block_dev.c 53.1 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
1 2 3 4 5 6 7 8 9 10 11 12 13
/*
 *  linux/fs/block_dev.c
 *
 *  Copyright (C) 1991, 1992  Linus Torvalds
 *  Copyright (C) 2001  Andrea Arcangeli <andrea@suse.de> SuSE
 */

#include <linux/init.h>
#include <linux/mm.h>
#include <linux/fcntl.h>
#include <linux/slab.h>
#include <linux/kmod.h>
#include <linux/major.h>
14
#include <linux/device_cgroup.h>
Linus Torvalds's avatar
Linus Torvalds committed
15 16
#include <linux/highmem.h>
#include <linux/blkdev.h>
17
#include <linux/backing-dev.h>
Linus Torvalds's avatar
Linus Torvalds committed
18 19
#include <linux/module.h>
#include <linux/blkpg.h>
20
#include <linux/magic.h>
21
#include <linux/dax.h>
Linus Torvalds's avatar
Linus Torvalds committed
22
#include <linux/buffer_head.h>
Al Viro's avatar
Al Viro committed
23
#include <linux/swap.h>
24
#include <linux/pagevec.h>
25
#include <linux/writeback.h>
Linus Torvalds's avatar
Linus Torvalds committed
26 27 28 29
#include <linux/mpage.h>
#include <linux/mount.h>
#include <linux/uio.h>
#include <linux/namei.h>
30
#include <linux/log2.h>
Al Viro's avatar
Al Viro committed
31
#include <linux/cleancache.h>
32
#include <linux/dax.h>
33
#include <linux/badblocks.h>
34
#include <linux/task_io_accounting_ops.h>
35
#include <linux/falloc.h>
36
#include <linux/uaccess.h>
37
#include "internal.h"
Linus Torvalds's avatar
Linus Torvalds committed
38 39 40 41 42 43

struct bdev_inode {
	struct block_device bdev;
	struct inode vfs_inode;
};

44 45
static const struct address_space_operations def_blk_aops;

Linus Torvalds's avatar
Linus Torvalds committed
46 47 48 49 50
static inline struct bdev_inode *BDEV_I(struct inode *inode)
{
	return container_of(inode, struct bdev_inode, vfs_inode);
}

51
struct block_device *I_BDEV(struct inode *inode)
Linus Torvalds's avatar
Linus Torvalds committed
52 53 54 55 56
{
	return &BDEV_I(inode)->bdev;
}
EXPORT_SYMBOL(I_BDEV);

57
static void bdev_write_inode(struct block_device *bdev)
58
{
59 60 61
	struct inode *inode = bdev->bd_inode;
	int ret;

62 63 64
	spin_lock(&inode->i_lock);
	while (inode->i_state & I_DIRTY) {
		spin_unlock(&inode->i_lock);
65 66 67 68 69 70 71
		ret = write_inode_now(inode, true);
		if (ret) {
			char name[BDEVNAME_SIZE];
			pr_warn_ratelimited("VFS: Dirty inode writeback failed "
					    "for block device %s (err=%d).\n",
					    bdevname(bdev, name), ret);
		}
72 73 74 75 76
		spin_lock(&inode->i_lock);
	}
	spin_unlock(&inode->i_lock);
}

77
/* Kill _all_ buffers and pagecache , dirty or not.. */
Al Viro's avatar
Al Viro committed
78
void kill_bdev(struct block_device *bdev)
Linus Torvalds's avatar
Linus Torvalds committed
79
{
Al Viro's avatar
Al Viro committed
80 81
	struct address_space *mapping = bdev->bd_inode->i_mapping;

82
	if (mapping->nrpages == 0 && mapping->nrexceptional == 0)
83
		return;
Al Viro's avatar
Al Viro committed
84

85
	invalidate_bh_lrus();
Al Viro's avatar
Al Viro committed
86
	truncate_inode_pages(mapping, 0);
Linus Torvalds's avatar
Linus Torvalds committed
87
}	
Al Viro's avatar
Al Viro committed
88 89 90 91 92 93 94
EXPORT_SYMBOL(kill_bdev);

/* Invalidate clean unused buffers and pagecache. */
void invalidate_bdev(struct block_device *bdev)
{
	struct address_space *mapping = bdev->bd_inode->i_mapping;

95 96 97 98 99
	if (mapping->nrpages) {
		invalidate_bh_lrus();
		lru_add_drain_all();	/* make sure all lru add caches are flushed */
		invalidate_mapping_pages(mapping, 0, -1);
	}
Al Viro's avatar
Al Viro committed
100 101 102
	/* 99% of the time, we don't need to flush the cleancache on the bdev.
	 * But, for the strange corners, lets be cautious
	 */
103
	cleancache_invalidate_inode(mapping);
Al Viro's avatar
Al Viro committed
104 105
}
EXPORT_SYMBOL(invalidate_bdev);
Linus Torvalds's avatar
Linus Torvalds committed
106 107 108 109

int set_blocksize(struct block_device *bdev, int size)
{
	/* Size must be a power of two, and between 512 and PAGE_SIZE */
110
	if (size > PAGE_SIZE || size < 512 || !is_power_of_2(size))
Linus Torvalds's avatar
Linus Torvalds committed
111 112 113
		return -EINVAL;

	/* Size cannot be smaller than the size supported by the device */
114
	if (size < bdev_logical_block_size(bdev))
Linus Torvalds's avatar
Linus Torvalds committed
115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135
		return -EINVAL;

	/* Don't change the size if it is same as current */
	if (bdev->bd_block_size != size) {
		sync_blockdev(bdev);
		bdev->bd_block_size = size;
		bdev->bd_inode->i_blkbits = blksize_bits(size);
		kill_bdev(bdev);
	}
	return 0;
}

EXPORT_SYMBOL(set_blocksize);

int sb_set_blocksize(struct super_block *sb, int size)
{
	if (set_blocksize(sb->s_bdev, size))
		return 0;
	/* If we get here, we know size is power of two
	 * and it's value is between 512 and PAGE_SIZE */
	sb->s_blocksize = size;
136
	sb->s_blocksize_bits = blksize_bits(size);
Linus Torvalds's avatar
Linus Torvalds committed
137 138 139 140 141 142 143
	return sb->s_blocksize;
}

EXPORT_SYMBOL(sb_set_blocksize);

int sb_min_blocksize(struct super_block *sb, int size)
{
144
	int minsize = bdev_logical_block_size(sb->s_bdev);
Linus Torvalds's avatar
Linus Torvalds committed
145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161
	if (size < minsize)
		size = minsize;
	return sb_set_blocksize(sb, size);
}

EXPORT_SYMBOL(sb_min_blocksize);

static int
blkdev_get_block(struct inode *inode, sector_t iblock,
		struct buffer_head *bh, int create)
{
	bh->b_bdev = I_BDEV(inode);
	bh->b_blocknr = iblock;
	set_buffer_mapped(bh);
	return 0;
}

162 163 164 165 166
static struct inode *bdev_file_inode(struct file *file)
{
	return file->f_mapping->host;
}

167 168 169 170 171 172 173 174 175 176
static unsigned int dio_bio_write_op(struct kiocb *iocb)
{
	unsigned int op = REQ_OP_WRITE | REQ_SYNC | REQ_IDLE;

