dm.c 59.4 KB
Newer Older
L
Linus Torvalds 已提交
1 2
/*
 * Copyright (C) 2001, 2002 Sistina Software (UK) Limited.
M
Milan Broz 已提交
3
 * Copyright (C) 2004-2008 Red Hat, Inc. All rights reserved.
L
Linus Torvalds 已提交
4 5 6 7 8
 *
 * This file is released under the GPL.
 */

#include "dm.h"
M
Mike Anderson 已提交
9
#include "dm-uevent.h"
L
Linus Torvalds 已提交
10 11 12

#include <linux/init.h>
#include <linux/module.h>
A
Arjan van de Ven 已提交
13
#include <linux/mutex.h>
L
Linus Torvalds 已提交
14 15 16 17 18 19 20
#include <linux/moduleparam.h>
#include <linux/blkpg.h>
#include <linux/bio.h>
#include <linux/buffer_head.h>
#include <linux/mempool.h>
#include <linux/slab.h>
#include <linux/idr.h>
D
Darrick J. Wong 已提交
21
#include <linux/hdreg.h>
22 23

#include <trace/events/block.h>
L
Linus Torvalds 已提交
24

25 26
#define DM_MSG_PREFIX "core"

M
Milan Broz 已提交
27 28 29 30 31 32 33
/*
 * Cookies are numeric values sent with CHANGE and REMOVE
 * uevents while resuming, removing or renaming the device.
 */
#define DM_COOKIE_ENV_VAR_NAME "DM_COOKIE"
#define DM_COOKIE_LENGTH 24

L
Linus Torvalds 已提交
34 35 36 37 38
static const char *_name = DM_NAME;

static unsigned int major = 0;
static unsigned int _major = 0;

39
static DEFINE_SPINLOCK(_minor_lock);
L
Linus Torvalds 已提交
40
/*
K
Kiyoshi Ueda 已提交
41
 * For bio-based dm.
L
Linus Torvalds 已提交
42 43 44 45 46 47
 * One of these is allocated per bio.
 */
struct dm_io {
	struct mapped_device *md;
	int error;
	atomic_t io_count;
R
Richard Kennedy 已提交
48
	struct bio *bio;
49
	unsigned long start_time;
L
Linus Torvalds 已提交
50 51 52
};

/*
K
Kiyoshi Ueda 已提交
53
 * For bio-based dm.
L
Linus Torvalds 已提交
54 55 56
 * One of these is allocated per target within a bio.  Hopefully
 * this will be simplified out one day.
 */
A
Alasdair G Kergon 已提交
57
struct dm_target_io {
L
Linus Torvalds 已提交
58 59 60 61 62
	struct dm_io *io;
	struct dm_target *ti;
	union map_info info;
};

K
Kiyoshi Ueda 已提交
63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80
/*
 * For request-based dm.
 * One of these is allocated per request.
 */
struct dm_rq_target_io {
	struct mapped_device *md;
	struct dm_target *ti;
	struct request *orig, clone;
	int error;
	union map_info info;
};

/*
 * For request-based dm.
 * One of these is allocated per bio.
 */
struct dm_rq_clone_bio_info {
	struct bio *orig;
81
	struct dm_rq_target_io *tio;
K
Kiyoshi Ueda 已提交
82 83
};

L
Linus Torvalds 已提交
84 85
union map_info *dm_get_mapinfo(struct bio *bio)
{
A
Alasdair G Kergon 已提交
86
	if (bio && bio->bi_private)
A
Alasdair G Kergon 已提交
87
		return &((struct dm_target_io *)bio->bi_private)->info;
A
Alasdair G Kergon 已提交
88
	return NULL;
L
Linus Torvalds 已提交
89 90
}

91 92 93 94 95 96 97 98
union map_info *dm_get_rq_mapinfo(struct request *rq)
{
	if (rq && rq->end_io_data)
		return &((struct dm_rq_target_io *)rq->end_io_data)->info;
	return NULL;
}
EXPORT_SYMBOL_GPL(dm_get_rq_mapinfo);

99 100
#define MINOR_ALLOCED ((void *)-1)

L
Linus Torvalds 已提交
101 102 103
/*
 * Bits for the md->flags field.
 */
104
#define DMF_BLOCK_IO_FOR_SUSPEND 0
L
Linus Torvalds 已提交
105
#define DMF_SUSPENDED 1
106
#define DMF_FROZEN 2
J
Jeff Mahoney 已提交
107
#define DMF_FREEING 3
108
#define DMF_DELETING 4
109
#define DMF_NOFLUSH_SUSPENDING 5
110
#define DMF_QUEUE_IO_TO_THREAD 6
L
Linus Torvalds 已提交
111

112 113 114
/*
 * Work processed by per-device workqueue.
 */
L
Linus Torvalds 已提交
115
struct mapped_device {
116
	struct rw_semaphore io_lock;
117
	struct mutex suspend_lock;
L
Linus Torvalds 已提交
118 119
	rwlock_t map_lock;
	atomic_t holders;
120
	atomic_t open_count;
L
Linus Torvalds 已提交
121 122 123

	unsigned long flags;

124
	struct request_queue *queue;
L
Linus Torvalds 已提交
125
	struct gendisk *disk;
M
Mike Anderson 已提交
126
	char name[16];
L
Linus Torvalds 已提交
127 128 129 130 131 132

	void *interface_ptr;

	/*
	 * A list of ios that arrived while we were suspended.
	 */
133
	atomic_t pending[2];
L
Linus Torvalds 已提交
134
	wait_queue_head_t wait;
135
	struct work_struct work;
K
Kiyoshi Ueda 已提交
136
	struct bio_list deferred;
137
	spinlock_t deferred_lock;
L
Linus Torvalds 已提交
138

139 140 141 142 143
	/*
	 * An error from the barrier request currently being processed.
	 */
	int barrier_error;

144 145 146 147 148
	/*
	 * Processing queue (flush/barriers)
	 */
	struct workqueue_struct *wq;

L
Linus Torvalds 已提交
149 150 151 152 153 154 155 156 157 158 159
	/*
	 * The current mapping.
	 */
	struct dm_table *map;

	/*
	 * io objects are allocated from here.
	 */
	mempool_t *io_pool;
	mempool_t *tio_pool;

S
Stefan Bader 已提交
160 161
	struct bio_set *bs;

L
Linus Torvalds 已提交
162 163 164 165 166
	/*
	 * Event handling.
	 */
	atomic_t event_nr;
	wait_queue_head_t eventq;
M
Mike Anderson 已提交
167 168 169
	atomic_t uevent_seq;
	struct list_head uevent_list;
	spinlock_t uevent_lock; /* Protect access to uevent_list */
L
Linus Torvalds 已提交
170 171 172 173 174

	/*
	 * freeze/thaw support require holding onto a super block
	 */
	struct super_block *frozen_sb;
175
	struct block_device *bdev;
D
Darrick J. Wong 已提交
176 177 178

	/* forced geometry settings */
	struct hd_geometry geometry;
M
Milan Broz 已提交
179

180 181 182 183 184 185
	/* marker of flush suspend for request-based dm */
	struct request suspend_rq;

	/* For saving the address of __make_request for request based dm */
	make_request_fn *saved_make_request_fn;

M
Milan Broz 已提交
186 187
	/* sysfs handle */
	struct kobject kobj;
188 189 190

	/* zero-length barrier that will be cloned and submitted to targets */
	struct bio barrier_bio;
L
Linus Torvalds 已提交
191 192
};

K
Kiyoshi Ueda 已提交
193 194 195 196 197 198 199 200 201
/*
 * For mempools pre-allocation at the table loading time.
 */
struct dm_md_mempools {
	mempool_t *io_pool;
	mempool_t *tio_pool;
	struct bio_set *bs;
};

L
Linus Torvalds 已提交
202
#define MIN_IOS 256
203 204
static struct kmem_cache *_io_cache;
static struct kmem_cache *_tio_cache;
K
Kiyoshi Ueda 已提交
205 206
static struct kmem_cache *_rq_tio_cache;
static struct kmem_cache *_rq_bio_info_cache;
L
Linus Torvalds 已提交
207 208 209

static int __init local_init(void)
{
K
Kiyoshi Ueda 已提交
210
	int r = -ENOMEM;
L
Linus Torvalds 已提交
211 212

	/* allocate a slab for the dm_ios */
A
Alasdair G Kergon 已提交
213
	_io_cache = KMEM_CACHE(dm_io, 0);
L
Linus Torvalds 已提交
214
	if (!_io_cache)
K
Kiyoshi Ueda 已提交
215
		return r;
L
Linus Torvalds 已提交
216 217

	/* allocate a slab for the target ios */
A
Alasdair G Kergon 已提交
218
	_tio_cache = KMEM_CACHE(dm_target_io, 0);
K
Kiyoshi Ueda 已提交
219 220
	if (!_tio_cache)
		goto out_free_io_cache;
L
Linus Torvalds 已提交
221

K
Kiyoshi Ueda 已提交
222 223 224 225 226 227 228 229
	_rq_tio_cache = KMEM_CACHE(dm_rq_target_io, 0);
	if (!_rq_tio_cache)
		goto out_free_tio_cache;

	_rq_bio_info_cache = KMEM_CACHE(dm_rq_clone_bio_info, 0);
	if (!_rq_bio_info_cache)
		goto out_free_rq_tio_cache;

M
Mike Anderson 已提交
230
	r = dm_uevent_init();
K
Kiyoshi Ueda 已提交
231
	if (r)
K
Kiyoshi Ueda 已提交
232
		goto out_free_rq_bio_info_cache;
M
Mike Anderson 已提交
233

L
Linus Torvalds 已提交
234 235
	_major = major;
	r = register_blkdev(_major, _name);
K
Kiyoshi Ueda 已提交
236 237
	if (r < 0)
		goto out_uevent_exit;
L
Linus Torvalds 已提交
238 239 240 241 242

	if (!_major)
		_major = r;

	return 0;
K
Kiyoshi Ueda 已提交
243 244 245

out_uevent_exit:
	dm_uevent_exit();
K
Kiyoshi Ueda 已提交
246 247 248 249
out_free_rq_bio_info_cache:
	kmem_cache_destroy(_rq_bio_info_cache);
out_free_rq_tio_cache:
	kmem_cache_destroy(_rq_tio_cache);
K
Kiyoshi Ueda 已提交
250 251 252 253 254 255
out_free_tio_cache:
	kmem_cache_destroy(_tio_cache);
out_free_io_cache:
	kmem_cache_destroy(_io_cache);

	return r;
L
Linus Torvalds 已提交
256 257 258 259
}

static void local_exit(void)
{
K
Kiyoshi Ueda 已提交
260 261
	kmem_cache_destroy(_rq_bio_info_cache);
	kmem_cache_destroy(_rq_tio_cache);
L
Linus Torvalds 已提交
262 263
	kmem_cache_destroy(_tio_cache);
	kmem_cache_destroy(_io_cache);
264
	unregister_blkdev(_major, _name);
M
Mike Anderson 已提交
265
	dm_uevent_exit();
L
Linus Torvalds 已提交
266 267 268 269 270 271

	_major = 0;

	DMINFO("cleaned up");
}

272
static int (*_inits[])(void) __initdata = {
L
Linus Torvalds 已提交
273 274 275 276
	local_init,
	dm_target_init,
	dm_linear_init,
	dm_stripe_init,
277
	dm_kcopyd_init,
L
Linus Torvalds 已提交
278 279 280
	dm_interface_init,
};

281
static void (*_exits[])(void) = {
L
Linus Torvalds 已提交
282 283 284 285
	local_exit,
	dm_target_exit,
	dm_linear_exit,
	dm_stripe_exit,
286
	dm_kcopyd_exit,
L
Linus Torvalds 已提交
287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321
	dm_interface_exit,
};

static int __init dm_init(void)
{
	const int count = ARRAY_SIZE(_inits);

	int r, i;

	for (i = 0; i < count; i++) {
		r = _inits[i]();
		if (r)
			goto bad;
	}

	return 0;

      bad:
	while (i--)
		_exits[i]();

	return r;
}

static void __exit dm_exit(void)
{
	int i = ARRAY_SIZE(_exits);

	while (i--)
		_exits[i]();
}

/*
 * Block device functions
 */
A
Al Viro 已提交
322
static int dm_blk_open(struct block_device *bdev, fmode_t mode)
L
Linus Torvalds 已提交
323 324 325
{
	struct mapped_device *md;

