f2fs: fix SSA updates resulting in corruption
The f2fs_collapse_range and f2fs_insert_range changes the block addresses
directly. But that can cause uncovered SSA updates.
In that case, we need to give up to change the block addresses and do buffered
writes to keep filesystem consistency.
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
index f05ae22..6f2310c 100644
--- a/fs/f2fs/f2fs.h
+++ b/fs/f2fs/f2fs.h
@@ -1233,6 +1233,16 @@
return sbi->total_valid_inode_count;
}
+static inline void f2fs_copy_page(struct page *src, struct page *dst)
+{
+ char *src_kaddr = kmap(src);
+ char *dst_kaddr = kmap(dst);
+
+ memcpy(dst_kaddr, src_kaddr, PAGE_SIZE);
+ kunmap(dst);
+ kunmap(src);
+}
+
static inline void f2fs_put_page(struct page *page, int unlock)
{
if (!page)
@@ -1754,6 +1764,7 @@
int create_flush_cmd_control(struct f2fs_sb_info *);
void destroy_flush_cmd_control(struct f2fs_sb_info *);
void invalidate_blocks(struct f2fs_sb_info *, block_t);
+bool is_checkpointed_data(struct f2fs_sb_info *, block_t);
void refresh_sit_entry(struct f2fs_sb_info *, block_t, block_t);
void clear_prefree_segments(struct f2fs_sb_info *, struct cp_control *);
void release_discard_addrs(struct f2fs_sb_info *);
diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
index 6d3cfd5..8d5583b 100644
--- a/fs/f2fs/file.c
+++ b/fs/f2fs/file.c
@@ -826,86 +826,100 @@
return ret;
}
-static int f2fs_do_collapse(struct inode *inode, pgoff_t start, pgoff_t end)
+static int __exchange_data_block(struct inode *inode, pgoff_t src,
+ pgoff_t dst, bool full)
{
struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
struct dnode_of_data dn;
+ block_t new_addr;
+ bool do_replace = false;
+ int ret;
+
+ set_new_dnode(&dn, inode, NULL, NULL, 0);
+ ret = get_dnode_of_data(&dn, src, LOOKUP_NODE_RA);
+ if (ret && ret != -ENOENT) {
+ return ret;
+ } else if (ret == -ENOENT) {
+ new_addr = NULL_ADDR;
+ } else {
+ new_addr = dn.data_blkaddr;
+ if (!is_checkpointed_data(sbi, new_addr)) {
+ dn.data_blkaddr = NULL_ADDR;
+ /* do not invalidate this block address */
+ set_data_blkaddr(&dn);
+ f2fs_update_extent_cache(&dn);
+ do_replace = true;
+ }
+ f2fs_put_dnode(&dn);
+ }
+
+ if (new_addr == NULL_ADDR)
+ return full ? truncate_hole(inode, dst, dst + 1) : 0;
+
+ if (do_replace) {
+ struct page *ipage = get_node_page(sbi, inode->i_ino);
+ struct node_info ni;
+
+ if (IS_ERR(ipage)) {
+ ret = PTR_ERR(ipage);
+ goto err_out;
+ }
+
+ set_new_dnode(&dn, inode, ipage, NULL, 0);
+ ret = f2fs_reserve_block(&dn, dst);
+ if (ret)
+ goto err_out;
+
+ truncate_data_blocks_range(&dn, 1);
+
+ get_node_info(sbi, dn.nid, &ni);
+ f2fs_replace_block(sbi, &dn, dn.data_blkaddr, new_addr,
+ ni.version, true);
+ f2fs_put_dnode(&dn);
+ } else {
+ struct page *psrc, *pdst;
+
+ psrc = get_lock_data_page(inode, src);
+ if (IS_ERR(psrc))
+ return PTR_ERR(psrc);
+ pdst = get_new_data_page(inode, NULL, dst, false);