	/* avoid the need for a I/O completion work item */
	if (iocb->ki_flags & IOCB_DSYNC)
		op |= REQ_FUA;
	return op;
}

177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192
#define DIO_INLINE_BIO_VECS 4

static void blkdev_bio_end_io_simple(struct bio *bio)
{
	struct task_struct *waiter = bio->bi_private;

	WRITE_ONCE(bio->bi_private, NULL);
	wake_up_process(waiter);
}

static ssize_t
__blkdev_direct_IO_simple(struct kiocb *iocb, struct iov_iter *iter,
		int nr_pages)
{
	struct file *file = iocb->ki_filp;
	struct block_device *bdev = I_BDEV(bdev_file_inode(file));
193
	struct bio_vec inline_vecs[DIO_INLINE_BIO_VECS], *vecs, *bvec;
194 195 196 197 198 199 200
	loff_t pos = iocb->ki_pos;
	bool should_dirty = false;
	struct bio bio;
	ssize_t ret;
	blk_qc_t qc;
	int i;

201 202
	if ((pos | iov_iter_alignment(iter)) &
	    (bdev_logical_block_size(bdev) - 1))
203 204
		return -EINVAL;

205 206 207 208 209 210 211 212
	if (nr_pages <= DIO_INLINE_BIO_VECS)
		vecs = inline_vecs;
	else {
		vecs = kmalloc(nr_pages * sizeof(struct bio_vec), GFP_KERNEL);
		if (!vecs)
			return -ENOMEM;
	}

213
	bio_init(&bio, vecs, nr_pages);
214
	bio_set_dev(&bio, bdev);
215
	bio.bi_iter.bi_sector = pos >> 9;
216
	bio.bi_write_hint = iocb->ki_hint;
217 218 219 220 221 222 223 224 225
	bio.bi_private = current;
	bio.bi_end_io = blkdev_bio_end_io_simple;

	ret = bio_iov_iter_get_pages(&bio, iter);
	if (unlikely(ret))
		return ret;
	ret = bio.bi_iter.bi_size;

	if (iov_iter_rw(iter) == READ) {
226
		bio.bi_opf = REQ_OP_READ;
227 228 229
		if (iter_is_iovec(iter))
			should_dirty = true;
	} else {
230
		bio.bi_opf = dio_bio_write_op(iocb);
231 232 233 234 235 236 237 238 239
		task_io_account_write(ret);
	}

	qc = submit_bio(&bio);
	for (;;) {
		set_current_state(TASK_UNINTERRUPTIBLE);
		if (!READ_ONCE(bio.bi_private))
			break;
		if (!(iocb->ki_flags & IOCB_HIPRI) ||
240
		    !blk_poll(bdev_get_queue(bdev), qc))
241 242 243 244 245 246 247 248 249 250
			io_schedule();
	}
	__set_current_state(TASK_RUNNING);

	bio_for_each_segment_all(bvec, &bio, i) {
		if (should_dirty && !PageCompound(bvec->bv_page))
			set_page_dirty_lock(bvec->bv_page);
		put_page(bvec->bv_page);
	}

251 252 253
	if (vecs != inline_vecs)
		kfree(vecs);

254
	if (unlikely(bio.bi_status))
255
		ret = blk_status_to_errno(bio.bi_status);
256 257 258

	bio_uninit(&bio);

259 260 261
	return ret;
}

262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282
struct blkdev_dio {
	union {
		struct kiocb		*iocb;
		struct task_struct	*waiter;
	};
	size_t			size;
	atomic_t		ref;
	bool			multi_bio : 1;
	bool			should_dirty : 1;
	bool			is_sync : 1;
	struct bio		bio;
};

static struct bio_set *blkdev_dio_pool __read_mostly;

static void blkdev_bio_end_io(struct bio *bio)
{
	struct blkdev_dio *dio = bio->bi_private;
	bool should_dirty = dio->should_dirty;

	if (dio->multi_bio && !atomic_dec_and_test(&dio->ref)) {
283 284
		if (bio->bi_status && !dio->bio.bi_status)
			dio->bio.bi_status = bio->bi_status;
285 286 287
	} else {
		if (!dio->is_sync) {
			struct kiocb *iocb = dio->iocb;
288
			ssize_t ret;
289

290
			if (likely(!dio->bio.bi_status)) {
291 292
				ret = dio->size;
				iocb->ki_pos += ret;
293 294
			} else {
				ret = blk_status_to_errno(dio->bio.bi_status);
295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318
			}

			dio->iocb->ki_complete(iocb, ret, 0);
			bio_put(&dio->bio);
		} else {
			struct task_struct *waiter = dio->waiter;

			WRITE_ONCE(dio->waiter, NULL);
			wake_up_process(waiter);
		}
	}

	if (should_dirty) {
		bio_check_pages_dirty(bio);
	} else {
		struct bio_vec *bvec;
		int i;

		bio_for_each_segment_all(bvec, bio, i)
			put_page(bvec->bv_page);
		bio_put(bio);
	}
}

319
static ssize_t
320
__blkdev_direct_IO(struct kiocb *iocb, struct iov_iter *iter, int nr_pages)
321 322
{
	struct file *file = iocb->ki_filp;
323
	struct inode *inode = bdev_file_inode(file);
324
	struct block_device *bdev = I_BDEV(inode);
325
	struct blk_plug plug;
326 327
	struct blkdev_dio *dio;
	struct bio *bio;
328
	bool is_read = (iov_iter_rw(iter) == READ), is_sync;
329 330
	loff_t pos = iocb->ki_pos;
	blk_qc_t qc = BLK_QC_T_NONE;
331
	int ret = 0;
332

333 334
	if ((pos | iov_iter_alignment(iter)) &
	    (bdev_logical_block_size(bdev) - 1))
335 336 337 338 339 340
		return -EINVAL;

	bio = bio_alloc_bioset(GFP_KERNEL, nr_pages, blkdev_dio_pool);
	bio_get(bio); /* extra ref for the completion handler */

	dio = container_of(bio, struct blkdev_dio, bio);
341
	dio->is_sync = is_sync = is_sync_kiocb(iocb);
342 343 344 345 346 347 348 349 350
	if (dio->is_sync)
		dio->waiter = current;
	else
		dio->iocb = iocb;

	dio->size = 0;
	dio->multi_bio = false;
	dio->should_dirty = is_read && (iter->type == ITER_IOVEC);

351
	blk_start_plug(&plug);
352
	for (;;) {
353
		bio_set_dev(bio, bdev);
354
		bio->bi_iter.bi_sector = pos >> 9;
355
		bio->bi_write_hint = iocb->ki_hint;
356 357 358 359 360
		bio->bi_private = dio;
		bio->bi_end_io = blkdev_bio_end_io;

		ret = bio_iov_iter_get_pages(bio, iter);
		if (unlikely(ret)) {
361
			bio->bi_status = BLK_STS_IOERR;
362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393
			bio_endio(bio);
			break;
		}

		if (is_read) {
			bio->bi_opf = REQ_OP_READ;
			if (dio->should_dirty)
				bio_set_pages_dirty(bio);
		} else {
			bio->bi_opf = dio_bio_write_op(iocb);
			task_io_account_write(bio->bi_iter.bi_size);
		}

		dio->size += bio->bi_iter.bi_size;
		pos += bio->bi_iter.bi_size;

		nr_pages = iov_iter_npages(iter, BIO_MAX_PAGES);
		if (!nr_pages) {
			qc = submit_bio(bio);
			break;
		}

		if (!dio->multi_bio) {
			dio->multi_bio = true;
			atomic_set(&dio->ref, 2);
		} else {
			atomic_inc(&dio->ref);
		}

		submit_bio(bio);
		bio = bio_alloc(GFP_KERNEL, nr_pages);
	}
394
	blk_finish_plug(&plug);
395