J
Jeff Mahoney 已提交
326 327
	spin_lock(&_minor_lock);

A
Al Viro 已提交
328
	md = bdev->bd_disk->private_data;
J
Jeff Mahoney 已提交
329 330 331
	if (!md)
		goto out;

332 333
	if (test_bit(DMF_FREEING, &md->flags) ||
	    test_bit(DMF_DELETING, &md->flags)) {
J
Jeff Mahoney 已提交
334 335 336 337
		md = NULL;
		goto out;
	}

L
Linus Torvalds 已提交
338
	dm_get(md);
339
	atomic_inc(&md->open_count);
J
Jeff Mahoney 已提交
340 341 342 343 344

out:
	spin_unlock(&_minor_lock);

	return md ? 0 : -ENXIO;
L
Linus Torvalds 已提交
345 346
}

A
Al Viro 已提交
347
static int dm_blk_close(struct gendisk *disk, fmode_t mode)
L
Linus Torvalds 已提交
348
{
A
Al Viro 已提交
349
	struct mapped_device *md = disk->private_data;
350
	atomic_dec(&md->open_count);
L
Linus Torvalds 已提交
351 352 353 354
	dm_put(md);
	return 0;
}

355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378
int dm_open_count(struct mapped_device *md)
{
	return atomic_read(&md->open_count);
}

/*
 * Guarantees nothing is using the device before it's deleted.
 */
int dm_lock_for_deletion(struct mapped_device *md)
{
	int r = 0;

	spin_lock(&_minor_lock);

	if (dm_open_count(md))
		r = -EBUSY;
	else
		set_bit(DMF_DELETING, &md->flags);

	spin_unlock(&_minor_lock);

	return r;
}

D
Darrick J. Wong 已提交
379 380 381 382 383 384 385
static int dm_blk_getgeo(struct block_device *bdev, struct hd_geometry *geo)
{
	struct mapped_device *md = bdev->bd_disk->private_data;

	return dm_get_geometry(md, geo);
}

A
Al Viro 已提交
386
static int dm_blk_ioctl(struct block_device *bdev, fmode_t mode,
387 388
			unsigned int cmd, unsigned long arg)
{
A
Al Viro 已提交
389 390
	struct mapped_device *md = bdev->bd_disk->private_data;
	struct dm_table *map = dm_get_table(md);
391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408
	struct dm_target *tgt;
	int r = -ENOTTY;

	if (!map || !dm_table_get_size(map))
		goto out;

	/* We only support devices that have a single target */
	if (dm_table_get_num_targets(map) != 1)
		goto out;

	tgt = dm_table_get_target(map, 0);

	if (dm_suspended(md)) {
		r = -EAGAIN;
		goto out;
	}

	if (tgt->type->ioctl)
409
		r = tgt->type->ioctl(tgt, cmd, arg);
410 411 412 413 414 415 416

out:
	dm_table_put(map);

	return r;
}

A
Alasdair G Kergon 已提交
417
static struct dm_io *alloc_io(struct mapped_device *md)
L
Linus Torvalds 已提交
418 419 420 421
{
	return mempool_alloc(md->io_pool, GFP_NOIO);
}

A
Alasdair G Kergon 已提交
422
static void free_io(struct mapped_device *md, struct dm_io *io)
L
Linus Torvalds 已提交
423 424 425 426
{
	mempool_free(io, md->io_pool);
}

A
Alasdair G Kergon 已提交
427
static void free_tio(struct mapped_device *md, struct dm_target_io *tio)
L
Linus Torvalds 已提交
428 429 430 431
{
	mempool_free(tio, md->tio_pool);
}

432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451
static struct dm_rq_target_io *alloc_rq_tio(struct mapped_device *md)
{
	return mempool_alloc(md->tio_pool, GFP_ATOMIC);
}

static void free_rq_tio(struct dm_rq_target_io *tio)
{
	mempool_free(tio, tio->md->tio_pool);
}

static struct dm_rq_clone_bio_info *alloc_bio_info(struct mapped_device *md)
{
	return mempool_alloc(md->io_pool, GFP_ATOMIC);
}

static void free_bio_info(struct dm_rq_clone_bio_info *info)
{
	mempool_free(info, info->tio->md->io_pool);
}

452 453 454
static void start_io_acct(struct dm_io *io)
{
	struct mapped_device *md = io->md;
T
Tejun Heo 已提交
455
	int cpu;
456
	int rw = bio_data_dir(io->bio);
457 458 459

	io->start_time = jiffies;

T
Tejun Heo 已提交
460 461 462
	cpu = part_stat_lock();
	part_round_stats(cpu, &dm_disk(md)->part0);
	part_stat_unlock();
463
	dm_disk(md)->part0.in_flight[rw] = atomic_inc_return(&md->pending[rw]);
464 465
}

466
static void end_io_acct(struct dm_io *io)
467 468 469 470
{
	struct mapped_device *md = io->md;
	struct bio *bio = io->bio;
	unsigned long duration = jiffies - io->start_time;
T
Tejun Heo 已提交
471
	int pending, cpu;
472 473
	int rw = bio_data_dir(bio);

T
Tejun Heo 已提交
474 475 476 477
	cpu = part_stat_lock();
	part_round_stats(cpu, &dm_disk(md)->part0);
	part_stat_add(cpu, &dm_disk(md)->part0, ticks[rw], duration);
	part_stat_unlock();
478

479 480 481 482
	/*
	 * After this is decremented the bio must not be touched if it is
	 * a barrier.
	 */
483 484 485
	dm_disk(md)->part0.in_flight[rw] = pending =
		atomic_dec_return(&md->pending[rw]);
	pending += atomic_read(&md->pending[rw^0x1]);
486

487 488 489
	/* nudge anyone waiting on suspend queue */
	if (!pending)
		wake_up(&md->wait);
490 491
}

L
Linus Torvalds 已提交
492 493 494
/*
 * Add the bio to the list of deferred io.
 */
M
Mikulas Patocka 已提交
495
static void queue_io(struct mapped_device *md, struct bio *bio)
L
Linus Torvalds 已提交
496
{
497
	down_write(&md->io_lock);
L
Linus Torvalds 已提交
498

499
	spin_lock_irq(&md->deferred_lock);
L
Linus Torvalds 已提交
500
	bio_list_add(&md->deferred, bio);
501
	spin_unlock_irq(&md->deferred_lock);
L
Linus Torvalds 已提交
502

M
Mikulas Patocka 已提交
503 504 505
	if (!test_and_set_bit(DMF_QUEUE_IO_TO_THREAD, &md->flags))
		queue_work(md->wq, &md->work);

506
	up_write(&md->io_lock);
L
Linus Torvalds 已提交
507 508 509 510 511 512 513 514 515 516
}

/*
 * Everyone (including functions in this file), should use this
 * function to access the md->map field, and make sure they call
 * dm_table_put() when finished.
 */
struct dm_table *dm_get_table(struct mapped_device *md)
{
	struct dm_table *t;
517
	unsigned long flags;
L
Linus Torvalds 已提交
518

519
	read_lock_irqsave(&md->map_lock, flags);
L
Linus Torvalds 已提交
520 521 522
	t = md->map;
	if (t)
		dm_table_get(t);
523
	read_unlock_irqrestore(&md->map_lock, flags);
L
Linus Torvalds 已提交
524 525 526 527

	return t;
}

D
Darrick J. Wong 已提交
528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554
/*
 * Get the geometry associated with a dm device
 */
int dm_get_geometry(struct mapped_device *md, struct hd_geometry *geo)
{
	*geo = md->geometry;

	return 0;
}

/*
 * Set the geometry of a device.
 */
int dm_set_geometry(struct mapped_device *md, struct hd_geometry *geo)
{
	sector_t sz = (sector_t)geo->cylinders * geo->heads * geo->sectors;

	if (geo->start > sz) {
		DMWARN("Start sector is beyond the geometry limits.");
		return -EINVAL;
	}

	md->geometry = *geo;

	return 0;
}

L
Linus Torvalds 已提交
555 556 557 558 559 560 561 562 563
/*-----------------------------------------------------------------
 * CRUD START:
 *   A more elegant soln is in the works that uses the queue
 *   merge fn, unfortunately there are a couple of changes to
 *   the block layer that I want to make for this.  So in the
 *   interests of getting something for people to use I give
 *   you this clearly demarcated crap.
 *---------------------------------------------------------------*/

564 565 566 567 568
static int __noflush_suspending(struct mapped_device *md)
{
	return test_bit(DMF_NOFLUSH_SUSPENDING, &md->flags);
}

L
Linus Torvalds 已提交
569 570 571 572
/*
 * Decrements the number of outstanding ios that a bio has been
 * cloned into, completing the original io if necc.
 */
573
static void dec_pending(struct dm_io *io, int error)
L
Linus Torvalds 已提交
574
{
575
	unsigned long flags;
576 577 578
	int io_error;
	struct bio *bio;
	struct mapped_device *md = io->md;
579 580

	/* Push-back supersedes any I/O errors */
581
	if (error && !(io->error > 0 && __noflush_suspending(md)))
L
Linus Torvalds 已提交
582 583 584
		io->error = error;

	if (atomic_dec_and_test(&io->io_count)) {
585 586 587 588
		if (io->error == DM_ENDIO_REQUEUE) {
			/*
			 * Target requested pushing back the I/O.
			 */
589
			spin_lock_irqsave(&md->deferred_lock, flags);
590
			if (__noflush_suspending(md)) {
591
				if (!bio_rw_flagged(io->bio, BIO_RW_BARRIER))
592 593 594
					bio_list_add_head(&md->deferred,
							  io->bio);
			} else
595 596
				/* noflush suspend was interrupted. */
				io->error = -EIO;
597
			spin_unlock_irqrestore(&md->deferred_lock, flags);
598 599
		}

600 601
		io_error = io->error;
		bio = io->bio;
602

603
		if (bio_rw_flagged(bio, BIO_RW_BARRIER)) {
604 605 606 607 608
			/*
			 * There can be just one barrier request so we use
			 * a per-device variable for error reporting.
			 * Note that you can't touch the bio after end_io_acct
			 */
609
			if (!md->barrier_error && io_error != -EOPNOTSUPP)
610
				md->barrier_error = io_error;
611 612 613
			end_io_acct(io);
		} else {
			end_io_acct(io);
614

615 616
			if (io_error != DM_ENDIO_REQUEUE) {
				trace_block_bio_complete(md->queue, bio);
617

618 619
				bio_endio(bio, io_error);
			}
620
		}
621 622

		free_io(md, io);
L
Linus Torvalds 已提交
623 624 625
	}
}

626
static void clone_endio(struct bio *bio, int error)
L
Linus Torvalds 已提交
627 628
{
	int r = 0;
A
Alasdair G Kergon 已提交
629
	struct dm_target_io *tio = bio->bi_private;
630
	struct dm_io *io = tio->io;
S
Stefan Bader 已提交
631
	struct mapped_device *md = tio->io->md;
L
Linus Torvalds 已提交
632 633 634 635 636 637 638
	dm_endio_fn endio = tio->ti->type->end_io;

	if (!bio_flagged(bio, BIO_UPTODATE) && !error)
		error = -EIO;

	if (endio) {
		r = endio(tio->ti, bio, error, &tio->info);
639 640 641 642 643
		if (r < 0 || r == DM_ENDIO_REQUEUE)
			/*
			 * error and requeue request are handled
			 * in dec_pending().
			 */
L
Linus Torvalds 已提交
644
			error = r;
645 646
		else if (r == DM_ENDIO_INCOMPLETE)
			/* The target will handle the io */
647
			return;
648 649 650 651
		else if (r) {
			DMWARN("unimplemented target endio return value: %d", r);
			BUG();
		}
L
Linus Torvalds 已提交
652 653
	}

S
Stefan Bader 已提交
654 655 656 657 658 659
	/*
	 * Store md for cleanup instead of tio which is about to get freed.
	 */
	bio->bi_private = md->bs;

	free_tio(md, tio);
660 661
	bio_put(bio);
	dec_pending(io, error);
L
Linus Torvalds 已提交
662 663
}

664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742
/*
 * Partial completion handling for request-based dm
 */
static void end_clone_bio(struct bio *clone, int error)
{
	struct dm_rq_clone_bio_info *info = clone->bi_private;
	struct dm_rq_target_io *tio = info->tio;
	struct bio *bio = info->orig;
	unsigned int nr_bytes = info->orig->bi_size;

	bio_put(clone);

	if (tio->error)
		/*
		 * An error has already been detected on the request.
		 * Once error occurred, just let clone->end_io() handle
		 * the remainder.
		 */
		return;
	else if (error) {
		/*
		 * Don't notice the error to the upper layer yet.
		 * The error handling decision is made by the target driver,
		 * when the request is completed.
		 */
		tio->error = error;
		return;
	}

