> The primary issue for 'routeadm -u', though, is one that affects > systems regardless of whether they run the 'nwam' service. All > systems get an IPv6 lo0 loopback interface, whether they need it or > not, and that change came in quite some time ago with NWAM Phase 0.
A tertiary issue is that we now have a bunch of code in routeadm that does nothing, as ra_numv6intfs() will always return > 0. A similar pile of code was pulled out of zoneadmd a couple builds ago, and I'm yanking a similar pile from the IP RCM module as part of the IPMP rearchitecture project. Doubtless there are others. (To be clear: I think having IPv6 loopback always plumbed is a step in the right direction -- it's just a IPC mechanism and should have no bearing on whether the system is connected to an IPv6 network. But having it plumbed all the time also invalidates any places where we historically used its presence to determine IPv6 network connectivity, such as in routeadm.) -- meem _______________________________________________ networking-discuss mailing list [email protected]