+ if (IS_ERR(pdst)) {
+ f2fs_put_page(psrc, 1);
+ return PTR_ERR(pdst);
+ }
+ f2fs_copy_page(psrc, pdst);
+ set_page_dirty(pdst);
+ f2fs_put_page(pdst, 1);
+ f2fs_put_page(psrc, 1);
+
+ return truncate_hole(inode, src, src + 1);
+ }
+ return 0;
+
+err_out:
+ if (!get_dnode_of_data(&dn, src, LOOKUP_NODE)) {
+ dn.data_blkaddr = new_addr;
+ set_data_blkaddr(&dn);
+ f2fs_update_extent_cache(&dn);
+ f2fs_put_dnode(&dn);
+ }
+ return ret;
+}
+
+static int f2fs_do_collapse(struct inode *inode, pgoff_t start, pgoff_t end)
+{
+ struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
pgoff_t nrpages = (i_size_read(inode) + PAGE_SIZE - 1) / PAGE_SIZE;
int ret = 0;
for (; end < nrpages; start++, end++) {
- block_t new_addr, old_addr;
-
+ f2fs_balance_fs(sbi);
f2fs_lock_op(sbi);
-
- set_new_dnode(&dn, inode, NULL, NULL, 0);
- ret = get_dnode_of_data(&dn, end, LOOKUP_NODE_RA);
- if (ret && ret != -ENOENT) {
- goto out;
- } else if (ret == -ENOENT) {
- new_addr = NULL_ADDR;
- } else {
- new_addr = dn.data_blkaddr;
- truncate_data_blocks_range(&dn, 1);
- f2fs_put_dnode(&dn);
- }
-
- if (new_addr == NULL_ADDR) {
- set_new_dnode(&dn, inode, NULL, NULL, 0);
- ret = get_dnode_of_data(&dn, start, LOOKUP_NODE_RA);
- if (ret && ret != -ENOENT) {
- goto out;
- } else if (ret == -ENOENT) {
- f2fs_unlock_op(sbi);
- continue;
- }
-
- if (dn.data_blkaddr == NULL_ADDR) {
- f2fs_put_dnode(&dn);
- f2fs_unlock_op(sbi);
- continue;
- } else {
- truncate_data_blocks_range(&dn, 1);
- }
-
- f2fs_put_dnode(&dn);
- } else {
- struct page *ipage;
-
- ipage = get_node_page(sbi, inode->i_ino);
- if (IS_ERR(ipage)) {
- ret = PTR_ERR(ipage);
- goto out;
- }
-
- set_new_dnode(&dn, inode, ipage, NULL, 0);
- ret = f2fs_reserve_block(&dn, start);
- if (ret)
- goto out;
-
- old_addr = dn.data_blkaddr;
- if (old_addr != NEW_ADDR && new_addr == NEW_ADDR) {
- dn.data_blkaddr = NULL_ADDR;
- f2fs_update_extent_cache(&dn);
- invalidate_blocks(sbi, old_addr);
-
- dn.data_blkaddr = new_addr;
- set_data_blkaddr(&dn);
- } else if (new_addr != NEW_ADDR) {
- struct node_info ni;
-
- get_node_info(sbi, dn.nid, &ni);
- f2fs_replace_block(sbi, &dn, old_addr, new_addr,
- ni.version, true);
- }
-
- f2fs_put_dnode(&dn);
- }
+ ret = __exchange_data_block(inode, end, start, true);
f2fs_unlock_op(sbi);
+ if (ret)
+ break;
}
- return 0;
-out:
- f2fs_unlock_op(sbi);
return ret;
}
@@ -944,7 +958,12 @@
if (ret)
return ret;
+ /* write out all moved pages, if possible */
+ filemap_write_and_wait_range(inode->i_mapping, offset, LLONG_MAX);
+ truncate_pagecache(inode, offset);
+
new_size = i_size_read(inode) - len;
+ truncate_pagecache(inode, new_size);
ret = truncate_blocks(inode, new_size, true);
if (!ret)
@@ -1067,7 +1086,7 @@
struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
pgoff_t pg_start, pg_end, delta, nrpages, idx;
loff_t new_size;
- int ret;
+ int ret = 0;
