data.c 39.2 KB
Newer Older
J
Jaegeuk Kim 已提交
1
/*
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
 * 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>
#include <linux/writeback.h>
#include <linux/backing-dev.h>
C
Chao Yu 已提交
17
#include <linux/pagevec.h>
18 19
#include <linux/blkdev.h>
#include <linux/bio.h>
20
#include <linux/prefetch.h>
21
#include <linux/uio.h>
J
Jaegeuk Kim 已提交
22
#include <linux/cleancache.h>
23 24 25 26

#include "f2fs.h"
#include "node.h"
#include "segment.h"
J
Jaegeuk Kim 已提交
27
#include "trace.h"
28
#include <trace/events/f2fs.h>
29

30
static void f2fs_read_end_io(struct bio *bio)
31
{
32 33
	struct bio_vec *bvec;
	int i;
34

35
	if (f2fs_bio_encrypted(bio)) {
36
		if (bio->bi_error) {
37 38 39 40 41 42 43
			f2fs_release_crypto_ctx(bio->bi_private);
		} else {
			f2fs_end_io_crypto_work(bio->bi_private, bio);
			return;
		}
	}

44 45
	bio_for_each_segment_all(bvec, bio, i) {
		struct page *page = bvec->bv_page;
J
Jaegeuk Kim 已提交
46

47
		if (!bio->bi_error) {
J
Jaegeuk Kim 已提交
48 49 50 51 52 53 54 55 56 57
			SetPageUptodate(page);
		} else {
			ClearPageUptodate(page);
			SetPageError(page);
		}
		unlock_page(page);
	}
	bio_put(bio);
}

58
static void f2fs_write_end_io(struct bio *bio)
59
{
60
	struct f2fs_sb_info *sbi = bio->bi_private;
61 62
	struct bio_vec *bvec;
	int i;
63

64
	bio_for_each_segment_all(bvec, bio, i) {
65 66
		struct page *page = bvec->bv_page;

67 68
		f2fs_restore_and_release_control_page(&page);

69
		if (unlikely(bio->bi_error)) {
70
			set_page_dirty(page);
71
			set_bit(AS_EIO, &page->mapping->flags);
72
			f2fs_stop_checkpoint(sbi);
73 74 75
		}
		end_page_writeback(page);
		dec_page_count(sbi, F2FS_WRITEBACK);
76
	}
77 78 79 80 81 82 83 84

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

	bio_put(bio);
}

85 86 87 88 89 90 91 92
/*
 * 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;

J
Jaegeuk Kim 已提交
93
	bio = f2fs_bio_alloc(npages);
94 95

	bio->bi_bdev = sbi->sb->s_bdev;
C
Chao Yu 已提交
96
	bio->bi_iter.bi_sector = SECTOR_FROM_BLOCK(blk_addr);
97
	bio->bi_end_io = is_read ? f2fs_read_end_io : f2fs_write_end_io;
98
	bio->bi_private = is_read ? NULL : sbi;
99 100 101 102

	return bio;
}

J
Jaegeuk Kim 已提交
103
static void __submit_merged_bio(struct f2fs_bio_info *io)
104
{
J
Jaegeuk Kim 已提交
105
	struct f2fs_io_info *fio = &io->fio;
106 107 108 109

	if (!io->bio)
		return;

110
	if (is_read_io(fio->rw))
111
		trace_f2fs_submit_read_bio(io->sbi->sb, fio, io->bio);
112
	else
113
		trace_f2fs_submit_write_bio(io->sbi->sb, fio, io->bio);
114

115
	submit_bio(fio->rw, io->bio);
116 117 118 119
	io->bio = NULL;
}

void f2fs_submit_merged_bio(struct f2fs_sb_info *sbi,
J
Jaegeuk Kim 已提交
120
				enum page_type type, int rw)
121 122 123 124 125 126
{
	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];

127
	down_write(&io->io_rwsem);
J
Jaegeuk Kim 已提交
128 129 130 131

	/* change META to META_FLUSH in the checkpoint procedure */
	if (type >= META_FLUSH) {
		io->fio.type = META_FLUSH;
J
Jaegeuk Kim 已提交
132 133 134 135
		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;
J
Jaegeuk Kim 已提交
136 137
	}
	__submit_merged_bio(io);
138
	up_write(&io->io_rwsem);
139 140 141 142 143 144
}

/*
 * Fill the locked page with data located in the block address.
 * Return unlocked page.
 */
145
int f2fs_submit_page_bio(struct f2fs_io_info *fio)
146 147
{
	struct bio *bio;
148
	struct page *page = fio->encrypted_page ? fio->encrypted_page : fio->page;
149

150
	trace_f2fs_submit_page_bio(page, fio);
151
	f2fs_trace_ios(fio, 0);
152 153

	/* Allocate a new bio */
154
	bio = __bio_alloc(fio->sbi, fio->blk_addr, 1, is_read_io(fio->rw));
155 156 157 158 159 160

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

161
	submit_bio(fio->rw, bio);
162 163 164
	return 0;
}

165
void f2fs_submit_page_mbio(struct f2fs_io_info *fio)
166
{
167
	struct f2fs_sb_info *sbi = fio->sbi;
J
Jaegeuk Kim 已提交
168
	enum page_type btype = PAGE_TYPE_OF_BIO(fio->type);
169
	struct f2fs_bio_info *io;
170
	bool is_read = is_read_io(fio->rw);
171
	struct page *bio_page;
172

173
	io = is_read ? &sbi->read_io : &sbi->write_io[btype];
174

175
	verify_block_addr(sbi, fio->blk_addr);
176

177
	down_write(&io->io_rwsem);
178

179
	if (!is_read)
180 181
		inc_page_count(sbi, F2FS_WRITEBACK);

182
	if (io->bio && (io->last_block_in_bio != fio->blk_addr - 1 ||
J
Jaegeuk Kim 已提交
183 184
						io->fio.rw != fio->rw))
		__submit_merged_bio(io);
185 186
alloc_new:
	if (io->bio == NULL) {
J
Jaegeuk Kim 已提交
187
		int bio_blocks = MAX_BIO_BLOCKS(sbi);
188

189
		io->bio = __bio_alloc(sbi, fio->blk_addr, bio_blocks, is_read);
J
Jaegeuk Kim 已提交
190
		io->fio = *fio;
191 192
	}

193 194 195
	bio_page = fio->encrypted_page ? fio->encrypted_page : fio->page;

	if (bio_add_page(io->bio, bio_page, PAGE_CACHE_SIZE, 0) <
196
							PAGE_CACHE_SIZE) {
J
Jaegeuk Kim 已提交
197
		__submit_merged_bio(io);
198 199 200
		goto alloc_new;
	}

201
	io->last_block_in_bio = fio->blk_addr;
202
	f2fs_trace_ios(fio, 0);
203

204
	up_write(&io->io_rwsem);
205
	trace_f2fs_submit_page_mbio(fio->page, fio);
206 207
}

J
Jaegeuk Kim 已提交
208
/*
209 210 211 212 213
 * Lock ordering for the change of data block address:
 * ->data_page
 *  ->node_page
 *    update block addresses in the node page
 */
214
void set_data_blkaddr(struct dnode_of_data *dn)
215 216 217 218 219 220
{
	struct f2fs_node *rn;
	__le32 *addr_array;
	struct page *node_page = dn->node_page;
	unsigned int ofs_in_node = dn->ofs_in_node;

