data.c 20.1 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

J
Jaegeuk Kim 已提交
27
/*
28 29 30 31 32 33 34 35 36 37 38 39
 * 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;

J
Jin Xu 已提交
40
	f2fs_wait_on_page_writeback(node_page, NODE, false);
41

42
	rn = F2FS_NODE(node_page);
43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58

	/* 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);

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

59 60
	trace_f2fs_reserve_new_block(dn->inode, dn->nid, dn->ofs_in_node);

61 62 63 64 65 66 67 68 69 70 71 72 73
	__set_data_blkaddr(dn, NEW_ADDR);
	dn->data_blkaddr = NEW_ADDR;
	sync_inode_page(dn);
	return 0;
}

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;

74 75 76
	if (is_inode_flag_set(fi, FI_NO_EXTENT))
		return 0;

77 78 79 80 81 82
	read_lock(&fi->ext.ext_lock);
	if (fi->ext.len == 0) {
		read_unlock(&fi->ext.ext_lock);
		return 0;
	}

83 84
	stat_inc_total_hit(inode->i_sb);

85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101
	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;

102
		stat_inc_read_hit(inode->i_sb);
103 104 105 106 107 108 109 110 111 112 113 114
		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;
115
	int need_update = true;
116

117
	f2fs_bug_on(blk_addr == NEW_ADDR);
118 119
	fofs = start_bidx_of_node(ofs_of_node(dn->node_page), fi) +
							dn->ofs_in_node;
120 121 122 123

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

124 125 126
	if (is_inode_flag_set(fi, FI_NO_EXTENT))
		return;

127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147
	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
148
	/* Front merge */
149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172
	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;
		}
173 174
	} else {
		need_update = false;
175 176
	}

177 178 179 180 181 182
	/* 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;
	}
183 184
end_update:
	write_unlock(&fi->ext.ext_lock);
185 186 187
	if (need_update)
		sync_inode_page(dn);
	return;
188 189
}

190
struct page *find_data_page(struct inode *inode, pgoff_t index, bool sync)
191 192 193 194 195 196 197 198 199 200 201 202 203
{
	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);
204
	err = get_dnode_of_data(&dn, index, LOOKUP_NODE);
205 206 207 208 209 210 211 212 213 214 215
	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 */
	if (dn.data_blkaddr == NEW_ADDR)
		return ERR_PTR(-EINVAL);

216
	page = grab_cache_page_write_begin(mapping, index, AOP_FLAG_NOFS);
217 218 219
	if (!page)
		return ERR_PTR(-ENOMEM);

220 221 222 223 224
	if (PageUptodate(page)) {
		unlock_page(page);
		return page;
	}

225 226 227 228 229 230 231 232
	err = f2fs_readpage(sbi, page, dn.data_blkaddr,
					sync ? READ_SYNC : READA);
	if (sync) {
		wait_on_page_locked(page);
		if (!PageUptodate(page)) {
			f2fs_put_page(page, 0);
			return ERR_PTR(-EIO);
		}
233 234 235 236
	}
	return page;
}

J
Jaegeuk Kim 已提交
237
/*
238 239 240 241 242 243 244 245 246 247 248 249
 * 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;

250
repeat:
251
	page = grab_cache_page_write_begin(mapping, index, AOP_FLAG_NOFS);
252 253 254
	if (!page)
		return ERR_PTR(-ENOMEM);

255
	set_new_dnode(&dn, inode, NULL, NULL, 0);
256
	err = get_dnode_of_data(&dn, index, LOOKUP_NODE);
257 258
	if (err) {
		f2fs_put_page(page, 1);
259
		return ERR_PTR(err);
260
	}
261 262
	f2fs_put_dnode(&dn);

263 264
	if (dn.data_blkaddr == NULL_ADDR) {
		f2fs_put_page(page, 1);
265
		return ERR_PTR(-ENOENT);
266
	}
267 268 269 270

	if (PageUptodate(page))
		return page;

