[deliverable/linux.git] / fs / f2fs / recovery.c

/*
 * fs/f2fs/recovery.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 "f2fs.h"
#include "node.h"
#include "segment.h"

static struct kmem_cache *fsync_entry_slab;

bool space_for_roll_forward(struct f2fs_sb_info *sbi)
{
	if (sbi->last_valid_block_count + sbi->alloc_valid_block_count
			> sbi->user_block_count)
		return false;
	return true;
}

static struct fsync_inode_entry *get_fsync_inode(struct list_head *head,
								nid_t ino)
{
	struct fsync_inode_entry *entry;

	list_for_each_entry(entry, head, list)
		if (entry->inode->i_ino == ino)
			return entry;

	return NULL;
}

static int recover_dentry(struct page *ipage, struct inode *inode)
{
	struct f2fs_inode *raw_inode = F2FS_INODE(ipage);
	nid_t pino = le32_to_cpu(raw_inode->i_pino);
	struct f2fs_dir_entry *de;
	struct qstr name;
	struct page *page;
	struct inode *dir, *einode;
	int err = 0;

	dir = f2fs_iget(inode->i_sb, pino);
	if (IS_ERR(dir)) {
		err = PTR_ERR(dir);
		goto out;
	}

	name.len = le32_to_cpu(raw_inode->i_namelen);
	name.name = raw_inode->i_name;

	if (unlikely(name.len > F2FS_NAME_LEN)) {
		WARN_ON(1);
		err = -ENAMETOOLONG;
		goto out_err;
	}
retry:
	de = f2fs_find_entry(dir, &name, &page);
	if (de && inode->i_ino == le32_to_cpu(de->ino)) {
		clear_inode_flag(F2FS_I(inode), FI_INC_LINK);
		goto out_unmap_put;
	}
	if (de) {
		einode = f2fs_iget(inode->i_sb, le32_to_cpu(de->ino));
		if (IS_ERR(einode)) {
			WARN_ON(1);
			err = PTR_ERR(einode);
			if (err == -ENOENT)
				err = -EEXIST;
			goto out_unmap_put;
		}
		err = acquire_orphan_inode(F2FS_SB(inode->i_sb));
		if (err) {
			iput(einode);
			goto out_unmap_put;
		}
		f2fs_delete_entry(de, page, einode);
		iput(einode);
		goto retry;
	}
	err = __f2fs_add_link(dir, &name, inode);
	if (err)
		goto out_err;

	if (is_inode_flag_set(F2FS_I(dir), FI_DELAY_IPUT)) {
		iput(dir);
	} else {
		add_dirty_dir_inode(dir);
		set_inode_flag(F2FS_I(dir), FI_DELAY_IPUT);
	}

	goto out;

out_unmap_put:
	kunmap(page);
	f2fs_put_page(page, 0);
out_err:
	iput(dir);
out:
	f2fs_msg(inode->i_sb, KERN_NOTICE,
			"%s: ino = %x, name = %s, dir = %lx, err = %d",
			__func__, ino_of_node(ipage), raw_inode->i_name,
			IS_ERR(dir) ? 0 : dir->i_ino, err);
	return err;
}

static int recover_inode(struct inode *inode, struct page *node_page)
{
	struct f2fs_inode *raw_inode = F2FS_INODE(node_page);

	if (!IS_INODE(node_page))
		return 0;

	inode->i_mode = le16_to_cpu(raw_inode->i_mode);
	i_size_write(inode, le64_to_cpu(raw_inode->i_size));
	inode->i_atime.tv_sec = le64_to_cpu(raw_inode->i_mtime);
	inode->i_ctime.tv_sec = le64_to_cpu(raw_inode->i_ctime);
	inode->i_mtime.tv_sec = le64_to_cpu(raw_inode->i_mtime);
	inode->i_atime.tv_nsec = le32_to_cpu(raw_inode->i_mtime_nsec);
	inode->i_ctime.tv_nsec = le32_to_cpu(raw_inode->i_ctime_nsec);
	inode->i_mtime.tv_nsec = le32_to_cpu(raw_inode->i_mtime_nsec);

	if (is_dent_dnode(node_page))
		return recover_dentry(node_page, inode);

	f2fs_msg(inode->i_sb, KERN_NOTICE, "recover_inode: ino = %x, name = %s",
			ino_of_node(node_page), raw_inode->i_name);
	return 0;
}

static int find_fsync_dnodes(struct f2fs_sb_info *sbi, struct list_head *head)
{
	unsigned long long cp_ver = cur_cp_version(F2FS_CKPT(sbi));
	struct curseg_info *curseg;
	struct page *page;
	block_t blkaddr;
	int err = 0;

	/* get node pages in the current segment */
	curseg = CURSEG_I(sbi, CURSEG_WARM_NODE);
	blkaddr = NEXT_FREE_BLKADDR(sbi, curseg);

