Merge branch 'master' into docuser

* New manpage: mtracebis.rst
* Makefile.am includes mtracebis.rst
* configure.ac lines removed
* Debian packaging files updated
* Fixed up manpage |seealso-programs| in the process
* Centos7 build package list updated to include systemd-devel
* New paragraph on netns vrf support in zebra manpage

Conflicts:
	configure.ac
	debianpkg/backports/ubuntu14.04/debian/frr.install
	debianpkg/frr.install
	doc/Makefile.am
	doc/developer/Building_FRR_on_CentOS7.rst
	doc/zebra.8.in

Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>

COMMUNITY.md

@@ -492,6 +492,36 @@ In all cases, compatibility pieces should be marked with compiler/preprocessor
 annotations to print warnings at compile time, pointing to the appropriate
 update path.  A `-Werror` build should fail if compatibility bits are used.
 
+### Release Process/Schedule
+
+FRR employs a <MAJOR>.<MINOR>.<BUGFIX> versioning scheme.
+
+* MAJOR - Significant new features or multiple minor features
+  A example of a MAJOR feature is a New Routing Protocol
+* MINOR - Smaller Features
+  A example of a MINOR feature is the addition of the BGP Shutdown feature.
+* BUGFIX - Fixes for actual bugs and/or security issues.
+
+We will pull a new development branch for the next release every 4 months.
+The current schedule is Feb/June/October 1.  The decision for a MAJOR/MINOR
+release is made at the time of branch pull based on what has been received
+the previous 4 months.  The branch name will be dev/MAJOR.MINOR.  At
+this point in time the master branch configure.ac and packaging systems
+will be updated to reflect the next possible release name to allow
+for easy distinguishing.  Additionally the new dev branch will have
+these files updated too.
+
+After one month the development branch will be renamed to
+stable/MAJOR.MINOR.  This process is not held up unless a crash or
+security issue has been found and needs to be addressed.  Issues
+being fixed will not cause a delay.
+
+Bug fix releases are at 1 month intervals until next MAJOR.MINOR is
+pulled.  Then at that time as needed for issues filed.
+
+Security issues are fixed for 1 year minimum on old releases and
+normal bug fixes for the current and previous release
+
 ### Miscellaneous
 
 When in doubt, follow the guidelines in the Linux kernel style guide, or ask on

bgpd/bgp_dump.c

@@ -465,7 +465,7 @@ static void bgp_dump_common(struct stream *obuf, struct peer *peer,
 	}
 
 	if (peer->su.sa.sa_family == AF_INET) {
-		stream_putw(obuf, peer->ifindex);
+		stream_putw(obuf, peer->ifp ? peer->ifp->ifindex : 0);
 		stream_putw(obuf, AFI_IP);
 
 		stream_put(obuf, &peer->su.sin.sin_addr, IPV4_MAX_BYTELEN);
@@ -477,7 +477,7 @@ static void bgp_dump_common(struct stream *obuf, struct peer *peer,
 			stream_put(obuf, empty, IPV4_MAX_BYTELEN);
 	} else if (peer->su.sa.sa_family == AF_INET6) {
 		/* Interface Index and Address family. */
-		stream_putw(obuf, peer->ifindex);
+		stream_putw(obuf, peer->ifp ? peer->ifp->ifindex : 0);
 		stream_putw(obuf, AFI_IP6);
 
 		/* Source IP Address and Destination IP Address. */

bgpd/bgp_evpn.c

@@ -666,10 +666,11 @@ static void build_evpn_route_extcomm(struct bgpevpn *vpn, struct attr *attr,
 	for (ALL_LIST_ELEMENTS(vpn->export_rtl, node, nnode, ecom))
 		attr->ecommunity = ecommunity_merge(attr->ecommunity, ecom);
 
-	/* Add the export RTs for L3VNI - currently only supported for IPV4 host
+	/*
-	 * routes
+	 * only attach l3-vni export rts for ipv4 address family and if we are
+	 * advertising both the labels in type-2 routes
 	 */
-	if (afi == AFI_IP) {
+	if (afi == AFI_IP && CHECK_FLAG(vpn->flags, VNI_FLAG_USE_TWO_LABELS)) {
 		vrf_export_rtl = bgpevpn_get_vrf_export_rtl(vpn);
 		if (vrf_export_rtl && !list_isempty(vrf_export_rtl)) {
 			for (ALL_LIST_ELEMENTS(vrf_export_rtl, node, nnode,
@@ -690,7 +691,12 @@ static void build_evpn_route_extcomm(struct bgpevpn *vpn, struct attr *attr,
 			ecommunity_merge(attr->ecommunity, &ecom_sticky);
 	}
 
-	if (afi == AFI_IP && !is_zero_mac(&attr->rmac)) {
+	/*
+	 * only attach l3-vni rmac for ipv4 address family and if we are
+	 * advertising both the labels in type-2 routes
+	 */
+	if (afi == AFI_IP && !is_zero_mac(&attr->rmac) &&
+	    CHECK_FLAG(vpn->flags, VNI_FLAG_USE_TWO_LABELS)) {
 		memset(&ecom_rmac, 0, sizeof(ecom_rmac));
 		encode_rmac_extcomm(&eval_rmac, &attr->rmac);
 		ecom_rmac.size = 1;
@@ -1189,8 +1195,13 @@ static int update_evpn_route_entry(struct bgp *bgp, struct bgpevpn *vpn,
 
 		/* The VNI goes into the 'label' field of the route */
 		vni2label(vpn->vni, &label[0]);
-		/* Type-2 routes may carry a second VNI - the L3-VNI */
+
-		if (evp->prefix.route_type == BGP_EVPN_MAC_IP_ROUTE) {
+		/* Type-2 routes may carry a second VNI - the L3-VNI.
+		 * Only attach second label if we are advertising two labels for
+		 * type-2 routes.
+		 */
+		if (evp->prefix.route_type == BGP_EVPN_MAC_IP_ROUTE &&
+		    CHECK_FLAG(vpn->flags, VNI_FLAG_USE_TWO_LABELS)) {
 			vni_t l3vni;
 
 			l3vni = bgpevpn_get_l3vni(vpn);
@@ -1209,6 +1220,24 @@ static int update_evpn_route_entry(struct bgp *bgp, struct bgpevpn *vpn,
 		    && !CHECK_FLAG(tmp_ri->flags, BGP_INFO_REMOVED))
 			route_change = 0;
 		else {
+			/*
+			 * The attributes have changed, type-2 routes needs to
+			 * be advertised with right labels.
+			 */
+			vni2label(vpn->vni, &label[0]);
+			if (evp->prefix.route_type == BGP_EVPN_MAC_IP_ROUTE &&
+			    CHECK_FLAG(vpn->flags, VNI_FLAG_USE_TWO_LABELS)) {
+				vni_t l3vni;
+
+				l3vni = bgpevpn_get_l3vni(vpn);
+				if (l3vni) {
+					vni2label(l3vni, &label[1]);
+					num_labels++;
+				}
+			}
+			memcpy(&tmp_ri->extra->label, label, sizeof(label));
+			tmp_ri->extra->num_labels = num_labels;
+
 			/* The attribute has changed. */
 			/* Add (or update) attribute to hash. */
 			attr_new = bgp_attr_intern(attr);
@@ -1254,7 +1283,13 @@ static int update_evpn_route(struct bgp *bgp, struct bgpevpn *vpn,
 	attr.mp_nexthop_len = BGP_ATTR_NHLEN_IPV4;
 	attr.sticky = CHECK_FLAG(flags, ZEBRA_MACIP_TYPE_STICKY) ? 1 : 0;
 	attr.default_gw = CHECK_FLAG(flags, ZEBRA_MACIP_TYPE_GW) ? 1 : 0;
+
+	/* PMSI is only needed for type-3 routes */
+	if (p->prefix.route_type == BGP_EVPN_IMET_ROUTE)
 		attr.flag |= ATTR_FLAG_BIT(BGP_ATTR_PMSI_TUNNEL);
+
+	/* router mac is only needed for type-2 and type-5  routes */
+	if (p->prefix.route_type == BGP_EVPN_MAC_IP_ROUTE)
 		bgpevpn_get_rmac(vpn, &attr.rmac);
 	vni2label(vpn->vni, &(attr.label));
 
@@ -3275,6 +3310,19 @@ void bgp_evpn_advertise_type5_routes(struct bgp *bgp_vrf,
 		 */
 		for (ri = rn->info; ri; ri = ri->next) {
 			if (CHECK_FLAG(ri->flags, BGP_INFO_SELECTED)) {
+
+				/* apply the route-map */
+				if (bgp_vrf->adv_cmd_rmap[afi][safi].map) {
+					int ret = 0;
+
+					ret =
+						route_map_apply(
+							bgp_vrf->adv_cmd_rmap[afi][safi].map,
+							&rn->p, RMAP_BGP, ri);
+					if (ret == RMAP_DENYMATCH)
+						continue;
+				}
+
 				bgp_evpn_advertise_type5_route(bgp_vrf, &rn->p,
 							       ri->attr,
 							       afi, safi);
@@ -4214,7 +4262,8 @@ static void link_l2vni_hash_to_l3vni(struct hash_backet *backet,
 int bgp_evpn_local_l3vni_add(vni_t l3vni,
 			     vrf_id_t vrf_id,
 			     struct ethaddr *rmac,
-			     struct in_addr originator_ip)
+			     struct in_addr originator_ip,
+			     int filter)
 {
 	struct bgp *bgp_vrf = NULL; /* bgp VRF instance */
 	struct bgp *bgp_def = NULL; /* default bgp instance */
@@ -4266,6 +4315,10 @@ int bgp_evpn_local_l3vni_add(vni_t l3vni,
 	/* set the originator ip */
 	bgp_vrf->originator_ip = originator_ip;
 
+	/* set the right filter - are we using l3vni only for prefix routes? */
+	if (filter)
+		SET_FLAG(bgp_vrf->vrf_flags, BGP_VRF_L3VNI_PREFIX_ROUTES_ONLY);
+
 	/* auto derive RD/RT */
 	if (!CHECK_FLAG(bgp_vrf->vrf_flags, BGP_VRF_IMPORT_RT_CFGD))
 		evpn_auto_rt_import_add_for_vrf(bgp_vrf);
@@ -4279,7 +4332,10 @@ int bgp_evpn_local_l3vni_add(vni_t l3vni,
 			link_l2vni_hash_to_l3vni,
 		     bgp_vrf);
 
-	/* updates all corresponding local mac-ip routes */
+	/* Only update all corresponding type-2 routes if we are advertising two
+	 * labels along with type-2 routes
+	 */
+	if (!filter)
 		for (ALL_LIST_ELEMENTS_RO(bgp_vrf->l2vnis, node, vpn))
 			update_routes_for_vni(bgp_def, vpn);
 
@@ -4340,9 +4396,12 @@ int bgp_evpn_local_l3vni_del(vni_t l3vni,
 	}
 
 	/* update all corresponding local mac-ip routes */
-	for (ALL_LIST_ELEMENTS_RO(bgp_vrf->l2vnis, node, vpn))
+	if (!CHECK_FLAG(bgp_vrf->vrf_flags, BGP_VRF_L3VNI_PREFIX_ROUTES_ONLY)) {
+		for (ALL_LIST_ELEMENTS_RO(bgp_vrf->l2vnis, node, vpn)) {
+			UNSET_FLAG(vpn->flags, VNI_FLAG_USE_TWO_LABELS);
 			update_routes_for_vni(bgp_def, vpn);
-
+		}
+	}
 
 	/* Delete the instance if it was autocreated */
 	if (CHECK_FLAG(bgp_vrf->vrf_flags, BGP_VRF_AUTO))

bgpd/bgp_evpn.h

@@ -91,7 +91,8 @@ extern int bgp_evpn_local_macip_add(struct bgp *bgp, vni_t vni,
 				    u_char flags);
 extern int bgp_evpn_local_l3vni_add(vni_t vni, vrf_id_t vrf_id,
 				    struct ethaddr *rmac,
-				    struct in_addr originator_ip);
+				    struct in_addr originator_ip,
+				    int filter);
 extern int bgp_evpn_local_l3vni_del(vni_t vni, vrf_id_t vrf_id);
 extern int bgp_evpn_local_vni_del(struct bgp *bgp, vni_t vni);
 extern int bgp_evpn_local_vni_add(struct bgp *bgp, vni_t vni,

bgpd/bgp_evpn_private.h

@@ -61,6 +61,10 @@ struct bgpevpn {
 #define VNI_FLAG_RD_CFGD           0x4  /* RD is user configured. */
 #define VNI_FLAG_IMPRT_CFGD        0x8  /* Import RT is user configured */
 #define VNI_FLAG_EXPRT_CFGD        0x10 /* Export RT is user configured */
+#define VNI_FLAG_USE_TWO_LABELS    0x20 /* Attach both L2-VNI and L3-VNI if
+					   needed for this VPN */
+
+	struct bgp *bgp_vrf; /* back pointer to the vrf instance */
 
 	/* Flag to indicate if we are advertising the g/w mac ip for this VNI*/
 	u_int8_t advertise_gw_macip;
@@ -132,66 +136,66 @@ static inline int bgp_evpn_vrf_rd_matches_existing(struct bgp *bgp_vrf,
 
 static inline vni_t bgpevpn_get_l3vni(struct bgpevpn *vpn)
 {
-	struct bgp *bgp_vrf = NULL;
+	return vpn->bgp_vrf ? vpn->bgp_vrf->l3vni : 0;
-
-	bgp_vrf = bgp_lookup_by_vrf_id(vpn->tenant_vrf_id);
-	if (!bgp_vrf)
-		return 0;
-
-	return bgp_vrf->l3vni;
 }
 
 static inline void bgpevpn_get_rmac(struct bgpevpn *vpn, struct ethaddr *rmac)
 {
-	struct bgp *bgp_vrf = NULL;
-
 	memset(rmac, 0, sizeof(struct ethaddr));
-	bgp_vrf = bgp_lookup_by_vrf_id(vpn->tenant_vrf_id);
+	if (!vpn->bgp_vrf)
-	if (!bgp_vrf)
 		return;
-	memcpy(rmac, &bgp_vrf->rmac, sizeof(struct ethaddr));
+	memcpy(rmac, &vpn->bgp_vrf->rmac, sizeof(struct ethaddr));
 }
 
 static inline struct list *bgpevpn_get_vrf_export_rtl(struct bgpevpn *vpn)
 {
-	struct bgp *bgp_vrf = NULL;
+	if (!vpn->bgp_vrf)
-
-	bgp_vrf = bgp_lookup_by_vrf_id(vpn->tenant_vrf_id);
-	if (!bgp_vrf)
 		return NULL;
 
-	return bgp_vrf->vrf_export_rtl;
+	return vpn->bgp_vrf->vrf_export_rtl;
 }
 
 static inline struct list *bgpevpn_get_vrf_import_rtl(struct bgpevpn *vpn)
 {
-	struct bgp *bgp_vrf = NULL;
+	if (!vpn->bgp_vrf)
-
-	bgp_vrf = bgp_lookup_by_vrf_id(vpn->tenant_vrf_id);
-	if (!bgp_vrf)
 		return NULL;
 
-	return bgp_vrf->vrf_import_rtl;
+	return vpn->bgp_vrf->vrf_import_rtl;
 }
 
 static inline void bgpevpn_unlink_from_l3vni(struct bgpevpn *vpn)
 {
-	struct bgp *bgp_vrf = NULL;
+	/* bail if vpn is not associated to bgp_vrf */
-
+	if (!vpn->bgp_vrf)
-	bgp_vrf = bgp_lookup_by_vrf_id(vpn->tenant_vrf_id);
-	if (!bgp_vrf || !bgp_vrf->l2vnis)
 		return;
-	listnode_delete(bgp_vrf->l2vnis, vpn);
+  
+	UNSET_FLAG(vpn->flags, VNI_FLAG_USE_TWO_LABELS);
+	listnode_delete(vpn->bgp_vrf->l2vnis, vpn);
+  
+	/* remove the backpointer to the vrf instance */
+	vpn->bgp_vrf = NULL;
 }
 
 static inline void bgpevpn_link_to_l3vni(struct bgpevpn *vpn)
 {
 	struct bgp *bgp_vrf = NULL;
 
-	bgp_vrf = bgp_lookup_by_vrf_id(vpn->tenant_vrf_id);
+	/* bail if vpn is already associated to vrf */
-	if (!bgp_vrf || !bgp_vrf->l2vnis)
+	if (vpn->bgp_vrf)
 		return;
+
+	bgp_vrf = bgp_lookup_by_vrf_id(vpn->tenant_vrf_id);
+	if (!bgp_vrf)
+		return;
+
+	/* associate the vpn to the bgp_vrf instance */
+	vpn->bgp_vrf = bgp_vrf;
 	listnode_add_sort(bgp_vrf->l2vnis, vpn);
+
+	/* check if we are advertising two labels for this vpn */
+	if (!CHECK_FLAG(bgp_vrf->vrf_flags,
+		       BGP_VRF_L3VNI_PREFIX_ROUTES_ONLY))
+		SET_FLAG(vpn->flags, VNI_FLAG_USE_TWO_LABELS);
 }
 
 static inline int is_vni_configured(struct bgpevpn *vpn)

bgpd/bgp_evpn_vty.c

@@ -83,7 +83,7 @@ static void display_vrf_import_rt(struct vty *vty,
 	case ECOMMUNITY_ENCODE_AS:
 		eas.as = (*pnt++ << 8);
 		eas.as |= (*pnt++);
-		pnt = ptr_get_be32(pnt, &eas.val);
+		ptr_get_be32(pnt, &eas.val);
 
 		snprintf(rt_buf, RT_ADDRSTRLEN, "%u:%u", eas.as, eas.val);
 
@@ -195,7 +195,7 @@ static void display_import_rt(struct vty *vty, struct irt_node *irt,
 	case ECOMMUNITY_ENCODE_AS:
 		eas.as = (*pnt++ << 8);
 		eas.as |= (*pnt++);
-		pnt = ptr_get_be32(pnt, &eas.val);
+		ptr_get_be32(pnt, &eas.val);
 
 		snprintf(rt_buf, RT_ADDRSTRLEN, "%u:%u", eas.as, eas.val);
 
@@ -491,53 +491,6 @@ static void display_vni(struct vty *vty, struct bgpevpn *vpn, json_object *json)
 		json_object_object_add(json, "exportRts", json_export_rtl);
 }
 
-static void evpn_show_vrf_routes(struct vty *vty,
-				 struct bgp *bgp_vrf)
-{
-	struct bgp *bgp_def = NULL;
-	struct bgp_node *rn;
-	struct bgp_info *ri;
-	int header = 1;
-	u_int32_t prefix_cnt, path_cnt;
-	struct bgp_table *table;
-
-	prefix_cnt = path_cnt = 0;
-	bgp_def = bgp_get_default();
-	if (!bgp_def)
-		return;
-
-	table = (struct bgp_table *)bgp_vrf->rib[AFI_L2VPN][SAFI_EVPN];
-	for (rn = bgp_table_top(table); rn; rn = bgp_route_next(rn)) {
-		char prefix_str[BUFSIZ];
-
-		bgp_evpn_route2str((struct prefix_evpn *)&rn->p, prefix_str,
-				   sizeof(prefix_str));
-
-		if (rn->info) {
-			/* Overall header/legend displayed once. */
-			if (header) {
-				bgp_evpn_show_route_header(vty, bgp_def, NULL);
-				header = 0;
-			}
-			prefix_cnt++;
-		}
-
-		/* For EVPN, the prefix is displayed for each path (to fit in
-		 * with code that already exists).
-		 */
-		for (ri = rn->info; ri; ri = ri->next) {
-			route_vty_out(vty, &rn->p, ri, 0, SAFI_EVPN, NULL);
-			path_cnt++;
-		}
-	}
-
-	if (prefix_cnt == 0)
-		vty_out(vty, "No EVPN prefixes exist for this VRF");
-	else
-		vty_out(vty, "\nDisplayed %u prefixes (%u paths)",
-			prefix_cnt, path_cnt);
-}
-
 static void show_vni_routes(struct bgp *bgp, struct bgpevpn *vpn, int type,
 			    struct vty *vty, struct in_addr vtep_ip,
 			    json_object *json)
@@ -2723,19 +2676,34 @@ DEFUN (no_bgp_evpn_advertise_vni_subnet,
 
 DEFUN (bgp_evpn_advertise_type5,
        bgp_evpn_advertise_type5_cmd,
-       "advertise " BGP_AFI_CMD_STR "" BGP_SAFI_CMD_STR,
+       "advertise " BGP_AFI_CMD_STR "" BGP_SAFI_CMD_STR " [route-map WORD]",
        "Advertise prefix routes\n"
        BGP_AFI_HELP_STR
-       BGP_SAFI_HELP_STR)
+       BGP_SAFI_HELP_STR
+       "route-map for filtering specific routes\n"
+       "Name of the route map\n")
 {
 	struct bgp *bgp_vrf = VTY_GET_CONTEXT(bgp); /* bgp vrf instance */
 	int idx_afi = 0;
 	int idx_safi = 0;
+	int idx_rmap = 0;
 	afi_t afi = 0;
 	safi_t safi = 0;
+	int ret = 0;
+	int rmap_changed = 0;
 
 	argv_find_and_parse_afi(argv, argc, &idx_afi, &afi);
 	argv_find_and_parse_safi(argv, argc, &idx_safi, &safi);
+	ret = argv_find(argv, argc, "route-map", &idx_rmap);
+	if (ret) {
+		if (!bgp_vrf->adv_cmd_rmap[afi][safi].name)
+			rmap_changed = 1;
+		else if (strcmp(argv[idx_rmap + 1]->arg,
+				bgp_vrf->adv_cmd_rmap[afi][safi].name) != 0)
+			rmap_changed = 1;
+	} else if (bgp_vrf->adv_cmd_rmap[afi][safi].name) {
+		rmap_changed = 1;
+	}
 
 	if (!(afi == AFI_IP) || (afi == AFI_IP6)) {
 		vty_out(vty,
@@ -2754,24 +2722,44 @@ DEFUN (bgp_evpn_advertise_type5,
 		/* if we are already advertising ipv4 prefix as type-5
 		 * nothing to do
 		 */
-		if (!CHECK_FLAG(bgp_vrf->vrf_flags,
+		if (!rmap_changed && CHECK_FLAG(bgp_vrf->vrf_flags,
-				BGP_VRF_ADVERTISE_IPV4_IN_EVPN)) {
+						BGP_VRF_ADVERTISE_IPV4_IN_EVPN))
+			return CMD_WARNING;
 		SET_FLAG(bgp_vrf->vrf_flags,
 			 BGP_VRF_ADVERTISE_IPV4_IN_EVPN);
-			bgp_evpn_advertise_type5_routes(bgp_vrf, afi, safi);
-		}
 	} else {
 
