file.c 16.4 KB
Newer Older
1
/*
2
 *  linux/fs/ext4/file.c
3 4 5 6 7 8 9 10 11 12 13 14
 *
 * Copyright (C) 1992, 1993, 1994, 1995
 * Remy Card (card@masi.ibp.fr)
 * Laboratoire MASI - Institut Blaise Pascal
 * Universite Pierre et Marie Curie (Paris VI)
 *
 *  from
 *
 *  linux/fs/minix/file.c
 *
 *  Copyright (C) 1991, 1992  Linus Torvalds
 *
15
 *  ext4 fs regular file handling primitives
16 17 18 19 20 21 22
 *
 *  64-bit file support on 64-bit platforms by Jakub Jelinek
 *	(jj@sunsite.ms.mff.cuni.cz)
 */

#include <linux/time.h>
#include <linux/fs.h>
23
#include <linux/jbd2.h>
24 25
#include <linux/mount.h>
#include <linux/path.h>
26
#include <linux/aio.h>
27
#include <linux/quotaops.h>
28
#include <linux/pagevec.h>
29 30
#include "ext4.h"
#include "ext4_jbd2.h"
31 32 33 34 35
#include "xattr.h"
#include "acl.h"

/*
 * Called when an inode is released. Note that this is different
36
 * from ext4_file_open: open gets called at every open, but release
37 38
 * gets called only when /all/ the files are closed.
 */
39
static int ext4_release_file(struct inode *inode, struct file *filp)
40
{
41
	if (ext4_test_inode_state(inode, EXT4_STATE_DA_ALLOC_CLOSE)) {
42
		ext4_alloc_da_blocks(inode);
43
		ext4_clear_inode_state(inode, EXT4_STATE_DA_ALLOC_CLOSE);
44
	}
45 46
	/* if we are the last writer on the inode, drop the block reservation */
	if ((filp->f_mode & FMODE_WRITE) &&
47 48
			(atomic_read(&inode->i_writecount) == 1) &&
		        !EXT4_I(inode)->i_reserved_data_blocks)
49
	{
50
		down_write(&EXT4_I(inode)->i_data_sem);
51
		ext4_discard_preallocations(inode);
52
		up_write(&EXT4_I(inode)->i_data_sem);
53 54
	}
	if (is_dx(inode) && filp->private_data)
55
		ext4_htree_free_dir_info(filp->private_data);
56 57 58 59

	return 0;
}

60
static void ext4_unwritten_wait(struct inode *inode)
61 62 63
{
	wait_queue_head_t *wq = ext4_ioend_wq(inode);

64
	wait_event(*wq, (atomic_read(&EXT4_I(inode)->i_unwritten) == 0));
65 66 67 68 69 70 71 72 73 74 75 76
}

/*
 * This tests whether the IO in question is block-aligned or not.
 * Ext4 utilizes unwritten extents when hole-filling during direct IO, and they
 * are converted to written only after the IO is complete.  Until they are
 * mapped, these blocks appear as holes, so dio_zero_block() will assume that
 * it needs to zero out portions of the start and/or end block.  If 2 AIO
 * threads are at work on the same unwritten block, they must be synchronized
 * or one thread will zero the other's data, causing corruption.
 */
static int
A
Al Viro 已提交
77
ext4_unaligned_aio(struct inode *inode, struct iov_iter *from, loff_t pos)
78 79 80 81
{
	struct super_block *sb = inode->i_sb;
	int blockmask = sb->s_blocksize - 1;

82
	if (pos >= i_size_read(inode))
83 84
		return 0;

A
Al Viro 已提交
85
	if ((pos | iov_iter_alignment(from)) & blockmask)
86 87 88 89 90
		return 1;

	return 0;
}

91
static ssize_t
A
Al Viro 已提交
92
ext4_file_write_iter(struct kiocb *iocb, struct iov_iter *from)
93
{
94
	struct file *file = iocb->ki_filp;
95
	struct inode *inode = file_inode(iocb->ki_filp);
96
	struct mutex *aio_mutex = NULL;
97
	struct blk_plug plug;
R
Ross Zwisler 已提交
98
	int o_direct = io_is_direct(file);
99
	int overwrite = 0;
A
Al Viro 已提交
100
	size_t length = iov_iter_count(from);
101
	ssize_t ret;
A
Al Viro 已提交
102
	loff_t pos = iocb->ki_pos;
103