396
	if (!is_sync)
397 398 399 400 401 402 403 404
		return -EIOCBQUEUED;

	for (;;) {
		set_current_state(TASK_UNINTERRUPTIBLE);
		if (!READ_ONCE(dio->waiter))
			break;

		if (!(iocb->ki_flags & IOCB_HIPRI) ||
405
		    !blk_poll(bdev_get_queue(bdev), qc))
406 407 408 409
			io_schedule();
	}
	__set_current_state(TASK_RUNNING);

410
	if (!ret)
411
		ret = blk_status_to_errno(dio->bio.bi_status);
412
	if (likely(!ret))
413 414 415 416 417 418 419 420 421
		ret = dio->size;

	bio_put(&dio->bio);
	return ret;
}

static ssize_t
blkdev_direct_IO(struct kiocb *iocb, struct iov_iter *iter)
{
422
	int nr_pages;
423

424
	nr_pages = iov_iter_npages(iter, BIO_MAX_PAGES + 1);
425 426
	if (!nr_pages)
		return 0;
427
	if (is_sync_kiocb(iocb) && nr_pages <= BIO_MAX_PAGES)
428
		return __blkdev_direct_IO_simple(iocb, iter, nr_pages);
429 430 431 432 433 434

	return __blkdev_direct_IO(iocb, iter, min(nr_pages, BIO_MAX_PAGES));
}

static __init int blkdev_init(void)
{
435
	blkdev_dio_pool = bioset_create(4, offsetof(struct blkdev_dio, bio), BIOSET_NEED_BVECS);
436 437 438
	if (!blkdev_dio_pool)
		return -ENOMEM;
	return 0;
439
}
440
module_init(blkdev_init);
441

442 443 444 445 446 447 448 449 450
int __sync_blockdev(struct block_device *bdev, int wait)
{
	if (!bdev)
		return 0;
	if (!wait)
		return filemap_flush(bdev->bd_inode->i_mapping);
	return filemap_write_and_wait(bdev->bd_inode->i_mapping);
}

451 452 453 454 455 456
/*
 * Write out and wait upon all the dirty data associated with a block
 * device via its mapping.  Does not take the superblock lock.
 */
int sync_blockdev(struct block_device *bdev)
{
457
	return __sync_blockdev(bdev, 1);
458 459 460 461 462 463 464 465 466 467 468 469
}
EXPORT_SYMBOL(sync_blockdev);

/*
 * Write out and wait upon all dirty data associated with this
 * device.   Filesystem data as well as the underlying block
 * device.  Takes the superblock lock.
 */
int fsync_bdev(struct block_device *bdev)
{
	struct super_block *sb = get_super(bdev);
	if (sb) {
470
		int res = sync_filesystem(sb);
471 472 473 474 475
		drop_super(sb);
		return res;
	}
	return sync_blockdev(bdev);
}
Al Viro's avatar
Al Viro committed
476
EXPORT_SYMBOL(fsync_bdev);
477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495

/**
 * freeze_bdev  --  lock a filesystem and force it into a consistent state
 * @bdev:	blockdevice to lock
 *
 * If a superblock is found on this device, we take the s_umount semaphore
 * on it to make sure nobody unmounts until the snapshot creation is done.
 * The reference counter (bd_fsfreeze_count) guarantees that only the last
 * unfreeze process can unfreeze the frozen filesystem actually when multiple
 * freeze requests arrive simultaneously. It counts up in freeze_bdev() and
 * count down in thaw_bdev(). When it becomes 0, thaw_bdev() will unfreeze
 * actually.
 */
struct super_block *freeze_bdev(struct block_device *bdev)
{
	struct super_block *sb;
	int error = 0;

	mutex_lock(&bdev->bd_fsfreeze_mutex);
496 497 498 499 500 501
	if (++bdev->bd_fsfreeze_count > 1) {
		/*
		 * We don't even need to grab a reference - the first call
		 * to freeze_bdev grab an active reference and only the last
		 * thaw_bdev drops it.
		 */
502
		sb = get_super(bdev);
503 504
		if (sb)
			drop_super(sb);
505 506 507 508 509 510 511
		mutex_unlock(&bdev->bd_fsfreeze_mutex);
		return sb;
	}

	sb = get_active_super(bdev);
	if (!sb)
		goto out;
512 513 514 515
	if (sb->s_op->freeze_super)
		error = sb->s_op->freeze_super(sb);
	else
		error = freeze_super(sb);
516 517 518
	if (error) {
		deactivate_super(sb);
		bdev->bd_fsfreeze_count--;
519
		mutex_unlock(&bdev->bd_fsfreeze_mutex);
520
		return ERR_PTR(error);
521
	}
522
	deactivate_super(sb);
523
 out:
524 525
	sync_blockdev(bdev);
	mutex_unlock(&bdev->bd_fsfreeze_mutex);
526
	return sb;	/* thaw_bdev releases s->s_umount */
527 528 529 530 531 532 533 534 535 536 537 538
}
EXPORT_SYMBOL(freeze_bdev);

/**
 * thaw_bdev  -- unlock filesystem
 * @bdev:	blockdevice to unlock
 * @sb:		associated superblock
 *
 * Unlocks the filesystem and marks it writeable again after freeze_bdev().
 */
int thaw_bdev(struct block_device *bdev, struct super_block *sb)
{
539
	int error = -EINVAL;
540 541

	mutex_lock(&bdev->bd_fsfreeze_mutex);
542
	if (!bdev->bd_fsfreeze_count)
543
		goto out;
544 545 546

	error = 0;
	if (--bdev->bd_fsfreeze_count > 0)
547
		goto out;
548 549

	if (!sb)
550
		goto out;
551

552 553 554 555
	if (sb->s_op->thaw_super)
		error = sb->s_op->thaw_super(sb);
	else
		error = thaw_super(sb);
556
	if (error)
557 558
		bdev->bd_fsfreeze_count++;
out:
559
	mutex_unlock(&bdev->bd_fsfreeze_mutex);
560
	return error;
561 562 563
}
EXPORT_SYMBOL(thaw_bdev);

Linus Torvalds's avatar
Linus Torvalds committed
564 565 566 567 568 569 570 571 572 573
static int blkdev_writepage(struct page *page, struct writeback_control *wbc)
{
	return block_write_full_page(page, blkdev_get_block, wbc);
}

static int blkdev_readpage(struct file * file, struct page * page)
{
	return block_read_full_page(page, blkdev_get_block);
}

574 575 576 577 578 579
static int blkdev_readpages(struct file *file, struct address_space *mapping,
			struct list_head *pages, unsigned nr_pages)
{
	return mpage_readpages(mapping, pages, nr_pages, blkdev_get_block);
}

