data.c 25.2 KB
Newer Older
J
Jaegeuk Kim 已提交
1
/*
2 3 4 5 6 7 8 9 10 11 12 13 14
 * fs/f2fs/data.c
 *
 * Copyright (c) 2012 Samsung Electronics Co., Ltd.
 *             http://www.samsung.com/
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as
 * published by the Free Software Foundation.
 */
#include <linux/fs.h>
#include <linux/f2fs_fs.h>
#include <linux/buffer_head.h>
#include <linux/mpage.h>
15
#include <linux/aio.h>
16 17 18 19
#include <linux/writeback.h>
#include <linux/backing-dev.h>
#include <linux/blkdev.h>
#include <linux/bio.h>
20
#include <linux/prefetch.h>
21 22 23 24

#include "f2fs.h"
#include "node.h"
#include "segment.h"
25
#include <trace/events/f2fs.h>
26

27 28
static void f2fs_read_end_io(struct bio *bio, int err)
{
29 30
	struct bio_vec *bvec;
	int i;
31

32
	bio_for_each_segment_all(bvec, bio, i) {
33 34
		struct page *page = bvec->bv_page;

35 36 37
		if (!err) {
			SetPageUptodate(page);
		} else {
38 39 40 41
			ClearPageUptodate(page);
			SetPageError(page);
		}
		unlock_page(page);
42
	}
43 44 45 46 47
	bio_put(bio);
}

static void f2fs_write_end_io(struct bio *bio, int err)
{
48
	struct f2fs_sb_info *sbi = bio->bi_private;
49 50
	struct bio_vec *bvec;
	int i;
51

52
	bio_for_each_segment_all(bvec, bio, i) {
53 54
		struct page *page = bvec->bv_page;

55
		if (unlikely(err)) {
56 57
			SetPageError(page);
			set_bit(AS_EIO, &page->mapping->flags);
58
			f2fs_stop_checkpoint(sbi);
59 60 61
		}
		end_page_writeback(page);
		dec_page_count(sbi, F2FS_WRITEBACK);
62
	}
63

64 65 66 67
	if (sbi->wait_io) {
		complete(sbi->wait_io);
		sbi->wait_io = NULL;
	}
68 69 70 71 72 73 74 75

	if (!get_pages(sbi, F2FS_WRITEBACK) &&
			!list_empty(&sbi->cp_wait.task_list))
		wake_up(&sbi->cp_wait);

	bio_put(bio);
}

76 77 78 79 80 81 82 83 84 85 86 87
/*
 * Low-level block read/write IO operations.
 */
static struct bio *__bio_alloc(struct f2fs_sb_info *sbi, block_t blk_addr,
				int npages, bool is_read)
{
	struct bio *bio;

	/* No failure on bio allocation */
	bio = bio_alloc(GFP_NOIO, npages);

	bio->bi_bdev = sbi->sb->s_bdev;
88
	bio->bi_iter.bi_sector = SECTOR_FROM_BLOCK(sbi, blk_addr);
89
	bio->bi_end_io = is_read ? f2fs_read_end_io : f2fs_write_end_io;
90
	bio->bi_private = sbi;
91 92 93 94

	return bio;
}

J
Jaegeuk Kim 已提交
95
static void __submit_merged_bio(struct f2fs_bio_info *io)
96
{
J
Jaegeuk Kim 已提交
97 98
	struct f2fs_io_info *fio = &io->fio;
	int rw;
99 100 101 102

	if (!io->bio)
		return;

103
	rw = fio->rw;
104 105

	if (is_read_io(rw)) {
106 107
		trace_f2fs_submit_read_bio(io->sbi->sb, rw,
						fio->type, io->bio);
J
Jaegeuk Kim 已提交
108
		submit_bio(rw, io->bio);
109
	} else {
110 111 112 113 114 115 116 117
		trace_f2fs_submit_write_bio(io->sbi->sb, rw,
						fio->type, io->bio);
		/*
		 * META_FLUSH is only from the checkpoint procedure, and we
		 * should wait this metadata bio for FS consistency.
		 */
		if (fio->type == META_FLUSH) {
			DECLARE_COMPLETION_ONSTACK(wait);
118
			io->sbi->wait_io = &wait;
119 120 121 122 123
			submit_bio(rw, io->bio);
			wait_for_completion(&wait);
		} else {
			submit_bio(rw, io->bio);
		}
124
	}
125

126 127 128 129
	io->bio = NULL;
}

void f2fs_submit_merged_bio(struct f2fs_sb_info *sbi,
J
Jaegeuk Kim 已提交
130
				enum page_type type, int rw)
131 132 133 134 135 136
{
	enum page_type btype = PAGE_TYPE_OF_BIO(type);
	struct f2fs_bio_info *io;

	io = is_read_io(rw) ? &sbi->read_io : &sbi->write_io[btype];

