data.c 28 KB
Newer Older
Jaegeuk Kim's avatar
Jaegeuk Kim committed
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"
Jaegeuk Kim's avatar
Jaegeuk Kim committed
25
#include "trace.h"
26
#include <trace/events/f2fs.h>
27

28 29
static void f2fs_read_end_io(struct bio *bio, int err)
{
30 31
	struct bio_vec *bvec;
	int i;
32

33
	bio_for_each_segment_all(bvec, bio, i) {
34 35
		struct page *page = bvec->bv_page;

36 37 38
		if (!err) {
			SetPageUptodate(page);
		} else {
39 40 41 42
			ClearPageUptodate(page);
			SetPageError(page);
		}
		unlock_page(page);
43
	}
44 45 46 47 48
	bio_put(bio);
}

static void f2fs_write_end_io(struct bio *bio, int err)
{
49
	struct f2fs_sb_info *sbi = bio->bi_private;
50 51
	struct bio_vec *bvec;
	int i;
52

53
	bio_for_each_segment_all(bvec, bio, i) {
54 55
		struct page *page = bvec->bv_page;

56
		if (unlikely(err)) {
57
			set_page_dirty(page);
58
			set_bit(AS_EIO, &page->mapping->flags);
59
			f2fs_stop_checkpoint(sbi);
60 61 62
		}
		end_page_writeback(page);
		dec_page_count(sbi, F2FS_WRITEBACK);
63
	}
64 65 66 67 68 69 70 71

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

	bio_put(bio);
}

72 73 74 75 76 77 78 79 80 81 82 83
/*
 * 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;
Chao Yu's avatar
Chao Yu committed
84
	bio->bi_iter.bi_sector = SECTOR_FROM_BLOCK(blk_addr);
85
	bio->bi_end_io = is_read ? f2fs_read_end_io : f2fs_write_end_io;
86
	bio->bi_private = sbi;
87 88 89 90

	return bio;
}

Jaegeuk Kim's avatar
Jaegeuk Kim committed
91
static void __submit_merged_bio(struct f2fs_bio_info *io)
92
{
Jaegeuk Kim's avatar
Jaegeuk Kim committed
93
	struct f2fs_io_info *fio = &io->fio;
94 95 96 97

	if (!io->bio)
		return;

98 99 100 101 102 103
	if (is_read_io(fio->rw))
		trace_f2fs_submit_read_bio(io->sbi->sb, fio->rw,
							fio->type, io->bio);
	else
		trace_f2fs_submit_write_bio(io->sbi->sb, fio->rw,
							fio->type, io->bio);
104

105
	submit_bio(fio->rw, io->bio);
106 107 108 109
	io->bio = NULL;
}

void f2fs_submit_merged_bio(struct f2fs_sb_info *sbi,
Jaegeuk Kim's avatar
Jaegeuk Kim committed
110
				enum page_type type, int rw)
111 112 113 114 115 116
{
	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];

117
	down_write(&io->io_rwsem);
Jaegeuk Kim's avatar
Jaegeuk Kim committed
118 119 120 121

	/* change META to META_FLUSH in the checkpoint procedure */
	if (type >= META_FLUSH) {
		io->fio.type = META_FLUSH;
Jaegeuk Kim's avatar
Jaegeuk Kim committed
122 123 124 125
		if (test_opt(sbi, NOBARRIER))
			io->fio.rw = WRITE_FLUSH | REQ_META | REQ_PRIO;
		else
			io->fio.rw = WRITE_FLUSH_FUA | REQ_META | REQ_PRIO;
Jaegeuk Kim's avatar
Jaegeuk Kim committed
126 127
	}
	__submit_merged_bio(io);
128
	up_write(&io->io_rwsem);
129 130 131 132 133 134 135
}

/*
 * 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,
136
					struct f2fs_io_info *fio)
137 138 139
{
	struct bio *bio;

140
	trace_f2fs_submit_page_bio(page, fio->blk_addr, fio->rw);
Jaegeuk Kim's avatar
Jaegeuk Kim committed
141
	f2fs_trace_ios(page, fio, 0);
142 143

	/* Allocate a new bio */
144
	bio = __bio_alloc(sbi, fio->blk_addr, 1, is_read_io(fio->rw));
145 146 147 148 149 150 151

