data.c

/*
 * 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/aio.h>
#include <linux/writeback.h>
#include <linux/backing-dev.h>
#include <linux/blkdev.h>
#include <linux/bio.h>
#include <linux/prefetch.h>

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

static void f2fs_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 (unlikely(!uptodate)) {
			ClearPageUptodate(page);
			SetPageError(page);
		} else {
			SetPageUptodate(page);
		}
		unlock_page(page);
	} while (bvec >= bio->bi_io_vec);

	bio_put(bio);
}

static void f2fs_write_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;
	struct f2fs_sb_info *sbi = F2FS_SB(bvec->bv_page->mapping->host->i_sb);

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

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

		if (unlikely(!uptodate)) {
			SetPageError(page);
			set_bit(AS_EIO, &page->mapping->flags);
			set_ckpt_flags(sbi->ckpt, CP_ERROR_FLAG);
			sbi->sb->s_flags |= MS_RDONLY;
		}
		end_page_writeback(page);
		dec_page_count(sbi, F2FS_WRITEBACK);
	} while (bvec >= bio->bi_io_vec);

	if (bio->bi_private)
		complete(bio->bi_private);

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

	bio_put(bio);
}

/*
 * 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;
	bio->bi_sector = SECTOR_FROM_BLOCK(sbi, blk_addr);
	bio->bi_end_io = is_read ? f2fs_read_end_io : f2fs_write_end_io;

	return bio;
}

static void __submit_merged_bio(struct f2fs_bio_info *io)
{
	struct f2fs_io_info *fio = &io->fio;
	int rw;

	if (!io->bio)
		return;

	rw = fio->rw;

	if (is_read_io(rw)) {
		trace_f2fs_submit_read_bio(io->sbi->sb, rw,
						fio->type, io->bio);
		submit_bio(rw, io->bio);
	} else {
		trace_f2fs_submit_write_bio(io->sbi->sb, rw,
						fio->type, io->bio);
		/*
		 * META_FLUSH is only from the checkpoint procedure, and we
		 * should wait this metadata bio for FS consistency.
		 */
		if (fio->type == META_FLUSH) {
			DECLARE_COMPLETION_ONSTACK(wait);
			io->bio->bi_private = &wait;
			submit_bio(rw, io->bio);
			wait_for_completion(&wait);
		} else {
			submit_bio(rw, io->bio);
		}
	}

	io->bio = NULL;
}

void f2fs_submit_merged_bio(struct f2fs_sb_info *sbi,
				enum page_type type, int rw)
{
	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];

	mutex_lock(&io->io_mutex);

	/* change META to META_FLUSH in the checkpoint procedure */
	if (type >= META_FLUSH) {
		io->fio.type = META_FLUSH;
		io->fio.rw = WRITE_FLUSH_FUA;
	}
	__submit_merged_bio(io);
	mutex_unlock(&io->io_mutex);
}

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

	trace_f2fs_submit_page_bio(page, blk_addr, rw);

	/* Allocate a new bio */
	bio = __bio_alloc(sbi, blk_addr, 1, is_read_io(rw));

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

	submit_bio(rw, bio);
	return 0;
}

void f2fs_submit_page_mbio(struct f2fs_sb_info *sbi, struct page *page,
			block_t blk_addr, struct f2fs_io_info *fio)
{
	enum page_type btype = PAGE_TYPE_OF_BIO(fio->type);
	struct f2fs_bio_info *io;
	bool is_read = is_read_io(fio->rw);

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

	verify_block_addr(sbi, blk_addr);

	mutex_lock(&io->io_mutex);

	if (!is_read)
		inc_page_count(sbi, F2FS_WRITEBACK);

	if (io->bio && (io->last_block_in_bio != blk_addr - 1 ||
						io->fio.rw != fio->rw))
		__submit_merged_bio(io);
alloc_new:
	if (io->bio == NULL) {
		int bio_blocks = MAX_BIO_BLOCKS(max_hw_blocks(sbi));

		io->bio = __bio_alloc(sbi, blk_addr, bio_blocks, is_read);
		io->fio = *fio;
	}

	if (bio_add_page(io->bio, page, PAGE_CACHE_SIZE, 0) <
							PAGE_CACHE_SIZE) {
		__submit_merged_bio(io);
		goto alloc_new;
	}

	io->last_block_in_bio = blk_addr;

	mutex_unlock(&io->io_mutex);
	trace_f2fs_submit_page_mbio(page, fio->rw, fio->type, blk_addr);
}

