block_dev.c 45.1 KB
Newer Older
L
Linus Torvalds 已提交
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>
L
Linus Torvalds 已提交
15 16 17 18
#include <linux/highmem.h>
#include <linux/blkdev.h>
#include <linux/module.h>
#include <linux/blkpg.h>
19
#include <linux/magic.h>
L
Linus Torvalds 已提交
20
#include <linux/buffer_head.h>
A
Al Viro 已提交
21
#include <linux/swap.h>
N
Nick Piggin 已提交
22
#include <linux/pagevec.h>
23
#include <linux/writeback.h>
L
Linus Torvalds 已提交
24 25 26 27
#include <linux/mpage.h>
#include <linux/mount.h>
#include <linux/uio.h>
#include <linux/namei.h>
28
#include <linux/log2.h>
A
Al Viro 已提交
29
#include <linux/cleancache.h>
L
Linus Torvalds 已提交
30
#include <asm/uaccess.h>
31
#include "internal.h"
L
Linus Torvalds 已提交
32 33 34 35 36 37

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

A
Adrian Bunk 已提交
38 39
static const struct address_space_operations def_blk_aops;

L
Linus Torvalds 已提交
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
static void bdev_write_inode(struct inode *inode)
{
	spin_lock(&inode->i_lock);
	while (inode->i_state & I_DIRTY) {
		spin_unlock(&inode->i_lock);
		WARN_ON_ONCE(write_inode_now(inode, true));
		spin_lock(&inode->i_lock);
	}
	spin_unlock(&inode->i_lock);
}

P
Peter Zijlstra 已提交
62
/* Kill _all_ buffers and pagecache , dirty or not.. */
A
Al Viro 已提交
63
void kill_bdev(struct block_device *bdev)
L
Linus Torvalds 已提交
64
{
A
Al Viro 已提交
65 66
	struct address_space *mapping = bdev->bd_inode->i_mapping;

67
	if (mapping->nrpages == 0 && mapping->nrshadows == 0)
P
Peter Zijlstra 已提交
68
		return;
A
Al Viro 已提交
69

P
Peter Zijlstra 已提交
70
	invalidate_bh_lrus();
A
Al Viro 已提交
71
	truncate_inode_pages(mapping, 0);
L
Linus Torvalds 已提交
72
}	
A
Al Viro 已提交
73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88
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;

	if (mapping->nrpages == 0)
		return;

	invalidate_bh_lrus();
	lru_add_drain_all();	/* make sure all lru add caches are flushed */
	invalidate_mapping_pages(mapping, 0, -1);
	/* 99% of the time, we don't need to flush the cleancache on the bdev.
	 * But, for the strange corners, lets be cautious
	 */
89
	cleancache_invalidate_inode(mapping);
A
Al Viro 已提交
90 91
}
EXPORT_SYMBOL(invalidate_bdev);
L
Linus Torvalds 已提交
92 93 94 95

int set_blocksize(struct block_device *bdev, int size)
{
	/* Size must be a power of two, and between 512 and PAGE_SIZE */
96
	if (size > PAGE_SIZE || size < 512 || !is_power_of_2(size))
L
Linus Torvalds 已提交
97 98 99
		return -EINVAL;

	/* Size cannot be smaller than the size supported by the device */
100
	if (size < bdev_logical_block_size(bdev))
L
Linus Torvalds 已提交
101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121
		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;
122
	sb->s_blocksize_bits = blksize_bits(size);
L
Linus Torvalds 已提交
123 124 125 126 127 128 129
	return sb->s_blocksize;
}

EXPORT_SYMBOL(sb_set_blocksize);

int sb_min_blocksize(struct super_block *sb, int size)
{
130
	int minsize = bdev_logical_block_size(sb->s_bdev);
L
Linus Torvalds 已提交
131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147
	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;
}

148
static ssize_t
149
blkdev_direct_IO(struct kiocb *iocb, struct iov_iter *iter, loff_t offset)
150 151 152 153
{
	struct file *file = iocb->ki_filp;
	struct inode *inode = file->f_mapping->host;

154 155
	return __blockdev_direct_IO(iocb, inode, I_BDEV(inode), iter, offset,
				    blkdev_get_block, NULL, NULL, 0);
156 157
}

158 159 160 161 162 163 164 165 166
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
167 168 169 170 171 172
/*
 * 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)
{
173
	return __sync_blockdev(bdev, 1);
N
Nick Piggin 已提交
174 175 176 177 178 179 180 181 182 183 184 185
}
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) {
186
		int res = sync_filesystem(sb);
N
Nick Piggin 已提交
187 188 189 190 191
		drop_super(sb);
		return res;
	}
	return sync_blockdev(bdev);
}
192
EXPORT_SYMBOL(fsync_bdev);
N
Nick Piggin 已提交
193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211

/**
 * 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);
212 213 214 215 216 217
	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
218
		sb = get_super(bdev);
219 220 221 222 223 224 225 226
		drop_super(sb);
		mutex_unlock(&bdev->bd_fsfreeze_mutex);
		return sb;
	}

	sb = get_active_super(bdev);
	if (!sb)
		goto out;
227 228 229 230
	if (sb->s_op->freeze_super)
		error = sb->s_op->freeze_super(sb);
	else
		error = freeze_super(sb);
231 232 233
	if (error) {
		deactivate_super(sb);
		bdev->bd_fsfreeze_count--;
N
Nick Piggin 已提交
234
		mutex_unlock(&bdev->bd_fsfreeze_mutex);
235
		return ERR_PTR(error);
N
Nick Piggin 已提交
236
	}
237
	deactivate_super(sb);
238
 out:
N
Nick Piggin 已提交
239 240
	sync_blockdev(bdev);
	mutex_unlock(&bdev->bd_fsfreeze_mutex);
241
	return sb;	/* thaw_bdev releases s->s_umount */
N
Nick Piggin 已提交
242 243 244 245 246 247 248 249 250 251 252 253
}
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)
{
254
	int error = -EINVAL;
N
Nick Piggin 已提交
255 256

	mutex_lock(&bdev->bd_fsfreeze_mutex);
257
	if (!bdev->bd_fsfreeze_count)
258
		goto out;
259 260 261

	error = 0;
	if (--bdev->bd_fsfreeze_count > 0)
262
		goto out;
263 264

	if (!sb)
265
		goto out;
266

267 268 269 270
	if (sb->s_op->thaw_super)
		error = sb->s_op->thaw_super(sb);
	else
		error = thaw_super(sb);
271 272 273 274 275 276
	if (error) {
		bdev->bd_fsfreeze_count++;
		mutex_unlock(&bdev->bd_fsfreeze_mutex);
		return error;
	}
out:
N
Nick Piggin 已提交
277 278 279 280 281
	mutex_unlock(&bdev->bd_fsfreeze_mutex);
	return 0;
}
EXPORT_SYMBOL(thaw_bdev);

