block_dev.c 35.9 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14
/*
 *  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>
#include <linux/smp_lock.h>
15
#include <linux/device_cgroup.h>
L
Linus Torvalds 已提交
16 17 18 19 20
#include <linux/highmem.h>
#include <linux/blkdev.h>
#include <linux/module.h>
#include <linux/blkpg.h>
#include <linux/buffer_head.h>
N
Nick Piggin 已提交
21
#include <linux/pagevec.h>
22
#include <linux/writeback.h>
L
Linus Torvalds 已提交
23 24 25 26
#include <linux/mpage.h>
#include <linux/mount.h>
#include <linux/uio.h>
#include <linux/namei.h>
27
#include <linux/log2.h>
28
#include <linux/kmemleak.h>
L
Linus Torvalds 已提交
29
#include <asm/uaccess.h>
30
#include "internal.h"
L
Linus Torvalds 已提交
31 32 33 34 35 36

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

A
Adrian Bunk 已提交
37 38
static const struct address_space_operations def_blk_aops;

L
Linus Torvalds 已提交
39 40 41 42 43 44 45 46 47 48 49 50
static inline struct bdev_inode *BDEV_I(struct inode *inode)
{
	return container_of(inode, struct bdev_inode, vfs_inode);
}

inline struct block_device *I_BDEV(struct inode *inode)
{
	return &BDEV_I(inode)->bdev;
}

EXPORT_SYMBOL(I_BDEV);

51 52 53 54 55 56 57 58 59 60 61
/*
 * move the inode from it's current bdi to the a new bdi. if the inode is dirty
 * we need to move it onto the dirty list of @dst so that the inode is always
 * on the right list.
 */
static void bdev_inode_switch_bdi(struct inode *inode,
			struct backing_dev_info *dst)
{
	spin_lock(&inode_lock);
	inode->i_data.backing_dev_info = dst;
	if (inode->i_state & I_DIRTY)
N
Nick Piggin 已提交
62
		list_move(&inode->i_wb_list, &dst->wb.b_dirty);
63 64 65
	spin_unlock(&inode_lock);
}

L
Linus Torvalds 已提交
66 67 68 69 70 71 72 73 74 75 76 77 78
static sector_t max_block(struct block_device *bdev)
{
	sector_t retval = ~((sector_t)0);
	loff_t sz = i_size_read(bdev->bd_inode);

	if (sz) {
		unsigned int size = block_size(bdev);
		unsigned int sizebits = blksize_bits(size);
		retval = (sz >> sizebits);
	}
	return retval;
}

P
Peter Zijlstra 已提交
79
/* Kill _all_ buffers and pagecache , dirty or not.. */
L
Linus Torvalds 已提交
80 81
static void kill_bdev(struct block_device *bdev)
{
P
Peter Zijlstra 已提交
82 83 84
	if (bdev->bd_inode->i_mapping->nrpages == 0)
		return;
	invalidate_bh_lrus();
L
Linus Torvalds 已提交
85 86 87 88 89 90
	truncate_inode_pages(bdev->bd_inode->i_mapping, 0);
}	

int set_blocksize(struct block_device *bdev, int size)
{
	/* Size must be a power of two, and between 512 and PAGE_SIZE */
91
	if (size > PAGE_SIZE || size < 512 || !is_power_of_2(size))
L
Linus Torvalds 已提交
92 93 94
		return -EINVAL;

	/* Size cannot be smaller than the size supported by the device */
95
	if (size < bdev_logical_block_size(bdev))
L
Linus Torvalds 已提交
96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116
		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;
117
	sb->s_blocksize_bits = blksize_bits(size);
L
Linus Torvalds 已提交
118 119 120 121 122 123 124
	return sb->s_blocksize;
}

EXPORT_SYMBOL(sb_set_blocksize);

int sb_min_blocksize(struct super_block *sb, int size)
{
125
	int minsize = bdev_logical_block_size(sb->s_bdev);
L
Linus Torvalds 已提交
126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154
	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)
{
	if (iblock >= max_block(I_BDEV(inode))) {
		if (create)
			return -EIO;

		/*
		 * for reads, we're just trying to fill a partial page.
		 * return a hole, they will have to call get_block again
		 * before they can fill it, and they will get -EIO at that
		 * time
		 */
		return 0;
	}
	bh->b_bdev = I_BDEV(inode);
	bh->b_blocknr = iblock;
	set_buffer_mapped(bh);
	return 0;
}

155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189
static int
blkdev_get_blocks(struct inode *inode, sector_t iblock,
		struct buffer_head *bh, int create)
{
	sector_t end_block = max_block(I_BDEV(inode));
	unsigned long max_blocks = bh->b_size >> inode->i_blkbits;

	if ((iblock + max_blocks) > end_block) {
		max_blocks = end_block - iblock;
		if ((long)max_blocks <= 0) {
			if (create)
				return -EIO;	/* write fully beyond EOF */
			/*
			 * It is a read which is fully beyond EOF.  We return
			 * a !buffer_mapped buffer
			 */
			max_blocks = 0;
		}
	}

	bh->b_bdev = I_BDEV(inode);
	bh->b_blocknr = iblock;
	bh->b_size = max_blocks << inode->i_blkbits;
	if (max_blocks)
		set_buffer_mapped(bh);
	return 0;
}

static ssize_t
blkdev_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov,
			loff_t offset, unsigned long nr_segs)
{
	struct file *file = iocb->ki_filp;
	struct inode *inode = file->f_mapping->host;

190 191
	return __blockdev_direct_IO(rw, iocb, inode, I_BDEV(inode), iov, offset,
				    nr_segs, blkdev_get_blocks, NULL, NULL, 0);
192 193
}

194 195 196 197 198 199 200 201 202
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);
}