	/* read node page */
	page = alloc_page(GFP_F2FS_ZERO);
	if (!page)
		return -ENOMEM;
	lock_page(page);

	while (1) {
		struct fsync_inode_entry *entry;

		err = f2fs_submit_page_bio(sbi, page, blkaddr, READ_SYNC);
		if (err)
			return err;

		lock_page(page);

		if (cp_ver != cpver_of_node(page))
			break;

		if (!is_fsync_dnode(page))
			goto next;

		entry = get_fsync_inode(head, ino_of_node(page));
		if (entry) {
			if (IS_INODE(page) && is_dent_dnode(page))
				set_inode_flag(F2FS_I(entry->inode),
							FI_INC_LINK);
		} else {
			if (IS_INODE(page) && is_dent_dnode(page)) {
				err = recover_inode_page(sbi, page);
				if (err)
					break;
			}

			/* add this fsync inode to the list */
			entry = kmem_cache_alloc(fsync_entry_slab, GFP_NOFS);
			if (!entry) {
				err = -ENOMEM;
				break;
			}

			entry->inode = f2fs_iget(sbi->sb, ino_of_node(page));
			if (IS_ERR(entry->inode)) {
				err = PTR_ERR(entry->inode);
				kmem_cache_free(fsync_entry_slab, entry);
				break;
			}
			list_add_tail(&entry->list, head);
		}
		entry->blkaddr = blkaddr;

		err = recover_inode(entry->inode, page);
		if (err && err != -ENOENT)
			break;
next:
		/* check next segment */
		blkaddr = next_blkaddr_of_node(page);
	}

	unlock_page(page);
	__free_pages(page, 0);

	return err;
}

static void destroy_fsync_dnodes(struct list_head *head)
{
	struct fsync_inode_entry *entry, *tmp;

	list_for_each_entry_safe(entry, tmp, head, list) {
		iput(entry->inode);
		list_del(&entry->list);
		kmem_cache_free(fsync_entry_slab, entry);
	}
}

static int check_index_in_prev_nodes(struct f2fs_sb_info *sbi,
			block_t blkaddr, struct dnode_of_data *dn)
{
	struct seg_entry *sentry;
	unsigned int segno = GET_SEGNO(sbi, blkaddr);
	unsigned short blkoff = GET_BLKOFF_FROM_SEG0(sbi, blkaddr);
	struct f2fs_summary_block *sum_node;
	struct f2fs_summary sum;
	struct page *sum_page, *node_page;
	nid_t ino, nid;
	struct inode *inode;
	unsigned int offset;
	block_t bidx;
	int i;

	sentry = get_seg_entry(sbi, segno);
	if (!f2fs_test_bit(blkoff, sentry->cur_valid_map))
		return 0;

	/* Get the previous summary */
	for (i = CURSEG_WARM_DATA; i <= CURSEG_COLD_DATA; i++) {
		struct curseg_info *curseg = CURSEG_I(sbi, i);
		if (curseg->segno == segno) {
			sum = curseg->sum_blk->entries[blkoff];
			goto got_it;
		}
	}

	sum_page = get_sum_page(sbi, segno);
	sum_node = (struct f2fs_summary_block *)page_address(sum_page);
	sum = sum_node->entries[blkoff];
	f2fs_put_page(sum_page, 1);
got_it:
	/* Use the locked dnode page and inode */
	nid = le32_to_cpu(sum.nid);
	if (dn->inode->i_ino == nid) {
		struct dnode_of_data tdn = *dn;
		tdn.nid = nid;
		tdn.node_page = dn->inode_page;
		tdn.ofs_in_node = le16_to_cpu(sum.ofs_in_node);
		truncate_data_blocks_range(&tdn, 1);
		return 0;
	} else if (dn->nid == nid) {
		struct dnode_of_data tdn = *dn;
		tdn.ofs_in_node = le16_to_cpu(sum.ofs_in_node);
		truncate_data_blocks_range(&tdn, 1);
		return 0;
	}

	/* Get the node page */
	node_page = get_node_page(sbi, nid);
	if (IS_ERR(node_page))
		return PTR_ERR(node_page);

	offset = ofs_of_node(node_page);
	ino = ino_of_node(node_page);
	f2fs_put_page(node_page, 1);

	/* Deallocate previous index in the node page */
	inode = f2fs_iget(sbi->sb, ino);
	if (IS_ERR(inode))
		return PTR_ERR(inode);

	bidx = start_bidx_of_node(offset, F2FS_I(inode)) +
					le16_to_cpu(sum.ofs_in_node);

	truncate_hole(inode, bidx, bidx + 1);
	iput(inode);
	return 0;
}

static int do_recover_data(struct f2fs_sb_info *sbi, struct inode *inode,
					struct page *page, block_t blkaddr)
{
	struct f2fs_inode_info *fi = F2FS_I(inode);
	unsigned int start, end;
	struct dnode_of_data dn;
	struct f2fs_summary sum;
	struct node_info ni;
	int err = 0, recovered = 0;