 		/* if we are already advertising ipv6 prefix as type-5
 		 * nothing to do
 		 */
-		if (!CHECK_FLAG(bgp_vrf->vrf_flags,
+		if (!rmap_changed && CHECK_FLAG(bgp_vrf->vrf_flags,
-				BGP_VRF_ADVERTISE_IPV6_IN_EVPN)) {
+						BGP_VRF_ADVERTISE_IPV6_IN_EVPN))
+			return CMD_WARNING;
 		SET_FLAG(bgp_vrf->vrf_flags,
 			 BGP_VRF_ADVERTISE_IPV6_IN_EVPN);
+	}
+
+	if (rmap_changed) {
+		bgp_evpn_withdraw_type5_routes(bgp_vrf, afi, safi);
+		if (bgp_vrf->adv_cmd_rmap[afi][safi].name) {
+			XFREE(MTYPE_ROUTE_MAP_NAME,
+			      bgp_vrf->adv_cmd_rmap[afi][safi].name);
+			bgp_vrf->adv_cmd_rmap[afi][safi].name = NULL;
+			bgp_vrf->adv_cmd_rmap[afi][safi].map = NULL;
+		}
+	}
+
+	/* set the route-map for advertise command */
+	if (ret && argv[idx_rmap + 1]->arg) {
+		bgp_vrf->adv_cmd_rmap[afi][safi].name =
+			XSTRDUP(MTYPE_ROUTE_MAP_NAME,
+				argv[idx_rmap + 1]->arg);
+		bgp_vrf->adv_cmd_rmap[afi][safi].map =
+			route_map_lookup_by_name(argv[idx_rmap + 1]->arg);
+	}
+
+	/* advertise type-5 routes */
 	bgp_evpn_advertise_type5_routes(bgp_vrf, afi, safi);
-		}
-	}
 	return CMD_SUCCESS;
 }
 
@@ -2827,6 +2815,15 @@ DEFUN (no_bgp_evpn_advertise_type5,
 				   BGP_VRF_ADVERTISE_IPV6_IN_EVPN);
 		}
 	}
+
+	/* clear the route-map information for advertise ipv4/ipv6 unicast */
+	if (bgp_vrf->adv_cmd_rmap[afi][safi].name) {
+		XFREE(MTYPE_ROUTE_MAP_NAME,
+		      bgp_vrf->adv_cmd_rmap[afi][safi].name);
+		bgp_vrf->adv_cmd_rmap[afi][safi].name = NULL;
+		bgp_vrf->adv_cmd_rmap[afi][safi].map = NULL;
+	}
+
 	return CMD_SUCCESS;
 }
 
@@ -2923,16 +2920,23 @@ DEFUN(show_bgp_l2vpn_evpn_vni,
  */
 DEFUN(show_bgp_l2vpn_evpn_summary,
       show_bgp_l2vpn_evpn_summary_cmd,
-      "show bgp l2vpn evpn summary [json]",
+      "show bgp [vrf VRFNAME] l2vpn evpn summary [json]",
       SHOW_STR
       BGP_STR
+      "bgp vrf\n"
+      "vrf name\n"
       L2VPN_HELP_STR
       EVPN_HELP_STR
       "Summary of BGP neighbor status\n"
       JSON_STR)
 {
+	int idx_vrf = 0;
 	u_char uj = use_json(argc, argv);
-	return bgp_show_summary_vty(vty, NULL, AFI_L2VPN, SAFI_EVPN, uj);
+	char *vrf = NULL;
+
+	if (argv_find(argv, argc, "vrf", &idx_vrf))
+		vrf = argv[++idx_vrf]->arg;
+	return bgp_show_summary_vty(vty, vrf, AFI_L2VPN, SAFI_EVPN, uj);
 }
 
 /*
@@ -3139,33 +3143,6 @@ DEFUN(show_bgp_l2vpn_evpn_route_rd_macip,
 	return CMD_SUCCESS;
 }
 
-/*
- * Display per-VRF EVPN routing table.
- */
-DEFUN(show_bgp_l2vpn_evpn_route_vrf, show_bgp_l2vpn_evpn_route_vrf_cmd,
-      "show bgp l2vpn evpn route vrf VRFNAME",
-      SHOW_STR
-      BGP_STR
-      L2VPN_HELP_STR
-      EVPN_HELP_STR
-      "EVPN route information\n"
-      "VRF\n"
-      "VRF Name\n")
-{
-	int vrf_idx = 6;
-	char *vrf_name = NULL;
-	struct bgp *bgp_vrf = NULL;
-
-	vrf_name = argv[vrf_idx]->arg;
-	bgp_vrf = bgp_lookup_by_name(vrf_name);
-	if (!bgp_vrf)
-		return CMD_WARNING;
-
-	evpn_show_vrf_routes(vty, bgp_vrf);
-
-	return CMD_SUCCESS;
-}
-
 /*
  * Display per-VNI EVPN routing table.
  */
@@ -3897,6 +3874,10 @@ DEFUN (show_bgp_vrf_l3vni_info,
 		vty_out(vty, "  L3-VNI: %u\n", bgp->l3vni);
 		vty_out(vty, "  Rmac: %s\n",
 			prefix_mac2str(&bgp->rmac, buf, sizeof(buf)));
+		vty_out(vty, "  VNI Filter: %s\n",
+			CHECK_FLAG(bgp->vrf_flags,
+				   BGP_VRF_L3VNI_PREFIX_ROUTES_ONLY) ?
+				"prefix-routes-only" : "none");
 		vty_out(vty, "  L2-VNI List:\n");
 		vty_out(vty, "    ");
 		for (ALL_LIST_ELEMENTS_RO(bgp->l2vnis, node, vpn))
@@ -3922,6 +3903,10 @@ DEFUN (show_bgp_vrf_l3vni_info,
 		json_object_string_add(json, "rmac",
 				       prefix_mac2str(&bgp->rmac, buf,
 						      sizeof(buf)));
+		json_object_string_add(json, "vniFilter",
+				       CHECK_FLAG(bgp->vrf_flags,
+						  BGP_VRF_L3VNI_PREFIX_ROUTES_ONLY)
+				       ? "prefix-routes-only" : "none");
 		/* list of l2vnis */
 		for (ALL_LIST_ELEMENTS_RO(bgp->l2vnis, node, vpn))
 			json_object_array_add(json_vnis,
@@ -4325,6 +4310,8 @@ DEFUN (no_bgp_evpn_vni_rt_without_val,
 void bgp_config_write_evpn_info(struct vty *vty, struct bgp *bgp, afi_t afi,
 				safi_t safi)
 {
+	char buf1[RD_ADDRSTRLEN];
+
 	if (bgp->vnihash)
 		hash_iterate(bgp->vnihash,
 			     (void (*)(struct hash_backet *,
@@ -4342,6 +4329,42 @@ void bgp_config_write_evpn_info(struct vty *vty, struct bgp *bgp, afi_t afi,
 
 	if (CHECK_FLAG(bgp->vrf_flags, BGP_VRF_ADVERTISE_IPV6_IN_EVPN))
 		vty_out(vty, "  advertise ipv6 unicast\n");
+
+	if (CHECK_FLAG(bgp->vrf_flags, BGP_VRF_RD_CFGD))
+		vty_out(vty, "   rd %s\n",
+			prefix_rd2str(&bgp->vrf_prd, buf1, sizeof(buf1)));
+
+	/* import route-target */
+	if (CHECK_FLAG(bgp->vrf_flags, BGP_VRF_IMPORT_RT_CFGD)) {
+		char *ecom_str;
+		struct listnode *node, *nnode;
+		struct ecommunity *ecom;
+
+		for (ALL_LIST_ELEMENTS(bgp->vrf_import_rtl, node, nnode,
+				       ecom)) {
+			ecom_str = ecommunity_ecom2str(
+				ecom, ECOMMUNITY_FORMAT_ROUTE_MAP, 0);
+			vty_out(vty, "   route-target import %s\n",
+				ecom_str);
+			XFREE(MTYPE_ECOMMUNITY_STR, ecom_str);
+		}
+	}
+
+	/* export route-target */
+	if (CHECK_FLAG(bgp->vrf_flags, BGP_VRF_EXPORT_RT_CFGD)) {
+		char *ecom_str;
+		struct listnode *node, *nnode;
+		struct ecommunity *ecom;
+
+		for (ALL_LIST_ELEMENTS(bgp->vrf_export_rtl, node, nnode,
+				       ecom)) {
+			ecom_str = ecommunity_ecom2str(
+				ecom, ECOMMUNITY_FORMAT_ROUTE_MAP, 0);
+			vty_out(vty, "   route-target export %s\n",
+				ecom_str);
+			XFREE(MTYPE_ECOMMUNITY_STR, ecom_str);
+		}
+	}
 }
 
 void bgp_ethernetvpn_init(void)
@@ -4379,7 +4402,6 @@ void bgp_ethernetvpn_init(void)
 	install_element(VIEW_NODE, &show_bgp_l2vpn_evpn_route_rd_cmd);
 	install_element(VIEW_NODE, &show_bgp_l2vpn_evpn_route_rd_macip_cmd);
 	install_element(VIEW_NODE, &show_bgp_l2vpn_evpn_route_vni_cmd);
-	install_element(VIEW_NODE, &show_bgp_l2vpn_evpn_route_vrf_cmd);
 	install_element(VIEW_NODE,
 			&show_bgp_l2vpn_evpn_route_vni_multicast_cmd);
 	install_element(VIEW_NODE, &show_bgp_l2vpn_evpn_route_vni_macip_cmd);

bgpd/bgp_fsm.c

@@ -289,7 +289,8 @@ void bgp_timer_set(struct peer *peer)
 		/* First entry point of peer's finite state machine.  In Idle
 		   status start timer is on unless peer is shutdown or peer is
 		   inactive.  All other timer must be turned off */
-		if (BGP_PEER_START_SUPPRESSED(peer) || !peer_active(peer)) {
+		if (BGP_PEER_START_SUPPRESSED(peer) || !peer_active(peer)
+		    || peer->bgp->vrf_id == VRF_UNKNOWN) {
 			BGP_TIMER_OFF(peer->t_start);
 		} else {
 			BGP_TIMER_ON(peer->t_start, bgp_start_timer,
@@ -1376,6 +1377,15 @@ int bgp_start(struct peer *peer)
 		return 0;
 	}
 
+	if (peer->bgp &&
+	    peer->bgp->vrf_id == VRF_UNKNOWN) {
+		if (bgp_debug_neighbor_events(peer))
+			zlog_err(
+				 "%s [FSM] In a VRF that is not initialised yet",
+				 peer->host);
+		return -1;
+	}
+
 	/* Register to be notified on peer up */
 	if (peer->sort == BGP_PEER_EBGP && peer->ttl == 1
 	    && !CHECK_FLAG(peer->flags, PEER_FLAG_DISABLE_CONNECTED_CHECK)

bgpd/bgp_keepalives.c

@@ -119,7 +119,7 @@ static void peer_process(struct hash_backet *hb, void *arg)
 	}
 
 	/* if calculated next update for this peer < current delay, use it */
-	if (next_update->tv_sec <= 0 || timercmp(&diff, next_update, <))
+	if (next_update->tv_sec < 0 || timercmp(&diff, next_update, <))
 		*next_update = diff;
 }
 

bgpd/bgp_main.c

@@ -41,6 +41,7 @@
 #include "vrf.h"
 #include "bfd.h"
 #include "libfrr.h"
+#include "ns.h"
 
 #include "bgpd/bgpd.h"
 #include "bgpd/bgp_attr.h"
@@ -57,6 +58,7 @@
 #include "bgpd/bgp_zebra.h"
 #include "bgpd/bgp_packet.h"
 #include "bgpd/bgp_keepalives.h"
+#include "bgpd/bgp_network.h"
 
 #ifdef ENABLE_BGP_VNC
 #include "bgpd/rfapi/rfapi_backend.h"
@@ -259,6 +261,7 @@ static int bgp_vrf_enable(struct vrf *vrf)
 		/* We have instance configured, link to VRF and make it "up". */
 		bgp_vrf_link(bgp, vrf);
 
+		bgp_handle_socket(bgp, vrf, old_vrf_id, true);
 		/* Update any redistribute vrf bitmaps if the vrf_id changed */
 		if (old_vrf_id != bgp->vrf_id)
 			bgp_update_redist_vrf_bitmaps(bgp, old_vrf_id);
@@ -282,6 +285,7 @@ static int bgp_vrf_disable(struct vrf *vrf)
 	bgp = bgp_lookup_by_name(vrf->name);
 	if (bgp) {
 		old_vrf_id = bgp->vrf_id;
+		bgp_handle_socket(bgp, vrf, VRF_UNKNOWN, false);
 		/* We have instance configured, unlink from VRF and make it
 		 * "down". */
 		bgp_vrf_unlink(bgp, vrf);

bgpd/bgp_mpath.c

@@ -92,6 +92,26 @@ int bgp_maximum_paths_unset(struct bgp *bgp, afi_t afi, safi_t safi,
 	return 0;
 }
 
+/*
+ * bgp_interface_same
+ *
+ * Return true if ifindex for ifp1 and ifp2 are the same, else return false.
+ */
+static int bgp_interface_same(struct interface *ifp1, struct interface *ifp2)
+{
+	if (!ifp1 && !ifp2)
+		return 1;
+
+	if (!ifp1 && ifp2)
+		return 0;
+
+	if (ifp1 && !ifp2)
+		return 0;
+
+	return (ifp1->ifindex == ifp2->ifindex);
+}
+
+
 /*
  * bgp_info_nexthop_cmp
  *
@@ -130,7 +150,8 @@ int bgp_info_nexthop_cmp(struct bgp_info *bi1, struct bgp_info *bi2)
 
 				if (!bi1->attr->mp_nexthop_prefer_global &&
 				    !bi2->attr->mp_nexthop_prefer_global)
-					compare = !(bi1->peer->ifindex == bi2->peer->ifindex);
+					compare = !bgp_interface_same(bi1->peer->ifp, bi2->peer->ifp);
+
 				if (!compare)
 					compare = IPV6_ADDR_CMP(&addr1, &addr2);
 				break;

bgpd/bgp_network.c

@@ -34,6 +34,7 @@
 #include "queue.h"
 #include "hash.h"
 #include "filter.h"
+#include "ns.h"
 
 #include "bgpd/bgpd.h"
 #include "bgpd/bgp_open.h"
@@ -44,13 +45,14 @@
 
 extern struct zebra_privs_t bgpd_privs;
 
-static int bgp_bind(struct peer *);
+static char *bgp_get_bound_name(struct peer *peer);
 
 /* BGP listening socket. */
 struct bgp_listener {
 	int fd;
 	union sockunion su;
 	struct thread *thread;
+	struct bgp *bgp;
 };
 
 /*
@@ -284,6 +286,7 @@ static int bgp_accept(struct thread *thread)
 		return -1;
 	}
 	listener->thread = NULL;
+
 	thread_add_read(bm->master, bgp_accept, listener, accept_sock,
 			&listener->thread);
 
@@ -296,8 +299,13 @@ static int bgp_accept(struct thread *thread)
 	}
 	set_nonblocking(bgp_sock);
 
-	/* Obtain BGP instance this connection is meant for. */
+	/* Obtain BGP instance this connection is meant for.
-	if (bgp_get_instance_for_inc_conn(bgp_sock, &bgp)) {
+	 * - if it is a VRF netns sock, then BGP is in listener structure
+	 * - otherwise, the bgp instance need to be demultiplexed
+	 */
+	if (listener->bgp)
+		bgp = listener->bgp;
+	else if (bgp_get_instance_for_inc_conn(bgp_sock, &bgp)) {
 		if (bgp_debug_neighbor_events(NULL))
 			zlog_debug(
 				"[Event] Could not get instance for incoming conn from %s",
@@ -407,7 +415,7 @@ static int bgp_accept(struct thread *thread)
 	peer->doppelganger = peer1;
 	peer1->doppelganger = peer;
 	peer->fd = bgp_sock;
-	bgp_bind(peer);
+	vrf_bind(peer->bgp->vrf_id, bgp_sock, bgp_get_bound_name(peer));
 	bgp_fsm_change_status(peer, Active);
 	BGP_TIMER_OFF(peer->t_start); /* created in peer_create() */
 
@@ -435,21 +443,20 @@ static int bgp_accept(struct thread *thread)
 }
 
 /* BGP socket bind. */
-static int bgp_bind(struct peer *peer)
+static char *bgp_get_bound_name(struct peer *peer)
 {
-#ifdef SO_BINDTODEVICE
-	int ret;
-	int myerrno;
 	char *name = NULL;
 
-	/* If not bound to an interface or part of a VRF, we don't care. */
+	if ((peer->bgp->vrf_id == VRF_DEFAULT) &&
-	if (!peer->bgp->vrf_id && !peer->ifname && !peer->conf_if)
+	    !peer->ifname && !peer->conf_if)
-		return 0;
+		return NULL;
 
 	if (peer->su.sa.sa_family != AF_INET
 	    && peer->su.sa.sa_family != AF_INET6)
-		return 0; // unexpected
+		return NULL; // unexpected
 
+	if (!peer)
+		return name;
 	/* For IPv6 peering, interface (unnumbered or link-local with interface)
 	 * takes precedence over VRF. For IPv4 peering, explicit interface or
 	 * VRF are the situations to bind.
@@ -461,30 +468,7 @@ static int bgp_bind(struct peer *peer)
 	else
 		name = peer->ifname ? peer->ifname : peer->bgp->name;
 
-	if (!name)
+	return name;
-		return 0;
-
-	if (bgp_debug_neighbor_events(peer))
-		zlog_debug("%s Binding to interface %s", peer->host, name);
-
-	if (bgpd_privs.change(ZPRIVS_RAISE))
-		zlog_err("bgp_bind: could not raise privs");
-
-	ret = setsockopt(peer->fd, SOL_SOCKET, SO_BINDTODEVICE, name,
-			 strlen(name));
-	myerrno = errno;
-
-	if (bgpd_privs.change(ZPRIVS_LOWER))
-		zlog_err("bgp_bind: could not lower privs");
-
-	if (ret < 0) {
-		if (bgp_debug_neighbor_events(peer))
-			zlog_debug("bind to interface %s failed, errno=%d",
-				   name, myerrno);
-		return ret;
-	}
-#endif /* SO_BINDTODEVICE */
-	return 0;
 }
 
 static int bgp_update_address(struct interface *ifp, const union sockunion *dst,
@@ -558,8 +542,13 @@ int bgp_connect(struct peer *peer)
 		zlog_debug("Peer address not learnt: Returning from connect");
 		return 0;
 	}
+	if (bgpd_privs.change(ZPRIVS_RAISE))
+		zlog_err("Can't raise privileges");
 	/* Make socket for the peer. */
-	peer->fd = sockunion_socket(&peer->su);
+	peer->fd = vrf_sockunion_socket(&peer->su, peer->bgp->vrf_id,
+					bgp_get_bound_name(peer));
+	if (bgpd_privs.change(ZPRIVS_LOWER))
+		zlog_err("Can't lower privileges");
 	if (peer->fd < 0)
 		return -1;
 
@@ -591,9 +580,6 @@ int bgp_connect(struct peer *peer)
 	if (peer->password)
 		bgp_md5_set_connect(peer->fd, &peer->su, peer->password);
 
-	/* Bind socket. */
-	bgp_bind(peer);
-
 	/* Update source bind. */
 	if (bgp_update_source(peer) < 0) {
 		return connect_error;
@@ -642,12 +628,12 @@ int bgp_getsockname(struct peer *peer)
 		return -1;
 #endif
 	}
-
 	return 0;
 }
 
 
-static int bgp_listener(int sock, struct sockaddr *sa, socklen_t salen)
+static int bgp_listener(int sock, struct sockaddr *sa, socklen_t salen,
+			struct bgp *bgp)
 {
 	struct bgp_listener *listener;
 	int ret, en;
@@ -683,8 +669,14 @@ static int bgp_listener(int sock, struct sockaddr *sa, socklen_t salen)
 		return ret;
 	}
 
-	listener = XMALLOC(MTYPE_BGP_LISTENER, sizeof(*listener));
+	listener = XCALLOC(MTYPE_BGP_LISTENER, sizeof(*listener));
 	listener->fd = sock;
+
+	/* this socket needs a change of ns. record bgp back pointer */
+	if (bgp->vrf_id != VRF_DEFAULT &&
+	    vrf_is_mapped_on_netns(bgp->vrf_id))
+		listener->bgp = bgp;
+
 	memcpy(&listener->su, sa, salen);
 	listener->thread = NULL;
 	thread_add_read(bm->master, bgp_accept, listener, sock,
@@ -695,7 +687,7 @@ static int bgp_listener(int sock, struct sockaddr *sa, socklen_t salen)
 }
 
 /* IPv6 supported version of BGP server socket setup.  */
-int bgp_socket(unsigned short port, const char *address)
+int bgp_socket(struct bgp *bgp, unsigned short port, const char *address)
 {
 	struct addrinfo *ainfo;
 	struct addrinfo *ainfo_save;
@@ -710,7 +702,12 @@ int bgp_socket(unsigned short port, const char *address)
 	snprintf(port_str, sizeof(port_str), "%d", port);
 	port_str[sizeof(port_str) - 1] = '\0';
 
-	ret = getaddrinfo(address, port_str, &req, &ainfo_save);
+	if (bgpd_privs.change(ZPRIVS_RAISE))
+		zlog_err("Can't raise privileges");
+	ret = vrf_getaddrinfo(address, port_str, &req,
+			      &ainfo_save, bgp->vrf_id);
+	if (bgpd_privs.change(ZPRIVS_LOWER))
+		zlog_err("Can't lower privileges");
 	if (ret != 0) {
 		zlog_err("getaddrinfo: %s", gai_strerror(ret));
 		return -1;
@@ -723,8 +720,13 @@ int bgp_socket(unsigned short port, const char *address)
 		if (ainfo->ai_family != AF_INET && ainfo->ai_family != AF_INET6)
 			continue;
 
-		sock = socket(ainfo->ai_family, ainfo->ai_socktype,
+		if (bgpd_privs.change(ZPRIVS_RAISE))
-			      ainfo->ai_protocol);
+			zlog_err("Can't raise privileges");
+		sock = vrf_socket(ainfo->ai_family, ainfo->ai_socktype,
+				  ainfo->ai_protocol, bgp->vrf_id,
+				  NULL);
+		if (bgpd_privs.change(ZPRIVS_LOWER))
+			zlog_err("Can't lower privileges");
 		if (sock < 0) {
 			zlog_err("socket: %s", safe_strerror(errno));
 			continue;
@@ -734,7 +736,8 @@ int bgp_socket(unsigned short port, const char *address)
 		 * ttl=255 */
 		sockopt_ttl(ainfo->ai_family, sock, MAXTTL);
 
-		ret = bgp_listener(sock, ainfo->ai_addr, ainfo->ai_addrlen);
+		ret = bgp_listener(sock, ainfo->ai_addr,
+				   ainfo->ai_addrlen, bgp);
 		if (ret == 0)
 			++count;
 		else
@@ -751,6 +754,32 @@ int bgp_socket(unsigned short port, const char *address)
 	return 0;
 }
 