580 581 582
static int blkdev_write_begin(struct file *file, struct address_space *mapping,
			loff_t pos, unsigned len, unsigned flags,
			struct page **pagep, void **fsdata)
Linus Torvalds's avatar
Linus Torvalds committed
583
{
584 585
	return block_write_begin(mapping, pos, len, flags, pagep,
				 blkdev_get_block);
Linus Torvalds's avatar
Linus Torvalds committed
586 587
}

588 589 590
static int blkdev_write_end(struct file *file, struct address_space *mapping,
			loff_t pos, unsigned len, unsigned copied,
			struct page *page, void *fsdata)
Linus Torvalds's avatar
Linus Torvalds committed
591
{
592 593 594 595
	int ret;
	ret = block_write_end(file, mapping, pos, len, copied, page, fsdata);

	unlock_page(page);
596
	put_page(page);
597 598

	return ret;
Linus Torvalds's avatar
Linus Torvalds committed
599 600 601 602
}

/*
 * private llseek:
Al Viro's avatar
Al Viro committed
603
 * for a block special file file_inode(file)->i_size is zero
Linus Torvalds's avatar
Linus Torvalds committed
604 605
 * so we compute the size by hand (just as in block_read/write above)
 */
606
static loff_t block_llseek(struct file *file, loff_t offset, int whence)
Linus Torvalds's avatar
Linus Torvalds committed
607
{
608
	struct inode *bd_inode = bdev_file_inode(file);
Linus Torvalds's avatar
Linus Torvalds committed
609 610
	loff_t retval;

Al Viro's avatar
Al Viro committed
611
	inode_lock(bd_inode);
612
	retval = fixed_size_llseek(file, offset, whence, i_size_read(bd_inode));
Al Viro's avatar
Al Viro committed
613
	inode_unlock(bd_inode);
Linus Torvalds's avatar
Linus Torvalds committed
614 615 616
	return retval;
}
	
617
int blkdev_fsync(struct file *filp, loff_t start, loff_t end, int datasync)
Linus Torvalds's avatar
Linus Torvalds committed
618
{
619
	struct inode *bd_inode = bdev_file_inode(filp);
620
	struct block_device *bdev = I_BDEV(bd_inode);
621
	int error;
622
	
623
	error = file_write_and_wait_range(filp, start, end);
624 625
	if (error)
		return error;
626

627 628 629 630 631
	/*
	 * There is no need to serialise calls to blkdev_issue_flush with
	 * i_mutex and doing so causes performance issues with concurrent
	 * O_SYNC writers to a block device.
	 */
632
	error = blkdev_issue_flush(bdev, GFP_KERNEL, NULL);
633 634
	if (error == -EOPNOTSUPP)
		error = 0;
635

636
	return error;
Linus Torvalds's avatar
Linus Torvalds committed
637
}
638
EXPORT_SYMBOL(blkdev_fsync);
Linus Torvalds's avatar
Linus Torvalds committed
639

640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659
/**
 * bdev_read_page() - Start reading a page from a block device
 * @bdev: The device to read the page from
 * @sector: The offset on the device to read the page to (need not be aligned)
 * @page: The page to read
 *
 * On entry, the page should be locked.  It will be unlocked when the page
 * has been read.  If the block driver implements rw_page synchronously,
 * that will be true on exit from this function, but it need not be.
 *
 * Errors returned by this function are usually "soft", eg out of memory, or
 * queue full; callers should try a different route to read this page rather
 * than propagate an error back up the stack.
 *
 * Return: negative errno if an error occurs, 0 if submission was successful.
 */
int bdev_read_page(struct block_device *bdev, sector_t sector,
			struct page *page)
{
	const struct block_device_operations *ops = bdev->bd_disk->fops;
660 661
	int result = -EOPNOTSUPP;

662
	if (!ops->rw_page || bdev_get_integrity(bdev))
663 664
		return result;

665
	result = blk_queue_enter(bdev->bd_queue, 0);
666 667
	if (result)
		return result;
668
	result = ops->rw_page(bdev, sector + get_start_sect(bdev), page, false);
669 670
	blk_queue_exit(bdev->bd_queue);
	return result;
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
}
EXPORT_SYMBOL_GPL(bdev_read_page);

/**
 * bdev_write_page() - Start writing a page to a block device
 * @bdev: The device to write the page to
 * @sector: The offset on the device to write the page to (need not be aligned)
 * @page: The page to write
 * @wbc: The writeback_control for the write
 *
 * On entry, the page should be locked and not currently under writeback.
 * On exit, if the write started successfully, the page will be unlocked and
 * under writeback.  If the write failed already (eg the driver failed to
 * queue the page to the device), the page will still be locked.  If the
 * caller is a ->writepage implementation, it will need to unlock the page.
 *
 * Errors returned by this function are usually "soft", eg out of memory, or
 * queue full; callers should try a different route to write this page rather
 * than propagate an error back up the stack.
 *
 * Return: negative errno if an error occurs, 0 if submission was successful.
 */
int bdev_write_page(struct block_device *bdev, sector_t sector,
			struct page *page, struct writeback_control *wbc)
{
	int result;
	const struct block_device_operations *ops = bdev->bd_disk->fops;
698

699
	if (!ops->rw_page || bdev_get_integrity(bdev))
700
		return -EOPNOTSUPP;
701
	result = blk_queue_enter(bdev->bd_queue, 0);
702 703 704
	if (result)
		return result;

705
	set_page_writeback(page);
706
	result = ops->rw_page(bdev, sector + get_start_sect(bdev), page, true);
707
	if (result) {
708
		end_page_writeback(page);
709 710
	} else {
		clean_page_buffers(page);
711
		unlock_page(page);
712
	}
713
	blk_queue_exit(bdev->bd_queue);
714 715 716 717
	return result;
}
EXPORT_SYMBOL_GPL(bdev_write_page);

Linus Torvalds's avatar
Linus Torvalds committed
718 719 720 721 722
/*
 * pseudo-fs
 */

static  __cacheline_aligned_in_smp DEFINE_SPINLOCK(bdev_lock);
723
static struct kmem_cache * bdev_cachep __read_mostly;
Linus Torvalds's avatar
Linus Torvalds committed
724 725 726

static struct inode *bdev_alloc_inode(struct super_block *sb)
{
727
	struct bdev_inode *ei = kmem_cache_alloc(bdev_cachep, GFP_KERNEL);
Linus Torvalds's avatar
Linus Torvalds committed
728 729 730 731 732
	if (!ei)
		return NULL;
	return &ei->vfs_inode;
}

Nick Piggin's avatar
Nick Piggin committed
733
static void bdev_i_callback(struct rcu_head *head)
Linus Torvalds's avatar
Linus Torvalds committed
734
{
Nick Piggin's avatar
Nick Piggin committed
735
	struct inode *inode = container_of(head, struct inode, i_rcu);
Linus Torvalds's avatar
Linus Torvalds committed
736 737 738 739 740
	struct bdev_inode *bdi = BDEV_I(inode);

	kmem_cache_free(bdev_cachep, bdi);
}

Nick Piggin's avatar
Nick Piggin committed
741 742 743 744 745
static void bdev_destroy_inode(struct inode *inode)
{
	call_rcu(&inode->i_rcu, bdev_i_callback);
}

746
static void init_once(void *foo)
Linus Torvalds's avatar
Linus Torvalds committed
747 748 749 750
{
	struct bdev_inode *ei = (struct bdev_inode *) foo;
	struct block_device *bdev = &ei->bdev;

