xfs: ensure truncate forces zeroed blocks to disk

A new fsync vs power fail test in xfstests indicated that XFS can
have unreliable data consistency when doing extending truncates that
require block zeroing. The blocks beyond EOF get zeroed in memory,
but we never force those changes to disk before we run the
transaction that extends the file size and exposes those blocks to
userspace. This can result in the blocks not being correctly zeroed
after a crash.

Because in-memory behaviour is correct, tools like fsx don't pick up
any coherency problems - it's not until the filesystem is shutdown
or the system crashes after writing the truncate transaction to the
journal but before the zeroed data in the page cache is flushed that
the issue is exposed.

Fix this by also flushing the dirty data in memory region between
the old size and new size when we've found blocks that need zeroing
in the truncate process.

Reported-by: Liu Bo <bo.li.liu@oracle.com>
cc: <stable@vger.kernel.org>
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>

1 files changed, 10 insertions, 4 deletions

diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c
index ce615d12fb44..a2e1cb8a568b 100644
--- a/fs/xfs/xfs_file.c
+++ b/fs/xfs/xfs_file.c
@@ -397,7 +397,8 @@ STATIC int				/* error (positive) */
 xfs_zero_last_block(
 	struct xfs_inode	*ip,
 	xfs_fsize_t		offset,
-	xfs_fsize_t		isize)
+	xfs_fsize_t		isize,
+	bool			*did_zeroing)
 {
 	struct xfs_mount	*mp = ip->i_mount;
 	xfs_fileoff_t		last_fsb = XFS_B_TO_FSBT(mp, isize);
@@ -425,6 +426,7 @@ xfs_zero_last_block(
 	zero_len = mp->m_sb.sb_blocksize - zero_offset;
 	if (isize + zero_len > offset)
 		zero_len = offset - isize;
+	*did_zeroing = true;
 	return xfs_iozero(ip, isize, zero_len);
 }
 
@@ -443,7 +445,8 @@ int					/* error (positive) */
 xfs_zero_eof(
 	struct xfs_inode	*ip,
 	xfs_off_t		offset,		/* starting I/O offset */
-	xfs_fsize_t		isize)		/* current inode size */
+	xfs_fsize_t		isize,		/* current inode size */
+	bool			*did_zeroing)
 {
 	struct xfs_mount	*mp = ip->i_mount;
 	xfs_fileoff_t		start_zero_fsb;
@@ -465,7 +468,7 @@ xfs_zero_eof(
 	 * We only zero a part of that block so it is handled specially.
 	 */
 	if (XFS_B_FSB_OFFSET(mp, isize) != 0) {
-		error = xfs_zero_last_block(ip, offset, isize);
+		error = xfs_zero_last_block(ip, offset, isize, did_zeroing);
 		if (error)
 			return error;
 	}
@@ -525,6 +528,7 @@ xfs_zero_eof(
 		if (error)
 			return error;
 
+		*did_zeroing = true;
 		start_zero_fsb = imap.br_startoff + imap.br_blockcount;
 		ASSERT(start_zero_fsb <= (end_zero_fsb + 1));
 	}
@@ -567,13 +571,15 @@ restart:
 	 * having to redo all checks before.
 	 */
 	if (*pos > i_size_read(inode)) {
+		bool	zero = false;
+
 		if (*iolock == XFS_IOLOCK_SHARED) {
 			xfs_rw_iunlock(ip, *iolock);
 			*iolock = XFS_IOLOCK_EXCL;
 			xfs_rw_ilock(ip, *iolock);
 			goto restart;
 		}
-		error = xfs_zero_eof(ip, *pos, i_size_read(inode));
+		error = xfs_zero_eof(ip, *pos, i_size_read(inode), &zero);
 		if (error)
 			return error;
 	}