On (06/01/17 00:41), Julian Anastasov wrote: > > So, we do not hold reference to neigh while accessing > its fields. I suspect we need to move the table lock from > neigh_remove_one here, for example:
Another thought is to have neigh_remove_one to remove a neigh only if it is NUD_FAILED - that may potentially remove more than one entry, but that's probably harmless? --Sowmini