	/* step 1: recover xattr */
	if (IS_INODE(page)) {
		recover_inline_xattr(inode, page);
	} else if (f2fs_has_xattr_block(ofs_of_node(page))) {
		recover_xattr_data(inode, page, blkaddr);
		goto out;
	}

	/* step 2: recover inline data */
	if (recover_inline_data(inode, page))
		goto out;

	/* step 3: recover data indices */
	start = start_bidx_of_node(ofs_of_node(page), fi);
	end = start + ADDRS_PER_PAGE(page, fi);

	f2fs_lock_op(sbi);

	set_new_dnode(&dn, inode, NULL, NULL, 0);

	err = get_dnode_of_data(&dn, start, ALLOC_NODE);
	if (err) {
		f2fs_unlock_op(sbi);
		goto out;
	}

	f2fs_wait_on_page_writeback(dn.node_page, NODE);

	get_node_info(sbi, dn.nid, &ni);
	f2fs_bug_on(ni.ino != ino_of_node(page));
	f2fs_bug_on(ofs_of_node(dn.node_page) != ofs_of_node(page));

	for (; start < end; start++) {
		block_t src, dest;

		src = datablock_addr(dn.node_page, dn.ofs_in_node);
		dest = datablock_addr(page, dn.ofs_in_node);

		if (src != dest && dest != NEW_ADDR && dest != NULL_ADDR) {
			if (src == NULL_ADDR) {
				err = reserve_new_block(&dn);
				/* We should not get -ENOSPC */
				f2fs_bug_on(err);
			}

			/* Check the previous node page having this index */
			err = check_index_in_prev_nodes(sbi, dest, &dn);
			if (err)
				goto err;

			set_summary(&sum, dn.nid, dn.ofs_in_node, ni.version);

			/* write dummy data page */
			recover_data_page(sbi, NULL, &sum, src, dest);
			update_extent_cache(dest, &dn);
			recovered++;
		}
		dn.ofs_in_node++;
	}

	/* write node page in place */
	set_summary(&sum, dn.nid, 0, 0);
	if (IS_INODE(dn.node_page))
		sync_inode_page(&dn);

	copy_node_footer(dn.node_page, page);
	fill_node_footer(dn.node_page, dn.nid, ni.ino,
					ofs_of_node(page), false);
	set_page_dirty(dn.node_page);
err:
	f2fs_put_dnode(&dn);
	f2fs_unlock_op(sbi);
out:
	f2fs_msg(sbi->sb, KERN_NOTICE,
		"recover_data: ino = %lx, recovered = %d blocks, err = %d",
		inode->i_ino, recovered, err);
	return err;
}

static int recover_data(struct f2fs_sb_info *sbi,
				struct list_head *head, int type)
{
	unsigned long long cp_ver = cur_cp_version(F2FS_CKPT(sbi));
	struct curseg_info *curseg;
	struct page *page;
	int err = 0;
	block_t blkaddr;

	/* get node pages in the current segment */
	curseg = CURSEG_I(sbi, type);
	blkaddr = NEXT_FREE_BLKADDR(sbi, curseg);

	/* read node page */
	page = alloc_page(GFP_F2FS_ZERO);
	if (!page)
		return -ENOMEM;

	lock_page(page);

	while (1) {
		struct fsync_inode_entry *entry;

		err = f2fs_submit_page_bio(sbi, page, blkaddr, READ_SYNC);
		if (err)
			return err;

		lock_page(page);

		if (cp_ver != cpver_of_node(page))
			break;

		entry = get_fsync_inode(head, ino_of_node(page));
		if (!entry)
			goto next;

		err = do_recover_data(sbi, entry->inode, page, blkaddr);
		if (err)
			break;

		if (entry->blkaddr == blkaddr) {
			iput(entry->inode);
			list_del(&entry->list);
			kmem_cache_free(fsync_entry_slab, entry);
		}
next:
		/* check next segment */
		blkaddr = next_blkaddr_of_node(page);
	}

	unlock_page(page);
	__free_pages(page, 0);

	if (!err)
		allocate_new_segments(sbi);
	return err;
}

int recover_fsync_data(struct f2fs_sb_info *sbi)
{
	struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_WARM_NODE);
	struct list_head inode_list;
	block_t blkaddr;
	int err;
	bool need_writecp = false;

	fsync_entry_slab = f2fs_kmem_cache_create("f2fs_fsync_inode_entry",
			sizeof(struct fsync_inode_entry));
	if (!fsync_entry_slab)
		return -ENOMEM;

	INIT_LIST_HEAD(&inode_list);

	/* step #1: find fsynced inode numbers */
	sbi->por_doing = true;

	/* prevent checkpoint */
	mutex_lock(&sbi->cp_mutex);

	blkaddr = NEXT_FREE_BLKADDR(sbi, curseg);

	err = find_fsync_dnodes(sbi, &inode_list);
	if (err)
		goto out;

	if (list_empty(&inode_list))
		goto out;

	need_writecp = true;

