make 203 pass on non-default block size filesystems

make 203 pass on non-default block size filesystems

The xfs_bmap output in the golden image is filesystem block size
dependent. Make all writes 64k to ensure that the allocation/hole
pattern is consistent across all supported filesystem block sizes.

Also, use the SCRATCH_DEV instead of the TEST_DEV so that we test
according to MKFS_OPTIONS rather than test on whatever setup the
TEST_DEV was created with.

Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Alex Elder <aelder@sgi.com>