+/* this function closes vrf socket
+ * this should be called only for vrf socket with netns backend
+ */
+void bgp_close_vrf_socket(struct bgp *bgp)
+{
+	struct listnode *node, *next;
+	struct bgp_listener *listener;
+
+	if (!bgp)
+		return;
+
+	if (bm->listen_sockets == NULL)
+		return;
+
+	for (ALL_LIST_ELEMENTS(bm->listen_sockets, node, next, listener)) {
+		if (listener->bgp == bgp) {
+			thread_cancel(listener->thread);
+			close(listener->fd);
+			listnode_delete(bm->listen_sockets, listener);
+			XFREE(MTYPE_BGP_LISTENER, listener);
+		}
+	}
+}
+
+/* this function closes main socket
+ */
 void bgp_close(void)
 {
 	struct listnode *node, *next;
@@ -760,6 +789,8 @@ void bgp_close(void)
 		return;
 
 	for (ALL_LIST_ELEMENTS(bm->listen_sockets, node, next, listener)) {
+		if (listener->bgp)
+			continue;
 		thread_cancel(listener->thread);
 		close(listener->fd);
 		listnode_delete(bm->listen_sockets, listener);

bgpd/bgp_network.h

@@ -23,7 +23,9 @@
 
 #define BGP_SOCKET_SNDBUF_SIZE 65536
 
-extern int bgp_socket(unsigned short, const char *);
+extern int bgp_socket(struct bgp *bgp, unsigned short port,
+		      const char *address);
+extern void bgp_close_vrf_socket(struct bgp *bgp);
 extern void bgp_close(void);
 extern int bgp_connect(struct peer *);
 extern int bgp_getsockname(struct peer *);

bgpd/bgp_routemap.c

@@ -3076,6 +3076,19 @@ static void bgp_route_map_process_update(struct bgp *bgp, const char *rmap_name,
 				}
 			}
 		}
+
+	/* for type5 command route-maps */
+	FOREACH_AFI_SAFI (afi, safi) {
+		if (bgp->adv_cmd_rmap[afi][safi].name &&
+		    strcmp(rmap_name, bgp->adv_cmd_rmap[afi][safi].name) == 0) {
+			if (BGP_DEBUG(zebra, ZEBRA))
+				zlog_debug(
+					   "Processing route_map %s update on advertise type5 route command",
+					   rmap_name);
+			bgp_evpn_withdraw_type5_routes(bgp, afi, safi);
+			bgp_evpn_advertise_type5_routes(bgp, afi, safi);
+		}
+	}
 }
 
 static int bgp_route_map_process_update_cb(char *rmap_name)

bgpd/bgp_vty.c

@@ -3395,7 +3395,7 @@ DEFUN (no_neighbor_set_peer_group,
 		return CMD_WARNING_CONFIG_FAILED;
 	}
 
-	ret = peer_group_unbind(bgp, peer, group);
+	ret = peer_delete(peer);
 
 	return bgp_vty_return(vty, ret);
 }
@@ -7278,8 +7278,7 @@ static void bgp_show_summary_afi_safi(struct vty *vty, struct bgp *bgp, int afi,
 				safi = SAFI_MAX;
 		}
 		afi++;
-		if (!afi_wildcard
+		if (!afi_wildcard)
-		    || afi == AFI_L2VPN) /* special case, not handled yet */
 			afi = AFI_MAX;
 	}
 

bgpd/bgp_zebra.c

@@ -1026,14 +1026,14 @@ void bgp_zebra_announce(struct bgp_node *rn, struct prefix *p,
 	if (peer->sort == BGP_PEER_IBGP || peer->sort == BGP_PEER_CONFED
 	    || info->sub_type == BGP_ROUTE_AGGREGATE) {
 		SET_FLAG(api.flags, ZEBRA_FLAG_IBGP);
-		SET_FLAG(api.flags, ZEBRA_FLAG_INTERNAL);
+		SET_FLAG(api.flags, ZEBRA_FLAG_ALLOW_RECURSION);
 	}
 
 	if ((peer->sort == BGP_PEER_EBGP && peer->ttl != 1)
 	    || CHECK_FLAG(peer->flags, PEER_FLAG_DISABLE_CONNECTED_CHECK)
 	    || bgp_flag_check(bgp, BGP_FLAG_DISABLE_NH_CONNECTED_CHK))
 
-		SET_FLAG(api.flags, ZEBRA_FLAG_INTERNAL);
+		SET_FLAG(api.flags, ZEBRA_FLAG_ALLOW_RECURSION);
 
 	/* Metric is currently based on the best-path only */
 	metric = info->attr->med;
@@ -1265,14 +1265,14 @@ void bgp_zebra_withdraw(struct prefix *p, struct bgp_info *info, safi_t safi)
 		SET_FLAG(api.flags, ZEBRA_FLAG_EVPN_ROUTE);
 
 	if (peer->sort == BGP_PEER_IBGP) {
-		SET_FLAG(api.flags, ZEBRA_FLAG_INTERNAL);
+		SET_FLAG(api.flags, ZEBRA_FLAG_ALLOW_RECURSION);
 		SET_FLAG(api.flags, ZEBRA_FLAG_IBGP);
 	}
 
 	if ((peer->sort == BGP_PEER_EBGP && peer->ttl != 1)
 	    || CHECK_FLAG(peer->flags, PEER_FLAG_DISABLE_CONNECTED_CHECK)
 	    || bgp_flag_check(peer->bgp, BGP_FLAG_DISABLE_NH_CONNECTED_CHK))
-		SET_FLAG(api.flags, ZEBRA_FLAG_INTERNAL);
+		SET_FLAG(api.flags, ZEBRA_FLAG_ALLOW_RECURSION);
 
 	if (bgp_debug_zebra(p)) {
 		char buf[PREFIX_STRLEN];
@@ -1726,6 +1726,7 @@ static void bgp_zebra_connected(struct zclient *zclient)
 static int bgp_zebra_process_local_l3vni(int cmd, struct zclient *zclient,
 					 zebra_size_t length, vrf_id_t vrf_id)
 {
+	int filter = 0;
 	char buf[ETHER_ADDR_STRLEN];
 	vni_t l3vni = 0;
 	struct ethaddr rmac;
@@ -1739,16 +1740,20 @@ static int bgp_zebra_process_local_l3vni(int cmd, struct zclient *zclient,
 	if (cmd == ZEBRA_L3VNI_ADD) {
 		stream_get(&rmac, s, sizeof(struct ethaddr));
 		originator_ip.s_addr = stream_get_ipv4(s);
+		stream_get(&filter, s, sizeof(int));
 	}
 
 	if (BGP_DEBUG(zebra, ZEBRA))
-		zlog_debug("Rx L3-VNI %s VRF %s VNI %u RMAC %s",
+		zlog_debug("Rx L3-VNI %s VRF %s VNI %u RMAC %s filter %s",
 			   (cmd == ZEBRA_L3VNI_ADD) ? "add" : "del",
-			   vrf_id_to_name(vrf_id), l3vni,
+			   vrf_id_to_name(vrf_id),
-			   prefix_mac2str(&rmac, buf, sizeof(buf)));
+			   l3vni,
+			   prefix_mac2str(&rmac, buf, sizeof(buf)),
+			   filter ? "prefix-routes-only" : "none");
 
 	if (cmd == ZEBRA_L3VNI_ADD)
-		bgp_evpn_local_l3vni_add(l3vni, vrf_id, &rmac, originator_ip);
+		bgp_evpn_local_l3vni_add(l3vni, vrf_id, &rmac, originator_ip,
+					 filter);
 	else
 		bgp_evpn_local_l3vni_del(l3vni, vrf_id);
 

bgpd/bgpd.c

@@ -101,6 +101,42 @@ static void bgp_if_finish(struct bgp *bgp);
 
 extern struct zclient *zclient;
 
+/* handle main socket creation or deletion */
+static int bgp_check_main_socket(bool create, struct bgp *bgp)
+{
+	static int bgp_server_main_created;
+	struct listnode *bgpnode, *nbgpnode;
+	struct bgp *bgp_temp;
+
+	if (bgp->inst_type == BGP_INSTANCE_TYPE_VRF)
+		return 0;
+	if (create == true) {
+		if (bgp_server_main_created)
+			return 0;
+		if (bgp_socket(bgp, bm->port, bm->address) < 0)
+			return BGP_ERR_INVALID_VALUE;
+		bgp_server_main_created = 1;
+		return 0;
+	}
+	if (!bgp_server_main_created)
+		return 0;
+	/* only delete socket on some cases */
+	for (ALL_LIST_ELEMENTS(bm->bgp, bgpnode, nbgpnode, bgp_temp)) {
+		/* do not count with current bgp */
+		if (bgp_temp == bgp)
+			continue;
+		/* if other instance non VRF, do not delete socket */
+		if (bgp_temp->inst_type == BGP_INSTANCE_TYPE_DEFAULT)
+			return 0;
+		/* vrf lite, do not delete socket */
+		if (!vrf_is_mapped_on_netns(bgp_temp->vrf_id))
+			return 0;
+	}
+	bgp_close();
+	bgp_server_main_created = 0;
+	return 0;
+}
+
 void bgp_session_reset(struct peer *peer)
 {
 	if (peer->doppelganger && (peer->doppelganger->status != Deleted)
@@ -1200,7 +1236,6 @@ void peer_xfer_config(struct peer *peer_dst, struct peer *peer_src)
 	peer_dst->config = peer_src->config;
 
 	peer_dst->local_as = peer_src->local_as;
-	peer_dst->ifindex = peer_src->ifindex;
 	peer_dst->port = peer_src->port;
 	(void)peer_sort(peer_dst);
 	peer_dst->rmap_type = peer_src->rmap_type;
@@ -2981,11 +3016,67 @@ struct bgp *bgp_lookup_by_vrf_id(vrf_id_t vrf_id)
 	return (vrf->info) ? (struct bgp *)vrf->info : NULL;
 }
 
+/* handle socket creation or deletion, if necessary
+ * this is called for all new BGP instances
+ */
+int bgp_handle_socket(struct bgp *bgp, struct vrf *vrf,
+			  vrf_id_t old_vrf_id, bool create)
+{
+	int ret = 0;
+
+	/* Create BGP server socket, if listen mode not disabled */
+	if (!bgp || bgp_option_check(BGP_OPT_NO_LISTEN))
+		return 0;
+	if (bgp->name
+	    && bgp->inst_type == BGP_INSTANCE_TYPE_VRF
+	    && vrf) {
+		/*
+		 * suppress vrf socket
+		 */
+		if (create == FALSE) {
+			if (vrf_is_mapped_on_netns(vrf->vrf_id))
+				bgp_close_vrf_socket(bgp);
+			else
+				ret = bgp_check_main_socket(create, bgp);
+			return ret;
+		}
+		/* do nothing
+		 * if vrf_id did not change
+		 */
+		if (vrf->vrf_id == old_vrf_id)
+			return 0;
+		if (old_vrf_id != VRF_UNKNOWN) {
+			/* look for old socket. close it. */
+			bgp_close_vrf_socket(bgp);
+		}
+		/* if backend is not yet identified ( VRF_UNKNOWN) then
+		 *   creation will be done later
+		 */
+		if (vrf->vrf_id == VRF_UNKNOWN)
+			return 0;
+		/* if BGP VRF instance requested
+		 * if backend is NETNS, create BGP server socket in the NETNS
+		 */
+		if (vrf_is_mapped_on_netns(bgp->vrf_id)) {
+			ret = bgp_socket(bgp, bm->port, bm->address);
+			if (ret < 0)
+				return BGP_ERR_INVALID_VALUE;
+			return 0;
+		}
+	}
+	/* if BGP VRF instance requested or VRF lite backend
+	 * if BGP non VRF instance, create it
+	 *  if not already done
+	 */
+	return bgp_check_main_socket(create, bgp);
+}
+
 /* Called from VTY commands. */
 int bgp_get(struct bgp **bgp_val, as_t *as, const char *name,
 	    enum bgp_instance_type inst_type)
 {
 	struct bgp *bgp;
+	struct vrf *vrf = NULL;
 
 	/* Multiple instance check. */
 	if (bgp_option_check(BGP_OPT_MULTIPLE_INSTANCE)) {
@@ -3033,25 +3124,19 @@ int bgp_get(struct bgp **bgp_val, as_t *as, const char *name,
 
 	bgp->t_rmap_def_originate_eval = NULL;
 
-	/* Create BGP server socket, if first instance.  */
-	if (list_isempty(bm->bgp) && !bgp_option_check(BGP_OPT_NO_LISTEN)) {
-		if (bgp_socket(bm->port, bm->address) < 0)
-			return BGP_ERR_INVALID_VALUE;
-	}
-
-	listnode_add(bm->bgp, bgp);
-
 	/* If Default instance or VRF, link to the VRF structure, if present. */
 	if (bgp->inst_type == BGP_INSTANCE_TYPE_DEFAULT
 	    || bgp->inst_type == BGP_INSTANCE_TYPE_VRF) {
-		struct vrf *vrf;
-
 		vrf = bgp_vrf_lookup_by_instance_type(bgp);
 		if (vrf)
 			bgp_vrf_link(bgp, vrf);
 	}
+	/* BGP server socket already processed if BGP instance
+	 * already part of the list
+	 */
+	bgp_handle_socket(bgp, vrf, VRF_UNKNOWN, true);
+	listnode_add(bm->bgp, bgp);
 
-	/* Register with Zebra, if needed */
 	if (IS_BGP_INST_KNOWN_TO_ZEBRA(bgp))
 		bgp_zebra_instance_register(bgp);
 
@@ -3188,8 +3273,6 @@ int bgp_delete(struct bgp *bgp)
 	 * routes to be processed still referencing the struct bgp.
 	 */
 	listnode_delete(bm->bgp, bgp);
-	if (list_isempty(bm->bgp))
-		bgp_close();
 
 	/* Deregister from Zebra, if needed */
 	if (IS_BGP_INST_KNOWN_TO_ZEBRA(bgp))
@@ -3199,6 +3282,7 @@ int bgp_delete(struct bgp *bgp)
 	bgp_if_finish(bgp);
 
 	vrf = bgp_vrf_lookup_by_instance_type(bgp);
+	bgp_handle_socket(bgp, vrf, VRF_UNKNOWN, false);
 	if (vrf)
 		bgp_vrf_unlink(bgp, vrf);
 
@@ -3337,11 +3421,12 @@ struct peer *peer_lookup(struct bgp *bgp, union sockunion *su)
 		struct listnode *bgpnode, *nbgpnode;
 
 		for (ALL_LIST_ELEMENTS(bm->bgp, bgpnode, nbgpnode, bgp)) {
-			/* Skip VRFs, this function will not be invoked without
+			/* Skip VRFs Lite only, this function will not be
-			 * an instance
+			 * invoked without an instance
 			 * when examining VRFs.
 			 */
-			if (bgp->inst_type == BGP_INSTANCE_TYPE_VRF)
+			if ((bgp->inst_type == BGP_INSTANCE_TYPE_VRF) &&
+			    !vrf_is_mapped_on_netns(bgp->vrf_id))
 				continue;
 
 			peer = hash_lookup(bgp->peerhash, &tmp_peer);
@@ -7315,38 +7400,6 @@ int bgp_config_write(struct vty *vty)
 		if (bgp_option_check(BGP_OPT_CONFIG_CISCO))
 			vty_out(vty, " no auto-summary\n");
 
-		/* import route-target */
-		if (CHECK_FLAG(bgp->vrf_flags, BGP_VRF_IMPORT_RT_CFGD)) {
-			char *ecom_str;
-			struct listnode *node, *nnode;
-			struct ecommunity *ecom;
-
-			for (ALL_LIST_ELEMENTS(bgp->vrf_import_rtl, node, nnode,
-					       ecom)) {
-				ecom_str = ecommunity_ecom2str(
-					ecom, ECOMMUNITY_FORMAT_ROUTE_MAP, 0);
-				vty_out(vty, "   route-target import %s\n",
-					ecom_str);
-				XFREE(MTYPE_ECOMMUNITY_STR, ecom_str);
-			}
-		}
-
-		/* export route-target */
-		if (CHECK_FLAG(bgp->vrf_flags, BGP_VRF_EXPORT_RT_CFGD)) {
-			char *ecom_str;
-			struct listnode *node, *nnode;
-			struct ecommunity *ecom;
-
-			for (ALL_LIST_ELEMENTS(bgp->vrf_export_rtl, node, nnode,
-					       ecom)) {
-				ecom_str = ecommunity_ecom2str(
-					ecom, ECOMMUNITY_FORMAT_ROUTE_MAP, 0);
-				vty_out(vty, "   route-target export %s\n",
-					ecom_str);
-				XFREE(MTYPE_ECOMMUNITY_STR, ecom_str);
-			}
-		}
-
 		/* IPv4 unicast configuration.  */
 		bgp_config_write_family(vty, bgp, AFI_IP, SAFI_UNICAST);
 

bgpd/bgpd.h

@@ -436,6 +436,7 @@ struct bgp {
 #define BGP_VRF_IMPORT_RT_CFGD              (1 << 3)
 #define BGP_VRF_EXPORT_RT_CFGD              (1 << 4)
 #define BGP_VRF_RD_CFGD                     (1 << 5)
+#define BGP_VRF_L3VNI_PREFIX_ROUTES_ONLY    (1 << 6)
 
 	/* unique ID for auto derivation of RD for this vrf */
 	uint16_t vrf_rd_id;
@@ -452,6 +453,9 @@ struct bgp {
 	/* list of corresponding l2vnis (struct bgpevpn) */
 	struct list *l2vnis;
 
+	/* route map for advertise ipv4/ipv6 unicast (type-5 routes) */
+	struct bgp_rmap adv_cmd_rmap[AFI_MAX][SAFI_MAX];
+
 	QOBJ_FIELDS
 };
 DECLARE_QOBJ_TYPE(bgp)
@@ -683,7 +687,6 @@ struct peer {
 	time_t readtime;     /* Last read time */
 	time_t resettime;    /* Last reset time */
 
-	ifindex_t ifindex;     /* ifindex of the BGP connection. */
 	char *conf_if;	 /* neighbor interface config name. */
 	struct interface *ifp; /* corresponding interface */
 	char *ifname;	  /* bind interface name. */
@@ -1351,6 +1354,9 @@ extern void bgp_instance_up(struct bgp *);
 extern void bgp_instance_down(struct bgp *);
 extern int bgp_delete(struct bgp *);
 
+extern int bgp_handle_socket(struct bgp *bgp, struct vrf *vrf,
+			     vrf_id_t old_vrf_id, bool create);
+
 extern int bgp_flag_set(struct bgp *, int);
 extern int bgp_flag_unset(struct bgp *, int);
 extern int bgp_flag_check(struct bgp *, int);

bgpd/rfapi/bgp_rfapi_cfg.c

@@ -2977,6 +2977,11 @@ DEFUN_NOSH (vnc_vrf_policy,
 	struct rfapi_nve_group_cfg *rfg;
 	VTY_DECLVAR_CONTEXT(bgp, bgp);
 
+	if (bgp->inst_type == BGP_INSTANCE_TYPE_VRF) {
+		vty_out(vty, "Can't configure vrf-policy within a BGP VRF instance\n");
+		return CMD_WARNING_CONFIG_FAILED;
+	}
+
 	/* Search for name */
 	rfg = bgp_rfapi_cfg_match_byname(bgp, argv[1]->arg,
 					 RFAPI_GROUP_CFG_VRF);
@@ -3007,6 +3012,10 @@ DEFUN (vnc_no_vrf_policy,
 {
 	VTY_DECLVAR_CONTEXT(bgp, bgp);
 
+	/* silently return */
+	if (bgp->inst_type == BGP_INSTANCE_TYPE_VRF)
+		return CMD_SUCCESS;
+
 	return bgp_rfapi_delete_named_nve_group(vty, bgp, argv[2]->arg,
 						RFAPI_GROUP_CFG_VRF);
 }

debianpkg/backports/ubuntu14.04/debian/frr.install

@@ -1,5 +1,6 @@
 etc/frr/
 usr/bin/vtysh
+usr/bin/mtracebis
 usr/include/frr/
 usr/lib/
 tools/frr etc/init.d/

debianpkg/frr.install

@@ -1,5 +1,6 @@
 etc/frr/
 usr/bin/vtysh
+usr/bin/mtracebis
 usr/include/frr/
 usr/lib/
 tools/frr usr/lib/frr

doc/Makefile.am

@@ -11,6 +11,7 @@ man_MANS = $(MANPAGE_BUILDDIR)/frr.1
 
 if PIMD
 man_MANS += $(MANPAGE_BUILDDIR)/pimd.8
+man_MANS += $(MANPAGE_BUILDDIR)/mtracebis.8
 endif
 
 if BGPD
@@ -108,6 +109,7 @@ EXTRA_DIST = manpages/defines.rst \
 	manpages/vtysh.rst \
 	manpages/nhrpd.rst \
 	manpages/pimd.rst \
+	manpages/mtracebis.rst \
 	manpages/ripngd.rst \
 	manpages/frr.rst \
 	manpages/conf.py \

doc/developer/Building_FRR_on_CentOS7.rst

@@ -22,7 +22,7 @@ Add packages:
     sudo yum install git autoconf automake libtool make gawk \
       readline-devel texinfo net-snmp-devel groff pkgconfig \
       json-c-devel pam-devel bison flex pytest c-ares-devel \
-      perl-XML-LibXML python-devel python-sphinx
+      perl-XML-LibXML python-devel systemd-devel python-sphinx
 
 Get FRR, compile it and install it (from Git)
 ---------------------------------------------
@@ -68,6 +68,7 @@ an example.)
         --enable-group=frr \
         --enable-vty-group=frrvt \
         --enable-rtadv \
+	--enable-systemd=yes \
         --disable-exampledir \
         --enable-watchfrr \
         --disable-ldpd \

doc/manpages/conf.py

@@ -318,6 +318,7 @@ man_pages = [
     ('ldpd', 'ldpd', fwfrr.format("an LDP "), [], 8),
     ('nhrpd', 'nhrpd', fwfrr.format("a Next Hop Routing Protocol "), [], 8),
     ('pimd', 'pimd', fwfrr.format("a PIM "), [], 8),
+    ('mtracebis', 'mtracebis', "a multicast trace client", [], 8),
     ('ripd', 'ripd', fwfrr.format("a RIP "), [], 8),
     ('ripngd', 'ripngd', fwfrr.format("a RIPNG "), [], 8),
     ('zebra', 'zebra', 'a routing manager for use with associated FRRouting components.', [], 8),

doc/manpages/defines.rst

@@ -1,3 +1,3 @@
 .. |synopsis-options| replace:: [-d|-t|-dt] [-C] [-f config-file] [-i pid-file] [-z zclient-path] [-u user] [-g group] [-A vty-addr] [-P vty-port] [-M module[:options]] [-N pathspace] [--vty_socket vty-path] [--moduledir module-path]
 .. |synopsis-options-hv| replace:: [-h] [-v]
-.. |seealso-programs| replace:: zebra(8), vtysh(1), ripd(8), ripngd(8), ospfd(8), ospf6d(8), bgpd(8), isisd(8), babeld(8), nhrpd(8), pimd(8), ldpd(8), eigrpd(8)
+.. |seealso-programs| replace:: zebra(8), vtysh(1), ripd(8), ripngd(8), ospfd(8), ospf6d(8), bgpd(8), isisd(8), babeld(8), nhrpd(8), pimd(8), ldpd(8), eigrpd(8), mtracebis(8)

doc/manpages/epilogue.rst

@@ -8,7 +8,7 @@ The daemon may log to standard output, to a VTY, to a log file, or through syslo
 
 SEE ALSO
 ========
-zebra(8), vtysh(1), ripd(8), ripngd(8), ospfd(8), ospf6d(8), bgpd(8), isisd(8), babeld(8), nhrpd(8), pimd(8), ldpd(8), eigrpd(8)
+|seealso-programs|
 |PACKAGE_URL|
 
 BUGS

doc/manpages/frr.rst

@@ -2,6 +2,8 @@
 FRR
 ***
 
+.. include:: defines.rst
+
 SYNOPSIS
 ========
 frr [ start ]

doc/manpages/index.rst

@@ -18,6 +18,7 @@ Welcome to FRR's documentation!
    ospfclient
    ospfd
    pimd
+   mtracebis
    ripd
    ripngd
    watchfrr

doc/manpages/mtracebis.rst

@@ -0,0 +1,39 @@
+*********
+MTRACEBIS
+*********
+
+.. include:: defines.rst
+.. |PROGRAM| replace:: mtracebis
+
+SYNOPSIS
+========
+|PROGRAM| |synopsis-options-hv|
+
+|PROGRAM| <multicast source>
+
+DESCRIPTION
+===========
+|PROGRAM| is a program to initiate multicast traceroute, or "mtrace", queries.
+
+The initial version of the program requires multicast source IP address and
+initiates a weak traceroute across the network. This tests whether the
+interfaces towards the source are multicast enabled. The first query sent is a
+full query, capable of crossing the network all the way to the source. If this
+fails, hop-by-hop queries are initiated.
+
+Hop-by-hop queries start by requesting only a response from the nearest router.
+Following that, next query is extended to the next two routers, and so on...
+until a set of routers is tested for connectivity.
+
+FILES
+=====
+
+|INSTALL_PREFIX_SBIN|/|PROGRAM|
+   The default location of the |PROGRAM| binary.
+
+.. include:: epilogue.rst
+
+AUTHORS
+=======
+
+Mladen Sablic

doc/manpages/ospfclient.rst

@@ -2,6 +2,8 @@
 OSPFCLIENT
 **********
 
+.. include:: defines.rst
+
 SYNOPSIS
 ========
 ospfclient <ospfd> <lsatype> <opaquetype> <opaqueid> <ifaddr> <areaid>

doc/manpages/zebra.rst

@@ -37,6 +37,11 @@ OPTIONS available for the |DAEMON| command:
 
    Note that this affects Linux only.
 
