file.c 19.5 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 24
#include <linux/mount.h>
#include <linux/path.h>
25
#include <linux/dax.h>
26
#include <linux/quotaops.h>
27
#include <linux/pagevec.h>
28
#include <linux/uio.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;
98
	int o_direct = iocb->ki_flags & IOCB_DIRECT;
99
	int overwrite = 0;
100
	ssize_t ret;
101

102 103 104 105 106 107 108
	/*
	 * 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) &&
109
	    (iocb->ki_flags & IOCB_APPEND ||
110
	     ext4_unaligned_aio(inode, from, iocb->ki_pos))) {
111 112 113 114 115 116
		aio_mutex = ext4_aio_mutex(inode);
		mutex_lock(aio_mutex);
		ext4_unwritten_wait(inode);
	}

	mutex_lock(&inode->i_mutex);
117 118
	ret = generic_write_checks(iocb, from);
	if (ret <= 0)
119
		goto out;
120

121 122 123 124
	/*
	 * If we have encountered a bitmap-format file, the size limit
	 * is smaller than s_maxbytes, which is for extent-mapped files.
	 */
125
	if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))) {
126
		struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
127

128
		if (iocb->ki_pos >= sbi->s_bitmap_maxbytes) {
129
			ret = -EFBIG;
130
			goto out;
131
		}
132
		iov_iter_truncate(from, sbi->s_bitmap_maxbytes - iocb->ki_pos);
133 134
	}

135
	iocb->private = &overwrite;
136
	if (o_direct) {
137 138
		size_t length = iov_iter_count(from);
		loff_t pos = iocb->ki_pos;
139 140 141
		blk_start_plug(&plug);

		/* check whether we do a DIO overwrite or not */
142
		if (ext4_should_dioread_nolock(inode) && !aio_mutex &&
143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168
		    !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;
		}
169
	}
170

A
Al Viro 已提交
171
	ret = __generic_file_write_iter(iocb, from);
172
	mutex_unlock(&inode->i_mutex);
173

174 175 176 177 178 179
	if (ret > 0) {
		ssize_t err;

		err = generic_write_sync(file, iocb->ki_pos - ret, ret);
		if (err < 0)
			ret = err;
180
	}
181 182
	if (o_direct)
		blk_finish_plug(&plug);
183

184 185 186 187 188 189
	if (aio_mutex)
		mutex_unlock(aio_mutex);
	return ret;

out:
	mutex_unlock(&inode->i_mutex);
190 191
	if (aio_mutex)
		mutex_unlock(aio_mutex);
192
	return ret;
193 194
}

R
Ross Zwisler 已提交
195
#ifdef CONFIG_FS_DAX
196 197 198
static void ext4_end_io_unwritten(struct buffer_head *bh, int uptodate)
{
	struct inode *inode = bh->b_assoc_map->host;
199
	/* XXX: breaks on 32-bit > 16TB. Is that even supported? */
200 201 202 203 204 205 206 207
	loff_t offset = (loff_t)(uintptr_t)bh->b_private << inode->i_blkbits;
	int err;
	if (!uptodate)
		return;
	WARN_ON(!buffer_unwritten(bh));
	err = ext4_convert_unwritten_extents(NULL, inode, offset, bh->b_size);
}

R
Ross Zwisler 已提交
208 209
static int ext4_dax_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
{
210 211
	int result;
	handle_t *handle = NULL;
212 213
	struct inode *inode = file_inode(vma->vm_file);
	struct super_block *sb = inode->i_sb;
214 215 216 217 218
	bool write = vmf->flags & FAULT_FLAG_WRITE;

	if (write) {
		sb_start_pagefault(sb);
		file_update_time(vma->vm_file);
219
		down_read(&EXT4_I(inode)->i_mmap_sem);
220 221
		handle = ext4_journal_start_sb(sb, EXT4_HT_WRITE_PAGE,
						EXT4_DATA_TRANS_BLOCKS(sb));
222 223
	} else
		down_read(&EXT4_I(inode)->i_mmap_sem);
224 225 226 227 228 229 230 231 232 233

	if (IS_ERR(handle))
		result = VM_FAULT_SIGBUS;
	else
		result = __dax_fault(vma, vmf, ext4_get_block_dax,
						ext4_end_io_unwritten);