	/* step #2: recover data */
	err = recover_data(sbi, &inode_list, CURSEG_WARM_NODE);
	if (!err)
		f2fs_bug_on(!list_empty(&inode_list));
out:
	destroy_fsync_dnodes(&inode_list);
	kmem_cache_destroy(fsync_entry_slab);

	if (err) {
		truncate_inode_pages_final(NODE_MAPPING(sbi));
		truncate_inode_pages_final(META_MAPPING(sbi));
	}

	sbi->por_doing = false;
	if (err) {
		discard_next_dnode(sbi, blkaddr);

		/* Flush all the NAT/SIT pages */
		while (get_pages(sbi, F2FS_DIRTY_META))
			sync_meta_pages(sbi, META, LONG_MAX);
		set_ckpt_flags(sbi->ckpt, CP_ERROR_FLAG);
		mutex_unlock(&sbi->cp_mutex);
	} else if (need_writecp) {
		mutex_unlock(&sbi->cp_mutex);
		write_checkpoint(sbi, false);
	} else {
		mutex_unlock(&sbi->cp_mutex);
	}
	return err;
}
Commit	Line	Data
0a8165d7	1	/*
d624c96f JK	2	* fs/f2fs/recovery.c
	3	*
	4	* Copyright (c) 2012 Samsung Electronics Co., Ltd.
	5	* http://www.samsung.com/
	6	*
	7	* This program is free software; you can redistribute it and/or modify
	8	* it under the terms of the GNU General Public License version 2 as
	9	* published by the Free Software Foundation.
	10	*/
	11	#include <linux/fs.h>
	12	#include <linux/f2fs_fs.h>
	13	#include "f2fs.h"
	14	#include "node.h"
	15	#include "segment.h"
	16
	17	static struct kmem_cache *fsync_entry_slab;
	18
	19	bool space_for_roll_forward(struct f2fs_sb_info *sbi)
	20	{
	21	if (sbi->last_valid_block_count + sbi->alloc_valid_block_count
	22	> sbi->user_block_count)
	23	return false;
	24	return true;
	25	}
	26
	27	static struct fsync_inode_entry get_fsync_inode(struct list_head head,
	28	nid_t ino)
	29	{
d624c96f JK	30	struct fsync_inode_entry *entry;
d624c96f JK	31
2d7b822a	32	list_for_each_entry(entry, head, list)
d624c96f JK	33	if (entry->inode->i_ino == ino)
d624c96f JK	34	return entry;
2d7b822a	35
d624c96f JK	36	return NULL;
	37	}
	38
	39	static int recover_dentry(struct page ipage, struct inode inode)
	40	{
58bfaf44	41	struct f2fs_inode *raw_inode = F2FS_INODE(ipage);
74d0b917	42	nid_t pino = le32_to_cpu(raw_inode->i_pino);
6b8213d9	43	struct f2fs_dir_entry *de;
b7f7a5e0	44	struct qstr name;
d624c96f	45	struct page *page;
6b8213d9	46	struct inode dir, einode;
d624c96f JK	47	int err = 0;
d624c96f JK	48
ed57c27f JK	49	dir = f2fs_iget(inode->i_sb, pino);
	50	if (IS_ERR(dir)) {
	51	err = PTR_ERR(dir);
	52	goto out;
	53	}
	54
b7f7a5e0 AV	55	name.len = le32_to_cpu(raw_inode->i_namelen);
b7f7a5e0 AV	56	name.name = raw_inode->i_name;
d96b1431 CY	57
	58	if (unlikely(name.len > F2FS_NAME_LEN)) {
	59	WARN_ON(1);
	60	err = -ENAMETOOLONG;
86928f98	61	goto out_err;
d96b1431	62	}
6b8213d9 JK	63	retry:
6b8213d9 JK	64	de = f2fs_find_entry(dir, &name, &page);
695facc0 JK	65	if (de && inode->i_ino == le32_to_cpu(de->ino)) {
695facc0 JK	66	clear_inode_flag(F2FS_I(inode), FI_INC_LINK);
2e5558f4	67	goto out_unmap_put;
695facc0	68	}
6b8213d9 JK	69	if (de) {
	70	einode = f2fs_iget(inode->i_sb, le32_to_cpu(de->ino));
	71	if (IS_ERR(einode)) {
	72	WARN_ON(1);
5c1f9927 CY	73	err = PTR_ERR(einode);
5c1f9927 CY	74	if (err == -ENOENT)
6b8213d9	75	err = -EEXIST;
2e5558f4 RK	76	goto out_unmap_put;
	77	}
	78	err = acquire_orphan_inode(F2FS_SB(inode->i_sb));
	79	if (err) {
	80	iput(einode);
	81	goto out_unmap_put;
6b8213d9 JK	82	}
	83	f2fs_delete_entry(de, page, einode);
	84	iput(einode);
	85	goto retry;
d624c96f	86	}
6b8213d9	87	err = __f2fs_add_link(dir, &name, inode);
86928f98 JK	88	if (err)
