linux/fs/f2fs/recovery.c
Jaegeuk Kim edc55aaf0d f2fs: avoid f2fs_bug_on if f2fs_get_meta_page_nofail got EIO
This patch avoids BUG_ON when f2fs_get_meta_page_nofail got EIO during
xfstests/generic/475.

Reviewed-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2018-09-28 10:39:58 -07:00

741 lines
17 KiB
C

// SPDX-License-Identifier: GPL-2.0
/*
* fs/f2fs/recovery.c
*
* Copyright (c) 2012 Samsung Electronics Co., Ltd.
* http://www.samsung.com/
*/
#include <linux/fs.h>
#include <linux/f2fs_fs.h>
#include "f2fs.h"
#include "node.h"
#include "segment.h"
/*
* Roll forward recovery scenarios.
*
* [Term] F: fsync_mark, D: dentry_mark
*
* 1. inode(x) | CP | inode(x) | dnode(F)
* -> Update the latest inode(x).
*
* 2. inode(x) | CP | inode(F) | dnode(F)
* -> No problem.
*
* 3. inode(x) | CP | dnode(F) | inode(x)
* -> Recover to the latest dnode(F), and drop the last inode(x)
*
* 4. inode(x) | CP | dnode(F) | inode(F)
* -> No problem.
*
* 5. CP | inode(x) | dnode(F)
* -> The inode(DF) was missing. Should drop this dnode(F).
*
* 6. CP | inode(DF) | dnode(F)
* -> No problem.
*
* 7. CP | dnode(F) | inode(DF)
* -> If f2fs_iget fails, then goto next to find inode(DF).
*
* 8. CP | dnode(F) | inode(x)
* -> If f2fs_iget fails, then goto next to find inode(DF).
* But it will fail due to no inode(DF).
*/
static struct kmem_cache *fsync_entry_slab;
bool f2fs_space_for_roll_forward(struct f2fs_sb_info *sbi)
{
s64 nalloc = percpu_counter_sum_positive(&sbi->alloc_valid_block_count);
if (sbi->last_valid_block_count + nalloc > 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 struct fsync_inode_entry *add_fsync_inode(struct f2fs_sb_info *sbi,
struct list_head *head, nid_t ino, bool quota_inode)
{
struct inode *inode;
struct fsync_inode_entry *entry;
int err;
inode = f2fs_iget_retry(sbi->sb, ino);
if (IS_ERR(inode))
return ERR_CAST(inode);
err = dquot_initialize(inode);
if (err)
goto err_out;
if (quota_inode) {
err = dquot_alloc_inode(inode);
if (err)
goto err_out;
}
entry = f2fs_kmem_cache_alloc(fsync_entry_slab, GFP_F2FS_ZERO);
entry->inode = inode;
list_add_tail(&entry->list, head);
return entry;
err_out:
iput(inode);
return ERR_PTR(err);
}
static void del_fsync_inode(struct fsync_inode_entry *entry)
{
iput(entry->inode);
list_del(&entry->list);
kmem_cache_free(fsync_entry_slab, entry);
}
static int recover_dentry(struct inode *inode, struct page *ipage,
struct list_head *dir_list)
{
struct f2fs_inode *raw_inode = F2FS_INODE(ipage);
nid_t pino = le32_to_cpu(raw_inode->i_pino);
struct f2fs_dir_entry *de;
struct fscrypt_name fname;
struct page *page;
struct inode *dir, *einode;
struct fsync_inode_entry *entry;
int err = 0;
char *name;
entry = get_fsync_inode(dir_list, pino);
if (!entry) {
entry = add_fsync_inode(F2FS_I_SB(inode), dir_list,
pino, false);
if (IS_ERR(entry)) {
dir = ERR_CAST(entry);
err = PTR_ERR(entry);
goto out;
}
}
dir = entry->inode;
memset(&fname, 0, sizeof(struct fscrypt_name));
fname.disk_name.len = le32_to_cpu(raw_inode->i_namelen);
fname.disk_name.name = raw_inode->i_name;
if (unlikely(fname.disk_name.len > F2FS_NAME_LEN)) {
WARN_ON(1);
err = -ENAMETOOLONG;
goto out;
}
retry:
de = __f2fs_find_entry(dir, &fname, &page);
if (de && inode->i_ino == le32_to_cpu(de->ino))
goto out_put;
if (de) {
einode = f2fs_iget_retry(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_put;
}
err = dquot_initialize(einode);