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

152
	submit_bio(fio->rw, bio);
153 154 155 156
	return 0;
}

void f2fs_submit_page_mbio(struct f2fs_sb_info *sbi, struct page *page,
157
					struct f2fs_io_info *fio)
158
{
Jaegeuk Kim's avatar
Jaegeuk Kim committed
159
	enum page_type btype = PAGE_TYPE_OF_BIO(fio->type);
160
	struct f2fs_bio_info *io;
161
	bool is_read = is_read_io(fio->rw);
162

163
	io = is_read ? &sbi->read_io : &sbi->write_io[btype];
164

165
	verify_block_addr(sbi, fio->blk_addr);
166

167
	down_write(&io->io_rwsem);
168

169
	if (!is_read)
170 171
		inc_page_count(sbi, F2FS_WRITEBACK);

172
	if (io->bio && (io->last_block_in_bio != fio->blk_addr - 1 ||
Jaegeuk Kim's avatar
Jaegeuk Kim committed
173 174
						io->fio.rw != fio->rw))
		__submit_merged_bio(io);
175 176
alloc_new:
	if (io->bio == NULL) {
Jaegeuk Kim's avatar
Jaegeuk Kim committed
177
		int bio_blocks = MAX_BIO_BLOCKS(sbi);
178

179
		io->bio = __bio_alloc(sbi, fio->blk_addr, bio_blocks, is_read);
Jaegeuk Kim's avatar
Jaegeuk Kim committed
180
		io->fio = *fio;
181 182 183 184
	}

	if (bio_add_page(io->bio, page, PAGE_CACHE_SIZE, 0) <
							PAGE_CACHE_SIZE) {
Jaegeuk Kim's avatar
Jaegeuk Kim committed
185
		__submit_merged_bio(io);
186 187 188
		goto alloc_new;
	}

189
	io->last_block_in_bio = fio->blk_addr;
Jaegeuk Kim's avatar
Jaegeuk Kim committed
190
	f2fs_trace_ios(page, fio, 0);
191

192
	up_write(&io->io_rwsem);
193
	trace_f2fs_submit_page_mbio(page, fio->rw, fio->type, fio->blk_addr);
194 195
}

Jaegeuk Kim's avatar
Jaegeuk Kim committed
196
/*
197 198 199 200 201 202 203 204 205 206 207 208
 * 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;

209
	f2fs_wait_on_page_writeback(node_page, NODE);
210

211
	rn = F2FS_NODE(node_page);
212 213 214 215 216 217 218 219 220

	/* 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)
{
221
	struct f2fs_sb_info *sbi = F2FS_I_SB(dn->inode);
222

223
	if (unlikely(is_inode_flag_set(F2FS_I(dn->inode), FI_NO_ALLOC)))
224
		return -EPERM;
225
	if (unlikely(!inc_valid_block_count(sbi, dn->inode, 1)))
226 227
		return -ENOSPC;

228 229
	trace_f2fs_reserve_new_block(dn->inode, dn->nid, dn->ofs_in_node);

230 231
	__set_data_blkaddr(dn, NEW_ADDR);
	dn->data_blkaddr = NEW_ADDR;
232
	mark_inode_dirty(dn->inode);
233 234 235 236
	sync_inode_page(dn);
	return 0;
}

237 238 239 240 241 242 243 244
int f2fs_reserve_block(struct dnode_of_data *dn, pgoff_t index)
{
	bool need_put = dn->inode_page ? false : true;
	int err;

	err = get_dnode_of_data(dn, index, ALLOC_NODE);
	if (err)
		return err;
245

246 247
	if (dn->data_blkaddr == NULL_ADDR)
		err = reserve_new_block(dn);
248
	if (err || need_put)
249 250 251 252
		f2fs_put_dnode(dn);
	return err;
}

253 254 255 256 257 258 259
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;

260 261 262
	if (is_inode_flag_set(fi, FI_NO_EXTENT))
		return 0;

263 264 265 266 267 268
	read_lock(&fi->ext.ext_lock);
	if (fi->ext.len == 0) {
		read_unlock(&fi->ext.ext_lock);
		return 0;
	}

269 270
	stat_inc_total_hit(inode->i_sb);

271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287
	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;

288
		stat_inc_read_hit(inode->i_sb);
289 290 291 292 293 294 295 296 297 298 299 300
		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;
301
	int need_update = true;
302

303
	f2fs_bug_on(F2FS_I_SB(dn->inode), blk_addr == NEW_ADDR);
304 305
	fofs = start_bidx_of_node(ofs_of_node(dn->node_page), fi) +
							dn->ofs_in_node;
306 307 308 309