	/*
	 * I/O for the bio successfully completed.
	 * Notice the data completion to the upper layer.
	 */

	/*
	 * bios are processed from the head of the list.
	 * So the completing bio should always be rq->bio.
	 * If it's not, something wrong is happening.
	 */
	if (tio->orig->bio != bio)
		DMERR("bio completion is going in the middle of the request");

	/*
	 * Update the original request.
	 * Do not use blk_end_request() here, because it may complete
	 * the original request before the clone, and break the ordering.
	 */
	blk_update_request(tio->orig, 0, nr_bytes);
}

/*
 * Don't touch any member of the md after calling this function because
 * the md may be freed in dm_put() at the end of this function.
 * Or do dm_get() before calling this function and dm_put() later.
 */
static void rq_completed(struct mapped_device *md, int run_queue)
{
	int wakeup_waiters = 0;
	struct request_queue *q = md->queue;
	unsigned long flags;

	spin_lock_irqsave(q->queue_lock, flags);
	if (!queue_in_flight(q))
		wakeup_waiters = 1;
	spin_unlock_irqrestore(q->queue_lock, flags);

	/* nudge anyone waiting on suspend queue */
	if (wakeup_waiters)
		wake_up(&md->wait);

	if (run_queue)
		blk_run_queue(q);

	/*
	 * dm_put() must be at the end of this function. See the comment above
	 */
	dm_put(md);
}

743 744 745 746 747 748 749 750
static void free_rq_clone(struct request *clone)
{
	struct dm_rq_target_io *tio = clone->end_io_data;

	blk_rq_unprep_clone(clone);
	free_rq_tio(tio);
}

751 752 753 754 755 756 757
static void dm_unprep_request(struct request *rq)
{
	struct request *clone = rq->special;

	rq->special = NULL;
	rq->cmd_flags &= ~REQ_DONTPREP;

758
	free_rq_clone(clone);
759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835
}

/*
 * Requeue the original request of a clone.
 */
void dm_requeue_unmapped_request(struct request *clone)
{
	struct dm_rq_target_io *tio = clone->end_io_data;
	struct mapped_device *md = tio->md;
	struct request *rq = tio->orig;
	struct request_queue *q = rq->q;
	unsigned long flags;

	dm_unprep_request(rq);

	spin_lock_irqsave(q->queue_lock, flags);
	if (elv_queue_empty(q))
		blk_plug_device(q);
	blk_requeue_request(q, rq);
	spin_unlock_irqrestore(q->queue_lock, flags);

	rq_completed(md, 0);
}
EXPORT_SYMBOL_GPL(dm_requeue_unmapped_request);

static void __stop_queue(struct request_queue *q)
{
	blk_stop_queue(q);
}

static void stop_queue(struct request_queue *q)
{
	unsigned long flags;

	spin_lock_irqsave(q->queue_lock, flags);
	__stop_queue(q);
	spin_unlock_irqrestore(q->queue_lock, flags);
}

static void __start_queue(struct request_queue *q)
{
	if (blk_queue_stopped(q))
		blk_start_queue(q);
}

static void start_queue(struct request_queue *q)
{
	unsigned long flags;

	spin_lock_irqsave(q->queue_lock, flags);
	__start_queue(q);
	spin_unlock_irqrestore(q->queue_lock, flags);
}

/*
 * Complete the clone and the original request.
 * Must be called without queue lock.
 */
static void dm_end_request(struct request *clone, int error)
{
	struct dm_rq_target_io *tio = clone->end_io_data;
	struct mapped_device *md = tio->md;
	struct request *rq = tio->orig;

	if (blk_pc_request(rq)) {
		rq->errors = clone->errors;
		rq->resid_len = clone->resid_len;

		if (rq->sense)
			/*
			 * We are using the sense buffer of the original
			 * request.
			 * So setting the length of the sense data is enough.
			 */
			rq->sense_len = clone->sense_len;
	}

836
	free_rq_clone(clone);
837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924

	blk_end_request_all(rq, error);

	rq_completed(md, 1);
}

/*
 * Request completion handler for request-based dm
 */
static void dm_softirq_done(struct request *rq)
{
	struct request *clone = rq->completion_data;
	struct dm_rq_target_io *tio = clone->end_io_data;
	dm_request_endio_fn rq_end_io = tio->ti->type->rq_end_io;
	int error = tio->error;

	if (!(rq->cmd_flags & REQ_FAILED) && rq_end_io)
		error = rq_end_io(tio->ti, clone, error, &tio->info);

	if (error <= 0)
		/* The target wants to complete the I/O */
		dm_end_request(clone, error);
	else if (error == DM_ENDIO_INCOMPLETE)
		/* The target will handle the I/O */
		return;
	else if (error == DM_ENDIO_REQUEUE)
		/* The target wants to requeue the I/O */
		dm_requeue_unmapped_request(clone);
	else {
		DMWARN("unimplemented target endio return value: %d", error);
		BUG();
	}
}

/*
 * Complete the clone and the original request with the error status
 * through softirq context.
 */
static void dm_complete_request(struct request *clone, int error)
{
	struct dm_rq_target_io *tio = clone->end_io_data;
	struct request *rq = tio->orig;

	tio->error = error;
	rq->completion_data = clone;
	blk_complete_request(rq);
}

/*
 * Complete the not-mapped clone and the original request with the error status
 * through softirq context.
 * Target's rq_end_io() function isn't called.
 * This may be used when the target's map_rq() function fails.
 */
void dm_kill_unmapped_request(struct request *clone, int error)
{
	struct dm_rq_target_io *tio = clone->end_io_data;
	struct request *rq = tio->orig;

	rq->cmd_flags |= REQ_FAILED;
	dm_complete_request(clone, error);
}
EXPORT_SYMBOL_GPL(dm_kill_unmapped_request);

/*
 * Called with the queue lock held
 */
static void end_clone_request(struct request *clone, int error)
{
	/*
	 * For just cleaning up the information of the queue in which
	 * the clone was dispatched.
	 * The clone is *NOT* freed actually here because it is alloced from
	 * dm own mempool and REQ_ALLOCED isn't set in clone->cmd_flags.
	 */
	__blk_put_request(clone->q, clone);

	/*
	 * Actual request completion is done in a softirq context which doesn't
	 * hold the queue lock.  Otherwise, deadlock could occur because:
	 *     - another request may be submitted by the upper level driver
	 *       of the stacking during the completion
	 *     - the submission which requires queue lock may be done
	 *       against this queue
	 */
	dm_complete_request(clone, error);
}

L
Linus Torvalds 已提交
925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945
static sector_t max_io_len(struct mapped_device *md,
			   sector_t sector, struct dm_target *ti)
{
	sector_t offset = sector - ti->begin;
	sector_t len = ti->len - offset;

	/*
	 * Does the target need to split even further ?
	 */
	if (ti->split_io) {
		sector_t boundary;
		boundary = ((offset + ti->split_io) & ~(ti->split_io - 1))
			   - offset;
		if (len > boundary)
			len = boundary;
	}

	return len;
}

static void __map_bio(struct dm_target *ti, struct bio *clone,
A
Alasdair G Kergon 已提交
946
		      struct dm_target_io *tio)
L
Linus Torvalds 已提交
947 948
{
	int r;
949
	sector_t sector;
S
Stefan Bader 已提交
950
	struct mapped_device *md;
L
Linus Torvalds 已提交
951 952 953 954 955 956 957 958 959 960

	clone->bi_end_io = clone_endio;
	clone->bi_private = tio;

	/*
	 * Map the clone.  If r == 0 we don't need to do
	 * anything, the target has assumed ownership of
	 * this io.
	 */
	atomic_inc(&tio->io->io_count);
961
	sector = clone->bi_sector;
L
Linus Torvalds 已提交
962
	r = ti->type->map(ti, clone, &tio->info);
963
	if (r == DM_MAPIO_REMAPPED) {
L
Linus Torvalds 已提交
964
		/* the bio has been remapped so dispatch it */
965

966
		trace_block_remap(bdev_get_queue(clone->bi_bdev), clone,
967
				    tio->io->bio->bi_bdev->bd_dev, sector);
968

L
Linus Torvalds 已提交
969
		generic_make_request(clone);
970 971
	} else if (r < 0 || r == DM_MAPIO_REQUEUE) {
		/* error the io and bail out, or requeue it if needed */
S
Stefan Bader 已提交
972 973 974 975 976 977
		md = tio->io->md;
		dec_pending(tio->io, r);
		/*
		 * Store bio_set for cleanup.
		 */
		clone->bi_private = md->bs;
L
Linus Torvalds 已提交
978
		bio_put(clone);
S
Stefan Bader 已提交
979
		free_tio(md, tio);
980 981 982
	} else if (r) {
		DMWARN("unimplemented target map return value: %d", r);
		BUG();
L
Linus Torvalds 已提交
983 984 985 986 987 988 989 990 991 992 993 994 995
	}
}

struct clone_info {
	struct mapped_device *md;
	struct dm_table *map;
	struct bio *bio;
	struct dm_io *io;
	sector_t sector;
	sector_t sector_count;
	unsigned short idx;
};

P
Peter Osterlund 已提交
996 997
static void dm_bio_destructor(struct bio *bio)
{
S
Stefan Bader 已提交
998 999 1000
	struct bio_set *bs = bio->bi_private;

	bio_free(bio, bs);
P
Peter Osterlund 已提交
1001 1002
}

L
Linus Torvalds 已提交
1003 1004 1005 1006 1007
/*
 * Creates a little bio that is just does part of a bvec.
 */
static struct bio *split_bvec(struct bio *bio, sector_t sector,
			      unsigned short idx, unsigned int offset,
S
Stefan Bader 已提交
1008
			      unsigned int len, struct bio_set *bs)
L
Linus Torvalds 已提交
1009 1010 1011 1012
{
	struct bio *clone;
	struct bio_vec *bv = bio->bi_io_vec + idx;

S
Stefan Bader 已提交
1013
	clone = bio_alloc_bioset(GFP_NOIO, 1, bs);
P
Peter Osterlund 已提交
1014
	clone->bi_destructor = dm_bio_destructor;
L
Linus Torvalds 已提交
1015 1016 1017 1018
	*clone->bi_io_vec = *bv;

	clone->bi_sector = sector;
	clone->bi_bdev = bio->bi_bdev;
1019
	clone->bi_rw = bio->bi_rw & ~(1 << BIO_RW_BARRIER);
L
Linus Torvalds 已提交
1020 1021 1022 1023
	clone->bi_vcnt = 1;
	clone->bi_size = to_bytes(len);
	clone->bi_io_vec->bv_offset = offset;
	clone->bi_io_vec->bv_len = clone->bi_size;
M
Martin K. Petersen 已提交
1024
	clone->bi_flags |= 1 << BIO_CLONED;
L
Linus Torvalds 已提交
1025

M
Martin K. Petersen 已提交
1026
	if (bio_integrity(bio)) {
1027
		bio_integrity_clone(clone, bio, GFP_NOIO, bs);
M
Martin K. Petersen 已提交
1028 1029 1030 1031
		bio_integrity_trim(clone,
				   bio_sector_offset(bio, idx, offset), len);
	}

L
Linus Torvalds 已提交
1032 1033 1034 1035 1036 1037 1038 1039
	return clone;
}

/*
 * Creates a bio that consists of range of complete bvecs.
 */
static struct bio *clone_bio(struct bio *bio, sector_t sector,
			     unsigned short idx, unsigned short bv_count,
S
Stefan Bader 已提交
1040
			     unsigned int len, struct bio_set *bs)
L
Linus Torvalds 已提交
1041 1042 1043
{
	struct bio *clone;

S
Stefan Bader 已提交
1044 1045
	clone = bio_alloc_bioset(GFP_NOIO, bio->bi_max_vecs, bs);
	__bio_clone(clone, bio);
1046
	clone->bi_rw &= ~(1 << BIO_RW_BARRIER);
S
Stefan Bader 已提交
1047
	clone->bi_destructor = dm_bio_destructor;
L
Linus Torvalds 已提交
1048 1049 1050 1051 1052 1053
	clone->bi_sector = sector;
	clone->bi_idx = idx;
	clone->bi_vcnt = idx + bv_count;
	clone->bi_size = to_bytes(len);
	clone->bi_flags &= ~(1 << BIO_SEG_VALID);