L
Linus Torvalds 已提交
282 283 284 285 286 287 288 289 290 291
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);
}

292 293 294 295 296 297
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);
}

N
Nick Piggin 已提交
298 299 300
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
301
{
302 303
	return block_write_begin(mapping, pos, len, flags, pagep,
				 blkdev_get_block);
L
Linus Torvalds 已提交
304 305
}

N
Nick Piggin 已提交
306 307 308
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
309
{
N
Nick Piggin 已提交
310 311 312 313 314 315 316
	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
317 318 319 320
}

/*
 * private llseek:
A
Al Viro 已提交
321
 * for a block special file file_inode(file)->i_size is zero
L
Linus Torvalds 已提交
322 323
 * so we compute the size by hand (just as in block_read/write above)
 */
324
static loff_t block_llseek(struct file *file, loff_t offset, int whence)
L
Linus Torvalds 已提交
325 326 327 328
{
	struct inode *bd_inode = file->f_mapping->host;
	loff_t retval;

329
	mutex_lock(&bd_inode->i_mutex);
330
	retval = fixed_size_llseek(file, offset, whence, i_size_read(bd_inode));
331
	mutex_unlock(&bd_inode->i_mutex);
L
Linus Torvalds 已提交
332 333 334
	return retval;
}
	
335
int blkdev_fsync(struct file *filp, loff_t start, loff_t end, int datasync)
L
Linus Torvalds 已提交
336
{
337 338
	struct inode *bd_inode = filp->f_mapping->host;
	struct block_device *bdev = I_BDEV(bd_inode);
339
	int error;
340 341 342 343
	
	error = filemap_write_and_wait_range(filp->f_mapping, start, end);
	if (error)
		return error;
344

345 346 347 348 349
	/*
	 * 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.
	 */
350
	error = blkdev_issue_flush(bdev, GFP_KERNEL, NULL);
351 352
	if (error == -EOPNOTSUPP)
		error = 0;
353

354
	return error;
L
Linus Torvalds 已提交
355
}
356
EXPORT_SYMBOL(blkdev_fsync);
L
Linus Torvalds 已提交
357

358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420
/**
 * 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;
	if (!ops->rw_page)
		return -EOPNOTSUPP;
	return ops->rw_page(bdev, sector + get_start_sect(bdev), page, READ);
}
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;
	int rw = (wbc->sync_mode == WB_SYNC_ALL) ? WRITE_SYNC : WRITE;
	const struct block_device_operations *ops = bdev->bd_disk->fops;
	if (!ops->rw_page)
		return -EOPNOTSUPP;
	set_page_writeback(page);
	result = ops->rw_page(bdev, sector + get_start_sect(bdev), page, rw);
	if (result)
		end_page_writeback(page);
	else
		unlock_page(page);
	return result;
}
EXPORT_SYMBOL_GPL(bdev_write_page);

421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460
/**
 * bdev_direct_access() - Get the address for directly-accessibly memory
 * @bdev: The device containing the memory
 * @sector: The offset within the device
 * @addr: Where to put the address of the memory
 * @pfn: The Page Frame Number for the memory
 * @size: The number of bytes requested
 *
 * If a block device is made up of directly addressable memory, this function
 * will tell the caller the PFN and the address of the memory.  The address
 * may be directly dereferenced within the kernel without the need to call
 * ioremap(), kmap() or similar.  The PFN is suitable for inserting into
 * page tables.
 *
 * Return: negative errno if an error occurs, otherwise the number of bytes
 * accessible at this address.
 */
long bdev_direct_access(struct block_device *bdev, sector_t sector,
			void **addr, unsigned long *pfn, long size)
{
	long avail;
	const struct block_device_operations *ops = bdev->bd_disk->fops;

	if (size < 0)
		return size;
	if (!ops->direct_access)
		return -EOPNOTSUPP;
	if ((sector + DIV_ROUND_UP(size, 512)) >
					part_nr_sects_read(bdev->bd_part))
		return -ERANGE;
	sector += get_start_sect(bdev);
	if (sector % (PAGE_SIZE / 512))
		return -EINVAL;
	avail = ops->direct_access(bdev, sector, addr, pfn, size);
	if (!avail)
		return -ERANGE;
	return min(avail, size);
}
EXPORT_SYMBOL_GPL(bdev_direct_access);

L
Linus Torvalds 已提交
461 462 463 464 465
/*
 * pseudo-fs
 */

static  __cacheline_aligned_in_smp DEFINE_SPINLOCK(bdev_lock);
466
static struct kmem_cache * bdev_cachep __read_mostly;
L
Linus Torvalds 已提交
467 468 469

static struct inode *bdev_alloc_inode(struct super_block *sb)
{
470
	struct bdev_inode *ei = kmem_cache_alloc(bdev_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
471 472 473 474 475
	if (!ei)
		return NULL;
	return &ei->vfs_inode;
}

N
Nick Piggin 已提交
476
static void bdev_i_callback(struct rcu_head *head)
L
Linus Torvalds 已提交
477
{
N
Nick Piggin 已提交
478
	struct inode *inode = container_of(head, struct inode, i_rcu);
L
Linus Torvalds 已提交
479 480 481 482 483
	struct bdev_inode *bdi = BDEV_I(inode);

	kmem_cache_free(bdev_cachep, bdi);
}

N
Nick Piggin 已提交
484 485 486 487 488
static void bdev_destroy_inode(struct inode *inode)
{
	call_rcu(&inode->i_rcu, bdev_i_callback);
}