	if (write) {
		if (!IS_ERR(handle))
			ext4_journal_stop(handle);
234
		up_read(&EXT4_I(inode)->i_mmap_sem);
235
		sb_end_pagefault(sb);
236 237
	} else
		up_read(&EXT4_I(inode)->i_mmap_sem);
238 239

	return result;
R
Ross Zwisler 已提交
240 241
}

M
Matthew Wilcox 已提交
242 243 244
static int ext4_dax_pmd_fault(struct vm_area_struct *vma, unsigned long addr,
						pmd_t *pmd, unsigned int flags)
{
245 246 247 248 249 250 251 252 253
	int result;
	handle_t *handle = NULL;
	struct inode *inode = file_inode(vma->vm_file);
	struct super_block *sb = inode->i_sb;
	bool write = flags & FAULT_FLAG_WRITE;

	if (write) {
		sb_start_pagefault(sb);
		file_update_time(vma->vm_file);
254
		down_read(&EXT4_I(inode)->i_mmap_sem);
255 256 257
		handle = ext4_journal_start_sb(sb, EXT4_HT_WRITE_PAGE,
				ext4_chunk_trans_blocks(inode,
							PMD_SIZE / PAGE_SIZE));
258 259
	} else
		down_read(&EXT4_I(inode)->i_mmap_sem);
260 261 262 263 264 265 266 267 268 269

	if (IS_ERR(handle))
		result = VM_FAULT_SIGBUS;
	else
		result = __dax_pmd_fault(vma, addr, pmd, flags,
				ext4_get_block_dax, ext4_end_io_unwritten);

	if (write) {
		if (!IS_ERR(handle))
			ext4_journal_stop(handle);
270
		up_read(&EXT4_I(inode)->i_mmap_sem);
271
		sb_end_pagefault(sb);
272 273
	} else
		up_read(&EXT4_I(inode)->i_mmap_sem);
274 275

	return result;
M
Matthew Wilcox 已提交
276 277
}

R
Ross Zwisler 已提交
278 279
static int ext4_dax_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
{
280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320
	int err;
	struct inode *inode = file_inode(vma->vm_file);

	sb_start_pagefault(inode->i_sb);
	file_update_time(vma->vm_file);
	down_read(&EXT4_I(inode)->i_mmap_sem);
	err = __dax_mkwrite(vma, vmf, ext4_get_block_dax,
			    ext4_end_io_unwritten);
	up_read(&EXT4_I(inode)->i_mmap_sem);
	sb_end_pagefault(inode->i_sb);

	return err;
}

/*
 * Handle write fault for VM_MIXEDMAP mappings. Similarly to ext4_dax_mkwrite()
 * handler we check for races agaist truncate. Note that since we cycle through
 * i_mmap_sem, we are sure that also any hole punching that began before we
 * were called is finished by now and so if it included part of the file we
 * are working on, our pte will get unmapped and the check for pte_same() in
 * wp_pfn_shared() fails. Thus fault gets retried and things work out as
 * desired.
 */
static int ext4_dax_pfn_mkwrite(struct vm_area_struct *vma,
				struct vm_fault *vmf)
{
	struct inode *inode = file_inode(vma->vm_file);
	struct super_block *sb = inode->i_sb;
	int ret = VM_FAULT_NOPAGE;
	loff_t size;

	sb_start_pagefault(sb);
	file_update_time(vma->vm_file);
	down_read(&EXT4_I(inode)->i_mmap_sem);
	size = (i_size_read(inode) + PAGE_SIZE - 1) >> PAGE_SHIFT;
	if (vmf->pgoff >= size)
		ret = VM_FAULT_SIGBUS;
	up_read(&EXT4_I(inode)->i_mmap_sem);
	sb_end_pagefault(sb);

	return ret;
R
Ross Zwisler 已提交
321 322 323 324
}

static const struct vm_operations_struct ext4_dax_vm_ops = {
	.fault		= ext4_dax_fault,
M
Matthew Wilcox 已提交
325
	.pmd_fault	= ext4_dax_pmd_fault,
R
Ross Zwisler 已提交
326
	.page_mkwrite	= ext4_dax_mkwrite,
327
	.pfn_mkwrite	= ext4_dax_pfn_mkwrite,
R
Ross Zwisler 已提交
328 329 330 331 332
};
#else
#define ext4_dax_vm_ops	ext4_file_vm_ops
#endif