221
	f2fs_wait_on_page_writeback(node_page, NODE);
222

223
	rn = F2FS_NODE(node_page);
224 225 226

	/* Get physical address of data block */
	addr_array = blkaddr_in_node(rn);
J
Jaegeuk Kim 已提交
227
	addr_array[ofs_in_node] = cpu_to_le32(dn->data_blkaddr);
228 229 230 231 232
	set_page_dirty(node_page);
}

int reserve_new_block(struct dnode_of_data *dn)
{
233
	struct f2fs_sb_info *sbi = F2FS_I_SB(dn->inode);
234

235
	if (unlikely(is_inode_flag_set(F2FS_I(dn->inode), FI_NO_ALLOC)))
236
		return -EPERM;
237
	if (unlikely(!inc_valid_block_count(sbi, dn->inode, 1)))
238 239
		return -ENOSPC;

240 241
	trace_f2fs_reserve_new_block(dn->inode, dn->nid, dn->ofs_in_node);

242
	dn->data_blkaddr = NEW_ADDR;
243
	set_data_blkaddr(dn);
244
	mark_inode_dirty(dn->inode);
245 246 247 248
	sync_inode_page(dn);
	return 0;
}

249 250 251 252 253 254 255 256
int f2fs_reserve_block(struct dnode_of_data *dn, pgoff_t index)
{
	bool need_put = dn->inode_page ? false : true;
	int err;

	err = get_dnode_of_data(dn, index, ALLOC_NODE);
	if (err)
		return err;
257

258 259
	if (dn->data_blkaddr == NULL_ADDR)
		err = reserve_new_block(dn);
260
	if (err || need_put)
261 262 263 264
		f2fs_put_dnode(dn);
	return err;
}

265
int f2fs_get_block(struct dnode_of_data *dn, pgoff_t index)
266
{
267
	struct extent_info ei;
268
	struct inode *inode = dn->inode;
269

270 271 272
	if (f2fs_lookup_extent_cache(inode, index, &ei)) {
		dn->data_blkaddr = ei.blk + index - ei.fofs;
		return 0;
273
	}
274

275
	return f2fs_reserve_block(dn, index);
276 277
}

278 279
struct page *get_read_data_page(struct inode *inode, pgoff_t index,
						int rw, bool for_write)
280 281 282 283
{
	struct address_space *mapping = inode->i_mapping;
	struct dnode_of_data dn;
	struct page *page;
C
Chao Yu 已提交
284
	struct extent_info ei;
285
	int err;
286
	struct f2fs_io_info fio = {
287
		.sbi = F2FS_I_SB(inode),
288
		.type = DATA,
289
		.rw = rw,
290
		.encrypted_page = NULL,
291
	};
292

293 294 295
	if (f2fs_encrypted_inode(inode) && S_ISREG(inode->i_mode))
		return read_mapping_page(mapping, index, NULL);

296
	page = f2fs_grab_cache_page(mapping, index, for_write);
297 298 299
	if (!page)
		return ERR_PTR(-ENOMEM);

C
Chao Yu 已提交
300 301 302 303 304
	if (f2fs_lookup_extent_cache(inode, index, &ei)) {
		dn.data_blkaddr = ei.blk + index - ei.fofs;
		goto got_it;
	}

305
	set_new_dnode(&dn, inode, NULL, NULL, 0);
306
	err = get_dnode_of_data(&dn, index, LOOKUP_NODE);
307 308
	if (err)
		goto put_err;
309 310
	f2fs_put_dnode(&dn);

311
	if (unlikely(dn.data_blkaddr == NULL_ADDR)) {
312 313
		err = -ENOENT;
		goto put_err;
314
	}
C
Chao Yu 已提交
315
got_it:
316 317
	if (PageUptodate(page)) {
		unlock_page(page);
318
		return page;
319
	}
320

J
Jaegeuk Kim 已提交
321 322 323 324 325 326 327 328 329
	/*
	 * 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);
330
		unlock_page(page);
J
Jaegeuk Kim 已提交
331 332
		return page;
	}
333

334
	fio.blk_addr = dn.data_blkaddr;
335 336
	fio.page = page;
	err = f2fs_submit_page_bio(&fio);
337
	if (err)
338
		goto put_err;
339
	return page;
340 341 342 343

put_err:
	f2fs_put_page(page, 1);
	return ERR_PTR(err);
344 345 346 347 348 349 350 351 352 353 354 355
}

struct page *find_data_page(struct inode *inode, pgoff_t index)
{
	struct address_space *mapping = inode->i_mapping;
	struct page *page;

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

356
	page = get_read_data_page(inode, index, READ_SYNC, false);
357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375
	if (IS_ERR(page))
		return page;

	if (PageUptodate(page))
		return page;

	wait_on_page_locked(page);
	if (unlikely(!PageUptodate(page))) {
		f2fs_put_page(page, 0);
		return ERR_PTR(-EIO);
	}
	return page;
}

/*
 * 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.
 */
376 377
struct page *get_lock_data_page(struct inode *inode, pgoff_t index,
							bool for_write)
378 379 380 381
{
	struct address_space *mapping = inode->i_mapping;
	struct page *page;
repeat:
382
	page = get_read_data_page(inode, index, READ_SYNC, for_write);
383 384
	if (IS_ERR(page))
		return page;
385

386
	/* wait for read completion */
387
	lock_page(page);
388
	if (unlikely(!PageUptodate(page))) {
389 390
		f2fs_put_page(page, 1);
		return ERR_PTR(-EIO);
391
	}
392
	if (unlikely(page->mapping != mapping)) {
393 394
		f2fs_put_page(page, 1);
		goto repeat;
395 396 397 398
	}
	return page;
}

J
Jaegeuk Kim 已提交
399
/*
400 401
 * Caller ensures that this data page is never allocated.
 * A new zero-filled data page is allocated in the page cache.
402
 *
C
Chao Yu 已提交
403 404
 * Also, caller should grab and release a rwsem by calling f2fs_lock_op() and
 * f2fs_unlock_op().
405 406
 * Note that, ipage is set only by make_empty_dir, and if any error occur,
 * ipage should be released by this function.
407
 */
408
struct page *get_new_data_page(struct inode *inode,
409
		struct page *ipage, pgoff_t index, bool new_i_size)
410 411 412 413 414
{
	struct address_space *mapping = inode->i_mapping;
	struct page *page;
	struct dnode_of_data dn;
	int err;
415
repeat:
416
	page = f2fs_grab_cache_page(mapping, index, true);
417 418 419 420 421 422
	if (!page) {
		/*
		 * before exiting, we should make sure ipage will be released
		 * if any error occur.
		 */
		f2fs_put_page(ipage, 1);
423
		return ERR_PTR(-ENOMEM);
424
	}
425

426
	set_new_dnode(&dn, inode, ipage, NULL, 0);
427
	err = f2fs_reserve_block(&dn, index);
428 429
	if (err) {
		f2fs_put_page(page, 1);
430
		return ERR_PTR(err);
431
	}
432 433
	if (!ipage)
		f2fs_put_dnode(&dn);
434 435

	if (PageUptodate(page))
436
		goto got_it;
437 438 439

	if (dn.data_blkaddr == NEW_ADDR) {
		zero_user_segment(page, 0, PAGE_CACHE_SIZE);
440
		SetPageUptodate(page);
441
	} else {
442
		f2fs_put_page(page, 1);
443

