;
168	goto out;
169	}
170
171	if (sscanf(argv[4], "%llu", &tmpll) != 1) {
172	ti->error = "Invalid write device sector";
173	goto bad;
174	}
175	dc->start_write = tmpll;
176
177	if (sscanf(argv[5], "%u", &dc->write_delay) != 1) {
178	ti->error = "Invalid write delay";
179	goto bad;
180	}
181
182	if (dm_get_device(ti, argv[3], dc->start_write, ti->len,
183	dm_table_get_mode(ti->table), &dc->dev_write)) {
184	ti->error = "Write device lookup failed";
185	dm_put_device(ti, dc->dev_read);
186	goto bad;
187	}
188
189	out:
190	dc->delayed_pool = mempool_create_slab_pool(128, delayed_cache);
191	if (!dc->delayed_pool) {
192	DMERR("Couldn't create delayed bio pool.");
193	goto bad;
194	}
195
ac818646	196	setup_timer(&dc->delay_timer, handle_delayed_timer, (unsigned long)dc);
26b9f228 HM	197
	198	INIT_WORK(&dc->flush_expired_bios, flush_expired_bios);
	199	INIT_LIST_HEAD(&dc->delayed_bios);
ac818646	200	mutex_init(&dc->timer_lock);
26b9f228 HM	201	atomic_set(&dc->may_delay, 1);
	202
	203	ti->private = dc;
	204	return 0;
	205
	206	bad:
	207	kfree(dc);
	208	return -EINVAL;
	209	}
	210
	211	static void delay_dtr(struct dm_target *ti)
	212	{
	213	struct delay_c *dc = ti->private;
	214
	215	flush_workqueue(kdelayd_wq);
	216
	217	dm_put_device(ti, dc->dev_read);
	218
	219	if (dc->dev_write)
	220	dm_put_device(ti, dc->dev_write);
	221
	222	mempool_destroy(dc->delayed_pool);
	223	kfree(dc);
	224	}
	225
	226	static int delay_bio(struct delay_c dc, int delay, struct bio bio)
	227	{
028867ac	228	struct dm_delay_info *delayed;
26b9f228 HM	229	unsigned long expires = 0;
	230
	231	if (!delay \|\| !atomic_read(&dc->may_delay))
	232	return 1;
	233
	234	delayed = mempool_alloc(dc->delayed_pool, GFP_NOIO);
	235
	236	delayed->context = dc;
	237	delayed->bio = bio;
	238	delayed->expires = expires = jiffies + (delay * HZ / 1000);
	239
	240	mutex_lock(&delayed_bios_lock);
	241
	242	if (bio_data_dir(bio) == WRITE)
	243	dc->writes++;
	244	else
	245	dc->reads++;
	246
	247	list_add_tail(&delayed->list, &dc->delayed_bios);
	248
	249	mutex_unlock(&delayed_bios_lock);
	250
	251	queue_timeout(dc, expires);
	252
	253	return 0;
	254	}
	255
	256	static void delay_presuspend(struct dm_target *ti)
	257	{
	258	struct delay_c *dc = ti->private;
	259
	260	atomic_set(&dc->may_delay, 0);
	261	del_timer_sync(&dc->delay_timer);
	262	flush_bios(flush_delayed_bios(dc, 1));
	263	}
	264
	265	static void delay_resume(struct dm_target *ti)
	266	{
	267	struct delay_c *dc = ti->private;
	268
	269	atomic_set(&dc->may_delay, 1);
	270	}
	271
	272	static int delay_map(struct dm_target ti, struct bio bio,
	273	union map_info *map_context)
	274	{
	275	struct delay_c *dc = ti->private;
	276
	277	if ((bio_data_dir(bio) == WRITE) && (dc->dev_write)) {
	278	bio->bi_bdev = dc->dev_write->bdev;
	279	bio->bi_sector = dc->start_write +
	280	(bio->bi_sector - ti->begin);
	281
	282	return delay_bio(dc, dc->write_delay, bio);
	283	}
	284
	285	bio->bi_bdev = dc->dev_read->bdev;
	286	bio->bi_sector = dc->start_read +
	287	(bio->bi_sector - ti->begin);
	288
	289	return delay_bio(dc, dc->read_delay, bio);
	290	}
	291
	292	static int delay_status(struct dm_target *ti, status_type_t type,
293	char *result, unsigned maxlen)
294	{
295	struct delay_c *dc = ti->private;
296	int sz = 0;
297
298	switch (type) {
299	case STATUSTYPE_INFO:
300	DMEMIT("%u %u", dc->reads, dc->writes);
301	break;
302
303	case STATUSTYPE_TABLE:
304	DMEMIT("%s %llu %u", dc->dev_read->name,
305	(unsigned long long) dc->start_read,
306	dc->read_delay);
307	if (dc->dev_write)
308	DMEMIT("%s %llu %u", dc->dev_write->name,
309	(unsigned long long) dc->start_write,
310	dc->write_delay);
311	break;
312	}
313
314	return 0;
315	}
316
317	static struct target_type delay_target = {
318	.name = "delay",
319	.version = {1, 0, 2},
320	.module = THIS_MODULE,
321	.ctr = delay_ctr,
322	.dtr = delay_dtr,
323	.map = delay_map,
324	.presuspend = delay_presuspend,
325	.resume = delay_resume,
326	.status = delay_status,
327	};
328
329	static int __init dm_delay_init(void)
330	{
331	int r = -ENOMEM;
332
333	kdelayd_wq = create_workqueue("kdelayd");
334	if (!kdelayd_wq) {
335	DMERR("Couldn't start kdelayd");
336	goto bad_queue;
337	}
338
028867ac	339	delayed_cache = KMEM_CACHE(dm_delay_info, 0);
26b9f228 HM	340	if (!delayed_cache) {
	341	DMERR("Couldn't create delayed bio cache.");
	342	goto bad_memcache;
	343	}
	344
	345	r = dm_register_target(&delay_target);
	346	if (r < 0) {
	347	DMERR("register failed %d", r);
	348	goto bad_register;
	349	}
	350
	351	return 0;
	352
	353	bad_register:
	354	kmem_cache_destroy(delayed_cache);
	355	bad_memcache:
	356	destroy_workqueue(kdelayd_wq);
	357	bad_queue:
	358	return r;
	359	}
	360
	361	static void __exit dm_delay_exit(void)
	362	{
	363	int r = dm_unregister_target(&delay_target);
	364
	365	if (r < 0)
	366	DMERR("unregister failed %d", r);
	367
	368	kmem_cache_destroy(delayed_cache);
	369	destroy_workqueue(kdelayd_wq);
	370	}
	371
	372	/* Module hooks */
	373	module_init(dm_delay_init);
	374	module_exit(dm_delay_exit);
	375
	376	MODULE_DESCRIPTION(DM_NAME " delay target");
	377	MODULE_AUTHOR("Heinz Mauelshagen <mauelshagen@redhat.com>");
	378	MODULE_LICENSE("GPL");