	89	goto out_err;
	90
	91	if (is_inode_flag_set(F2FS_I(dir), FI_DELAY_IPUT)) {
	92	iput(dir);
	93	} else {
	94	add_dirty_dir_inode(dir);
	95	set_inode_flag(F2FS_I(dir), FI_DELAY_IPUT);
	96	}
	97
2e5558f4 RK	98	goto out;
	99
	100	out_unmap_put:
	101	kunmap(page);
	102	f2fs_put_page(page, 0);
86928f98 JK	103	out_err:
86928f98 JK	104	iput(dir);
d624c96f	105	out:
6c311ec6 CF	106	f2fs_msg(inode->i_sb, KERN_NOTICE,
	107	"%s: ino = %x, name = %s, dir = %lx, err = %d",
	108	__func__, ino_of_node(ipage), raw_inode->i_name,
f28c06fa	109	IS_ERR(dir) ? 0 : dir->i_ino, err);
d624c96f JK	110	return err;
	111	}
	112
	113	static int recover_inode(struct inode inode, struct page node_page)
	114	{
58bfaf44	115	struct f2fs_inode *raw_inode = F2FS_INODE(node_page);
d624c96f	116
f356fe0c JK	117	if (!IS_INODE(node_page))
	118	return 0;
	119
25ca923b	120	inode->i_mode = le16_to_cpu(raw_inode->i_mode);
d624c96f JK	121	i_size_write(inode, le64_to_cpu(raw_inode->i_size));
	122	inode->i_atime.tv_sec = le64_to_cpu(raw_inode->i_mtime);
	123	inode->i_ctime.tv_sec = le64_to_cpu(raw_inode->i_ctime);
	124	inode->i_mtime.tv_sec = le64_to_cpu(raw_inode->i_mtime);
	125	inode->i_atime.tv_nsec = le32_to_cpu(raw_inode->i_mtime_nsec);
	126	inode->i_ctime.tv_nsec = le32_to_cpu(raw_inode->i_ctime_nsec);
	127	inode->i_mtime.tv_nsec = le32_to_cpu(raw_inode->i_mtime_nsec);
	128
f356fe0c JK	129	if (is_dent_dnode(node_page))
	130	return recover_dentry(node_page, inode);
	131
	132	f2fs_msg(inode->i_sb, KERN_NOTICE, "recover_inode: ino = %x, name = %s",
	133	ino_of_node(node_page), raw_inode->i_name);
	134	return 0;
d624c96f JK	135	}
	136
	137	static int find_fsync_dnodes(struct f2fs_sb_info sbi, struct list_head head)
	138	{
d71b5564	139	unsigned long long cp_ver = cur_cp_version(F2FS_CKPT(sbi));
d624c96f JK	140	struct curseg_info *curseg;
	141	struct page *page;
	142	block_t blkaddr;
	143	int err = 0;
	144
	145	/* get node pages in the current segment */
	146	curseg = CURSEG_I(sbi, CURSEG_WARM_NODE);
695fd1ed	147	blkaddr = NEXT_FREE_BLKADDR(sbi, curseg);
d624c96f JK	148
	149	/* read node page */
	150	page = alloc_page(GFP_F2FS_ZERO);
e27dae4d DC	151	if (!page)
e27dae4d DC	152	return -ENOMEM;
d624c96f JK	153	lock_page(page);
	154
	155	while (1) {
	156	struct fsync_inode_entry *entry;
	157
93dfe2ac	158	err = f2fs_submit_page_bio(sbi, page, blkaddr, READ_SYNC);
393ff91f	159	if (err)
b9987a27	160	return err;
d624c96f	161
393ff91f JK	162	lock_page(page);
393ff91f JK	163
6ead1142	164	if (cp_ver != cpver_of_node(page))
f356fe0c	165	break;
d624c96f JK	166
	167	if (!is_fsync_dnode(page))
	168	goto next;
	169
	170	entry = get_fsync_inode(head, ino_of_node(page));
	171	if (entry) {
d624c96f JK	172	if (IS_INODE(page) && is_dent_dnode(page))
	173	set_inode_flag(F2FS_I(entry->inode),
	174	FI_INC_LINK);
	175	} else {
	176	if (IS_INODE(page) && is_dent_dnode(page)) {
6ead1142 JK	177	err = recover_inode_page(sbi, page);
6ead1142 JK	178	if (err)
f356fe0c	179	break;
d624c96f JK	180	}
	181
	182	/* add this fsync inode to the list */
	183	entry = kmem_cache_alloc(fsync_entry_slab, GFP_NOFS);
	184	if (!entry) {
	185	err = -ENOMEM;
f356fe0c	186	break;
d624c96f JK	187	}
d624c96f JK	188
d624c96f JK	189	entry->inode = f2fs_iget(sbi->sb, ino_of_node(page));
	190	if (IS_ERR(entry->inode)) {
	191	err = PTR_ERR(entry->inode);
fd8bb65f	192	kmem_cache_free(fsync_entry_slab, entry);