137
	down_write(&io->io_rwsem);
J
Jaegeuk Kim 已提交
138 139 140 141

	/* change META to META_FLUSH in the checkpoint procedure */
	if (type >= META_FLUSH) {
		io->fio.type = META_FLUSH;
142
		io->fio.rw = WRITE_FLUSH_FUA | REQ_META | REQ_PRIO;
J
Jaegeuk Kim 已提交
143 144
	}
	__submit_merged_bio(io);
145
	up_write(&io->io_rwsem);
146 147 148 149 150 151 152 153 154 155 156 157 158 159
}

/*
 * Fill the locked page with data located in the block address.
 * Return unlocked page.
 */
int f2fs_submit_page_bio(struct f2fs_sb_info *sbi, struct page *page,
					block_t blk_addr, int rw)
{
	struct bio *bio;

	trace_f2fs_submit_page_bio(page, blk_addr, rw);

	/* Allocate a new bio */
160
	bio = __bio_alloc(sbi, blk_addr, 1, is_read_io(rw));
161 162 163 164 165 166 167 168 169 170 171 172

	if (bio_add_page(bio, page, PAGE_CACHE_SIZE, 0) < PAGE_CACHE_SIZE) {
		bio_put(bio);
		f2fs_put_page(page, 1);
		return -EFAULT;
	}

	submit_bio(rw, bio);
	return 0;
}

void f2fs_submit_page_mbio(struct f2fs_sb_info *sbi, struct page *page,
J
Jaegeuk Kim 已提交
173
			block_t blk_addr, struct f2fs_io_info *fio)
174
{
J
Jaegeuk Kim 已提交
175
	enum page_type btype = PAGE_TYPE_OF_BIO(fio->type);
176
	struct f2fs_bio_info *io;
177
	bool is_read = is_read_io(fio->rw);
178

179
	io = is_read ? &sbi->read_io : &sbi->write_io[btype];
180 181 182

	verify_block_addr(sbi, blk_addr);

183
	down_write(&io->io_rwsem);
184

185
	if (!is_read)
186 187
		inc_page_count(sbi, F2FS_WRITEBACK);

188
	if (io->bio && (io->last_block_in_bio != blk_addr - 1 ||
J
Jaegeuk Kim 已提交
189 190
						io->fio.rw != fio->rw))
		__submit_merged_bio(io);
191 192
alloc_new:
	if (io->bio == NULL) {
193 194 195
		int bio_blocks = MAX_BIO_BLOCKS(max_hw_blocks(sbi));

		io->bio = __bio_alloc(sbi, blk_addr, bio_blocks, is_read);
J
Jaegeuk Kim 已提交
196
		io->fio = *fio;
197 198 199 200
	}

	if (bio_add_page(io->bio, page, PAGE_CACHE_SIZE, 0) <
							PAGE_CACHE_SIZE) {
J
Jaegeuk Kim 已提交
201
		__submit_merged_bio(io);
202 203 204 205 206
		goto alloc_new;
	}

	io->last_block_in_bio = blk_addr;

207
	up_write(&io->io_rwsem);
J
Jaegeuk Kim 已提交
208
	trace_f2fs_submit_page_mbio(page, fio->rw, fio->type, blk_addr);
209 210
}

J
Jaegeuk Kim 已提交
211
/*
212 213 214 215 216 217 218 219 220 221 222 223
 * Lock ordering for the change of data block address:
 * ->data_page
 *  ->node_page
 *    update block addresses in the node page
 */
static void __set_data_blkaddr(struct dnode_of_data *dn, block_t new_addr)
{
	struct f2fs_node *rn;
	__le32 *addr_array;
	struct page *node_page = dn->node_page;
	unsigned int ofs_in_node = dn->ofs_in_node;

224
	f2fs_wait_on_page_writeback(node_page, NODE);
225

226
	rn = F2FS_NODE(node_page);
227 228 229 230 231 232 233 234 235 236 237

	/* Get physical address of data block */
	addr_array = blkaddr_in_node(rn);
	addr_array[ofs_in_node] = cpu_to_le32(new_addr);
	set_page_dirty(node_page);
}

int reserve_new_block(struct dnode_of_data *dn)
{
	struct f2fs_sb_info *sbi = F2FS_SB(dn->inode->i_sb);

238
	if (unlikely(is_inode_flag_set(F2FS_I(dn->inode), FI_NO_ALLOC)))
239
		return -EPERM;
240
	if (unlikely(!inc_valid_block_count(sbi, dn->inode, 1)))
241 242
		return -ENOSPC;

243 244
	trace_f2fs_reserve_new_block(dn->inode, dn->nid, dn->ofs_in_node);

245 246
	__set_data_blkaddr(dn, NEW_ADDR);
	dn->data_blkaddr = NEW_ADDR;
247
	mark_inode_dirty(dn->inode);
248 249 250 251
	sync_inode_page(dn);
	return 0;
}

252 253 254 255 256
int f2fs_reserve_block(struct dnode_of_data *dn, pgoff_t index)
{
	bool need_put = dn->inode_page ? false : true;
	int err;

257 258 259
	/* if inode_page exists, index should be zero */
	f2fs_bug_on(!need_put && index);

260 261 262
	err = get_dnode_of_data(dn, index, ALLOC_NODE);
	if (err)
		return err;
263