M
Martin K. Petersen 已提交
1054
	if (bio_integrity(bio)) {
1055
		bio_integrity_clone(clone, bio, GFP_NOIO, bs);
M
Martin K. Petersen 已提交
1056 1057 1058 1059 1060 1061

		if (idx != bio->bi_idx || clone->bi_size < bio->bi_size)
			bio_integrity_trim(clone,
					   bio_sector_offset(bio, idx, 0), len);
	}

L
Linus Torvalds 已提交
1062 1063 1064
	return clone;
}

1065 1066
static struct dm_target_io *alloc_tio(struct clone_info *ci,
				      struct dm_target *ti)
1067
{
1068
	struct dm_target_io *tio = mempool_alloc(ci->md->tio_pool, GFP_NOIO);
1069 1070 1071 1072

	tio->io = ci->io;
	tio->ti = ti;
	memset(&tio->info, 0, sizeof(tio->info));
1073 1074 1075 1076 1077 1078 1079 1080 1081 1082

	return tio;
}

static void __flush_target(struct clone_info *ci, struct dm_target *ti,
			  unsigned flush_nr)
{
	struct dm_target_io *tio = alloc_tio(ci, ti);
	struct bio *clone;

1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106
	tio->info.flush_request = flush_nr;

	clone = bio_alloc_bioset(GFP_NOIO, 0, ci->md->bs);
	__bio_clone(clone, ci->bio);
	clone->bi_destructor = dm_bio_destructor;

	__map_bio(ti, clone, tio);
}

static int __clone_and_map_empty_barrier(struct clone_info *ci)
{
	unsigned target_nr = 0, flush_nr;
	struct dm_target *ti;

	while ((ti = dm_table_get_target(ci->map, target_nr++)))
		for (flush_nr = 0; flush_nr < ti->num_flush_requests;
		     flush_nr++)
			__flush_target(ci, ti, flush_nr);

	ci->sector_count = 0;

	return 0;
}

1107
static int __clone_and_map(struct clone_info *ci)
L
Linus Torvalds 已提交
1108 1109
{
	struct bio *clone, *bio = ci->bio;
1110 1111
	struct dm_target *ti;
	sector_t len = 0, max;
A
Alasdair G Kergon 已提交
1112
	struct dm_target_io *tio;
L
Linus Torvalds 已提交
1113

1114 1115 1116
	if (unlikely(bio_empty_barrier(bio)))
		return __clone_and_map_empty_barrier(ci);

1117 1118 1119 1120 1121 1122
	ti = dm_table_find_target(ci->map, ci->sector);
	if (!dm_target_is_valid(ti))
		return -EIO;

	max = max_io_len(ci->md, ci->sector, ti);

L
Linus Torvalds 已提交
1123 1124 1125
	/*
	 * Allocate a target io object.
	 */
1126
	tio = alloc_tio(ci, ti);
L
Linus Torvalds 已提交
1127 1128 1129 1130 1131 1132 1133

	if (ci->sector_count <= max) {
		/*
		 * Optimise for the simple case where we can do all of
		 * the remaining io with a single clone.
		 */
		clone = clone_bio(bio, ci->sector, ci->idx,
S
Stefan Bader 已提交
1134 1135
				  bio->bi_vcnt - ci->idx, ci->sector_count,
				  ci->md->bs);
L
Linus Torvalds 已提交
1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157
		__map_bio(ti, clone, tio);
		ci->sector_count = 0;

	} else if (to_sector(bio->bi_io_vec[ci->idx].bv_len) <= max) {
		/*
		 * There are some bvecs that don't span targets.
		 * Do as many of these as possible.
		 */
		int i;
		sector_t remaining = max;
		sector_t bv_len;

		for (i = ci->idx; remaining && (i < bio->bi_vcnt); i++) {
			bv_len = to_sector(bio->bi_io_vec[i].bv_len);

			if (bv_len > remaining)
				break;

			remaining -= bv_len;
			len += bv_len;
		}

S
Stefan Bader 已提交
1158 1159
		clone = clone_bio(bio, ci->sector, ci->idx, i - ci->idx, len,
				  ci->md->bs);
L
Linus Torvalds 已提交
1160 1161 1162 1163 1164 1165 1166 1167
		__map_bio(ti, clone, tio);

		ci->sector += len;
		ci->sector_count -= len;
		ci->idx = i;

	} else {
		/*
1168
		 * Handle a bvec that must be split between two or more targets.
L
Linus Torvalds 已提交
1169 1170
		 */
		struct bio_vec *bv = bio->bi_io_vec + ci->idx;
1171 1172
		sector_t remaining = to_sector(bv->bv_len);
		unsigned int offset = 0;
L
Linus Torvalds 已提交
1173

1174 1175 1176
		do {
			if (offset) {
				ti = dm_table_find_target(ci->map, ci->sector);
1177 1178 1179
				if (!dm_target_is_valid(ti))
					return -EIO;

1180
				max = max_io_len(ci->md, ci->sector, ti);
L
Linus Torvalds 已提交
1181

1182
				tio = alloc_tio(ci, ti);
1183 1184 1185 1186 1187
			}

			len = min(remaining, max);

			clone = split_bvec(bio, ci->sector, ci->idx,
S
Stefan Bader 已提交
1188 1189
					   bv->bv_offset + offset, len,
					   ci->md->bs);
1190 1191 1192 1193 1194 1195 1196

			__map_bio(ti, clone, tio);

			ci->sector += len;
			ci->sector_count -= len;
			offset += to_bytes(len);
		} while (remaining -= len);
L
Linus Torvalds 已提交
1197 1198 1199

		ci->idx++;
	}
1200 1201

	return 0;
L
Linus Torvalds 已提交
1202 1203 1204
}

/*
M
Mikulas Patocka 已提交
1205
 * Split the bio into several clones and submit it to targets.
L
Linus Torvalds 已提交
1206
 */
1207
static void __split_and_process_bio(struct mapped_device *md, struct bio *bio)
L
Linus Torvalds 已提交
1208 1209
{
	struct clone_info ci;
1210
	int error = 0;
L
Linus Torvalds 已提交
1211 1212

	ci.map = dm_get_table(md);
1213
	if (unlikely(!ci.map)) {
1214
		if (!bio_rw_flagged(bio, BIO_RW_BARRIER))
1215 1216
			bio_io_error(bio);
		else
1217 1218
			if (!md->barrier_error)
				md->barrier_error = -EIO;
1219 1220
		return;
	}
1221

L
Linus Torvalds 已提交
1222 1223 1224 1225 1226 1227 1228 1229 1230
	ci.md = md;
	ci.bio = bio;
	ci.io = alloc_io(md);
	ci.io->error = 0;
	atomic_set(&ci.io->io_count, 1);
	ci.io->bio = bio;
	ci.io->md = md;
	ci.sector = bio->bi_sector;
	ci.sector_count = bio_sectors(bio);
1231 1232
	if (unlikely(bio_empty_barrier(bio)))
		ci.sector_count = 1;
L
Linus Torvalds 已提交
1233 1234
	ci.idx = bio->bi_idx;

1235
	start_io_acct(ci.io);
1236 1237
	while (ci.sector_count && !error)
		error = __clone_and_map(&ci);
L
Linus Torvalds 已提交
1238 1239

	/* drop the extra reference count */
1240
	dec_pending(ci.io, error);
L
Linus Torvalds 已提交
1241 1242 1243 1244 1245 1246
	dm_table_put(ci.map);
}
/*-----------------------------------------------------------------
 * CRUD END
 *---------------------------------------------------------------*/

M
Milan Broz 已提交
1247 1248 1249 1250 1251 1252 1253 1254
static int dm_merge_bvec(struct request_queue *q,
			 struct bvec_merge_data *bvm,
			 struct bio_vec *biovec)
{
	struct mapped_device *md = q->queuedata;
	struct dm_table *map = dm_get_table(md);
	struct dm_target *ti;
	sector_t max_sectors;
1255
	int max_size = 0;
M
Milan Broz 已提交
1256 1257

	if (unlikely(!map))
1258
		goto out;
M
Milan Broz 已提交
1259 1260

	ti = dm_table_find_target(map, bvm->bi_sector);
1261 1262
	if (!dm_target_is_valid(ti))
		goto out_table;
M
Milan Broz 已提交
1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279

	/*
	 * Find maximum amount of I/O that won't need splitting
	 */
	max_sectors = min(max_io_len(md, bvm->bi_sector, ti),
			  (sector_t) BIO_MAX_SECTORS);
	max_size = (max_sectors << SECTOR_SHIFT) - bvm->bi_size;
	if (max_size < 0)
		max_size = 0;

	/*
	 * merge_bvec_fn() returns number of bytes
	 * it can accept at this offset
	 * max is precomputed maximal io size
	 */
	if (max_size && ti->type->merge)
		max_size = ti->type->merge(ti, bvm, biovec, max_size);
1280 1281 1282 1283 1284 1285 1286 1287 1288 1289
	/*
	 * If the target doesn't support merge method and some of the devices
	 * provided their merge_bvec method (we know this by looking at
	 * queue_max_hw_sectors), then we can't allow bios with multiple vector
	 * entries.  So always set max_size to 0, and the code below allows
	 * just one page.
	 */
	else if (queue_max_hw_sectors(q) <= PAGE_SIZE >> 9)

		max_size = 0;
M
Milan Broz 已提交
1290

1291
out_table:
1292 1293 1294
	dm_table_put(map);

out:
M
Milan Broz 已提交
1295 1296 1297 1298 1299 1300 1301 1302 1303
	/*
	 * Always allow an entire first page
	 */
	if (max_size <= biovec->bv_len && !(bvm->bi_size >> SECTOR_SHIFT))
		max_size = biovec->bv_len;

	return max_size;
}

L
Linus Torvalds 已提交
1304 1305 1306 1307
/*
 * The request function that just remaps the bio built up by
 * dm_merge_bvec.
 */
1308
static int _dm_request(struct request_queue *q, struct bio *bio)
L
Linus Torvalds 已提交
1309
{
1310
	int rw = bio_data_dir(bio);
L
Linus Torvalds 已提交
1311
	struct mapped_device *md = q->queuedata;
T
Tejun Heo 已提交
1312
	int cpu;
L
Linus Torvalds 已提交
1313

1314
	down_read(&md->io_lock);
L
Linus Torvalds 已提交
1315

T
Tejun Heo 已提交
1316 1317 1318 1319
	cpu = part_stat_lock();
	part_stat_inc(cpu, &dm_disk(md)->part0, ios[rw]);
	part_stat_add(cpu, &dm_disk(md)->part0, sectors[rw], bio_sectors(bio));
	part_stat_unlock();
1320

L
Linus Torvalds 已提交
1321
	/*
1322 1323
	 * If we're suspended or the thread is processing barriers
	 * we have to queue this io for later.
L
Linus Torvalds 已提交
1324
	 */
1325
	if (unlikely(test_bit(DMF_QUEUE_IO_TO_THREAD, &md->flags)) ||
1326
	    unlikely(bio_rw_flagged(bio, BIO_RW_BARRIER))) {
1327
		up_read(&md->io_lock);
L
Linus Torvalds 已提交
1328

A
Alasdair G Kergon 已提交
1329 1330 1331 1332 1333
		if (unlikely(test_bit(DMF_BLOCK_IO_FOR_SUSPEND, &md->flags)) &&
		    bio_rw(bio) == READA) {
			bio_io_error(bio);
			return 0;
		}
L
Linus Torvalds 已提交
1334

M
Mikulas Patocka 已提交
1335
		queue_io(md, bio);
L
Linus Torvalds 已提交
1336

M
Mikulas Patocka 已提交
1337
		return 0;
L
Linus Torvalds 已提交
1338 1339
	}

1340
	__split_and_process_bio(md, bio);
1341
	up_read(&md->io_lock);
1342
	return 0;
L
Linus Torvalds 已提交
1343 1344
}

1345 1346 1347 1348
static int dm_make_request(struct request_queue *q, struct bio *bio)
{
	struct mapped_device *md = q->queuedata;