N
Nick Piggin 已提交
203 204 205 206 207 208
/*
 * 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)
{
209
	return __sync_blockdev(bdev, 1);
N
Nick Piggin 已提交
210 211 212 213 214 215 216 217 218 219 220 221
}
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) {
222
		int res = sync_filesystem(sb);
N
Nick Piggin 已提交
223 224 225 226 227
		drop_super(sb);
		return res;
	}
	return sync_blockdev(bdev);
}
228
EXPORT_SYMBOL(fsync_bdev);
N
Nick Piggin 已提交
229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247

/**
 * 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);
248 249 250 251 252 253
	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.
		 */
N
Nick Piggin 已提交
254
		sb = get_super(bdev);
255 256 257 258 259 260 261 262
		drop_super(sb);
		mutex_unlock(&bdev->bd_fsfreeze_mutex);
		return sb;
	}

	sb = get_active_super(bdev);
	if (!sb)
		goto out;
263 264 265 266
	error = freeze_super(sb);
	if (error) {
		deactivate_super(sb);
		bdev->bd_fsfreeze_count--;
N
Nick Piggin 已提交
267
		mutex_unlock(&bdev->bd_fsfreeze_mutex);
268
		return ERR_PTR(error);
N
Nick Piggin 已提交
269
	}
270
	deactivate_super(sb);
271
 out:
N
Nick Piggin 已提交
272 273
	sync_blockdev(bdev);
	mutex_unlock(&bdev->bd_fsfreeze_mutex);
274
	return sb;	/* thaw_bdev releases s->s_umount */
N
Nick Piggin 已提交
275 276 277 278 279 280 281 282 283 284 285 286
}
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)
{
287
	int error = -EINVAL;
N
Nick Piggin 已提交
288 289

	mutex_lock(&bdev->bd_fsfreeze_mutex);
290
	if (!bdev->bd_fsfreeze_count)
291
		goto out;
292 293 294

	error = 0;
	if (--bdev->bd_fsfreeze_count > 0)
295
		goto out;
296 297

	if (!sb)
298
		goto out;
299

300 301 302 303 304 305 306
	error = thaw_super(sb);
	if (error) {
		bdev->bd_fsfreeze_count++;
		mutex_unlock(&bdev->bd_fsfreeze_mutex);
		return error;
	}
out:
N
Nick Piggin 已提交
307 308 309 310 311
	mutex_unlock(&bdev->bd_fsfreeze_mutex);
	return 0;
}
EXPORT_SYMBOL(thaw_bdev);

L
Linus Torvalds 已提交
312 313 314 315 316 317 318 319 320 321
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);
}

N
Nick Piggin 已提交
322 323 324
static int blkdev_write_begin(struct file *file, struct address_space *mapping,
			loff_t pos, unsigned len, unsigned flags,
			struct page **pagep, void **fsdata)
L
Linus Torvalds 已提交
325
{
326 327
	return block_write_begin(mapping, pos, len, flags, pagep,
				 blkdev_get_block);
L
Linus Torvalds 已提交
328 329
}

N
Nick Piggin 已提交
330 331 332
static int blkdev_write_end(struct file *file, struct address_space *mapping,
			loff_t pos, unsigned len, unsigned copied,
			struct page *page, void *fsdata)
L
Linus Torvalds 已提交
333
{
N
Nick Piggin 已提交
334 335 336 337 338 339 340
	int ret;
	ret = block_write_end(file, mapping, pos, len, copied, page, fsdata);

	unlock_page(page);
	page_cache_release(page);

	return ret;
L
Linus Torvalds 已提交
341 342 343 344
}

/*
 * private llseek:
345
 * for a block special file file->f_path.dentry->d_inode->i_size is zero
L
Linus Torvalds 已提交
346 347 348 349 350 351 352 353
 * so we compute the size by hand (just as in block_read/write above)
 */
static loff_t block_llseek(struct file *file, loff_t offset, int origin)
{
	struct inode *bd_inode = file->f_mapping->host;
	loff_t size;
	loff_t retval;

354
	mutex_lock(&bd_inode->i_mutex);
L
Linus Torvalds 已提交
355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370
	size = i_size_read(bd_inode);

	switch (origin) {
		case 2:
			offset += size;
			break;
		case 1:
			offset += file->f_pos;
	}
	retval = -EINVAL;
	if (offset >= 0 && offset <= size) {
		if (offset != file->f_pos) {
			file->f_pos = offset;
		}
		retval = offset;
	}
371
	mutex_unlock(&bd_inode->i_mutex);
L
Linus Torvalds 已提交
372 373 374
	return retval;
}
	
375
int blkdev_fsync(struct file *filp, int datasync)
L
Linus Torvalds 已提交
376
{
377 378
	struct inode *bd_inode = filp->f_mapping->host;
	struct block_device *bdev = I_BDEV(bd_inode);
379 380
	int error;

381 382 383 384 385 386 387
	/*
	 * 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.
	 */
	mutex_unlock(&bd_inode->i_mutex);

388
	error = blkdev_issue_flush(bdev, GFP_KERNEL, NULL);
389 390
	if (error == -EOPNOTSUPP)
		error = 0;
391 392 393

	mutex_lock(&bd_inode->i_mutex);

394
	return error;
L
Linus Torvalds 已提交
395
}
396
EXPORT_SYMBOL(blkdev_fsync);
L
Linus Torvalds 已提交
397 398 399 400 401 402

/*
 * pseudo-fs
 */

static  __cacheline_aligned_in_smp DEFINE_SPINLOCK(bdev_lock);
403
static struct kmem_cache * bdev_cachep __read_mostly;
L
Linus Torvalds 已提交
404 405 406

static struct inode *bdev_alloc_inode(struct super_block *sb)
{
407
	struct bdev_inode *ei = kmem_cache_alloc(bdev_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
408 409 410 411 412 413 414 415 416 417 418 419
	if (!ei)
		return NULL;
	return &ei->vfs_inode;
}

static void bdev_destroy_inode(struct inode *inode)
{
	struct bdev_inode *bdi = BDEV_I(inode);

	kmem_cache_free(bdev_cachep, bdi);
}

420
static void init_once(void *foo)
L
Linus Torvalds 已提交
421 422 423 424
{
	struct bdev_inode *ei = (struct bdev_inode *) foo;
	struct block_device *bdev = &ei->bdev;