489
static void init_once(void *foo)
L
Linus Torvalds 已提交
490 491 492 493
{
	struct bdev_inode *ei = (struct bdev_inode *) foo;
	struct block_device *bdev = &ei->bdev;

C
Christoph Lameter 已提交
494 495 496 497
	memset(bdev, 0, sizeof(*bdev));
	mutex_init(&bdev->bd_mutex);
	INIT_LIST_HEAD(&bdev->bd_inodes);
	INIT_LIST_HEAD(&bdev->bd_list);
498 499 500
#ifdef CONFIG_SYSFS
	INIT_LIST_HEAD(&bdev->bd_holder_disks);
#endif
C
Christoph Lameter 已提交
501
	inode_init_once(&ei->vfs_inode);
502 503
	/* Initialize mutex for freeze. */
	mutex_init(&bdev->bd_fsfreeze_mutex);
L
Linus Torvalds 已提交
504 505 506 507 508 509 510 511 512
}

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

513
static void bdev_evict_inode(struct inode *inode)
L
Linus Torvalds 已提交
514 515 516
{
	struct block_device *bdev = &BDEV_I(inode)->bdev;
	struct list_head *p;
517
	truncate_inode_pages_final(&inode->i_data);
518
	invalidate_inode_buffers(inode); /* is it needed here? */
519
	clear_inode(inode);
L
Linus Torvalds 已提交
520 521 522 523 524 525 526 527
	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);
}

528
static const struct super_operations bdev_sops = {
L
Linus Torvalds 已提交
529 530 531 532
	.statfs = simple_statfs,
	.alloc_inode = bdev_alloc_inode,
	.destroy_inode = bdev_destroy_inode,
	.drop_inode = generic_delete_inode,
533
	.evict_inode = bdev_evict_inode,
L
Linus Torvalds 已提交
534 535
};

A
Al Viro 已提交
536 537
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
538
{
539
	return mount_pseudo(fs_type, "bdev:", &bdev_sops, NULL, BDEVFS_MAGIC);
L
Linus Torvalds 已提交
540 541 542 543
}

static struct file_system_type bd_type = {
	.name		= "bdev",
A
Al Viro 已提交
544
	.mount		= bd_mount,
L
Linus Torvalds 已提交
545 546 547
	.kill_sb	= kill_anon_super,
};

A
Al Viro 已提交
548
static struct super_block *blockdev_superblock __read_mostly;
L
Linus Torvalds 已提交
549 550 551 552

void __init bdev_cache_init(void)
{
	int err;
553
	static struct vfsmount *bd_mnt;
554

L
Linus Torvalds 已提交
555
	bdev_cachep = kmem_cache_create("bdev_cache", sizeof(struct bdev_inode),
556 557
			0, (SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|
				SLAB_MEM_SPREAD|SLAB_PANIC),
558
			init_once);
L
Linus Torvalds 已提交
559 560 561 562 563 564
	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");
565
	blockdev_superblock = bd_mnt->mnt_sb;   /* For writeback */
L
Linus Torvalds 已提交
566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595
}

/*
 * 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;

596
	inode = iget5_locked(blockdev_superblock, hash(dev),
L
Linus Torvalds 已提交
597 598 599 600 601 602 603 604 605
			bdev_test, bdev_set, &dev);

	if (!inode)
		return NULL;

	bdev = &BDEV_I(inode)->bdev;

	if (inode->i_state & I_NEW) {
		bdev->bd_contains = NULL;
606
		bdev->bd_super = NULL;
L
Linus Torvalds 已提交
607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625
		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);
		spin_lock(&bdev_lock);
		list_add(&bdev->bd_list, &all_bdevs);
		spin_unlock(&bdev_lock);
		unlock_new_inode(inode);
	}
	return bdev;
}

EXPORT_SYMBOL(bdget);

626 627 628 629 630 631
/**
 * 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
632
	ihold(bdev->bd_inode);
633 634
	return bdev;
}
635
EXPORT_SYMBOL(bdgrab);
636

L
Linus Torvalds 已提交
637 638
long nr_blockdev_pages(void)
{
639
	struct block_device *bdev;
L
Linus Torvalds 已提交
640 641
	long ret = 0;
	spin_lock(&bdev_lock);
642
	list_for_each_entry(bdev, &all_bdevs, bd_list) {
L
Linus Torvalds 已提交
643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658
		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;
659

L
Linus Torvalds 已提交
660 661
	spin_lock(&bdev_lock);
	bdev = inode->i_bdev;
662
	if (bdev) {
A
Al Viro 已提交
663
		ihold(bdev->bd_inode);
L
Linus Torvalds 已提交
664 665 666 667
		spin_unlock(&bdev_lock);
		return bdev;
	}
	spin_unlock(&bdev_lock);
668

L
Linus Torvalds 已提交
669 670 671
	bdev = bdget(inode->i_rdev);
	if (bdev) {
		spin_lock(&bdev_lock);
672 673
		if (!inode->i_bdev) {
			/*
A
Al Viro 已提交
674
			 * We take an additional reference to bd_inode,
675 676 677 678
			 * 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
679
			ihold(bdev->bd_inode);
680 681 682 683
			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
684 685 686 687 688
		spin_unlock(&bdev_lock);
	}
	return bdev;
}

689
int sb_is_blkdev_sb(struct super_block *sb)
A
Al Viro 已提交
690 691 692 693
{
	return sb == blockdev_superblock;
}

L
Linus Torvalds 已提交
694 695 696 697
/* Call when you free inode */

void bd_forget(struct inode *inode)
{
698 699
	struct block_device *bdev = NULL;