264 265
	if (dn->data_blkaddr == NULL_ADDR)
		err = reserve_new_block(dn);
266
	if (err || need_put)
267 268 269 270
		f2fs_put_dnode(dn);
	return err;
}

271 272 273 274 275 276 277
static int check_extent_cache(struct inode *inode, pgoff_t pgofs,
					struct buffer_head *bh_result)
{
	struct f2fs_inode_info *fi = F2FS_I(inode);
	pgoff_t start_fofs, end_fofs;
	block_t start_blkaddr;

278 279 280
	if (is_inode_flag_set(fi, FI_NO_EXTENT))
		return 0;

281 282 283 284 285 286
	read_lock(&fi->ext.ext_lock);
	if (fi->ext.len == 0) {
		read_unlock(&fi->ext.ext_lock);
		return 0;
	}

287 288
	stat_inc_total_hit(inode->i_sb);

289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305
	start_fofs = fi->ext.fofs;
	end_fofs = fi->ext.fofs + fi->ext.len - 1;
	start_blkaddr = fi->ext.blk_addr;

	if (pgofs >= start_fofs && pgofs <= end_fofs) {
		unsigned int blkbits = inode->i_sb->s_blocksize_bits;
		size_t count;

		clear_buffer_new(bh_result);
		map_bh(bh_result, inode->i_sb,
				start_blkaddr + pgofs - start_fofs);
		count = end_fofs - pgofs + 1;
		if (count < (UINT_MAX >> blkbits))
			bh_result->b_size = (count << blkbits);
		else
			bh_result->b_size = UINT_MAX;

306
		stat_inc_read_hit(inode->i_sb);
307 308 309 310 311 312 313 314 315 316 317 318
		read_unlock(&fi->ext.ext_lock);
		return 1;
	}
	read_unlock(&fi->ext.ext_lock);
	return 0;
}

void update_extent_cache(block_t blk_addr, struct dnode_of_data *dn)
{
	struct f2fs_inode_info *fi = F2FS_I(dn->inode);
	pgoff_t fofs, start_fofs, end_fofs;
	block_t start_blkaddr, end_blkaddr;
319
	int need_update = true;
320

321
	f2fs_bug_on(blk_addr == NEW_ADDR);
322 323
	fofs = start_bidx_of_node(ofs_of_node(dn->node_page), fi) +
							dn->ofs_in_node;
324 325 326 327

	/* Update the page address in the parent node */
	__set_data_blkaddr(dn, blk_addr);

328 329 330
	if (is_inode_flag_set(fi, FI_NO_EXTENT))
		return;

331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351
	write_lock(&fi->ext.ext_lock);

	start_fofs = fi->ext.fofs;
	end_fofs = fi->ext.fofs + fi->ext.len - 1;
	start_blkaddr = fi->ext.blk_addr;
	end_blkaddr = fi->ext.blk_addr + fi->ext.len - 1;

	/* Drop and initialize the matched extent */
	if (fi->ext.len == 1 && fofs == start_fofs)
		fi->ext.len = 0;

	/* Initial extent */
	if (fi->ext.len == 0) {
		if (blk_addr != NULL_ADDR) {
			fi->ext.fofs = fofs;
			fi->ext.blk_addr = blk_addr;
			fi->ext.len = 1;
		}
		goto end_update;
	}

N
Namjae Jeon 已提交
352
	/* Front merge */
353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376
	if (fofs == start_fofs - 1 && blk_addr == start_blkaddr - 1) {
		fi->ext.fofs--;
		fi->ext.blk_addr--;
		fi->ext.len++;
		goto end_update;
	}

	/* Back merge */
	if (fofs == end_fofs + 1 && blk_addr == end_blkaddr + 1) {
		fi->ext.len++;
		goto end_update;
	}

	/* Split the existing extent */
	if (fi->ext.len > 1 &&
		fofs >= start_fofs && fofs <= end_fofs) {
		if ((end_fofs - fofs) < (fi->ext.len >> 1)) {
			fi->ext.len = fofs - start_fofs;
		} else {
			fi->ext.fofs = fofs + 1;
			fi->ext.blk_addr = start_blkaddr +
					fofs - start_fofs + 1;
			fi->ext.len -= fofs - start_fofs + 1;
		}
377 378
	} else {
		need_update = false;
379 380
	}

381 382 383 384 385 386
	/* Finally, if the extent is very fragmented, let's drop the cache. */
	if (fi->ext.len < F2FS_MIN_EXTENT_LEN) {
		fi->ext.len = 0;
		set_inode_flag(fi, FI_NO_EXTENT);
		need_update = true;
	}
387 388
end_update:
	write_unlock(&fi->ext.ext_lock);
389 390 391
	if (need_update)
		sync_inode_page(dn);
	return;
392 393
}

394
struct page *find_data_page(struct inode *inode, pgoff_t index, bool sync)
395 396 397 398 399 400 401 402 403 404 405 406 407
{
	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
	struct address_space *mapping = inode->i_mapping;
	struct dnode_of_data dn;
	struct page *page;
	int err;

	page = find_get_page(mapping, index);
	if (page && PageUptodate(page))
		return page;
	f2fs_put_page(page, 0);

	set_new_dnode(&dn, inode, NULL, NULL, 0);
408
	err = get_dnode_of_data(&dn, index, LOOKUP_NODE);
409 410 411 412 413 414 415 416
	if (err)
		return ERR_PTR(err);
	f2fs_put_dnode(&dn);

	if (dn.data_blkaddr == NULL_ADDR)
		return ERR_PTR(-ENOENT);