new_size = i_size_read(inode) + len;
if (new_size > inode->i_sb->s_maxbytes)
@@ -1105,57 +1124,19 @@
nrpages = (i_size_read(inode) + PAGE_SIZE - 1) / PAGE_SIZE;
for (idx = nrpages - 1; idx >= pg_start && idx != -1; idx--) {
- struct dnode_of_data dn;
- struct page *ipage;
- block_t new_addr, old_addr;
-
f2fs_lock_op(sbi);
-
- set_new_dnode(&dn, inode, NULL, NULL, 0);
- ret = get_dnode_of_data(&dn, idx, LOOKUP_NODE_RA);
- if (ret && ret != -ENOENT) {
- goto out;
- } else if (ret == -ENOENT) {
- goto next;
- } else if (dn.data_blkaddr == NULL_ADDR) {
- f2fs_put_dnode(&dn);
- goto next;
- } else {
- new_addr = dn.data_blkaddr;
- truncate_data_blocks_range(&dn, 1);
- f2fs_put_dnode(&dn);
- }
-
- ipage = get_node_page(sbi, inode->i_ino);
- if (IS_ERR(ipage)) {
- ret = PTR_ERR(ipage);
- goto out;
- }
-
- set_new_dnode(&dn, inode, ipage, NULL, 0);
- ret = f2fs_reserve_block(&dn, idx + delta);
- if (ret)
- goto out;
-
- old_addr = dn.data_blkaddr;
- f2fs_bug_on(sbi, old_addr != NEW_ADDR);
-
- if (new_addr != NEW_ADDR) {
- struct node_info ni;
-
- get_node_info(sbi, dn.nid, &ni);
- f2fs_replace_block(sbi, &dn, old_addr, new_addr,
- ni.version, true);
- }
- f2fs_put_dnode(&dn);
-next:
+ ret = __exchange_data_block(inode, idx, idx + delta, false);
f2fs_unlock_op(sbi);
+ if (ret)
+ break;
}
- i_size_write(inode, new_size);
- return 0;
-out:
- f2fs_unlock_op(sbi);
+ /* write out all moved pages, if possible */
+ filemap_write_and_wait_range(inode->i_mapping, offset, LLONG_MAX);
+ truncate_pagecache(inode, offset);
+
+ if (!ret)
+ i_size_write(inode, new_size);
return ret;
}
diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index 1d86a35..581a9af 100644
--- a/fs/f2fs/segment.c
+++ b/fs/f2fs/segment.c
@@ -768,6 +768,30 @@
mutex_unlock(&sit_i->sentry_lock);
}
+bool is_checkpointed_data(struct f2fs_sb_info *sbi, block_t blkaddr)
+{
+ struct sit_info *sit_i = SIT_I(sbi);
+ unsigned int segno, offset;
+ struct seg_entry *se;
+ bool is_cp = false;
+
+ if (blkaddr == NEW_ADDR || blkaddr == NULL_ADDR)
+ return true;
+
+ mutex_lock(&sit_i->sentry_lock);
+
+ segno = GET_SEGNO(sbi, blkaddr);
+ se = get_seg_entry(sbi, segno);
+ offset = GET_BLKOFF_FROM_SEG0(sbi, blkaddr);
+
+ if (f2fs_test_bit(offset, se->ckpt_valid_map))
+ is_cp = true;
+
+ mutex_unlock(&sit_i->sentry_lock);
+
+ return is_cp;
+}
+
/*
* This function should be resided under the curseg_mutex lock
*/
@@ -1370,7 +1394,14 @@
curseg->next_blkoff = GET_BLKOFF_FROM_SEG0(sbi, new_blkaddr);
__add_sum_entry(sbi, type, sum);
- refresh_sit_entry(sbi, old_blkaddr, new_blkaddr);
+ if (!recover_curseg)
+ update_sit_entry(sbi, new_blkaddr, 1);
+ if (GET_SEGNO(sbi, old_blkaddr) != NULL_SEGNO)
+ update_sit_entry(sbi, old_blkaddr, -1);
+
+ locate_dirty_segment(sbi, GET_SEGNO(sbi, old_blkaddr));
+ locate_dirty_segment(sbi, GET_SEGNO(sbi, new_blkaddr));
+
locate_dirty_segment(sbi, old_cursegno);
if (recover_curseg) {