J
Jaegeuk Kim 已提交
271 272 273 274 275 276 277 278 279 280 281
	/*
	 * 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;
	}
282 283

	err = f2fs_readpage(sbi, page, dn.data_blkaddr, READ_SYNC);
284
	if (err)
285
		return ERR_PTR(err);
286 287 288 289 290

	lock_page(page);
	if (!PageUptodate(page)) {
		f2fs_put_page(page, 1);
		return ERR_PTR(-EIO);
291
	}
292 293 294
	if (page->mapping != mapping) {
		f2fs_put_page(page, 1);
		goto repeat;
295 296 297 298
	}
	return page;
}

J
Jaegeuk Kim 已提交
299
/*
300 301
 * Caller ensures that this data page is never allocated.
 * A new zero-filled data page is allocated in the page cache.
302 303 304
 *
 * Also, caller should grab and release a mutex by calling mutex_lock_op() and
 * mutex_unlock_op().
305
 * Note that, npage is set only by make_empty_dir.
306
 */
307 308
struct page *get_new_data_page(struct inode *inode,
		struct page *npage, pgoff_t index, bool new_i_size)
309 310 311 312 313 314 315
{
	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;

316
	set_new_dnode(&dn, inode, npage, npage, 0);
317
	err = get_dnode_of_data(&dn, index, ALLOC_NODE);
318 319 320 321 322
	if (err)
		return ERR_PTR(err);

	if (dn.data_blkaddr == NULL_ADDR) {
		if (reserve_new_block(&dn)) {
323 324
			if (!npage)
				f2fs_put_dnode(&dn);
325 326 327
			return ERR_PTR(-ENOSPC);
		}
	}
328 329
	if (!npage)
		f2fs_put_dnode(&dn);
330
repeat:
331 332 333 334 335 336 337 338 339
	page = grab_cache_page(mapping, index);
	if (!page)
		return ERR_PTR(-ENOMEM);

	if (PageUptodate(page))
		return page;

	if (dn.data_blkaddr == NEW_ADDR) {
		zero_user_segment(page, 0, PAGE_CACHE_SIZE);
340
		SetPageUptodate(page);
341 342
	} else {
		err = f2fs_readpage(sbi, page, dn.data_blkaddr, READ_SYNC);
343
		if (err)
344
			return ERR_PTR(err);
345 346 347 348
		lock_page(page);
		if (!PageUptodate(page)) {
			f2fs_put_page(page, 1);
			return ERR_PTR(-EIO);
349
		}
350 351 352
		if (page->mapping != mapping) {
			f2fs_put_page(page, 1);
			goto repeat;
353 354 355 356 357 358
		}
	}

	if (new_i_size &&
		i_size_read(inode) < ((index + 1) << PAGE_CACHE_SHIFT)) {
		i_size_write(inode, ((index + 1) << PAGE_CACHE_SHIFT));
359 360
		/* Only the directory inode sets new_i_size */
		set_inode_flag(F2FS_I(inode), FI_UPDATE_DIR);
361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387
		mark_inode_dirty_sync(inode);
	}
	return page;
}

static void read_end_io(struct bio *bio, int err)
{
	const int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
	struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;

	do {
		struct page *page = bvec->bv_page;

		if (--bvec >= bio->bi_io_vec)
			prefetchw(&bvec->bv_page->flags);

		if (uptodate) {
			SetPageUptodate(page);
		} else {
			ClearPageUptodate(page);
			SetPageError(page);
		}
		unlock_page(page);
	} while (bvec >= bio->bi_io_vec);
	bio_put(bio);
}

J
Jaegeuk Kim 已提交
388
/*
389
 * Fill the locked page with data located in the block address.
390
 * Return unlocked page.
391 392 393 394 395 396 397
 */
int f2fs_readpage(struct f2fs_sb_info *sbi, struct page *page,
					block_t blk_addr, int type)
{
	struct block_device *bdev = sbi->sb->s_bdev;
	struct bio *bio;

398
	trace_f2fs_readpage(page, blk_addr, type);
399 400

	/* Allocate a new bio */
401
	bio = f2fs_bio_alloc(bdev, 1);
402 403

	/* Initialize the bio */
404
	bio->bi_sector = SECTOR_FROM_BLOCK(sbi, blk_addr);
405
	bio->bi_end_io = read_end_io;
406