1349
	if (unlikely(bio_rw_flagged(bio, BIO_RW_BARRIER))) {
1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603
		bio_endio(bio, -EOPNOTSUPP);
		return 0;
	}

	return md->saved_make_request_fn(q, bio); /* call __make_request() */
}

static int dm_request_based(struct mapped_device *md)
{
	return blk_queue_stackable(md->queue);
}

static int dm_request(struct request_queue *q, struct bio *bio)
{
	struct mapped_device *md = q->queuedata;

	if (dm_request_based(md))
		return dm_make_request(q, bio);

	return _dm_request(q, bio);
}

void dm_dispatch_request(struct request *rq)
{
	int r;

	if (blk_queue_io_stat(rq->q))
		rq->cmd_flags |= REQ_IO_STAT;

	rq->start_time = jiffies;
	r = blk_insert_cloned_request(rq->q, rq);
	if (r)
		dm_complete_request(rq, r);
}
EXPORT_SYMBOL_GPL(dm_dispatch_request);

static void dm_rq_bio_destructor(struct bio *bio)
{
	struct dm_rq_clone_bio_info *info = bio->bi_private;
	struct mapped_device *md = info->tio->md;

	free_bio_info(info);
	bio_free(bio, md->bs);
}

static int dm_rq_bio_constructor(struct bio *bio, struct bio *bio_orig,
				 void *data)
{
	struct dm_rq_target_io *tio = data;
	struct mapped_device *md = tio->md;
	struct dm_rq_clone_bio_info *info = alloc_bio_info(md);

	if (!info)
		return -ENOMEM;

	info->orig = bio_orig;
	info->tio = tio;
	bio->bi_end_io = end_clone_bio;
	bio->bi_private = info;
	bio->bi_destructor = dm_rq_bio_destructor;

	return 0;
}

static int setup_clone(struct request *clone, struct request *rq,
		       struct dm_rq_target_io *tio)
{
	int r = blk_rq_prep_clone(clone, rq, tio->md->bs, GFP_ATOMIC,
				  dm_rq_bio_constructor, tio);

	if (r)
		return r;

	clone->cmd = rq->cmd;
	clone->cmd_len = rq->cmd_len;
	clone->sense = rq->sense;
	clone->buffer = rq->buffer;
	clone->end_io = end_clone_request;
	clone->end_io_data = tio;

	return 0;
}

static int dm_rq_flush_suspending(struct mapped_device *md)
{
	return !md->suspend_rq.special;
}

/*
 * Called with the queue lock held.
 */
static int dm_prep_fn(struct request_queue *q, struct request *rq)
{
	struct mapped_device *md = q->queuedata;
	struct dm_rq_target_io *tio;
	struct request *clone;

	if (unlikely(rq == &md->suspend_rq)) {
		if (dm_rq_flush_suspending(md))
			return BLKPREP_OK;
		else
			/* The flush suspend was interrupted */
			return BLKPREP_KILL;
	}

	if (unlikely(rq->special)) {
		DMWARN("Already has something in rq->special.");
		return BLKPREP_KILL;
	}

	tio = alloc_rq_tio(md); /* Only one for each original request */
	if (!tio)
		/* -ENOMEM */
		return BLKPREP_DEFER;

	tio->md = md;
	tio->ti = NULL;
	tio->orig = rq;
	tio->error = 0;
	memset(&tio->info, 0, sizeof(tio->info));

	clone = &tio->clone;
	if (setup_clone(clone, rq, tio)) {
		/* -ENOMEM */
		free_rq_tio(tio);
		return BLKPREP_DEFER;
	}

	rq->special = clone;
	rq->cmd_flags |= REQ_DONTPREP;

	return BLKPREP_OK;
}

static void map_request(struct dm_target *ti, struct request *rq,
			struct mapped_device *md)
{
	int r;
	struct request *clone = rq->special;
	struct dm_rq_target_io *tio = clone->end_io_data;

	/*
	 * Hold the md reference here for the in-flight I/O.
	 * We can't rely on the reference count by device opener,
	 * because the device may be closed during the request completion
	 * when all bios are completed.
	 * See the comment in rq_completed() too.
	 */
	dm_get(md);

	tio->ti = ti;
	r = ti->type->map_rq(ti, clone, &tio->info);
	switch (r) {
	case DM_MAPIO_SUBMITTED:
		/* The target has taken the I/O to submit by itself later */
		break;
	case DM_MAPIO_REMAPPED:
		/* The target has remapped the I/O so dispatch it */
		dm_dispatch_request(clone);
		break;
	case DM_MAPIO_REQUEUE:
		/* The target wants to requeue the I/O */
		dm_requeue_unmapped_request(clone);
		break;
	default:
		if (r > 0) {
			DMWARN("unimplemented target map return value: %d", r);
			BUG();
		}

		/* The target wants to complete the I/O */
		dm_kill_unmapped_request(clone, r);
		break;
	}
}

/*
 * q->request_fn for request-based dm.
 * Called with the queue lock held.
 */
static void dm_request_fn(struct request_queue *q)
{
	struct mapped_device *md = q->queuedata;
	struct dm_table *map = dm_get_table(md);
	struct dm_target *ti;
	struct request *rq;

	/*
	 * For noflush suspend, check blk_queue_stopped() to immediately
	 * quit I/O dispatching.
	 */
	while (!blk_queue_plugged(q) && !blk_queue_stopped(q)) {
		rq = blk_peek_request(q);
		if (!rq)
			goto plug_and_out;

		if (unlikely(rq == &md->suspend_rq)) { /* Flush suspend maker */
			if (queue_in_flight(q))
				/* Not quiet yet.  Wait more */
				goto plug_and_out;

			/* This device should be quiet now */
			__stop_queue(q);
			blk_start_request(rq);
			__blk_end_request_all(rq, 0);
			wake_up(&md->wait);
			goto out;
		}

		ti = dm_table_find_target(map, blk_rq_pos(rq));
		if (ti->type->busy && ti->type->busy(ti))
			goto plug_and_out;

		blk_start_request(rq);
		spin_unlock(q->queue_lock);
		map_request(ti, rq, md);
		spin_lock_irq(q->queue_lock);
	}

	goto out;

plug_and_out:
	if (!elv_queue_empty(q))
		/* Some requests still remain, retry later */
		blk_plug_device(q);

out:
	dm_table_put(map);

	return;
}

int dm_underlying_device_busy(struct request_queue *q)
{
	return blk_lld_busy(q);
}
EXPORT_SYMBOL_GPL(dm_underlying_device_busy);

static int dm_lld_busy(struct request_queue *q)
{
	int r;
	struct mapped_device *md = q->queuedata;
	struct dm_table *map = dm_get_table(md);

	if (!map || test_bit(DMF_BLOCK_IO_FOR_SUSPEND, &md->flags))
		r = 1;
	else
		r = dm_table_any_busy_target(map);

	dm_table_put(map);

	return r;
}

1604
static void dm_unplug_all(struct request_queue *q)
L
Linus Torvalds 已提交
1605 1606 1607 1608 1609
{
	struct mapped_device *md = q->queuedata;
	struct dm_table *map = dm_get_table(md);

	if (map) {
1610 1611 1612
		if (dm_request_based(md))
			generic_unplug_device(q);

L
Linus Torvalds 已提交
1613 1614 1615 1616 1617 1618 1619
		dm_table_unplug_all(map);
		dm_table_put(map);
	}
}

static int dm_any_congested(void *congested_data, int bdi_bits)
{
1620 1621 1622
	int r = bdi_bits;
	struct mapped_device *md = congested_data;
	struct dm_table *map;
L
Linus Torvalds 已提交
1623

1624
	if (!test_bit(DMF_BLOCK_IO_FOR_SUSPEND, &md->flags)) {
1625 1626
		map = dm_get_table(md);
		if (map) {
1627 1628 1629 1630 1631 1632 1633 1634 1635 1636
			/*
			 * Request-based dm cares about only own queue for
			 * the query about congestion status of request_queue
			 */
			if (dm_request_based(md))
				r = md->queue->backing_dev_info.state &
				    bdi_bits;
			else
				r = dm_table_any_congested(map, bdi_bits);

1637 1638 1639 1640
			dm_table_put(map);
		}
	}

L
Linus Torvalds 已提交
1641 1642 1643 1644 1645 1646 1647 1648
	return r;
}

/*-----------------------------------------------------------------
 * An IDR is used to keep track of allocated minor numbers.
 *---------------------------------------------------------------*/
static DEFINE_IDR(_minor_idr);

1649
static void free_minor(int minor)
L
Linus Torvalds 已提交
1650
{
1651
	spin_lock(&_minor_lock);
L
Linus Torvalds 已提交
1652
	idr_remove(&_minor_idr, minor);
1653
	spin_unlock(&_minor_lock);
L
Linus Torvalds 已提交
1654 1655 1656 1657 1658
}

/*
 * See if the device with a specific minor # is free.
 */
1659
static int specific_minor(int minor)
L
Linus Torvalds 已提交
1660 1661 1662 1663 1664 1665
{
	int r, m;

	if (minor >= (1 << MINORBITS))
		return -EINVAL;

J
Jeff Mahoney 已提交
1666 1667 1668 1669
	r = idr_pre_get(&_minor_idr, GFP_KERNEL);
	if (!r)
		return -ENOMEM;

1670
	spin_lock(&_minor_lock);
L
Linus Torvalds 已提交
1671 1672 1673 1674 1675 1676

	if (idr_find(&_minor_idr, minor)) {
		r = -EBUSY;
		goto out;
	}

1677
	r = idr_get_new_above(&_minor_idr, MINOR_ALLOCED, minor, &m);
J
Jeff Mahoney 已提交
1678
	if (r)
L
Linus Torvalds 已提交
1679 1680 1681 1682 1683 1684 1685 1686 1687
		goto out;

	if (m != minor) {
		idr_remove(&_minor_idr, m);
		r = -EBUSY;
		goto out;
	}

out:
1688
	spin_unlock(&_minor_lock);
L
Linus Torvalds 已提交
1689 1690 1691
	return r;
}

1692
static int next_free_minor(int *minor)
L
Linus Torvalds 已提交
1693
{
1694
	int r, m;
L
Linus Torvalds 已提交
1695 1696

	r = idr_pre_get(&_minor_idr, GFP_KERNEL);
J
Jeff Mahoney 已提交
1697 1698 1699
	if (!r)
		return -ENOMEM;

1700
	spin_lock(&_minor_lock);
L
Linus Torvalds 已提交
1701

1702
	r = idr_get_new(&_minor_idr, MINOR_ALLOCED, &m);
1703
	if (r)
L
Linus Torvalds 已提交
1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714
		goto out;

	if (m >= (1 << MINORBITS)) {
		idr_remove(&_minor_idr, m);
		r = -ENOSPC;
		goto out;
	}

	*minor = m;

out:
1715
	spin_unlock(&_minor_lock);
L
Linus Torvalds 已提交
1716 1717 1718
	return r;
}

1719
static const struct block_device_operations dm_blk_dops;
L
Linus Torvalds 已提交
1720

1721 1722
static void dm_wq_work(struct work_struct *work);

L
Linus Torvalds 已提交
1723 1724 1725
/*
 * Allocate and initialise a blank device with a given minor.
 */
1726
static struct mapped_device *alloc_dev(int minor)
L
Linus Torvalds 已提交
1727 1728
{
	int r;
1729
	struct mapped_device *md = kzalloc(sizeof(*md), GFP_KERNEL);
1730
	void *old_md;
L
Linus Torvalds 已提交
1731 1732 1733 1734 1735 1736

	if (!md) {
		DMWARN("unable to allocate device, out of memory.");
		return NULL;
	}

1737
	if (!try_module_get(THIS_MODULE))
M
Milan Broz 已提交
1738
		goto bad_module_get;
1739

L
Linus Torvalds 已提交
1740
	/* get a minor number for the dev */
1741
	if (minor == DM_ANY_MINOR)
1742
		r = next_free_minor(&minor);
1743
	else
1744
		r = specific_minor(minor);
L
Linus Torvalds 已提交
1745
	if (r < 0)
M
Milan Broz 已提交
1746
		goto bad_minor;
L
Linus Torvalds 已提交
1747

1748
	init_rwsem(&md->io_lock);
1749
	mutex_init(&md->suspend_lock);
1750
	spin_lock_init(&md->deferred_lock);
L
Linus Torvalds 已提交
1751 1752
	rwlock_init(&md->map_lock);
	atomic_set(&md->holders, 1);
1753
	atomic_set(&md->open_count, 0);
L
Linus Torvalds 已提交
1754
	atomic_set(&md->event_nr, 0);
M
Mike Anderson 已提交
1755 1756 1757
	atomic_set(&md->uevent_seq, 0);
	INIT_LIST_HEAD(&md->uevent_list);
	spin_lock_init(&md->uevent_lock);
L
Linus Torvalds 已提交
1758