751 752 753
	memset(bdev, 0, sizeof(*bdev));
	mutex_init(&bdev->bd_mutex);
	INIT_LIST_HEAD(&bdev->bd_list);
754 755 756
#ifdef CONFIG_SYSFS
	INIT_LIST_HEAD(&bdev->bd_holder_disks);
#endif
757
	bdev->bd_bdi = &noop_backing_dev_info;
758
	inode_init_once(&ei->vfs_inode);
759 760
	/* Initialize mutex for freeze. */
	mutex_init(&bdev->bd_fsfreeze_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
761 762
}

763
static void bdev_evict_inode(struct inode *inode)
Linus Torvalds's avatar
Linus Torvalds committed
764 765
{
	struct block_device *bdev = &BDEV_I(inode)->bdev;
766
	truncate_inode_pages_final(&inode->i_data);
767
	invalidate_inode_buffers(inode); /* is it needed here? */
768
	clear_inode(inode);
Linus Torvalds's avatar
Linus Torvalds committed
769 770 771
	spin_lock(&bdev_lock);
	list_del_init(&bdev->bd_list);
	spin_unlock(&bdev_lock);
772 773
	/* Detach inode from wb early as bdi_put() may free bdi->wb */
	inode_detach_wb(inode);
774
	if (bdev->bd_bdi != &noop_backing_dev_info) {
775
		bdi_put(bdev->bd_bdi);
776 777
		bdev->bd_bdi = &noop_backing_dev_info;
	}
Linus Torvalds's avatar
Linus Torvalds committed
778 779
}

780
static const struct super_operations bdev_sops = {
Linus Torvalds's avatar
Linus Torvalds committed
781 782 783 784
	.statfs = simple_statfs,
	.alloc_inode = bdev_alloc_inode,
	.destroy_inode = bdev_destroy_inode,
	.drop_inode = generic_delete_inode,
785
	.evict_inode = bdev_evict_inode,
Linus Torvalds's avatar
Linus Torvalds committed
786 787
};

Al Viro's avatar
Al Viro committed
788 789
static struct dentry *bd_mount(struct file_system_type *fs_type,
	int flags, const char *dev_name, void *data)
Linus Torvalds's avatar
Linus Torvalds committed
790
{
791 792
	struct dentry *dent;
	dent = mount_pseudo(fs_type, "bdev:", &bdev_sops, NULL, BDEVFS_MAGIC);
793
	if (!IS_ERR(dent))
794 795
		dent->d_sb->s_iflags |= SB_I_CGROUPWB;
	return dent;
Linus Torvalds's avatar
Linus Torvalds committed
796 797 798 799
}

static struct file_system_type bd_type = {
	.name		= "bdev",
Al Viro's avatar
Al Viro committed
800
	.mount		= bd_mount,
Linus Torvalds's avatar
Linus Torvalds committed
801 802 803
	.kill_sb	= kill_anon_super,
};

804 805
struct super_block *blockdev_superblock __read_mostly;
EXPORT_SYMBOL_GPL(blockdev_superblock);
Linus Torvalds's avatar
Linus Torvalds committed
806 807 808 809

void __init bdev_cache_init(void)
{
	int err;
810
	static struct vfsmount *bd_mnt;
811

Linus Torvalds's avatar
Linus Torvalds committed
812
	bdev_cachep = kmem_cache_create("bdev_cache", sizeof(struct bdev_inode),
813
			0, (SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|
814
				SLAB_MEM_SPREAD|SLAB_ACCOUNT|SLAB_PANIC),
815
			init_once);
Linus Torvalds's avatar
Linus Torvalds committed
816 817 818 819 820 821
	err = register_filesystem(&bd_type);
	if (err)
		panic("Cannot register bdev pseudo-fs");
	bd_mnt = kern_mount(&bd_type);
	if (IS_ERR(bd_mnt))
		panic("Cannot create bdev pseudo-fs");
822
	blockdev_superblock = bd_mnt->mnt_sb;   /* For writeback */
Linus Torvalds's avatar
Linus Torvalds committed
823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847
}

/*
 * Most likely _very_ bad one - but then it's hardly critical for small
 * /dev and can be fixed when somebody will need really large one.
 * Keep in mind that it will be fed through icache hash function too.
 */
static inline unsigned long hash(dev_t dev)
{
	return MAJOR(dev)+MINOR(dev);
}

static int bdev_test(struct inode *inode, void *data)
{
	return BDEV_I(inode)->bdev.bd_dev == *(dev_t *)data;
}

static int bdev_set(struct inode *inode, void *data)
{
	BDEV_I(inode)->bdev.bd_dev = *(dev_t *)data;
	return 0;
}

static LIST_HEAD(all_bdevs);

848 849 850 851 852 853 854 855 856 857 858 859 860 861 862
/*
 * If there is a bdev inode for this device, unhash it so that it gets evicted
 * as soon as last inode reference is dropped.
 */
void bdev_unhash_inode(dev_t dev)
{
	struct inode *inode;

	inode = ilookup5(blockdev_superblock, hash(dev), bdev_test, &dev);
	if (inode) {
		remove_inode_hash(inode);
		iput(inode);
	}
}

Linus Torvalds's avatar
Linus Torvalds committed
863 864 865 866 867
struct block_device *bdget(dev_t dev)
{
	struct block_device *bdev;
	struct inode *inode;

868
	inode = iget5_locked(blockdev_superblock, hash(dev),
Linus Torvalds's avatar
Linus Torvalds committed
869 870 871 872 873 874 875 876 877
			bdev_test, bdev_set, &dev);

	if (!inode)
		return NULL;

	bdev = &BDEV_I(inode)->bdev;

	if (inode->i_state & I_NEW) {
		bdev->bd_contains = NULL;
878
		bdev->bd_super = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
879
		bdev->bd_inode = inode;
Fabian Frederick's avatar
Fabian Frederick committed
880
		bdev->bd_block_size = i_blocksize(inode);
Linus Torvalds's avatar
Linus Torvalds committed
881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897
		bdev->bd_part_count = 0;
		bdev->bd_invalidated = 0;
		inode->i_mode = S_IFBLK;
		inode->i_rdev = dev;
		inode->i_bdev = bdev;
		inode->i_data.a_ops = &def_blk_aops;
		mapping_set_gfp_mask(&inode->i_data, GFP_USER);
		spin_lock(&bdev_lock);
		list_add(&bdev->bd_list, &all_bdevs);
		spin_unlock(&bdev_lock);
		unlock_new_inode(inode);
	}
	return bdev;
}

EXPORT_SYMBOL(bdget);

898 899 900 901 902 903
/**
 * bdgrab -- Grab a reference to an already referenced block device
 * @bdev:	Block device to grab a reference to.
 */
struct block_device *bdgrab(struct block_device *bdev)
{
Al Viro's avatar
Al Viro committed
904
	ihold(bdev->bd_inode);
905 906
	return bdev;
}
907
EXPORT_SYMBOL(bdgrab);
908