L
Linus Torvalds 已提交
700
	spin_lock(&bdev_lock);
701 702 703
	if (!sb_is_blkdev_sb(inode->i_sb))
		bdev = inode->i_bdev;
	__bd_forget(inode);
L
Linus Torvalds 已提交
704
	spin_unlock(&bdev_lock);
705 706 707

	if (bdev)
		iput(bdev->bd_inode);
L
Linus Torvalds 已提交
708 709
}

T
Tejun Heo 已提交
710 711 712 713 714 715
/**
 * 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
 *
L
Lucas De Marchi 已提交
716
 * Test whether @bdev can be claimed by @holder.
T
Tejun Heo 已提交
717 718 719 720 721 722 723 724 725
 *
 * 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
726 727
{
	if (bdev->bd_holder == holder)
T
Tejun Heo 已提交
728
		return true;	 /* already a holder */
L
Linus Torvalds 已提交
729
	else if (bdev->bd_holder != NULL)
T
Tejun Heo 已提交
730
		return false; 	 /* held by someone else */
L
Linus Torvalds 已提交
731
	else if (bdev->bd_contains == bdev)
T
Tejun Heo 已提交
732
		return true;  	 /* is a whole device which isn't held */
L
Linus Torvalds 已提交
733

734
	else if (whole->bd_holder == bd_may_claim)
T
Tejun Heo 已提交
735 736 737
		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
738
	else
T
Tejun Heo 已提交
739 740 741
		return true;	 /* is a partition of an un-held device */
}

742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767
/**
 * 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;

768 769
	/* if claiming is already in progress, wait for it to finish */
	if (whole->bd_claiming) {
770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792
		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
793 794 795 796 797 798 799
 * 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.
800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824
 *
 * 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);

825 826 827 828 829 830 831 832 833 834 835 836 837
	/*
	 * 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);

838
	module_put(disk->fops->owner);
839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857
	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);
	}
}

858
#ifdef CONFIG_SYSFS
859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875
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;
}

876
static int add_symlink(struct kobject *from, struct kobject *to)
877
{
878
	return sysfs_create_link(from, to, kobject_name(to));
879 880 881 882 883 884 885
}

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

J
Jun'ichi Nomura 已提交
886
/**
887 888 889
 * 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
890
 *
891 892
 * DON'T USE THIS UNLESS YOU'RE ALREADY USING IT.
 *
893
 * This functions creates the following sysfs symlinks.
894
 *
895 896
 * - from "slaves" directory of the holder @disk to the claimed @bdev
 * - from "holders" directory of the @bdev to the holder @disk
897
 *
898 899
 * For example, if /dev/dm-0 maps to /dev/sda and disk for dm-0 is
 * passed to bd_link_disk_holder(), then:
900
 *
901 902
 *   /sys/block/dm-0/slaves/sda --> /sys/block/sda
 *   /sys/block/sda/holders/dm-0 --> /sys/block/dm-0
903
 *
904 905 906
 * 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.
907
 *
908 909
 * CONTEXT:
 * Might sleep.
910
 *
911 912
 * RETURNS:
 * 0 on success, -errno on failure.
913
 */
914
int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk)
915
{
916
	struct bd_holder_disk *holder;
917
	int ret = 0;
918

919
	mutex_lock(&bdev->bd_mutex);
J
Jun'ichi Nomura 已提交
920

921
	WARN_ON_ONCE(!bdev->bd_holder);
922

923 924 925
	/* FIXME: remove the following once add_disk() handles errors */
	if (WARN_ON(!disk->slave_dir || !bdev->bd_part->holder_dir))
		goto out_unlock;
926

927 928 929
	holder = bd_find_holder_disk(bdev, disk);
	if (holder) {
		holder->refcnt++;
930
		goto out_unlock;
931
	}
932

933 934 935
	holder = kzalloc(sizeof(*holder), GFP_KERNEL);
	if (!holder) {
		ret = -ENOMEM;
936 937
		goto out_unlock;
	}
938

939 940 941 942 943 944 945 946 947 948 949
	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;
950 951 952 953 954
	/*
	 * bdev could be deleted beneath us which would implicitly destroy
	 * the holder directory.  Hold on to it.
	 */
	kobject_get(bdev->bd_part->holder_dir);
955 956 957 958 959 960 961 962

	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);
963
out_unlock:
964
	mutex_unlock(&bdev->bd_mutex);
965
	return ret;
966
}
967
EXPORT_SYMBOL_GPL(bd_link_disk_holder);
968

969 970 971 972 973 974 975 976 977 978 979
/**
 * 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)
980
{
981
	struct bd_holder_disk *holder;
982

983
	mutex_lock(&bdev->bd_mutex);
984

985 986 987 988 989 990
	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);
991
		kobject_put(bdev->bd_part->holder_dir);
992 993 994 995 996
		list_del_init(&holder->list);
		kfree(holder);
	}

	mutex_unlock(&bdev->bd_mutex);
L
Linus Torvalds 已提交
997
}
998
EXPORT_SYMBOL_GPL(bd_unlink_disk_holder);
999
#endif
L
Linus Torvalds 已提交
1000

1001 1002 1003 1004
/**
 * flush_disk - invalidates all buffer-cache entries on a disk
 *
 * @bdev:      struct block device to be flushed
1005
 * @kill_dirty: flag to guide handling of dirty inodes
1006 1007 1008 1009 1010
 *
 * 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.
 */
1011
static void flush_disk(struct block_device *bdev, bool kill_dirty)
1012
{
1013
	if (__invalidate_device(bdev, kill_dirty)) {
1014 1015 1016 1017 1018 1019 1020 1021 1022 1023
		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;
T
Tejun Heo 已提交
1024
	if (disk_part_scan_enabled(bdev->bd_disk))
1025 1026 1027
		bdev->bd_invalidated = 1;
}

1028
/**
1029
 * check_disk_size_change - checks for disk size change and adjusts bdev size.
1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049
 * @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);
1050
		flush_disk(bdev, false);
1051 1052 1053 1054
	}
}
EXPORT_SYMBOL(check_disk_size_change);