+
+.. option:: -n, --vrfwnetns
+
+   Enable namespace VRF backend. By default, the VRF backend relies on VRF-lite support from the Linux kernel. This option permits discovering Linux named network namespaces and mapping it to FRR VRF contexts.
+
 FILES
 =====
 

eigrpd/eigrp_zebra.c

@@ -94,13 +94,14 @@ static int eigrp_router_id_update_zebra(int command, struct zclient *zclient,
 	return 0;
 }
 
-static int eigrp_zebra_notify_owner(int command, struct zclient *zclient,
+static int eigrp_zebra_route_notify_owner(int command, struct zclient *zclient,
 					  zebra_size_t length, vrf_id_t vrf_id)
 {
 	struct prefix p;
 	enum zapi_route_notify_owner note;
+	uint32_t table;
 
-	if (!zapi_route_notify_decode(zclient->ibuf, &p, &note))
+	if (!zapi_route_notify_decode(zclient->ibuf, &p, &table, &note))
 		return -1;
 
 	return 0;
@@ -128,7 +129,7 @@ void eigrp_zebra_init(void)
 	zclient->interface_address_delete = eigrp_interface_address_delete;
 	zclient->redistribute_route_add = eigrp_zebra_read_route;
 	zclient->redistribute_route_del = eigrp_zebra_read_route;
-	zclient->notify_owner = eigrp_zebra_notify_owner;
+	zclient->route_notify_owner = eigrp_zebra_route_notify_owner;
 }
 
 

include/linux/net_namespace.h

@@ -0,0 +1,23 @@
+/* Copyright (c) 2015 6WIND S.A.
+ * Author: Nicolas Dichtel <nicolas.dichtel@6wind.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ */
+#ifndef _LINUX_NET_NAMESPACE_H_
+#define _LINUX_NET_NAMESPACE_H_
+
+/* Attributes of RTM_NEWNSID/RTM_GETNSID messages */
+enum {
+	NETNSA_NONE,
+#define NETNSA_NSID_NOT_ASSIGNED -1
+	NETNSA_NSID,
+	NETNSA_PID,
+	NETNSA_FD,
+	__NETNSA_MAX,
+};
+
+#define NETNSA_MAX		(__NETNSA_MAX - 1)
+
+#endif /* _LINUX_NET_NAMESPACE_H_ */

include/subdir.am

@@ -6,4 +6,5 @@ noinst_HEADERS += \
 	include/linux/neighbour.h \
 	include/linux/rtnetlink.h \
 	include/linux/socket.h \
+	include/linux/net_namespace.h \
 	# end

isisd/isis_circuit.c

@@ -1168,7 +1168,6 @@ void isis_circuit_af_set(struct isis_circuit *circuit, bool ip_router,
 	struct isis_area *area = circuit->area;
 	bool change = circuit->ip_router != ip_router
 		      || circuit->ipv6_router != ipv6_router;
-	bool was_enabled = !!circuit->area;
 
 	area->ip_circuits += ip_router - circuit->ip_router;
 	area->ipv6_circuits += ipv6_router - circuit->ipv6_router;
@@ -1182,8 +1181,6 @@ void isis_circuit_af_set(struct isis_circuit *circuit, bool ip_router,
 
 	if (!ip_router && !ipv6_router)
 		isis_csm_state_change(ISIS_DISABLE, circuit, area);
-	else if (!was_enabled)
-		isis_csm_state_change(ISIS_ENABLE, circuit, area);
 	else
 		lsp_regenerate_schedule(circuit->area, circuit->is_type, 0);
 }

ldpd/interface.c

@@ -306,8 +306,11 @@ if_reset(struct iface *iface, int af)
 	ia = iface_af_get(iface, af);
 	if_stop_hello_timer(ia);
 
-	while ((adj = RB_ROOT(ia_adj_head, &ia->adj_tree)) != NULL)
+	while (!RB_EMPTY(ia_adj_head, &ia->adj_tree)) {
+		adj = RB_ROOT(ia_adj_head, &ia->adj_tree);
+
 		adj_del(adj, S_SHUTDOWN);
+	}
 
 	/* try to cleanup */
 	switch (af) {

ldpd/l2vpn.c

@@ -76,16 +76,21 @@ l2vpn_del(struct l2vpn *l2vpn)
 	struct l2vpn_if		*lif;
 	struct l2vpn_pw		*pw;
 
-	while ((lif = RB_ROOT(l2vpn_if_head, &l2vpn->if_tree)) != NULL) {
+	while (!RB_EMPTY(l2vpn_if_head, &l2vpn->if_tree)) {
+		lif = RB_ROOT(l2vpn_if_head, &l2vpn->if_tree);
+
 		RB_REMOVE(l2vpn_if_head, &l2vpn->if_tree, lif);
 		free(lif);
 	}
-	while ((pw = RB_ROOT(l2vpn_pw_head, &l2vpn->pw_tree)) != NULL) {
+	while (!RB_EMPTY(l2vpn_pw_head, &l2vpn->pw_tree)) {
+		pw = RB_ROOT(l2vpn_pw_head, &l2vpn->pw_tree);
+
 		RB_REMOVE(l2vpn_pw_head, &l2vpn->pw_tree, pw);
 		free(pw);
 	}
-	while ((pw = RB_ROOT(l2vpn_pw_head,
+	while (!RB_EMPTY(l2vpn_pw_head, &l2vpn->pw_inactive_tree)) {
-	    &l2vpn->pw_inactive_tree)) != NULL) {
+		pw = RB_ROOT(l2vpn_pw_head, &l2vpn->pw_inactive_tree);
+
 		RB_REMOVE(l2vpn_pw_head, &l2vpn->pw_inactive_tree, pw);
 		free(pw);
 	}

ldpd/lde.c

@@ -1324,8 +1324,11 @@ lde_nbr_clear(void)
 {
 	struct lde_nbr	*ln;
 
-	 while ((ln = RB_ROOT(nbr_tree, &lde_nbrs)) != NULL)
+	while (!RB_EMPTY(nbr_tree, &lde_nbrs)) {
+		ln = RB_ROOT(nbr_tree, &lde_nbrs);
+
 		lde_nbr_del(ln);
+	}
 }
 
 static void

ldpd/lde_lib.c

@@ -129,7 +129,9 @@ fec_clear(struct fec_tree *fh, void (*free_cb)(void *))
 {
 	struct fec	*f;
 
-	while ((f = RB_ROOT(fec_tree, fh)) != NULL) {
+	while (!RB_EMPTY(fec_tree, fh)) {
+		f = RB_ROOT(fec_tree, fh);
+
 		fec_remove(fh, f);
 		free_cb(f);
 	}

ldpd/ldp_vty_conf.c

@@ -1475,18 +1475,23 @@ l2vpn_del_api(struct ldpd_conf *conf, struct l2vpn *l2vpn)
 	struct l2vpn_if		*lif;
 	struct l2vpn_pw		*pw;
 
-	while ((lif = RB_ROOT(l2vpn_if_head, &l2vpn->if_tree)) != NULL) {
+	while (!RB_EMPTY(l2vpn_if_head, &l2vpn->if_tree)) {
+		lif = RB_ROOT(l2vpn_if_head, &l2vpn->if_tree);
+
 		QOBJ_UNREG(lif);
 		RB_REMOVE(l2vpn_if_head, &l2vpn->if_tree, lif);
 		free(lif);
 	}
-	while ((pw = RB_ROOT(l2vpn_pw_head, &l2vpn->pw_tree)) != NULL) {
+	while (!RB_EMPTY(l2vpn_pw_head, &l2vpn->pw_tree)) {
+		pw = RB_ROOT(l2vpn_pw_head, &l2vpn->pw_tree);
+
 		QOBJ_UNREG(pw);
 		RB_REMOVE(l2vpn_pw_head, &l2vpn->pw_tree, pw);
 		free(pw);
 	}
-	while ((pw = RB_ROOT(l2vpn_pw_head,
+	while (!RB_EMPTY(l2vpn_pw_head, &l2vpn->pw_inactive_tree)) {
-	    &l2vpn->pw_inactive_tree)) != NULL) {
+		pw = RB_ROOT(l2vpn_pw_head, &l2vpn->pw_inactive_tree);
+
 		QOBJ_UNREG(pw);
 		RB_REMOVE(l2vpn_pw_head, &l2vpn->pw_inactive_tree, pw);
 		free(pw);

ldpd/ldpd.c

@@ -1066,13 +1066,17 @@ ldp_config_reset_main(struct ldpd_conf *conf)
 	struct iface		*iface;
 	struct nbr_params	*nbrp;
 
-	while ((iface = RB_ROOT(iface_head, &conf->iface_tree)) != NULL) {
+	while (!RB_EMPTY(iface_head, &conf->iface_tree)) {
+		iface = RB_ROOT(iface_head, &conf->iface_tree);
+
 		QOBJ_UNREG(iface);
 		RB_REMOVE(iface_head, &conf->iface_tree, iface);
 		free(iface);
 	}
 
-	while ((nbrp = RB_ROOT(nbrp_head, &conf->nbrp_tree)) != NULL) {
+	while (!RB_EMPTY(nbrp_head, &conf->nbrp_tree)) {
+		nbrp = RB_ROOT(nbrp_head, &conf->nbrp_tree);
+
 		QOBJ_UNREG(nbrp);
 		RB_REMOVE(nbrp_head, &conf->nbrp_tree, nbrp);
 		free(nbrp);
@@ -1128,20 +1132,25 @@ ldp_config_reset_l2vpns(struct ldpd_conf *conf)
 	struct l2vpn_if		*lif;
 	struct l2vpn_pw		*pw;
 
-	while ((l2vpn = RB_ROOT(l2vpn_head, &conf->l2vpn_tree)) != NULL) {
+	while (!RB_EMPTY(l2vpn_head, &conf->l2vpn_tree)) {
-		while ((lif = RB_ROOT(l2vpn_if_head,
+		l2vpn = RB_ROOT(l2vpn_head, &conf->l2vpn_tree);
-		    &l2vpn->if_tree)) != NULL) {
+		while (!RB_EMPTY(l2vpn_if_head, &l2vpn->if_tree)) {
+			lif = RB_ROOT(l2vpn_if_head, &l2vpn->if_tree);
+
 			QOBJ_UNREG(lif);
 			RB_REMOVE(l2vpn_if_head, &l2vpn->if_tree, lif);
 			free(lif);
 		}
-		while ((pw = RB_ROOT(l2vpn_pw_head, &l2vpn->pw_tree)) != NULL) {
+		while (!RB_EMPTY(l2vpn_pw_head, &l2vpn->pw_tree)) {
+			pw = RB_ROOT(l2vpn_pw_head, &l2vpn->pw_tree);
+
 			QOBJ_UNREG(pw);
 			RB_REMOVE(l2vpn_pw_head, &l2vpn->pw_tree, pw);
 			free(pw);
 		}
-		while ((pw = RB_ROOT(l2vpn_pw_head,
+		while (!RB_EMPTY(l2vpn_pw_head, &l2vpn->pw_inactive_tree)) {
-		    &l2vpn->pw_inactive_tree)) != NULL) {
+			pw = RB_ROOT(l2vpn_pw_head, &l2vpn->pw_inactive_tree);
+
 			QOBJ_UNREG(pw);
 			RB_REMOVE(l2vpn_pw_head, &l2vpn->pw_inactive_tree, pw);
 			free(pw);
@@ -1160,19 +1169,27 @@ ldp_clear_config(struct ldpd_conf *xconf)
 	struct nbr_params	*nbrp;
 	struct l2vpn		*l2vpn;
 
-	while ((iface = RB_ROOT(iface_head, &xconf->iface_tree)) != NULL) {
+	while (!RB_EMPTY(iface_head, &xconf->iface_tree)) {
+		iface = RB_ROOT(iface_head, &xconf->iface_tree);
+
 		RB_REMOVE(iface_head, &xconf->iface_tree, iface);
 		free(iface);
 	}
-	while ((tnbr = RB_ROOT(tnbr_head, &xconf->tnbr_tree)) != NULL) {
+	while (!RB_EMPTY(tnbr_head, &xconf->tnbr_tree)) {
+		tnbr = RB_ROOT(tnbr_head, &xconf->tnbr_tree);
+
 		RB_REMOVE(tnbr_head, &xconf->tnbr_tree, tnbr);
 		free(tnbr);
 	}
-	while ((nbrp = RB_ROOT(nbrp_head, &xconf->nbrp_tree)) != NULL) {
+	while (!RB_EMPTY(nbrp_head, &xconf->nbrp_tree)) {
+		nbrp = RB_ROOT(nbrp_head, &xconf->nbrp_tree);
+
 		RB_REMOVE(nbrp_head, &xconf->nbrp_tree, nbrp);
 		free(nbrp);
 	}
-	while ((l2vpn = RB_ROOT(l2vpn_head, &xconf->l2vpn_tree)) != NULL) {
+	while (!RB_EMPTY(l2vpn_head, &xconf->l2vpn_tree)) {
+		l2vpn = RB_ROOT(l2vpn_head, &xconf->l2vpn_tree);
+
 		RB_REMOVE(l2vpn_head, &xconf->l2vpn_tree, l2vpn);
 		l2vpn_del(l2vpn);
 	}

ldpd/ldpe.c

@@ -219,8 +219,11 @@ ldpe_shutdown(void)
 		assert(if_addr != LIST_FIRST(&global.addr_list));
 		free(if_addr);
 	}
-	while ((adj = RB_ROOT(global_adj_head, &global.adj_tree)) != NULL)
+	while (!RB_EMPTY(global_adj_head, &global.adj_tree)) {
+		adj = RB_ROOT(global_adj_head, &global.adj_tree);
+
 		adj_del(adj, S_SHUTDOWN);
+	}
 
 	/* clean up */
 	if (iev_lde)

ldpd/neighbor.c

@@ -316,7 +316,9 @@ nbr_del(struct nbr *nbr)
 	mapping_list_clr(&nbr->release_list);
 	mapping_list_clr(&nbr->abortreq_list);
 
-	while ((adj = RB_ROOT(nbr_adj_head, &nbr->adj_tree)) != NULL) {
+	while (!RB_EMPTY(nbr_adj_head, &nbr->adj_tree)) {
+		adj = RB_ROOT(nbr_adj_head, &nbr->adj_tree);
+
 		adj->nbr = NULL;
 		RB_REMOVE(nbr_adj_head, &nbr->adj_tree, adj);
 	}

lib/command.c

@@ -62,7 +62,7 @@ const char *node_names[] = {
 	"aaa",			    // AAA_NODE,
 	"keychain",		    // KEYCHAIN_NODE,
 	"keychain key",		    // KEYCHAIN_KEY_NODE,
-	"logical-router",	    // NS_NODE,
+	"logical-router",	    // LOGICALROUTER_NODE,
 	"vrf",			    // VRF_NODE,
 	"interface",		    // INTERFACE_NODE,
 	"zebra",		    // ZEBRA_NODE,
@@ -1291,7 +1291,7 @@ void cmd_exit(struct vty *vty)
 		break;
 	case INTERFACE_NODE:
 	case PW_NODE:
-	case NS_NODE:
+	case LOGICALROUTER_NODE:
 	case VRF_NODE:
 	case ZEBRA_NODE:
 	case BGP_NODE:
@@ -1376,7 +1376,7 @@ DEFUN (config_end,
 	case CONFIG_NODE:
 	case INTERFACE_NODE:
 	case PW_NODE:
-	case NS_NODE:
+	case LOGICALROUTER_NODE:
 	case VRF_NODE:
 	case ZEBRA_NODE:
 	case RIP_NODE:

lib/command.h

@@ -85,7 +85,7 @@ enum node_type {
 	AAA_NODE,		/* AAA node. */
 	KEYCHAIN_NODE,		/* Key-chain node. */
 	KEYCHAIN_KEY_NODE,      /* Key-chain key node. */
-	NS_NODE,		/* Logical-Router node. */
+	LOGICALROUTER_NODE,	/* Logical-Router node. */
 	VRF_NODE,		/* VRF mode node. */
 	INTERFACE_NODE,		/* Interface mode node. */
 	ZEBRA_NODE,		/* zebra connection node. */

lib/if.c

@@ -384,29 +384,35 @@ struct interface *if_get_by_name(const char *name, vrf_id_t vrf_id, int vty)
 {
 	struct interface *ifp;
 
-	ifp = if_lookup_by_name_all_vrf(name);
+	ifp = if_lookup_by_name(name, vrf_id);
-	if (ifp) {
+	if (ifp)
-		if (ifp->vrf_id == vrf_id)
 		return ifp;
-
+	/* Not Found on same VRF. If the interface command
-		/* Found a match on a different VRF. If the interface command
 	 * was entered in vty without a VRF (passed as VRF_DEFAULT),
 	 * accept the ifp we found. If a vrf was entered and there is
-		 * a mismatch, reject it if from vty. If it came from the kernel
+	 * a mismatch, reject it if from vty.
-		 * or by way of zclient, believe it and update the ifp
-		 * accordingly.
 	 */
+	ifp = if_lookup_by_name_all_vrf(name);
+	if (!ifp)
+		return if_create(name, vrf_id);
 	if (vty) {
 		if (vrf_id == VRF_DEFAULT)
 			return ifp;
 		return NULL;
-		} else {
+	}
+	/* if vrf backend uses NETNS, then
+	 * this should not be considered as an update
+	 * then create the new interface
+	 */
+	if (ifp->vrf_id != vrf_id &&
+	    vrf_is_mapped_on_netns(vrf_id))
+		return if_create(name, vrf_id);
+	/* If it came from the kernel
+	 * or by way of zclient, believe it and update
+	 * the ifp accordingly.
+	 */
 	if_update_to_new_vrf(ifp, vrf_id);
 	return ifp;
-		}
-	}
-
-	return if_create(name, vrf_id);
 }
 
 void if_set_index(struct interface *ifp, ifindex_t ifindex)
@@ -1078,7 +1084,9 @@ void if_terminate(struct vrf *vrf)
 {
 	struct interface *ifp;
 