if (err) {
iput(einode);
goto out_put;
}
err = f2fs_acquire_orphan_inode(F2FS_I_SB(inode));
if (err) {
iput(einode);
goto out_put;
}
f2fs_delete_entry(de, page, dir, einode);
iput(einode);
goto retry;
} else if (IS_ERR(page)) {
err = PTR_ERR(page);
} else {
err = f2fs_add_dentry(dir, &fname, inode,
inode->i_ino, inode->i_mode);
}
if (err == -ENOMEM)
goto retry;
goto out;
out_put:
f2fs_put_page(page, 0);
out:
if (file_enc_name(inode))
name = "<encrypted>";
else
name = raw_inode->i_name;
f2fs_msg(inode->i_sb, KERN_NOTICE,
"%s: ino = %x, name = %s, dir = %lx, err = %d",
__func__, ino_of_node(ipage), name,
IS_ERR(dir) ? 0 : dir->i_ino, err);
return err;
}
static void recover_inline_flags(struct inode *inode, struct f2fs_inode *ri)
{
if (ri->i_inline & F2FS_PIN_FILE)
set_inode_flag(inode, FI_PIN_FILE);
else
clear_inode_flag(inode, FI_PIN_FILE);
if (ri->i_inline & F2FS_DATA_EXIST)
set_inode_flag(inode, FI_DATA_EXIST);
else
clear_inode_flag(inode, FI_DATA_EXIST);
}
static void recover_inode(struct inode *inode, struct page *page)
{
struct f2fs_inode *raw = F2FS_INODE(page);
char *name;
inode->i_mode = le16_to_cpu(raw->i_mode);
i_uid_write(inode, le32_to_cpu(raw->i_uid));
i_gid_write(inode, le32_to_cpu(raw->i_gid));
if (raw->i_inline & F2FS_EXTRA_ATTR) {
if (f2fs_sb_has_project_quota(F2FS_I_SB(inode)->sb) &&
F2FS_FITS_IN_INODE(raw, le16_to_cpu(raw->i_extra_isize),
i_projid)) {
projid_t i_projid;
i_projid = (projid_t)le32_to_cpu(raw->i_projid);
F2FS_I(inode)->i_projid =
make_kprojid(&init_user_ns, i_projid);
}
}
f2fs_i_size_write(inode, le64_to_cpu(raw->i_size));
inode->i_atime.tv_sec = le64_to_cpu(raw->i_atime);
inode->i_ctime.tv_sec = le64_to_cpu(raw->i_ctime);
inode->i_mtime.tv_sec = le64_to_cpu(raw->i_mtime);
inode->i_atime.tv_nsec = le32_to_cpu(raw->i_atime_nsec);
inode->i_ctime.tv_nsec = le32_to_cpu(raw->i_ctime_nsec);
inode->i_mtime.tv_nsec = le32_to_cpu(raw->i_mtime_nsec);
F2FS_I(inode)->i_advise = raw->i_advise;
F2FS_I(inode)->i_flags = le32_to_cpu(raw->i_flags);
F2FS_I(inode)->i_gc_failures[GC_FAILURE_PIN] =
le16_to_cpu(raw->i_gc_failures);
recover_inline_flags(inode, raw);
f2fs_mark_inode_dirty_sync(inode, true);
if (file_enc_name(inode))
name = "<encrypted>";
else
name = F2FS_INODE(page)->i_name;
f2fs_msg(inode->i_sb, KERN_NOTICE,
"recover_inode: ino = %x, name = %s, inline = %x",
ino_of_node(page), name, raw->i_inline);
}
static int find_fsync_dnodes(struct f2fs_sb_info *sbi, struct list_head *head,
bool check_only)
{
struct curseg_info *curseg;
struct page *page = NULL;
block_t blkaddr;
unsigned int loop_cnt = 0;
unsigned int free_blocks = MAIN_SEGS(sbi) * sbi->blocks_per_seg -
valid_user_blocks(sbi);
int err = 0;
/* get node pages in the current segment */
curseg = CURSEG_I(sbi, CURSEG_WARM_NODE);
blkaddr = NEXT_FREE_BLKADDR(sbi, curseg);
while (1) {
struct fsync_inode_entry *entry;
if (!f2fs_is_valid_blkaddr(sbi, blkaddr, META_POR))
return 0;
page = f2fs_get_tmp_page(sbi, blkaddr);
if (IS_ERR(page)) {
err = PTR_ERR(page);
break;
}
if (!is_recoverable_dnode(page))
break;
if (!is_fsync_dnode(page))
goto next;
entry = get_fsync_inode(head, ino_of_node(page));
if (!entry) {
bool quota_inode = false;
if (!check_only &&
IS_INODE(page) && is_dent_dnode(page)) {
err = f2fs_recover_inode_page(sbi, page);
if (err)
break;
quota_inode = true;
}
/*
* CP | dnode(F) | inode(DF)
* For this case, we should not give up now.