Linus Torvalds's avatar
Linus Torvalds committed
909 910
long nr_blockdev_pages(void)
{
911
	struct block_device *bdev;
Linus Torvalds's avatar
Linus Torvalds committed
912 913
	long ret = 0;
	spin_lock(&bdev_lock);
914
	list_for_each_entry(bdev, &all_bdevs, bd_list) {
Linus Torvalds's avatar
Linus Torvalds committed
915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930
		ret += bdev->bd_inode->i_mapping->nrpages;
	}
	spin_unlock(&bdev_lock);
	return ret;
}

void bdput(struct block_device *bdev)
{
	iput(bdev->bd_inode);
}

EXPORT_SYMBOL(bdput);
 
static struct block_device *bd_acquire(struct inode *inode)
{
	struct block_device *bdev;
931

Linus Torvalds's avatar
Linus Torvalds committed
932 933
	spin_lock(&bdev_lock);
	bdev = inode->i_bdev;
934
	if (bdev && !inode_unhashed(bdev->bd_inode)) {
935
		bdgrab(bdev);
Linus Torvalds's avatar
Linus Torvalds committed
936 937 938 939
		spin_unlock(&bdev_lock);
		return bdev;
	}
	spin_unlock(&bdev_lock);
940

941 942 943 944 945 946 947 948 949
	/*
	 * i_bdev references block device inode that was already shut down
	 * (corresponding device got removed).  Remove the reference and look
	 * up block device inode again just in case new device got
	 * reestablished under the same device number.
	 */
	if (bdev)
		bd_forget(inode);

Linus Torvalds's avatar
Linus Torvalds committed
950 951 952
	bdev = bdget(inode->i_rdev);
	if (bdev) {
		spin_lock(&bdev_lock);
953 954
		if (!inode->i_bdev) {
			/*
Al Viro's avatar
Al Viro committed
955
			 * We take an additional reference to bd_inode,
956 957 958 959
			 * and it's released in clear_inode() of inode.
			 * So, we can access it via ->i_mapping always
			 * without igrab().
			 */
960
			bdgrab(bdev);
961 962 963
			inode->i_bdev = bdev;
			inode->i_mapping = bdev->bd_inode->i_mapping;
		}
Linus Torvalds's avatar
Linus Torvalds committed
964 965 966 967 968 969 970 971 972
		spin_unlock(&bdev_lock);
	}
	return bdev;
}

/* Call when you free inode */

void bd_forget(struct inode *inode)
{
973 974
	struct block_device *bdev = NULL;

Linus Torvalds's avatar
Linus Torvalds committed
975
	spin_lock(&bdev_lock);
976 977
	if (!sb_is_blkdev_sb(inode->i_sb))
		bdev = inode->i_bdev;
Al Viro's avatar
Al Viro committed
978 979
	inode->i_bdev = NULL;
	inode->i_mapping = &inode->i_data;
Linus Torvalds's avatar
Linus Torvalds committed
980
	spin_unlock(&bdev_lock);
981 982

	if (bdev)
983
		bdput(bdev);
Linus Torvalds's avatar
Linus Torvalds committed
984 985
}

986 987 988 989 990 991
/**
 * bd_may_claim - test whether a block device can be claimed
 * @bdev: block device of interest
 * @whole: whole block device containing @bdev, may equal @bdev
 * @holder: holder trying to claim @bdev
 *
992
 * Test whether @bdev can be claimed by @holder.
993 994 995 996 997 998 999 1000 1001
 *
 * CONTEXT:
 * spin_lock(&bdev_lock).
 *
 * RETURNS:
 * %true if @bdev can be claimed, %false otherwise.
 */
static bool bd_may_claim(struct block_device *bdev, struct block_device *whole,
			 void *holder)
Linus Torvalds's avatar
Linus Torvalds committed
1002 1003
{
	if (bdev->bd_holder == holder)
1004
		return true;	 /* already a holder */
Linus Torvalds's avatar
Linus Torvalds committed
1005
	else if (bdev->bd_holder != NULL)
1006
		return false; 	 /* held by someone else */
1007
	else if (whole == bdev)
1008
		return true;  	 /* is a whole device which isn't held */
Linus Torvalds's avatar
Linus Torvalds committed
1009

1010
	else if (whole->bd_holder == bd_may_claim)
1011 1012 1013
		return true; 	 /* is a partition of a device that is being partitioned */
	else if (whole->bd_holder != NULL)
		return false;	 /* is a partition of a held device */
Linus Torvalds's avatar
Linus Torvalds committed
1014
	else
1015 1016 1017
		return true;	 /* is a partition of an un-held device */
}

1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043
/**
 * bd_prepare_to_claim - prepare to claim a block device
 * @bdev: block device of interest
 * @whole: the whole device containing @bdev, may equal @bdev
 * @holder: holder trying to claim @bdev
 *
 * Prepare to claim @bdev.  This function fails if @bdev is already
 * claimed by another holder and waits if another claiming is in
 * progress.  This function doesn't actually claim.  On successful
 * return, the caller has ownership of bd_claiming and bd_holder[s].
 *
 * CONTEXT:
 * spin_lock(&bdev_lock).  Might release bdev_lock, sleep and regrab
 * it multiple times.
 *
 * RETURNS:
 * 0 if @bdev can be claimed, -EBUSY otherwise.
 */
static int bd_prepare_to_claim(struct block_device *bdev,
			       struct block_device *whole, void *holder)
{
retry:
	/* if someone else claimed, fail */
	if (!bd_may_claim(bdev, whole, holder))
		return -EBUSY;

1044 1045
	/* if claiming is already in progress, wait for it to finish */
	if (whole->bd_claiming) {
1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068
		wait_queue_head_t *wq = bit_waitqueue(&whole->bd_claiming, 0);
		DEFINE_WAIT(wait);

		prepare_to_wait(wq, &wait, TASK_UNINTERRUPTIBLE);
		spin_unlock(&bdev_lock);
		schedule();
		finish_wait(wq, &wait);
		spin_lock(&bdev_lock);
		goto retry;
	}

	/* yay, all mine */
	return 0;
}

/**
 * bd_start_claiming - start claiming a block device
 * @bdev: block device of interest
 * @holder: holder trying to claim @bdev
 *
 * @bdev is about to be opened exclusively.  Check @bdev can be opened
 * exclusively and mark that an exclusive open is in progress.  Each
 * successful call to this function must be matched with a call to
1069 1070 1071 1072 1073 1074 1075
 * either bd_finish_claiming() or bd_abort_claiming() (which do not
 * fail).
 *
 * This function is used to gain exclusive access to the block device
 * without actually causing other exclusive open attempts to fail. It
 * should be used when the open sequence itself requires exclusive
 * access but may subsequently fail.
1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100
 *
 * CONTEXT:
 * Might sleep.
 *
 * RETURNS:
 * Pointer to the block device containing @bdev on success, ERR_PTR()
 * value on failure.
 */
static struct block_device *bd_start_claiming(struct block_device *bdev,
					      void *holder)
{
	struct gendisk *disk;
	struct block_device *whole;
	int partno, err;

	might_sleep();

	/*
	 * @bdev might not have been initialized properly yet, look up
	 * and grab the outer block device the hard way.
	 */
	disk = get_gendisk(bdev->bd_dev, &partno);
	if (!disk)
		return ERR_PTR(-ENXIO);