/*
 * 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;

	f2fs_wait_on_page_writeback(node_page, NODE, false);

	rn = F2FS_NODE(node_page);

	/* 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 (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;

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

	__set_data_blkaddr(dn, NEW_ADDR);
	dn->data_blkaddr = NEW_ADDR;
	sync_inode_page(dn);
	return 0;
}

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;
	if (dn->data_blkaddr == NULL_ADDR)
		err = reserve_new_block(dn);

	if (need_put)
		f2fs_put_dnode(dn);
	return err;
}

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;

	if (is_inode_flag_set(fi, FI_NO_EXTENT))
		return 0;

	read_lock(&fi->ext.ext_lock);
	if (fi->ext.len == 0) {
		read_unlock(&fi->ext.ext_lock);
		return 0;
	}

	stat_inc_total_hit(inode->i_sb);

	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;

		stat_inc_read_hit(inode->i_sb);
		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;
	int need_update = true;

	f2fs_bug_on(blk_addr == NEW_ADDR);
	fofs = start_bidx_of_node(ofs_of_node(dn->node_page), fi) +
							dn->ofs_in_node;

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

	if (is_inode_flag_set(fi, FI_NO_EXTENT))
		return;

	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;
	}

	/* Front merge */
	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;
		}
	} else {
		need_update = false;
	}

	/* 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;
	}
end_update:
	write_unlock(&fi->ext.ext_lock);
	if (need_update)
		sync_inode_page(dn);
	return;
}

struct page *find_data_page(struct inode *inode, pgoff_t index, bool sync)
{
	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);
	err = get_dnode_of_data(&dn, index, LOOKUP_NODE);
	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 (unlikely(dn.data_blkaddr == NEW_ADDR))
		return ERR_PTR(-EINVAL);

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

	if (PageUptodate(page)) {
		unlock_page(page);
		return page;
	}

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

	if (sync) {
		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.
 */
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;

repeat:
	page = grab_cache_page_write_begin(mapping, index, AOP_FLAG_NOFS);
	if (!page)
		return ERR_PTR(-ENOMEM);

	set_new_dnode(&dn, inode, NULL, NULL, 0);
	err = get_dnode_of_data(&dn, index, LOOKUP_NODE);
	if (err) {
		f2fs_put_page(page, 1);
		return ERR_PTR(err);
	}
	f2fs_put_dnode(&dn);

	if (unlikely(dn.data_blkaddr == NULL_ADDR)) {
		f2fs_put_page(page, 1);
		return ERR_PTR(-ENOENT);
	}

	if (PageUptodate(page))
		return page;

	/*
	 * 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;
	}

	err = f2fs_submit_page_bio(sbi, page, dn.data_blkaddr, READ_SYNC);
	if (err)
		return ERR_PTR(err);

	lock_page(page);
	if (unlikely(!PageUptodate(page))) {
		f2fs_put_page(page, 1);
		return ERR_PTR(-EIO);
	}
	if (unlikely(page->mapping != mapping)) {
		f2fs_put_page(page, 1);
		goto repeat;
	}
	return page;
}

/*
 * Caller ensures that this data page is never allocated.
 * A new zero-filled data page is allocated in the page cache.
 *
 * Also, caller should grab and release a rwsem by calling f2fs_lock_op() and
 * f2fs_unlock_op().
 * Note that, npage is set only by make_empty_dir.
 */
struct page *get_new_data_page(struct inode *inode,
		struct page *npage, pgoff_t index, bool new_i_size)
{
	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;

	set_new_dnode(&dn, inode, npage, npage, 0);
	err = f2fs_reserve_block(&dn, index);
	if (err)
		return ERR_PTR(err);
repeat:
	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);
		SetPageUptodate(page);
	} else {
		err = f2fs_submit_page_bio(sbi, page, dn.data_blkaddr,
								READ_SYNC);
		if (err)
			return ERR_PTR(err);
		lock_page(page);
		if (unlikely(!PageUptodate(page))) {
			f2fs_put_page(page, 1);
			return ERR_PTR(-EIO);
		}
		if (unlikely(page->mapping != mapping)) {
			f2fs_put_page(page, 1);
			goto repeat;
		}
	}

	if (new_i_size &&
		i_size_read(inode) < ((index + 1) << PAGE_CACHE_SHIFT)) {
		i_size_write(inode, ((index + 1) << PAGE_CACHE_SHIFT));
		/* Only the directory inode sets new_i_size */
		set_inode_flag(F2FS_I(inode), FI_UPDATE_DIR);
		mark_inode_dirty_sync(inode);
	}
	return page;
}

static int __allocate_data_block(struct dnode_of_data *dn)
{
	struct f2fs_sb_info *sbi = F2FS_SB(dn->inode->i_sb);
	struct f2fs_summary sum;
	block_t new_blkaddr;
	struct node_info ni;
	int type;

