Re: [PATCH 4/9] xfs: only reserve quota blocks for bmbt changes if we're changing the data fork

[Brian Foster <bfoster@redhat.com>]

On Wed, Jun 24, 2020 at 06:18:06PM -0700, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
> 
> Now that we've reworked xfs_reflink_remap_extent to remap only one
> extent per transaction, we actually know if the extent being removed is
> an allocated mapping.  This means that we now know ahead of time if
> we're going to be touching the data fork.
> 
> Since we only need blocks for a bmbt split if we're going to update the
> data fork, we only need to get quota reservation if we know we're going
> to touch the data fork.
> 
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> ---

That addresses my question on the previous patch. Looks like the qres
check was just misplaced. NBD:

Reviewed-by: Brian Foster <bfoster@redhat.com>

>  fs/xfs/xfs_reflink.c |   11 ++++++++---
>  1 file changed, 8 insertions(+), 3 deletions(-)
> 
> 
> diff --git a/fs/xfs/xfs_reflink.c b/fs/xfs/xfs_reflink.c
> index c593d52156df..9cc1c340d0ec 100644
> --- a/fs/xfs/xfs_reflink.c
> +++ b/fs/xfs/xfs_reflink.c
> @@ -1043,7 +1043,11 @@ xfs_reflink_remap_extent(
>  	 * Compute quota reservation if we think the quota block counter for
>  	 * this file could increase.
>  	 *
> -	 * We start by reserving enough blocks to handle a bmbt split.
> +	 * Adding a written extent to the extent map can cause a bmbt split,
> +	 * and removing a mapped extent from the extent can cause a bmbt split.
> +	 * The two operations cannot both cause a split since they operate on
> +	 * the same index in the bmap btree, so we only need a reservation for
> +	 * one bmbt split if either thing is happening.
>  	 *
>  	 * If we are mapping a written extent into the file, we need to have
>  	 * enough quota block count reservation to handle the blocks in that
> @@ -1056,8 +1060,9 @@ xfs_reflink_remap_extent(
>  	 * before we started.  That should have removed all the delalloc
>  	 * reservations, but we code defensively.
>  	 */
> -	qdelta = 0;
> -	qres = XFS_EXTENTADD_SPACE_RES(mp, XFS_DATA_FORK);
> +	qres = qdelta = 0;
> +	if (smap_mapped || dmap_written)
> +		qres = XFS_EXTENTADD_SPACE_RES(mp, XFS_DATA_FORK);
>  	if (dmap_written)
>  		qres += dmap->br_blockcount;
>  	if (qres > 0) {
>