>>>>> On Thu, 22 Mar 2001 16:02:10 -0500,
>>>>> "David E. Cross" <[EMAIL PROTECTED]> said:
> Why is routing done via the ::1 and 127.0.0.1 network addresses? I notice
> for "normal" interfaces it is bound directly to "link#2" and such.
It's just a characteristic (or ristriction if you want to say that) of
BSD's IPv4 routing on point-to-point interfaces. I don't know the
deep rationale.
> I realize I don't really know what I am talking about here, but, it
> seems that binding it to the link is more efficient than having it go
> through the loopback interface.
I'm not sure what you mean "efficient" here. But using addresses to
install a route to the loopback interface does not decrease the output
performance.
> Also, it will work in cases where the
> loopback is not defined (don't ask... just don't ask)
That's not true. In the BSD's routing architecture, if you want to
install a route of a particular address family to an interface, you
need at least one interface address of the address family on the
interface. The address need not to be the well-defined "loopback
address", though.
JINMEI, Tatuya
Communication Platform Lab.
Corporate R&D Center, Toshiba Corp.
[EMAIL PROTECTED]
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-net" in the body of the message