K
Kiyoshi Ueda 已提交
1759
	md->queue = blk_init_queue(dm_request_fn, NULL);
L
Linus Torvalds 已提交
1760
	if (!md->queue)
M
Milan Broz 已提交
1761
		goto bad_queue;
L
Linus Torvalds 已提交
1762

K
Kiyoshi Ueda 已提交
1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774
	/*
	 * Request-based dm devices cannot be stacked on top of bio-based dm
	 * devices.  The type of this dm device has not been decided yet,
	 * although we initialized the queue using blk_init_queue().
	 * The type is decided at the first table loading time.
	 * To prevent problematic device stacking, clear the queue flag
	 * for request stacking support until then.
	 *
	 * This queue is new, so no concurrency on the queue_flags.
	 */
	queue_flag_clear_unlocked(QUEUE_FLAG_STACKABLE, md->queue);
	md->saved_make_request_fn = md->queue->make_request_fn;
L
Linus Torvalds 已提交
1775 1776 1777 1778
	md->queue->queuedata = md;
	md->queue->backing_dev_info.congested_fn = dm_any_congested;
	md->queue->backing_dev_info.congested_data = md;
	blk_queue_make_request(md->queue, dm_request);
1779
	blk_queue_bounce_limit(md->queue, BLK_BOUNCE_ANY);
L
Linus Torvalds 已提交
1780
	md->queue->unplug_fn = dm_unplug_all;
M
Milan Broz 已提交
1781
	blk_queue_merge_bvec(md->queue, dm_merge_bvec);
K
Kiyoshi Ueda 已提交
1782 1783 1784
	blk_queue_softirq_done(md->queue, dm_softirq_done);
	blk_queue_prep_rq(md->queue, dm_prep_fn);
	blk_queue_lld_busy(md->queue, dm_lld_busy);
S
Stefan Bader 已提交
1785

L
Linus Torvalds 已提交
1786 1787
	md->disk = alloc_disk(1);
	if (!md->disk)
M
Milan Broz 已提交
1788
		goto bad_disk;
L
Linus Torvalds 已提交
1789

1790 1791
	atomic_set(&md->pending[0], 0);
	atomic_set(&md->pending[1], 0);
1792
	init_waitqueue_head(&md->wait);
1793
	INIT_WORK(&md->work, dm_wq_work);
1794 1795
	init_waitqueue_head(&md->eventq);

L
Linus Torvalds 已提交
1796 1797 1798 1799 1800 1801 1802
	md->disk->major = _major;
	md->disk->first_minor = minor;
	md->disk->fops = &dm_blk_dops;
	md->disk->queue = md->queue;
	md->disk->private_data = md;
	sprintf(md->disk->disk_name, "dm-%d", minor);
	add_disk(md->disk);
M
Mike Anderson 已提交
1803
	format_dev_t(md->name, MKDEV(_major, minor));
L
Linus Torvalds 已提交
1804

1805 1806 1807 1808
	md->wq = create_singlethread_workqueue("kdmflush");
	if (!md->wq)
		goto bad_thread;

M
Mikulas Patocka 已提交
1809 1810 1811 1812
	md->bdev = bdget_disk(md->disk, 0);
	if (!md->bdev)
		goto bad_bdev;

1813
	/* Populate the mapping, nobody knows we exist yet */
1814
	spin_lock(&_minor_lock);
1815
	old_md = idr_replace(&_minor_idr, md, minor);
1816
	spin_unlock(&_minor_lock);
1817 1818 1819

	BUG_ON(old_md != MINOR_ALLOCED);

L
Linus Torvalds 已提交
1820 1821
	return md;

M
Mikulas Patocka 已提交
1822 1823
bad_bdev:
	destroy_workqueue(md->wq);
1824 1825
bad_thread:
	put_disk(md->disk);
M
Milan Broz 已提交
1826
bad_disk:
1827
	blk_cleanup_queue(md->queue);
M
Milan Broz 已提交
1828
bad_queue:
L
Linus Torvalds 已提交
1829
	free_minor(minor);
M
Milan Broz 已提交
1830
bad_minor:
1831
	module_put(THIS_MODULE);
M
Milan Broz 已提交
1832
bad_module_get:
L
Linus Torvalds 已提交
1833 1834 1835 1836
	kfree(md);
	return NULL;
}

J
Jun'ichi Nomura 已提交
1837 1838
static void unlock_fs(struct mapped_device *md);

L
Linus Torvalds 已提交
1839 1840
static void free_dev(struct mapped_device *md)
{
1841
	int minor = MINOR(disk_devt(md->disk));
1842

M
Mikulas Patocka 已提交
1843 1844
	unlock_fs(md);
	bdput(md->bdev);
1845
	destroy_workqueue(md->wq);
K
Kiyoshi Ueda 已提交
1846 1847 1848 1849 1850 1851
	if (md->tio_pool)
		mempool_destroy(md->tio_pool);
	if (md->io_pool)
		mempool_destroy(md->io_pool);
	if (md->bs)
		bioset_free(md->bs);
M
Martin K. Petersen 已提交
1852
	blk_integrity_unregister(md->disk);
L
Linus Torvalds 已提交
1853
	del_gendisk(md->disk);
1854
	free_minor(minor);
J
Jeff Mahoney 已提交
1855 1856 1857 1858 1859

	spin_lock(&_minor_lock);
	md->disk->private_data = NULL;
	spin_unlock(&_minor_lock);

L
Linus Torvalds 已提交
1860
	put_disk(md->disk);
1861
	blk_cleanup_queue(md->queue);
1862
	module_put(THIS_MODULE);
L
Linus Torvalds 已提交
1863 1864 1865
	kfree(md);
}

K
Kiyoshi Ueda 已提交
1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888
static void __bind_mempools(struct mapped_device *md, struct dm_table *t)
{
	struct dm_md_mempools *p;

	if (md->io_pool && md->tio_pool && md->bs)
		/* the md already has necessary mempools */
		goto out;

	p = dm_table_get_md_mempools(t);
	BUG_ON(!p || md->io_pool || md->tio_pool || md->bs);

	md->io_pool = p->io_pool;
	p->io_pool = NULL;
	md->tio_pool = p->tio_pool;
	p->tio_pool = NULL;
	md->bs = p->bs;
	p->bs = NULL;

out:
	/* mempool bind completed, now no need any mempools in the table */
	dm_table_free_md_mempools(t);
}

L
Linus Torvalds 已提交
1889 1890 1891 1892 1893
/*
 * Bind a table to the device.
 */
static void event_callback(void *context)
{
M
Mike Anderson 已提交
1894 1895
	unsigned long flags;
	LIST_HEAD(uevents);
L
Linus Torvalds 已提交
1896 1897
	struct mapped_device *md = (struct mapped_device *) context;

M
Mike Anderson 已提交
1898 1899 1900 1901
	spin_lock_irqsave(&md->uevent_lock, flags);
	list_splice_init(&md->uevent_list, &uevents);
	spin_unlock_irqrestore(&md->uevent_lock, flags);

1902
	dm_send_uevents(&uevents, &disk_to_dev(md->disk)->kobj);
M
Mike Anderson 已提交
1903

L
Linus Torvalds 已提交
1904 1905 1906 1907
	atomic_inc(&md->event_nr);
	wake_up(&md->eventq);
}

1908
static void __set_size(struct mapped_device *md, sector_t size)
L
Linus Torvalds 已提交
1909
{
1910
	set_capacity(md->disk, size);
L
Linus Torvalds 已提交
1911

1912 1913 1914
	mutex_lock(&md->bdev->bd_inode->i_mutex);
	i_size_write(md->bdev->bd_inode, (loff_t)size << SECTOR_SHIFT);
	mutex_unlock(&md->bdev->bd_inode->i_mutex);
L
Linus Torvalds 已提交
1915 1916
}

1917 1918
static int __bind(struct mapped_device *md, struct dm_table *t,
		  struct queue_limits *limits)
L
Linus Torvalds 已提交
1919
{
1920
	struct request_queue *q = md->queue;
L
Linus Torvalds 已提交
1921
	sector_t size;
1922
	unsigned long flags;
L
Linus Torvalds 已提交
1923 1924

	size = dm_table_get_size(t);
D
Darrick J. Wong 已提交
1925 1926 1927 1928 1929 1930 1931

	/*
	 * Wipe any geometry if the size of the table changed.
	 */
	if (size != get_capacity(md->disk))
		memset(&md->geometry, 0, sizeof(md->geometry));

M
Mikulas Patocka 已提交
1932
	__set_size(md, size);
1933 1934 1935

	if (!size) {
		dm_table_destroy(t);
L
Linus Torvalds 已提交
1936
		return 0;
1937
	}
L
Linus Torvalds 已提交
1938

1939 1940
	dm_table_event_callback(t, event_callback, md);

K
Kiyoshi Ueda 已提交
1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952
	/*
	 * The queue hasn't been stopped yet, if the old table type wasn't
	 * for request-based during suspension.  So stop it to prevent
	 * I/O mapping before resume.
	 * This must be done before setting the queue restrictions,
	 * because request-based dm may be run just after the setting.
	 */
	if (dm_table_request_based(t) && !blk_queue_stopped(q))
		stop_queue(q);

	__bind_mempools(md, t);

1953
	write_lock_irqsave(&md->map_lock, flags);
L
Linus Torvalds 已提交
1954
	md->map = t;
1955
	dm_table_set_restrictions(t, q, limits);
1956
	write_unlock_irqrestore(&md->map_lock, flags);
L
Linus Torvalds 已提交
1957 1958 1959 1960 1961 1962 1963

	return 0;
}

static void __unbind(struct mapped_device *md)
{
	struct dm_table *map = md->map;
1964
	unsigned long flags;
L
Linus Torvalds 已提交
1965 1966 1967 1968 1969

	if (!map)
		return;

	dm_table_event_callback(map, NULL, NULL);
1970
	write_lock_irqsave(&md->map_lock, flags);
L
Linus Torvalds 已提交
1971
	md->map = NULL;
1972
	write_unlock_irqrestore(&md->map_lock, flags);
1973
	dm_table_destroy(map);
L
Linus Torvalds 已提交
1974 1975 1976 1977 1978
}

/*
 * Constructor for a new device.
 */
1979
int dm_create(int minor, struct mapped_device **result)
L
Linus Torvalds 已提交
1980 1981 1982
{
	struct mapped_device *md;

1983
	md = alloc_dev(minor);
L
Linus Torvalds 已提交
1984 1985 1986
	if (!md)
		return -ENXIO;

M
Milan Broz 已提交
1987 1988
	dm_sysfs_init(md);

L
Linus Torvalds 已提交
1989 1990 1991 1992
	*result = md;
	return 0;
}

1993
static struct mapped_device *dm_find_md(dev_t dev)
L
Linus Torvalds 已提交
1994 1995 1996 1997 1998 1999 2000
{
	struct mapped_device *md;
	unsigned minor = MINOR(dev);

	if (MAJOR(dev) != _major || minor >= (1 << MINORBITS))
		return NULL;

2001
	spin_lock(&_minor_lock);
L
Linus Torvalds 已提交
2002 2003

	md = idr_find(&_minor_idr, minor);
J
Jeff Mahoney 已提交
2004
	if (md && (md == MINOR_ALLOCED ||
2005
		   (MINOR(disk_devt(dm_disk(md))) != minor) ||
A
Alasdair G Kergon 已提交
2006
		   test_bit(DMF_FREEING, &md->flags))) {
2007
		md = NULL;
J
Jeff Mahoney 已提交
2008 2009
		goto out;
	}
L
Linus Torvalds 已提交
2010

J
Jeff Mahoney 已提交
2011
out:
2012
	spin_unlock(&_minor_lock);
L
Linus Torvalds 已提交
2013

2014 2015 2016
	return md;
}

2017 2018 2019 2020 2021 2022 2023 2024 2025 2026
struct mapped_device *dm_get_md(dev_t dev)
{
	struct mapped_device *md = dm_find_md(dev);

	if (md)
		dm_get(md);

	return md;
}

A
Alasdair G Kergon 已提交
2027
void *dm_get_mdptr(struct mapped_device *md)
2028
{
A
Alasdair G Kergon 已提交
2029
	return md->interface_ptr;
L
Linus Torvalds 已提交
2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041
}

void dm_set_mdptr(struct mapped_device *md, void *ptr)
{
	md->interface_ptr = ptr;
}

void dm_get(struct mapped_device *md)
{
	atomic_inc(&md->holders);
}

2042 2043 2044 2045 2046 2047
const char *dm_device_name(struct mapped_device *md)
{
	return md->name;
}
EXPORT_SYMBOL_GPL(dm_device_name);