f356fe0c	193	break;
d624c96f	194	}
fd8bb65f	195	list_add_tail(&entry->list, head);
d624c96f	196	}
addbe45b JK	197	entry->blkaddr = blkaddr;
addbe45b JK	198
f356fe0c JK	199	err = recover_inode(entry->inode, page);
	200	if (err && err != -ENOENT)
	201	break;
d624c96f JK	202	next:
	203	/* check next segment */
	204	blkaddr = next_blkaddr_of_node(page);
d624c96f	205	}
b9987a27	206
d624c96f JK	207	unlock_page(page);
d624c96f JK	208	__free_pages(page, 0);
b9987a27	209
d624c96f JK	210	return err;
	211	}
	212
5ebefc5b	213	static void destroy_fsync_dnodes(struct list_head *head)
d624c96f	214	{
d8b79b2f DC	215	struct fsync_inode_entry entry, tmp;
	216
	217	list_for_each_entry_safe(entry, tmp, head, list) {
d624c96f JK	218	iput(entry->inode);
	219	list_del(&entry->list);
	220	kmem_cache_free(fsync_entry_slab, entry);
	221	}
	222	}
	223
39cf72cf	224	static int check_index_in_prev_nodes(struct f2fs_sb_info *sbi,
b292dcab	225	block_t blkaddr, struct dnode_of_data *dn)
d624c96f JK	226	{
	227	struct seg_entry *sentry;
	228	unsigned int segno = GET_SEGNO(sbi, blkaddr);
491c0854	229	unsigned short blkoff = GET_BLKOFF_FROM_SEG0(sbi, blkaddr);
f6517cfc	230	struct f2fs_summary_block *sum_node;
d624c96f	231	struct f2fs_summary sum;
f6517cfc	232	struct page sum_page, node_page;
b292dcab	233	nid_t ino, nid;
d624c96f	234	struct inode *inode;
de93653f	235	unsigned int offset;
d624c96f JK	236	block_t bidx;
	237	int i;
	238
	239	sentry = get_seg_entry(sbi, segno);
	240	if (!f2fs_test_bit(blkoff, sentry->cur_valid_map))
39cf72cf	241	return 0;
d624c96f JK	242
	243	/* Get the previous summary */
	244	for (i = CURSEG_WARM_DATA; i <= CURSEG_COLD_DATA; i++) {
	245	struct curseg_info *curseg = CURSEG_I(sbi, i);
	246	if (curseg->segno == segno) {
	247	sum = curseg->sum_blk->entries[blkoff];
f6517cfc	248	goto got_it;
d624c96f JK	249	}
d624c96f JK	250	}
d624c96f	251
f6517cfc JK	252	sum_page = get_sum_page(sbi, segno);
	253	sum_node = (struct f2fs_summary_block *)page_address(sum_page);
	254	sum = sum_node->entries[blkoff];
	255	f2fs_put_page(sum_page, 1);
	256	got_it:
b292dcab JK	257	/* Use the locked dnode page and inode */
	258	nid = le32_to_cpu(sum.nid);
	259	if (dn->inode->i_ino == nid) {
	260	struct dnode_of_data tdn = *dn;
	261	tdn.nid = nid;
	262	tdn.node_page = dn->inode_page;
060dd67b	263	tdn.ofs_in_node = le16_to_cpu(sum.ofs_in_node);
b292dcab	264	truncate_data_blocks_range(&tdn, 1);
39cf72cf	265	return 0;
b292dcab JK	266	} else if (dn->nid == nid) {
b292dcab JK	267	struct dnode_of_data tdn = *dn;
060dd67b	268	tdn.ofs_in_node = le16_to_cpu(sum.ofs_in_node);
b292dcab	269	truncate_data_blocks_range(&tdn, 1);
39cf72cf	270	return 0;
b292dcab JK	271	}
b292dcab JK	272
d624c96f	273	/* Get the node page */
b292dcab	274	node_page = get_node_page(sbi, nid);
39cf72cf JK	275	if (IS_ERR(node_page))
39cf72cf JK	276	return PTR_ERR(node_page);
de93653f JK	277
de93653f JK	278	offset = ofs_of_node(node_page);
d624c96f JK	279	ino = ino_of_node(node_page);
	280	f2fs_put_page(node_page, 1);
	281
	282	/* Deallocate previous index in the node page */
d4686d56	283	inode = f2fs_iget(sbi->sb, ino);
06025f4d	284	if (IS_ERR(inode))
39cf72cf	285	return PTR_ERR(inode);
06025f4d	286
de93653f JK	287	bidx = start_bidx_of_node(offset, F2FS_I(inode)) +
	288	le16_to_cpu(sum.ofs_in_node);
	289
d624c96f JK	290	truncate_hole(inode, bidx, bidx + 1);
d624c96f JK	291	iput(inode);
39cf72cf	292	return 0;
