[mirror_ubuntu-bionic-kernel.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 <linux/device-mapper.h>

#define DM_MSG_PREFIX "delay"

struct delay_c {
	struct timer_list delay_timer;
	struct mutex timer_lock;
	struct workqueue_struct *kdelayd_wq;
	struct work_struct flush_expired_bios;
	struct list_head delayed_bios;
	atomic_t may_delay;

	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;
	unsigned long expires;
};

static DEFINE_MUTEX(delayed_bios_lock);

static void handle_delayed_timer(struct timer_list *t)
{
	struct delay_c *dc = from_timer(dc, t, delay_timer);

	queue_work(dc->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;
	struct 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)) {
			struct bio *bio = dm_bio_from_per_bio_data(delayed,
						sizeof(struct dm_delay_info));
			list_del(&delayed->list);
			bio_list_add(&flush_bios, bio);
			if ((bio_data_dir(bio) == WRITE))
				delayed->context->writes--;
			else
				delayed->context->reads--;
			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.
 * Offsets are specified in sectors.
 * 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;
	char dummy;
	int ret;

	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;

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

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

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

	ret = -EINVAL;
	dc->dev_write = NULL;
	if (argc == 3)
		goto out;

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

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

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

out:
	ret = -EINVAL;
	dc->kdelayd_wq = alloc_workqueue("kdelayd", WQ_MEM_RECLAIM, 0);
	if (!dc->kdelayd_wq) {
		DMERR("Couldn't start kdelayd");
		goto bad_queue;
	}

	timer_setup(&dc->delay_timer, handle_delayed_timer, 0);

	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->num_flush_bios = 1;
	ti->num_discard_bios = 1;
	ti->per_io_data_size = sizeof(struct dm_delay_info);
	ti->private = dc;
	return 0;

bad_queue:
	if (dc->dev_write)
		dm_put_device(ti, dc->dev_write);
bad_dev_read:
	dm_put_device(ti, dc->dev_read);
bad:
	kfree(dc);
	return ret;
}

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

	if (dc->kdelayd_wq)
		destroy_workqueue(dc->kdelayd_wq);

	dm_put_device(ti, dc->dev_read);

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

	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 DM_MAPIO_REMAPPED;

	delayed = dm_per_bio_data(bio, sizeof(struct dm_delay_info));

	delayed->context = dc;
	delayed->expires = expires = jiffies + msecs_to_jiffies(delay);

	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 DM_MAPIO_SUBMITTED;
}

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)
{
	struct delay_c *dc = ti->private;

	if ((bio_data_dir(bio) == WRITE) && (dc->dev_write)) {
		bio_set_dev(bio, dc->dev_write->bdev);
		if (bio_sectors(bio))
			bio->bi_iter.bi_sector = dc->start_write +
				dm_target_offset(ti, bio->bi_iter.bi_sector);

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

	bio_set_dev(bio, dc->dev_read->bdev);
	bio->bi_iter.bi_sector = dc->start_read +
		dm_target_offset(ti, bio->bi_iter.bi_sector);

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

static void delay_status(struct dm_target *ti, status_type_t type,
			 unsigned status_flags, 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;
	}
}

static int delay_iterate_devices(struct dm_target *ti,
				 iterate_devices_callout_fn fn, void *data)
{
	struct delay_c *dc = ti->private;
	int ret = 0;

	ret = fn(ti, dc->dev_read, dc->start_read, ti->len, data);
	if (ret)
		goto out;

	if (dc->dev_write)
		ret = fn(ti, dc->dev_write, dc->start_write, ti->len, data);

out:
	return ret;
}

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

static int __init dm_delay_init(void)
{
	int r;

