Scott Lamb writes: > Looks like I have a slightly different situation than you're describing. > I'm not specifying a "set ipcp ranges" anywhere. The peer chooses an > inside address that's different from but within the same /24 as its > outside address. I'm trying to run everything within two /16s over the > VPN (which include those addresses). So I guess I'm getting the same > routing loop but in a slightly more complicated way.
Yes.. the problem happens when the route to the outside IP address goes over the PPTP tunnel.. you get an "infinite encapsulation loop". > Taking out my "set iface route" lines in mpd.conf made it no longer > crash and I can talk to just the VPN host. So I tried next a "route add > vpnhost mygw" before starting the VPN. It worked. So that's definitely > the problem. It would be nice if there were a slicker way of doing this > (mpd automatically adding that route if this situation occurs?), > especially since I'm on DHCP. And it does make me nervous that mpd can > make the kernel crash at all. But I can at least work around this now. You are right, kernel crashing is always bogus no matter how it happens. Mpd adding a host route for the remote IP is a good idea, I didn't think of that. I'll try to add something to that effect when time permits. Really there are two things to fix: mpd as you describe, and ng_ksocket, which is where the crash originates. -Archie __________________________________________________________________________ Archie Cobbs * Packet Design * http://www.packetdesign.com To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-net" in the body of the message
