Kacheong Poon wrote:
> Erik Nordmark wrote:
>
>> The strong end-system model (see RFC 1122) talks about restricting
>> things both at the receiving and sending ends, and we have some RFEs
>> asking for that. What this means in practice is that the source IP
>> address constrains the interface on which the packet can be sent out,
>> whether the source IP address was set by the application (doing a
>> bind()), or set as part of accepting a new TCP/SCTP connection.
>
>
> I guess an app will get a network unreachable error if
> there is no route to a destination using the bound interface
> even when there is a route using another interface?
Correct.
> Can
> an app enforce this restriction even if the system does not
> (ip[6]_strict_src_multihoming is not set)? Via a socket
> option?
We could invent such a socket option (or a pair with IP_ and IPV6_ as
different prefixes to be consistent with the way we currently do
IPv4/IPv6 options.)
Do you know if such a socket option already exists for other OSes so we
don't have to invent our own name?
Erik