	/* By fallocate(), there is no cached page, but with NEW_ADDR */
417
	if (unlikely(dn.data_blkaddr == NEW_ADDR))
418 419
		return ERR_PTR(-EINVAL);

420
	page = grab_cache_page_write_begin(mapping, index, AOP_FLAG_NOFS);
421 422 423
	if (!page)
		return ERR_PTR(-ENOMEM);

424 425 426 427 428
	if (PageUptodate(page)) {
		unlock_page(page);
		return page;
	}

429
	err = f2fs_submit_page_bio(sbi, page, dn.data_blkaddr,
430
					sync ? READ_SYNC : READA);
431 432 433
	if (err)
		return ERR_PTR(err);

434 435
	if (sync) {
		wait_on_page_locked(page);
436
		if (unlikely(!PageUptodate(page))) {
437 438 439
			f2fs_put_page(page, 0);
			return ERR_PTR(-EIO);
		}
440 441 442 443
	}
	return page;
}

J
Jaegeuk Kim 已提交
444
/*
445 446 447 448 449 450 451 452 453 454 455 456
 * If it tries to access a hole, return an error.
 * Because, the callers, functions in dir.c and GC, should be able to know
 * whether this page exists or not.
 */
struct page *get_lock_data_page(struct inode *inode, pgoff_t index)
{
	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
	struct address_space *mapping = inode->i_mapping;
	struct dnode_of_data dn;
	struct page *page;
	int err;

457
repeat:
458
	page = grab_cache_page_write_begin(mapping, index, AOP_FLAG_NOFS);
459 460 461
	if (!page)
		return ERR_PTR(-ENOMEM);

462
	set_new_dnode(&dn, inode, NULL, NULL, 0);
463
	err = get_dnode_of_data(&dn, index, LOOKUP_NODE);
464 465
	if (err) {
		f2fs_put_page(page, 1);
466
		return ERR_PTR(err);
467
	}
468 469
	f2fs_put_dnode(&dn);

470
	if (unlikely(dn.data_blkaddr == NULL_ADDR)) {
471
		f2fs_put_page(page, 1);
472
		return ERR_PTR(-ENOENT);
473
	}
474 475 476 477

	if (PageUptodate(page))
		return page;

J
Jaegeuk Kim 已提交
478 479 480 481 482 483 484 485 486 487 488
	/*
	 * A new dentry page is allocated but not able to be written, since its
	 * new inode page couldn't be allocated due to -ENOSPC.
	 * In such the case, its blkaddr can be remained as NEW_ADDR.
	 * see, f2fs_add_link -> get_new_data_page -> init_inode_metadata.
	 */
	if (dn.data_blkaddr == NEW_ADDR) {
		zero_user_segment(page, 0, PAGE_CACHE_SIZE);
		SetPageUptodate(page);
		return page;
	}
489

490
	err = f2fs_submit_page_bio(sbi, page, dn.data_blkaddr, READ_SYNC);
491
	if (err)
492
		return ERR_PTR(err);
493 494

	lock_page(page);
495
	if (unlikely(!PageUptodate(page))) {
496 497
		f2fs_put_page(page, 1);
		return ERR_PTR(-EIO);
498
	}
499
	if (unlikely(page->mapping != mapping)) {
500 501
		f2fs_put_page(page, 1);
		goto repeat;
502 503 504 505
	}
	return page;
}

J
Jaegeuk Kim 已提交
506
/*
507 508
 * Caller ensures that this data page is never allocated.
 * A new zero-filled data page is allocated in the page cache.
509
 *
C
Chao Yu 已提交
510 511
 * Also, caller should grab and release a rwsem by calling f2fs_lock_op() and
 * f2fs_unlock_op().
512
 * Note that, ipage is set only by make_empty_dir.
513
 */
514
struct page *get_new_data_page(struct inode *inode,
515
		struct page *ipage, pgoff_t index, bool new_i_size)
516 517 518 519 520 521 522
{
	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
	struct address_space *mapping = inode->i_mapping;
	struct page *page;
	struct dnode_of_data dn;
	int err;

523
	set_new_dnode(&dn, inode, ipage, NULL, 0);
524
	err = f2fs_reserve_block(&dn, index);
525 526
	if (err)
		return ERR_PTR(err);
527
repeat:
528
	page = grab_cache_page(mapping, index);
529 530 531 532
	if (!page) {
		err = -ENOMEM;
		goto put_err;
	}
533 534 535 536 537 538

	if (PageUptodate(page))
		return page;

	if (dn.data_blkaddr == NEW_ADDR) {
		zero_user_segment(page, 0, PAGE_CACHE_SIZE);
539
		SetPageUptodate(page);
540
	} else {
541 542
		err = f2fs_submit_page_bio(sbi, page, dn.data_blkaddr,
								READ_SYNC);
543
		if (err)
544 545
			goto put_err;

