Date: Tue, 14 Jul 2009 10:10:37 +0000 (UTC) From: "Bjoern A. Zeeb" <bz@FreeBSD.org> To: Qing Li <qingli@FreeBSD.org> Cc: svn-src-head@freebsd.org, svn-src-all@freebsd.org, src-committers@freebsd.org Subject: Re: svn commit: r195643 - head/sys/netinet6 Message-ID: <20090714100524.P245@maildrop.int.zabbadoz.net> In-Reply-To: <200907121920.n6CJKtf1092990@svn.freebsd.org> References: <200907121920.n6CJKtf1092990@svn.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, 12 Jul 2009, Qing Li wrote: > Author: qingli > Date: Sun Jul 12 19:20:55 2009 > New Revision: 195643 > URL: http://svn.freebsd.org/changeset/base/195643 > > Log: > This patch adds a host route to an interface address (that is assigned > to a non loopback/ppp link type) through the loopback interface. Prior > to the new L2/L3 rewrite, this host route was explicitly created when > processing the IPv6 address assignment. This loopback host route is > deleted when that IPv6 address is removed from the interface. Thanks a lot for fixing this. Now I am reviewing another patch and I found that this here used to be configurable actually by sysctls: nd6_useloopback for v6 and the v4 version from r192011 used to use useloopback; see the lines following pointing at the 7-STABLE tree to have a tree before new-arp: http://fxr.watson.org/fxr/source/netinet/if_ether.c?v=FREEBSD7#L259 http://fxr.watson.org/fxr/source/netinet6/nd6.c?v=FREEBSD7#L1320 Not sure if we'll want to restore that behaviour or remove the sysctls? /bz > Reviewed by: bz, gnn > Approved by: re > > Modified: > head/sys/netinet6/in6.c > > Modified: head/sys/netinet6/in6.c > ============================================================================== > --- head/sys/netinet6/in6.c Sun Jul 12 17:07:35 2009 (r195642) > +++ head/sys/netinet6/in6.c Sun Jul 12 19:20:55 2009 (r195643) > @@ -1194,6 +1194,25 @@ in6_purgeaddr(struct ifaddr *ifa) > ifa_ref(ifa0); > IF_ADDR_UNLOCK(ifp); > > + if (!(ia->ia_ifp->if_flags & (IFF_LOOPBACK | IFF_POINTOPOINT))) { > + struct rt_addrinfo info; > + struct sockaddr_dl null_sdl; > + > + bzero(&null_sdl, sizeof(null_sdl)); > + null_sdl.sdl_len = sizeof(null_sdl); > + null_sdl.sdl_family = AF_LINK; > + null_sdl.sdl_type = V_loif->if_type; > + null_sdl.sdl_index = V_loif->if_index; > + bzero(&info, sizeof(info)); > + info.rti_flags = ia->ia_flags | RTF_HOST | RTF_STATIC; > + info.rti_info[RTAX_DST] = (struct sockaddr *)&ia->ia_addr; > + info.rti_info[RTAX_GATEWAY] = (struct sockaddr *)&null_sdl; > + error = rtrequest1_fib(RTM_DELETE, &info, NULL, 0); > + > + if (error != 0) > + log(LOG_INFO, "in6_purgeaddr: deletion failed\n"); > + } > + > /* stop DAD processing */ > nd6_dad_stop(ifa); > > @@ -1755,6 +1774,33 @@ in6_ifinit(struct ifnet *ifp, struct in6 > ia->ia_flags |= IFA_ROUTE; > } > > + /* > + * add a loopback route to self > + */ > + if (!(ifp->if_flags & (IFF_LOOPBACK | IFF_POINTOPOINT))) { > + struct rt_addrinfo info; > + struct rtentry *rt = NULL; > + static struct sockaddr_dl null_sdl = {sizeof(null_sdl), AF_LINK}; > + > + bzero(&info, sizeof(info)); > + info.rti_ifp = V_loif; > + info.rti_flags = ia->ia_flags | RTF_HOST | RTF_STATIC; > + info.rti_info[RTAX_DST] = (struct sockaddr *)&ia->ia_addr; > + info.rti_info[RTAX_GATEWAY] = (struct sockaddr *)&null_sdl; > + error = rtrequest1_fib(RTM_ADD, &info, &rt, 0); > + > + if (error == 0 && rt != NULL) { > + RT_LOCK(rt); > + ((struct sockaddr_dl *)rt->rt_gateway)->sdl_type = > + rt->rt_ifp->if_type; > + ((struct sockaddr_dl *)rt->rt_gateway)->sdl_index = > + rt->rt_ifp->if_index; > + RT_REMREF(rt); > + RT_UNLOCK(rt); > + } else if (error != 0) > + log(LOG_INFO, "in6_ifinit: insertion failed\n"); > + } > + > /* Add ownaddr as loopback rtentry, if necessary (ex. on p2p link). */ > if (newhost) { > struct llentry *ln; > -- Bjoern A. Zeeb The greatest risk is not taking one.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20090714100524.P245>