xfs: don't allocate COW blocks for zeroing holes or unwritten extents
authorChristoph Hellwig <hch@lst.de>
Thu, 1 Mar 2018 22:10:31 +0000 (14:10 -0800)
committerDarrick J. Wong <darrick.wong@oracle.com>
Thu, 1 Mar 2018 22:10:31 +0000 (14:10 -0800)
The iomap zeroing interface is smart enough to skip zeroing holes or
unwritten extents.  Don't subvert this logic for reflink files.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
fs/xfs/xfs_iomap.c

index 66e1edb..4e771e0 100644 (file)
@@ -955,6 +955,13 @@ static inline bool imap_needs_alloc(struct inode *inode,
                (IS_DAX(inode) && imap->br_state == XFS_EXT_UNWRITTEN);
 }
 
+static inline bool needs_cow_for_zeroing(struct xfs_bmbt_irec *imap, int nimaps)
+{
+       return nimaps &&
+               imap->br_startblock != HOLESTARTBLOCK &&
+               imap->br_state != XFS_EXT_UNWRITTEN;
+}
+
 static inline bool need_excl_ilock(struct xfs_inode *ip, unsigned flags)
 {
        /*
@@ -1024,7 +1031,9 @@ xfs_file_iomap_begin(
                        goto out_unlock;
        }
 
-       if ((flags & (IOMAP_WRITE | IOMAP_ZERO)) && xfs_is_reflink_inode(ip)) {
+       if (xfs_is_reflink_inode(ip) &&
+           ((flags & IOMAP_WRITE) ||
+            ((flags & IOMAP_ZERO) && needs_cow_for_zeroing(&imap, nimaps)))) {
                if (flags & IOMAP_DIRECT) {
                        /*
                         * A reflinked inode will result in CoW alloc.