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

	return 0;

bad_register:
	return r;
}

static void __exit dm_delay_exit(void)
{
	dm_unregister_target(&delay_target);
}

/* 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
	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 <linux/device-mapper.h>
	17
	18	#define DM_MSG_PREFIX "delay"
	19
	20	struct delay_c {
	21	struct timer_list delay_timer;
	22	struct mutex timer_lock;
	23	struct workqueue_struct *kdelayd_wq;
	24	struct work_struct flush_expired_bios;
	25	struct list_head delayed_bios;
	26	atomic_t may_delay;
	27
	28	struct dm_dev *dev_read;
	29	sector_t start_read;
	30	unsigned read_delay;
	31	unsigned reads;
	32
	33	struct dm_dev *dev_write;
	34	sector_t start_write;
	35	unsigned write_delay;
	36	unsigned writes;
	37	};
	38
	39	struct dm_delay_info {
	40	struct delay_c *context;
	41	struct list_head list;
	42	unsigned long expires;
	43	};
	44
	45	static DEFINE_MUTEX(delayed_bios_lock);
	46
	47	static void handle_delayed_timer(struct timer_list *t)
	48	{
	49	struct delay_c *dc = from_timer(dc, t, delay_timer);
	50
	51	queue_work(dc->kdelayd_wq, &dc->flush_expired_bios);
	52	}
	53
	54	static void queue_timeout(struct delay_c *dc, unsigned long expires)
	55	{
	56	mutex_lock(&dc->timer_lock);
	57
	58	if (!timer_pending(&dc->delay_timer) \|\| expires < dc->delay_timer.expires)
	59	mod_timer(&dc->delay_timer, expires);
	60
	61	mutex_unlock(&dc->timer_lock);
	62	}
	63
	64	static void flush_bios(struct bio *bio)
	65	{
	66	struct bio *n;
	67
	68	while (bio) {
	69	n = bio->bi_next;
	70	bio->bi_next = NULL;
	71	generic_make_request(bio);
	72	bio = n;
	73	}
	74	}
	75
	76	static struct bio flush_delayed_bios(struct delay_c dc, int flush_all)
	77	{
	78	struct dm_delay_info delayed, next;
	79	unsigned long next_expires = 0;
	80	int start_timer = 0;
	81	struct bio_list flush_bios = { };
	82
	83	mutex_lock(&delayed_bios_lock);
	84	list_for_each_entry_safe(delayed, next, &dc->delayed_bios, list) {
	85	if (flush_all \|\| time_after_eq(jiffies, delayed->expires)) {
	86	struct bio *bio = dm_bio_from_per_bio_data(delayed,
	87	sizeof(struct dm_delay_info));
	88	list_del(&delayed->list);
	89	bio_list_add(&flush_bios, bio);
	90	if ((bio_data_dir(bio) == WRITE))
	91	delayed->context->writes--;
	92	else
	93	delayed->context->reads--;
	94	continue;
	95	}
	96
	97	if (!start_timer) {
	98	start_timer = 1;
	99	next_expires = delayed->expires;
	100	} else
	101	next_expires = min(next_expires, delayed->expires);
	102	}
	103
	104	mutex_unlock(&delayed_bios_lock);
	105
	106	if (start_timer)
	107	queue_timeout(dc, next_expires);
	108
	109	return bio_list_get(&flush_bios);
	110	}
	111
	112	static void flush_expired_bios(struct work_struct *work)
	113	{
	114	struct delay_c *dc;
	115
	116	dc = container_of(work, struct delay_c, flush_expired_bios);
	117	flush_bios(flush_delayed_bios(dc, 0));
	118	}
	119
	120	/*
	121	* Mapping parameters:
	122	* <device> <offset> <delay> [<write_device> <write_offset> <write_delay>]
	123	*
	124	* With separate write parameters, the first set is only used for reads.
	125	* Offsets are specified in sectors.
	126	* Delays are specified in milliseconds.
	127	*/
