> Under what valid use case will multiple interfaces have the same network > address? > Hi, Neil. I'm not sure the specific valid use case.
The point is, do we trust the sctp global addr list has no duplicate address ? In one of users' computer, he found hundreds of duplicate addresses in INIT_ACK packets. (https://bugzilla.redhat.com/show_bug.cgi?id=1308362) If we think NETDEV_UP/DOWN event in addrs chain always come in pairs, I can not explain how come this issue happened. But from sctp end, we can avoid it with this patch. after all binding any duplicate addresses is illegal, just like what sctp_bind does. what do you think ?