546
		lock_page(page);
547
		if (unlikely(!PageUptodate(page))) {
548
			f2fs_put_page(page, 1);
549 550
			err = -EIO;
			goto put_err;
551
		}
552
		if (unlikely(page->mapping != mapping)) {
553 554
			f2fs_put_page(page, 1);
			goto repeat;
555 556 557 558 559 560
		}
	}

	if (new_i_size &&
		i_size_read(inode) < ((index + 1) << PAGE_CACHE_SHIFT)) {
		i_size_write(inode, ((index + 1) << PAGE_CACHE_SHIFT));
561 562
		/* Only the directory inode sets new_i_size */
		set_inode_flag(F2FS_I(inode), FI_UPDATE_DIR);
563 564
	}
	return page;
565 566 567 568

put_err:
	f2fs_put_dnode(&dn);
	return ERR_PTR(err);
569 570
}

571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602
static int __allocate_data_block(struct dnode_of_data *dn)
{
	struct f2fs_sb_info *sbi = F2FS_SB(dn->inode->i_sb);
	struct f2fs_summary sum;
	block_t new_blkaddr;
	struct node_info ni;
	int type;

	if (unlikely(is_inode_flag_set(F2FS_I(dn->inode), FI_NO_ALLOC)))
		return -EPERM;
	if (unlikely(!inc_valid_block_count(sbi, dn->inode, 1)))
		return -ENOSPC;

	__set_data_blkaddr(dn, NEW_ADDR);
	dn->data_blkaddr = NEW_ADDR;

	get_node_info(sbi, dn->nid, &ni);
	set_summary(&sum, dn->nid, dn->ofs_in_node, ni.version);

	type = CURSEG_WARM_DATA;

	allocate_data_block(sbi, NULL, NULL_ADDR, &new_blkaddr, &sum, type);

	/* direct IO doesn't use extent cache to maximize the performance */
	set_inode_flag(F2FS_I(dn->inode), FI_NO_EXTENT);
	update_extent_cache(new_blkaddr, dn);
	clear_inode_flag(F2FS_I(dn->inode), FI_NO_EXTENT);

	dn->data_blkaddr = new_blkaddr;
	return 0;
}

J
Jaegeuk Kim 已提交
603
/*
C
Chao Yu 已提交
604 605 606 607 608 609
 * get_data_block() now supported readahead/bmap/rw direct_IO with mapped bh.
 * If original data blocks are allocated, then give them to blockdev.
 * Otherwise,
 *     a. preallocate requested block addresses
 *     b. do not use extent cache for better performance
 *     c. give the block addresses to blockdev
610
 */
611
static int get_data_block(struct inode *inode, sector_t iblock,
612 613
			struct buffer_head *bh_result, int create)
{
614
	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
615 616 617
	unsigned int blkbits = inode->i_sb->s_blocksize_bits;
	unsigned maxblocks = bh_result->b_size >> blkbits;
	struct dnode_of_data dn;
618 619 620 621
	int mode = create ? ALLOC_NODE : LOOKUP_NODE_RA;
	pgoff_t pgofs, end_offset;
	int err = 0, ofs = 1;
	bool allocated = false;
622 623 624 625

	/* Get the page offset from the block offset(iblock) */
	pgofs =	(pgoff_t)(iblock >> (PAGE_CACHE_SHIFT - blkbits));

626 627 628 629 630
	if (check_extent_cache(inode, pgofs, bh_result))
		goto out;

	if (create)
		f2fs_lock_op(sbi);
631 632 633

	/* When reading holes, we need its node page */
	set_new_dnode(&dn, inode, NULL, NULL, 0);
634
	err = get_dnode_of_data(&dn, pgofs, mode);
635
	if (err) {
636 637 638
		if (err == -ENOENT)
			err = 0;
		goto unlock_out;
639
	}
640 641
	if (dn.data_blkaddr == NEW_ADDR)
		goto put_out;
642

643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669
	if (dn.data_blkaddr != NULL_ADDR) {
		map_bh(bh_result, inode->i_sb, dn.data_blkaddr);
	} else if (create) {
		err = __allocate_data_block(&dn);
		if (err)
			goto put_out;
		allocated = true;
		map_bh(bh_result, inode->i_sb, dn.data_blkaddr);
	} else {
		goto put_out;
	}

	end_offset = IS_INODE(dn.node_page) ?
			ADDRS_PER_INODE(F2FS_I(inode)) : ADDRS_PER_BLOCK;
	bh_result->b_size = (((size_t)1) << blkbits);
	dn.ofs_in_node++;
	pgofs++;

get_next:
	if (dn.ofs_in_node >= end_offset) {
		if (allocated)
			sync_inode_page(&dn);
		allocated = false;
		f2fs_put_dnode(&dn);

		set_new_dnode(&dn, inode, NULL, NULL, 0);
		err = get_dnode_of_data(&dn, pgofs, mode);
670
		if (err) {
671 672 673 674
			if (err == -ENOENT)
				err = 0;
			goto unlock_out;
		}
675 676 677
		if (dn.data_blkaddr == NEW_ADDR)
			goto put_out;