L
Linus Torvalds 已提交
2048 2049
void dm_put(struct mapped_device *md)
{
M
Mike Anderson 已提交
2050
	struct dm_table *map;
L
Linus Torvalds 已提交
2051

J
Jeff Mahoney 已提交
2052 2053
	BUG_ON(test_bit(DMF_FREEING, &md->flags));

2054
	if (atomic_dec_and_lock(&md->holders, &_minor_lock)) {
M
Mike Anderson 已提交
2055
		map = dm_get_table(md);
2056 2057
		idr_replace(&_minor_idr, MINOR_ALLOCED,
			    MINOR(disk_devt(dm_disk(md))));
J
Jeff Mahoney 已提交
2058
		set_bit(DMF_FREEING, &md->flags);
2059
		spin_unlock(&_minor_lock);
2060
		if (!dm_suspended(md)) {
L
Linus Torvalds 已提交
2061 2062 2063
			dm_table_presuspend_targets(map);
			dm_table_postsuspend_targets(map);
		}
M
Milan Broz 已提交
2064
		dm_sysfs_exit(md);
M
Mike Anderson 已提交
2065
		dm_table_put(map);
2066
		__unbind(md);
L
Linus Torvalds 已提交
2067 2068 2069
		free_dev(md);
	}
}
E
Edward Goggin 已提交
2070
EXPORT_SYMBOL_GPL(dm_put);
L
Linus Torvalds 已提交
2071

2072
static int dm_wait_for_completion(struct mapped_device *md, int interruptible)
2073 2074
{
	int r = 0;
2075
	DECLARE_WAITQUEUE(wait, current);
2076 2077
	struct request_queue *q = md->queue;
	unsigned long flags;
2078 2079 2080 2081

	dm_unplug_all(md->queue);

	add_wait_queue(&md->wait, &wait);
2082 2083

	while (1) {
2084
		set_current_state(interruptible);
2085 2086

		smp_mb();
2087 2088 2089 2090 2091 2092 2093
		if (dm_request_based(md)) {
			spin_lock_irqsave(q->queue_lock, flags);
			if (!queue_in_flight(q) && blk_queue_stopped(q)) {
				spin_unlock_irqrestore(q->queue_lock, flags);
				break;
			}
			spin_unlock_irqrestore(q->queue_lock, flags);
2094 2095
		} else if (!atomic_read(&md->pending[0]) &&
					!atomic_read(&md->pending[1]))
2096 2097
			break;

2098 2099
		if (interruptible == TASK_INTERRUPTIBLE &&
		    signal_pending(current)) {
2100 2101 2102 2103 2104 2105 2106 2107
			r = -EINTR;
			break;
		}

		io_schedule();
	}
	set_current_state(TASK_RUNNING);

2108 2109
	remove_wait_queue(&md->wait, &wait);

2110 2111 2112
	return r;
}

M
Mikulas Patocka 已提交
2113
static void dm_flush(struct mapped_device *md)
2114 2115
{
	dm_wait_for_completion(md, TASK_UNINTERRUPTIBLE);
2116 2117 2118 2119 2120 2121 2122

	bio_init(&md->barrier_bio);
	md->barrier_bio.bi_bdev = md->bdev;
	md->barrier_bio.bi_rw = WRITE_BARRIER;
	__split_and_process_bio(md, &md->barrier_bio);

	dm_wait_for_completion(md, TASK_UNINTERRUPTIBLE);
2123 2124 2125 2126
}

static void process_barrier(struct mapped_device *md, struct bio *bio)
{
2127 2128
	md->barrier_error = 0;

M
Mikulas Patocka 已提交
2129
	dm_flush(md);
2130

2131 2132 2133
	if (!bio_empty_barrier(bio)) {
		__split_and_process_bio(md, bio);
		dm_flush(md);
2134 2135 2136
	}

	if (md->barrier_error != DM_ENDIO_REQUEUE)
M
Mikulas Patocka 已提交
2137
		bio_endio(bio, md->barrier_error);
2138 2139 2140 2141 2142
	else {
		spin_lock_irq(&md->deferred_lock);
		bio_list_add_head(&md->deferred, bio);
		spin_unlock_irq(&md->deferred_lock);
	}
2143 2144
}

L
Linus Torvalds 已提交
2145 2146 2147
/*
 * Process the deferred bios
 */
2148
static void dm_wq_work(struct work_struct *work)
L
Linus Torvalds 已提交
2149
{
2150 2151
	struct mapped_device *md = container_of(work, struct mapped_device,
						work);
2152
	struct bio *c;
L
Linus Torvalds 已提交
2153

2154 2155
	down_write(&md->io_lock);

2156
	while (!test_bit(DMF_BLOCK_IO_FOR_SUSPEND, &md->flags)) {
A
Alasdair G Kergon 已提交
2157 2158 2159 2160 2161
		spin_lock_irq(&md->deferred_lock);
		c = bio_list_pop(&md->deferred);
		spin_unlock_irq(&md->deferred_lock);

		if (!c) {
2162
			clear_bit(DMF_QUEUE_IO_TO_THREAD, &md->flags);
A
Alasdair G Kergon 已提交
2163 2164
			break;
		}
2165

2166 2167
		up_write(&md->io_lock);

K
Kiyoshi Ueda 已提交
2168 2169 2170
		if (dm_request_based(md))
			generic_make_request(c);
		else {
2171
			if (bio_rw_flagged(c, BIO_RW_BARRIER))
K
Kiyoshi Ueda 已提交
2172 2173 2174 2175
				process_barrier(md, c);
			else
				__split_and_process_bio(md, c);
		}
2176 2177

		down_write(&md->io_lock);
2178
	}
M
Milan Broz 已提交
2179

2180
	up_write(&md->io_lock);
L
Linus Torvalds 已提交
2181 2182
}

2183
static void dm_queue_flush(struct mapped_device *md)
2184
{
2185 2186
	clear_bit(DMF_BLOCK_IO_FOR_SUSPEND, &md->flags);
	smp_mb__after_clear_bit();
2187
	queue_work(md->wq, &md->work);
2188 2189
}

L
Linus Torvalds 已提交
2190 2191 2192 2193 2194
/*
 * Swap in a new table (destroying old one).
 */
int dm_swap_table(struct mapped_device *md, struct dm_table *table)
{
2195
	struct queue_limits limits;
2196
	int r = -EINVAL;
L
Linus Torvalds 已提交
2197

2198
	mutex_lock(&md->suspend_lock);
L
Linus Torvalds 已提交
2199 2200

	/* device must be suspended */
2201
	if (!dm_suspended(md))
2202
		goto out;
L
Linus Torvalds 已提交
2203

2204 2205 2206 2207
	r = dm_calculate_queue_limits(table, &limits);
	if (r)
		goto out;

K
Kiyoshi Ueda 已提交
2208 2209 2210 2211 2212 2213 2214
	/* cannot change the device type, once a table is bound */
	if (md->map &&
	    (dm_table_get_type(md->map) != dm_table_get_type(table))) {
		DMWARN("can't change the device type after a table is bound");
		goto out;
	}

L
Linus Torvalds 已提交
2215
	__unbind(md);
2216
	r = __bind(md, table, &limits);
L
Linus Torvalds 已提交
2217

2218
out:
2219
	mutex_unlock(&md->suspend_lock);
2220
	return r;
L
Linus Torvalds 已提交
2221 2222
}

2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283
static void dm_rq_invalidate_suspend_marker(struct mapped_device *md)
{
	md->suspend_rq.special = (void *)0x1;
}

static void dm_rq_abort_suspend(struct mapped_device *md, int noflush)
{
	struct request_queue *q = md->queue;
	unsigned long flags;

	spin_lock_irqsave(q->queue_lock, flags);
	if (!noflush)
		dm_rq_invalidate_suspend_marker(md);
	__start_queue(q);
	spin_unlock_irqrestore(q->queue_lock, flags);
}

static void dm_rq_start_suspend(struct mapped_device *md, int noflush)
{
	struct request *rq = &md->suspend_rq;
	struct request_queue *q = md->queue;

	if (noflush)
		stop_queue(q);
	else {
		blk_rq_init(q, rq);
		blk_insert_request(q, rq, 0, NULL);
	}
}

static int dm_rq_suspend_available(struct mapped_device *md, int noflush)
{
	int r = 1;
	struct request *rq = &md->suspend_rq;
	struct request_queue *q = md->queue;
	unsigned long flags;

	if (noflush)
		return r;

	/* The marker must be protected by queue lock if it is in use */
	spin_lock_irqsave(q->queue_lock, flags);
	if (unlikely(rq->ref_count)) {
		/*
		 * This can happen, when the previous flush suspend was
		 * interrupted, the marker is still in the queue and
		 * this flush suspend has been invoked, because we don't
		 * remove the marker at the time of suspend interruption.
		 * We have only one marker per mapped_device, so we can't
		 * start another flush suspend while it is in use.
		 */
		BUG_ON(!rq->special); /* The marker should be invalidated */
		DMWARN("Invalidating the previous flush suspend is still in"
		       " progress.  Please retry later.");
		r = 0;
	}
	spin_unlock_irqrestore(q->queue_lock, flags);

	return r;
}

L
Linus Torvalds 已提交
2284 2285 2286 2287
/*
 * Functions to lock and unlock any filesystem running on the
 * device.
 */
2288
static int lock_fs(struct mapped_device *md)
L
Linus Torvalds 已提交
2289
{
2290
	int r;
L
Linus Torvalds 已提交
2291 2292

	WARN_ON(md->frozen_sb);
2293

2294
	md->frozen_sb = freeze_bdev(md->bdev);
2295
	if (IS_ERR(md->frozen_sb)) {
2296
		r = PTR_ERR(md->frozen_sb);
2297 2298
		md->frozen_sb = NULL;
		return r;
2299 2300
	}

2301 2302
	set_bit(DMF_FROZEN, &md->flags);

L
Linus Torvalds 已提交
2303 2304 2305
	return 0;
}

2306
static void unlock_fs(struct mapped_device *md)
L
Linus Torvalds 已提交
2307
{
2308 2309 2310
	if (!test_bit(DMF_FROZEN, &md->flags))
		return;

2311
	thaw_bdev(md->bdev, md->frozen_sb);
L
Linus Torvalds 已提交
2312
	md->frozen_sb = NULL;
2313
	clear_bit(DMF_FROZEN, &md->flags);
L
Linus Torvalds 已提交
2314 2315 2316 2317 2318 2319 2320 2321 2322
}

/*
 * We need to be able to change a mapping table under a mounted
 * filesystem.  For example we might want to move some data in
 * the background.  Before the table can be swapped with
 * dm_bind_table, dm_suspend must be called to flush any in
 * flight bios and ensure that any further io gets deferred.
 */
2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369
/*
 * Suspend mechanism in request-based dm.
 *
 * After the suspend starts, further incoming requests are kept in
 * the request_queue and deferred.
 * Remaining requests in the request_queue at the start of suspend are flushed
 * if it is flush suspend.
 * The suspend completes when the following conditions have been satisfied,
 * so wait for it:
 *    1. q->in_flight is 0 (which means no in_flight request)
 *    2. queue has been stopped (which means no request dispatching)
 *
 *
 * Noflush suspend
 * ---------------
 * Noflush suspend doesn't need to dispatch remaining requests.
 * So stop the queue immediately.  Then, wait for all in_flight requests
 * to be completed or requeued.
 *
 * To abort noflush suspend, start the queue.
 *
 *
 * Flush suspend
 * -------------
 * Flush suspend needs to dispatch remaining requests.  So stop the queue
 * after the remaining requests are completed. (Requeued request must be also
 * re-dispatched and completed.  Until then, we can't stop the queue.)
 *
 * During flushing the remaining requests, further incoming requests are also
 * inserted to the same queue.  To distinguish which requests are to be
 * flushed, we insert a marker request to the queue at the time of starting
 * flush suspend, like a barrier.
 * The dispatching is blocked when the marker is found on the top of the queue.
 * And the queue is stopped when all in_flight requests are completed, since
 * that means the remaining requests are completely flushed.
 * Then, the marker is removed from the queue.
 *
 * To abort flush suspend, we also need to take care of the marker, not only
 * starting the queue.
 * We don't remove the marker forcibly from the queue since it's against
 * the block-layer manner.  Instead, we put a invalidated mark on the marker.
 * When the invalidated marker is found on the top of the queue, it is
 * immediately removed from the queue, so it doesn't block dispatching.
 * Because we have only one marker per mapped_device, we can't start another
 * flush suspend until the invalidated marker is removed from the queue.
 * So fail and return with -EBUSY in such a case.
 */
2370
int dm_suspend(struct mapped_device *md, unsigned suspend_flags)
L
Linus Torvalds 已提交
2371
{
2372
	struct dm_table *map = NULL;
2373
	int r = 0;
2374
	int do_lockfs = suspend_flags & DM_SUSPEND_LOCKFS_FLAG ? 1 : 0;
2375
	int noflush = suspend_flags & DM_SUSPEND_NOFLUSH_FLAG ? 1 : 0;
L
Linus Torvalds 已提交
2376

