rculfhash: merge duplicated code for bucket lookup

[ Edit by Mathieu Desnoyers: also remove unnecessary lookups that end up
  returning an already looked up node. Same behavior is guaranteed
  because the hash and size stay invariant across retry. ]

Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>