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

310 311 312
	if (is_inode_flag_set(fi, FI_NO_EXTENT))
		return;

313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333
	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;
	}

Namjae Jeon's avatar
Namjae Jeon committed
334
	/* Front merge */
335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358
	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;
		}
359 360
	} else {
		need_update = false;
361 362
	}

363 364 365 366 367 368
	/* 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;
	}
369 370
end_update:
	write_unlock(&fi->ext.ext_lock);
371 372 373
	if (need_update)
		sync_inode_page(dn);
	return;
374 375
}

376
struct page *find_data_page(struct inode *inode, pgoff_t index, bool sync)
377 378 379 380 381
{
	struct address_space *mapping = inode->i_mapping;
	struct dnode_of_data dn;
	struct page *page;
	int err;
382 383 384 385
	struct f2fs_io_info fio = {
		.type = DATA,
		.rw = sync ? READ_SYNC : READA,
	};
386 387 388 389 390 391 392

	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);
393
	err = get_dnode_of_data(&dn, index, LOOKUP_NODE);
394 395 396 397 398 399 400 401
	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 */
402
	if (unlikely(dn.data_blkaddr == NEW_ADDR))
403 404
		return ERR_PTR(-EINVAL);

405
	page = grab_cache_page(mapping, index);
406 407 408
	if (!page)
		return ERR_PTR(-ENOMEM);

409 410 411 412 413
	if (PageUptodate(page)) {
		unlock_page(page);
		return page;
	}

414 415
	fio.blk_addr = dn.data_blkaddr;
	err = f2fs_submit_page_bio(F2FS_I_SB(inode), page, &fio);
416 417 418
	if (err)
		return ERR_PTR(err);

419 420
	if (sync) {
		wait_on_page_locked(page);
421
		if (unlikely(!PageUptodate(page))) {
422 423 424
			f2fs_put_page(page, 0);
			return ERR_PTR(-EIO);
		}
425 426 427 428
	}
	return page;
}

Jaegeuk Kim's avatar
Jaegeuk Kim committed
429
/*
430 431 432 433 434 435 436 437 438 439
 * 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 address_space *mapping = inode->i_mapping;
	struct dnode_of_data dn;
	struct page *page;
	int err;
440 441 442 443
	struct f2fs_io_info fio = {
		.type = DATA,
		.rw = READ_SYNC,
	};
444
repeat:
445
	page = grab_cache_page(mapping, index);
446 447 448
	if (!page)
		return ERR_PTR(-ENOMEM);

449
	set_new_dnode(&dn, inode, NULL, NULL, 0);
450
	err = get_dnode_of_data(&dn, index, LOOKUP_NODE);
451 452
	if (err) {
		f2fs_put_page(page, 1);
453
		return ERR_PTR(err);
454
	}
455 456
	f2fs_put_dnode(&dn);

457
	if (unlikely(dn.data_blkaddr == NULL_ADDR)) {
458
		f2fs_put_page(page, 1);
459
		return ERR_PTR(-ENOENT);
460
	}
461 462 463 464

	if (PageUptodate(page))
		return page;

Jaegeuk Kim's avatar
Jaegeuk Kim committed
465 466 467 468 469 470 471 472 473 474 475
	/*
	 * 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;
	}
476

477 478
	fio.blk_addr = dn.data_blkaddr;
	err = f2fs_submit_page_bio(F2FS_I_SB(inode), page, &fio);
479
	if (err)
480
		return ERR_PTR(err);
481 482

	lock_page(page);
483
	if (unlikely(!PageUptodate(page))) {
484 485
		f2fs_put_page(page, 1);
		return ERR_PTR(-EIO);
486
	}
487
	if (unlikely(page->mapping != mapping)) {
488 489
		f2fs_put_page(page, 1);
		goto repeat;
490 491 492 493
	}
	return page;
}

Jaegeuk Kim's avatar
Jaegeuk Kim committed
494
/*
495 496
 * Caller ensures that this data page is never allocated.
 * A new zero-filled data page is allocated in the page cache.
497
 *
Chao Yu's avatar
Chao Yu committed
498 499
 * Also, caller should grab and release a rwsem by calling f2fs_lock_op() and
 * f2fs_unlock_op().
500
 * Note that, ipage is set only by make_empty_dir.
501
 */
502
struct page *get_new_data_page(struct inode *inode,
503
		struct page *ipage, pgoff_t index, bool new_i_size)
504 505 506 507 508 509
{
	struct address_space *mapping = inode->i_mapping;
	struct page *page;
	struct dnode_of_data dn;
	int err;