1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113
	/*
	 * Normally, @bdev should equal what's returned from bdget_disk()
	 * if partno is 0; however, some drivers (floppy) use multiple
	 * bdev's for the same physical device and @bdev may be one of the
	 * aliases.  Keep @bdev if partno is 0.  This means claimer
	 * tracking is broken for those devices but it has always been that
	 * way.
	 */
	if (partno)
		whole = bdget_disk(disk, 0);
	else
		whole = bdgrab(bdev);

1114
	module_put(disk->fops->owner);
1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133
	put_disk(disk);
	if (!whole)
		return ERR_PTR(-ENOMEM);

	/* prepare to claim, if successful, mark claiming in progress */
	spin_lock(&bdev_lock);

	err = bd_prepare_to_claim(bdev, whole, holder);
	if (err == 0) {
		whole->bd_claiming = holder;
		spin_unlock(&bdev_lock);
		return whole;
	} else {
		spin_unlock(&bdev_lock);
		bdput(whole);
		return ERR_PTR(err);
	}
}

1134
#ifdef CONFIG_SYSFS
1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151
struct bd_holder_disk {
	struct list_head	list;
	struct gendisk		*disk;
	int			refcnt;
};

static struct bd_holder_disk *bd_find_holder_disk(struct block_device *bdev,
						  struct gendisk *disk)
{
	struct bd_holder_disk *holder;

	list_for_each_entry(holder, &bdev->bd_holder_disks, list)
		if (holder->disk == disk)
			return holder;
	return NULL;
}

1152
static int add_symlink(struct kobject *from, struct kobject *to)
1153
{
1154
	return sysfs_create_link(from, to, kobject_name(to));
1155 1156 1157 1158 1159 1160 1161
}

static void del_symlink(struct kobject *from, struct kobject *to)
{
	sysfs_remove_link(from, kobject_name(to));
}

1162
/**
1163 1164 1165
 * bd_link_disk_holder - create symlinks between holding disk and slave bdev
 * @bdev: the claimed slave bdev
 * @disk: the holding disk
1166
 *
1167 1168
 * DON'T USE THIS UNLESS YOU'RE ALREADY USING IT.
 *
1169
 * This functions creates the following sysfs symlinks.
1170
 *
1171 1172
 * - from "slaves" directory of the holder @disk to the claimed @bdev
 * - from "holders" directory of the @bdev to the holder @disk
1173
 *
1174 1175
 * For example, if /dev/dm-0 maps to /dev/sda and disk for dm-0 is
 * passed to bd_link_disk_holder(), then:
1176
 *
1177 1178
 *   /sys/block/dm-0/slaves/sda --> /sys/block/sda
 *   /sys/block/sda/holders/dm-0 --> /sys/block/dm-0
1179
 *
1180 1181 1182
 * The caller must have claimed @bdev before calling this function and
 * ensure that both @bdev and @disk are valid during the creation and
 * lifetime of these symlinks.
1183
 *
1184 1185
 * CONTEXT:
 * Might sleep.
1186
 *
1187 1188
 * RETURNS:
 * 0 on success, -errno on failure.
1189
 */
1190
int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk)
1191
{
1192
	struct bd_holder_disk *holder;
1193
	int ret = 0;
1194

1195
	mutex_lock(&bdev->bd_mutex);
1196

1197
	WARN_ON_ONCE(!bdev->bd_holder);
1198

1199 1200 1201
	/* FIXME: remove the following once add_disk() handles errors */
	if (WARN_ON(!disk->slave_dir || !bdev->bd_part->holder_dir))
		goto out_unlock;
1202

1203 1204 1205
	holder = bd_find_holder_disk(bdev, disk);
	if (holder) {
		holder->refcnt++;
1206
		goto out_unlock;
1207
	}
1208

1209 1210 1211
	holder = kzalloc(sizeof(*holder), GFP_KERNEL);
	if (!holder) {
		ret = -ENOMEM;
1212 1213
		goto out_unlock;
	}
1214

1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225
	INIT_LIST_HEAD(&holder->list);
	holder->disk = disk;
	holder->refcnt = 1;

	ret = add_symlink(disk->slave_dir, &part_to_dev(bdev->bd_part)->kobj);
	if (ret)
		goto out_free;

	ret = add_symlink(bdev->bd_part->holder_dir, &disk_to_dev(disk)->kobj);
	if (ret)
		goto out_del;
1226 1227 1228 1229 1230
	/*
	 * bdev could be deleted beneath us which would implicitly destroy
	 * the holder directory.  Hold on to it.
	 */
	kobject_get(bdev->bd_part->holder_dir);
1231 1232 1233 1234 1235 1236 1237 1238

	list_add(&holder->list, &bdev->bd_holder_disks);
	goto out_unlock;

out_del:
	del_symlink(disk->slave_dir, &part_to_dev(bdev->bd_part)->kobj);
out_free:
	kfree(holder);
1239
out_unlock:
1240
	mutex_unlock(&bdev->bd_mutex);
1241
	return ret;
1242
}
1243
EXPORT_SYMBOL_GPL(bd_link_disk_holder);
1244

1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255
/**
 * bd_unlink_disk_holder - destroy symlinks created by bd_link_disk_holder()
 * @bdev: the calimed slave bdev
 * @disk: the holding disk
 *
 * DON'T USE THIS UNLESS YOU'RE ALREADY USING IT.
 *
 * CONTEXT:
 * Might sleep.
 */
void bd_unlink_disk_holder(struct block_device *bdev, struct gendisk *disk)
1256
{
1257
	struct bd_holder_disk *holder;
1258

1259
	mutex_lock(&bdev->bd_mutex);
1260

1261 1262 1263 1264 1265 1266
	holder = bd_find_holder_disk(bdev, disk);

	if (!WARN_ON_ONCE(holder == NULL) && !--holder->refcnt) {
		del_symlink(disk->slave_dir, &part_to_dev(bdev->bd_part)->kobj);
		del_symlink(bdev->bd_part->holder_dir,
			    &disk_to_dev(disk)->kobj);
1267
		kobject_put(bdev->bd_part->holder_dir);
1268 1269 1270 1271 1272
		list_del_init(&holder->list);
		kfree(holder);
	}

	mutex_unlock(&bdev->bd_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
1273
}
1274
EXPORT_SYMBOL_GPL(bd_unlink_disk_holder);
1275
#endif
Linus Torvalds's avatar
Linus Torvalds committed
1276

1277 1278 1279 1280
/**
 * flush_disk - invalidates all buffer-cache entries on a disk
 *
 * @bdev:      struct block device to be flushed
1281
 * @kill_dirty: flag to guide handling of dirty inodes
1282 1283 1284 1285 1286
 *
 * Invalidates all buffer-cache entries on a disk. It should be called
 * when a disk has been changed -- either by a media change or online
 * resize.
 */
1287
static void flush_disk(struct block_device *bdev, bool kill_dirty)
1288
{
1289
	if (__invalidate_device(bdev, kill_dirty)) {
1290
		printk(KERN_WARNING "VFS: busy inodes on changed media or "
1291 1292
		       "resized disk %s\n",
		       bdev->bd_disk ? bdev->bd_disk->disk_name : "");
1293 1294 1295 1296
	}

	if (!bdev->bd_disk)
		return;
1297
	if (disk_part_scan_enabled(bdev->bd_disk))
1298 1299 1300
		bdev->bd_invalidated = 1;
}

