IPv6 config review

Andy Whitcroft apw at canonical.com
Fri Jan 27 16:54:05 UTC 2012


For foundation's blueprint "IPv6 healthcheck for Precise" we were tasked
with reviewing all of the IPv6 configuration options.

I have now reviewed all the IPv6 kernel options, most are enabled and
those enabled seem appropriate:

    CONFIG_INET6_AH=m
    CONFIG_INET6_ESP=m
    CONFIG_INET6_IPCOMP=m
    CONFIG_INET6_TUNNEL=m
    CONFIG_INET6_XFRM_MODE_BEET=m
    CONFIG_INET6_XFRM_MODE_ROUTEOPTIMIZATION=m
    CONFIG_INET6_XFRM_MODE_TRANSPORT=m
    CONFIG_INET6_XFRM_MODE_TUNNEL=m
    CONFIG_INET6_XFRM_TUNNEL=m
    CONFIG_IPV6_MIP6=m
    CONFIG_IPV6_MULTIPLE_TABLES=y
    CONFIG_IPV6_NDISC_NODETYPE=y
    CONFIG_IPV6_PRIVACY=y
    CONFIG_IPV6_ROUTER_PREF=y
    CONFIG_IPV6_ROUTE_INFO=y
    CONFIG_IPV6_SIT=m
    CONFIG_IPV6_SIT_6RD=y
    CONFIG_IPV6_SUBTREES=y
    CONFIG_IPV6_TUNNEL=m

There are three options which are not enabled:

    # CONFIG_IPV6_OPTIMISTIC_DAD is not set

This seems to have some issues known in its implementation, so for now
it seems safer to leave it off as it is only a time optimisation and
enabling it later should be fine.

    # CONFIG_IPV6_MROUTE is not set
    # CONFIG_IPV6_PIMSM_V2 is not set
    # CONFIG_IPV6_MROUTE_MULTIPLE_TABLES is not set

This add support for multicast in IPv6.  Although it is marked EXPERIMENTAL
it does sound like something we should be enabling for testing in precise.

I have therefore prepared a patch to update these final three and pushed
it to precise.

-apw




More information about the kernel-team mailing list