678
		end_offset = IS_INODE(dn.node_page) ?
679 680
			ADDRS_PER_INODE(F2FS_I(inode)) : ADDRS_PER_BLOCK;
	}
681

682 683 684 685 686 687 688 689 690
	if (maxblocks > (bh_result->b_size >> blkbits)) {
		block_t blkaddr = datablock_addr(dn.node_page, dn.ofs_in_node);
		if (blkaddr == NULL_ADDR && create) {
			err = __allocate_data_block(&dn);
			if (err)
				goto sync_out;
			allocated = true;
			blkaddr = dn.data_blkaddr;
		}
691
		/* Give more consecutive addresses for the read ahead */
692 693 694 695 696 697 698
		if (blkaddr == (bh_result->b_blocknr + ofs)) {
			ofs++;
			dn.ofs_in_node++;
			pgofs++;
			bh_result->b_size += (((size_t)1) << blkbits);
			goto get_next;
		}
699
	}
700 701 702 703
sync_out:
	if (allocated)
		sync_inode_page(&dn);
put_out:
704
	f2fs_put_dnode(&dn);
705 706 707 708 709 710
unlock_out:
	if (create)
		f2fs_unlock_op(sbi);
out:
	trace_f2fs_get_data_block(inode, iblock, bh_result, err);
	return err;
711 712 713 714
}

static int f2fs_read_data_page(struct file *file, struct page *page)
{
H
Huajun Li 已提交
715 716 717 718 719 720 721 722 723 724
	struct inode *inode = page->mapping->host;
	int ret;

	/* If the file has inline data, try to read it directlly */
	if (f2fs_has_inline_data(inode))
		ret = f2fs_read_inline_data(inode, page);
	else
		ret = mpage_readpage(page, get_data_block);

	return ret;
725 726 727 728 729 730
}

static int f2fs_read_data_pages(struct file *file,
			struct address_space *mapping,
			struct list_head *pages, unsigned nr_pages)
{
H
Huajun Li 已提交
731 732 733 734 735 736
	struct inode *inode = file->f_mapping->host;

	/* If the file has inline data, skip readpages */
	if (f2fs_has_inline_data(inode))
		return 0;

737
	return mpage_readpages(mapping, pages, nr_pages, get_data_block);
738 739
}

J
Jaegeuk Kim 已提交
740
int do_write_data_page(struct page *page, struct f2fs_io_info *fio)
741 742
{
	struct inode *inode = page->mapping->host;
J
Jaegeuk Kim 已提交
743
	block_t old_blkaddr, new_blkaddr;
744 745 746 747
	struct dnode_of_data dn;
	int err = 0;

	set_new_dnode(&dn, inode, NULL, NULL, 0);
748
	err = get_dnode_of_data(&dn, page->index, LOOKUP_NODE);
749 750 751
	if (err)
		return err;

J
Jaegeuk Kim 已提交
752
	old_blkaddr = dn.data_blkaddr;
753 754

	/* This page is already truncated */
J
Jaegeuk Kim 已提交
755
	if (old_blkaddr == NULL_ADDR)
756 757 758 759 760 761 762 763
		goto out_writepage;

	set_page_writeback(page);

	/*
	 * If current allocation needs SSR,
	 * it had better in-place writes for updated data.
	 */
J
Jaegeuk Kim 已提交
764
	if (unlikely(old_blkaddr != NEW_ADDR &&
765 766
			!is_cold_data(page) &&
			need_inplace_update(inode))) {
J
Jaegeuk Kim 已提交
767
		rewrite_data_page(page, old_blkaddr, fio);
768
	} else {
J
Jaegeuk Kim 已提交
769 770
		write_data_page(page, &dn, &new_blkaddr, fio);
		update_extent_cache(new_blkaddr, &dn);
771 772 773 774 775 776 777 778 779 780 781 782 783 784
	}
out_writepage:
	f2fs_put_dnode(&dn);
	return err;
}

static int f2fs_write_data_page(struct page *page,
					struct writeback_control *wbc)
{
	struct inode *inode = page->mapping->host;
	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
	loff_t i_size = i_size_read(inode);
	const pgoff_t end_index = ((unsigned long long) i_size)
							>> PAGE_CACHE_SHIFT;
H
Huajun Li 已提交
785
	unsigned offset = 0;
786
	bool need_balance_fs = false;
787
	int err = 0;
J
Jaegeuk Kim 已提交
788 789
	struct f2fs_io_info fio = {
		.type = DATA,
C
Chris Fries 已提交
790
		.rw = (wbc->sync_mode == WB_SYNC_ALL) ? WRITE_SYNC : WRITE,
J
Jaegeuk Kim 已提交
791
	};
792 793

	if (page->index < end_index)
794
		goto write;
795 796 797 798 799 800 801

	/*
	 * If the offset is out-of-range of file size,
	 * this page does not have to be written to disk.
	 */
	offset = i_size & (PAGE_CACHE_SIZE - 1);
	if ((page->index >= end_index + 1) || !offset) {
802
		inode_dec_dirty_dents(inode);
803
		goto out;
804 805 806
	}

	zero_user_segment(page, offset, PAGE_CACHE_SIZE);
807
write:
808
	if (unlikely(sbi->por_doing))
809 810
		goto redirty_out;

