xfs: set IOMAP_F_NEW more carefully

Don't set IOMAP_F_NEW if we COW over an existing allocated range, as
these aren't strictly new allocations.  This is required to be able to
use IOMAP_F_NEW to zero newly allocated blocks, which is required for
the iomap code to fully support file systems that don't do delayed
allocations or use unwritten extents.

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>

diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c
index 699bbb8..f7b8b13 100644 (file)
--- a/fs/xfs/xfs_iomap.c
+++ b/fs/xfs/xfs_iomap.c
@@ -707,9 +707,12 @@ retry:
         * Flag newly allocated delalloc blocks with IOMAP_F_NEW so we punch
         * them out if the write happens to fail.
         */
-       iomap_flags |= IOMAP_F_NEW;
-       trace_xfs_iomap_alloc(ip, offset, count, whichfork,
-                       whichfork == XFS_DATA_FORK ? &imap : &cmap);
+       if (whichfork == XFS_DATA_FORK) {
+               iomap_flags |= IOMAP_F_NEW;
+               trace_xfs_iomap_alloc(ip, offset, count, whichfork, &imap);
+       } else {
+               trace_xfs_iomap_alloc(ip, offset, count, whichfork, &cmap);
+       }
 done:
        if (whichfork == XFS_COW_FORK) {
                if (imap.br_startoff > offset_fsb) {