C
Christoph Lameter 已提交
425 426 427 428 429
	memset(bdev, 0, sizeof(*bdev));
	mutex_init(&bdev->bd_mutex);
	INIT_LIST_HEAD(&bdev->bd_inodes);
	INIT_LIST_HEAD(&bdev->bd_list);
	inode_init_once(&ei->vfs_inode);
430 431
	/* Initialize mutex for freeze. */
	mutex_init(&bdev->bd_fsfreeze_mutex);
L
Linus Torvalds 已提交
432 433 434 435 436 437 438 439 440
}

static inline void __bd_forget(struct inode *inode)
{
	list_del_init(&inode->i_devices);
	inode->i_bdev = NULL;
	inode->i_mapping = &inode->i_data;
}

441
static void bdev_evict_inode(struct inode *inode)
L
Linus Torvalds 已提交
442 443 444
{
	struct block_device *bdev = &BDEV_I(inode)->bdev;
	struct list_head *p;
445 446 447
	truncate_inode_pages(&inode->i_data, 0);
	invalidate_inode_buffers(inode); /* is it needed here? */
	end_writeback(inode);
L
Linus Torvalds 已提交
448 449 450 451 452 453 454 455
	spin_lock(&bdev_lock);
	while ( (p = bdev->bd_inodes.next) != &bdev->bd_inodes ) {
		__bd_forget(list_entry(p, struct inode, i_devices));
	}
	list_del_init(&bdev->bd_list);
	spin_unlock(&bdev_lock);
}

456
static const struct super_operations bdev_sops = {
L
Linus Torvalds 已提交
457 458 459 460
	.statfs = simple_statfs,
	.alloc_inode = bdev_alloc_inode,
	.destroy_inode = bdev_destroy_inode,
	.drop_inode = generic_delete_inode,
461
	.evict_inode = bdev_evict_inode,
L
Linus Torvalds 已提交
462 463
};

A
Al Viro 已提交
464 465
static struct dentry *bd_mount(struct file_system_type *fs_type,
	int flags, const char *dev_name, void *data)
L
Linus Torvalds 已提交
466
{
A
Al Viro 已提交
467
	return mount_pseudo(fs_type, "bdev:", &bdev_sops, 0x62646576);
L
Linus Torvalds 已提交
468 469 470 471
}

static struct file_system_type bd_type = {
	.name		= "bdev",
A
Al Viro 已提交
472
	.mount		= bd_mount,
L
Linus Torvalds 已提交
473 474 475
	.kill_sb	= kill_anon_super,
};

476
struct super_block *blockdev_superblock __read_mostly;
L
Linus Torvalds 已提交
477 478 479 480

void __init bdev_cache_init(void)
{
	int err;
481 482
	struct vfsmount *bd_mnt;

L
Linus Torvalds 已提交
483
	bdev_cachep = kmem_cache_create("bdev_cache", sizeof(struct bdev_inode),
484 485
			0, (SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|
				SLAB_MEM_SPREAD|SLAB_PANIC),
486
			init_once);
L
Linus Torvalds 已提交
487 488 489 490 491 492
	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");
493 494 495 496 497
	/*
	 * This vfsmount structure is only used to obtain the
	 * blockdev_superblock, so tell kmemleak not to report it.
	 */
	kmemleak_not_leak(bd_mnt);
L
Linus Torvalds 已提交
498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528
	blockdev_superblock = bd_mnt->mnt_sb;	/* For writeback */
}

/*
 * 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);

struct block_device *bdget(dev_t dev)
{
	struct block_device *bdev;
	struct inode *inode;

529
	inode = iget5_locked(blockdev_superblock, hash(dev),
L
Linus Torvalds 已提交
530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558
			bdev_test, bdev_set, &dev);

	if (!inode)
		return NULL;

	bdev = &BDEV_I(inode)->bdev;

	if (inode->i_state & I_NEW) {
		bdev->bd_contains = NULL;
		bdev->bd_inode = inode;
		bdev->bd_block_size = (1 << inode->i_blkbits);
		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);
		inode->i_data.backing_dev_info = &default_backing_dev_info;
		spin_lock(&bdev_lock);
		list_add(&bdev->bd_list, &all_bdevs);
		spin_unlock(&bdev_lock);
		unlock_new_inode(inode);
	}
	return bdev;
}

EXPORT_SYMBOL(bdget);

559 560 561 562 563 564
/**
 * 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)
{
A
Al Viro 已提交
565
	ihold(bdev->bd_inode);
566 567 568
	return bdev;
}

L
Linus Torvalds 已提交
569 570
long nr_blockdev_pages(void)
{
571
	struct block_device *bdev;
L
Linus Torvalds 已提交
572 573
	long ret = 0;
	spin_lock(&bdev_lock);
574
	list_for_each_entry(bdev, &all_bdevs, bd_list) {
L
Linus Torvalds 已提交
575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590
		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;
591

L
Linus Torvalds 已提交
592 593
	spin_lock(&bdev_lock);
	bdev = inode->i_bdev;
594
	if (bdev) {
A
Al Viro 已提交
595
		ihold(bdev->bd_inode);
L
Linus Torvalds 已提交
596 597 598 599
		spin_unlock(&bdev_lock);
		return bdev;
	}
	spin_unlock(&bdev_lock);
600

L
Linus Torvalds 已提交
601 602 603
	bdev = bdget(inode->i_rdev);
	if (bdev) {
		spin_lock(&bdev_lock);
604 605
		if (!inode->i_bdev) {
			/*
A
Al Viro 已提交
606
			 * We take an additional reference to bd_inode,
607 608 609 610
			 * and it's released in clear_inode() of inode.
			 * So, we can access it via ->i_mapping always
			 * without igrab().
			 */
A
Al Viro 已提交
611
			ihold(bdev->bd_inode);
612 613 614 615
			inode->i_bdev = bdev;
			inode->i_mapping = bdev->bd_inode->i_mapping;
			list_add(&inode->i_devices, &bdev->bd_inodes);
		}
L
Linus Torvalds 已提交
616 617 618 619 620 621 622 623 624
		spin_unlock(&bdev_lock);
	}
	return bdev;
}