*/
entry = add_fsync_inode(sbi, head, ino_of_node(page),
quota_inode);
if (IS_ERR(entry)) {
err = PTR_ERR(entry);
if (err == -ENOENT) {
err = 0;
goto next;
}
break;
}
}
entry->blkaddr = blkaddr;
if (IS_INODE(page) && is_dent_dnode(page))
entry->last_dentry = blkaddr;
next:
/* sanity check in order to detect looped node chain */
if (++loop_cnt >= free_blocks ||
blkaddr == next_blkaddr_of_node(page)) {
f2fs_msg(sbi->sb, KERN_NOTICE,
"%s: detect looped node chain, "
"blkaddr:%u, next:%u",
__func__, blkaddr, next_blkaddr_of_node(page));
err = -EINVAL;
break;
}
/* check next segment */
blkaddr = next_blkaddr_of_node(page);
f2fs_put_page(page, 1);
f2fs_ra_meta_pages_cond(sbi, blkaddr);
}
f2fs_put_page(page, 1);
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)
del_fsync_inode(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;
struct dnode_of_data tdn = *dn;
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_HOT_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 = f2fs_get_sum_page(sbi, segno);
if (IS_ERR(sum_page))
return PTR_ERR(sum_page);
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) {
tdn.nid = nid;
if (!dn->inode_page_locked)
lock_page(dn->inode_page);
tdn.node_page = dn->inode_page;
tdn.ofs_in_node = le16_to_cpu(sum.ofs_in_node);
goto truncate_out;
} else if (dn->nid == nid) {
tdn.ofs_in_node = le16_to_cpu(sum.ofs_in_node);
goto truncate_out;
}
/* Get the node page */
node_page = f2fs_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);
if (ino != dn->inode->i_ino) {
int ret;
/* Deallocate previous index in the node page */
inode = f2fs_iget_retry(sbi->sb, ino);
if (IS_ERR(inode))
return PTR_ERR(inode);
ret = dquot_initialize(inode);
if (ret) {
iput(inode);
return ret;
}
} else {
inode = dn->inode;
}
bidx = f2fs_start_bidx_of_node(offset, inode) +
le16_to_cpu(sum.ofs_in_node);
/*
* if inode page is locked, unlock temporarily, but its reference
* count keeps alive.
