On Wed, Jun 24, 2020 at 06:17:45PM -0700, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@xxxxxxxxxx>
>
> Quota reservations are supposed to account for the blocks that might be
> allocated due to a bmap btree split. Reflink doesn't do this, so fix
> this to make the quota accounting more accurate before we start
> rearranging things.
>
> Fixes: 862bb360ef56 ("xfs: reflink extents from one file to another")
> Signed-off-by: Darrick J. Wong <darrick.wong@xxxxxxxxxx>
> ---
Reviewed-by: Brian Foster <bfoster@xxxxxxxxxx>
> fs/xfs/xfs_reflink.c | 21 ++++++++++++++-------
> 1 file changed, 14 insertions(+), 7 deletions(-)
>
>
> diff --git a/fs/xfs/xfs_reflink.c b/fs/xfs/xfs_reflink.c
> index 107bf2a2f344..d89201d40891 100644
> --- a/fs/xfs/xfs_reflink.c
> +++ b/fs/xfs/xfs_reflink.c
> @@ -1003,6 +1003,7 @@ xfs_reflink_remap_extent(
> xfs_filblks_t rlen;
> xfs_filblks_t unmap_len;
> xfs_off_t newlen;
> + int64_t qres;
> int error;
>
> unmap_len = irec->br_startoff + irec->br_blockcount - destoff;
> @@ -1025,13 +1026,19 @@ xfs_reflink_remap_extent(
> xfs_ilock(ip, XFS_ILOCK_EXCL);
> xfs_trans_ijoin(tp, ip, 0);
>
> - /* If we're not just clearing space, then do we have enough quota? */
> - if (real_extent) {
> - error = xfs_trans_reserve_quota_nblks(tp, ip,
> - irec->br_blockcount, 0, XFS_QMOPT_RES_REGBLKS);
> - if (error)
> - goto out_cancel;
> - }
> + /*
> + * Reserve quota for this operation. We don't know if the first unmap
> + * in the dest file will cause a bmap btree split, so we always reserve
> + * at least enough blocks for that split. If the extent being mapped
> + * in is written, we need to reserve quota for that too.
> + */
> + qres = XFS_EXTENTADD_SPACE_RES(mp, XFS_DATA_FORK);
> + if (real_extent)
> + qres += irec->br_blockcount;
> + error = xfs_trans_reserve_quota_nblks(tp, ip, qres, 0,
> + XFS_QMOPT_RES_REGBLKS);
> + if (error)
> + goto out_cancel;
>
> trace_xfs_reflink_remap(ip, irec->br_startoff,
> irec->br_blockcount, irec->br_startblock);
>