-	while ((ifp = RB_ROOT(if_name_head, &vrf->ifaces_by_name)) != NULL) {
+	while (!RB_EMPTY(if_name_head, &vrf->ifaces_by_name)) {
+		ifp = RB_ROOT(if_name_head, &vrf->ifaces_by_name);
+
 		if (ifp->node) {
 			ifp->node->info = NULL;
 			route_unlock_node(ifp->node);

lib/logicalrouter.c

@@ -0,0 +1,159 @@
+/*
+ * Logical Router functions.
+ * Copyright (C) 2018 6WIND S.A.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; see the file COPYING; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include <zebra.h>
+
+#include "ns.h"
+#include "log.h"
+#include "memory.h"
+
+#include "command.h"
+#include "vty.h"
+#include "logicalrouter.h"
+
+/* Comment that useless define to avoid compilation error
+ * in order to use it, one could provide the kind of NETNS to NS backend
+ * so that the allocation will match the logical router
+ * DEFINE_MTYPE_STATIC(LIB, LOGICALROUTER, "LogicalRouter Context")
+ */
+DEFINE_MTYPE_STATIC(LIB, LOGICALROUTER_NAME, "Logical Router Name")
+
+/* Logical Router node has no interface. */
+static struct cmd_node logicalrouter_node = {LOGICALROUTER_NODE, "",
+					     1};
+
+static int logicalrouter_backend;
+
+/* Get a NS. If not found, create one. */
+static struct ns *logicalrouter_get(ns_id_t ns_id)
+{
+	struct ns *ns;
+
+	ns = ns_lookup(ns_id);
+	if (ns)
+		return (ns);
+	ns = ns_get_created(ns, NULL, ns_id);
+	return ns;
+}
+
+static int logicalrouter_is_backend_netns(void)
+{
+	return (logicalrouter_backend == LOGICALROUTER_BACKEND_NETNS);
+}
+
+
+DEFUN_NOSH (logicalrouter,
+       logicalrouter_cmd,
+       "logical-router (1-65535) ns NAME",
+       "Enable a logical-router\n"
+       "Specify the logical-router indentifier\n"
+       "The Name Space\n"
+       "The file name in " NS_RUN_DIR ", or a full pathname\n")
+{
+	int idx_number = 1;
+	int idx_name = 3;
+	ns_id_t ns_id;
+	struct ns *ns = NULL;
+	char *pathname = ns_netns_pathname(vty, argv[idx_name]->arg);
+
+	if (!pathname)
+		return CMD_WARNING_CONFIG_FAILED;
+
+	ns_id = strtoul(argv[idx_number]->arg, NULL, 10);
+	ns = logicalrouter_get(ns_id);
+
+	if (ns->name && strcmp(ns->name, pathname) != 0) {
+		vty_out(vty, "NS %u is already configured with NETNS %s\n",
+			ns->ns_id, ns->name);
+		return CMD_WARNING;
+	}
+
+	if (!ns->name)
+		ns->name = XSTRDUP(MTYPE_LOGICALROUTER_NAME, pathname);
+
+	if (!ns_enable(ns, NULL)) {
+		vty_out(vty, "Can not associate NS %u with NETNS %s\n",
+			ns->ns_id, ns->name);
+		return CMD_WARNING_CONFIG_FAILED;
+	}
+
+	return CMD_SUCCESS;
+}
+
+DEFUN (no_logicalrouter,
+       no_logicalrouter_cmd,
+       "no logical-router (1-65535) ns NAME",
+       NO_STR
+       "Enable a Logical-Router\n"
+       "Specify the Logical-Router identifier\n"
+       "The Name Space\n"
+       "The file name in " NS_RUN_DIR ", or a full pathname\n")
+{
+	int idx_number = 2;
+	int idx_name = 4;
+	ns_id_t ns_id;
+	struct ns *ns = NULL;
+	char *pathname = ns_netns_pathname(vty, argv[idx_name]->arg);
+
+	if (!pathname)
+		return CMD_WARNING_CONFIG_FAILED;
+
+	ns_id = strtoul(argv[idx_number]->arg, NULL, 10);
+	ns = ns_lookup(ns_id);
+
+	if (!ns) {
+		vty_out(vty, "NS %u is not found\n", ns_id);
+		return CMD_SUCCESS;
+	}
+
+	if (ns->name && strcmp(ns->name, pathname) != 0) {
+		vty_out(vty, "Incorrect NETNS file name\n");
+		return CMD_WARNING_CONFIG_FAILED;
+	}
+
+	ns_disable(ns);
+
+	if (ns->name) {
+		XFREE(MTYPE_LOGICALROUTER_NAME, ns->name);
+		ns->name = NULL;
+	}
+
+	return CMD_SUCCESS;
+}
+
+/* Initialize NS module. */
+void logicalrouter_init(int (*writefunc)(struct vty *vty))
+{
+	if (ns_have_netns() && logicalrouter_is_backend_netns()) {
+		/* Install LogicalRouter commands. */
+		install_node(&logicalrouter_node, writefunc);
+		install_element(CONFIG_NODE, &logicalrouter_cmd);
+		install_element(CONFIG_NODE, &no_logicalrouter_cmd);
+	}
+}
+
+void logicalrouter_terminate(void)
+{
+	ns_terminate();
+}
+
+void logicalrouter_configure_backend(int backend_netns)
+{
+	logicalrouter_backend = backend_netns;
+}

lib/logicalrouter.h

@@ -0,0 +1,41 @@
+/*
+ * Logical Router related header.
+ * Copyright (C) 2018 6WIND S.A.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; see the file COPYING; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#ifndef _ZEBRA_LOGICAL_ROUTER_H
+#define _ZEBRA_LOGICAL_ROUTER_H
+
+/* Logical Router Backend defines */
+#define LOGICALROUTER_BACKEND_OFF   0
+#define LOGICALROUTER_BACKEND_NETNS 1
+
+/*
+ * Logical Router initializer/destructor
+ */
+extern void logicalrouter_init(int (*writefunc)(struct vty *vty));
+extern void logicalrouter_terminate(void);
+
+/* used to configure backend for logical router
+ * Currently, the whole NETNS feature is exclusively shared
+ * between logical router and VRF backend NETNS
+ * However, when logical router feature will be available,
+ * one can think of having exclusivity only per NETNS
+ */
+extern void logicalrouter_configure_backend(int backend_netns);
+
+#endif /*_ZEBRA_LOGICAL_ROUTER_H*/

lib/netns_linux.c

@@ -0,0 +1,539 @@
+/*
+ * NS functions.
+ * Copyright (C) 2014 6WIND S.A.
+ *
+ * This file is part of GNU Zebra.
+ *
+ * GNU Zebra is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published
+ * by the Free Software Foundation; either version 2, or (at your
+ * option) any later version.
+ *
+ * GNU Zebra is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; see the file COPYING; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include <zebra.h>
+
+#ifdef HAVE_NETNS
+#undef _GNU_SOURCE
+#define _GNU_SOURCE
+
+#include <sched.h>
+#endif
+
+/* for basename */
+#include <libgen.h>
+
+#include "if.h"
+#include "ns.h"
+#include "log.h"
+#include "memory.h"
+
+#include "command.h"
+#include "vty.h"
+#include "vrf.h"
+
+DEFINE_MTYPE_STATIC(LIB, NS, "NetNS Context")
+DEFINE_MTYPE_STATIC(LIB, NS_NAME, "NetNS Name")
+
+static inline int ns_compare(const struct ns *ns, const struct ns *ns2);
+static struct ns *ns_lookup_name_internal(const char *name);
+
+RB_GENERATE(ns_head, ns, entry, ns_compare)
+
+struct ns_head ns_tree = RB_INITIALIZER(&ns_tree);
+
+static struct ns *default_ns;
+static int ns_current_ns_fd;
+static int ns_default_ns_fd;
+
+static int ns_debug;
+
+#ifndef CLONE_NEWNET
+#define CLONE_NEWNET 0x40000000
+/* New network namespace (lo, device, names sockets, etc) */
+#endif
+
+#ifndef HAVE_SETNS
+static inline int setns(int fd, int nstype)
+{
+#ifdef __NR_setns
+	return syscall(__NR_setns, fd, nstype);
+#else
+	errno = EINVAL;
+	return -1;
+#endif
+}
+#endif /* !HAVE_SETNS */
+
+#ifdef HAVE_NETNS
+static int have_netns_enabled = -1;
+#endif /* HAVE_NETNS */
+
+/* default NS ID value used when VRF backend is not NETNS */
+#define NS_DEFAULT_INTERNAL 0
+
+static int have_netns(void)
+{
+#ifdef HAVE_NETNS
+	if (have_netns_enabled < 0) {
+		int fd = open(NS_DEFAULT_NAME, O_RDONLY);
+
+		if (fd < 0)
+			have_netns_enabled = 0;
+		else {
+			have_netns_enabled = 1;
+			close(fd);
+		}
+	}
+	return have_netns_enabled;
+#else
+	return 0;
+#endif
+}
+
+/* Holding NS hooks  */
+struct ns_master {
+	int (*ns_new_hook)(struct ns *ns);
+	int (*ns_delete_hook)(struct ns *ns);
+	int (*ns_enable_hook)(struct ns *ns);
+	int (*ns_disable_hook)(struct ns *ns);
+} ns_master = {
+	0,
+};
+
+static int ns_is_enabled(struct ns *ns);
+
+static inline int ns_compare(const struct ns *a, const struct ns *b)
+{
+	return (a->ns_id - b->ns_id);
+}
+
+/* Look up a NS by identifier. */
+static struct ns *ns_lookup_internal(ns_id_t ns_id)
+{
+	struct ns ns;
+
+	ns.ns_id = ns_id;
+	return RB_FIND(ns_head, &ns_tree, &ns);
+}
+
+/* Look up a NS by name */
+static struct ns *ns_lookup_name_internal(const char *name)
+{
+	struct ns *ns = NULL;
+
+	RB_FOREACH (ns, ns_head, &ns_tree) {
+		if (ns->name != NULL) {
+			if (strcmp(name, ns->name) == 0)
+				return ns;
+		}
+	}
+	return NULL;
+}
+
+static struct ns *ns_get_created_internal(struct ns *ns, char *name,
+					  ns_id_t ns_id)
+{
+	int created = 0;
+	/*
+	 * Initialize interfaces.
+	 */
+	if (!ns && !name && ns_id != NS_UNKNOWN)
+		ns = ns_lookup_internal(ns_id);
+	if (!ns && name)
+		ns = ns_lookup_name_internal(name);
+	if (!ns) {
+		ns = XCALLOC(MTYPE_NS, sizeof(struct ns));
+		ns->ns_id = ns_id;
+		if (name)
+			ns->name = XSTRDUP(MTYPE_NS_NAME, name);
+		ns->fd = -1;
+		RB_INSERT(ns_head, &ns_tree, ns);
+		created = 1;
+	}
+	if (ns_id != ns->ns_id) {
+		RB_REMOVE(ns_head, &ns_tree, ns);
+		ns->ns_id = ns_id;
+		RB_INSERT(ns_head, &ns_tree, ns);
+	}
+	if (!created)
+		return ns;
+	if (ns_debug) {
+		if (ns->ns_id != NS_UNKNOWN)
+			zlog_info("NS %u is created.", ns->ns_id);
+		else
+			zlog_info("NS %s is created.", ns->name);
+	}
+	if (ns_master.ns_new_hook)
+		(*ns_master.ns_new_hook) (ns);
+	return ns;
+}
+
+/*
+ * Enable a NS - that is, let the NS be ready to use.
+ * The NS_ENABLE_HOOK callback will be called to inform
+ * that they can allocate resources in this NS.
+ *
+ * RETURN: 1 - enabled successfully; otherwise, 0.
+ */
+static int ns_enable_internal(struct ns *ns, void (*func)(ns_id_t, void *))
+{
+	if (!ns_is_enabled(ns)) {
+		if (have_netns()) {
+			ns->fd = open(ns->name, O_RDONLY);
+		} else {
+			ns->fd = -2;
+			/* Remember ns_enable_hook has been called */
+			errno = -ENOTSUP;
+		}
+
+		if (!ns_is_enabled(ns)) {
+			zlog_err("Can not enable NS %u: %s!", ns->ns_id,
+				 safe_strerror(errno));
+			return 0;
+		}
+
+		/* Non default NS. leave */
+		if (ns->ns_id == NS_UNKNOWN) {
+			zlog_err("Can not enable NS %s %u: Invalid NSID",
+				 ns->name, ns->ns_id);
+			return 0;
+		}
+		if (func)
+			func(ns->ns_id, (void *)ns->vrf_ctxt);
+		if (ns_debug) {
+			if (have_netns())
+				zlog_info("NS %u is associated with NETNS %s.",
+					  ns->ns_id, ns->name);
+			zlog_info("NS %u is enabled.", ns->ns_id);
+		}
+		/* zebra first receives NS enable event,
+		 * then VRF enable event
+		 */
+		if (ns_master.ns_enable_hook)
+			(*ns_master.ns_enable_hook)(ns);
+	}
+
+	return 1;
+}
+
+/*
+ * Check whether the NS is enabled - that is, whether the NS
+ * is ready to allocate resources. Currently there's only one
+ * type of resource: socket.
+ */
+static int ns_is_enabled(struct ns *ns)
+{
+	if (have_netns())
+		return ns && ns->fd >= 0;
+	else
+		return ns && ns->fd == -2 && ns->ns_id == NS_DEFAULT;
+}
+
+/*
+ * Disable a NS - that is, let the NS be unusable.
+ * The NS_DELETE_HOOK callback will be called to inform
+ * that they must release the resources in the NS.
+ */
+static void ns_disable_internal(struct ns *ns)
+{
+	if (ns_is_enabled(ns)) {
+		if (ns_debug)
+			zlog_info("NS %u is to be disabled.",
+				  ns->ns_id);
+
+		if (ns_master.ns_disable_hook)
+			(*ns_master.ns_disable_hook)(ns);
+
+		if (have_netns())
+			close(ns->fd);
+
+		ns->fd = -1;
+	}
+}
+
+struct ns *ns_get_created(struct ns *ns, char *name, ns_id_t ns_id)
+{
+	return ns_get_created_internal(ns, name, ns_id);
+}
+
+int ns_have_netns(void)
+{
+	return have_netns();
+}
+
+/* Delete a NS. This is called in ns_terminate(). */
+void ns_delete(struct ns *ns)
+{
+	if (ns_debug)
+		zlog_info("NS %u is to be deleted.", ns->ns_id);
+
+	ns_disable(ns);
+
+	if (ns_master.ns_delete_hook)
+		(*ns_master.ns_delete_hook)(ns);
+
+	/*
+	 * I'm not entirely sure if the vrf->iflist
+	 * needs to be moved into here or not.
+	 */
+	// if_terminate (&ns->iflist);
+
+	RB_REMOVE(ns_head, &ns_tree, ns);
+	if (ns->name)
+		XFREE(MTYPE_NS_NAME, ns->name);
+
+	XFREE(MTYPE_NS, ns);
+}
+
+/* Look up the data pointer of the specified VRF. */
+void *
+ns_info_lookup(ns_id_t ns_id)
+{
+	struct ns *ns = ns_lookup_internal(ns_id);
+
+	return ns ? ns->info : NULL;
+}
+
+/* Look up a NS by name */
+struct ns *ns_lookup_name(const char *name)
+{
+	return ns_lookup_name_internal(name);
+}
+
+int ns_enable(struct ns *ns, void (*func)(ns_id_t, void *))
+{
+	return ns_enable_internal(ns, func);
+}
+
+void ns_disable(struct ns *ns)
+{
+	return ns_disable_internal(ns);
+}
+
+struct ns *ns_lookup(ns_id_t ns_id)
+{
+	return ns_lookup_internal(ns_id);
+}
+
+void ns_walk_func(int (*func)(struct ns *))
+{
+	struct ns *ns = NULL;
+
+	RB_FOREACH (ns, ns_head, &ns_tree)
+		func(ns);
+}
+
+const char *ns_get_name(struct ns *ns)
+{
+	if (!ns)
+		return NULL;
+	return ns->name;
+}
+
+/* Add a NS hook. Please add hooks before calling ns_init(). */
+void ns_add_hook(int type, int (*func)(struct ns *))
+{
+	switch (type) {
+	case NS_NEW_HOOK:
+		ns_master.ns_new_hook = func;
+		break;
+	case NS_DELETE_HOOK:
+		ns_master.ns_delete_hook = func;
+		break;
+	case NS_ENABLE_HOOK:
+		ns_master.ns_enable_hook = func;
+		break;
+	case NS_DISABLE_HOOK:
+		ns_master.ns_disable_hook = func;
+		break;
+	default:
+		break;
+	}
+}
+
+/*
+ * NS realization with NETNS
+ */
+
+char *ns_netns_pathname(struct vty *vty, const char *name)
+{
+	static char pathname[PATH_MAX];
+	char *result;
+	char *check_base;
+
+	if (name[0] == '/') /* absolute pathname */
+		result = realpath(name, pathname);
+	else {
+		/* relevant pathname */
+		char tmp_name[PATH_MAX];
+
+		snprintf(tmp_name, PATH_MAX, "%s/%s", NS_RUN_DIR, name);
+		result = realpath(tmp_name, pathname);
+	}
+
+	if (!result) {
+		if (vty)
+			vty_out(vty, "Invalid pathname: %s\n",
+				safe_strerror(errno));
+		else
+			zlog_warn("Invalid pathname: %s",
+				  safe_strerror(errno));
+		return NULL;
+	}
+	check_base = basename(pathname);
+	if (check_base != NULL && strlen(check_base) + 1 > NS_NAMSIZ) {
+		if (vty)
+			vty_out(vty, "NS name (%s) invalid: too long (>%d)\n",
+				check_base, NS_NAMSIZ-1);
+		else
+			zlog_warn("NS name (%s) invalid: too long (>%d)",
+				  check_base, NS_NAMSIZ-1);
+		return NULL;
+	}
+	return pathname;
+}
+
+void ns_init(void)
+{
+	static int ns_initialised;
+
+	ns_debug = 0;
+	/* silently return as initialisation done */
+	if (ns_initialised == 1)
+		return;
+	errno = 0;
+#ifdef HAVE_NETNS
+	if (have_netns_enabled < 0)
+		ns_default_ns_fd = open(NS_DEFAULT_NAME, O_RDONLY);
+	else
+		ns_default_ns_fd = -1;
+#else
+	ns_default_ns_fd = -1;
+	default_ns = NULL;
+#endif /* HAVE_NETNS */
+	if (ns_default_ns_fd == -1)
+		zlog_err("NS initialisation failure (%s)",
+			 safe_strerror(errno));
+	ns_current_ns_fd = -1;
+	ns_initialised = 1;
+}
+
+/* Initialize NS module. */
+void ns_init_management(ns_id_t default_ns_id)
+{
+	int fd;
+
+	ns_init();
+	default_ns = ns_get_created_internal(NULL, NULL, default_ns_id);
+	if (!default_ns) {
+		zlog_err("%s: failed to create the default NS!",
+			 __func__);
+		exit(1);
+	}
+	if (have_netns()) {
+		fd = open(NS_DEFAULT_NAME, O_RDONLY);
+		default_ns->fd = fd;
+	}
+	/* Set the default NS name. */
+	default_ns->name = XSTRDUP(MTYPE_NS_NAME, NS_DEFAULT_NAME);
+	if (ns_debug)
+		zlog_info("%s: default NSID is %u",
+			  __func__, default_ns->ns_id);
+
+	/* Enable the default NS. */
+	if (!ns_enable(default_ns, NULL)) {
+		zlog_err("%s: failed to enable the default NS!",
+			 __func__);
+		exit(1);
+	}
+}
+
+/* Terminate NS module. */
+void ns_terminate(void)
+{
+	struct ns *ns;
+
+	while (!RB_EMPTY(ns_head, &ns_tree)) {
+		ns = RB_ROOT(ns_head, &ns_tree);
+
+		ns_delete(ns);
+	}
+}
+
+int ns_switch_to_netns(const char *name)
+{
+	int ret;
+	int fd;
+
+	if (name == NULL)
+		return -1;
+	if (ns_default_ns_fd == -1)
+		return -1;
+	fd = open(name, O_RDONLY);
+	if (fd == -1) {
+		errno = EINVAL;
+		return -1;
+	}
+	ret = setns(fd, CLONE_NEWNET);
+	ns_current_ns_fd = fd;
+	close(fd);
+	return ret;
+}
+
+/* returns 1 if switch() was not called before
+ * return status of setns() otherwise
+ */
+int ns_switchback_to_initial(void)
+{
+	if (ns_current_ns_fd != -1 && ns_default_ns_fd != -1) {
+		int ret;
+
+		ret = setns(ns_default_ns_fd, CLONE_NEWNET);
+		ns_current_ns_fd = -1;
+		return ret;
+	}
+	/* silently ignore if setns() is not called */
+	return 1;
+}
+
+/* Create a socket for the NS. */
+int ns_socket(int domain, int type, int protocol, ns_id_t ns_id)
+{
+	struct ns *ns = ns_lookup(ns_id);
+	int ret;
+
+	if (!ns || !ns_is_enabled(ns)) {
+		errno = EINVAL;
+		return -1;
+	}
+	if (have_netns()) {
+		ret = (ns_id != NS_DEFAULT) ? setns(ns->fd, CLONE_NEWNET) : 0;
+		if (ret >= 0) {
+			ret = socket(domain, type, protocol);
+			if (ns_id != NS_DEFAULT) {
+				setns(ns_lookup(NS_DEFAULT)->fd, CLONE_NEWNET);
+				ns_current_ns_fd = ns_id;
+			}
+		}
+	} else
+		ret = socket(domain, type, protocol);
+
+	return ret;
+}
+
+ns_id_t ns_get_default_id(void)
+{
+	if (default_ns)
+		return default_ns->ns_id;
+	return NS_UNKNOWN;
+}
+

lib/netns_other.c

@@ -0,0 +1,165 @@
+/*
+ * NetNS backend for non Linux systems
+ * Copyright (C) 2018 6WIND S.A.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; see the file COPYING; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+
+#if !defined(GNU_LINUX) && (defined(SUNOS_5) || defined(OPEN_BSD))
+/* SUNOS_5 or OPEN_BSD */
+
+#include <zebra.h>
+#include "ns.h"
+#include "log.h"
+#include "memory.h"
+
+DEFINE_MTYPE_STATIC(LIB, NS, "NetNS Context")
+DEFINE_MTYPE_STATIC(LIB, NS_NAME, "NetNS Name")
+
+
+static inline int ns_compare(const struct ns *ns, const struct ns *ns2);
+
+RB_GENERATE(ns_head, ns, entry, ns_compare)
+
+struct ns_head ns_tree = RB_INITIALIZER(&ns_tree);
+
+static inline int ns_compare(const struct ns *a, const struct ns *b)
+{
+	return (a->ns_id - b->ns_id);
+}
+
+void ns_terminate(void)
+{
+}
+
+/* API to initialize NETNS managerment
+ * parameter is the default ns_id
+ */
+void ns_init_management(ns_id_t ns_id)
+{
+}
+
+
+/*
+ * NS utilities
+ */
+
+/* Create a socket serving for the given NS
+ */
+int ns_socket(int domain, int type, int protocol, ns_id_t ns_id)
+{
+	return -1;
+}
+
+/* return the path of the NETNS */
+char *ns_netns_pathname(struct vty *vty, const char *name)
+{
+	return NULL;
+}
+
+/* Parse and execute a function on all the NETNS */
+void ns_walk_func(int (*func)(struct ns *))
+{
+}
+
+/* API to get the NETNS name, from the ns pointer */
+const char *ns_get_name(struct ns *ns)
+{
+	return NULL;
+}
+
+/* only called from vrf ( when removing netns from vrf)
+ * or at VRF or logical router termination
+ */
+void ns_delete(struct ns *ns)
+{
+}
+
+/* return > 0 if netns is available
+ * called by VRF to check netns backend is available for VRF
+ */
+int ns_have_netns(void)
+{
+	return 0;
+}
+
+/* API to get context information of a NS */
+void *ns_info_lookup(ns_id_t ns_id)
+{
+	return NULL;
+}
+
+/*
+ * NS init routine
+ * should be called from backendx
+ */
+void ns_init(void)
+{
+}
+
+/* API to retrieve default NS */
+ns_id_t ns_get_default_id(void)
+{
+	return NS_UNKNOWN;
+}
+
+
+/* API that can be used to change from NS */
+int ns_switchback_to_initial(void)
+{
+	return 0;
+}
+int ns_switch_to_netns(const char *netns_name)
+{
+	return 0;
+}
+
+/*
+ * NS handling routines.
+ * called by modules that use NS backend
+ */
+
+/* API to search for already present NETNS */
+struct ns *ns_lookup(ns_id_t ns_id)
+{
+	return NULL;
+}
+
+struct ns *ns_lookup_name(const char *name)
+{
+	return NULL;
+}
+
+/* API to handle NS : creation, enable, disable
+ * for enable, a callback function is passed as parameter
+ * the callback belongs to the module that uses NS as backend
+ * upon enabling the NETNS, the upper layer is informed
+ */
+int ns_enable(struct ns *ns, int (*func)(ns_id_t, void *))
+{
+	return 0;
+}
+
+struct ns *ns_get_created(struct ns *ns, char *name, ns_id_t ns_id)
+{
+	return NULL;
+}
+
+void ns_disable(struct ns *ns)
+{
+}
+
+#endif /* !GNU_LINUX */