/* Call when you free inode */

void bd_forget(struct inode *inode)
{
625 626
	struct block_device *bdev = NULL;

L
Linus Torvalds 已提交
627
	spin_lock(&bdev_lock);
628
	if (inode->i_bdev) {
629
		if (!sb_is_blkdev_sb(inode->i_sb))
630
			bdev = inode->i_bdev;
L
Linus Torvalds 已提交
631
		__bd_forget(inode);
632
	}
L
Linus Torvalds 已提交
633
	spin_unlock(&bdev_lock);
634 635 636

	if (bdev)
		iput(bdev->bd_inode);
L
Linus Torvalds 已提交
637 638
}

T
Tejun Heo 已提交
639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654
/**
 * 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
 *
 * Test whther @bdev can be claimed by @holder.
 *
 * 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)
L
Linus Torvalds 已提交
655 656
{
	if (bdev->bd_holder == holder)
T
Tejun Heo 已提交
657
		return true;	 /* already a holder */
L
Linus Torvalds 已提交
658
	else if (bdev->bd_holder != NULL)
T
Tejun Heo 已提交
659
		return false; 	 /* held by someone else */
L
Linus Torvalds 已提交
660
	else if (bdev->bd_contains == bdev)
T
Tejun Heo 已提交
661
		return true;  	 /* is a whole device which isn't held */
L
Linus Torvalds 已提交
662

663
	else if (whole->bd_holder == bd_may_claim)
T
Tejun Heo 已提交
664 665 666
		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 */
L
Linus Torvalds 已提交
667
	else
T
Tejun Heo 已提交
668 669 670
		return true;	 /* is a partition of an un-held device */
}

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
/**
 * 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;

697 698
	/* if claiming is already in progress, wait for it to finish */
	if (whole->bd_claiming) {
699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721
		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
N
Nick Piggin 已提交
722 723 724 725 726 727 728
 * 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.
729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754
 *
 * 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);

	whole = bdget_disk(disk, 0);
755
	module_put(disk->fops->owner);
756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774
	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);
	}
}

775
#ifdef CONFIG_SYSFS
776
static int add_symlink(struct kobject *from, struct kobject *to)
777
{
778
	return sysfs_create_link(from, to, kobject_name(to));
779 780 781 782 783 784 785
}

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

J
Jun'ichi Nomura 已提交
786
/**
787 788 789
 * bd_link_disk_holder - create symlinks between holding disk and slave bdev
 * @bdev: the claimed slave bdev
 * @disk: the holding disk
J
Jun'ichi Nomura 已提交
790
 *
791
 * This functions creates the following sysfs symlinks.
J
Jun'ichi Nomura 已提交
792
 *
793 794
 * - from "slaves" directory of the holder @disk to the claimed @bdev
 * - from "holders" directory of the @bdev to the holder @disk
795
 *
796 797
 * For example, if /dev/dm-0 maps to /dev/sda and disk for dm-0 is
 * passed to bd_link_disk_holder(), then:
798
 *
799 800
 *   /sys/block/dm-0/slaves/sda --> /sys/block/sda
 *   /sys/block/sda/holders/dm-0 --> /sys/block/dm-0
801
 *
802 803 804
 * 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.
805
 *
806 807
 * CONTEXT:
 * Might sleep.
808
 *
809 810
 * RETURNS:
 * 0 on success, -errno on failure.
811
 */
812
int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk)
813
{
814
	int ret = 0;
815

816
	mutex_lock(&bdev->bd_mutex);
J
Jun'ichi Nomura 已提交
817

818
	WARN_ON_ONCE(!bdev->bd_holder || bdev->bd_holder_disk);
819

820 821 822
	/* FIXME: remove the following once add_disk() handles errors */
	if (WARN_ON(!disk->slave_dir || !bdev->bd_part->holder_dir))
		goto out_unlock;
823

824 825 826
	ret = add_symlink(disk->slave_dir, &part_to_dev(bdev->bd_part)->kobj);
	if (ret)
		goto out_unlock;
827

828 829 830 831 832
	ret = add_symlink(bdev->bd_part->holder_dir, &disk_to_dev(disk)->kobj);
	if (ret) {
		del_symlink(disk->slave_dir, &part_to_dev(bdev->bd_part)->kobj);
		goto out_unlock;
	}
833

834 835
	bdev->bd_holder_disk = disk;
out_unlock:
836
	mutex_unlock(&bdev->bd_mutex);
837
	return ret;
838
}
839
EXPORT_SYMBOL_GPL(bd_link_disk_holder);
840

841
static void bd_unlink_disk_holder(struct block_device *bdev)
842
{
843
	struct gendisk *disk = bdev->bd_holder_disk;
844

845 846 847 848 849 850
	bdev->bd_holder_disk = NULL;
	if (!disk)
		return;

	del_symlink(disk->slave_dir, &part_to_dev(bdev->bd_part)->kobj);
	del_symlink(bdev->bd_part->holder_dir, &disk_to_dev(disk)->kobj);
851
}
852 853 854
#else
static inline void bd_unlink_disk_holder(struct block_device *bdev)
{ }
855 856
#endif

L
Linus Torvalds 已提交
857 858 859 860 861 862 863
/*
 * Tries to open block device by device number.  Use it ONLY if you
 * really do not have anything better - i.e. when you are behind a
 * truly sucky interface and all you are given is a device number.  _Never_
 * to be used for internal purposes.  If you ever need it - reconsider
 * your API.
 */
864
struct block_device *open_by_devnum(dev_t dev, fmode_t mode, void *holder)
L
Linus Torvalds 已提交
865 866 867 868
{
	struct block_device *bdev = bdget(dev);
	int err = -ENOMEM;
	if (bdev)
869
		err = blkdev_get(bdev, mode, holder);
L
Linus Torvalds 已提交
870 871 872 873 874
	return err ? ERR_PTR(err) : bdev;
}