1055
/**
1056
 * revalidate_disk - wrapper for lower-level driver's revalidate_disk call-back
1057 1058 1059 1060 1061 1062 1063 1064
 * @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)
{
1065
	struct block_device *bdev;
1066 1067 1068 1069 1070
	int ret = 0;

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

1071 1072 1073 1074 1075 1076
	bdev = bdget_disk(disk, 0);
	if (!bdev)
		return ret;

	mutex_lock(&bdev->bd_mutex);
	check_disk_size_change(disk, bdev);
1077
	bdev->bd_invalidated = 0;
1078 1079
	mutex_unlock(&bdev->bd_mutex);
	bdput(bdev);
1080 1081 1082 1083
	return ret;
}
EXPORT_SYMBOL(revalidate_disk);

L
Linus Torvalds 已提交
1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095
/*
 * 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;
1096
	const struct block_device_operations *bdops = disk->fops;
1097
	unsigned int events;
L
Linus Torvalds 已提交
1098

1099 1100 1101
	events = disk_clear_events(disk, DISK_EVENT_MEDIA_CHANGE |
				   DISK_EVENT_EJECT_REQUEST);
	if (!(events & DISK_EVENT_MEDIA_CHANGE))
L
Linus Torvalds 已提交
1102 1103
		return 0;

1104
	flush_disk(bdev, true);
L
Linus Torvalds 已提交
1105 1106 1107 1108 1109 1110 1111 1112 1113
	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)
{
1114
	unsigned bsize = bdev_logical_block_size(bdev);
L
Linus Torvalds 已提交
1115

1116 1117 1118
	mutex_lock(&bdev->bd_inode->i_mutex);
	i_size_write(bdev->bd_inode, size);
	mutex_unlock(&bdev->bd_inode->i_mutex);
L
Linus Torvalds 已提交
1119 1120 1121 1122 1123 1124 1125 1126 1127 1128
	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
1129
static void __blkdev_put(struct block_device *bdev, fmode_t mode, int for_part);
1130

1131 1132 1133 1134 1135 1136 1137
/*
 * bd_mutex locking:
 *
 *  mutex_lock(part->bd_mutex)
 *    mutex_lock_nested(whole->bd_mutex, 1)
 */

1138
static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part)
L
Linus Torvalds 已提交
1139 1140
{
	struct gendisk *disk;
1141
	struct module *owner;
1142
	int ret;
1143
	int partno;
1144 1145
	int perm = 0;

1146
	if (mode & FMODE_READ)
1147
		perm |= MAY_READ;
1148
	if (mode & FMODE_WRITE)
1149 1150 1151 1152
		perm |= MAY_WRITE;
	/*
	 * hooks: /n/, see "layering violations".
	 */
1153 1154 1155 1156 1157 1158
	if (!for_part) {
		ret = devcgroup_inode_permission(bdev->bd_inode, perm);
		if (ret != 0) {
			bdput(bdev);
			return ret;
		}
1159
	}
1160

1161
 restart:
T
Tejun Heo 已提交
1162

1163
	ret = -ENXIO;
1164
	disk = get_gendisk(bdev->bd_dev, &partno);
T
Tejun Heo 已提交
1165
	if (!disk)
1166
		goto out;
1167
	owner = disk->fops->owner;
L
Linus Torvalds 已提交
1168

1169
	disk_block_events(disk);
1170
	mutex_lock_nested(&bdev->bd_mutex, for_part);
L
Linus Torvalds 已提交
1171 1172
	if (!bdev->bd_openers) {
		bdev->bd_disk = disk;
1173
		bdev->bd_queue = disk->queue;
L
Linus Torvalds 已提交
1174
		bdev->bd_contains = bdev;
1175
		if (!partno) {
1176 1177 1178 1179 1180
			ret = -ENXIO;
			bdev->bd_part = disk_get_part(disk, partno);
			if (!bdev->bd_part)
				goto out_clear;

1181
			ret = 0;
L
Linus Torvalds 已提交
1182
			if (disk->fops->open) {
1183
				ret = disk->fops->open(bdev, mode);
1184 1185 1186 1187 1188 1189 1190 1191
				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;
1192
					bdev->bd_queue = NULL;
1193
					mutex_unlock(&bdev->bd_mutex);
1194 1195
					disk_unblock_events(disk);
					put_disk(disk);
1196
					module_put(owner);
1197 1198
					goto restart;
				}
L
Linus Torvalds 已提交
1199
			}
1200

1201
			if (!ret)
1202 1203
				bd_set_size(bdev,(loff_t)get_capacity(disk)<<9);

1204 1205 1206 1207 1208 1209
			/*
			 * If the device is invalidated, rescan partition
			 * if open succeeded or failed with -ENOMEDIUM.
			 * The latter is necessary to prevent ghost
			 * partitions on a removed medium.
			 */
1210 1211 1212 1213 1214 1215
			if (bdev->bd_invalidated) {
				if (!ret)
					rescan_partitions(disk, bdev);
				else if (ret == -ENOMEDIUM)
					invalidate_partitions(disk, bdev);
			}
1216 1217
			if (ret)
				goto out_clear;
L
Linus Torvalds 已提交
1218 1219 1220 1221 1222
		} else {
			struct block_device *whole;
			whole = bdget_disk(disk, 0);
			ret = -ENOMEM;
			if (!whole)
T
Tejun Heo 已提交
1223
				goto out_clear;
1224
			BUG_ON(for_part);
1225
			ret = __blkdev_get(whole, mode, 1);
L
Linus Torvalds 已提交
1226
			if (ret)
T
Tejun Heo 已提交
1227
				goto out_clear;
L
Linus Torvalds 已提交
1228
			bdev->bd_contains = whole;
1229
			bdev->bd_part = disk_get_part(disk, partno);
1230
			if (!(disk->flags & GENHD_FL_UP) ||
1231
			    !bdev->bd_part || !bdev->bd_part->nr_sects) {
L
Linus Torvalds 已提交
1232
				ret = -ENXIO;
T
Tejun Heo 已提交
1233
				goto out_clear;
L
Linus Torvalds 已提交
1234
			}
1235
			bd_set_size(bdev, (loff_t)bdev->bd_part->nr_sects << 9);
L
Linus Torvalds 已提交
1236 1237 1238
		}
	} else {
		if (bdev->bd_contains == bdev) {
1239 1240
			ret = 0;
			if (bdev->bd_disk->fops->open)
1241
				ret = bdev->bd_disk->fops->open(bdev, mode);
1242
			/* the same as first opener case, read comment there */
1243 1244 1245 1246 1247 1248
			if (bdev->bd_invalidated) {
				if (!ret)
					rescan_partitions(bdev->bd_disk, bdev);
				else if (ret == -ENOMEDIUM)
					invalidate_partitions(bdev->bd_disk, bdev);
			}
1249 1250
			if (ret)
				goto out_unlock_bdev;
L
Linus Torvalds 已提交
1251
		}
1252 1253
		/* only one opener holds refs to the module and disk */
		put_disk(disk);
1254
		module_put(owner);
L
Linus Torvalds 已提交
1255 1256
	}
	bdev->bd_openers++;
1257 1258
	if (for_part)
		bdev->bd_part_count++;
1259
	mutex_unlock(&bdev->bd_mutex);
1260
	disk_unblock_events(disk);
L
Linus Torvalds 已提交
1261 1262
	return 0;