444
		page = get_read_data_page(inode, index, READ_SYNC, true);
445
		if (IS_ERR(page))
446
			goto repeat;
447 448 449

		/* wait for read completion */
		lock_page(page);
450
	}
451
got_it:
C
Chao Yu 已提交
452 453 454
	if (new_i_size && i_size_read(inode) <
				((loff_t)(index + 1) << PAGE_CACHE_SHIFT)) {
		i_size_write(inode, ((loff_t)(index + 1) << PAGE_CACHE_SHIFT));
455 456
		/* Only the directory inode sets new_i_size */
		set_inode_flag(F2FS_I(inode), FI_UPDATE_DIR);
457 458 459 460
	}
	return page;
}

461 462
static int __allocate_data_block(struct dnode_of_data *dn)
{
463
	struct f2fs_sb_info *sbi = F2FS_I_SB(dn->inode);
464
	struct f2fs_inode_info *fi = F2FS_I(dn->inode);
465 466
	struct f2fs_summary sum;
	struct node_info ni;
467
	int seg = CURSEG_WARM_DATA;
468
	pgoff_t fofs;
469 470 471

	if (unlikely(is_inode_flag_set(F2FS_I(dn->inode), FI_NO_ALLOC)))
		return -EPERM;
472 473 474 475 476

	dn->data_blkaddr = datablock_addr(dn->node_page, dn->ofs_in_node);
	if (dn->data_blkaddr == NEW_ADDR)
		goto alloc;

477 478 479
	if (unlikely(!inc_valid_block_count(sbi, dn->inode, 1)))
		return -ENOSPC;

480
alloc:
481 482 483
	get_node_info(sbi, dn->nid, &ni);
	set_summary(&sum, dn->nid, dn->ofs_in_node, ni.version);

484 485 486
	if (dn->ofs_in_node == 0 && dn->inode_page == dn->node_page)
		seg = CURSEG_DIRECT_IO;

487 488
	allocate_data_block(sbi, NULL, dn->data_blkaddr, &dn->data_blkaddr,
								&sum, seg);
489
	set_data_blkaddr(dn);
490

491 492 493
	/* update i_size */
	fofs = start_bidx_of_node(ofs_of_node(dn->node_page), fi) +
							dn->ofs_in_node;
C
Chao Yu 已提交
494 495 496
	if (i_size_read(dn->inode) < ((loff_t)(fofs + 1) << PAGE_CACHE_SHIFT))
		i_size_write(dn->inode,
				((loff_t)(fofs + 1) << PAGE_CACHE_SHIFT));
497

J
Jaegeuk Kim 已提交
498
	/* direct IO doesn't use extent cache to maximize the performance */
499
	f2fs_drop_largest_extent(dn->inode, fofs);
J
Jaegeuk Kim 已提交
500

501 502 503
	return 0;
}

504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526
static void __allocate_data_blocks(struct inode *inode, loff_t offset,
							size_t count)
{
	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
	struct dnode_of_data dn;
	u64 start = F2FS_BYTES_TO_BLK(offset);
	u64 len = F2FS_BYTES_TO_BLK(count);
	bool allocated;
	u64 end_offset;

	while (len) {
		f2fs_balance_fs(sbi);
		f2fs_lock_op(sbi);

		/* When reading holes, we need its node page */
		set_new_dnode(&dn, inode, NULL, NULL, 0);
		if (get_dnode_of_data(&dn, start, ALLOC_NODE))
			goto out;

		allocated = false;
		end_offset = ADDRS_PER_PAGE(dn.node_page, F2FS_I(inode));

		while (dn.ofs_in_node < end_offset && len) {
527 528
			block_t blkaddr;

529 530 531
			if (unlikely(f2fs_cp_error(sbi)))
				goto sync_out;

532
			blkaddr = datablock_addr(dn.node_page, dn.ofs_in_node);
533
			if (blkaddr == NULL_ADDR || blkaddr == NEW_ADDR) {
534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559
				if (__allocate_data_block(&dn))
					goto sync_out;
				allocated = true;
			}
			len--;
			start++;
			dn.ofs_in_node++;
		}

		if (allocated)
			sync_inode_page(&dn);

		f2fs_put_dnode(&dn);
		f2fs_unlock_op(sbi);
	}
	return;

sync_out:
	if (allocated)
		sync_inode_page(&dn);
	f2fs_put_dnode(&dn);
out:
	f2fs_unlock_op(sbi);
	return;
}

J
Jaegeuk Kim 已提交
560
/*
J
Jaegeuk Kim 已提交
561 562
 * f2fs_map_blocks() now supported readahead/bmap/rw direct_IO with
 * f2fs_map_blocks structure.
C
Chao Yu 已提交
563 564 565 566 567
 * 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
568
 */
J
Jaegeuk Kim 已提交
569
static int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map,
C
Chao Yu 已提交
570
						int create, int flag)
571
{
J
Jaegeuk Kim 已提交
572
	unsigned int maxblocks = map->m_len;
573
	struct dnode_of_data dn;
574
	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
575 576 577
	int mode = create ? ALLOC_NODE : LOOKUP_NODE_RA;
	pgoff_t pgofs, end_offset;
	int err = 0, ofs = 1;
578
	struct extent_info ei;
579
	bool allocated = false;
580

J
Jaegeuk Kim 已提交
581 582 583 584 585
	map->m_len = 0;
	map->m_flags = 0;

	/* it only supports block size == page size */
	pgofs =	(pgoff_t)map->m_lblk;
586

587
	if (f2fs_lookup_extent_cache(inode, pgofs, &ei)) {
J
Jaegeuk Kim 已提交
588 589 590
		map->m_pblk = ei.blk + pgofs - ei.fofs;
		map->m_len = min((pgoff_t)maxblocks, ei.fofs + ei.len - pgofs);
		map->m_flags = F2FS_MAP_MAPPED;
591
		goto out;
592
	}
593

594
	if (create)
595
		f2fs_lock_op(F2FS_I_SB(inode));
596 597 598

	/* When reading holes, we need its node page */
	set_new_dnode(&dn, inode, NULL, NULL, 0);
599
	err = get_dnode_of_data(&dn, pgofs, mode);
600
	if (err) {
601 602 603
		if (err == -ENOENT)
			err = 0;
		goto unlock_out;
604
	}
C
Chao Yu 已提交
605 606 607

	if (dn.data_blkaddr == NEW_ADDR || dn.data_blkaddr == NULL_ADDR) {
		if (create) {
608 609 610 611
			if (unlikely(f2fs_cp_error(sbi))) {
				err = -EIO;
				goto put_out;
			}
C
Chao Yu 已提交
612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630
			err = __allocate_data_block(&dn);
			if (err)
				goto put_out;
			allocated = true;
			map->m_flags = F2FS_MAP_NEW;
		} else {
			if (flag != F2FS_GET_BLOCK_FIEMAP ||
						dn.data_blkaddr != NEW_ADDR) {
				if (flag == F2FS_GET_BLOCK_BMAP)
					err = -ENOENT;
				goto put_out;
			}