EXPORT_SYMBOL(open_by_devnum);

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
/**
 * flush_disk - invalidates all buffer-cache entries on a disk
 *
 * @bdev:      struct block device to be flushed
 *
 * 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.
 */
static void flush_disk(struct block_device *bdev)
{
	if (__invalidate_device(bdev)) {
		char name[BDEVNAME_SIZE] = "";

		if (bdev->bd_disk)
			disk_name(bdev->bd_disk, 0, name);
		printk(KERN_WARNING "VFS: busy inodes on changed media or "
		       "resized disk %s\n", name);
	}

	if (!bdev->bd_disk)
		return;
	if (disk_partitionable(bdev->bd_disk))
		bdev->bd_invalidated = 1;
}

901
/**
902
 * check_disk_size_change - checks for disk size change and adjusts bdev size.
903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922
 * @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) {
		char name[BDEVNAME_SIZE];

		disk_name(disk, 0, name);
		printk(KERN_INFO
		       "%s: detected capacity change from %lld to %lld\n",
		       name, bdev_size, disk_size);
		i_size_write(bdev->bd_inode, disk_size);
923
		flush_disk(bdev);
924 925 926 927
	}
}
EXPORT_SYMBOL(check_disk_size_change);

928
/**
929
 * revalidate_disk - wrapper for lower-level driver's revalidate_disk call-back
930 931 932 933 934 935 936 937
 * @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)
{
938
	struct block_device *bdev;
939 940 941 942 943
	int ret = 0;

	if (disk->fops->revalidate_disk)
		ret = disk->fops->revalidate_disk(disk);

944 945 946 947 948 949 950 951
	bdev = bdget_disk(disk, 0);
	if (!bdev)
		return ret;

	mutex_lock(&bdev->bd_mutex);
	check_disk_size_change(disk, bdev);
	mutex_unlock(&bdev->bd_mutex);
	bdput(bdev);
952 953 954 955
	return ret;
}
EXPORT_SYMBOL(revalidate_disk);

L
Linus Torvalds 已提交
956 957 958 959 960 961 962 963 964 965 966 967
/*
 * 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;
968
	const struct block_device_operations *bdops = disk->fops;
L
Linus Torvalds 已提交
969 970 971 972 973 974

	if (!bdops->media_changed)
		return 0;
	if (!bdops->media_changed(bdev->bd_disk))
		return 0;

975
	flush_disk(bdev);
L
Linus Torvalds 已提交
976 977 978 979 980 981 982 983 984
	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)
{
985
	unsigned bsize = bdev_logical_block_size(bdev);
L
Linus Torvalds 已提交
986 987 988 989 990 991 992 993 994 995 996 997

	bdev->bd_inode->i_size = size;
	while (bsize < PAGE_CACHE_SIZE) {
		if (size & bsize)
			break;
		bsize <<= 1;
	}
	bdev->bd_block_size = bsize;
	bdev->bd_inode->i_blkbits = blksize_bits(bsize);
}
EXPORT_SYMBOL(bd_set_size);

A
Al Viro 已提交
998
static int __blkdev_put(struct block_device *bdev, fmode_t mode, int for_part);
999

1000 1001 1002 1003 1004 1005 1006
/*
 * bd_mutex locking:
 *
 *  mutex_lock(part->bd_mutex)
 *    mutex_lock_nested(whole->bd_mutex, 1)
 */

1007
static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part)
L
Linus Torvalds 已提交
1008 1009
{
	struct gendisk *disk;
1010
	int ret;
1011
	int partno;
1012 1013
	int perm = 0;

1014
	if (mode & FMODE_READ)
1015
		perm |= MAY_READ;
1016
	if (mode & FMODE_WRITE)
1017 1018 1019 1020
		perm |= MAY_WRITE;
	/*
	 * hooks: /n/, see "layering violations".
	 */
1021 1022 1023 1024 1025 1026
	if (!for_part) {
		ret = devcgroup_inode_permission(bdev->bd_inode, perm);
		if (ret != 0) {
			bdput(bdev);
			return ret;
		}
1027
	}
1028

1029
 restart:
T
Tejun Heo 已提交
1030

1031
	ret = -ENXIO;
1032
	disk = get_gendisk(bdev->bd_dev, &partno);
T
Tejun Heo 已提交
1033
	if (!disk)
1034
		goto out;
L
Linus Torvalds 已提交
1035

1036
	mutex_lock_nested(&bdev->bd_mutex, for_part);
L
Linus Torvalds 已提交
1037 1038 1039
	if (!bdev->bd_openers) {
		bdev->bd_disk = disk;
		bdev->bd_contains = bdev;
1040
		if (!partno) {
L
Linus Torvalds 已提交
1041
			struct backing_dev_info *bdi;
1042 1043 1044 1045 1046 1047

			ret = -ENXIO;
			bdev->bd_part = disk_get_part(disk, partno);
			if (!bdev->bd_part)
				goto out_clear;

