Re: [Patch net-next v4 3/5] vxlan: add ipv6 support

[Cong Wang <amwang@redhat.com>]

On Wed, 2013-04-17 at 11:15 +0300, Mike Rapoport wrote:
> On Wed, Apr 17, 2013 at 01:10:20PM +0800, Cong Wang wrote:
> >  #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
> > @@ -42,6 +41,11 @@
> >  #include <net/inet_ecn.h>
> >  #include <net/net_namespace.h>
> >  #include <net/netns/generic.h>
> > +#if IS_ENABLED(CONFIG_IPV6)
> 
> It seems to me that some of the #ifdefs may be removed, especially on
> the rtnetlink path.

Which one? If you mean the one above, I know it can be removed, but it
can save some compile time if we have it.

> 
> > +#include <net/addrconf.h>
> > +#include <net/ip6_route.h>
> > +#include <net/ip6_tunnel.h>
> > +#endif
> >
...
> >
> > +	if (data[IFLA_VXLAN_REMOTE]) {
> > +		__be32 gaddr = nla_get_be32(data[IFLA_VXLAN_REMOTE]);
> > +		if (!IN_MULTICAST(ntohl(gaddr))) {
> > +			pr_debug("group address is not IPv4 multicast\n");
> > +			return -EADDRNOTAVAIL;
> > +		}
> > +	} else if (data[IFLA_VXLAN_REMOTE6]) {
> > +#if IS_ENABLED(CONFIG_IPV6)
> > +		struct in6_addr gaddr;
> > +		nla_memcpy(&gaddr, data[IFLA_VXLAN_REMOTE6], sizeof(gaddr));
> > +		if (!ipv6_addr_is_multicast(&gaddr)) {
> > +			pr_debug("group address is not IPv6 multicast\n");
> > +			return -EADDRNOTAVAIL;
> > +		}
> > +#else
> > +		return -EPFNOSUPPORT;
> > +#endif
> > +	}
> 
> Adding back the test for multicast address contradicts the ability to
> use unicast as default remote destination.

Oops, it is definitely my mistake when handling the conflicts here.
Fixed now.

Thanks.