	128	static int delay_ctr(struct dm_target ti, unsigned int argc, char *argv)
	129	{
	130	struct delay_c *dc;
	131	unsigned long long tmpll;
	132	char dummy;
	133	int ret;
	134
	135	if (argc != 3 && argc != 6) {
	136	ti->error = "Requires exactly 3 or 6 arguments";
	137	return -EINVAL;
	138	}
	139
	140	dc = kmalloc(sizeof(*dc), GFP_KERNEL);
	141	if (!dc) {
	142	ti->error = "Cannot allocate context";
	143	return -ENOMEM;
	144	}
	145
	146	dc->reads = dc->writes = 0;
	147
	148	ret = -EINVAL;
	149	if (sscanf(argv[1], "%llu%c", &tmpll, &dummy) != 1 \|\| tmpll != (sector_t)tmpll) {
	150	ti->error = "Invalid device sector";
	151	goto bad;
	152	}
	153	dc->start_read = tmpll;
	154
	155	if (sscanf(argv[2], "%u%c", &dc->read_delay, &dummy) != 1) {
	156	ti->error = "Invalid delay";
	157	goto bad;
	158	}
	159
	160	ret = dm_get_device(ti, argv[0], dm_table_get_mode(ti->table),
	161	&dc->dev_read);
	162	if (ret) {
	163	ti->error = "Device lookup failed";
	164	goto bad;
	165	}
	166
	167	ret = -EINVAL;
	168	dc->dev_write = NULL;
	169	if (argc == 3)
	170	goto out;
	171
	172	if (sscanf(argv[4], "%llu%c", &tmpll, &dummy) != 1) {
	173	ti->error = "Invalid write device sector";
	174	goto bad_dev_read;
	175	}
	176	dc->start_write = tmpll;
	177
	178	if (sscanf(argv[5], "%u%c", &dc->write_delay, &dummy) != 1) {
	179	ti->error = "Invalid write delay";
	180	goto bad_dev_read;
	181	}
	182
	183	ret = dm_get_device(ti, argv[3], dm_table_get_mode(ti->table),
	184	&dc->dev_write);
	185	if (ret) {
	186	ti->error = "Write device lookup failed";
	187	goto bad_dev_read;
	188	}
	189
	190	out:
	191	ret = -EINVAL;
	192	dc->kdelayd_wq = alloc_workqueue("kdelayd", WQ_MEM_RECLAIM, 0);
	193	if (!dc->kdelayd_wq) {
	194	DMERR("Couldn't start kdelayd");
	195	goto bad_queue;
	196	}
	197
	198	timer_setup(&dc->delay_timer, handle_delayed_timer, 0);
	199
	200	INIT_WORK(&dc->flush_expired_bios, flush_expired_bios);
	201	INIT_LIST_HEAD(&dc->delayed_bios);
	202	mutex_init(&dc->timer_lock);
	203	atomic_set(&dc->may_delay, 1);
	204
	205	ti->num_flush_bios = 1;
	206	ti->num_discard_bios = 1;
	207	ti->per_io_data_size = sizeof(struct dm_delay_info);
	208	ti->private = dc;
	209	return 0;
	210
	211	bad_queue:
	212	if (dc->dev_write)
	213	dm_put_device(ti, dc->dev_write);
	214	bad_dev_read:
	215	dm_put_device(ti, dc->dev_read);
	216	bad:
	217	kfree(dc);
	218	return ret;
	219	}
	220
	221	static void delay_dtr(struct dm_target *ti)
	222	{
	223	struct delay_c *dc = ti->private;
	224
	225	if (dc->kdelayd_wq)
	226	destroy_workqueue(dc->kdelayd_wq);
	227
	228	dm_put_device(ti, dc->dev_read);
	229
	230	if (dc->dev_write)
	231	dm_put_device(ti, dc->dev_write);
	232
	233	kfree(dc);
	234	}
	235
	236	static int delay_bio(struct delay_c dc, int delay, struct bio bio)
	237	{
	238	struct dm_delay_info *delayed;
	239	unsigned long expires = 0;
	240
	241	if (!delay \|\| !atomic_read(&dc->may_delay))
	242	return DM_MAPIO_REMAPPED;
	243
	244	delayed = dm_per_bio_data(bio, sizeof(struct dm_delay_info));
	245
	246	delayed->context = dc;
	247	delayed->expires = expires = jiffies + msecs_to_jiffies(delay);
	248
	249	mutex_lock(&delayed_bios_lock);
	250
	251	if (bio_data_dir(bio) == WRITE)
	252	dc->writes++;
	253	else
	254	dc->reads++;
	255