L
Linus Torvalds 已提交
1048
			if (disk->fops->open) {
1049
				ret = disk->fops->open(bdev, mode);
1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062
				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;
					module_put(disk->fops->owner);
					put_disk(disk);
					bdev->bd_disk = NULL;
					mutex_unlock(&bdev->bd_mutex);
					goto restart;
				}
L
Linus Torvalds 已提交
1063
				if (ret)
T
Tejun Heo 已提交
1064
					goto out_clear;
L
Linus Torvalds 已提交
1065 1066 1067 1068 1069 1070
			}
			if (!bdev->bd_openers) {
				bd_set_size(bdev,(loff_t)get_capacity(disk)<<9);
				bdi = blk_get_backing_dev_info(bdev);
				if (bdi == NULL)
					bdi = &default_backing_dev_info;
1071
				bdev_inode_switch_bdi(bdev->bd_inode, bdi);
L
Linus Torvalds 已提交
1072 1073 1074 1075 1076 1077 1078 1079
			}
			if (bdev->bd_invalidated)
				rescan_partitions(disk, bdev);
		} else {
			struct block_device *whole;
			whole = bdget_disk(disk, 0);
			ret = -ENOMEM;
			if (!whole)
T
Tejun Heo 已提交
1080
				goto out_clear;
1081
			BUG_ON(for_part);
1082
			ret = __blkdev_get(whole, mode, 1);
L
Linus Torvalds 已提交
1083
			if (ret)
T
Tejun Heo 已提交
1084
				goto out_clear;
L
Linus Torvalds 已提交
1085
			bdev->bd_contains = whole;
1086 1087
			bdev_inode_switch_bdi(bdev->bd_inode,
				whole->bd_inode->i_data.backing_dev_info);
1088
			bdev->bd_part = disk_get_part(disk, partno);
1089
			if (!(disk->flags & GENHD_FL_UP) ||
1090
			    !bdev->bd_part || !bdev->bd_part->nr_sects) {
L
Linus Torvalds 已提交
1091
				ret = -ENXIO;
T
Tejun Heo 已提交
1092
				goto out_clear;
L
Linus Torvalds 已提交
1093
			}
1094
			bd_set_size(bdev, (loff_t)bdev->bd_part->nr_sects << 9);
L
Linus Torvalds 已提交
1095 1096
		}
	} else {
T
Tejun Heo 已提交
1097
		module_put(disk->fops->owner);
1098
		put_disk(disk);
T
Tejun Heo 已提交
1099
		disk = NULL;
L
Linus Torvalds 已提交
1100 1101
		if (bdev->bd_contains == bdev) {
			if (bdev->bd_disk->fops->open) {
1102
				ret = bdev->bd_disk->fops->open(bdev, mode);
L
Linus Torvalds 已提交
1103
				if (ret)
T
Tejun Heo 已提交
1104
					goto out_unlock_bdev;
L
Linus Torvalds 已提交
1105 1106 1107 1108 1109 1110
			}
			if (bdev->bd_invalidated)
				rescan_partitions(bdev->bd_disk, bdev);
		}
	}
	bdev->bd_openers++;
1111 1112
	if (for_part)
		bdev->bd_part_count++;
1113
	mutex_unlock(&bdev->bd_mutex);
L
Linus Torvalds 已提交
1114 1115
	return 0;

T
Tejun Heo 已提交
1116
 out_clear:
1117
	disk_put_part(bdev->bd_part);
L
Linus Torvalds 已提交
1118
	bdev->bd_disk = NULL;
T
Tejun Heo 已提交
1119
	bdev->bd_part = NULL;
1120
	bdev_inode_switch_bdi(bdev->bd_inode, &default_backing_dev_info);
L
Linus Torvalds 已提交
1121
	if (bdev != bdev->bd_contains)
1122
		__blkdev_put(bdev->bd_contains, mode, 1);
L
Linus Torvalds 已提交
1123
	bdev->bd_contains = NULL;
T
Tejun Heo 已提交
1124
 out_unlock_bdev:
1125
	mutex_unlock(&bdev->bd_mutex);
1126
 out:
T
Tejun Heo 已提交
1127 1128 1129 1130 1131
	if (disk)
		module_put(disk->fops->owner);
	put_disk(disk);
	bdput(bdev);

L
Linus Torvalds 已提交
1132 1133 1134
	return ret;
}

1135
int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder)
L
Linus Torvalds 已提交
1136
{
1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151
	struct block_device *whole = NULL;
	int res;

	WARN_ON_ONCE((mode & FMODE_EXCL) && !holder);

	if ((mode & FMODE_EXCL) && holder) {
		whole = bd_start_claiming(bdev, holder);
		if (IS_ERR(whole)) {
			bdput(bdev);
			return PTR_ERR(whole);
		}
	}

	res = __blkdev_get(bdev, mode, 0);

1152 1153 1154 1155 1156 1157
	/* __blkdev_get() may alter read only status, check it afterwards */
	if (!res && (mode & FMODE_WRITE) && bdev_read_only(bdev)) {
		__blkdev_put(bdev, mode, 0);
		res = -EACCES;
	}

1158
	if (whole) {
1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182
		/* finish claiming */
		spin_lock(&bdev_lock);

		if (res == 0) {
			BUG_ON(!bd_may_claim(bdev, whole, holder));
			/*
			 * Note that for a whole device bd_holders
			 * will be incremented twice, and bd_holder
			 * will be set to bd_may_claim before being
			 * set to holder
			 */
			whole->bd_holders++;
			whole->bd_holder = bd_may_claim;
			bdev->bd_holders++;
			bdev->bd_holder = holder;
		}

		/* tell others that we're done */
		BUG_ON(whole->bd_claiming != holder);
		whole->bd_claiming = NULL;
		wake_up_bit(&whole->bd_claiming, 0);

		spin_unlock(&bdev_lock);
		bdput(whole);
1183 1184 1185
	}

	return res;
1186
}
L
Linus Torvalds 已提交
1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200
EXPORT_SYMBOL(blkdev_get);

static int blkdev_open(struct inode * inode, struct file * filp)
{
	struct block_device *bdev;

	/*
	 * Preserve backwards compatibility and allow large file access
	 * even if userspace doesn't ask for it explicitly. Some mkfs
	 * binary needs it. We might want to drop this workaround
	 * during an unstable branch.
	 */
	filp->f_flags |= O_LARGEFILE;

1201 1202 1203 1204 1205 1206 1207
	if (filp->f_flags & O_NDELAY)
		filp->f_mode |= FMODE_NDELAY;
	if (filp->f_flags & O_EXCL)
		filp->f_mode |= FMODE_EXCL;
	if ((filp->f_flags & O_ACCMODE) == 3)
		filp->f_mode |= FMODE_WRITE_IOCTL;