*/
if (ino == dn->inode->i_ino && dn->inode_page_locked)
unlock_page(dn->inode_page);
set_new_dnode(&tdn, inode, NULL, NULL, 0);
if (f2fs_get_dnode_of_data(&tdn, bidx, LOOKUP_NODE))
goto out;
if (tdn.data_blkaddr == blkaddr)
f2fs_truncate_data_blocks_range(&tdn, 1);
f2fs_put_dnode(&tdn);
out:
if (ino != dn->inode->i_ino)
iput(inode);
else if (dn->inode_page_locked)
lock_page(dn->inode_page);
return 0;
truncate_out:
if (datablock_addr(tdn.inode, tdn.node_page,
tdn.ofs_in_node) == blkaddr)
f2fs_truncate_data_blocks_range(&tdn, 1);
if (dn->inode->i_ino == nid && !dn->inode_page_locked)
unlock_page(dn->inode_page);
return 0;
}
static int do_recover_data(struct f2fs_sb_info *sbi, struct inode *inode,
struct page *page)
{
struct dnode_of_data dn;
struct node_info ni;
unsigned int start, end;
int err = 0, recovered = 0;
/* step 1: recover xattr */
if (IS_INODE(page)) {
f2fs_recover_inline_xattr(inode, page);
} else if (f2fs_has_xattr_block(ofs_of_node(page))) {
err = f2fs_recover_xattr_data(inode, page);
if (!err)
recovered++;
goto out;
}
/* step 2: recover inline data */
if (f2fs_recover_inline_data(inode, page))
goto out;
/* step 3: recover data indices */
start = f2fs_start_bidx_of_node(ofs_of_node(page), inode);
end = start + ADDRS_PER_PAGE(page, inode);
set_new_dnode(&dn, inode, NULL, NULL, 0);
retry_dn:
err = f2fs_get_dnode_of_data(&dn, start, ALLOC_NODE);
if (err) {
if (err == -ENOMEM) {
congestion_wait(BLK_RW_ASYNC, HZ/50);
goto retry_dn;
}
goto out;
}
f2fs_wait_on_page_writeback(dn.node_page, NODE, true);
err = f2fs_get_node_info(sbi, dn.nid, &ni);
if (err)
goto err;
f2fs_bug_on(sbi, ni.ino != ino_of_node(page));
f2fs_bug_on(sbi, ofs_of_node(dn.node_page) != ofs_of_node(page));
for (; start < end; start++, dn.ofs_in_node++) {
block_t src, dest;
src = datablock_addr(dn.inode, dn.node_page, dn.ofs_in_node);
dest = datablock_addr(dn.inode, page, dn.ofs_in_node);
/* skip recovering if dest is the same as src */
if (src == dest)
continue;
/* dest is invalid, just invalidate src block */
if (dest == NULL_ADDR) {
f2fs_truncate_data_blocks_range(&dn, 1);
continue;
}
if (!file_keep_isize(inode) &&
(i_size_read(inode) <= ((loff_t)start << PAGE_SHIFT)))
f2fs_i_size_write(inode,
(loff_t)(start + 1) << PAGE_SHIFT);
/*
* dest is reserved block, invalidate src block
* and then reserve one new block in dnode page.
*/
if (dest == NEW_ADDR) {
f2fs_truncate_data_blocks_range(&dn, 1);
f2fs_reserve_new_block(&dn);
continue;
}
/* dest is valid block, try to recover from src to dest */
if (f2fs_is_valid_blkaddr(sbi, dest, META_POR)) {
if (src == NULL_ADDR) {
err = f2fs_reserve_new_block(&dn);
while (err &&
IS_ENABLED(CONFIG_F2FS_FAULT_INJECTION))
err = f2fs_reserve_new_block(&dn);
/* We should not get -ENOSPC */
f2fs_bug_on(sbi, err);
if (err)
goto err;
}
retry_prev:
/* Check the previous node page having this index */
err = check_index_in_prev_nodes(sbi, dest, &dn);
if (err) {
if (err == -ENOMEM) {
congestion_wait(BLK_RW_ASYNC, HZ/50);
goto retry_prev;
}
goto err;
}
/* write dummy data page */
f2fs_replace_block(sbi, &dn, src, dest,
ni.version, false, false);
recovered++;
}
}
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);
out:
f2fs_msg(sbi->sb, KERN_NOTICE,
"recover_data: ino = %lx (i_size: %s) recovered = %d, err = %d",
inode->i_ino,
file_keep_isize(inode) ? "keep" : "recover",
recovered, err);
return err;
}
static int recover_data(struct f2fs_sb_info *sbi, struct list_head *inode_list,
struct list_head *dir_list)
{
struct curseg_info *curseg;
struct page *page = NULL;
int err = 0;
block_t blkaddr;
/* get node pages in the current segment */
curseg = CURSEG_I(sbi, CURSEG_WARM_NODE);
blkaddr = NEXT_FREE_BLKADDR(sbi, curseg);
while (1) {
struct fsync_inode_entry *entry;
if (!f2fs_is_valid_blkaddr(sbi, blkaddr, META_POR))
break;
f2fs_ra_meta_pages_cond(sbi, blkaddr);
page = f2fs_get_tmp_page(sbi, blkaddr);
if (IS_ERR(page)) {
err = PTR_ERR(page);
break;
}
if (!is_recoverable_dnode(page)) {
f2fs_put_page(page, 1);
break;
}
entry = get_fsync_inode(inode_list, ino_of_node(page));
if (!entry)
goto next;
/*
* inode(x) | CP | inode(x) | dnode(F)
* In this case, we can lose the latest inode(x).
