提交 67a7d5f5 编写于 作者: J Jan Kara 提交者: Theodore Ts'o

ext4: fix fdatasync(2) after extent manipulation operations

Currently, extent manipulation operations such as hole punch, range
zeroing, or extent shifting do not record the fact that file data has
changed and thus fdatasync(2) has a work to do. As a result if we crash
e.g. after a punch hole and fdatasync, user can still possibly see the
punched out data after journal replay. Test generic/392 fails due to
these problems.

Fix the problem by properly marking that file data has changed in these
operations.

CC: stable@vger.kernel.org
Fixes: a4bb6b64Signed-off-by: NJan Kara <jack@suse.cz>
Signed-off-by: NTheodore Ts'o <tytso@mit.edu>
上级 a056bdaa
...@@ -4877,6 +4877,8 @@ static long ext4_zero_range(struct file *file, loff_t offset, ...@@ -4877,6 +4877,8 @@ static long ext4_zero_range(struct file *file, loff_t offset,
/* Zero out partial block at the edges of the range */ /* Zero out partial block at the edges of the range */
ret = ext4_zero_partial_blocks(handle, inode, offset, len); ret = ext4_zero_partial_blocks(handle, inode, offset, len);
if (ret >= 0)
ext4_update_inode_fsync_trans(handle, inode, 1);
if (file->f_flags & O_SYNC) if (file->f_flags & O_SYNC)
ext4_handle_sync(handle); ext4_handle_sync(handle);
...@@ -5563,6 +5565,7 @@ int ext4_collapse_range(struct inode *inode, loff_t offset, loff_t len) ...@@ -5563,6 +5565,7 @@ int ext4_collapse_range(struct inode *inode, loff_t offset, loff_t len)
ext4_handle_sync(handle); ext4_handle_sync(handle);
inode->i_mtime = inode->i_ctime = current_time(inode); inode->i_mtime = inode->i_ctime = current_time(inode);
ext4_mark_inode_dirty(handle, inode); ext4_mark_inode_dirty(handle, inode);
ext4_update_inode_fsync_trans(handle, inode, 1);
out_stop: out_stop:
ext4_journal_stop(handle); ext4_journal_stop(handle);
...@@ -5736,6 +5739,8 @@ int ext4_insert_range(struct inode *inode, loff_t offset, loff_t len) ...@@ -5736,6 +5739,8 @@ int ext4_insert_range(struct inode *inode, loff_t offset, loff_t len)
up_write(&EXT4_I(inode)->i_data_sem); up_write(&EXT4_I(inode)->i_data_sem);
if (IS_SYNC(inode)) if (IS_SYNC(inode))
ext4_handle_sync(handle); ext4_handle_sync(handle);
if (ret >= 0)
ext4_update_inode_fsync_trans(handle, inode, 1);
out_stop: out_stop:
ext4_journal_stop(handle); ext4_journal_stop(handle);
......
...@@ -4218,6 +4218,8 @@ int ext4_punch_hole(struct inode *inode, loff_t offset, loff_t length) ...@@ -4218,6 +4218,8 @@ int ext4_punch_hole(struct inode *inode, loff_t offset, loff_t length)
inode->i_mtime = inode->i_ctime = current_time(inode); inode->i_mtime = inode->i_ctime = current_time(inode);
ext4_mark_inode_dirty(handle, inode); ext4_mark_inode_dirty(handle, inode);
if (ret >= 0)
ext4_update_inode_fsync_trans(handle, inode, 1);
out_stop: out_stop:
ext4_journal_stop(handle); ext4_journal_stop(handle);
out_dio: out_dio:
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册