104 105 106 107 108 109 110 111
	/*
	 * Unaligned direct AIO must be serialized; see comment above
	 * In the case of O_APPEND, assume that we must always serialize
	 */
	if (o_direct &&
	    ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS) &&
	    !is_sync_kiocb(iocb) &&
	    (file->f_flags & O_APPEND ||
A
Al Viro 已提交
112
	     ext4_unaligned_aio(inode, from, pos))) {
113 114 115 116 117 118 119 120 121
		aio_mutex = ext4_aio_mutex(inode);
		mutex_lock(aio_mutex);
		ext4_unwritten_wait(inode);
	}

	mutex_lock(&inode->i_mutex);
	if (file->f_flags & O_APPEND)
		iocb->ki_pos = pos = i_size_read(inode);

122 123 124 125
	/*
	 * If we have encountered a bitmap-format file, the size limit
	 * is smaller than s_maxbytes, which is for extent-mapped files.
	 */
126
	if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))) {
127
		struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
128

129 130 131 132 133 134
		if ((pos > sbi->s_bitmap_maxbytes) ||
		    (pos == sbi->s_bitmap_maxbytes && length > 0)) {
			mutex_unlock(&inode->i_mutex);
			ret = -EFBIG;
			goto errout;
		}
135

A
Al Viro 已提交
136 137
		if (pos + length > sbi->s_bitmap_maxbytes)
			iov_iter_truncate(from, sbi->s_bitmap_maxbytes - pos);
138 139
	}

140
	iocb->private = &overwrite;
141
	if (o_direct) {
142 143 144 145
		blk_start_plug(&plug);


		/* check whether we do a DIO overwrite or not */
146
		if (ext4_should_dioread_nolock(inode) && !aio_mutex &&
147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172
		    !file->f_mapping->nrpages && pos + length <= i_size_read(inode)) {
			struct ext4_map_blocks map;
			unsigned int blkbits = inode->i_blkbits;
			int err, len;

			map.m_lblk = pos >> blkbits;
			map.m_len = (EXT4_BLOCK_ALIGN(pos + length, blkbits) >> blkbits)
				- map.m_lblk;
			len = map.m_len;

			err = ext4_map_blocks(NULL, inode, &map, 0);
			/*
			 * 'err==len' means that all of blocks has
			 * been preallocated no matter they are
			 * initialized or not.  For excluding
			 * unwritten extents, we need to check
			 * m_flags.  There are two conditions that
			 * indicate for initialized extents.  1) If we
			 * hit extent cache, EXT4_MAP_MAPPED flag is
			 * returned; 2) If we do a real lookup,
			 * non-flags are returned.  So we should check
			 * these two conditions.
			 */
			if (err == len && (map.m_flags & EXT4_MAP_MAPPED))
				overwrite = 1;
		}
173
	}
174

A
Al Viro 已提交
175
	ret = __generic_file_write_iter(iocb, from);
176
	mutex_unlock(&inode->i_mutex);
177

178 179 180 181 182 183
	if (ret > 0) {
		ssize_t err;

		err = generic_write_sync(file, iocb->ki_pos - ret, ret);
		if (err < 0)
			ret = err;
184
	}
185 186
	if (o_direct)
		blk_finish_plug(&plug);
187

188
errout:
189 190
	if (aio_mutex)
		mutex_unlock(aio_mutex);
191
	return ret;
192 193
}