2377
	mutex_lock(&md->suspend_lock);
2378

M
Milan Broz 已提交
2379 2380
	if (dm_suspended(md)) {
		r = -EINVAL;
2381
		goto out_unlock;
M
Milan Broz 已提交
2382
	}
L
Linus Torvalds 已提交
2383

2384 2385 2386 2387 2388
	if (dm_request_based(md) && !dm_rq_suspend_available(md, noflush)) {
		r = -EBUSY;
		goto out_unlock;
	}

L
Linus Torvalds 已提交
2389 2390
	map = dm_get_table(md);

2391 2392 2393 2394 2395 2396 2397
	/*
	 * DMF_NOFLUSH_SUSPENDING must be set before presuspend.
	 * This flag is cleared before dm_suspend returns.
	 */
	if (noflush)
		set_bit(DMF_NOFLUSH_SUSPENDING, &md->flags);

2398 2399 2400
	/* This does not get reverted if there's an error later. */
	dm_table_presuspend_targets(map);

M
Mikulas Patocka 已提交
2401 2402 2403 2404 2405 2406 2407
	/*
	 * Flush I/O to the device. noflush supersedes do_lockfs,
	 * because lock_fs() needs to flush I/Os.
	 */
	if (!noflush && do_lockfs) {
		r = lock_fs(md);
		if (r)
K
Kiyoshi Ueda 已提交
2408
			goto out;
2409
	}
L
Linus Torvalds 已提交
2410 2411

	/*
2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425
	 * Here we must make sure that no processes are submitting requests
	 * to target drivers i.e. no one may be executing
	 * __split_and_process_bio. This is called from dm_request and
	 * dm_wq_work.
	 *
	 * To get all processes out of __split_and_process_bio in dm_request,
	 * we take the write lock. To prevent any process from reentering
	 * __split_and_process_bio from dm_request, we set
	 * DMF_QUEUE_IO_TO_THREAD.
	 *
	 * To quiesce the thread (dm_wq_work), we set DMF_BLOCK_IO_FOR_SUSPEND
	 * and call flush_workqueue(md->wq). flush_workqueue will wait until
	 * dm_wq_work exits and DMF_BLOCK_IO_FOR_SUSPEND will prevent any
	 * further calls to __split_and_process_bio from dm_wq_work.
L
Linus Torvalds 已提交
2426
	 */
2427
	down_write(&md->io_lock);
2428 2429
	set_bit(DMF_BLOCK_IO_FOR_SUSPEND, &md->flags);
	set_bit(DMF_QUEUE_IO_TO_THREAD, &md->flags);
2430
	up_write(&md->io_lock);
L
Linus Torvalds 已提交
2431

2432 2433
	flush_workqueue(md->wq);

2434 2435 2436
	if (dm_request_based(md))
		dm_rq_start_suspend(md, noflush);

L
Linus Torvalds 已提交
2437
	/*
2438 2439 2440
	 * At this point no more requests are entering target request routines.
	 * We call dm_wait_for_completion to wait for all existing requests
	 * to finish.
L
Linus Torvalds 已提交
2441
	 */
2442
	r = dm_wait_for_completion(md, TASK_INTERRUPTIBLE);
L
Linus Torvalds 已提交
2443

2444
	down_write(&md->io_lock);
2445
	if (noflush)
2446
		clear_bit(DMF_NOFLUSH_SUSPENDING, &md->flags);
2447
	up_write(&md->io_lock);
2448

L
Linus Torvalds 已提交
2449
	/* were we interrupted ? */
2450
	if (r < 0) {
2451
		dm_queue_flush(md);
M
Milan Broz 已提交
2452

2453 2454 2455
		if (dm_request_based(md))
			dm_rq_abort_suspend(md, noflush);

2456
		unlock_fs(md);
2457
		goto out; /* pushback list is already flushed, so skip flush */
2458
	}
L
Linus Torvalds 已提交
2459

2460 2461 2462 2463 2464 2465
	/*
	 * If dm_wait_for_completion returned 0, the device is completely
	 * quiescent now. There is no request-processing activity. All new
	 * requests are being added to md->deferred list.
	 */

2466
	dm_table_postsuspend_targets(map);
L
Linus Torvalds 已提交
2467

2468
	set_bit(DMF_SUSPENDED, &md->flags);
2469

2470 2471
out:
	dm_table_put(map);
2472 2473

out_unlock:
2474
	mutex_unlock(&md->suspend_lock);
2475
	return r;
L
Linus Torvalds 已提交
2476 2477 2478 2479
}

int dm_resume(struct mapped_device *md)
{
2480 2481
	int r = -EINVAL;
	struct dm_table *map = NULL;
L
Linus Torvalds 已提交
2482

2483
	mutex_lock(&md->suspend_lock);
2484
	if (!dm_suspended(md))
2485 2486 2487
		goto out;

	map = dm_get_table(md);
2488
	if (!map || !dm_table_get_size(map))
2489
		goto out;
L
Linus Torvalds 已提交
2490

2491 2492 2493
	r = dm_table_resume_targets(map);
	if (r)
		goto out;
2494

2495
	dm_queue_flush(md);
2496

2497 2498 2499 2500 2501 2502 2503 2504
	/*
	 * Flushing deferred I/Os must be done after targets are resumed
	 * so that mapping of targets can work correctly.
	 * Request-based dm is queueing the deferred I/Os in its request_queue.
	 */
	if (dm_request_based(md))
		start_queue(md->queue);

2505 2506 2507 2508
	unlock_fs(md);

	clear_bit(DMF_SUSPENDED, &md->flags);

L
Linus Torvalds 已提交
2509
	dm_table_unplug_all(map);
2510 2511 2512
	r = 0;
out:
	dm_table_put(map);
2513
	mutex_unlock(&md->suspend_lock);
2514

2515
	return r;
L
Linus Torvalds 已提交
2516 2517 2518 2519 2520
}

/*-----------------------------------------------------------------
 * Event notification.
 *---------------------------------------------------------------*/
M
Milan Broz 已提交
2521 2522
void dm_kobject_uevent(struct mapped_device *md, enum kobject_action action,
		       unsigned cookie)
2523
{
M
Milan Broz 已提交
2524 2525 2526 2527 2528 2529 2530 2531 2532 2533
	char udev_cookie[DM_COOKIE_LENGTH];
	char *envp[] = { udev_cookie, NULL };

	if (!cookie)
		kobject_uevent(&disk_to_dev(md->disk)->kobj, action);
	else {
		snprintf(udev_cookie, DM_COOKIE_LENGTH, "%s=%u",
			 DM_COOKIE_ENV_VAR_NAME, cookie);
		kobject_uevent_env(&disk_to_dev(md->disk)->kobj, action, envp);
	}
2534 2535
}

M
Mike Anderson 已提交
2536 2537 2538 2539 2540
uint32_t dm_next_uevent_seq(struct mapped_device *md)
{
	return atomic_add_return(1, &md->uevent_seq);
}

L
Linus Torvalds 已提交
2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551
uint32_t dm_get_event_nr(struct mapped_device *md)
{
	return atomic_read(&md->event_nr);
}

int dm_wait_event(struct mapped_device *md, int event_nr)
{
	return wait_event_interruptible(md->eventq,
			(event_nr != atomic_read(&md->event_nr)));
}

M
Mike Anderson 已提交
2552 2553 2554 2555 2556 2557 2558 2559 2560
void dm_uevent_add(struct mapped_device *md, struct list_head *elist)
{
	unsigned long flags;

	spin_lock_irqsave(&md->uevent_lock, flags);
	list_add(elist, &md->uevent_list);
	spin_unlock_irqrestore(&md->uevent_lock, flags);
}

L
Linus Torvalds 已提交
2561 2562 2563 2564 2565 2566 2567 2568 2569
/*
 * The gendisk is only valid as long as you have a reference
 * count on 'md'.
 */
struct gendisk *dm_disk(struct mapped_device *md)
{
	return md->disk;
}

M
Milan Broz 已提交
2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586
struct kobject *dm_kobject(struct mapped_device *md)
{
	return &md->kobj;
}

/*
 * struct mapped_device should not be exported outside of dm.c
 * so use this check to verify that kobj is part of md structure
 */
struct mapped_device *dm_get_from_kobject(struct kobject *kobj)
{
	struct mapped_device *md;

	md = container_of(kobj, struct mapped_device, kobj);
	if (&md->kobj != kobj)
		return NULL;

2587 2588 2589 2590
	if (test_bit(DMF_FREEING, &md->flags) ||
	    test_bit(DMF_DELETING, &md->flags))
		return NULL;

M
Milan Broz 已提交
2591 2592 2593 2594
	dm_get(md);
	return md;
}

L
Linus Torvalds 已提交
2595 2596 2597 2598 2599
int dm_suspended(struct mapped_device *md)
{
	return test_bit(DMF_SUSPENDED, &md->flags);
}

2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610
int dm_noflush_suspending(struct dm_target *ti)
{
	struct mapped_device *md = dm_table_get_md(ti->table);
	int r = __noflush_suspending(md);

	dm_put(md);

	return r;
}
EXPORT_SYMBOL_GPL(dm_noflush_suspending);

K
Kiyoshi Ueda 已提交
2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665
struct dm_md_mempools *dm_alloc_md_mempools(unsigned type)
{
	struct dm_md_mempools *pools = kmalloc(sizeof(*pools), GFP_KERNEL);

	if (!pools)
		return NULL;

	pools->io_pool = (type == DM_TYPE_BIO_BASED) ?
			 mempool_create_slab_pool(MIN_IOS, _io_cache) :
			 mempool_create_slab_pool(MIN_IOS, _rq_bio_info_cache);
	if (!pools->io_pool)
		goto free_pools_and_out;

	pools->tio_pool = (type == DM_TYPE_BIO_BASED) ?
			  mempool_create_slab_pool(MIN_IOS, _tio_cache) :
			  mempool_create_slab_pool(MIN_IOS, _rq_tio_cache);
	if (!pools->tio_pool)
		goto free_io_pool_and_out;

	pools->bs = (type == DM_TYPE_BIO_BASED) ?
		    bioset_create(16, 0) : bioset_create(MIN_IOS, 0);
	if (!pools->bs)
		goto free_tio_pool_and_out;

	return pools;

free_tio_pool_and_out:
	mempool_destroy(pools->tio_pool);

free_io_pool_and_out:
	mempool_destroy(pools->io_pool);

free_pools_and_out:
	kfree(pools);

	return NULL;
}

void dm_free_md_mempools(struct dm_md_mempools *pools)
{
	if (!pools)
		return;

	if (pools->io_pool)
		mempool_destroy(pools->io_pool);

	if (pools->tio_pool)
		mempool_destroy(pools->tio_pool);

	if (pools->bs)
		bioset_free(pools->bs);

	kfree(pools);
}

2666
static const struct block_device_operations dm_blk_dops = {
L
Linus Torvalds 已提交
2667 2668
	.open = dm_blk_open,
	.release = dm_blk_close,
2669
	.ioctl = dm_blk_ioctl,
D
Darrick J. Wong 已提交
2670
	.getgeo = dm_blk_getgeo,
L
Linus Torvalds 已提交
2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686
	.owner = THIS_MODULE
};

EXPORT_SYMBOL(dm_get_mapinfo);

/*
 * module hooks
 */
module_init(dm_init);
module_exit(dm_exit);

module_param(major, uint, 0);
MODULE_PARM_DESC(major, "The major number of the device mapper");
MODULE_DESCRIPTION(DM_NAME " driver");
MODULE_AUTHOR("Joe Thornber <dm-devel@redhat.com>");
MODULE_LICENSE("GPL");