lib/ns.c

@@ -1,453 +0,0 @@
-/*
- * NS functions.
- * Copyright (C) 2014 6WIND S.A.
- *
- * This file is part of GNU Zebra.
- *
- * GNU Zebra is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published
- * by the Free Software Foundation; either version 2, or (at your
- * option) any later version.
- *
- * GNU Zebra is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
- * General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License along
- * with this program; see the file COPYING; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
- */
-
-#include <zebra.h>
-
-#ifdef HAVE_NETNS
-#undef _GNU_SOURCE
-#define _GNU_SOURCE
-
-#include <sched.h>
-#endif
-
-#include "if.h"
-#include "ns.h"
-#include "log.h"
-#include "memory.h"
-
-#include "command.h"
-#include "vty.h"
-
-DEFINE_MTYPE_STATIC(LIB, NS, "Logical-Router")
-DEFINE_MTYPE_STATIC(LIB, NS_NAME, "Logical-Router Name")
-
-static __inline int ns_compare(const struct ns *, const struct ns *);
-static struct ns *ns_lookup(ns_id_t);
-
-RB_GENERATE(ns_head, ns, entry, ns_compare)
-
-struct ns_head ns_tree = RB_INITIALIZER(&ns_tree);
-
-#ifndef CLONE_NEWNET
-#define CLONE_NEWNET 0x40000000 /* New network namespace (lo, device, names sockets, etc) */
-#endif
-
-#ifndef HAVE_SETNS
-static inline int setns(int fd, int nstype)
-{
-#ifdef __NR_setns
-	return syscall(__NR_setns, fd, nstype);
-#else
-	errno = ENOSYS;
-	return -1;
-#endif
-}
-#endif /* HAVE_SETNS */
-
-#ifdef HAVE_NETNS
-
-#define NS_DEFAULT_NAME    "/proc/self/ns/net"
-static int have_netns_enabled = -1;
-
-#else  /* !HAVE_NETNS */
-
-#define NS_DEFAULT_NAME    "Default-logical-router"
-
-#endif /* HAVE_NETNS */
-
-static int have_netns(void)
-{
-#ifdef HAVE_NETNS
-	if (have_netns_enabled < 0) {
-		int fd = open(NS_DEFAULT_NAME, O_RDONLY);
-
-		if (fd < 0)
-			have_netns_enabled = 0;
-		else {
-			have_netns_enabled = 1;
-			close(fd);
-		}
-	}
-	return have_netns_enabled;
-#else
-	return 0;
-#endif
-}
-
-/* Holding NS hooks  */
-struct ns_master {
-	int (*ns_new_hook)(ns_id_t, void **);
-	int (*ns_delete_hook)(ns_id_t, void **);
-	int (*ns_enable_hook)(ns_id_t, void **);
-	int (*ns_disable_hook)(ns_id_t, void **);
-} ns_master = {
-	0,
-};
-
-static int ns_is_enabled(struct ns *ns);
-static int ns_enable(struct ns *ns);
-static void ns_disable(struct ns *ns);
-
-static __inline int ns_compare(const struct ns *a, const struct ns *b)
-{
-	return (a->ns_id - b->ns_id);
-}
-
-/* Get a NS. If not found, create one. */
-static struct ns *ns_get(ns_id_t ns_id)
-{
-	struct ns *ns;
-
-	ns = ns_lookup(ns_id);
-	if (ns)
-		return (ns);
-
-	ns = XCALLOC(MTYPE_NS, sizeof(struct ns));
-	ns->ns_id = ns_id;
-	ns->fd = -1;
-	RB_INSERT(ns_head, &ns_tree, ns);
-
-	/*
-	 * Initialize interfaces.
-	 *
-	 * I'm not sure if this belongs here or in
-	 * the vrf code.
-	 */
-	// if_init (&ns->iflist);
-
-	zlog_info("NS %u is created.", ns_id);
-
-	if (ns_master.ns_new_hook)
-		(*ns_master.ns_new_hook)(ns_id, &ns->info);
-
-	return ns;
-}
-
-/* Delete a NS. This is called in ns_terminate(). */
-static void ns_delete(struct ns *ns)
-{
-	zlog_info("NS %u is to be deleted.", ns->ns_id);
-
-	ns_disable(ns);
-
-	if (ns_master.ns_delete_hook)
-		(*ns_master.ns_delete_hook)(ns->ns_id, &ns->info);
-
-	/*
-	 * I'm not entirely sure if the vrf->iflist
-	 * needs to be moved into here or not.
-	 */
-	// if_terminate (&ns->iflist);
-
-	RB_REMOVE(ns_head, &ns_tree, ns);
-	if (ns->name)
-		XFREE(MTYPE_NS_NAME, ns->name);
-
-	XFREE(MTYPE_NS, ns);
-}
-
-/* Look up a NS by identifier. */
-static struct ns *ns_lookup(ns_id_t ns_id)
-{
-	struct ns ns;
-	ns.ns_id = ns_id;
-	return (RB_FIND(ns_head, &ns_tree, &ns));
-}
-
-/*
- * Check whether the NS is enabled - that is, whether the NS
- * is ready to allocate resources. Currently there's only one
- * type of resource: socket.
- */
-static int ns_is_enabled(struct ns *ns)
-{
-	if (have_netns())
-		return ns && ns->fd >= 0;
-	else
-		return ns && ns->fd == -2 && ns->ns_id == NS_DEFAULT;
-}
-
-/*
- * Enable a NS - that is, let the NS be ready to use.
- * The NS_ENABLE_HOOK callback will be called to inform
- * that they can allocate resources in this NS.
- *
- * RETURN: 1 - enabled successfully; otherwise, 0.
- */
-static int ns_enable(struct ns *ns)
-{
-
-	if (!ns_is_enabled(ns)) {
-		if (have_netns()) {
-			ns->fd = open(ns->name, O_RDONLY);
-		} else {
-			ns->fd = -2; /* Remember that ns_enable_hook has been
-					called */
-			errno = -ENOTSUP;
-		}
-
-		if (!ns_is_enabled(ns)) {
-			zlog_err("Can not enable NS %u: %s!", ns->ns_id,
-				 safe_strerror(errno));
-			return 0;
-		}
-
-		if (have_netns())
-			zlog_info("NS %u is associated with NETNS %s.",
-				  ns->ns_id, ns->name);
-
-		zlog_info("NS %u is enabled.", ns->ns_id);
-		if (ns_master.ns_enable_hook)
-			(*ns_master.ns_enable_hook)(ns->ns_id, &ns->info);
-	}
-
-	return 1;
-}
-
-/*
- * Disable a NS - that is, let the NS be unusable.
- * The NS_DELETE_HOOK callback will be called to inform
- * that they must release the resources in the NS.
- */
-static void ns_disable(struct ns *ns)
-{
-	if (ns_is_enabled(ns)) {
-		zlog_info("NS %u is to be disabled.", ns->ns_id);
-
-		if (ns_master.ns_disable_hook)
-			(*ns_master.ns_disable_hook)(ns->ns_id, &ns->info);
-
-		if (have_netns())
-			close(ns->fd);
-
-		ns->fd = -1;
-	}
-}
-
-
-/* Add a NS hook. Please add hooks before calling ns_init(). */
-void ns_add_hook(int type, int (*func)(ns_id_t, void **))
-{
-	switch (type) {
-	case NS_NEW_HOOK:
-		ns_master.ns_new_hook = func;
-		break;
-	case NS_DELETE_HOOK:
-		ns_master.ns_delete_hook = func;
-		break;
-	case NS_ENABLE_HOOK:
-		ns_master.ns_enable_hook = func;
-		break;
-	case NS_DISABLE_HOOK:
-		ns_master.ns_disable_hook = func;
-		break;
-	default:
-		break;
-	}
-}
-
-/*
- * NS realization with NETNS
- */
-
-static char *ns_netns_pathname(struct vty *vty, const char *name)
-{
-	static char pathname[PATH_MAX];
-	char *result;
-
-	if (name[0] == '/') /* absolute pathname */
-		result = realpath(name, pathname);
-	else /* relevant pathname */
-	{
-		char tmp_name[PATH_MAX];
-		snprintf(tmp_name, PATH_MAX, "%s/%s", NS_RUN_DIR, name);
-		result = realpath(tmp_name, pathname);
-	}
-
-	if (!result) {
-		vty_out(vty, "Invalid pathname: %s\n", safe_strerror(errno));
-		return NULL;
-	}
-	return pathname;
-}
-
-DEFUN_NOSH (ns_netns,
-       ns_netns_cmd,
-       "logical-router (1-65535) ns NAME",
-       "Enable a logical-router\n"
-       "Specify the logical-router indentifier\n"
-       "The Name Space\n"
-       "The file name in " NS_RUN_DIR ", or a full pathname\n")
-{
-	int idx_number = 1;
-	int idx_name = 3;
-	ns_id_t ns_id = NS_DEFAULT;
-	struct ns *ns = NULL;
-	char *pathname = ns_netns_pathname(vty, argv[idx_name]->arg);
-
-	if (!pathname)
-		return CMD_WARNING_CONFIG_FAILED;
-
-	ns_id = strtoul(argv[idx_number]->arg, NULL, 10);
-	ns = ns_get(ns_id);
-
-	if (ns->name && strcmp(ns->name, pathname) != 0) {
-		vty_out(vty, "NS %u is already configured with NETNS %s\n",
-			ns->ns_id, ns->name);
-		return CMD_WARNING;
-	}
-
-	if (!ns->name)
-		ns->name = XSTRDUP(MTYPE_NS_NAME, pathname);
-
-	if (!ns_enable(ns)) {
-		vty_out(vty, "Can not associate NS %u with NETNS %s\n",
-			ns->ns_id, ns->name);
-		return CMD_WARNING_CONFIG_FAILED;
-	}
-
-	return CMD_SUCCESS;
-}
-
-DEFUN (no_ns_netns,
-       no_ns_netns_cmd,
-       "no logical-router (1-65535) ns NAME",
-       NO_STR
-       "Enable a Logical-Router\n"
-       "Specify the Logical-Router identifier\n"
-       "The Name Space\n"
-       "The file name in " NS_RUN_DIR ", or a full pathname\n")
-{
-	int idx_number = 2;
-	int idx_name = 4;
-	ns_id_t ns_id = NS_DEFAULT;
-	struct ns *ns = NULL;
-	char *pathname = ns_netns_pathname(vty, argv[idx_name]->arg);
-
-	if (!pathname)
-		return CMD_WARNING_CONFIG_FAILED;
-
-	ns_id = strtoul(argv[idx_number]->arg, NULL, 10);
-	ns = ns_lookup(ns_id);
-
-	if (!ns) {
-		vty_out(vty, "NS %u is not found\n", ns_id);
-		return CMD_SUCCESS;
-	}
-
-	if (ns->name && strcmp(ns->name, pathname) != 0) {
-		vty_out(vty, "Incorrect NETNS file name\n");
-		return CMD_WARNING_CONFIG_FAILED;
-	}
-
-	ns_disable(ns);
-
-	if (ns->name) {
-		XFREE(MTYPE_NS_NAME, ns->name);
-		ns->name = NULL;
-	}
-
-	return CMD_SUCCESS;
-}
-
-/* NS node. */
-static struct cmd_node ns_node = {NS_NODE, "", /* NS node has no interface. */
-				  1};
-
-/* NS configuration write function. */
-static int ns_config_write(struct vty *vty)
-{
-	struct ns *ns;
-	int write = 0;
-
-	RB_FOREACH (ns, ns_head, &ns_tree) {
-		if (ns->ns_id == NS_DEFAULT || ns->name == NULL)
-			continue;
-
-		vty_out(vty, "logical-router %u netns %s\n", ns->ns_id,
-			ns->name);
-		write = 1;
-	}
-
-	return write;
-}
-
-/* Initialize NS module. */
-void ns_init(void)
-{
-	struct ns *default_ns;
-
-	/* The default NS always exists. */
-	default_ns = ns_get(NS_DEFAULT);
-	if (!default_ns) {
-		zlog_err("ns_init: failed to create the default NS!");
-		exit(1);
-	}
-
-	/* Set the default NS name. */
-	default_ns->name = XSTRDUP(MTYPE_NS_NAME, NS_DEFAULT_NAME);
-
-	/* Enable the default NS. */
-	if (!ns_enable(default_ns)) {
-		zlog_err("ns_init: failed to enable the default NS!");
-		exit(1);
-	}
-
-	if (have_netns()) {
-		/* Install NS commands. */
-		install_node(&ns_node, ns_config_write);
-		install_element(CONFIG_NODE, &ns_netns_cmd);
-		install_element(CONFIG_NODE, &no_ns_netns_cmd);
-	}
-}
-
-/* Terminate NS module. */
-void ns_terminate(void)
-{
-	struct ns *ns;
-
-	while ((ns = RB_ROOT(ns_head, &ns_tree)) != NULL)
-		ns_delete(ns);
-}
-
-/* Create a socket for the NS. */
-int ns_socket(int domain, int type, int protocol, ns_id_t ns_id)
-{
-	struct ns *ns = ns_lookup(ns_id);
-	int ret = -1;
-
-	if (!ns_is_enabled(ns)) {
-		errno = ENOSYS;
-		return -1;
-	}
-
-	if (have_netns()) {
-		ret = (ns_id != NS_DEFAULT) ? setns(ns->fd, CLONE_NEWNET) : 0;
-		if (ret >= 0) {
-			ret = socket(domain, type, protocol);
-			if (ns_id != NS_DEFAULT)
-				setns(ns_lookup(NS_DEFAULT)->fd, CLONE_NEWNET);
-		}
-	} else
-		ret = socket(domain, type, protocol);
-
-	return ret;
-}

lib/ns.h

@@ -24,16 +24,22 @@
 
 #include "openbsd-tree.h"
 #include "linklist.h"
+#include "vty.h"
 
 typedef u_int32_t ns_id_t;
 
 /* the default NS ID */
-#define NS_DEFAULT 0
 #define NS_UNKNOWN UINT32_MAX
 
 /* Default netns directory (Linux) */
 #define NS_RUN_DIR         "/var/run/netns"
 
+#ifdef HAVE_NETNS
+#define NS_DEFAULT_NAME    "/proc/self/ns/net"
+#else  /* !HAVE_NETNS */
+#define NS_DEFAULT_NAME    "Default-logical-router"
+#endif /* HAVE_NETNS */
+
 struct ns {
 	RB_ENTRY(ns) entry;
 
@@ -49,6 +55,9 @@ struct ns {
 	/* Master list of interfaces belonging to this NS */
 	struct list *iflist;
 
+	/* Back Pointer to VRF */
+	void *vrf_ctxt;
+
 	/* User data */
 	void *info;
 };
@@ -57,6 +66,11 @@ RB_PROTOTYPE(ns_head, ns, entry, ns_compare)
 
 extern struct ns_head ns_tree;
 
+/*
+ * API for managing NETNS. eg from zebra daemon
+ * one want to manage the list of NETNS, etc...
+ */
+
 /*
  * NS hooks
  */
@@ -74,20 +88,82 @@ extern struct ns_head ns_tree;
  *          - param 2: the address of the user data pointer (the user data
  *                     can be stored in or freed from there)
  */
-extern void ns_add_hook(int, int (*)(ns_id_t, void **));
+extern void ns_add_hook(int type, int (*)(struct ns *));
+
 
 /*
  * NS initializer/destructor
  */
-/* Please add hooks before calling ns_init(). */
+
-extern void ns_init(void);
 extern void ns_terminate(void);
 
+/* API to initialize NETNS managerment
+ * parameter is the default ns_id
+ */
+extern void ns_init_management(ns_id_t ns_id);
+
+
 /*
  * NS utilities
  */
 
-/* Create a socket serving for the given NS */
+/* Create a socket serving for the given NS
-extern int ns_socket(int, int, int, ns_id_t);
+ */
+int ns_socket(int domain, int type, int protocol, ns_id_t ns_id);
+
+/* return the path of the NETNS */
+extern char *ns_netns_pathname(struct vty *vty, const char *name);
+
+/* Parse and execute a function on all the NETNS */
+extern void ns_walk_func(int (*func)(struct ns *));
+
+/* API to get the NETNS name, from the ns pointer */
+extern const char *ns_get_name(struct ns *ns);
+
+/* only called from vrf ( when removing netns from vrf)
+ * or at VRF or logical router termination
+ */
+extern void ns_delete(struct ns *ns);
+
+/* return > 0 if netns is available
+ * called by VRF to check netns backend is available for VRF
+ */
+extern int ns_have_netns(void);
+
+/* API to get context information of a NS */
+extern void *ns_info_lookup(ns_id_t ns_id);
+
+/*
+ * NS init routine
+ * should be called from backendx
+ */
+extern void ns_init(void);
+
+/* API to retrieve default NS */
+extern ns_id_t ns_get_default_id(void);
+
+#define NS_DEFAULT ns_get_default_id()
+
+/* API that can be used to change from NS */
+extern int ns_switchback_to_initial(void);
+extern int ns_switch_to_netns(const char *netns_name);
+
+/*
+ * NS handling routines.
+ * called by modules that use NS backend
+ */
+
+/* API to search for already present NETNS */
+extern struct ns *ns_lookup(ns_id_t ns_id);
+extern struct ns *ns_lookup_name(const char *name);
+
+/* API to handle NS : creation, enable, disable
+ * for enable, a callback function is passed as parameter
+ * the callback belongs to the module that uses NS as backend
+ * upon enabling the NETNS, the upper layer is informed
+ */
+extern int ns_enable(struct ns *ns, void (*func)(ns_id_t, void *));
+extern struct ns *ns_get_created(struct ns *ns, char *name, ns_id_t ns_id);
+extern void ns_disable(struct ns *ns);
 
 #endif /*_ZEBRA_NS_H*/

lib/prefix.h

@@ -239,6 +239,7 @@ static inline void ipv4_addr_copy(struct in_addr *dst,
 #define IPV4_NET127(a)  ((((u_int32_t) (a)) & 0xff000000) == 0x7f000000)
 #define IPV4_LINKLOCAL(a) ((((u_int32_t) (a)) & 0xffff0000) == 0xa9fe0000)
 #define IPV4_CLASS_DE(a)  ((((u_int32_t) (a)) & 0xe0000000) == 0xe0000000)
+#define IPV4_MC_LINKLOCAL(a)  ((((u_int32_t) (a)) & 0xffffff00) == 0xe0000000)
 
 /* Max bit/byte length of IPv6 address. */
 #define IPV6_MAX_BYTELEN    16

lib/subdir.am

@@ -42,7 +42,8 @@ lib_libfrr_la_SOURCES = \
 	lib/module.c \
 	lib/network.c \
 	lib/nexthop.c \
-	lib/ns.c \
+	lib/netns_linux.c \
+	lib/netns_other.c \
 	lib/openbsd-tree.c \
 	lib/pid_output.c \
 	lib/plist.c \
@@ -74,6 +75,7 @@ lib_libfrr_la_SOURCES = \
 	lib/wheel.c \
 	lib/workqueue.c \
 	lib/zclient.c \
+	lib/logicalrouter.c \
 	# end
 
 lib/plist_clippy.c: $(CLIPPY_DEPS)
@@ -158,6 +160,7 @@ pkginclude_HEADERS += \
 	lib/zassert.h \
 	lib/zclient.h \
 	lib/zebra.h \
+	lib/logicalrouter.h \
 	# end
 
 nodist_pkginclude_HEADERS += \

lib/vrf.c

@@ -21,6 +21,9 @@
 
 #include <zebra.h>
 
+/* for basename */
+#include <libgen.h>
+
 #include "if.h"
 #include "vrf.h"
 #include "vrf_int.h"
@@ -29,6 +32,10 @@
 #include "log.h"
 #include "memory.h"
 #include "command.h"
+#include "ns.h"
+
+/* default VRF ID value used when VRF backend is not NETNS */
+#define VRF_DEFAULT_INTERNAL 0
 
 DEFINE_MTYPE_STATIC(LIB, VRF, "VRF")
 DEFINE_MTYPE_STATIC(LIB, VRF_BITMAP, "VRF bit-map")
@@ -44,6 +51,8 @@ RB_GENERATE(vrf_name_head, vrf, name_entry, vrf_name_compare);
 struct vrf_id_head vrfs_by_id = RB_INITIALIZER(&vrfs_by_id);
 struct vrf_name_head vrfs_by_name = RB_INITIALIZER(&vrfs_by_name);
 
+static int vrf_backend;
+
 /*
  * Turn on/off debug code
  * for vrf.
@@ -61,7 +70,6 @@ struct vrf_master {
 };
 
 static int vrf_is_enabled(struct vrf *vrf);
-static void vrf_disable(struct vrf *vrf);
 
 /* VRF list existance check by name. */
 struct vrf *vrf_lookup_by_name(const char *name)
@@ -81,6 +89,54 @@ static int vrf_name_compare(const struct vrf *a, const struct vrf *b)
 	return strcmp(a->name, b->name);
 }
 
+/* if ns_id is different and not VRF_UNKNOWN,
+ * then update vrf identifier, and enable VRF
+ */
+static void vrf_update_vrf_id(ns_id_t ns_id, void *opaqueptr)
+{
+	ns_id_t vrf_id = (vrf_id_t)ns_id;
+	vrf_id_t old_vrf_id;
+	struct vrf *vrf = (struct vrf *)opaqueptr;
+
+	if (!vrf)
+		return;
+	old_vrf_id = vrf->vrf_id;
+	if (vrf_id == vrf->vrf_id)
+		return;
+	if (vrf->vrf_id != VRF_UNKNOWN)
+		RB_REMOVE(vrf_id_head, &vrfs_by_id, vrf);
+	vrf->vrf_id = vrf_id;
+	RB_INSERT(vrf_id_head, &vrfs_by_id, vrf);
+	if (old_vrf_id == VRF_UNKNOWN)
+		vrf_enable((struct vrf *)vrf);
+}
+
+int vrf_switch_to_netns(vrf_id_t vrf_id)
+{
+	char *name;
+	struct vrf *vrf = vrf_lookup_by_id(vrf_id);
+
+	/* VRF is default VRF. silently ignore */
+	if (!vrf || vrf->vrf_id == VRF_DEFAULT)
+		return 0;
+	/* VRF has no NETNS backend. silently ignore */
+	if (vrf->data.l.netns_name[0] == '\0')
+		return 0;
+	name = ns_netns_pathname(NULL, vrf->data.l.netns_name);
+	if (debug_vrf)
+		zlog_debug("VRF_SWITCH: %s(%u)", name, vrf->vrf_id);
+	return ns_switch_to_netns(name);
+}
+
+int vrf_switchback_to_initial(void)
+{
+	int ret = ns_switchback_to_initial();
+
+	if (ret == 0 && debug_vrf)
+		zlog_debug("VRF_SWITCHBACK");
+	return ret;
+}
+
 /* Get a VRF. If not found, create one.
  * Arg:
  *   name   - The name of the vrf.  May be NULL if unknown.
@@ -94,7 +150,8 @@ struct vrf *vrf_get(vrf_id_t vrf_id, const char *name)
 	int new = 0;
 
 	if (debug_vrf)
-		zlog_debug("VRF_GET: %s(%u)", name, vrf_id);
+		zlog_debug("VRF_GET: %s(%u)",
+			   name == NULL ? "(NULL)" : name, vrf_id);
 