R
Ross Zwisler 已提交
194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213
#ifdef CONFIG_FS_DAX
static int ext4_dax_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
{
	return dax_fault(vma, vmf, ext4_get_block);
					/* Is this the right get_block? */
}

static int ext4_dax_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
{
	return dax_mkwrite(vma, vmf, ext4_get_block);
}

static const struct vm_operations_struct ext4_dax_vm_ops = {
	.fault		= ext4_dax_fault,
	.page_mkwrite	= ext4_dax_mkwrite,
};
#else
#define ext4_dax_vm_ops	ext4_file_vm_ops
#endif

214
static const struct vm_operations_struct ext4_file_vm_ops = {
215
	.fault		= filemap_fault,
216
	.map_pages	= filemap_map_pages,
217 218 219 220 221 222
	.page_mkwrite   = ext4_page_mkwrite,
};

static int ext4_file_mmap(struct file *file, struct vm_area_struct *vma)
{
	file_accessed(file);
R
Ross Zwisler 已提交
223 224 225 226 227 228
	if (IS_DAX(file_inode(file))) {
		vma->vm_ops = &ext4_dax_vm_ops;
		vma->vm_flags |= VM_MIXEDMAP;
	} else {
		vma->vm_ops = &ext4_file_vm_ops;
	}
229 230 231
	return 0;
}

232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249
static int ext4_file_open(struct inode * inode, struct file * filp)
{
	struct super_block *sb = inode->i_sb;
	struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
	struct vfsmount *mnt = filp->f_path.mnt;
	struct path path;
	char buf[64], *cp;

	if (unlikely(!(sbi->s_mount_flags & EXT4_MF_MNTDIR_SAMPLED) &&
		     !(sb->s_flags & MS_RDONLY))) {
		sbi->s_mount_flags |= EXT4_MF_MNTDIR_SAMPLED;
		/*
		 * Sample where the filesystem has been mounted and
		 * store it in the superblock for sysadmin convenience
		 * when trying to sort through large numbers of block
		 * devices or filesystem images.
		 */
		memset(buf, 0, sizeof(buf));
250 251
		path.mnt = mnt;
		path.dentry = mnt->mnt_root;
252 253
		cp = d_path(&path, buf, sizeof(buf));
		if (!IS_ERR(cp)) {
254 255 256
			handle_t *handle;
			int err;

257
			handle = ext4_journal_start_sb(sb, EXT4_HT_MISC, 1);
258 259
			if (IS_ERR(handle))
				return PTR_ERR(handle);
260
			BUFFER_TRACE(sbi->s_sbh, "get_write_access");
261 262 263 264 265
			err = ext4_journal_get_write_access(handle, sbi->s_sbh);
			if (err) {
				ext4_journal_stop(handle);
				return err;
			}
266 267
			strlcpy(sbi->s_es->s_last_mounted, cp,
				sizeof(sbi->s_es->s_last_mounted));
268 269
			ext4_handle_dirty_super(handle, sb);
			ext4_journal_stop(handle);
270 271
		}
	}
272 273 274 275
	/*
	 * Set up the jbd2_inode if we are opening the inode for
	 * writing and the journal is present
	 */
276 277 278 279
	if (filp->f_mode & FMODE_WRITE) {
		int ret = ext4_inode_attach_jinode(inode);
		if (ret < 0)
			return ret;
280
	}
281
	return dquot_file_open(inode, filp);
282 283
}

284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299
/*
 * Here we use ext4_map_blocks() to get a block mapping for a extent-based
 * file rather than ext4_ext_walk_space() because we can introduce
 * SEEK_DATA/SEEK_HOLE for block-mapped and extent-mapped file at the same
 * function.  When extent status tree has been fully implemented, it will
 * track all extent status for a file and we can directly use it to
 * retrieve the offset for SEEK_DATA/SEEK_HOLE.
 */

/*
 * When we retrieve the offset for SEEK_DATA/SEEK_HOLE, we would need to
 * lookup page cache to check whether or not there has some data between
 * [startoff, endoff] because, if this range contains an unwritten extent,
 * we determine this extent as a data or a hole according to whether the
 * page cache has data or not.
 */
300 301 302 303
static int ext4_find_unwritten_pgoff(struct inode *inode,
				     int whence,
				     struct ext4_map_blocks *map,
				     loff_t *offset)
304 305
{
	struct pagevec pvec;
306
	unsigned int blkbits;
307 308
	pgoff_t index;
	pgoff_t end;
309
	loff_t endoff;
310 311 312 313
	loff_t startoff;
	loff_t lastoff;
	int found = 0;