			/*
			 * preallocated unwritten block should be mapped
			 * for fiemap.
			 */
			if (dn.data_blkaddr == NEW_ADDR)
				map->m_flags = F2FS_MAP_UNWRITTEN;
C
Chao Yu 已提交
631 632
		}
	}
633

C
Chao Yu 已提交
634 635 636
	map->m_flags |= F2FS_MAP_MAPPED;
	map->m_pblk = dn.data_blkaddr;
	map->m_len = 1;
637

638
	end_offset = ADDRS_PER_PAGE(dn.node_page, F2FS_I(inode));
639 640 641 642 643 644 645 646 647 648 649 650
	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);
651
		if (err) {
652 653 654 655
			if (err == -ENOENT)
				err = 0;
			goto unlock_out;
		}
C
Chao Yu 已提交
656

657
		end_offset = ADDRS_PER_PAGE(dn.node_page, F2FS_I(inode));
658
	}
659

J
Jaegeuk Kim 已提交
660
	if (maxblocks > map->m_len) {
661
		block_t blkaddr = datablock_addr(dn.node_page, dn.ofs_in_node);
C
Chao Yu 已提交
662 663 664

		if (blkaddr == NEW_ADDR || blkaddr == NULL_ADDR) {
			if (create) {
665 666 667 668
				if (unlikely(f2fs_cp_error(sbi))) {
					err = -EIO;
					goto sync_out;
				}
C
Chao Yu 已提交
669 670 671 672 673 674 675 676 677 678 679 680 681 682 683
				err = __allocate_data_block(&dn);
				if (err)
					goto sync_out;
				allocated = true;
				map->m_flags |= F2FS_MAP_NEW;
				blkaddr = dn.data_blkaddr;
			} else {
				/*
				 * we only merge preallocated unwritten blocks
				 * for fiemap.
				 */
				if (flag != F2FS_GET_BLOCK_FIEMAP ||
						blkaddr != NEW_ADDR)
					goto sync_out;
			}
684
		}
C
Chao Yu 已提交
685

A
arter97 已提交
686
		/* Give more consecutive addresses for the readahead */
687 688 689 690
		if ((map->m_pblk != NEW_ADDR &&
				blkaddr == (map->m_pblk + ofs)) ||
				(map->m_pblk == NEW_ADDR &&
				blkaddr == NEW_ADDR)) {
691 692 693
			ofs++;
			dn.ofs_in_node++;
			pgofs++;
J
Jaegeuk Kim 已提交
694
			map->m_len++;
695 696
			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
out:
J
Jaegeuk Kim 已提交
707
	trace_f2fs_map_blocks(inode, map, err);
708
	return err;
709 710
}

J
Jaegeuk Kim 已提交
711
static int __get_data_block(struct inode *inode, sector_t iblock,
C
Chao Yu 已提交
712
			struct buffer_head *bh, int create, int flag)
J
Jaegeuk Kim 已提交
713 714 715 716 717 718 719
{
	struct f2fs_map_blocks map;
	int ret;

	map.m_lblk = iblock;
	map.m_len = bh->b_size >> inode->i_blkbits;

C
Chao Yu 已提交
720
	ret = f2fs_map_blocks(inode, &map, create, flag);
J
Jaegeuk Kim 已提交
721 722 723 724 725 726 727 728
	if (!ret) {
		map_bh(bh, inode->i_sb, map.m_pblk);
		bh->b_state = (bh->b_state & ~F2FS_MAP_FLAGS) | map.m_flags;
		bh->b_size = map.m_len << inode->i_blkbits;
	}
	return ret;
}

729
static int get_data_block(struct inode *inode, sector_t iblock,
C
Chao Yu 已提交
730 731 732 733 734 735
			struct buffer_head *bh_result, int create, int flag)
{
	return __get_data_block(inode, iblock, bh_result, create, flag);
}

static int get_data_block_dio(struct inode *inode, sector_t iblock,
736 737
			struct buffer_head *bh_result, int create)
{
C
Chao Yu 已提交
738 739
	return __get_data_block(inode, iblock, bh_result, create,
						F2FS_GET_BLOCK_DIO);
740 741
}

C
Chao Yu 已提交
742
static int get_data_block_bmap(struct inode *inode, sector_t iblock,
743 744
			struct buffer_head *bh_result, int create)
{
C
Chao Yu 已提交
745 746
	return __get_data_block(inode, iblock, bh_result, create,
						F2FS_GET_BLOCK_BMAP);
747 748
}

749 750 751 752 753 754 755 756 757 758
static inline sector_t logical_to_blk(struct inode *inode, loff_t offset)
{
	return (offset >> inode->i_blkbits);
}

static inline loff_t blk_to_logical(struct inode *inode, sector_t blk)
{
	return (blk << inode->i_blkbits);
}

J
Jaegeuk Kim 已提交
759 760 761
int f2fs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
		u64 start, u64 len)
{
762 763 764 765 766 767 768 769 770 771 772 773
	struct buffer_head map_bh;
	sector_t start_blk, last_blk;
	loff_t isize = i_size_read(inode);
	u64 logical = 0, phys = 0, size = 0;
	u32 flags = 0;
	bool past_eof = false, whole_file = false;
	int ret = 0;

	ret = fiemap_check_flags(fieinfo, FIEMAP_FLAG_SYNC);
	if (ret)
		return ret;

J
Jaegeuk Kim 已提交
774 775 776 777 778 779
	if (f2fs_has_inline_data(inode)) {
		ret = f2fs_inline_data_fiemap(inode, fieinfo, start, len);
		if (ret != -EAGAIN)
			return ret;
	}

780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795
	mutex_lock(&inode->i_mutex);

	if (len >= isize) {
		whole_file = true;
		len = isize;
	}

	if (logical_to_blk(inode, len) == 0)
		len = blk_to_logical(inode, 1);

	start_blk = logical_to_blk(inode, start);
	last_blk = logical_to_blk(inode, start + len - 1);
next:
	memset(&map_bh, 0, sizeof(struct buffer_head));
	map_bh.b_size = len;

C
Chao Yu 已提交
796 797
	ret = get_data_block(inode, start_blk, &map_bh, 0,
					F2FS_GET_BLOCK_FIEMAP);
798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866
	if (ret)
		goto out;

	/* HOLE */
	if (!buffer_mapped(&map_bh)) {
		start_blk++;

		if (!past_eof && blk_to_logical(inode, start_blk) >= isize)
			past_eof = 1;

		if (past_eof && size) {
			flags |= FIEMAP_EXTENT_LAST;
			ret = fiemap_fill_next_extent(fieinfo, logical,
					phys, size, flags);
		} else if (size) {
			ret = fiemap_fill_next_extent(fieinfo, logical,
					phys, size, flags);
			size = 0;
		}

		/* if we have holes up to/past EOF then we're done */
		if (start_blk > last_blk || past_eof || ret)
			goto out;
	} else {
		if (start_blk > last_blk && !whole_file) {
			ret = fiemap_fill_next_extent(fieinfo, logical,
					phys, size, flags);
			goto out;
		}