510
	set_new_dnode(&dn, inode, ipage, NULL, 0);
511
	err = f2fs_reserve_block(&dn, index);
512 513
	if (err)
		return ERR_PTR(err);
514
repeat:
515
	page = grab_cache_page(mapping, index);
516 517 518 519
	if (!page) {
		err = -ENOMEM;
		goto put_err;
	}
520 521 522 523 524 525

	if (PageUptodate(page))
		return page;

	if (dn.data_blkaddr == NEW_ADDR) {
		zero_user_segment(page, 0, PAGE_CACHE_SIZE);
526
		SetPageUptodate(page);
527
	} else {
528 529 530 531 532 533
		struct f2fs_io_info fio = {
			.type = DATA,
			.rw = READ_SYNC,
			.blk_addr = dn.data_blkaddr,
		};
		err = f2fs_submit_page_bio(F2FS_I_SB(inode), page, &fio);
534
		if (err)
535 536
			goto put_err;

537
		lock_page(page);
538
		if (unlikely(!PageUptodate(page))) {
539
			f2fs_put_page(page, 1);
540 541
			err = -EIO;
			goto put_err;
542
		}
543
		if (unlikely(page->mapping != mapping)) {
544 545
			f2fs_put_page(page, 1);
			goto repeat;
546 547 548 549 550 551
		}
	}

	if (new_i_size &&
		i_size_read(inode) < ((index + 1) << PAGE_CACHE_SHIFT)) {
		i_size_write(inode, ((index + 1) << PAGE_CACHE_SHIFT));
552 553
		/* Only the directory inode sets new_i_size */
		set_inode_flag(F2FS_I(inode), FI_UPDATE_DIR);
554 555
	}
	return page;
556 557 558 559

put_err:
	f2fs_put_dnode(&dn);
	return ERR_PTR(err);
560 561
}

562 563
static int __allocate_data_block(struct dnode_of_data *dn)
{
564
	struct f2fs_sb_info *sbi = F2FS_I_SB(dn->inode);
565
	struct f2fs_inode_info *fi = F2FS_I(dn->inode);
566 567 568
	struct f2fs_summary sum;
	block_t new_blkaddr;
	struct node_info ni;
569
	pgoff_t fofs;
570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591
	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);

592 593 594 595 596 597
	/* update i_size */
	fofs = start_bidx_of_node(ofs_of_node(dn->node_page), fi) +
							dn->ofs_in_node;
	if (i_size_read(dn->inode) < ((fofs + 1) << PAGE_CACHE_SHIFT))
		i_size_write(dn->inode, ((fofs + 1) << PAGE_CACHE_SHIFT));

598 599 600 601
	dn->data_blkaddr = new_blkaddr;
	return 0;
}

Jaegeuk Kim's avatar
Jaegeuk Kim committed
602
/*
Chao Yu's avatar
Chao Yu committed
603 604 605 606 607 608
 * 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
609
 */
610 611
static int __get_data_block(struct inode *inode, sector_t iblock,
			struct buffer_head *bh_result, int create, bool fiemap)
612 613 614 615
{
	unsigned int blkbits = inode->i_sb->s_blocksize_bits;
	unsigned maxblocks = bh_result->b_size >> blkbits;
	struct dnode_of_data dn;
616 617 618 619
	int mode = create ? ALLOC_NODE : LOOKUP_NODE_RA;
	pgoff_t pgofs, end_offset;
	int err = 0, ofs = 1;
	bool allocated = false;
620 621 622 623

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

624 625 626
	if (check_extent_cache(inode, pgofs, bh_result))
		goto out;

627
	if (create) {
628 629
		f2fs_balance_fs(F2FS_I_SB(inode));
		f2fs_lock_op(F2FS_I_SB(inode));
630
	}
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
	if (dn.data_blkaddr == NEW_ADDR && !fiemap)
641
		goto put_out;
642

643 644 645 646 647 648 649 650 651 652 653 654
	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;
	}

655
	end_offset = ADDRS_PER_PAGE(dn.node_page, F2FS_I(inode));
656 657 658 659 660 661 662 663 664 665 666 667 668
	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);
669
		if (err) {
670 671 672 673
			if (err == -ENOENT)
				err = 0;
			goto unlock_out;
		}
674
		if (dn.data_blkaddr == NEW_ADDR && !fiemap)
675 676
			goto put_out;

