Currently the block sender does not indicate if the final batch of blocks sent is all the block the far end has. Adding a flag to the protocol and having the client check for that to inform further sync behaviour would be useful. Currently the node waits for a new block hash gossip to try to sync again, which may take several minutes.