		/*
		 * if size != 0 then we know we already have an extent
		 * to add, so add it.
		 */
		if (size) {
			ret = fiemap_fill_next_extent(fieinfo, logical,
					phys, size, flags);
			if (ret)
				goto out;
		}

		logical = blk_to_logical(inode, start_blk);
		phys = blk_to_logical(inode, map_bh.b_blocknr);
		size = map_bh.b_size;
		flags = 0;
		if (buffer_unwritten(&map_bh))
			flags = FIEMAP_EXTENT_UNWRITTEN;

		start_blk += logical_to_blk(inode, size);

		/*
		 * If we are past the EOF, then we need to make sure as
		 * soon as we find a hole that the last extent we found
		 * is marked with FIEMAP_EXTENT_LAST
		 */
		if (!past_eof && logical + size >= isize)
			past_eof = true;
	}
	cond_resched();
	if (fatal_signal_pending(current))
		ret = -EINTR;
	else
		goto next;
out:
	if (ret == 1)
		ret = 0;

	mutex_unlock(&inode->i_mutex);
	return ret;
J
Jaegeuk Kim 已提交
867 868
}

J
Jaegeuk Kim 已提交
869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930
/*
 * This function was originally taken from fs/mpage.c, and customized for f2fs.
 * Major change was from block_size == page_size in f2fs by default.
 */
static int f2fs_mpage_readpages(struct address_space *mapping,
			struct list_head *pages, struct page *page,
			unsigned nr_pages)
{
	struct bio *bio = NULL;
	unsigned page_idx;
	sector_t last_block_in_bio = 0;
	struct inode *inode = mapping->host;
	const unsigned blkbits = inode->i_blkbits;
	const unsigned blocksize = 1 << blkbits;
	sector_t block_in_file;
	sector_t last_block;
	sector_t last_block_in_file;
	sector_t block_nr;
	struct block_device *bdev = inode->i_sb->s_bdev;
	struct f2fs_map_blocks map;

	map.m_pblk = 0;
	map.m_lblk = 0;
	map.m_len = 0;
	map.m_flags = 0;

	for (page_idx = 0; nr_pages; page_idx++, nr_pages--) {

		prefetchw(&page->flags);
		if (pages) {
			page = list_entry(pages->prev, struct page, lru);
			list_del(&page->lru);
			if (add_to_page_cache_lru(page, mapping,
						  page->index, GFP_KERNEL))
				goto next_page;
		}

		block_in_file = (sector_t)page->index;
		last_block = block_in_file + nr_pages;
		last_block_in_file = (i_size_read(inode) + blocksize - 1) >>
								blkbits;
		if (last_block > last_block_in_file)
			last_block = last_block_in_file;

		/*
		 * Map blocks using the previous result first.
		 */
		if ((map.m_flags & F2FS_MAP_MAPPED) &&
				block_in_file > map.m_lblk &&
				block_in_file < (map.m_lblk + map.m_len))
			goto got_it;

		/*
		 * Then do more f2fs_map_blocks() calls until we are
		 * done with this page.
		 */
		map.m_flags = 0;

		if (block_in_file < last_block) {
			map.m_lblk = block_in_file;
			map.m_len = last_block - block_in_file;

931 932
			if (f2fs_map_blocks(inode, &map, 0,
							F2FS_GET_BLOCK_READ))
J
Jaegeuk Kim 已提交
933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960
				goto set_error_page;
		}
got_it:
		if ((map.m_flags & F2FS_MAP_MAPPED)) {
			block_nr = map.m_pblk + block_in_file - map.m_lblk;
			SetPageMappedToDisk(page);

			if (!PageUptodate(page) && !cleancache_get_page(page)) {
				SetPageUptodate(page);
				goto confused;
			}
		} else {
			zero_user_segment(page, 0, PAGE_CACHE_SIZE);
			SetPageUptodate(page);
			unlock_page(page);
			goto next_page;
		}

		/*
		 * This page will go to BIO.  Do we need to send this
		 * BIO off first?
		 */
		if (bio && (last_block_in_bio != block_nr - 1)) {
submit_and_realloc:
			submit_bio(READ, bio);
			bio = NULL;
		}
		if (bio == NULL) {
961 962 963 964 965 966 967 968 969 970
			struct f2fs_crypto_ctx *ctx = NULL;

			if (f2fs_encrypted_inode(inode) &&
					S_ISREG(inode->i_mode)) {

				ctx = f2fs_get_crypto_ctx(inode);
				if (IS_ERR(ctx))
					goto set_error_page;

				/* wait the page to be moved by cleaning */
971 972
				f2fs_wait_on_encrypted_page_writeback(
						F2FS_I_SB(inode), block_nr);
973 974
			}

J
Jaegeuk Kim 已提交
975
			bio = bio_alloc(GFP_KERNEL,
976
				min_t(int, nr_pages, BIO_MAX_PAGES));
977 978 979
			if (!bio) {
				if (ctx)
					f2fs_release_crypto_ctx(ctx);
J
Jaegeuk Kim 已提交
980
				goto set_error_page;
981
			}
J
Jaegeuk Kim 已提交
982 983
			bio->bi_bdev = bdev;
			bio->bi_iter.bi_sector = SECTOR_FROM_BLOCK(block_nr);
984
			bio->bi_end_io = f2fs_read_end_io;
985
			bio->bi_private = ctx;
J
Jaegeuk Kim 已提交
986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013
		}

		if (bio_add_page(bio, page, blocksize, 0) < blocksize)
			goto submit_and_realloc;

		last_block_in_bio = block_nr;
		goto next_page;
set_error_page:
		SetPageError(page);
		zero_user_segment(page, 0, PAGE_CACHE_SIZE);
		unlock_page(page);
		goto next_page;
confused:
		if (bio) {
			submit_bio(READ, bio);
			bio = NULL;
		}
		unlock_page(page);
next_page:
		if (pages)
			page_cache_release(page);
	}
	BUG_ON(pages && !list_empty(pages));
	if (bio)
		submit_bio(READ, bio);
	return 0;
}

1014 1015
static int f2fs_read_data_page(struct file *file, struct page *page)
{
H
Huajun Li 已提交
1016
	struct inode *inode = page->mapping->host;
1017
	int ret = -EAGAIN;
H
Huajun Li 已提交
1018

1019 1020
	trace_f2fs_readpage(page, DATA);

A
arter97 已提交
1021
	/* If the file has inline data, try to read it directly */
H
Huajun Li 已提交
1022 1023
	if (f2fs_has_inline_data(inode))
		ret = f2fs_read_inline_data(inode, page);
1024
	if (ret == -EAGAIN)
J
Jaegeuk Kim 已提交
1025
		ret = f2fs_mpage_readpages(page->mapping, NULL, page, 1);
H
Huajun Li 已提交
1026
	return ret;
1027 1028 1029 1030 1031 1032
}

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
1033
	struct inode *inode = file->f_mapping->host;
1034 1035 1036
	struct page *page = list_entry(pages->prev, struct page, lru);

	trace_f2fs_readpages(inode, page, nr_pages);
H
Huajun Li 已提交
1037 1038 1039 1040 1041

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

J
Jaegeuk Kim 已提交
1042
	return f2fs_mpage_readpages(mapping, pages, NULL, nr_pages);
1043 1044
}