677
		end_offset = ADDRS_PER_PAGE(dn.node_page, F2FS_I(inode));
678
	}
679

680 681 682 683 684 685 686 687 688
	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;
		}
arter97's avatar
arter97 committed
689
		/* Give more consecutive addresses for the readahead */
690 691 692 693 694 695 696
		if (blkaddr == (bh_result->b_blocknr + ofs)) {
			ofs++;
			dn.ofs_in_node++;
			pgofs++;
			bh_result->b_size += (((size_t)1) << blkbits);
			goto get_next;
		}
697
	}
698 699 700 701
sync_out:
	if (allocated)
		sync_inode_page(&dn);
put_out:
702
	f2fs_put_dnode(&dn);
703 704
unlock_out:
	if (create)
705
		f2fs_unlock_op(F2FS_I_SB(inode));
706 707 708
out:
	trace_f2fs_get_data_block(inode, iblock, bh_result, err);
	return err;
709 710
}

711 712 713 714 715 716 717 718 719 720 721 722
static int get_data_block(struct inode *inode, sector_t iblock,
			struct buffer_head *bh_result, int create)
{
	return __get_data_block(inode, iblock, bh_result, create, false);
}

static int get_data_block_fiemap(struct inode *inode, sector_t iblock,
			struct buffer_head *bh_result, int create)
{
	return __get_data_block(inode, iblock, bh_result, create, true);
}

Jaegeuk Kim's avatar
Jaegeuk Kim committed
723 724 725
int f2fs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
		u64 start, u64 len)
{
726 727
	return generic_block_fiemap(inode, fieinfo,
				start, len, get_data_block_fiemap);
Jaegeuk Kim's avatar
Jaegeuk Kim committed
728 729
}

730 731
static int f2fs_read_data_page(struct file *file, struct page *page)
{
732
	struct inode *inode = page->mapping->host;
733
	int ret = -EAGAIN;
734

735 736
	trace_f2fs_readpage(page, DATA);

arter97's avatar
arter97 committed
737
	/* If the file has inline data, try to read it directly */
738 739
	if (f2fs_has_inline_data(inode))
		ret = f2fs_read_inline_data(inode, page);
740
	if (ret == -EAGAIN)
741 742 743
		ret = mpage_readpage(page, get_data_block);

	return ret;
744 745 746 747 748 749
}

static int f2fs_read_data_pages(struct file *file,
			struct address_space *mapping,
			struct list_head *pages, unsigned nr_pages)
{
750 751 752 753 754 755
	struct inode *inode = file->f_mapping->host;

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

756
	return mpage_readpages(mapping, pages, nr_pages, get_data_block);
757 758
}

Jaegeuk Kim's avatar
Jaegeuk Kim committed
759
int do_write_data_page(struct page *page, struct f2fs_io_info *fio)
760 761 762 763 764 765
{
	struct inode *inode = page->mapping->host;
	struct dnode_of_data dn;
	int err = 0;

	set_new_dnode(&dn, inode, NULL, NULL, 0);
766
	err = get_dnode_of_data(&dn, page->index, LOOKUP_NODE);
767 768 769
	if (err)
		return err;

770
	fio->blk_addr = dn.data_blkaddr;
771 772

	/* This page is already truncated */
773
	if (fio->blk_addr == NULL_ADDR)
774 775 776 777 778 779 780 781
		goto out_writepage;

	set_page_writeback(page);

	/*
	 * If current allocation needs SSR,
	 * it had better in-place writes for updated data.
	 */
782
	if (unlikely(fio->blk_addr != NEW_ADDR &&
783 784
			!is_cold_data(page) &&
			need_inplace_update(inode))) {
785
		rewrite_data_page(page, fio);
786
		set_inode_flag(F2FS_I(inode), FI_UPDATE_WRITE);
787
	} else {
788 789
		write_data_page(page, &dn, fio);
		update_extent_cache(fio->blk_addr, &dn);
790
		set_inode_flag(F2FS_I(inode), FI_APPEND_WRITE);
791 792 793 794 795 796 797 798 799 800
	}
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;
801
	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
802 803 804
	loff_t i_size = i_size_read(inode);
	const pgoff_t end_index = ((unsigned long long) i_size)
							>> PAGE_CACHE_SHIFT;
805
	unsigned offset = 0;
806
	bool need_balance_fs = false;
807
	int err = 0;
Jaegeuk Kim's avatar
Jaegeuk Kim committed
808 809
	struct f2fs_io_info fio = {
		.type = DATA,
Chris Fries's avatar
Chris Fries committed
810
		.rw = (wbc->sync_mode == WB_SYNC_ALL) ? WRITE_SYNC : WRITE,
Jaegeuk Kim's avatar
Jaegeuk Kim committed
811
	};
812