333
static const struct vm_operations_struct ext4_file_vm_ops = {
334
	.fault		= ext4_filemap_fault,
335
	.map_pages	= filemap_map_pages,
336 337 338 339 340
	.page_mkwrite   = ext4_page_mkwrite,
};

static int ext4_file_mmap(struct file *file, struct vm_area_struct *vma)
{
341 342 343
	struct inode *inode = file->f_mapping->host;

	if (ext4_encrypted_inode(inode)) {
344
		int err = ext4_get_encryption_info(inode);
345 346
		if (err)
			return 0;
347 348
		if (ext4_encryption_info(inode) == NULL)
			return -ENOKEY;
349
	}
350
	file_accessed(file);
R
Ross Zwisler 已提交
351 352
	if (IS_DAX(file_inode(file))) {
		vma->vm_ops = &ext4_dax_vm_ops;
M
Matthew Wilcox 已提交
353
		vma->vm_flags |= VM_MIXEDMAP | VM_HUGEPAGE;
R
Ross Zwisler 已提交
354 355 356
	} else {
		vma->vm_ops = &ext4_file_vm_ops;
	}
357 358 359
	return 0;
}

360 361 362 363 364 365 366
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;
367
	int ret;
368 369 370 371 372 373 374 375 376 377 378

	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));
379 380
		path.mnt = mnt;
		path.dentry = mnt->mnt_root;
381 382
		cp = d_path(&path, buf, sizeof(buf));
		if (!IS_ERR(cp)) {
383 384 385
			handle_t *handle;
			int err;

386
			handle = ext4_journal_start_sb(sb, EXT4_HT_MISC, 1);
387 388
			if (IS_ERR(handle))
				return PTR_ERR(handle);
389
			BUFFER_TRACE(sbi->s_sbh, "get_write_access");
390 391 392 393 394
			err = ext4_journal_get_write_access(handle, sbi->s_sbh);
			if (err) {
				ext4_journal_stop(handle);
				return err;
			}
395 396
			strlcpy(sbi->s_es->s_last_mounted, cp,
				sizeof(sbi->s_es->s_last_mounted));
397 398
			ext4_handle_dirty_super(handle, sb);
			ext4_journal_stop(handle);
399 400
		}
	}
401 402 403 404 405 406 407
	if (ext4_encrypted_inode(inode)) {
		ret = ext4_get_encryption_info(inode);
		if (ret)
			return -EACCES;
		if (ext4_encryption_info(inode) == NULL)
			return -ENOKEY;
	}
408 409 410 411
	/*
	 * Set up the jbd2_inode if we are opening the inode for
	 * writing and the journal is present
	 */
412
	if (filp->f_mode & FMODE_WRITE) {
413
		ret = ext4_inode_attach_jinode(inode);
414 415
		if (ret < 0)
			return ret;
416
	}
417
	return dquot_file_open(inode, filp);
418 419
}

420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435
/*
 * 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.
 */
436 437 438 439
static int ext4_find_unwritten_pgoff(struct inode *inode,
				     int whence,
				     struct ext4_map_blocks *map,
				     loff_t *offset)
440 441
{
	struct pagevec pvec;
442
	unsigned int blkbits;
443 444
	pgoff_t index;
	pgoff_t end;
445
	loff_t endoff;
446 447 448 449
	loff_t startoff;
	loff_t lastoff;
	int found = 0;

450
	blkbits = inode->i_sb->s_blocksize_bits;
451 452
	startoff = *offset;
	lastoff = startoff;
453
	endoff = (loff_t)(map->m_lblk + map->m_len) << blkbits;
454 455 456 457 458 459 460 461 462 463 464 465 466

	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) {
467
			if (whence == SEEK_DATA)
468 469
				break;

470
			BUG_ON(whence != SEEK_HOLE);
471 472 473 474 475 476 477 478 479 480 481 482 483 484 485
			/*
			 * 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.
		 */
486
		if (lastoff == startoff && whence == SEEK_HOLE &&
487 488 489 490 491 492 493 494 495 496 497 498 499
		    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.
			 */
500
			if (lastoff < endoff && whence == SEEK_HOLE &&
501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524
			    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)) {
525
						if (whence == SEEK_DATA)
526 527
							found = 1;
					} else {
528
						if (whence == SEEK_HOLE)
529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549
							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.
		 */
550
		if (nr_pages < num && whence == SEEK_HOLE) {
551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570
			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;
571 572 573 574 575 576
	struct ext4_map_blocks map;
	struct extent_status es;
	ext4_lblk_t start, last, end;
	loff_t dataoff, isize;
	int blkbits;
	int ret = 0;
577 578

	mutex_lock(&inode->i_mutex);
579 580 581

	isize = i_size_read(inode);
	if (offset >= isize) {
582 583 584
		mutex_unlock(&inode->i_mutex);
		return -ENXIO;
	}
585 586 587 588 589 590 591 592 593 594 595 596 597 598