314
	blkbits = inode->i_sb->s_blocksize_bits;
315 316
	startoff = *offset;
	lastoff = startoff;
317
	endoff = (loff_t)(map->m_lblk + map->m_len) << blkbits;
318 319 320 321 322 323 324 325 326 327 328 329 330

	index = startoff >> PAGE_CACHE_SHIFT;
	end = endoff >> PAGE_CACHE_SHIFT;

	pagevec_init(&pvec, 0);
	do {
		int i, num;
		unsigned long nr_pages;

		num = min_t(pgoff_t, end - index, PAGEVEC_SIZE);
		nr_pages = pagevec_lookup(&pvec, inode->i_mapping, index,
					  (pgoff_t)num);
		if (nr_pages == 0) {
331
			if (whence == SEEK_DATA)
332 333
				break;

334
			BUG_ON(whence != SEEK_HOLE);
335 336 337 338 339 340 341 342 343 344 345 346 347 348 349
			/*
			 * If this is the first time to go into the loop and
			 * offset is not beyond the end offset, it will be a
			 * hole at this offset
			 */
			if (lastoff == startoff || lastoff < endoff)
				found = 1;
			break;
		}

		/*
		 * If this is the first time to go into the loop and
		 * offset is smaller than the first page offset, it will be a
		 * hole at this offset.
		 */
350
		if (lastoff == startoff && whence == SEEK_HOLE &&
351 352 353 354 355 356 357 358 359 360 361 362 363
		    lastoff < page_offset(pvec.pages[0])) {
			found = 1;
			break;
		}

		for (i = 0; i < nr_pages; i++) {
			struct page *page = pvec.pages[i];
			struct buffer_head *bh, *head;

			/*
			 * If the current offset is not beyond the end of given
			 * range, it will be a hole.
			 */
364
			if (lastoff < endoff && whence == SEEK_HOLE &&
365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388
			    page->index > end) {
				found = 1;
				*offset = lastoff;
				goto out;
			}

			lock_page(page);

			if (unlikely(page->mapping != inode->i_mapping)) {
				unlock_page(page);
				continue;
			}

			if (!page_has_buffers(page)) {
				unlock_page(page);
				continue;
			}

			if (page_has_buffers(page)) {
				lastoff = page_offset(page);
				bh = head = page_buffers(page);
				do {
					if (buffer_uptodate(bh) ||
					    buffer_unwritten(bh)) {
389
						if (whence == SEEK_DATA)
390 391
							found = 1;
					} else {
392
						if (whence == SEEK_HOLE)
393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413
							found = 1;
					}
					if (found) {
						*offset = max_t(loff_t,
							startoff, lastoff);
						unlock_page(page);
						goto out;
					}
					lastoff += bh->b_size;
					bh = bh->b_this_page;
				} while (bh != head);
			}

			lastoff = page_offset(page) + PAGE_SIZE;
			unlock_page(page);
		}

		/*
		 * The no. of pages is less than our desired, that would be a
		 * hole in there.
		 */
414
		if (nr_pages < num && whence == SEEK_HOLE) {
415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434
			found = 1;
			*offset = lastoff;
			break;
		}

		index = pvec.pages[i - 1]->index + 1;
		pagevec_release(&pvec);
	} while (index <= end);

out:
	pagevec_release(&pvec);
	return found;
}

/*
 * ext4_seek_data() retrieves the offset for SEEK_DATA.
 */
static loff_t ext4_seek_data(struct file *file, loff_t offset, loff_t maxsize)
{
	struct inode *inode = file->f_mapping->host;
435 436 437 438 439 440
	struct ext4_map_blocks map;
	struct extent_status es;
	ext4_lblk_t start, last, end;
	loff_t dataoff, isize;
	int blkbits;
	int ret = 0;
441 442

