osd: avoid multi set osd_op.outdata in tier pool
There are two read op on the same object for ec pool. First op read
miss happend, call do_proxy_read and promote_object, The second op only
do_proxy_read. but before call process_copy_chunk for first op and then
finish_proxy_read for second op, first op receive reply from base pool
first and then second received. so the second op reset the field "outdata"
in m->ops first. And then call process_copy_chunk for first op,
cancel_proxy_read for second op and then requeue it. At last in do_osd_ops,
it append outdata field.
PR https://github.com/ceph/ceph/pull/4895 is pending for review.