407 408
	if (bio_add_page(bio, page, PAGE_CACHE_SIZE, 0) < PAGE_CACHE_SIZE) {
		bio_put(bio);
409
		f2fs_put_page(page, 1);
410 411 412 413 414 415 416
		return -EFAULT;
	}

	submit_bio(type, bio);
	return 0;
}

417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464
void f2fs_submit_read_bio(struct f2fs_sb_info *sbi, int rw)
{
	struct f2fs_bio_info *io = &sbi->read_io;

	if (!io->bio)
		return;

	mutex_lock(&io->io_mutex);
	if (io->bio) {
		submit_bio(rw, io->bio);
		io->bio = NULL;
	}
	mutex_unlock(&io->io_mutex);
}

void submit_read_page(struct f2fs_sb_info *sbi, struct page *page,
					block_t blk_addr, int rw)
{
	struct block_device *bdev = sbi->sb->s_bdev;
	struct f2fs_bio_info *io = &sbi->read_io;
	int bio_blocks;

	verify_block_addr(sbi, blk_addr);

	mutex_lock(&io->io_mutex);

	if (io->bio && io->last_block_in_bio != blk_addr - 1) {
		submit_bio(rw, io->bio);
		io->bio = NULL;
	}
alloc_new:
	if (io->bio == NULL) {
		bio_blocks = MAX_BIO_BLOCKS(max_hw_blocks(sbi));
		io->bio = f2fs_bio_alloc(bdev, bio_blocks);
		io->bio->bi_sector = SECTOR_FROM_BLOCK(sbi, blk_addr);
		io->bio->bi_end_io = read_end_io;
	}

	if (bio_add_page(io->bio, page, PAGE_CACHE_SIZE, 0) <
							PAGE_CACHE_SIZE) {
		submit_bio(rw, io->bio);
		io->bio = NULL;
		goto alloc_new;
	}

	io->last_block_in_bio = blk_addr;

	mutex_unlock(&io->io_mutex);
465
	trace_f2fs_submit_read_page(page, rw, META, blk_addr);
466 467
}

J
Jaegeuk Kim 已提交
468
/*
469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485
 * This function should be used by the data read flow only where it
 * does not check the "create" flag that indicates block allocation.
 * The reason for this special functionality is to exploit VFS readahead
 * mechanism.
 */
static int get_data_block_ro(struct inode *inode, sector_t iblock,
			struct buffer_head *bh_result, int create)
{
	unsigned int blkbits = inode->i_sb->s_blocksize_bits;
	unsigned maxblocks = bh_result->b_size >> blkbits;
	struct dnode_of_data dn;
	pgoff_t pgofs;
	int err;

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

486 487
	if (check_extent_cache(inode, pgofs, bh_result)) {
		trace_f2fs_get_data_block(inode, iblock, bh_result, 0);
488
		return 0;
489
	}
490 491 492

	/* When reading holes, we need its node page */
	set_new_dnode(&dn, inode, NULL, NULL, 0);
493
	err = get_dnode_of_data(&dn, pgofs, LOOKUP_NODE_RA);
494 495
	if (err) {
		trace_f2fs_get_data_block(inode, iblock, bh_result, err);
496
		return (err == -ENOENT) ? 0 : err;
497
	}
498 499

	/* It does not support data allocation */
500
	f2fs_bug_on(create);
501 502 503 504 505 506

	if (dn.data_blkaddr != NEW_ADDR && dn.data_blkaddr != NULL_ADDR) {
		int i;
		unsigned int end_offset;

		end_offset = IS_INODE(dn.node_page) ?
507
				ADDRS_PER_INODE(F2FS_I(inode)) :
508 509 510 511 512 513 514 515 516 517 518 519 520 521
				ADDRS_PER_BLOCK;

		clear_buffer_new(bh_result);