	256	list_add_tail(&delayed->list, &dc->delayed_bios);
	257
	258	mutex_unlock(&delayed_bios_lock);
	259
	260	queue_timeout(dc, expires);
	261
	262	return DM_MAPIO_SUBMITTED;
	263	}
	264
	265	static void delay_presuspend(struct dm_target *ti)
	266	{
	267	struct delay_c *dc = ti->private;
	268
	269	atomic_set(&dc->may_delay, 0);
	270	del_timer_sync(&dc->delay_timer);
	271	flush_bios(flush_delayed_bios(dc, 1));
	272	}
	273
	274	static void delay_resume(struct dm_target *ti)
	275	{
	276	struct delay_c *dc = ti->private;
	277
	278	atomic_set(&dc->may_delay, 1);
	279	}
	280
	281	static int delay_map(struct dm_target ti, struct bio bio)
	282	{
	283	struct delay_c *dc = ti->private;
	284
	285	if ((bio_data_dir(bio) == WRITE) && (dc->dev_write)) {
	286	bio_set_dev(bio, dc->dev_write->bdev);
	287	if (bio_sectors(bio))
	288	bio->bi_iter.bi_sector = dc->start_write +
	289	dm_target_offset(ti, bio->bi_iter.bi_sector);
	290
	291	return delay_bio(dc, dc->write_delay, bio);
	292	}
	293
	294	bio_set_dev(bio, dc->dev_read->bdev);
	295	bio->bi_iter.bi_sector = dc->start_read +
	296	dm_target_offset(ti, bio->bi_iter.bi_sector);
	297
	298	return delay_bio(dc, dc->read_delay, bio);
	299	}
	300
	301	static void delay_status(struct dm_target *ti, status_type_t type,
	302	unsigned status_flags, char *result, unsigned maxlen)
	303	{
	304	struct delay_c *dc = ti->private;
	305	int sz = 0;
	306
	307	switch (type) {
	308	case STATUSTYPE_INFO:
	309	DMEMIT("%u %u", dc->reads, dc->writes);
	310	break;
	311
	312	case STATUSTYPE_TABLE:
	313	DMEMIT("%s %llu %u", dc->dev_read->name,
	314	(unsigned long long) dc->start_read,
	315	dc->read_delay);
	316	if (dc->dev_write)
	317	DMEMIT(" %s %llu %u", dc->dev_write->name,
	318	(unsigned long long) dc->start_write,
	319	dc->write_delay);
	320	break;
	321	}
	322	}
	323
	324	static int delay_iterate_devices(struct dm_target *ti,
	325	iterate_devices_callout_fn fn, void *data)
	326	{
	327	struct delay_c *dc = ti->private;
	328	int ret = 0;
	329
	330	ret = fn(ti, dc->dev_read, dc->start_read, ti->len, data);
	331	if (ret)
	332	goto out;
	333
	334	if (dc->dev_write)
	335	ret = fn(ti, dc->dev_write, dc->start_write, ti->len, data);
	336
	337	out:
	338	return ret;
	339	}
	340
	341	static struct target_type delay_target = {
	342	.name = "delay",
	343	.version = {1, 2, 1},
	344	.features = DM_TARGET_PASSES_INTEGRITY,
	345	.module = THIS_MODULE,
	346	.ctr = delay_ctr,
	347	.dtr = delay_dtr,
	348	.map = delay_map,
	349	.presuspend = delay_presuspend,
	350	.resume = delay_resume,
	351	.status = delay_status,
	352	.iterate_devices = delay_iterate_devices,
	353	};
	354
	355	static int __init dm_delay_init(void)
	356	{
	357	int r;
	358
	359	r = dm_register_target(&delay_target);
	360	if (r < 0) {
	361	DMERR("register failed %d", r);
	362	goto bad_register;
	363	}
	364
	365	return 0;
	366
	367	bad_register:
	368	return r;
	369	}
	370
	371	static void __exit dm_delay_exit(void)
	372	{
	373	dm_unregister_target(&delay_target);
	374	}
	375
	376	/* Module hooks */
	377	module_init(dm_delay_init);
	378	module_exit(dm_delay_exit);
	379
	380	MODULE_DESCRIPTION(DM_NAME " delay target");
	381	MODULE_AUTHOR("Heinz Mauelshagen <mauelshagen@redhat.com>");
	382	MODULE_LICENSE("GPL");