L
Linus Torvalds 已提交
1208
	bdev = bd_acquire(inode);
1209 1210
	if (bdev == NULL)
		return -ENOMEM;
L
Linus Torvalds 已提交
1211

1212 1213
	filp->f_mapping = bdev->bd_inode->i_mapping;

1214
	return blkdev_get(bdev, filp->f_mode, filp);
L
Linus Torvalds 已提交
1215 1216
}

A
Al Viro 已提交
1217
static int __blkdev_put(struct block_device *bdev, fmode_t mode, int for_part)
1218 1219 1220
{
	int ret = 0;
	struct gendisk *disk = bdev->bd_disk;
1221
	struct block_device *victim = NULL;
1222

1223
	mutex_lock_nested(&bdev->bd_mutex, for_part);
1224 1225 1226
	if (for_part)
		bdev->bd_part_count--;

1227
	if (!--bdev->bd_openers) {
1228
		WARN_ON_ONCE(bdev->bd_holders);
1229 1230 1231 1232 1233
		sync_blockdev(bdev);
		kill_bdev(bdev);
	}
	if (bdev->bd_contains == bdev) {
		if (disk->fops->release)
A
Al Viro 已提交
1234
			ret = disk->fops->release(disk, mode);
1235 1236 1237 1238 1239 1240
	}
	if (!bdev->bd_openers) {
		struct module *owner = disk->fops->owner;

		put_disk(disk);
		module_put(owner);
T
Tejun Heo 已提交
1241 1242
		disk_put_part(bdev->bd_part);
		bdev->bd_part = NULL;
1243
		bdev->bd_disk = NULL;
1244 1245
		bdev_inode_switch_bdi(bdev->bd_inode,
					&default_backing_dev_info);
1246 1247
		if (bdev != bdev->bd_contains)
			victim = bdev->bd_contains;
1248 1249 1250 1251
		bdev->bd_contains = NULL;
	}
	mutex_unlock(&bdev->bd_mutex);
	bdput(bdev);
1252
	if (victim)
A
Al Viro 已提交
1253
		__blkdev_put(victim, mode, 1);
1254 1255 1256
	return ret;
}

A
Al Viro 已提交
1257
int blkdev_put(struct block_device *bdev, fmode_t mode)
1258
{
1259
	if (mode & FMODE_EXCL) {
1260 1261 1262 1263 1264 1265 1266
		bool bdev_free;

		/*
		 * Release a claim on the device.  The holder fields
		 * are protected with bdev_lock.  bd_mutex is to
		 * synchronize disk_holder unlinking.
		 */
1267
		mutex_lock(&bdev->bd_mutex);
1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282
		spin_lock(&bdev_lock);

		WARN_ON_ONCE(--bdev->bd_holders < 0);
		WARN_ON_ONCE(--bdev->bd_contains->bd_holders < 0);

		/* bd_contains might point to self, check in a separate step */
		if ((bdev_free = !bdev->bd_holders))
			bdev->bd_holder = NULL;
		if (!bdev->bd_contains->bd_holders)
			bdev->bd_contains->bd_holder = NULL;

		spin_unlock(&bdev_lock);

		/* if this was the last claim, holder link should go too */
		if (bdev_free)
1283
			bd_unlink_disk_holder(bdev);
1284

1285 1286
		mutex_unlock(&bdev->bd_mutex);
	}
A
Al Viro 已提交
1287
	return __blkdev_put(bdev, mode, 0);
1288
}
1289 1290
EXPORT_SYMBOL(blkdev_put);

L
Linus Torvalds 已提交
1291 1292 1293
static int blkdev_close(struct inode * inode, struct file * filp)
{
	struct block_device *bdev = I_BDEV(filp->f_mapping->host);
1294

A
Al Viro 已提交
1295
	return blkdev_put(bdev, filp->f_mode);
L
Linus Torvalds 已提交
1296 1297
}

1298
static long block_ioctl(struct file *file, unsigned cmd, unsigned long arg)
L
Linus Torvalds 已提交
1299
{
1300 1301
	struct block_device *bdev = I_BDEV(file->f_mapping->host);
	fmode_t mode = file->f_mode;
1302 1303 1304 1305 1306

	/*
	 * O_NDELAY can be altered using fcntl(.., F_SETFL, ..), so we have
	 * to updated it before every ioctl.
	 */
1307
	if (file->f_flags & O_NDELAY)
1308 1309 1310 1311
		mode |= FMODE_NDELAY;
	else
		mode &= ~FMODE_NDELAY;

1312
	return blkdev_ioctl(bdev, mode, cmd, arg);
L
Linus Torvalds 已提交
1313 1314
}

1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341
/*
 * Write data to the block device.  Only intended for the block device itself
 * and the raw driver which basically is a fake block device.
 *
 * Does not take i_mutex for the write and thus is not for general purpose
 * use.
 */
ssize_t blkdev_aio_write(struct kiocb *iocb, const struct iovec *iov,
			 unsigned long nr_segs, loff_t pos)
{
	struct file *file = iocb->ki_filp;
	ssize_t ret;

	BUG_ON(iocb->ki_pos != pos);

	ret = __generic_file_aio_write(iocb, iov, nr_segs, &iocb->ki_pos);
	if (ret > 0 || ret == -EIOCBQUEUED) {
		ssize_t err;

		err = generic_write_sync(file, pos, ret);
		if (err < 0 && ret > 0)
			ret = err;
	}
	return ret;
}
EXPORT_SYMBOL_GPL(blkdev_aio_write);