		/* Give more consecutive addresses for the read ahead */
		for (i = 0; i < end_offset - dn.ofs_in_node; i++)
			if (((datablock_addr(dn.node_page,
							dn.ofs_in_node + i))
				!= (dn.data_blkaddr + i)) || maxblocks == i)
				break;
		map_bh(bh_result, inode->i_sb, dn.data_blkaddr);
		bh_result->b_size = (i << blkbits);
	}
	f2fs_put_dnode(&dn);
522
	trace_f2fs_get_data_block(inode, iblock, bh_result, 0);
523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545
	return 0;
}

static int f2fs_read_data_page(struct file *file, struct page *page)
{
	return mpage_readpage(page, get_data_block_ro);
}

static int f2fs_read_data_pages(struct file *file,
			struct address_space *mapping,
			struct list_head *pages, unsigned nr_pages)
{
	return mpage_readpages(mapping, pages, nr_pages, get_data_block_ro);
}

int do_write_data_page(struct page *page)
{
	struct inode *inode = page->mapping->host;
	block_t old_blk_addr, new_blk_addr;
	struct dnode_of_data dn;
	int err = 0;

	set_new_dnode(&dn, inode, NULL, NULL, 0);
546
	err = get_dnode_of_data(&dn, page->index, LOOKUP_NODE);
547 548 549 550 551 552 553 554 555 556 557 558 559 560 561
	if (err)
		return err;

	old_blk_addr = dn.data_blkaddr;

	/* This page is already truncated */
	if (old_blk_addr == NULL_ADDR)
		goto out_writepage;

	set_page_writeback(page);

	/*
	 * If current allocation needs SSR,
	 * it had better in-place writes for updated data.
	 */
562 563 564
	if (unlikely(old_blk_addr != NEW_ADDR &&
			!is_cold_data(page) &&
			need_inplace_update(inode))) {
565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585
		rewrite_data_page(F2FS_SB(inode->i_sb), page,
						old_blk_addr);
	} else {
		write_data_page(inode, page, &dn,
				old_blk_addr, &new_blk_addr);
		update_extent_cache(new_blk_addr, &dn);
	}
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;
	unsigned offset;
586
	bool need_balance_fs = false;
587 588 589
	int err = 0;

	if (page->index < end_index)
590
		goto write;
591 592 593 594 595 596 597 598 599 600 601

	/*
	 * 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) {
		if (S_ISDIR(inode->i_mode)) {
			dec_page_count(sbi, F2FS_DIRTY_DENTS);
			inode_dec_dirty_dents(inode);
		}
602
		goto out;
603 604 605
	}

	zero_user_segment(page, offset, PAGE_CACHE_SIZE);
606 607 608
write:
	if (sbi->por_doing) {
		err = AOP_WRITEPAGE_ACTIVATE;
609
		goto redirty_out;
610
	}
611

612
	/* Dentry blocks are controlled by checkpoint */
613 614 615
	if (S_ISDIR(inode->i_mode)) {
		dec_page_count(sbi, F2FS_DIRTY_DENTS);
		inode_dec_dirty_dents(inode);
616 617
		err = do_write_data_page(page);
	} else {
618
		f2fs_lock_op(sbi);
619
		err = do_write_data_page(page);
620
		f2fs_unlock_op(sbi);
621
		need_balance_fs = true;
622
	}
623 624 625 626
	if (err == -ENOENT)
		goto out;
	else if (err)
		goto redirty_out;
627 628 629 630 631

	if (wbc->for_reclaim)
		f2fs_submit_bio(sbi, DATA, true);

	clear_cold_data(page);
632
out:
633
	unlock_page(page);
634
	if (need_balance_fs)
635 636 637 638 639 640
		f2fs_balance_fs(sbi);
	return 0;

redirty_out:
	wbc->pages_skipped++;
	set_page_dirty(page);
641
	return err;
642 643 644 645
}

#define MAX_DESIRED_PAGES_WP	4096

646 647 648 649 650 651 652 653 654
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;
}

655
static int f2fs_write_data_pages(struct address_space *mapping,
656 657 658 659
			    struct writeback_control *wbc)
{
	struct inode *inode = mapping->host;
	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
660
	bool locked = false;
661 662 663
	int ret;
	long excess_nrtw = 0, desired_nrtw;

P
P J P 已提交
664 665 666 667
	/* deal with chardevs and other special file */
	if (!mapping->a_ops->writepage)
		return 0;

