On Sun, Feb 24, 2019 at 12:31:05PM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@xxxxxxxxxx>
>
> Rework the data flow in xfs_file_iomap_begin where we decide if we have
> to break shared extents.
>
> Signed-off-by: Darrick J. Wong <darrick.wong@xxxxxxxxxx>
> ---
Reviewed-by: Brian Foster <bfoster@xxxxxxxxxx>
> fs/xfs/xfs_iomap.c | 13 ++++++++-----
> 1 file changed, 8 insertions(+), 5 deletions(-)
>
>
> diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c
> index a376fb2d4b94..a3043b1bcd9a 100644
> --- a/fs/xfs/xfs_iomap.c
> +++ b/fs/xfs/xfs_iomap.c
> @@ -982,7 +982,7 @@ xfs_file_iomap_begin(
> * been done up front, so we don't need to do them here.
> */
> if (xfs_is_cow_inode(ip)) {
> - struct xfs_bmbt_irec orig = imap;
> + struct xfs_bmbt_irec cmap;
> bool convert_now = (flags & IOMAP_DIRECT);
>
> /* if zeroing doesn't need COW allocation, then we are done. */
> @@ -991,7 +991,8 @@ xfs_file_iomap_begin(
> goto out_found;
>
> /* may drop and re-acquire the ilock */
> - error = xfs_reflink_allocate_cow(ip, &imap, &shared, &lockmode,
> + cmap = imap;
> + error = xfs_reflink_allocate_cow(ip, &cmap, &shared, &lockmode,
> convert_now);
> if (error)
> goto out_unlock;
> @@ -1001,10 +1002,12 @@ xfs_file_iomap_begin(
> * previous block (if there was any) so that the higher level
> * write code can perform read-modify-write operations. For
> * direct I/O code, which must be block aligned we need to
> - * report the newly allocated address.
> + * report the newly allocated address. If the data fork has
> + * a hole, copy the COW fork mapping to avoid allocating to
> + * the data fork.
> */
> - if (!convert_now && orig.br_startblock != HOLESTARTBLOCK)
> - imap = orig;
> + if (convert_now || imap.br_startblock == HOLESTARTBLOCK)
> + imap = cmap;
>
> end_fsb = imap.br_startoff + imap.br_blockcount;
> length = XFS_FSB_TO_B(mp, end_fsb) - offset;
>