* So, call recover_inode for the inode update.
*/
if (IS_INODE(page))
recover_inode(entry->inode, page);
if (entry->last_dentry == blkaddr) {
err = recover_dentry(entry->inode, page, dir_list);
if (err) {
f2fs_put_page(page, 1);
break;
}
}
err = do_recover_data(sbi, entry->inode, page);
if (err) {
f2fs_put_page(page, 1);
break;
}
if (entry->blkaddr == blkaddr)
del_fsync_inode(entry);
next:
/* check next segment */
blkaddr = next_blkaddr_of_node(page);
f2fs_put_page(page, 1);
}
if (!err)
f2fs_allocate_new_segments(sbi);
return err;
}
int f2fs_recover_fsync_data(struct f2fs_sb_info *sbi, bool check_only)
{
struct list_head inode_list;
struct list_head dir_list;
int err;
int ret = 0;
unsigned long s_flags = sbi->sb->s_flags;
bool need_writecp = false;
#ifdef CONFIG_QUOTA
int quota_enabled;
#endif
if (s_flags & SB_RDONLY) {
f2fs_msg(sbi->sb, KERN_INFO,
"recover fsync data on readonly fs");
sbi->sb->s_flags &= ~SB_RDONLY;
}
#ifdef CONFIG_QUOTA
/* Needed for iput() to work correctly and not trash data */
sbi->sb->s_flags |= SB_ACTIVE;
/* Turn on quotas so that they are updated correctly */
quota_enabled = f2fs_enable_quota_files(sbi, s_flags & SB_RDONLY);
#endif
fsync_entry_slab = f2fs_kmem_cache_create("f2fs_fsync_inode_entry",
sizeof(struct fsync_inode_entry));
if (!fsync_entry_slab) {
err = -ENOMEM;
goto out;
}
INIT_LIST_HEAD(&inode_list);
INIT_LIST_HEAD(&dir_list);
/* prevent checkpoint */
mutex_lock(&sbi->cp_mutex);
/* step #1: find fsynced inode numbers */
err = find_fsync_dnodes(sbi, &inode_list, check_only);
if (err || list_empty(&inode_list))
goto skip;
if (check_only) {
ret = 1;
goto skip;
}
need_writecp = true;
/* step #2: recover data */
err = recover_data(sbi, &inode_list, &dir_list);
if (!err)
f2fs_bug_on(sbi, !list_empty(&inode_list));
skip:
destroy_fsync_dnodes(&inode_list);
/* truncate meta pages to be used by the recovery */
truncate_inode_pages_range(META_MAPPING(sbi),
(loff_t)MAIN_BLKADDR(sbi) << PAGE_SHIFT, -1);
if (err) {
truncate_inode_pages_final(NODE_MAPPING(sbi));
truncate_inode_pages_final(META_MAPPING(sbi));
}
clear_sbi_flag(sbi, SBI_POR_DOING);
mutex_unlock(&sbi->cp_mutex);
/* let's drop all the directory inodes for clean checkpoint */
destroy_fsync_dnodes(&dir_list);
if (need_writecp) {
set_sbi_flag(sbi, SBI_IS_RECOVERED);
if (!err) {
struct cp_control cpc = {
.reason = CP_RECOVERY,
};
err = f2fs_write_checkpoint(sbi, &cpc);
}
}
kmem_cache_destroy(fsync_entry_slab);
out:
#ifdef CONFIG_QUOTA
/* Turn quotas off */
if (quota_enabled)
f2fs_quota_off_umount(sbi->sb);
#endif
sbi->sb->s_flags = s_flags; /* Restore SB_RDONLY status */
return ret ? ret: err;
}