T
Tejun Heo 已提交
1263
 out_clear:
1264
	disk_put_part(bdev->bd_part);
L
Linus Torvalds 已提交
1265
	bdev->bd_disk = NULL;
T
Tejun Heo 已提交
1266
	bdev->bd_part = NULL;
1267
	bdev->bd_queue = NULL;
L
Linus Torvalds 已提交
1268
	if (bdev != bdev->bd_contains)
1269
		__blkdev_put(bdev->bd_contains, mode, 1);
L
Linus Torvalds 已提交
1270
	bdev->bd_contains = NULL;
T
Tejun Heo 已提交
1271
 out_unlock_bdev:
1272
	mutex_unlock(&bdev->bd_mutex);
1273
	disk_unblock_events(disk);
T
Tejun Heo 已提交
1274
	put_disk(disk);
1275
	module_put(owner);
1276
 out:
T
Tejun Heo 已提交
1277 1278
	bdput(bdev);

L
Linus Torvalds 已提交
1279 1280 1281
	return ret;
}

1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300
/**
 * blkdev_get - open a block device
 * @bdev: block_device to open
 * @mode: FMODE_* mask
 * @holder: exclusive holder identifier
 *
 * Open @bdev with @mode.  If @mode includes %FMODE_EXCL, @bdev is
 * open with exclusive access.  Specifying %FMODE_EXCL with %NULL
 * @holder is invalid.  Exclusive opens may nest for the same @holder.
 *
 * On success, the reference count of @bdev is unchanged.  On failure,
 * @bdev is put.
 *
 * CONTEXT:
 * Might sleep.
 *
 * RETURNS:
 * 0 on success, -errno on failure.
 */
1301
int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder)
L
Linus Torvalds 已提交
1302
{
1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318
	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);

	if (whole) {
1319 1320
		struct gendisk *disk = whole->bd_disk;

1321
		/* finish claiming */
1322
		mutex_lock(&bdev->bd_mutex);
1323 1324
		spin_lock(&bdev_lock);

1325
		if (!res) {
1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344
			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);
1345 1346

		/*
1347 1348 1349 1350 1351
		 * Block event polling for write claims if requested.  Any
		 * write holder makes the write_holder state stick until
		 * all are released.  This is good enough and tracking
		 * individual writeable reference is too fragile given the
		 * way @mode is used in blkdev_get/put().
1352
		 */
1353 1354
		if (!res && (mode & FMODE_WRITE) && !bdev->bd_write_holder &&
		    (disk->flags & GENHD_FL_BLOCK_EVENTS_ON_EXCL_WRITE)) {
1355
			bdev->bd_write_holder = true;
1356
			disk_block_events(disk);
1357 1358 1359
		}

		mutex_unlock(&bdev->bd_mutex);
1360
		bdput(whole);
1361 1362 1363
	}

	return res;
1364
}
L
Linus Torvalds 已提交
1365 1366
EXPORT_SYMBOL(blkdev_get);

1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397
/**
 * blkdev_get_by_path - open a block device by name
 * @path: path to the block device to open
 * @mode: FMODE_* mask
 * @holder: exclusive holder identifier
 *
 * Open the blockdevice described by the device file at @path.  @mode
 * and @holder are identical to blkdev_get().
 *
 * On success, the returned block_device has reference count of one.
 *
 * CONTEXT:
 * Might sleep.
 *
 * RETURNS:
 * Pointer to block_device on success, ERR_PTR(-errno) on failure.
 */
struct block_device *blkdev_get_by_path(const char *path, fmode_t mode,
					void *holder)
{
	struct block_device *bdev;
	int err;

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

	err = blkdev_get(bdev, mode, holder);
	if (err)
		return ERR_PTR(err);

1398 1399 1400 1401 1402
	if ((mode & FMODE_WRITE) && bdev_read_only(bdev)) {
		blkdev_put(bdev, mode);
		return ERR_PTR(-EACCES);
	}

1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445
	return bdev;
}
EXPORT_SYMBOL(blkdev_get_by_path);

/**
 * blkdev_get_by_dev - open a block device by device number
 * @dev: device number of block device to open
 * @mode: FMODE_* mask
 * @holder: exclusive holder identifier
 *
 * Open the blockdevice described by device number @dev.  @mode and
 * @holder are identical to blkdev_get().
 *
 * 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.
 *
 * On success, the returned block_device has reference count of one.
 *
 * CONTEXT:
 * Might sleep.
 *
 * RETURNS:
 * Pointer to block_device on success, ERR_PTR(-errno) on failure.
 */
struct block_device *blkdev_get_by_dev(dev_t dev, fmode_t mode, void *holder)
{
	struct block_device *bdev;
	int err;

	bdev = bdget(dev);
	if (!bdev)
		return ERR_PTR(-ENOMEM);

	err = blkdev_get(bdev, mode, holder);
	if (err)
		return ERR_PTR(err);

	return bdev;
}
EXPORT_SYMBOL(blkdev_get_by_dev);