d624c96f JK	293	}
d624c96f JK	294
6ead1142	295	static int do_recover_data(struct f2fs_sb_info sbi, struct inode inode,
d624c96f JK	296	struct page *page, block_t blkaddr)
d624c96f JK	297	{
de93653f	298	struct f2fs_inode_info *fi = F2FS_I(inode);
d624c96f JK	299	unsigned int start, end;
	300	struct dnode_of_data dn;
	301	struct f2fs_summary sum;
	302	struct node_info ni;
f356fe0c	303	int err = 0, recovered = 0;
d624c96f	304
1c35a90e JK	305	/* step 1: recover xattr */
	306	if (IS_INODE(page)) {
	307	recover_inline_xattr(inode, page);
	308	} else if (f2fs_has_xattr_block(ofs_of_node(page))) {
	309	recover_xattr_data(inode, page, blkaddr);
1e1bb4ba	310	goto out;
1c35a90e	311	}
1e1bb4ba	312
1c35a90e JK	313	/* step 2: recover inline data */
1c35a90e JK	314	if (recover_inline_data(inode, page))
abb2366c JK	315	goto out;
abb2366c JK	316
1c35a90e	317	/* step 3: recover data indices */
de93653f	318	start = start_bidx_of_node(ofs_of_node(page), fi);
6403eb1f	319	end = start + ADDRS_PER_PAGE(page, fi);
d624c96f	320
e479556b	321	f2fs_lock_op(sbi);
1e1bb4ba	322
d624c96f	323	set_new_dnode(&dn, inode, NULL, NULL, 0);
39936837	324
6ead1142	325	err = get_dnode_of_data(&dn, start, ALLOC_NODE);
39936837	326	if (err) {
e479556b	327	f2fs_unlock_op(sbi);
1e1bb4ba	328	goto out;
39936837	329	}
d624c96f	330
3cb5ad15	331	f2fs_wait_on_page_writeback(dn.node_page, NODE);
d624c96f JK	332
d624c96f JK	333	get_node_info(sbi, dn.nid, &ni);
5d56b671 JK	334	f2fs_bug_on(ni.ino != ino_of_node(page));
5d56b671 JK	335	f2fs_bug_on(ofs_of_node(dn.node_page) != ofs_of_node(page));
d624c96f JK	336
	337	for (; start < end; start++) {
	338	block_t src, dest;
	339
	340	src = datablock_addr(dn.node_page, dn.ofs_in_node);
	341	dest = datablock_addr(page, dn.ofs_in_node);
	342
	343	if (src != dest && dest != NEW_ADDR && dest != NULL_ADDR) {
	344	if (src == NULL_ADDR) {
5d56b671	345	err = reserve_new_block(&dn);
d624c96f	346	/* We should not get -ENOSPC */
5d56b671	347	f2fs_bug_on(err);
d624c96f JK	348	}
	349
	350	/* Check the previous node page having this index */
39cf72cf JK	351	err = check_index_in_prev_nodes(sbi, dest, &dn);
	352	if (err)
	353	goto err;
d624c96f JK	354
	355	set_summary(&sum, dn.nid, dn.ofs_in_node, ni.version);
	356
	357	/* write dummy data page */
	358	recover_data_page(sbi, NULL, &sum, src, dest);
	359	update_extent_cache(dest, &dn);
f356fe0c	360	recovered++;
d624c96f JK	361	}
	362	dn.ofs_in_node++;
	363	}
	364
	365	/* write node page in place */
	366	set_summary(&sum, dn.nid, 0, 0);
	367	if (IS_INODE(dn.node_page))
	368	sync_inode_page(&dn);
	369
	370	copy_node_footer(dn.node_page, page);
	371	fill_node_footer(dn.node_page, dn.nid, ni.ino,
	372	ofs_of_node(page), false);
	373	set_page_dirty(dn.node_page);
39cf72cf	374	err:
d624c96f	375	f2fs_put_dnode(&dn);
e479556b	376	f2fs_unlock_op(sbi);
1e1bb4ba	377	out:
6c311ec6 CF	378	f2fs_msg(sbi->sb, KERN_NOTICE,
	379	"recover_data: ino = %lx, recovered = %d blocks, err = %d",
	380	inode->i_ino, recovered, err);
39cf72cf	381	return err;
d624c96f JK	382	}
d624c96f JK	383
6ead1142	384	static int recover_data(struct f2fs_sb_info *sbi,
d624c96f JK	385	struct list_head *head, int type)
d624c96f JK	386	{
d71b5564	387	unsigned long long cp_ver = cur_cp_version(F2FS_CKPT(sbi));
d624c96f JK	388	struct curseg_info *curseg;
d624c96f JK	389	struct page *page;
6ead1142	390	int err = 0;
d624c96f JK	391	block_t blkaddr;