811
	/* Dentry blocks are controlled by checkpoint */
812 813
	if (S_ISDIR(inode->i_mode)) {
		inode_dec_dirty_dents(inode);
J
Jaegeuk Kim 已提交
814
		err = do_write_data_page(page, &fio);
815 816
		goto done;
	}
H
Huajun Li 已提交
817

818
	if (!wbc->for_reclaim)
819
		need_balance_fs = true;
820
	else if (has_not_enough_free_secs(sbi, 0))
821
		goto redirty_out;
822

823 824 825 826 827 828 829 830 831
	f2fs_lock_op(sbi);
	if (f2fs_has_inline_data(inode) || f2fs_may_inline(inode))
		err = f2fs_write_inline_data(inode, page, offset);
	else
		err = do_write_data_page(page, &fio);
	f2fs_unlock_op(sbi);
done:
	if (err && err != -ENOENT)
		goto redirty_out;
832 833

	clear_cold_data(page);
834
out:
835
	unlock_page(page);
836
	if (need_balance_fs)
837 838 839 840 841
		f2fs_balance_fs(sbi);
	return 0;

redirty_out:
	wbc->pages_skipped++;
842
	account_page_redirty(page);
843
	set_page_dirty(page);
844
	return AOP_WRITEPAGE_ACTIVATE;
845 846
}

847 848 849 850 851 852 853 854 855
static int __f2fs_writepage(struct page *page, struct writeback_control *wbc,
			void *data)
{
	struct address_space *mapping = data;
	int ret = mapping->a_ops->writepage(page, wbc);
	mapping_set_error(mapping, ret);
	return ret;
}

856
static int f2fs_write_data_pages(struct address_space *mapping,
857 858 859 860
			    struct writeback_control *wbc)
{
	struct inode *inode = mapping->host;
	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
861
	bool locked = false;
862
	int ret;
863
	long diff;
864

P
P J P 已提交
865 866 867 868
	/* deal with chardevs and other special file */
	if (!mapping->a_ops->writepage)
		return 0;

869 870
	if (S_ISDIR(inode->i_mode) && wbc->sync_mode == WB_SYNC_NONE &&
			get_dirty_dents(inode) < nr_pages_to_skip(sbi, DATA))
871
		goto skip_write;
872

873
	diff = nr_pages_to_write(sbi, DATA, wbc);
874

875
	if (!S_ISDIR(inode->i_mode)) {
876
		mutex_lock(&sbi->writepages);
877 878
		locked = true;
	}
879
	ret = write_cache_pages(mapping, wbc, __f2fs_writepage, mapping);
880
	if (locked)
881
		mutex_unlock(&sbi->writepages);
J
Jaegeuk Kim 已提交
882 883

	f2fs_submit_merged_bio(sbi, DATA, WRITE);
884 885 886

	remove_dirty_dir_inode(inode);

887
	wbc->nr_to_write = max((long)0, wbc->nr_to_write - diff);
888
	return ret;
889 890 891 892

skip_write:
	wbc->pages_skipped += get_dirty_dents(inode);
	return 0;
893 894 895 896 897 898 899 900 901 902 903 904 905 906
}

static int f2fs_write_begin(struct file *file, struct address_space *mapping,
		loff_t pos, unsigned len, unsigned flags,
		struct page **pagep, void **fsdata)
{
	struct inode *inode = mapping->host;
	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
	struct page *page;
	pgoff_t index = ((unsigned long long) pos) >> PAGE_CACHE_SHIFT;
	struct dnode_of_data dn;
	int err = 0;

	f2fs_balance_fs(sbi);
907
repeat:
908 909 910 911
	err = f2fs_convert_inline_data(inode, pos + len);
	if (err)
		return err;

912 913 914 915 916
	page = grab_cache_page_write_begin(mapping, index, flags);
	if (!page)
		return -ENOMEM;
	*pagep = page;

917 918
	if (f2fs_has_inline_data(inode) && (pos + len) <= MAX_INLINE_DATA)
		goto inline_data;
H
Huajun Li 已提交
919

920
	f2fs_lock_op(sbi);
921
	set_new_dnode(&dn, inode, NULL, NULL, 0);
922
	err = f2fs_reserve_block(&dn, index);
923
	f2fs_unlock_op(sbi);
924

925 926 927 928
	if (err) {
		f2fs_put_page(page, 1);
		return err;
	}
H
Huajun Li 已提交
929
inline_data:
930 931 932 933 934 935 936 937 938
	if ((len == PAGE_CACHE_SIZE) || PageUptodate(page))
		return 0;

	if ((pos & PAGE_CACHE_MASK) >= i_size_read(inode)) {
		unsigned start = pos & (PAGE_CACHE_SIZE - 1);
		unsigned end = start + len;