L
Linus Torvalds 已提交
1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457
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;

1458 1459 1460 1461 1462 1463 1464
	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
1465
	bdev = bd_acquire(inode);
1466 1467
	if (bdev == NULL)
		return -ENOMEM;
L
Linus Torvalds 已提交
1468

1469 1470
	filp->f_mapping = bdev->bd_inode->i_mapping;

1471
	return blkdev_get(bdev, filp->f_mode, filp);
L
Linus Torvalds 已提交
1472 1473
}

A
Al Viro 已提交
1474
static void __blkdev_put(struct block_device *bdev, fmode_t mode, int for_part)
1475 1476
{
	struct gendisk *disk = bdev->bd_disk;
1477
	struct block_device *victim = NULL;
1478

1479
	mutex_lock_nested(&bdev->bd_mutex, for_part);
1480 1481 1482
	if (for_part)
		bdev->bd_part_count--;

1483
	if (!--bdev->bd_openers) {
1484
		WARN_ON_ONCE(bdev->bd_holders);
1485 1486
		sync_blockdev(bdev);
		kill_bdev(bdev);
1487 1488 1489
		/*
		 * ->release can cause the queue to disappear, so flush all
		 * dirty data before.
1490
		 */
1491
		bdev_write_inode(bdev->bd_inode);
1492 1493 1494
	}
	if (bdev->bd_contains == bdev) {
		if (disk->fops->release)
1495
			disk->fops->release(disk, mode);
1496 1497 1498 1499
	}
	if (!bdev->bd_openers) {
		struct module *owner = disk->fops->owner;

T
Tejun Heo 已提交
1500 1501
		disk_put_part(bdev->bd_part);
		bdev->bd_part = NULL;
1502
		bdev->bd_disk = NULL;
1503 1504
		if (bdev != bdev->bd_contains)
			victim = bdev->bd_contains;
1505
		bdev->bd_contains = NULL;
1506 1507 1508

		put_disk(disk);
		module_put(owner);
1509 1510 1511
	}
	mutex_unlock(&bdev->bd_mutex);
	bdput(bdev);
1512
	if (victim)
A
Al Viro 已提交
1513
		__blkdev_put(victim, mode, 1);
1514 1515
}

A
Al Viro 已提交
1516
void blkdev_put(struct block_device *bdev, fmode_t mode)
1517
{
1518 1519
	mutex_lock(&bdev->bd_mutex);

1520
	if (mode & FMODE_EXCL) {
1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540
		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.
		 */
		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);

1541 1542 1543 1544
		/*
		 * If this was the last claim, remove holder link and
		 * unblock evpoll if it was a write holder.
		 */
1545 1546 1547
		if (bdev_free && bdev->bd_write_holder) {
			disk_unblock_events(bdev->bd_disk);
			bdev->bd_write_holder = false;
1548
		}
1549
	}
1550

1551 1552 1553 1554 1555 1556 1557 1558 1559
	/*
	 * Trigger event checking and tell drivers to flush MEDIA_CHANGE
	 * event.  This is to ensure detection of media removal commanded
	 * from userland - e.g. eject(1).
	 */
	disk_flush_events(bdev->bd_disk, DISK_EVENT_MEDIA_CHANGE);

	mutex_unlock(&bdev->bd_mutex);

A
Al Viro 已提交
1560
	__blkdev_put(bdev, mode, 0);
1561
}
1562 1563
EXPORT_SYMBOL(blkdev_put);

L
Linus Torvalds 已提交
1564 1565 1566
static int blkdev_close(struct inode * inode, struct file * filp)
{
	struct block_device *bdev = I_BDEV(filp->f_mapping->host);
A
Al Viro 已提交
1567 1568
	blkdev_put(bdev, filp->f_mode);
	return 0;
L
Linus Torvalds 已提交
1569 1570
}

1571
static long block_ioctl(struct file *file, unsigned cmd, unsigned long arg)
L
Linus Torvalds 已提交
1572
{
1573 1574
	struct block_device *bdev = I_BDEV(file->f_mapping->host);
	fmode_t mode = file->f_mode;
1575 1576 1577 1578 1579

	/*
	 * O_NDELAY can be altered using fcntl(.., F_SETFL, ..), so we have
	 * to updated it before every ioctl.
	 */
1580
	if (file->f_flags & O_NDELAY)
1581 1582 1583 1584
		mode |= FMODE_NDELAY;
	else
		mode &= ~FMODE_NDELAY;

1585
	return blkdev_ioctl(bdev, mode, cmd, arg);
L
Linus Torvalds 已提交
1586 1587
}

1588 1589 1590 1591 1592 1593 1594
/*
 * 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.
 */
1595
ssize_t blkdev_write_iter(struct kiocb *iocb, struct iov_iter *from)
1596 1597
{
	struct file *file = iocb->ki_filp;
1598 1599
	struct inode *bd_inode = file->f_mapping->host;
	loff_t size = i_size_read(bd_inode);
1600
	struct blk_plug plug;
1601
	ssize_t ret;
1602

1603 1604
	if (bdev_read_only(I_BDEV(bd_inode)))
		return -EPERM;
1605

1606
	if (!iov_iter_count(from))
1607 1608
		return 0;

1609 1610 1611 1612
	if (iocb->ki_pos >= size)
		return -ENOSPC;

	iov_iter_truncate(from, size - iocb->ki_pos);
1613

1614
	blk_start_plug(&plug);
1615
	ret = __generic_file_write_iter(iocb, from);
1616
	if (ret > 0) {
1617
		ssize_t err;
1618
		err = generic_write_sync(file, iocb->ki_pos - ret, ret);
1619
		if (err < 0)
1620 1621
			ret = err;
	}
1622
	blk_finish_plug(&plug);
1623 1624
	return ret;
}
1625
EXPORT_SYMBOL_GPL(blkdev_write_iter);
1626

