> Yes, it does seem broken, kind of surprising we haven't seen this very > often. But instead of introducing a new ipif_need_up field, can't you > deduce it from the already available information ?
We already went through ip_sioctl_flags() -> ipif_down() and cleared IPIF_UP before calling ip_sioctl_flags_restart(), so I don't think there's enough information left. -- meem