1045
int do_write_data_page(struct f2fs_io_info *fio)
1046
{
1047
	struct page *page = fio->page;
1048 1049 1050 1051 1052
	struct inode *inode = page->mapping->host;
	struct dnode_of_data dn;
	int err = 0;

	set_new_dnode(&dn, inode, NULL, NULL, 0);
1053
	err = get_dnode_of_data(&dn, page->index, LOOKUP_NODE);
1054 1055 1056
	if (err)
		return err;

1057
	fio->blk_addr = dn.data_blkaddr;
1058 1059

	/* This page is already truncated */
1060 1061
	if (fio->blk_addr == NULL_ADDR) {
		ClearPageUptodate(page);
1062
		goto out_writepage;
1063
	}
1064

1065
	if (f2fs_encrypted_inode(inode) && S_ISREG(inode->i_mode)) {
1066 1067 1068 1069 1070

		/* wait for GCed encrypted page writeback */
		f2fs_wait_on_encrypted_page_writeback(F2FS_I_SB(inode),
							fio->blk_addr);

1071 1072 1073 1074 1075 1076 1077
		fio->encrypted_page = f2fs_encrypt(inode, fio->page);
		if (IS_ERR(fio->encrypted_page)) {
			err = PTR_ERR(fio->encrypted_page);
			goto out_writepage;
		}
	}

1078 1079 1080 1081 1082 1083
	set_page_writeback(page);

	/*
	 * If current allocation needs SSR,
	 * it had better in-place writes for updated data.
	 */
1084
	if (unlikely(fio->blk_addr != NEW_ADDR &&
1085
			!is_cold_data(page) &&
1086
			!IS_ATOMIC_WRITTEN_PAGE(page) &&
1087
			need_inplace_update(inode))) {
1088
		rewrite_data_page(fio);
1089
		set_inode_flag(F2FS_I(inode), FI_UPDATE_WRITE);
1090
		trace_f2fs_do_write_data_page(page, IPU);
1091
	} else {
1092
		write_data_page(&dn, fio);
1093
		set_data_blkaddr(&dn);
1094
		f2fs_update_extent_cache(&dn);
1095
		trace_f2fs_do_write_data_page(page, OPU);
1096
		set_inode_flag(F2FS_I(inode), FI_APPEND_WRITE);
1097 1098
		if (page->index == 0)
			set_inode_flag(F2FS_I(inode), FI_FIRST_BLOCK_WRITTEN);
1099 1100 1101 1102 1103 1104 1105 1106 1107 1108
	}
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;
1109
	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
1110 1111 1112
	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
1113
	unsigned offset = 0;
1114
	bool need_balance_fs = false;
1115
	int err = 0;
J
Jaegeuk Kim 已提交
1116
	struct f2fs_io_info fio = {
1117
		.sbi = sbi,
J
Jaegeuk Kim 已提交
1118
		.type = DATA,
C
Chris Fries 已提交
1119
		.rw = (wbc->sync_mode == WB_SYNC_ALL) ? WRITE_SYNC : WRITE,
1120
		.page = page,
1121
		.encrypted_page = NULL,
J
Jaegeuk Kim 已提交
1122
	};
1123

1124 1125
	trace_f2fs_writepage(page, DATA);

1126
	if (page->index < end_index)
1127
		goto write;
1128 1129 1130 1131 1132 1133

	/*
	 * 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);
1134
	if ((page->index >= end_index + 1) || !offset)
1135
		goto out;
1136 1137

	zero_user_segment(page, offset, PAGE_CACHE_SIZE);
1138
write:
1139
	if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING)))
1140
		goto redirty_out;
1141 1142 1143 1144 1145
	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;
1146

1147
	/* Dentry blocks are controlled by checkpoint */
1148
	if (S_ISDIR(inode->i_mode)) {
1149 1150
		if (unlikely(f2fs_cp_error(sbi)))
			goto redirty_out;
1151
		err = do_write_data_page(&fio);
1152 1153
		goto done;
	}
H
Huajun Li 已提交
1154

1155 1156 1157
	/* we should bypass data pages to proceed the kworkder jobs */
	if (unlikely(f2fs_cp_error(sbi))) {
		SetPageError(page);
1158
		goto out;
1159 1160
	}

1161
	if (!wbc->for_reclaim)
1162
		need_balance_fs = true;
1163
	else if (has_not_enough_free_secs(sbi, 0))
1164
		goto redirty_out;
1165

1166
	err = -EAGAIN;
1167
	f2fs_lock_op(sbi);
1168 1169 1170
	if (f2fs_has_inline_data(inode))
		err = f2fs_write_inline_data(inode, page);
	if (err == -EAGAIN)
1171
		err = do_write_data_page(&fio);
1172 1173 1174 1175
	f2fs_unlock_op(sbi);
done:
	if (err && err != -ENOENT)
		goto redirty_out;
1176 1177

	clear_cold_data(page);
1178
out:
1179
	inode_dec_dirty_pages(inode);
1180 1181
	if (err)
		ClearPageUptodate(page);
1182
	unlock_page(page);
1183
	if (need_balance_fs)
1184
		f2fs_balance_fs(sbi);
1185 1186
	if (wbc->for_reclaim)
		f2fs_submit_merged_bio(sbi, DATA, WRITE);
1187 1188 1189
	return 0;

redirty_out:
1190
	redirty_page_for_writepage(wbc, page);
1191
	return AOP_WRITEPAGE_ACTIVATE;
1192 1193
}

1194 1195 1196 1197 1198 1199 1200 1201 1202
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;
}

C
Chao Yu 已提交
1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280
/*
 * This function was copied from write_cche_pages from mm/page-writeback.c.
 * The major change is making write step of cold data page separately from
 * warm/hot data page.
 */
static int f2fs_write_cache_pages(struct address_space *mapping,
			struct writeback_control *wbc, writepage_t writepage,
			void *data)
{
	int ret = 0;
	int done = 0;
	struct pagevec pvec;
	int nr_pages;
	pgoff_t uninitialized_var(writeback_index);
	pgoff_t index;
	pgoff_t end;		/* Inclusive */
	pgoff_t done_index;
	int cycled;
	int range_whole = 0;
	int tag;
	int step = 0;

	pagevec_init(&pvec, 0);
next:
	if (wbc->range_cyclic) {
		writeback_index = mapping->writeback_index; /* prev offset */
		index = writeback_index;
		if (index == 0)
			cycled = 1;
		else
			cycled = 0;
		end = -1;
	} else {
		index = wbc->range_start >> PAGE_CACHE_SHIFT;
		end = wbc->range_end >> PAGE_CACHE_SHIFT;
		if (wbc->range_start == 0 && wbc->range_end == LLONG_MAX)
			range_whole = 1;
		cycled = 1; /* ignore range_cyclic tests */
	}
	if (wbc->sync_mode == WB_SYNC_ALL || wbc->tagged_writepages)
		tag = PAGECACHE_TAG_TOWRITE;
	else
		tag = PAGECACHE_TAG_DIRTY;
retry:
	if (wbc->sync_mode == WB_SYNC_ALL || wbc->tagged_writepages)
		tag_pages_for_writeback(mapping, index, end);
	done_index = index;
	while (!done && (index <= end)) {
		int i;

		nr_pages = pagevec_lookup_tag(&pvec, mapping, &index, tag,
			      min(end - index, (pgoff_t)PAGEVEC_SIZE - 1) + 1);
		if (nr_pages == 0)
			break;