1627
ssize_t blkdev_read_iter(struct kiocb *iocb, struct iov_iter *to)
1628 1629 1630 1631
{
	struct file *file = iocb->ki_filp;
	struct inode *bd_inode = file->f_mapping->host;
	loff_t size = i_size_read(bd_inode);
1632
	loff_t pos = iocb->ki_pos;
1633 1634 1635 1636 1637

	if (pos >= size)
		return 0;

	size -= pos;
1638 1639
	iov_iter_truncate(to, size);
	return generic_file_read_iter(iocb, to);
1640
}
1641
EXPORT_SYMBOL_GPL(blkdev_read_iter);
1642

1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656
/*
 * 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
1657
static const struct address_space_operations def_blk_aops = {
L
Linus Torvalds 已提交
1658
	.readpage	= blkdev_readpage,
1659
	.readpages	= blkdev_readpages,
L
Linus Torvalds 已提交
1660
	.writepage	= blkdev_writepage,
N
Nick Piggin 已提交
1661 1662
	.write_begin	= blkdev_write_begin,
	.write_end	= blkdev_write_end,
L
Linus Torvalds 已提交
1663
	.writepages	= generic_writepages,
1664
	.releasepage	= blkdev_releasepage,
L
Linus Torvalds 已提交
1665
	.direct_IO	= blkdev_direct_IO,
1666
	.is_dirty_writeback = buffer_check_dirty_writeback,
L
Linus Torvalds 已提交
1667 1668
};

1669
const struct file_operations def_blk_fops = {
L
Linus Torvalds 已提交
1670 1671 1672
	.open		= blkdev_open,
	.release	= blkdev_close,
	.llseek		= block_llseek,
1673
	.read_iter	= blkdev_read_iter,
1674
	.write_iter	= blkdev_write_iter,
1675
	.mmap		= generic_file_mmap,
1676
	.fsync		= blkdev_fsync,
1677
	.unlocked_ioctl	= block_ioctl,
L
Linus Torvalds 已提交
1678 1679 1680
#ifdef CONFIG_COMPAT
	.compat_ioctl	= compat_blkdev_ioctl,
#endif
1681
	.splice_read	= generic_file_splice_read,
A
Al Viro 已提交
1682
	.splice_write	= iter_file_splice_write,
L
Linus Torvalds 已提交
1683 1684 1685 1686 1687 1688 1689
};

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);
1690
	res = blkdev_ioctl(bdev, 0, cmd, arg);
L
Linus Torvalds 已提交
1691 1692 1693 1694 1695 1696 1697 1698
	set_fs(old_fs);
	return res;
}

EXPORT_SYMBOL(ioctl_by_bdev);

/**
 * lookup_bdev  - lookup a struct block_device by name
1699
 * @pathname:	special file representing the block device
L
Linus Torvalds 已提交
1700
 *
1701
 * Get a reference to the blockdevice at @pathname in the current
L
Linus Torvalds 已提交
1702 1703 1704
 * namespace if possible and return it.  Return ERR_PTR(error)
 * otherwise.
 */
1705
struct block_device *lookup_bdev(const char *pathname)
L
Linus Torvalds 已提交
1706 1707 1708
{
	struct block_device *bdev;
	struct inode *inode;
1709
	struct path path;
L
Linus Torvalds 已提交
1710 1711
	int error;

1712
	if (!pathname || !*pathname)
L
Linus Torvalds 已提交
1713 1714
		return ERR_PTR(-EINVAL);

1715
	error = kern_path(pathname, LOOKUP_FOLLOW, &path);
L
Linus Torvalds 已提交
1716 1717 1718
	if (error)
		return ERR_PTR(error);

1719
	inode = path.dentry->d_inode;
L
Linus Torvalds 已提交
1720 1721 1722 1723
	error = -ENOTBLK;
	if (!S_ISBLK(inode->i_mode))
		goto fail;
	error = -EACCES;
1724
	if (path.mnt->mnt_flags & MNT_NODEV)
L
Linus Torvalds 已提交
1725 1726 1727 1728 1729 1730
		goto fail;
	error = -ENOMEM;
	bdev = bd_acquire(inode);
	if (!bdev)
		goto fail;
out:
1731
	path_put(&path);
L
Linus Torvalds 已提交
1732 1733 1734 1735 1736
	return bdev;
fail:
	bdev = ERR_PTR(error);
	goto out;
}
1737
EXPORT_SYMBOL(lookup_bdev);
L
Linus Torvalds 已提交
1738

1739
int __invalidate_device(struct block_device *bdev, bool kill_dirty)
1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751
{
	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);
1752
		res = invalidate_inodes(sb, kill_dirty);
1753 1754
		drop_super(sb);
	}
1755
	invalidate_bdev(bdev);
1756 1757 1758
	return res;
}
EXPORT_SYMBOL(__invalidate_device);
1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794

void iterate_bdevs(void (*func)(struct block_device *, void *), void *arg)
{
	struct inode *inode, *old_inode = NULL;

	spin_lock(&inode_sb_list_lock);
	list_for_each_entry(inode, &blockdev_superblock->s_inodes, i_sb_list) {
		struct address_space *mapping = inode->i_mapping;

		spin_lock(&inode->i_lock);
		if (inode->i_state & (I_FREEING|I_WILL_FREE|I_NEW) ||
		    mapping->nrpages == 0) {
			spin_unlock(&inode->i_lock);
			continue;
		}
		__iget(inode);
		spin_unlock(&inode->i_lock);
		spin_unlock(&inode_sb_list_lock);
		/*
		 * We hold a reference to 'inode' so it couldn't have been
		 * removed from s_inodes list while we dropped the
		 * inode_sb_list_lock.  We cannot iput the inode now as we can
		 * be holding the last reference and we cannot iput it under
		 * inode_sb_list_lock. So we keep the reference and iput it
		 * later.
		 */
		iput(old_inode);
		old_inode = inode;

		func(I_BDEV(inode), arg);

		spin_lock(&inode_sb_list_lock);
	}
	spin_unlock(&inode_sb_list_lock);
	iput(old_inode);
}