	392
	393	/* get node pages in the current segment */
	394	curseg = CURSEG_I(sbi, type);
	395	blkaddr = NEXT_FREE_BLKADDR(sbi, curseg);
	396
	397	/* read node page */
a0acdfe0	398	page = alloc_page(GFP_F2FS_ZERO);
e27dae4d	399	if (!page)
6ead1142 JK	400	return -ENOMEM;
6ead1142 JK	401
d624c96f JK	402	lock_page(page);
	403
	404	while (1) {
	405	struct fsync_inode_entry *entry;
	406
93dfe2ac	407	err = f2fs_submit_page_bio(sbi, page, blkaddr, READ_SYNC);
6ead1142	408	if (err)
b9987a27	409	return err;
d624c96f	410
393ff91f JK	411	lock_page(page);
393ff91f JK	412
d624c96f	413	if (cp_ver != cpver_of_node(page))
45856aff	414	break;
d624c96f JK	415
	416	entry = get_fsync_inode(head, ino_of_node(page));
	417	if (!entry)
	418	goto next;
	419
6ead1142 JK	420	err = do_recover_data(sbi, entry->inode, page, blkaddr);
6ead1142 JK	421	if (err)
45856aff	422	break;
d624c96f JK	423
	424	if (entry->blkaddr == blkaddr) {
	425	iput(entry->inode);
	426	list_del(&entry->list);
	427	kmem_cache_free(fsync_entry_slab, entry);
	428	}
	429	next:
	430	/* check next segment */
	431	blkaddr = next_blkaddr_of_node(page);
d624c96f	432	}
b9987a27	433
d624c96f JK	434	unlock_page(page);
	435	__free_pages(page, 0);
	436
6ead1142 JK	437	if (!err)
	438	allocate_new_segments(sbi);
	439	return err;
d624c96f JK	440	}
d624c96f JK	441
6ead1142	442	int recover_fsync_data(struct f2fs_sb_info *sbi)
d624c96f	443	{
cf2271e7	444	struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_WARM_NODE);
d624c96f	445	struct list_head inode_list;
cf2271e7	446	block_t blkaddr;
6ead1142	447	int err;
aabe5136	448	bool need_writecp = false;
d624c96f JK	449
d624c96f JK	450	fsync_entry_slab = f2fs_kmem_cache_create("f2fs_fsync_inode_entry",
e8512d2e	451	sizeof(struct fsync_inode_entry));
6bacf52f	452	if (!fsync_entry_slab)
6ead1142	453	return -ENOMEM;
d624c96f JK	454
	455	INIT_LIST_HEAD(&inode_list);
	456
	457	/* step #1: find fsynced inode numbers */
aabe5136	458	sbi->por_doing = true;
cf2271e7	459
14f4e690 JK	460	/* prevent checkpoint */
	461	mutex_lock(&sbi->cp_mutex);
	462
cf2271e7 JK	463	blkaddr = NEXT_FREE_BLKADDR(sbi, curseg);
cf2271e7 JK	464
6ead1142 JK	465	err = find_fsync_dnodes(sbi, &inode_list);
6ead1142 JK	466	if (err)
d624c96f JK	467	goto out;
	468
	469	if (list_empty(&inode_list))
	470	goto out;
	471
aabe5136	472	need_writecp = true;
691c6fd2	473
d624c96f	474	/* step #2: recover data */
6ead1142	475	err = recover_data(sbi, &inode_list, CURSEG_WARM_NODE);
b307384e JK	476	if (!err)
b307384e JK	477	f2fs_bug_on(!list_empty(&inode_list));
d624c96f	478	out:
5ebefc5b	479	destroy_fsync_dnodes(&inode_list);
d624c96f	480	kmem_cache_destroy(fsync_entry_slab);
cf2271e7 JK	481
	482	if (err) {
	483	truncate_inode_pages_final(NODE_MAPPING(sbi));
	484	truncate_inode_pages_final(META_MAPPING(sbi));
	485	}
	486
aabe5136	487	sbi->por_doing = false;
cf2271e7 JK	488	if (err) {
	489	discard_next_dnode(sbi, blkaddr);
	490
	491	/* Flush all the NAT/SIT pages */
	492	while (get_pages(sbi, F2FS_DIRTY_META))
	493	sync_meta_pages(sbi, META, LONG_MAX);
14f4e690 JK	494	set_ckpt_flags(sbi->ckpt, CP_ERROR_FLAG);
14f4e690 JK	495	mutex_unlock(&sbi->cp_mutex);
cf2271e7	496	} else if (need_writecp) {
14f4e690	497	mutex_unlock(&sbi->cp_mutex);
2c2c149f	498	write_checkpoint(sbi, false);
14f4e690 JK	499	} else {
14f4e690 JK	500	mutex_unlock(&sbi->cp_mutex);
cf2271e7	501	}
6ead1142	502	return err;
d624c96f	503	}