1301
/**
1302
 * check_disk_size_change - checks for disk size change and adjusts bdev size.
1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317
 * @disk: struct gendisk to check
 * @bdev: struct bdev to adjust.
 *
 * This routine checks to see if the bdev size does not match the disk size
 * and adjusts it if it differs.
 */
void check_disk_size_change(struct gendisk *disk, struct block_device *bdev)
{
	loff_t disk_size, bdev_size;

	disk_size = (loff_t)get_capacity(disk) << 9;
	bdev_size = i_size_read(bdev->bd_inode);
	if (disk_size != bdev_size) {
		printk(KERN_INFO
		       "%s: detected capacity change from %lld to %lld\n",
1318
		       disk->disk_name, bdev_size, disk_size);
1319
		i_size_write(bdev->bd_inode, disk_size);
1320
		flush_disk(bdev, false);
1321 1322 1323 1324
	}
}
EXPORT_SYMBOL(check_disk_size_change);

1325
/**
1326
 * revalidate_disk - wrapper for lower-level driver's revalidate_disk call-back
1327 1328 1329 1330 1331 1332 1333 1334
 * @disk: struct gendisk to be revalidated
 *
 * This routine is a wrapper for lower-level driver's revalidate_disk
 * call-backs.  It is used to do common pre and post operations needed
 * for all revalidate_disk operations.
 */
int revalidate_disk(struct gendisk *disk)
{
1335
	struct block_device *bdev;
1336 1337 1338 1339
	int ret = 0;

	if (disk->fops->revalidate_disk)
		ret = disk->fops->revalidate_disk(disk);
1340 1341 1342 1343 1344 1345
	bdev = bdget_disk(disk, 0);
	if (!bdev)
		return ret;

	mutex_lock(&bdev->bd_mutex);
	check_disk_size_change(disk, bdev);
1346
	bdev->bd_invalidated = 0;
1347 1348
	mutex_unlock(&bdev->bd_mutex);
	bdput(bdev);
1349 1350 1351 1352
	return ret;
}
EXPORT_SYMBOL(revalidate_disk);

Linus Torvalds's avatar
Linus Torvalds committed
1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364
/*
 * This routine checks whether a removable media has been changed,
 * and invalidates all buffer-cache-entries in that case. This
 * is a relatively slow routine, so we have to try to minimize using
 * it. Thus it is called only upon a 'mount' or 'open'. This
 * is the best way of combining speed and utility, I think.
 * People changing diskettes in the middle of an operation deserve
 * to lose :-)
 */
int check_disk_change(struct block_device *bdev)
{
	struct gendisk *disk = bdev->bd_disk;
1365
	const struct block_device_operations *bdops = disk->fops;
1366
	unsigned int events;
Linus Torvalds's avatar
Linus Torvalds committed
1367

1368 1369 1370
	events = disk_clear_events(disk, DISK_EVENT_MEDIA_CHANGE |
				   DISK_EVENT_EJECT_REQUEST);
	if (!(events & DISK_EVENT_MEDIA_CHANGE))
Linus Torvalds's avatar
Linus Torvalds committed
1371 1372
		return 0;

1373
	flush_disk(bdev, true);
Linus Torvalds's avatar
Linus Torvalds committed
1374 1375 1376 1377 1378 1379 1380 1381 1382
	if (bdops->revalidate_disk)
		bdops->revalidate_disk(bdev->bd_disk);
	return 1;
}

EXPORT_SYMBOL(check_disk_change);

void bd_set_size(struct block_device *bdev, loff_t size)
{
1383
	unsigned bsize = bdev_logical_block_size(bdev);
Linus Torvalds's avatar
Linus Torvalds committed
1384

Al Viro's avatar
Al Viro committed
1385
	inode_lock(bdev->bd_inode);
1386
	i_size_write(bdev->bd_inode, size);
Al Viro's avatar
Al Viro committed
1387
	inode_unlock(bdev->bd_inode);
1388
	while (bsize < PAGE_SIZE) {
Linus Torvalds's avatar
Linus Torvalds committed
1389 1390 1391 1392 1393 1394 1395 1396 1397
		if (size & bsize)
			break;
		bsize <<= 1;
	}
	bdev->bd_block_size = bsize;
	bdev->bd_inode->i_blkbits = blksize_bits(bsize);
}
EXPORT_SYMBOL(bd_set_size);

Al Viro's avatar
Al Viro committed
1398
static void __blkdev_put(struct block_device *bdev, fmode_t mode, int for_part);
1399

1400 1401 1402 1403 1404 1405 1406
/*
 * bd_mutex locking:
 *
 *  mutex_lock(part->bd_mutex)
 *    mutex_lock_nested(whole->bd_mutex, 1)
 */

1407
static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part)
Linus Torvalds's avatar
Linus Torvalds committed
1408 1409
{
	struct gendisk *disk;
1410
	struct module *owner;
1411
	int ret;
1412
	int partno;
1413 1414
	int perm = 0;

1415
	if (mode & FMODE_READ)
1416
		perm |= MAY_READ;
1417
	if (mode & FMODE_WRITE)
1418 1419 1420 1421
		perm |= MAY_WRITE;
	/*
	 * hooks: /n/, see "layering violations".
	 */
1422 1423 1424 1425 1426 1427
	if (!for_part) {
		ret = devcgroup_inode_permission(bdev->bd_inode, perm);
		if (ret != 0) {
			bdput(bdev);
			return ret;
		}
1428
	}
1429

1430
 restart:
1431

1432
	ret = -ENXIO;
1433
	disk = get_gendisk(bdev->bd_dev, &partno);
1434
	if (!disk)
1435
		goto out;
1436
	owner = disk->fops->owner;
Linus Torvalds's avatar
Linus Torvalds committed
1437

1438
	disk_block_events(disk);
1439
	mutex_lock_nested(&bdev->bd_mutex, for_part);
Linus Torvalds's avatar
Linus Torvalds committed
1440 1441
	if (!bdev->bd_openers) {
		bdev->bd_disk = disk;
1442
		bdev->bd_queue = disk->queue;
Linus Torvalds's avatar
Linus Torvalds committed
1443
		bdev->bd_contains = bdev;
1444
		bdev->bd_partno = partno;
1445

1446
		if (!partno) {
1447 1448 1449 1450 1451
			ret = -ENXIO;
			bdev->bd_part = disk_get_part(disk, partno);
			if (!bdev->bd_part)
				goto out_clear;

1452
			ret = 0;
Linus Torvalds's avatar
Linus Torvalds committed
1453
			if (disk->fops->open) {
1454
				ret = disk->fops->open(bdev, mode);
1455 1456 1457 1458 1459 1460 1461 1462
				if (ret == -ERESTARTSYS) {
					/* Lost a race with 'disk' being
					 * deleted, try again.
					 * See md.c
					 */
					disk_put_part(bdev->bd_part);
					bdev->bd_part = NULL;
					bdev->bd_disk = NULL;
1463
					bdev->bd_queue = NULL;
1464
					mutex_unlock(&bdev->bd_mutex);
1465 1466
					disk_unblock_events(disk);
					put_disk(disk);
1467
					module_put(owner);
1468 1469
					goto restart;
				}
Linus Torvalds's avatar
Linus Torvalds committed
1470
			}
1471

1472
			if (!ret)
1473 1474
				bd_set_size(bdev,(loff_t)get_capacity(disk)<<9);

1475 1476 1477