Dan Smith wrote:
> BH> Is using IFA_F_PERMANENT correct here? Should you save the flags
> BH> from the address when checkpointing? Permanent means it was added
> BH> by the user, not by the kernel, so you could be changing things
> BH> slightly.
>
> Does the kernel create global scope addresses? Maybe it does in some
> more advanced IPv6 environments, but it seemed like excluding global
> scope addresses in checkpoint meant that we only saved (and thus
> restore) the permanent ones anyway.
It adds global-scope addresses through the auto-configuration process,
i.e. by receiving a prefix in a router advertisement.
> I guess it's a better idea to just save the flags anyhow now that I
> have a way to restore them.
Yes, and I just realized something else. This code:
+ ret = inet6_addr_add(net, dev->ifindex, &addr->inet6_addr,
+ addr->inet6_prefix_len, IFA_F_PERMANENT,
+ INFINITY_LIFE_TIME, INFINITY_LIFE_TIME);
isn't using the saved lifetimes either, so it won't ever go away.
And calling inet6_addr_add() isn't correct in all cases - you can
use it for manually-configured addresses (marked permanent), but
not for those added through address-autoconfiguration - for those
you'll want to use ipv6_add_addr(). But if you do that you'll
need to duplicate what's done after the add succeeds:
ifp = ipv6_add_addr(idev, pfx, plen, addr_scope, addr_flags);
if (!IS_ERR(ifp)) {
spin_lock_bh(&ifp->lock);
ifp->valid_lft = valid_lft;
ifp->prefered_lft = prefered_lft;
ifp->tstamp = jiffies;
spin_unlock_bh(&ifp->lock);
addrconf_prefix_route(&ifp->addr, ifp->prefix_len, dev,
expires, flags);
/*
* Note that section 3.1 of RFC 4429 indicates
* that the Optimistic flag should not be set for
* manually configured addresses
*/
addrconf_dad_start(ifp, 0);
in6_ifa_put(ifp);
addrconf_verify(0);
return 0;
}
That's just an example, not exactly correct.
-Brian
_______________________________________________
Containers mailing list
Containers@xxxxxxxxxxxxxxxxxxxxxxxxxx
https://lists.linux-foundation.org/mailman/listinfo/containers