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

			if (page->index > end) {
				done = 1;
				break;
			}

			done_index = page->index;

			lock_page(page);

			if (unlikely(page->mapping != mapping)) {
continue_unlock:
				unlock_page(page);
				continue;
			}

			if (!PageDirty(page)) {
				/* someone wrote it for us */
				goto continue_unlock;
			}

1281
			if (step == is_cold_data(page))
C
Chao Yu 已提交
1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333
				goto continue_unlock;

			if (PageWriteback(page)) {
				if (wbc->sync_mode != WB_SYNC_NONE)
					f2fs_wait_on_page_writeback(page, DATA);
				else
					goto continue_unlock;
			}

			BUG_ON(PageWriteback(page));
			if (!clear_page_dirty_for_io(page))
				goto continue_unlock;

			ret = (*writepage)(page, wbc, data);
			if (unlikely(ret)) {
				if (ret == AOP_WRITEPAGE_ACTIVATE) {
					unlock_page(page);
					ret = 0;
				} else {
					done_index = page->index + 1;
					done = 1;
					break;
				}
			}

			if (--wbc->nr_to_write <= 0 &&
			    wbc->sync_mode == WB_SYNC_NONE) {
				done = 1;
				break;
			}
		}
		pagevec_release(&pvec);
		cond_resched();
	}

	if (step < 1) {
		step++;
		goto next;
	}

	if (!cycled && !done) {
		cycled = 1;
		index = 0;
		end = writeback_index - 1;
		goto retry;
	}
	if (wbc->range_cyclic || (range_whole && wbc->nr_to_write > 0))
		mapping->writeback_index = done_index;

	return ret;
}

1334
static int f2fs_write_data_pages(struct address_space *mapping,
1335 1336 1337
			    struct writeback_control *wbc)
{
	struct inode *inode = mapping->host;
1338
	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
1339
	bool locked = false;
1340
	int ret;
1341
	long diff;
1342

1343 1344
	trace_f2fs_writepages(mapping->host, wbc, DATA);

P
P J P 已提交
1345 1346 1347 1348
	/* deal with chardevs and other special file */
	if (!mapping->a_ops->writepage)
		return 0;

1349 1350 1351 1352
	/* skip writing if there is no dirty page in this inode */
	if (!get_dirty_pages(inode) && wbc->sync_mode == WB_SYNC_NONE)
		return 0;

1353 1354 1355 1356 1357
	if (S_ISDIR(inode->i_mode) && wbc->sync_mode == WB_SYNC_NONE &&
			get_dirty_pages(inode) < nr_pages_to_skip(sbi, DATA) &&
			available_free_memory(sbi, DIRTY_DENTS))
		goto skip_write;

1358 1359 1360 1361
	/* during POR, we don't need to trigger writepage at all. */
	if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING)))
		goto skip_write;

1362
	diff = nr_pages_to_write(sbi, DATA, wbc);
1363

1364 1365 1366 1367
	if (!S_ISDIR(inode->i_mode)) {
		mutex_lock(&sbi->writepages);
		locked = true;
	}
C
Chao Yu 已提交
1368
	ret = f2fs_write_cache_pages(mapping, wbc, __f2fs_writepage, mapping);
C
Chao Yu 已提交
1369
	f2fs_submit_merged_bio(sbi, DATA, WRITE);
1370 1371
	if (locked)
		mutex_unlock(&sbi->writepages);
J
Jaegeuk Kim 已提交
1372

1373 1374
	remove_dirty_dir_inode(inode);

1375
	wbc->nr_to_write = max((long)0, wbc->nr_to_write - diff);
1376
	return ret;
1377 1378

skip_write:
1379
	wbc->pages_skipped += get_dirty_pages(inode);
1380
	return 0;
1381 1382
}

1383 1384 1385 1386 1387 1388
static void f2fs_write_failed(struct address_space *mapping, loff_t to)
{
	struct inode *inode = mapping->host;

	if (to > inode->i_size) {
		truncate_pagecache(inode, inode->i_size);
1389
		truncate_blocks(inode, inode->i_size, true);
1390 1391 1392
	}
}

1393 1394 1395 1396 1397
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;
1398
	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
1399 1400
	struct page *page = NULL;
	struct page *ipage;
1401 1402 1403 1404
	pgoff_t index = ((unsigned long long) pos) >> PAGE_CACHE_SHIFT;
	struct dnode_of_data dn;
	int err = 0;

1405 1406
	trace_f2fs_write_begin(inode, pos, len, flags);

1407
	f2fs_balance_fs(sbi);
1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418

	/*
	 * We should check this at this moment to avoid deadlock on inode page
	 * and #0 page. The locking rule for inline_data conversion should be:
	 * lock_page(page #0) -> lock_page(inode_page)
	 */
	if (index != 0) {
		err = f2fs_convert_inline_inode(inode);
		if (err)
			goto fail;
	}
1419
repeat:
1420
	page = grab_cache_page_write_begin(mapping, index, flags);
1421 1422 1423 1424
	if (!page) {
		err = -ENOMEM;
		goto fail;
	}
1425

1426 1427
	*pagep = page;

1428
	f2fs_lock_op(sbi);
1429 1430 1431

	/* check inline_data */
	ipage = get_node_page(sbi, inode->i_ino);
1432 1433
	if (IS_ERR(ipage)) {
		err = PTR_ERR(ipage);
1434
		goto unlock_fail;
1435
	}
1436

1437 1438
	set_new_dnode(&dn, inode, ipage, ipage, 0);

1439
	if (f2fs_has_inline_data(inode)) {
1440 1441 1442 1443 1444 1445
		if (pos + len <= MAX_INLINE_DATA) {
			read_inline_data(page, ipage);
			set_inode_flag(F2FS_I(inode), FI_DATA_EXIST);
			sync_inode_page(&dn);
			goto put_next;
		}
1446 1447 1448
		err = f2fs_convert_inline_page(&dn, page);
		if (err)
			goto put_fail;
1449
	}
1450 1451

	err = f2fs_get_block(&dn, index);
1452
	if (err)
1453
		goto put_fail;
1454
put_next:
1455 1456 1457
	f2fs_put_dnode(&dn);
	f2fs_unlock_op(sbi);

1458 1459
	f2fs_wait_on_page_writeback(page, DATA);

1460 1461 1462 1463
	/* wait for GCed encrypted page writeback */
	if (f2fs_encrypted_inode(inode) && S_ISREG(inode->i_mode))
		f2fs_wait_on_encrypted_page_writeback(sbi, dn.data_blkaddr);

C
Chao Yu 已提交
1464 1465 1466 1467
	if (len == PAGE_CACHE_SIZE)
		goto out_update;
	if (PageUptodate(page))
		goto out_clear;
1468 1469 1470 1471 1472 1473 1474

	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);
C
Chao Yu 已提交
1475
		goto out_update;
1476 1477
	}

1478
	if (dn.data_blkaddr == NEW_ADDR) {
1479 1480
		zero_user_segment(page, 0, PAGE_CACHE_SIZE);
	} else {
1481
		struct f2fs_io_info fio = {
1482
			.sbi = sbi,
1483 1484 1485
			.type = DATA,
			.rw = READ_SYNC,
			.blk_addr = dn.data_blkaddr,
1486
			.page = page,
1487
			.encrypted_page = NULL,
1488
		};
1489
		err = f2fs_submit_page_bio(&fio);
1490 1491
		if (err)
			goto fail;
1492