	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;
599
			break;
600
		}
601

602 603 604 605 606 607 608 609
		/*
		 * 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;
610 611 612
			break;
		}

613 614 615 616 617 618 619 620 621 622 623 624
		/*
		 * 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;
		}
625

626 627 628
		last++;
		dataoff = (loff_t)last << blkbits;
	} while (last <= end);
629 630 631

	mutex_unlock(&inode->i_mutex);

632 633 634 635
	if (dataoff > isize)
		return -ENXIO;

	return vfs_setpos(file, dataoff, maxsize);
636 637 638
}

/*
639
 * ext4_seek_hole() retrieves the offset for SEEK_HOLE.
640 641 642 643
 */
static loff_t ext4_seek_hole(struct file *file, loff_t offset, loff_t maxsize)
{
	struct inode *inode = file->f_mapping->host;
644 645 646 647 648 649
	struct ext4_map_blocks map;
	struct extent_status es;
	ext4_lblk_t start, last, end;
	loff_t holeoff, isize;
	int blkbits;
	int ret = 0;
650 651

	mutex_lock(&inode->i_mutex);
652 653 654

	isize = i_size_read(inode);
	if (offset >= isize) {
655 656 657 658
		mutex_unlock(&inode->i_mutex);
		return -ENXIO;
	}

659 660 661 662 663
	blkbits = inode->i_sb->s_blocksize_bits;
	start = offset >> blkbits;
	last = start;
	end = isize >> blkbits;
	holeoff = offset;
664

665 666 667 668 669 670 671 672 673
	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;
		}
674

675 676 677 678 679 680 681 682 683 684
		/*
		 * 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;
		}
685

686 687 688 689 690 691 692 693 694 695 696 697
		/*
		 * 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;
698 699
				continue;
			}
700
		}
701 702 703 704 705

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

706 707
	mutex_unlock(&inode->i_mutex);

708 709 710 711
	if (holeoff > isize)
		holeoff = isize;

	return vfs_setpos(file, holeoff, maxsize);
712 713
}

714
/*
715 716 717
 * ext4_llseek() handles both block-mapped and extent-mapped maxbytes values
 * by calling generic_file_llseek_size() with the appropriate maxbytes
 * value for each.
718
 */
719
loff_t ext4_llseek(struct file *file, loff_t offset, int whence)
720 721 722 723 724 725 726 727 728
{
	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;

729
	switch (whence) {
730 731 732
	case SEEK_SET:
	case SEEK_CUR:
	case SEEK_END:
733
		return generic_file_llseek_size(file, offset, whence,
734 735 736 737 738 739 740 741
						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;
742 743
}

744
const struct file_operations ext4_file_operations = {
745
	.llseek		= ext4_llseek,
746
	.read_iter	= generic_file_read_iter,
A
Al Viro 已提交
747
	.write_iter	= ext4_file_write_iter,
A
Andi Kleen 已提交
748
	.unlocked_ioctl = ext4_ioctl,
749
#ifdef CONFIG_COMPAT
750
	.compat_ioctl	= ext4_compat_ioctl,
751
#endif
752
	.mmap		= ext4_file_mmap,
753
	.open		= ext4_file_open,
754 755
	.release	= ext4_release_file,
	.fsync		= ext4_sync_file,
756
	.splice_read	= generic_file_splice_read,
A
Al Viro 已提交
757
	.splice_write	= iter_file_splice_write,
758
	.fallocate	= ext4_fallocate,
759 760
};

761
const struct inode_operations ext4_file_inode_operations = {
762
	.setattr	= ext4_setattr,
763
	.getattr	= ext4_getattr,
764 765
	.setxattr	= generic_setxattr,
	.getxattr	= generic_getxattr,
766
	.listxattr	= ext4_listxattr,
767
	.removexattr	= generic_removexattr,
768
	.get_acl	= ext4_get_acl,
769
	.set_acl	= ext4_set_acl,
770
	.fiemap		= ext4_fiemap,
771 772
};