 	/* Nothing to see, move along here */
 	if (!name && vrf_id == VRF_UNKNOWN)
@@ -109,8 +166,6 @@ struct vrf *vrf_get(vrf_id_t vrf_id, const char *name)
 	if (vrf == NULL) {
 		vrf = XCALLOC(MTYPE_VRF, sizeof(struct vrf));
 		vrf->vrf_id = VRF_UNKNOWN;
-		RB_INIT(if_name_head, &vrf->ifaces_by_name);
-		RB_INIT(if_index_head, &vrf->ifaces_by_index);
 		QOBJ_REG(vrf, vrf);
 		new = 1;
 
@@ -134,7 +189,6 @@ struct vrf *vrf_get(vrf_id_t vrf_id, const char *name)
 		strlcpy(vrf->name, name, sizeof(vrf->name));
 		RB_INSERT(vrf_name_head, &vrfs_by_name, vrf);
 	}
-
 	if (new &&vrf_master.vrf_new_hook)
 		(*vrf_master.vrf_new_hook)(vrf);
 
@@ -219,7 +273,7 @@ int vrf_enable(struct vrf *vrf)
  * The VRF_DELETE_HOOK callback will be called to inform
  * that they must release the resources in the VRF.
  */
-static void vrf_disable(struct vrf *vrf)
+void vrf_disable(struct vrf *vrf)
 {
 	if (!vrf_is_enabled(vrf))
 		return;
@@ -385,6 +439,8 @@ void vrf_init(int (*create)(struct vrf *), int (*enable)(struct vrf *),
 {
 	struct vrf *default_vrf;
 
+	/* initialise NS, in case VRF backend if NETNS */
+	ns_init();
 	if (debug_vrf)
 		zlog_debug("%s: Initializing VRF subsystem",
 			   __PRETTY_FUNCTION__);
@@ -419,12 +475,17 @@ void vrf_terminate(void)
 		zlog_debug("%s: Shutting down vrf subsystem",
 			   __PRETTY_FUNCTION__);
 
-	while ((vrf = RB_ROOT(vrf_id_head, &vrfs_by_id)) != NULL) {
+	while (!RB_EMPTY(vrf_id_head, &vrfs_by_id)) {
+		vrf = RB_ROOT(vrf_id_head, &vrfs_by_id);
+
 		/* Clear configured flag and invoke delete. */
 		UNSET_FLAG(vrf->status, VRF_CONFIGURED);
 		vrf_delete(vrf);
 	}
-	while ((vrf = RB_ROOT(vrf_name_head, &vrfs_by_name)) != NULL) {
+
+	while (!RB_EMPTY(vrf_name_head, &vrfs_by_name)) {
+		vrf = RB_ROOT(vrf_name_head, &vrfs_by_name);
+
 		/* Clear configured flag and invoke delete. */
 		UNSET_FLAG(vrf->status, VRF_CONFIGURED);
 		vrf_delete(vrf);
@@ -432,13 +493,148 @@ void vrf_terminate(void)
 }
 
 /* Create a socket for the VRF. */
-int vrf_socket(int domain, int type, int protocol, vrf_id_t vrf_id)
+int vrf_socket(int domain, int type, int protocol, vrf_id_t vrf_id,
+	       char *interfacename)
 {
-	int ret = -1;
+	int ret, save_errno, ret2;
 
+	ret = vrf_switch_to_netns(vrf_id);
+	if (ret < 0)
+		zlog_err("%s: Can't switch to VRF %u (%s)",
+			 __func__, vrf_id, safe_strerror(errno));
 	ret = socket(domain, type, protocol);
-
+	save_errno = errno;
+	ret2 = vrf_switchback_to_initial();
+	if (ret2 < 0)
+		zlog_err("%s: Can't switchback from VRF %u (%s)",
+			 __func__, vrf_id, safe_strerror(errno));
+	errno = save_errno;
+	if (ret <= 0)
 		return ret;
+	ret2 = vrf_bind(vrf_id, ret, interfacename);
+	if (ret2 < 0) {
+		close(ret);
+		ret = ret2;
+	}
+	return ret;
+}
+
+int vrf_is_backend_netns(void)
+{
+	return (vrf_backend == VRF_BACKEND_NETNS);
+}
+
+int vrf_get_backend(void)
+{
+	return vrf_backend;
+}
+
+void vrf_configure_backend(int vrf_backend_netns)
+{
+	vrf_backend = vrf_backend_netns;
+}
+
+int vrf_handler_create(struct vty *vty, const char *vrfname, struct vrf **vrf)
+{
+	struct vrf *vrfp;
+
+	if (strlen(vrfname) > VRF_NAMSIZ) {
+		if (vty)
+			vty_out(vty,
+			   "%% VRF name %s invalid: length exceeds %d bytes\n",
+			   vrfname, VRF_NAMSIZ);
+		else
+			zlog_warn(
+			   "%% VRF name %s invalid: length exceeds %d bytes\n",
+			   vrfname, VRF_NAMSIZ);
+		return CMD_WARNING_CONFIG_FAILED;
+	}
+
+	vrfp = vrf_get(VRF_UNKNOWN, vrfname);
+
+	if (vty)
+		VTY_PUSH_CONTEXT(VRF_NODE, vrfp);
+
+	if (vrf)
+		*vrf = vrfp;
+	return CMD_SUCCESS;
+}
+
+int vrf_netns_handler_create(struct vty *vty, struct vrf *vrf,
+			     char *pathname, ns_id_t ns_id)
+{
+	struct ns *ns = NULL;
+
+	if (!vrf)
+		return CMD_WARNING_CONFIG_FAILED;
+	if (vrf->vrf_id != VRF_UNKNOWN && vrf->ns_ctxt == NULL) {
+		if (vty)
+			vty_out(vty,
+				"VRF %u is already configured with VRF %s\n",
+				vrf->vrf_id, vrf->name);
+		else
+			zlog_warn("VRF %u is already configured with VRF %s\n",
+				  vrf->vrf_id, vrf->name);
+		return CMD_WARNING_CONFIG_FAILED;
+	}
+	if (vrf->ns_ctxt != NULL) {
+		ns = (struct ns *) vrf->ns_ctxt;
+		if (ns && 0 != strcmp(ns->name, pathname)) {
+			if (vty)
+				vty_out(vty,
+				   "VRF %u already configured with NETNS %s\n",
+				   vrf->vrf_id, ns->name);
+			else
+				zlog_warn(
+				  "VRF %u already configured with NETNS %s",
+				  vrf->vrf_id, ns->name);
+			return CMD_WARNING_CONFIG_FAILED;
+		}
+	}
+	ns = ns_lookup_name(pathname);
+	if (ns && ns->vrf_ctxt) {
+		struct vrf *vrf2 = (struct vrf *)ns->vrf_ctxt;
+
+		if (vrf2 == vrf)
+			return CMD_SUCCESS;
+		if (vty)
+			vty_out(vty, "NS %s is already configured"
+				" with VRF %u(%s)\n",
+			    ns->name, vrf2->vrf_id, vrf2->name);
+		else
+			zlog_warn("NS %s is already configured with VRF %u(%s)",
+				  ns->name, vrf2->vrf_id, vrf2->name);
+		return CMD_WARNING_CONFIG_FAILED;
+	}
+	ns = ns_get_created(ns, pathname, ns_id);
+	ns->vrf_ctxt = (void *)vrf;
+	vrf->ns_ctxt = (void *)ns;
+	/* update VRF netns NAME */
+	if (vrf)
+		strlcpy(vrf->data.l.netns_name, basename(pathname), NS_NAMSIZ);
+
+	if (!ns_enable(ns, vrf_update_vrf_id)) {
+		if (vty)
+			vty_out(vty, "Can not associate NS %u with NETNS %s\n",
+			    ns->ns_id, ns->name);
+		else
+			zlog_warn("Can not associate NS %u with NETNS %s",
+				  ns->ns_id, ns->name);
+		return CMD_WARNING_CONFIG_FAILED;
+	}
+
+	return CMD_SUCCESS;
+}
+
+int vrf_is_mapped_on_netns(vrf_id_t vrf_id)
+{
+	struct vrf *vrf = vrf_lookup_by_id(vrf_id);
+
+	if (!vrf || vrf->data.l.netns_name[0] == '\0')
+		return 0;
+	if (vrf->vrf_id == VRF_DEFAULT)
+		return 0;
+	return 1;
 }
 
 /* vrf CLI commands */
@@ -450,21 +646,8 @@ DEFUN_NOSH (vrf,
 {
 	int idx_name = 1;
 	const char *vrfname = argv[idx_name]->arg;
-	struct vrf *vrfp;
 
-	if (strlen(vrfname) > VRF_NAMSIZ) {
+	return vrf_handler_create(vty, vrfname, NULL);
-		vty_out(vty,
-			"%% VRF name %s is invalid: length exceeds "
-			"%d characters\n",
-			vrfname, VRF_NAMSIZ);
-		return CMD_WARNING_CONFIG_FAILED;
-	}
-
-	vrfp = vrf_get(VRF_UNKNOWN, vrfname);
-
-	VTY_PUSH_CONTEXT(VRF_NODE, vrfp);
-
-	return CMD_SUCCESS;
 }
 
 DEFUN_NOSH (no_vrf,
@@ -500,6 +683,55 @@ DEFUN_NOSH (no_vrf,
 
 struct cmd_node vrf_node = {VRF_NODE, "%s(config-vrf)# ", 1};
 
+DEFUN_NOSH (vrf_netns,
+	    vrf_netns_cmd,
+	    "netns NAME",
+	    "Attach VRF to a Namespace\n"
+	    "The file name in " NS_RUN_DIR ", or a full pathname\n")
+{
+	int idx_name = 1;
+	char *pathname = ns_netns_pathname(vty, argv[idx_name]->arg);
+
+	VTY_DECLVAR_CONTEXT(vrf, vrf);
+
+	if (!pathname)
+		return CMD_WARNING_CONFIG_FAILED;
+	return vrf_netns_handler_create(vty, vrf, pathname, NS_UNKNOWN);
+}
+
+DEFUN (no_vrf_netns,
+	no_vrf_netns_cmd,
+	"no netns [NAME]",
+	NO_STR
+	"Detach VRF from a Namespace\n"
+	"The file name in " NS_RUN_DIR ", or a full pathname\n")
+{
+	struct ns *ns = NULL;
+
+	VTY_DECLVAR_CONTEXT(vrf, vrf);
+
+	if (!vrf_is_backend_netns()) {
+		vty_out(vty, "VRF backend is not Netns. Aborting\n");
+		return CMD_WARNING_CONFIG_FAILED;
+	}
+	if (!vrf->ns_ctxt) {
+		vty_out(vty, "VRF %s(%u) is not configured with NetNS\n",
+			vrf->name, vrf->vrf_id);
+		return CMD_WARNING_CONFIG_FAILED;
+	}
+
+	ns = (struct ns *)vrf->ns_ctxt;
+
+	ns->vrf_ctxt = NULL;
+	vrf_disable(vrf);
+	/* vrf ID from VRF is necessary for Zebra
+	 * so that propagate to other clients is done
+	 */
+	ns_delete(ns);
+	vrf->ns_ctxt = NULL;
+	return CMD_SUCCESS;
+}
+
 /*
  * Debug CLI for vrf's
  */
@@ -552,4 +784,105 @@ void vrf_cmd_init(int (*writefunc)(struct vty *vty))
 	install_element(CONFIG_NODE, &no_vrf_cmd);
 	install_node(&vrf_node, writefunc);
 	install_default(VRF_NODE);
+	if (vrf_is_backend_netns() && ns_have_netns()) {
+		/* Install NS commands. */
+		install_element(VRF_NODE, &vrf_netns_cmd);
+		install_element(VRF_NODE, &no_vrf_netns_cmd);
+	}
+}
+
+vrf_id_t vrf_get_default_id(void)
+{
+	struct vrf *vrf = vrf_lookup_by_name(VRF_DEFAULT_NAME);
+
+	if (vrf)
+		return vrf->vrf_id;
+	if (vrf_is_backend_netns())
+		return ns_get_default_id();
+	else
+		return VRF_DEFAULT_INTERNAL;
+}
+
+int vrf_bind(vrf_id_t vrf_id, int fd, char *name)
+{
+	int ret = 0;
+
+	if (fd < 0 || name == NULL)
+		return fd;
+	if (vrf_is_mapped_on_netns(vrf_id))
+		return fd;
+#ifdef SO_BINDTODEVICE
+	ret = setsockopt(fd, SOL_SOCKET, SO_BINDTODEVICE, name,
+			 strlen(name));
+	if (ret < 0)
+		zlog_debug("bind to interface %s failed, errno=%d",
+			   name, errno);
+#endif /* SO_BINDTODEVICE */
+	return ret;
+}
+int vrf_getaddrinfo(const char *node, const char *service,
+		    const struct addrinfo *hints,
+		    struct addrinfo **res, vrf_id_t vrf_id)
+{
+	int ret, ret2, save_errno;
+
+	ret = vrf_switch_to_netns(vrf_id);
+	if (ret < 0)
+		zlog_err("%s: Can't switch to VRF %u (%s)",
+			 __func__, vrf_id, safe_strerror(errno));
+	ret = getaddrinfo(node, service, hints, res);
+	save_errno = errno;
+	ret2 = vrf_switchback_to_initial();
+	if (ret2 < 0)
+		zlog_err("%s: Can't switchback from VRF %u (%s)",
+			 __func__, vrf_id, safe_strerror(errno));
+	errno = save_errno;
+	return ret;
+}
+
+int vrf_ioctl(vrf_id_t vrf_id, int d, unsigned long request, char *params)
+{
+	int ret, saved_errno, rc;
+
+	ret = vrf_switch_to_netns(vrf_id);
+	if (ret < 0) {
+		zlog_err("%s: Can't switch to VRF %u (%s)",
+			 __func__, vrf_id, safe_strerror(errno));
+		return 0;
+	}
+	rc = ioctl(d, request, params);
+	saved_errno = errno;
+	ret = vrf_switchback_to_initial();
+	if (ret < 0)
+		zlog_err("%s: Can't switchback from VRF %u (%s)",
+			 __func__, vrf_id, safe_strerror(errno));
+	errno = saved_errno;
+	return rc;
+}
+
+int vrf_sockunion_socket(const union sockunion *su, vrf_id_t vrf_id,
+			 char *interfacename)
+{
+	int ret, save_errno, ret2;
+
+	ret = vrf_switch_to_netns(vrf_id);
+	if (ret < 0)
+		zlog_err("%s: Can't switch to VRF %u (%s)",
+			 __func__, vrf_id, safe_strerror(errno));
+	ret = sockunion_socket(su);
+	save_errno = errno;
+	ret2 = vrf_switchback_to_initial();
+	if (ret2 < 0)
+		zlog_err("%s: Can't switchback from VRF %u (%s)",
+			 __func__, vrf_id, safe_strerror(errno));
+	errno = save_errno;
+
+	if (ret <= 0)
+		return ret;
+	ret2 = vrf_bind(vrf_id, ret, interfacename);
+	if (ret2 < 0) {
+		close(ret);
+		ret = ret2;
+	}
+	return ret;
 }

lib/vrf.h

@@ -26,12 +26,9 @@
 #include "linklist.h"
 #include "qobj.h"
 #include "vty.h"
-
+#include "ns.h"
-/* The default NS ID */
-#define NS_DEFAULT 0
 
 /* The default VRF ID */
-#define VRF_DEFAULT 0
 #define VRF_UNKNOWN UINT32_MAX
 
 /* Pending: May need to refine this. */
@@ -42,6 +39,7 @@ enum { IFLA_VRF_UNSPEC, IFLA_VRF_TABLE, __IFLA_VRF_MAX };
 #endif
 
 #define VRF_NAMSIZ      36
+#define NS_NAMSIZ       16
 
 #define VRF_DEFAULT_NAME    "Default-IP-Routing-Table"
 
@@ -60,6 +58,7 @@ struct vrf_data {
 	union {
 		struct {
 			uint32_t table_id;
+			char netns_name[NS_NAMSIZ];
 		} l;
 	};
 };
@@ -88,6 +87,9 @@ struct vrf {
 	/* The table_id from the kernel */
 	struct vrf_data data;
 
+	/* Back pointer to namespace context */
+	void *ns_ctxt;
+
 	QOBJ_FIELDS
 };
 RB_HEAD(vrf_id_head, vrf);
@@ -96,6 +98,9 @@ RB_HEAD(vrf_name_head, vrf);
 RB_PROTOTYPE(vrf_name_head, vrf, name_entry, vrf_name_compare)
 DECLARE_QOBJ_TYPE(vrf)
 
+/* Allow VRF with netns as backend */
+#define VRF_BACKEND_VRF_LITE 0
+#define VRF_BACKEND_NETNS    1
 
 extern struct vrf_id_head vrfs_by_id;
 extern struct vrf_name_head vrfs_by_name;
@@ -195,17 +200,85 @@ extern void vrf_init(int (*create)(struct vrf *), int (*enable)(struct vrf *),
  */
 extern void vrf_terminate(void);
 
+/*
+ * Utilities to create networks objects,
+ * or call network operations
+ */
+
+/* Create a socket serving for the given VRF */
+extern int vrf_socket(int domain, int type,
+		      int protocol, vrf_id_t vrf_id,
+		      char *name);
+
+extern int vrf_sockunion_socket(const union sockunion *su,
+				vrf_id_t vrf_id, char *name);
+
+extern int vrf_bind(vrf_id_t vrf_id, int fd, char *name);
+
+/* VRF ioctl operations */
+extern int vrf_getaddrinfo(const char *node, const char *service,
+		    const struct addrinfo *hints,
+		    struct addrinfo **res, vrf_id_t vrf_id);
+
+extern int vrf_ioctl(vrf_id_t vrf_id, int d, unsigned long request, char *args);
+
+/* function called by macro VRF_DEFAULT
+ * to get the default VRF_ID
+ */
+extern vrf_id_t vrf_get_default_id(void);
+/* The default VRF ID */
+#define VRF_DEFAULT vrf_get_default_id()
+
+/* VRF is mapped on netns or not ? */
+int vrf_is_mapped_on_netns(vrf_id_t vrf_id);
+
+/* VRF switch from NETNS */
+extern int vrf_switch_to_netns(vrf_id_t vrf_id);
+extern int vrf_switchback_to_initial(void);
+
+/*
+ * VRF backend routines
+ * should be called from zebra only
+ */
+
+/* VRF vty command initialisation
+ */
 extern void vrf_cmd_init(int (*writefunc)(struct vty *vty));
 
+/* VRF vty debugging
+ */
+extern void vrf_install_commands(void);
+
 /*
  * VRF utilities
  */
 
-/* Create a socket serving for the given VRF */
+/* API for configuring VRF backend
-extern int vrf_socket(int, int, int, vrf_id_t);
+ * should be called from zebra only
-
-/*
- * VRF Debugging
  */
-extern void vrf_install_commands(void);
+extern void vrf_configure_backend(int vrf_backend_netns);
+extern int vrf_get_backend(void);
+extern int vrf_is_backend_netns(void);
+
+
+/* API to create a VRF. either from vty
+ * or through discovery
+ */
+extern int vrf_handler_create(struct vty *vty,
+			      const char *name,
+			      struct vrf **vrf);
+
+/* API to associate a VRF with a NETNS.
+ * called either from vty or through discovery
+ * should be called from zebra only
+ */
+extern int vrf_netns_handler_create(struct vty *vty, struct vrf *vrf,
+			     char *pathname, ns_id_t ns_id);
+
+/* used internally to enable or disable VRF.
+ * Notify a change in the VRF ID of the VRF
+ */
+extern void vrf_disable(struct vrf *vrf);
+extern int vrf_enable(struct vrf *vrf);
+
 #endif /*_ZEBRA_VRF_H*/

lib/zclient.c

@@ -363,7 +363,7 @@ static int zebra_hello_send(struct zclient *zclient)
 	return 0;
 }
 
-void zclient_send_vrf_label(struct zclient *zclient, vrf_id_t vrf_id,
+void zclient_send_vrf_label(struct zclient *zclient, vrf_id_t vrf_id, afi_t afi,
 			    mpls_label_t label, enum lsp_types_t ltype)
 {
 	struct stream *s;
@@ -373,6 +373,7 @@ void zclient_send_vrf_label(struct zclient *zclient, vrf_id_t vrf_id,
 
 	zclient_create_header(s, ZEBRA_VRF_LABEL, vrf_id);
 	stream_putl(s, label);
+	stream_putc(s, afi);
 	stream_putc(s, ltype);
 	stream_putw_at(s, 0, stream_get_endp(s));
 	zclient_send_message(zclient);
@@ -1209,14 +1210,20 @@ stream_failure:
 }
 
 bool zapi_route_notify_decode(struct stream *s, struct prefix *p,
+			      uint32_t *tableid,
 			      enum zapi_route_notify_owner *note)
 {
+	uint32_t t;
+
 	STREAM_GET(note, s, sizeof(*note));
 
 	STREAM_GETC(s, p->family);
 	STREAM_GETC(s, p->prefixlen);
 	STREAM_GET(&p->u.prefix, s,
-		   PSIZE(p->prefixlen));
+		   prefix_blen(p));
+	STREAM_GETL(s, t);
+
+	*tableid = t;
 
 	return true;
 
@@ -1401,8 +1408,8 @@ static void zclient_vrf_add(struct zclient *zclient, vrf_id_t vrf_id)
 
 	/* Lookup/create vrf by vrf_id. */
 	vrf = vrf_get(vrf_id, vrfname_tmp);
-	vrf->data = data;
+	vrf->data.l.table_id = data.l.table_id;
-
+	memcpy(vrf->data.l.netns_name, data.l.netns_name, NS_NAMSIZ);
 	vrf_enable(vrf);
 }
 