668 669 670 671 672 673
	if (wbc->nr_to_write < MAX_DESIRED_PAGES_WP) {
		desired_nrtw = MAX_DESIRED_PAGES_WP;
		excess_nrtw = desired_nrtw - wbc->nr_to_write;
		wbc->nr_to_write = desired_nrtw;
	}

674
	if (!S_ISDIR(inode->i_mode)) {
675
		mutex_lock(&sbi->writepages);
676 677
		locked = true;
	}
678
	ret = write_cache_pages(mapping, wbc, __f2fs_writepage, mapping);
679
	if (locked)
680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700
		mutex_unlock(&sbi->writepages);
	f2fs_submit_bio(sbi, DATA, (wbc->sync_mode == WB_SYNC_ALL));

	remove_dirty_dir_inode(inode);

	wbc->nr_to_write -= excess_nrtw;
	return ret;
}

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);
701
repeat:
702 703 704 705 706
	page = grab_cache_page_write_begin(mapping, index, flags);
	if (!page)
		return -ENOMEM;
	*pagep = page;

707
	f2fs_lock_op(sbi);
708 709

	set_new_dnode(&dn, inode, NULL, NULL, 0);
710
	err = get_dnode_of_data(&dn, index, ALLOC_NODE);
711 712
	if (err)
		goto err;
713

714
	if (dn.data_blkaddr == NULL_ADDR)
715
		err = reserve_new_block(&dn);
716

717
	f2fs_put_dnode(&dn);
718 719
	if (err)
		goto err;
720

721
	f2fs_unlock_op(sbi);
722 723 724 725 726 727 728 729 730 731

	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);
732
		goto out;
733 734 735 736 737 738
	}

	if (dn.data_blkaddr == NEW_ADDR) {
		zero_user_segment(page, 0, PAGE_CACHE_SIZE);
	} else {
		err = f2fs_readpage(sbi, page, dn.data_blkaddr, READ_SYNC);
739
		if (err)
740
			return err;
741 742 743 744
		lock_page(page);
		if (!PageUptodate(page)) {
			f2fs_put_page(page, 1);
			return -EIO;
745
		}
746 747 748
		if (page->mapping != mapping) {
			f2fs_put_page(page, 1);
			goto repeat;
749 750
		}
	}
751
out:
752 753 754
	SetPageUptodate(page);
	clear_cold_data(page);
	return 0;
755 756

err:
757
	f2fs_unlock_op(sbi);
758 759
	f2fs_put_page(page, 1);
	return err;
760 761
}

762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777
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);
	}

778
	f2fs_put_page(page, 1);
779 780 781
	return copied;
}

782 783 784 785 786 787 788 789 790 791 792 793 794 795
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;

	if (rw == WRITE)
		return 0;

	/* Needs synchronization with the cleaner */
	return blockdev_direct_IO(rw, iocb, inode, iov, offset, nr_segs,
						  get_data_block_ro);
}

796 797
static void f2fs_invalidate_data_page(struct page *page, unsigned int offset,
				      unsigned int length)
798 799 800 801 802 803 804 805 806 807 808 809 810
{
	struct inode *inode = page->mapping->host;
	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
	if (S_ISDIR(inode->i_mode) && PageDirty(page)) {
		dec_page_count(sbi, F2FS_DIRTY_DENTS);
		inode_dec_dirty_dents(inode);
	}
	ClearPagePrivate(page);
}

static int f2fs_release_data_page(struct page *page, gfp_t wait)
{
	ClearPagePrivate(page);
811
	return 1;
812 813 814 815 816 817 818
}

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

819 820
	trace_f2fs_set_page_dirty(page, DATA);

821 822 823 824 825 826 827 828 829
	SetPageUptodate(page);
	if (!PageDirty(page)) {
		__set_page_dirty_nobuffers(page);
		set_dirty_dir_page(inode, page);
		return 1;
	}
	return 0;
}

J
Jaegeuk Kim 已提交
830 831 832 833 834
static sector_t f2fs_bmap(struct address_space *mapping, sector_t block)
{
	return generic_block_bmap(mapping, block, get_data_block_ro);
}

835 836 837 838 839 840
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,
841
	.write_end	= f2fs_write_end,
842 843 844 845
	.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
846
	.bmap		= f2fs_bmap,
847
};