		/* Reading beyond i_size is simple: memset to zero */
		zero_user_segments(page, 0, start, end, PAGE_CACHE_SIZE);
939
		goto out;
940 941 942 943 944
	}

	if (dn.data_blkaddr == NEW_ADDR) {
		zero_user_segment(page, 0, PAGE_CACHE_SIZE);
	} else {
945
		if (f2fs_has_inline_data(inode)) {
H
Huajun Li 已提交
946
			err = f2fs_read_inline_data(inode, page);
947 948 949 950 951
			if (err) {
				page_cache_release(page);
				return err;
			}
		} else {
H
Huajun Li 已提交
952
			err = f2fs_submit_page_bio(sbi, page, dn.data_blkaddr,
953
							READ_SYNC);
954 955 956 957
			if (err)
				return err;
		}

958
		lock_page(page);
959
		if (unlikely(!PageUptodate(page))) {
960 961
			f2fs_put_page(page, 1);
			return -EIO;
962
		}
963
		if (unlikely(page->mapping != mapping)) {
964 965
			f2fs_put_page(page, 1);
			goto repeat;
966 967
		}
	}
968
out:
969 970 971 972 973
	SetPageUptodate(page);
	clear_cold_data(page);
	return 0;
}

974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989
static int f2fs_write_end(struct file *file,
			struct address_space *mapping,
			loff_t pos, unsigned len, unsigned copied,
			struct page *page, void *fsdata)
{
	struct inode *inode = page->mapping->host;

	SetPageUptodate(page);
	set_page_dirty(page);

	if (pos + copied > i_size_read(inode)) {
		i_size_write(inode, pos + copied);
		mark_inode_dirty(inode);
		update_inode_page(inode);
	}

990
	f2fs_put_page(page, 1);
991 992 993
	return copied;
}

994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011
static int check_direct_IO(struct inode *inode, int rw,
		const struct iovec *iov, loff_t offset, unsigned long nr_segs)
{
	unsigned blocksize_mask = inode->i_sb->s_blocksize - 1;
	int i;

	if (rw == READ)
		return 0;

	if (offset & blocksize_mask)
		return -EINVAL;

	for (i = 0; i < nr_segs; i++)
		if (iov[i].iov_len & blocksize_mask)
			return -EINVAL;
	return 0;
}

1012 1013 1014 1015 1016
static ssize_t f2fs_direct_IO(int rw, struct kiocb *iocb,
		const struct iovec *iov, loff_t offset, unsigned long nr_segs)
{
	struct file *file = iocb->ki_filp;
	struct inode *inode = file->f_mapping->host;
1017

H
Huajun Li 已提交
1018 1019 1020 1021
	/* Let buffer I/O handle the inline data case. */
	if (f2fs_has_inline_data(inode))
		return 0;

1022 1023 1024
	if (check_direct_IO(inode, rw, iov, offset, nr_segs))
		return 0;

1025
	return blockdev_direct_IO(rw, iocb, inode, iov, offset, nr_segs,
1026
							get_data_block);
1027 1028
}

1029 1030
static void f2fs_invalidate_data_page(struct page *page, unsigned int offset,
				      unsigned int length)
1031 1032
{
	struct inode *inode = page->mapping->host;
1033
	if (PageDirty(page))
1034 1035 1036 1037 1038 1039 1040
		inode_dec_dirty_dents(inode);
	ClearPagePrivate(page);
}

static int f2fs_release_data_page(struct page *page, gfp_t wait)
{
	ClearPagePrivate(page);
1041
	return 1;
1042 1043 1044 1045 1046 1047 1048
}

static int f2fs_set_data_page_dirty(struct page *page)
{
	struct address_space *mapping = page->mapping;
	struct inode *inode = mapping->host;

1049 1050
	trace_f2fs_set_page_dirty(page, DATA);

1051
	SetPageUptodate(page);
1052 1053
	mark_inode_dirty(inode);

1054 1055 1056 1057 1058 1059 1060 1061
	if (!PageDirty(page)) {
		__set_page_dirty_nobuffers(page);
		set_dirty_dir_page(inode, page);
		return 1;
	}
	return 0;
}

J
Jaegeuk Kim 已提交
1062 1063
static sector_t f2fs_bmap(struct address_space *mapping, sector_t block)
{
1064
	return generic_block_bmap(mapping, block, get_data_block);
J
Jaegeuk Kim 已提交
1065 1066
}

1067 1068 1069 1070 1071 1072
const struct address_space_operations f2fs_dblock_aops = {
	.readpage	= f2fs_read_data_page,
	.readpages	= f2fs_read_data_pages,
	.writepage	= f2fs_write_data_page,
	.writepages	= f2fs_write_data_pages,
	.write_begin	= f2fs_write_begin,
1073
	.write_end	= f2fs_write_end,
1074 1075 1076 1077
	.set_page_dirty	= f2fs_set_data_page_dirty,
	.invalidatepage	= f2fs_invalidate_data_page,
	.releasepage	= f2fs_release_data_page,
	.direct_IO	= f2fs_direct_IO,
J
Jaegeuk Kim 已提交
1078
	.bmap		= f2fs_bmap,
1079
};