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

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

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

	type = CURSEG_WARM_DATA;

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

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

	dn->data_blkaddr = new_blkaddr;
	return 0;
}

/*
 * 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
 */
static int get_data_block(struct inode *inode, sector_t iblock,
			struct buffer_head *bh_result, int create)
{
	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
	unsigned int blkbits = inode->i_sb->s_blocksize_bits;
	unsigned maxblocks = bh_result->b_size >> blkbits;
	struct dnode_of_data dn;
	int mode = create ? ALLOC_NODE : LOOKUP_NODE_RA;
	pgoff_t pgofs, end_offset;
	int err = 0, ofs = 1;
	bool allocated = false;

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

	if (check_extent_cache(inode, pgofs, bh_result))
		goto out;

	if (create)
		f2fs_lock_op(sbi);

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

	if (dn.data_blkaddr != NULL_ADDR) {
		map_bh(bh_result, inode->i_sb, dn.data_blkaddr);
	} else if (create) {
		err = __allocate_data_block(&dn);
		if (err)
			goto put_out;
		allocated = true;
		map_bh(bh_result, inode->i_sb, dn.data_blkaddr);
	} else {
		goto put_out;
	}

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

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

		set_new_dnode(&dn, inode, NULL, NULL, 0);
		err = get_dnode_of_data(&dn, pgofs, mode);
		if (err) {
			if (err == -ENOENT)
				err = 0;
			goto unlock_out;
		}
		if (dn.data_blkaddr == NEW_ADDR)
			goto put_out;

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

	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;
		}
		/* Give more consecutive addresses for the read ahead */
		if (blkaddr == (bh_result->b_blocknr + ofs)) {
			ofs++;
			dn.ofs_in_node++;
			pgofs++;
			bh_result->b_size += (((size_t)1) << blkbits);
			goto get_next;
		}
	}
sync_out:
	if (allocated)
		sync_inode_page(&dn);
put_out:
	f2fs_put_dnode(&dn);
unlock_out:
	if (create)
		f2fs_unlock_op(sbi);
out:
	trace_f2fs_get_data_block(inode, iblock, bh_result, err);
	return err;
}

static int f2fs_read_data_page(struct file *file, struct page *page)
{
	struct inode *inode = page->mapping->host;
	int ret;

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

	return ret;
}

static int f2fs_read_data_pages(struct file *file,
			struct address_space *mapping,
			struct list_head *pages, unsigned nr_pages)
{
	struct inode *inode = file->f_mapping->host;

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

	return mpage_readpages(mapping, pages, nr_pages, get_data_block);
}

int do_write_data_page(struct page *page, struct f2fs_io_info *fio)
{
	struct inode *inode = page->mapping->host;
	block_t old_blkaddr, new_blkaddr;
	struct dnode_of_data dn;
	int err = 0;

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

	old_blkaddr = dn.data_blkaddr;

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

	set_page_writeback(page);

	/*
	 * If current allocation needs SSR,
	 * it had better in-place writes for updated data.
	 */
	if (unlikely(old_blkaddr != NEW_ADDR &&
			!is_cold_data(page) &&
			need_inplace_update(inode))) {
		rewrite_data_page(page, old_blkaddr, fio);
	} else {
		write_data_page(page, &dn, &new_blkaddr, fio);
		update_extent_cache(new_blkaddr, &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 = 0;
	bool need_balance_fs = false;
	int err = 0;
	struct f2fs_io_info fio = {
		.type = DATA,
		.rw = (wbc->sync_mode == WB_SYNC_ALL) ? WRITE_SYNC: WRITE,
	};

	if (page->index < end_index)
		goto write;

	/*
	 * 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);
		}
		goto out;
	}

	zero_user_segment(page, offset, PAGE_CACHE_SIZE);
write:
	if (unlikely(sbi->por_doing)) {
		err = AOP_WRITEPAGE_ACTIVATE;
		goto redirty_out;
	}