813 814
	trace_f2fs_writepage(page, DATA);

815
	if (page->index < end_index)
816
		goto write;
817 818 819 820 821 822

	/*
	 * 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);
823
	if ((page->index >= end_index + 1) || !offset)
824
		goto out;
825 826

	zero_user_segment(page, offset, PAGE_CACHE_SIZE);
827
write:
828
	if (unlikely(sbi->por_doing))
829
		goto redirty_out;
830 831 832 833 834
	if (f2fs_is_drop_cache(inode))
		goto out;
	if (f2fs_is_volatile_file(inode) && !wbc->for_reclaim &&
			available_free_memory(sbi, BASE_CHECK))
		goto redirty_out;
835

836
	/* Dentry blocks are controlled by checkpoint */
837
	if (S_ISDIR(inode->i_mode)) {
838 839
		if (unlikely(f2fs_cp_error(sbi)))
			goto redirty_out;
Jaegeuk Kim's avatar
Jaegeuk Kim committed
840
		err = do_write_data_page(page, &fio);
841 842
		goto done;
	}
843

844 845 846 847
	/* we should bypass data pages to proceed the kworkder jobs */
	if (unlikely(f2fs_cp_error(sbi))) {
		SetPageError(page);
		unlock_page(page);
848
		goto out;
849 850
	}

851
	if (!wbc->for_reclaim)
852
		need_balance_fs = true;
853
	else if (has_not_enough_free_secs(sbi, 0))
854
		goto redirty_out;
855

856
	err = -EAGAIN;
857
	f2fs_lock_op(sbi);
858 859 860
	if (f2fs_has_inline_data(inode))
		err = f2fs_write_inline_data(inode, page);
	if (err == -EAGAIN)
861 862 863 864 865
		err = do_write_data_page(page, &fio);
	f2fs_unlock_op(sbi);
done:
	if (err && err != -ENOENT)
		goto redirty_out;
866 867

	clear_cold_data(page);
868
out:
869
	inode_dec_dirty_pages(inode);
870
	unlock_page(page);
871
	if (need_balance_fs)
872
		f2fs_balance_fs(sbi);
873 874
	if (wbc->for_reclaim)
		f2fs_submit_merged_bio(sbi, DATA, WRITE);
875 876 877
	return 0;

redirty_out:
878
	redirty_page_for_writepage(wbc, page);
879
	return AOP_WRITEPAGE_ACTIVATE;
880 881
}

882 883 884 885 886 887 888 889 890
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;
}

891
static int f2fs_write_data_pages(struct address_space *mapping,
892 893 894
			    struct writeback_control *wbc)
{
	struct inode *inode = mapping->host;
895
	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
896
	bool locked = false;
897
	int ret;
898
	long diff;
899

900 901
	trace_f2fs_writepages(mapping->host, wbc, DATA);

P J P's avatar
P J P committed
902 903 904 905
	/* deal with chardevs and other special file */
	if (!mapping->a_ops->writepage)
		return 0;

906
	if (S_ISDIR(inode->i_mode) && wbc->sync_mode == WB_SYNC_NONE &&
907
			get_dirty_pages(inode) < nr_pages_to_skip(sbi, DATA) &&
908
			available_free_memory(sbi, DIRTY_DENTS))
909
		goto skip_write;
910

911
	diff = nr_pages_to_write(sbi, DATA, wbc);
912

913
	if (!S_ISDIR(inode->i_mode)) {
914
		mutex_lock(&sbi->writepages);
915 916
		locked = true;
	}
917
	ret = write_cache_pages(mapping, wbc, __f2fs_writepage, mapping);
918
	if (locked)
919
		mutex_unlock(&sbi->writepages);
Jaegeuk Kim's avatar
Jaegeuk Kim committed
920 921

	f2fs_submit_merged_bio(sbi, DATA, WRITE);
922 923 924

	remove_dirty_dir_inode(inode);

925
	wbc->nr_to_write = max((long)0, wbc->nr_to_write - diff);