1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355
/*
 * Try to release a page associated with block device when the system
 * is under memory pressure.
 */
static int blkdev_releasepage(struct page *page, gfp_t wait)
{
	struct super_block *super = BDEV_I(page->mapping->host)->bdev.bd_super;

	if (super && super->s_op->bdev_try_to_free_page)
		return super->s_op->bdev_try_to_free_page(super, page, wait);

	return try_to_free_buffers(page);
}

A
Adrian Bunk 已提交
1356
static const struct address_space_operations def_blk_aops = {
L
Linus Torvalds 已提交
1357 1358 1359
	.readpage	= blkdev_readpage,
	.writepage	= blkdev_writepage,
	.sync_page	= block_sync_page,
N
Nick Piggin 已提交
1360 1361
	.write_begin	= blkdev_write_begin,
	.write_end	= blkdev_write_end,
L
Linus Torvalds 已提交
1362
	.writepages	= generic_writepages,
1363
	.releasepage	= blkdev_releasepage,
L
Linus Torvalds 已提交
1364 1365 1366
	.direct_IO	= blkdev_direct_IO,
};

1367
const struct file_operations def_blk_fops = {
L
Linus Torvalds 已提交
1368 1369 1370
	.open		= blkdev_open,
	.release	= blkdev_close,
	.llseek		= block_llseek,
1371 1372
	.read		= do_sync_read,
	.write		= do_sync_write,
L
Linus Torvalds 已提交
1373
  	.aio_read	= generic_file_aio_read,
1374
	.aio_write	= blkdev_aio_write,
L
Linus Torvalds 已提交
1375
	.mmap		= generic_file_mmap,
1376
	.fsync		= blkdev_fsync,
1377
	.unlocked_ioctl	= block_ioctl,
L
Linus Torvalds 已提交
1378 1379 1380
#ifdef CONFIG_COMPAT
	.compat_ioctl	= compat_blkdev_ioctl,
#endif
1381 1382
	.splice_read	= generic_file_splice_read,
	.splice_write	= generic_file_splice_write,
L
Linus Torvalds 已提交
1383 1384 1385 1386 1387 1388 1389
};

int ioctl_by_bdev(struct block_device *bdev, unsigned cmd, unsigned long arg)
{
	int res;
	mm_segment_t old_fs = get_fs();
	set_fs(KERNEL_DS);
1390
	res = blkdev_ioctl(bdev, 0, cmd, arg);
L
Linus Torvalds 已提交
1391 1392 1393 1394 1395 1396 1397 1398
	set_fs(old_fs);
	return res;
}

EXPORT_SYMBOL(ioctl_by_bdev);

/**
 * lookup_bdev  - lookup a struct block_device by name
1399
 * @pathname:	special file representing the block device
L
Linus Torvalds 已提交
1400
 *
1401
 * Get a reference to the blockdevice at @pathname in the current
L
Linus Torvalds 已提交
1402 1403 1404
 * namespace if possible and return it.  Return ERR_PTR(error)
 * otherwise.
 */
1405
struct block_device *lookup_bdev(const char *pathname)
L
Linus Torvalds 已提交
1406 1407 1408
{
	struct block_device *bdev;
	struct inode *inode;
1409
	struct path path;
L
Linus Torvalds 已提交
1410 1411
	int error;

1412
	if (!pathname || !*pathname)
L
Linus Torvalds 已提交
1413 1414
		return ERR_PTR(-EINVAL);

1415
	error = kern_path(pathname, LOOKUP_FOLLOW, &path);
L
Linus Torvalds 已提交
1416 1417 1418
	if (error)
		return ERR_PTR(error);

1419
	inode = path.dentry->d_inode;
L
Linus Torvalds 已提交
1420 1421 1422 1423
	error = -ENOTBLK;
	if (!S_ISBLK(inode->i_mode))
		goto fail;
	error = -EACCES;
1424
	if (path.mnt->mnt_flags & MNT_NODEV)
L
Linus Torvalds 已提交
1425 1426 1427 1428 1429 1430
		goto fail;
	error = -ENOMEM;
	bdev = bd_acquire(inode);
	if (!bdev)
		goto fail;
out:
1431
	path_put(&path);
L
Linus Torvalds 已提交
1432 1433 1434 1435 1436
	return bdev;
fail:
	bdev = ERR_PTR(error);
	goto out;
}
1437
EXPORT_SYMBOL(lookup_bdev);
L
Linus Torvalds 已提交
1438 1439

/**
1440
 * open_bdev_exclusive  -  open a block device by name and set it up for use
L
Linus Torvalds 已提交
1441 1442
 *
 * @path:	special file representing the block device
1443
 * @mode:	FMODE_... combination to pass be used
L
Linus Torvalds 已提交
1444 1445 1446 1447 1448
 * @holder:	owner for exclusion
 *
 * Open the blockdevice described by the special file at @path, claim it
 * for the @holder.
 */
1449
struct block_device *open_bdev_exclusive(const char *path, fmode_t mode, void *holder)
L
Linus Torvalds 已提交
1450
{
1451
	struct block_device *bdev;
1452
	int error;
L
Linus Torvalds 已提交
1453 1454 1455 1456 1457

	bdev = lookup_bdev(path);
	if (IS_ERR(bdev))
		return bdev;

1458
	error = blkdev_get(bdev, mode | FMODE_EXCL, holder);
L
Linus Torvalds 已提交
1459
	if (error)
1460
		return ERR_PTR(error);
1461

L
Linus Torvalds 已提交
1462 1463 1464
	return bdev;
}

1465
EXPORT_SYMBOL(open_bdev_exclusive);
L
Linus Torvalds 已提交
1466

1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482
int __invalidate_device(struct block_device *bdev)
{
	struct super_block *sb = get_super(bdev);
	int res = 0;

	if (sb) {
		/*
		 * no need to lock the super, get_super holds the
		 * read mutex so the filesystem cannot go away
		 * under us (->put_super runs with the write lock
		 * hold).
		 */
		shrink_dcache_sb(sb);
		res = invalidate_inodes(sb);
		drop_super(sb);
	}
1483
	invalidate_bdev(bdev);
1484 1485 1486
	return res;
}
EXPORT_SYMBOL(__invalidate_device);