	/* Dentry blocks are controlled by checkpoint */
	if (S_ISDIR(inode->i_mode)) {
		dec_page_count(sbi, F2FS_DIRTY_DENTS);
		inode_dec_dirty_dents(inode);
		err = do_write_data_page(page, &fio);
	} else {
		f2fs_lock_op(sbi);

		if (f2fs_has_inline_data(inode) || f2fs_may_inline(inode)) {
			err = f2fs_write_inline_data(inode, page, offset);
			f2fs_unlock_op(sbi);
			goto out;
		} else {
			err = do_write_data_page(page, &fio);
		}

		f2fs_unlock_op(sbi);
		need_balance_fs = true;
	}
	if (err == -ENOENT)
		goto out;
	else if (err)
		goto redirty_out;

	if (wbc->for_reclaim)
		f2fs_submit_merged_bio(sbi, DATA, WRITE);

	clear_cold_data(page);
out:
	unlock_page(page);
	if (need_balance_fs)
		f2fs_balance_fs(sbi);
	return 0;

redirty_out:
	wbc->pages_skipped++;
	set_page_dirty(page);
	return err;
}

#define MAX_DESIRED_PAGES_WP	4096

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;
}

static int f2fs_write_data_pages(struct address_space *mapping,
			    struct writeback_control *wbc)
{
	struct inode *inode = mapping->host;
	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
	bool locked = false;
	int ret;
	long excess_nrtw = 0, desired_nrtw;

	/* deal with chardevs and other special file */
	if (!mapping->a_ops->writepage)
		return 0;

	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;
	}

	if (!S_ISDIR(inode->i_mode)) {
		mutex_lock(&sbi->writepages);
		locked = true;
	}
	ret = write_cache_pages(mapping, wbc, __f2fs_writepage, mapping);
	if (locked)
		mutex_unlock(&sbi->writepages);

	f2fs_submit_merged_bio(sbi, DATA, WRITE);

	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);
repeat:
	err = f2fs_convert_inline_data(inode, pos + len);
	if (err)
		return err;

	page = grab_cache_page_write_begin(mapping, index, flags);
	if (!page)
		return -ENOMEM;
	*pagep = page;

	if (f2fs_has_inline_data(inode) && (pos + len) <= MAX_INLINE_DATA)
		goto inline_data;

	f2fs_lock_op(sbi);
	set_new_dnode(&dn, inode, NULL, NULL, 0);
	err = f2fs_reserve_block(&dn, index);
	f2fs_unlock_op(sbi);

	if (err) {
		f2fs_put_page(page, 1);
		return err;
	}
inline_data:
	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);
		goto out;
	}

	if (dn.data_blkaddr == NEW_ADDR) {
		zero_user_segment(page, 0, PAGE_CACHE_SIZE);
	} else {
		if (f2fs_has_inline_data(inode))
			err = f2fs_read_inline_data(inode, page);
		else
			err = f2fs_submit_page_bio(sbi, page, dn.data_blkaddr,
							READ_SYNC);
		if (err)
			return err;
		lock_page(page);
		if (unlikely(!PageUptodate(page))) {
			f2fs_put_page(page, 1);
			return -EIO;
		}
		if (unlikely(page->mapping != mapping)) {
			f2fs_put_page(page, 1);
			goto repeat;
		}
	}
out:
	SetPageUptodate(page);
	clear_cold_data(page);
	return 0;
}

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);
	}

	f2fs_put_page(page, 1);
	return copied;
}

static int check_direct_IO(struct inode *inode, int rw,
		const struct iovec *iov, loff_t offset, unsigned long nr_segs)
{
	unsigned blocksize_mask = inode->i_sb->s_blocksize - 1;
	int i;

	if (rw == READ)
		return 0;

	if (offset & blocksize_mask)
		return -EINVAL;

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

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;

	/* Let buffer I/O handle the inline data case. */
	if (f2fs_has_inline_data(inode))
		return 0;

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

	return blockdev_direct_IO(rw, iocb, inode, iov, offset, nr_segs,
							get_data_block);
}

static void f2fs_invalidate_data_page(struct page *page, unsigned int offset,
				      unsigned int length)
{
	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);
	return 1;
}

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

	trace_f2fs_set_page_dirty(page, DATA);

	SetPageUptodate(page);
	if (!PageDirty(page)) {
		__set_page_dirty_nobuffers(page);
		set_dirty_dir_page(inode, page);
		return 1;
	}
	return 0;
}

static sector_t f2fs_bmap(struct address_space *mapping, sector_t block)
{
	return generic_block_bmap(mapping, block, get_data_block);
}

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,
	.write_end	= f2fs_write_end,
	.set_page_dirty	= f2fs_set_data_page_dirty,
	.invalidatepage	= f2fs_invalidate_data_page,
	.releasepage	= f2fs_release_data_page,
	.direct_IO	= f2fs_direct_IO,
	.bmap		= f2fs_bmap,
};