	mutex_lock(&inode->i_mutex);
443 444 445

	isize = i_size_read(inode);
	if (offset >= isize) {
446 447 448
		mutex_unlock(&inode->i_mutex);
		return -ENXIO;
	}
449 450 451 452 453 454 455 456 457 458 459 460 461 462

	blkbits = inode->i_sb->s_blocksize_bits;
	start = offset >> blkbits;
	last = start;
	end = isize >> blkbits;
	dataoff = offset;

	do {
		map.m_lblk = last;
		map.m_len = end - last + 1;
		ret = ext4_map_blocks(NULL, inode, &map, 0);
		if (ret > 0 && !(map.m_flags & EXT4_MAP_UNWRITTEN)) {
			if (last != start)
				dataoff = (loff_t)last << blkbits;
463
			break;
464
		}
465

466 467 468 469 470 471 472 473
		/*
		 * If there is a delay extent at this offset,
		 * it will be as a data.
		 */
		ext4_es_find_delayed_extent_range(inode, last, last, &es);
		if (es.es_len != 0 && in_range(last, es.es_lblk, es.es_len)) {
			if (last != start)
				dataoff = (loff_t)last << blkbits;
474 475 476
			break;
		}

477 478 479 480 481 482 483 484 485 486 487 488
		/*
		 * If there is a unwritten extent at this offset,
		 * it will be as a data or a hole according to page
		 * cache that has data or not.
		 */
		if (map.m_flags & EXT4_MAP_UNWRITTEN) {
			int unwritten;
			unwritten = ext4_find_unwritten_pgoff(inode, SEEK_DATA,
							      &map, &dataoff);
			if (unwritten)
				break;
		}
489

490 491 492
		last++;
		dataoff = (loff_t)last << blkbits;
	} while (last <= end);
493 494 495

	mutex_unlock(&inode->i_mutex);

496 497 498 499
	if (dataoff > isize)
		return -ENXIO;

	return vfs_setpos(file, dataoff, maxsize);
500 501 502
}

/*
503
 * ext4_seek_hole() retrieves the offset for SEEK_HOLE.
504 505 506 507
 */
static loff_t ext4_seek_hole(struct file *file, loff_t offset, loff_t maxsize)
{
	struct inode *inode = file->f_mapping->host;
508 509 510 511 512 513
	struct ext4_map_blocks map;
	struct extent_status es;
	ext4_lblk_t start, last, end;
	loff_t holeoff, isize;
	int blkbits;
	int ret = 0;
514 515

	mutex_lock(&inode->i_mutex);
516 517 518

	isize = i_size_read(inode);
	if (offset >= isize) {
519 520 521 522
		mutex_unlock(&inode->i_mutex);
		return -ENXIO;
	}

523 524 525 526 527
	blkbits = inode->i_sb->s_blocksize_bits;
	start = offset >> blkbits;
	last = start;
	end = isize >> blkbits;
	holeoff = offset;
528

529 530 531 532 533 534 535 536 537
	do {
		map.m_lblk = last;
		map.m_len = end - last + 1;
		ret = ext4_map_blocks(NULL, inode, &map, 0);
		if (ret > 0 && !(map.m_flags & EXT4_MAP_UNWRITTEN)) {
			last += ret;
			holeoff = (loff_t)last << blkbits;
			continue;
		}
538

539 540 541 542 543 544 545 546 547 548
		/*
		 * If there is a delay extent at this offset,
		 * we will skip this extent.
		 */
		ext4_es_find_delayed_extent_range(inode, last, last, &es);
		if (es.es_len != 0 && in_range(last, es.es_lblk, es.es_len)) {
			last = es.es_lblk + es.es_len;
			holeoff = (loff_t)last << blkbits;
			continue;
		}
549