1493
		lock_page(page);
1494
		if (unlikely(!PageUptodate(page))) {
1495 1496
			err = -EIO;
			goto fail;
1497
		}
1498
		if (unlikely(page->mapping != mapping)) {
1499 1500
			f2fs_put_page(page, 1);
			goto repeat;
1501
		}
1502 1503 1504 1505

		/* avoid symlink page */
		if (f2fs_encrypted_inode(inode) && S_ISREG(inode->i_mode)) {
			err = f2fs_decrypt_one(inode, page);
1506
			if (err)
1507 1508
				goto fail;
		}
1509
	}
C
Chao Yu 已提交
1510
out_update:
1511
	SetPageUptodate(page);
C
Chao Yu 已提交
1512
out_clear:
1513 1514
	clear_cold_data(page);
	return 0;
1515

1516 1517
put_fail:
	f2fs_put_dnode(&dn);
1518 1519
unlock_fail:
	f2fs_unlock_op(sbi);
1520
fail:
1521
	f2fs_put_page(page, 1);
1522 1523
	f2fs_write_failed(mapping, pos + len);
	return err;
1524 1525
}

1526 1527 1528 1529 1530 1531 1532
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;

1533 1534
	trace_f2fs_write_end(inode, pos, len, copied);

1535
	set_page_dirty(page);
1536 1537 1538 1539 1540 1541 1542

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

1543
	f2fs_put_page(page, 1);
1544 1545 1546
	return copied;
}

1547 1548
static int check_direct_IO(struct inode *inode, struct iov_iter *iter,
			   loff_t offset)
1549 1550 1551 1552 1553 1554
{
	unsigned blocksize_mask = inode->i_sb->s_blocksize - 1;

	if (offset & blocksize_mask)
		return -EINVAL;

A
Al Viro 已提交
1555 1556 1557
	if (iov_iter_alignment(iter) & blocksize_mask)
		return -EINVAL;

1558 1559 1560
	return 0;
}

1561 1562
static ssize_t f2fs_direct_IO(struct kiocb *iocb, struct iov_iter *iter,
			      loff_t offset)
1563 1564
{
	struct file *file = iocb->ki_filp;
1565 1566 1567 1568
	struct address_space *mapping = file->f_mapping;
	struct inode *inode = mapping->host;
	size_t count = iov_iter_count(iter);
	int err;
1569

1570 1571 1572 1573 1574 1575
	/* we don't need to use inline_data strictly */
	if (f2fs_has_inline_data(inode)) {
		err = f2fs_convert_inline_inode(inode);
		if (err)
			return err;
	}
H
Huajun Li 已提交
1576

1577 1578 1579
	if (f2fs_encrypted_inode(inode) && S_ISREG(inode->i_mode))
		return 0;

1580 1581 1582
	err = check_direct_IO(inode, iter, offset);
	if (err)
		return err;
1583

1584
	trace_f2fs_direct_IO_enter(inode, offset, count, iov_iter_rw(iter));
1585

1586
	if (iov_iter_rw(iter) == WRITE) {
1587
		__allocate_data_blocks(inode, offset, count);
1588 1589 1590 1591 1592
		if (unlikely(f2fs_cp_error(F2FS_I_SB(inode)))) {
			err = -EIO;
			goto out;
		}
	}
1593

C
Chao Yu 已提交
1594
	err = blockdev_direct_IO(iocb, inode, iter, offset, get_data_block_dio);
1595
out:
1596
	if (err < 0 && iov_iter_rw(iter) == WRITE)
1597
		f2fs_write_failed(mapping, offset + count);
1598

1599
	trace_f2fs_direct_IO_exit(inode, offset, count, iov_iter_rw(iter), err);
1600

1601
	return err;
1602 1603
}

1604 1605
void f2fs_invalidate_page(struct page *page, unsigned int offset,
							unsigned int length)
1606 1607
{
	struct inode *inode = page->mapping->host;
1608
	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
1609

1610 1611
	if (inode->i_ino >= F2FS_ROOT_INO(sbi) &&
		(offset % PAGE_CACHE_SIZE || length != PAGE_CACHE_SIZE))
1612 1613
		return;

1614 1615 1616 1617 1618 1619 1620 1621
	if (PageDirty(page)) {
		if (inode->i_ino == F2FS_META_INO(sbi))
			dec_page_count(sbi, F2FS_DIRTY_META);
		else if (inode->i_ino == F2FS_NODE_INO(sbi))
			dec_page_count(sbi, F2FS_DIRTY_NODES);
		else
			inode_dec_dirty_pages(inode);
	}
C
Chao Yu 已提交
1622 1623 1624 1625 1626

	/* This is atomic written page, keep Private */
	if (IS_ATOMIC_WRITTEN_PAGE(page))
		return;

1627 1628 1629
	ClearPagePrivate(page);
}

1630
int f2fs_release_page(struct page *page, gfp_t wait)
1631
{
1632 1633 1634 1635
	/* If this is dirty page, keep PagePrivate */
	if (PageDirty(page))
		return 0;

C
Chao Yu 已提交
1636 1637 1638 1639
	/* This is atomic written page, keep Private */
	if (IS_ATOMIC_WRITTEN_PAGE(page))
		return 0;

1640
	ClearPagePrivate(page);
1641
	return 1;
1642 1643 1644 1645 1646 1647 1648
}

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

1649 1650
	trace_f2fs_set_page_dirty(page, DATA);

1651
	SetPageUptodate(page);
1652

1653
	if (f2fs_is_atomic_file(inode)) {
C
Chao Yu 已提交
1654 1655 1656 1657 1658 1659 1660 1661 1662
		if (!IS_ATOMIC_WRITTEN_PAGE(page)) {
			register_inmem_page(inode, page);
			return 1;
		}
		/*
		 * Previously, this page has been registered, we just
		 * return here.
		 */
		return 0;
1663 1664
	}

1665 1666
	if (!PageDirty(page)) {
		__set_page_dirty_nobuffers(page);
1667
		update_dirty_page(inode, page);
1668 1669 1670 1671 1672
		return 1;
	}
	return 0;
}

J
Jaegeuk Kim 已提交
1673 1674
static sector_t f2fs_bmap(struct address_space *mapping, sector_t block)
{
1675 1676
	struct inode *inode = mapping->host;

J
Jaegeuk Kim 已提交
1677 1678 1679 1680 1681 1682 1683
	if (f2fs_has_inline_data(inode))
		return 0;

	/* make sure allocating whole blocks */
	if (mapping_tagged(mapping, PAGECACHE_TAG_DIRTY))
		filemap_write_and_wait(mapping);

C
Chao Yu 已提交
1684
	return generic_block_bmap(mapping, block, get_data_block_bmap);
1685 1686
}

1687 1688 1689 1690 1691 1692
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,
1693
	.write_end	= f2fs_write_end,
1694
	.set_page_dirty	= f2fs_set_data_page_dirty,
1695 1696
	.invalidatepage	= f2fs_invalidate_page,
	.releasepage	= f2fs_release_page,
1697
	.direct_IO	= f2fs_direct_IO,
J
Jaegeuk Kim 已提交
1698
	.bmap		= f2fs_bmap,
1699
};