550 551 552 553 554 555 556 557 558 559 560 561
		/*
		 * If there is a unwritten extent at this offset,
		 * it will be as a data or a hole according to page
		 * cache that has data or not.
		 */
		if (map.m_flags & EXT4_MAP_UNWRITTEN) {
			int unwritten;
			unwritten = ext4_find_unwritten_pgoff(inode, SEEK_HOLE,
							      &map, &holeoff);
			if (!unwritten) {
				last += ret;
				holeoff = (loff_t)last << blkbits;
562 563
				continue;
			}
564
		}
565 566 567 568 569

		/* find a hole */
		break;
	} while (last <= end);

570 571
	mutex_unlock(&inode->i_mutex);

572 573 574 575
	if (holeoff > isize)
		holeoff = isize;

	return vfs_setpos(file, holeoff, maxsize);
576 577
}

578
/*
579 580 581
 * ext4_llseek() handles both block-mapped and extent-mapped maxbytes values
 * by calling generic_file_llseek_size() with the appropriate maxbytes
 * value for each.
582
 */
583
loff_t ext4_llseek(struct file *file, loff_t offset, int whence)
584 585 586 587 588 589 590 591 592
{
	struct inode *inode = file->f_mapping->host;
	loff_t maxbytes;

	if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)))
		maxbytes = EXT4_SB(inode->i_sb)->s_bitmap_maxbytes;
	else
		maxbytes = inode->i_sb->s_maxbytes;

593
	switch (whence) {
594 595 596
	case SEEK_SET:
	case SEEK_CUR:
	case SEEK_END:
597
		return generic_file_llseek_size(file, offset, whence,
598 599 600 601 602 603 604 605
						maxbytes, i_size_read(inode));
	case SEEK_DATA:
		return ext4_seek_data(file, offset, maxbytes);
	case SEEK_HOLE:
		return ext4_seek_hole(file, offset, maxbytes);
	}

	return -EINVAL;
606 607
}

608
const struct file_operations ext4_file_operations = {
609
	.llseek		= ext4_llseek,
610
	.read		= new_sync_read,
A
Al Viro 已提交
611
	.write		= new_sync_write,
612
	.read_iter	= generic_file_read_iter,
A
Al Viro 已提交
613
	.write_iter	= ext4_file_write_iter,
A
Andi Kleen 已提交
614
	.unlocked_ioctl = ext4_ioctl,
615
#ifdef CONFIG_COMPAT
616
	.compat_ioctl	= ext4_compat_ioctl,
617
#endif
618
	.mmap		= ext4_file_mmap,
619
	.open		= ext4_file_open,
620 621
	.release	= ext4_release_file,
	.fsync		= ext4_sync_file,
622
	.splice_read	= generic_file_splice_read,
A
Al Viro 已提交
623
	.splice_write	= iter_file_splice_write,
624
	.fallocate	= ext4_fallocate,
625 626
};

R
Ross Zwisler 已提交
627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646
#ifdef CONFIG_FS_DAX
const struct file_operations ext4_dax_file_operations = {
	.llseek		= ext4_llseek,
	.read		= new_sync_read,
	.write		= new_sync_write,
	.read_iter	= generic_file_read_iter,
	.write_iter	= ext4_file_write_iter,
	.unlocked_ioctl = ext4_ioctl,
#ifdef CONFIG_COMPAT
	.compat_ioctl	= ext4_compat_ioctl,
#endif
	.mmap		= ext4_file_mmap,
	.open		= ext4_file_open,
	.release	= ext4_release_file,
	.fsync		= ext4_sync_file,
	/* Splice not yet supported with DAX */
	.fallocate	= ext4_fallocate,
};
#endif

647
const struct inode_operations ext4_file_inode_operations = {
648
	.setattr	= ext4_setattr,
649
	.getattr	= ext4_getattr,
650 651
	.setxattr	= generic_setxattr,
	.getxattr	= generic_getxattr,
652
	.listxattr	= ext4_listxattr,
653
	.removexattr	= generic_removexattr,
654
	.get_acl	= ext4_get_acl,
655
	.set_acl	= ext4_set_acl,
656
	.fiemap		= ext4_fiemap,
657 658
};