@@ -2362,9 +2369,9 @@ static int zclient_read(struct thread *thread)
 						     vrf_id);
 		break;
 	case ZEBRA_ROUTE_NOTIFY_OWNER:
-		if (zclient->notify_owner)
+		if (zclient->route_notify_owner)
-			(*zclient->notify_owner)(command, zclient,
+			(*zclient->route_notify_owner)(command, zclient, length,
-						 length, vrf_id);
+						       vrf_id);
 		break;
 	default:
 		break;

lib/zclient.h

@@ -213,7 +213,7 @@ struct zclient {
 	int (*local_macip_add)(int, struct zclient *, uint16_t, vrf_id_t);
 	int (*local_macip_del)(int, struct zclient *, uint16_t, vrf_id_t);
 	int (*pw_status_update)(int, struct zclient *, uint16_t, vrf_id_t);
-	int (*notify_owner)(int command, struct zclient *zclient,
+	int (*route_notify_owner)(int command, struct zclient *zclient,
 				  uint16_t length, vrf_id_t vrf_id);
 };
 
@@ -390,9 +390,15 @@ extern void redist_del_instance(struct redist_proto *, u_short);
  * label for lookup.  If you pass in MPLS_LABEL_NONE
  * we will cause a delete action and remove this label pop
  * operation.
+ *
+ * The underlying AF_MPLS doesn't care about afi's
+ * but we can make the zebra_vrf keep track of what
+ * we have installed and play some special games
+ * to get them both installed.
  */
 extern void zclient_send_vrf_label(struct zclient *zclient, vrf_id_t vrf_id,
-				   mpls_label_t label, enum lsp_types_t ltype);
+				   afi_t afi, mpls_label_t label,
+				   enum lsp_types_t ltype);
 
 extern void zclient_send_reg_requests(struct zclient *, vrf_id_t);
 extern void zclient_send_dereg_requests(struct zclient *, vrf_id_t);
@@ -507,6 +513,7 @@ extern int zclient_send_rnh(struct zclient *zclient, int command,
 extern int zapi_route_encode(u_char, struct stream *, struct zapi_route *);
 extern int zapi_route_decode(struct stream *, struct zapi_route *);
 bool zapi_route_notify_decode(struct stream *s, struct prefix *p,
+			      uint32_t *tableid,
 			      enum zapi_route_notify_owner *note);
 extern struct nexthop *nexthop_from_zapi_nexthop(struct zapi_nexthop *znh);
 extern bool zapi_nexthop_update_decode(struct stream *s,

lib/zebra.h

@@ -402,7 +402,7 @@ extern const char *zserv_command_string(unsigned int command);
 #define strmatch(a,b) (!strcmp((a), (b)))
 
 /* Zebra message flags */
-#define ZEBRA_FLAG_INTERNAL           0x01
+#define ZEBRA_FLAG_ALLOW_RECURSION    0x01
 #define ZEBRA_FLAG_SELFROUTE          0x02
 #define ZEBRA_FLAG_IBGP               0x08
 #define ZEBRA_FLAG_SELECTED           0x10

nhrpd/nhrp_route.c

@@ -114,7 +114,7 @@ void nhrp_route_announce(int add, enum nhrp_cache_type type, const struct prefix
 		SET_FLAG(api.flags, ZEBRA_FLAG_FIB_OVERRIDE);
 		break;
 	}
-	SET_FLAG(api.flags, ZEBRA_FLAG_INTERNAL);
+	SET_FLAG(api.flags, ZEBRA_FLAG_ALLOW_RECURSION);
 
 	SET_FLAG(api.message, ZAPI_MESSAGE_NEXTHOP);
 	api.nexthop_num = 1;

ospf6d/ospf6_area.h

@@ -99,6 +99,8 @@ struct ospf6_area {
 
 	/* Time stamps. */
 	struct timeval ts_spf; /* SPF calculation time stamp. */
+
+	uint32_t full_nbrs; /* Fully adjacent neighbors. */
 };
 
 #define OSPF6_AREA_ENABLE     0x01

ospf6d/ospf6_neighbor.c

@@ -193,7 +193,11 @@ static void ospf6_neighbor_state_change(u_char next_state,
 		if (prev_state == OSPF6_NEIGHBOR_LOADING &&
 		    next_state == OSPF6_NEIGHBOR_FULL) {
 			OSPF6_AS_EXTERN_LSA_SCHEDULE(on->ospf6_if);
+			on->ospf6_if->area->full_nbrs++;
 		}
+
+		if (prev_state == OSPF6_NEIGHBOR_FULL)
+			on->ospf6_if->area->full_nbrs--;
 	}
 
 	if ((prev_state == OSPF6_NEIGHBOR_EXCHANGE

ospf6d/ospf6_spf.c

@@ -1029,19 +1029,22 @@ struct ospf6_lsa *ospf6_create_single_router_lsa(struct ospf6_area *area,
 
 	/* Fill Larger LSA Payload */
 	end = ospf6_lsdb_head(lsdb, 2, type, adv_router, &rtr_lsa);
-	if (rtr_lsa) {
+
+	/*
+	 * We assume at this point in time that rtr_lsa is
+	 * a valid pointer.
+	 */
+	assert(rtr_lsa);
 	if (!OSPF6_LSA_IS_MAXAGE(rtr_lsa)) {
 		/* Append first Link State ID LSA */
 		lsa_header = (struct ospf6_lsa_header *)rtr_lsa->header;
-			memcpy(new_header, lsa_header,
+		memcpy(new_header, lsa_header, ntohs(lsa_header->length));
-				ntohs(lsa_header->length));
 		/* Assign new lsa length as aggregated length. */
 		((struct ospf6_lsa_header *)new_header)->length =
 			htons(total_lsa_length);
 		new_header += ntohs(lsa_header->length);
 		num_lsa--;
 	}
-	}
 
 	/* Print LSA Name */
 	ospf6_lsa_printbuf(lsa, lsa->name, sizeof(lsa->name));

ospf6d/ospf6_top.c

@@ -335,6 +335,8 @@ DEFUN(ospf6_router_id,
 	int ret;
 	const char *router_id_str;
 	u_int32_t router_id;
+	struct ospf6_area *oa;
+	struct listnode *node;
 
 	argv_find(argv, argc, "A.B.C.D", &idx);
 	router_id_str = argv[idx]->arg;
@@ -346,7 +348,16 @@ DEFUN(ospf6_router_id,
 	}
 
 	o->router_id_static = router_id;
-	if (o->router_id == 0)
+
+	for (ALL_LIST_ELEMENTS_RO(o->area_list, node, oa)) {
+		if (oa->full_nbrs) {
+			vty_out(vty,
+				"For this router-id change to take effect,"
+				" save config and restart ospf6d\n");
+			return CMD_SUCCESS;
+		}
+	}
+
 	o->router_id = router_id;
 
 	return CMD_SUCCESS;
@@ -360,8 +371,22 @@ DEFUN(no_ospf6_router_id,
       V4NOTATION_STR)
 {
 	VTY_DECLVAR_CONTEXT(ospf6, o);
+	struct ospf6_area *oa;
+	struct listnode *node;
+
 	o->router_id_static = 0;
+
+	for (ALL_LIST_ELEMENTS_RO(o->area_list, node, oa)) {
+		if (oa->full_nbrs) {
+			vty_out(vty,
+				"For this router-id change to take effect,"
+				" save config and restart ospf6d\n");
+			return CMD_SUCCESS;
+		}
+	}
 	o->router_id = 0;
+	if (o->router_id_zebra.s_addr)
+		o->router_id = (uint32_t)o->router_id_zebra.s_addr;
 
 	return CMD_SUCCESS;
 }
@@ -523,6 +548,10 @@ DEFUN (ospf6_distance_ospf6,
 	VTY_DECLVAR_CONTEXT(ospf6, o);
 	int idx = 0;
 
+	o->distance_intra = 0;
+	o->distance_inter = 0;
+	o->distance_external = 0;
+
 	if (argv_find(argv, argc, "intra-area", &idx))
 		o->distance_intra = atoi(argv[idx + 1]->arg);
 	idx = 0;

ospf6d/ospf6_top.h

@@ -32,6 +32,8 @@ struct ospf6 {
 	/* static router id */
 	u_int32_t router_id_static;
 
+	struct in_addr router_id_zebra;
+
 	/* start time */
 	struct timeval starttime;
 

ospf6d/ospf6_zebra.c

@@ -46,8 +46,6 @@ unsigned char conf_debug_ospf6_zebra = 0;
 /* information about zebra. */
 struct zclient *zclient = NULL;
 
-struct in_addr router_id_zebra;
-
 /* Router-id update message from zebra. */
 static int ospf6_router_id_update_zebra(int command, struct zclient *zclient,
 					zebra_size_t length, vrf_id_t vrf_id)
@@ -56,13 +54,14 @@ static int ospf6_router_id_update_zebra(int command, struct zclient *zclient,
 	struct ospf6 *o = ospf6;
 
 	zebra_router_id_update_read(zclient->ibuf, &router_id);
-	router_id_zebra = router_id.u.prefix4;
 
 	if (o == NULL)
 		return 0;
 
+	o->router_id_zebra = router_id.u.prefix4;
+
 	if (o->router_id == 0)
-		o->router_id = (u_int32_t)router_id_zebra.s_addr;
+		o->router_id = (uint32_t)o->router_id_zebra.s_addr;
 
 	return 0;
 }

ospfd/ospf_ext.c

@@ -175,7 +175,7 @@ void ospf_ext_term(void)
 {
 
 	if ((OspfEXT.scope != OSPF_OPAQUE_AREA_LSA)
-	    || (OspfEXT.scope != OSPF_OPAQUE_AS_LSA))
+	    && (OspfEXT.scope != OSPF_OPAQUE_AS_LSA))
 		zlog_warn(
 			"EXT: Unable to unregister Extended Prefix "
 			"Opaque LSA functions: Wrong scope!");
@@ -334,10 +334,12 @@ static void set_prefix_sid(struct ext_itf *exti, uint8_t algorithm,
 
 	/* Set Label or Index value */
 	if (value_type == SID_LABEL) {
-		TLV_LEN(exti->node_sid) = htons(SID_LABEL_SIZE);
+		TLV_LEN(exti->node_sid) =
+			htons(SID_LABEL_SIZE(EXT_SUBTLV_PREFIX_SID_SIZE));
 		exti->node_sid.value = htonl(SET_LABEL(value));
 	} else {
-		TLV_LEN(exti->node_sid) = htons(SID_INDEX_SIZE);
+		TLV_LEN(exti->node_sid) =
+			htons(SID_INDEX_SIZE(EXT_SUBTLV_PREFIX_SID_SIZE));
 		exti->node_sid.value = htonl(value);
 	}
 
@@ -370,11 +372,13 @@ static void set_adj_sid(struct ext_itf *exti, bool backup, uint32_t value,
 	/* Adjust Length, Flags and Value depending on the type of Label */
 	if (value_type == SID_LABEL) {
 		SET_FLAG(flags, EXT_SUBTLV_LINK_ADJ_SID_VFLG);
-		TLV_LEN(exti->adj_sid[index]) = htons(SID_LABEL_SIZE);
+		TLV_LEN(exti->adj_sid[index]) =
+			htons(SID_LABEL_SIZE(EXT_SUBTLV_ADJ_SID_SIZE));
 		exti->adj_sid[index].value = htonl(SET_LABEL(value));
 	} else {
 		UNSET_FLAG(flags, EXT_SUBTLV_LINK_ADJ_SID_VFLG);
-		TLV_LEN(exti->adj_sid[index]) = htons(SID_INDEX_SIZE);
+		TLV_LEN(exti->adj_sid[index]) =
+			htons(SID_INDEX_SIZE(EXT_SUBTLV_ADJ_SID_SIZE));
 		exti->adj_sid[index].value = htonl(value);
 	}
 
@@ -402,7 +406,7 @@ static void set_lan_adj_sid(struct ext_itf *exti, bool backup, uint32_t value,
 	}
 
 	/* Set Header */
-	TLV_TYPE(exti->lan_sid[index]) = htons(EXT_SUBTLV_ADJ_SID);
+	TLV_TYPE(exti->lan_sid[index]) = htons(EXT_SUBTLV_LAN_ADJ_SID);
 
 	/* Only Local ADJ-SID is supported for the moment */
 	SET_FLAG(flags, EXT_SUBTLV_LINK_ADJ_SID_LFLG);
@@ -410,11 +414,13 @@ static void set_lan_adj_sid(struct ext_itf *exti, bool backup, uint32_t value,
 	/* Adjust Length, Flags and Value depending on the type of Label */
 	if (value_type == SID_LABEL) {
 		SET_FLAG(flags, EXT_SUBTLV_LINK_ADJ_SID_VFLG);
-		TLV_LEN(exti->lan_sid[index]) = htons(SID_LABEL_SIZE);
+		TLV_LEN(exti->lan_sid[index]) =
+			htons(SID_LABEL_SIZE(EXT_SUBTLV_PREFIX_RANGE_SIZE));
 		exti->lan_sid[index].value = htonl(SET_LABEL(value));
 	} else {
 		UNSET_FLAG(flags, EXT_SUBTLV_LINK_ADJ_SID_VFLG);
-		TLV_LEN(exti->lan_sid[index]) = htons(SID_INDEX_SIZE);
+		TLV_LEN(exti->lan_sid[index]) =
+			htons(SID_INDEX_SIZE(EXT_SUBTLV_PREFIX_RANGE_SIZE));
 		exti->lan_sid[index].value = htonl(value);
 	}
 
@@ -906,7 +912,7 @@ static void ospf_ext_link_lsa_body_set(struct stream *s, struct ext_itf *exti)
 		/* Build LSA body for an Extended Link TLV with Adj. SID */
 		build_tlv_header(s, &exti->link.header);
 		stream_put(s, TLV_DATA(&exti->link.header), EXT_TLV_LINK_SIZE);
-		/* then add Ajacency SubTLVs */
+		/* then add Adjacency SubTLVs */
 		build_tlv(s, &exti->adj_sid[1].header);
 		build_tlv(s, &exti->adj_sid[0].header);
 
@@ -921,8 +927,8 @@ static void ospf_ext_link_lsa_body_set(struct stream *s, struct ext_itf *exti)
 
 		/* Build LSA body for an Extended Link TLV with LAN SID */
 		build_tlv_header(s, &exti->link.header);
-		stream_put(s, &exti->link.header, EXT_TLV_LINK_SIZE);
+		stream_put(s, TLV_DATA(&exti->link.header), EXT_TLV_LINK_SIZE);
-		/* then add LAN-Ajacency SubTLVs */
+		/* then add LAN-Adjacency SubTLVs */
 		build_tlv(s, &exti->lan_sid[1].header);
 		build_tlv(s, &exti->lan_sid[0].header);
 	}
@@ -1671,8 +1677,9 @@ static uint16_t show_vty_ext_link_lan_adj_sid(struct vty *vty,
 	vty_out(vty,
 		"  LAN-Adj-SID Sub-TLV: Length %u\n\tFlags: "
 		"0x%x\n\tMT-ID:0x%x\n\tWeight: 0x%x\n\tNeighbor ID: "
-		"%s\n\tLabel: %u\n",
+		"%s\n\t%s: %u\n",
 		ntohs(top->header.length), top->flags, top->mtid, top->weight,
+		inet_ntoa(top->neighbor_id),
 		CHECK_FLAG(top->flags, EXT_SUBTLV_LINK_ADJ_SID_VFLG) ? "Label"
 								     : "Index",
 		CHECK_FLAG(top->flags, EXT_SUBTLV_LINK_ADJ_SID_VFLG)

ospfd/ospf_flood.c

@@ -563,6 +563,7 @@ int ospf_flood_through_area(struct ospf_area *area, struct ospf_neighbor *inbr,
 	struct ospf_interface *oi;
 	int lsa_ack_flag = 0;
 
+	assert(area);
 	/* All other types are specific to a single area (Area A).  The
 	   eligible interfaces are all those interfaces attaching to the
 	   Area A.  If Area A is the backbone, this includes all the virtual

ospfd/ospf_network.c

@@ -169,42 +169,27 @@ int ospf_if_ipmulticast(struct ospf *top, struct prefix *p, ifindex_t ifindex)
 	return ret;
 }
 
-int ospf_bind_vrfdevice(struct ospf *ospf, int ospf_sock)
-{
-	int ret = 0;
-
-#ifdef SO_BINDTODEVICE
-
-	if (ospf && ospf->vrf_id != VRF_DEFAULT &&
-	    ospf->vrf_id != VRF_UNKNOWN) {
-		ret = setsockopt(ospf_sock, SOL_SOCKET, SO_BINDTODEVICE,
-				 ospf->name,
-				 strlen(ospf->name));
-		if (ret < 0) {
-			int save_errno = errno;
-
-			zlog_warn("%s: Could not setsockopt SO_BINDTODEVICE %s",
-					__PRETTY_FUNCTION__,
-					safe_strerror(save_errno));
-		}
-
-	}
-#endif
-	return ret;
-}
-
 int ospf_sock_init(struct ospf *ospf)
 {
 	int ospf_sock;
 	int ret, hincl = 1;
 	int bufsize = (8 * 1024 * 1024);
 
+	/* silently ignore. already done */
+	if (ospf->fd > 0)
+		return -1;
+
+	if (ospf->vrf_id == VRF_UNKNOWN) {
+		/* silently return since VRF is not ready */
+		return -1;
+	}
 	if (ospfd_privs.change(ZPRIVS_RAISE)) {
 		zlog_err("ospf_sock_init: could not raise privs, %s",
 			 safe_strerror(errno));
 	}
 
-	ospf_sock = socket(AF_INET, SOCK_RAW, IPPROTO_OSPFIGP);
+	ospf_sock = vrf_socket(AF_INET, SOCK_RAW, IPPROTO_OSPFIGP,
+			       ospf->vrf_id, ospf->name);
 	if (ospf_sock < 0) {
 		int save_errno = errno;
 
@@ -216,12 +201,6 @@ int ospf_sock_init(struct ospf *ospf)
 		exit(1);
 	}
 
-	ret = ospf_bind_vrfdevice(ospf, ospf_sock);
-	if (ret < 0) {
-		close(ospf_sock);
-		goto out;
-	}
-
 #ifdef IP_HDRINCL
 	/* we will include IP header with packet */
 	ret = setsockopt(ospf_sock, IPPROTO_IP, IP_HDRINCL, &hincl,

ospfd/ospf_network.h

@@ -30,6 +30,5 @@ extern int ospf_if_add_alldrouters(struct ospf *, struct prefix *, ifindex_t);
 extern int ospf_if_drop_alldrouters(struct ospf *, struct prefix *, ifindex_t);
 extern int ospf_if_ipmulticast(struct ospf *, struct prefix *, ifindex_t);
 extern int ospf_sock_init(struct ospf *ospf);
-extern int ospf_bind_vrfdevice(struct ospf *, int);
 
 #endif /* _ZEBRA_OSPF_NETWORK_H */

ospfd/ospf_opaque.c

@@ -75,6 +75,7 @@ static void ospf_opaque_funclist_init(void);
 static void ospf_opaque_funclist_term(void);
 static void free_opaque_info_per_type(void *val);
 static void free_opaque_info_per_id(void *val);
+static void free_opaque_info_owner(void *val);
 static int ospf_opaque_lsa_install_hook(struct ospf_lsa *lsa);
 static int ospf_opaque_lsa_delete_hook(struct ospf_lsa *lsa);
 
@@ -439,9 +440,11 @@ void ospf_delete_opaque_functab(u_char lsa_type, u_char opaque_type)
 			if (functab->opaque_type == opaque_type) {
 				/* Cleanup internal control information, if it
 				 * still remains. */
-				if (functab->oipt != NULL)
+				if (functab->oipt != NULL) {
 					free_opaque_info_per_type(
 						functab->oipt);
+					free_opaque_info_owner(functab->oipt);
+				}
 
 				/* Dequeue listnode entry from the list. */
 				listnode_delete(funclist, functab);
@@ -572,6 +575,7 @@ register_opaque_info_per_type(struct ospf_opaque_functab *functab,
 		top = ospf_lookup_by_vrf_id(new->vrf_id);
 		if (new->area != NULL && (top = new->area->ospf) == NULL) {
 			free_opaque_info_per_type((void *)oipt);
+			free_opaque_info_owner(oipt);
 			oipt = NULL;
 			goto out; /* This case may not exist. */
 		}
@@ -583,6 +587,7 @@ register_opaque_info_per_type(struct ospf_opaque_functab *functab,
 			"register_opaque_info_per_type: Unexpected LSA-type(%u)",
 			new->data->type);
 		free_opaque_info_per_type((void *)oipt);
+		free_opaque_info_owner(oipt);
 		oipt = NULL;
 		goto out; /* This case may not exist. */
 	}
@@ -600,6 +605,35 @@ out:
 	return oipt;
 }
 
+/* Remove "oipt" from its owner's self-originated LSA list. */
+static void free_opaque_info_owner(void *val)
+{
+	struct opaque_info_per_type *oipt = (struct opaque_info_per_type *)val;
+
+        switch (oipt->lsa_type) {
+        case OSPF_OPAQUE_LINK_LSA: {
+                struct ospf_interface *oi =
+                        (struct ospf_interface *)(oipt->owner);
+                listnode_delete(oi->opaque_lsa_self, oipt);
+                break;
+        }
+        case OSPF_OPAQUE_AREA_LSA: {
+                struct ospf_area *area = (struct ospf_area *)(oipt->owner);
+                listnode_delete(area->opaque_lsa_self, oipt);
+                break;
+        }
+        case OSPF_OPAQUE_AS_LSA: {
+                struct ospf *top = (struct ospf *)(oipt->owner);
+                listnode_delete(top->opaque_lsa_self, oipt);
+                break;
+        }
+        default:
+                zlog_warn("free_opaque_info_owner: Unexpected LSA-type(%u)",
+                          oipt->lsa_type);
+                break; /* This case may not exist. */
+        }
+}
+
 static void free_opaque_info_per_type(void *val)
 {
 	struct opaque_info_per_type *oipt = (struct opaque_info_per_type *)val;

ospfd/ospf_packet.c

@@ -3897,6 +3897,10 @@ static void ospf_ls_upd_queue_send(struct ospf_interface *oi,
 		zlog_debug("listcount = %d, [%s]dst %s", listcount(update),
 			   IF_NAME(oi), inet_ntoa(addr));
 
+	/* Check that we have really something to process */
+	if (listcount(update) == 0)
+		return;
+
 	op = ospf_ls_upd_packet_new(update, oi);
 
 	/* Prepare OSPF common header. */

ospfd/ospf_routemap.c

@@ -337,6 +337,7 @@ static struct route_map_rule_cmd route_match_tag_cmd = {
 };
 
 struct ospf_metric {
+	enum { metric_increment, metric_decrement, metric_absolute } type;
 	bool used;
 	u_int32_t metric;
 };
@@ -356,8 +357,19 @@ static route_map_result_t route_set_metric(void *rule, struct prefix *prefix,
 		ei = object;
 
 		/* Set metric out value. */
-		if (metric->used)
+		if (!metric->used)
+			return RMAP_OKAY;
+		if (metric->type == metric_increment)
+			ei->route_map_set.metric += metric->metric;
+		if (metric->type == metric_decrement)
+			ei->route_map_set.metric -= metric->metric;
+		if (metric->type == metric_absolute)
 			ei->route_map_set.metric = metric->metric;
+
+		if ((signed int)ei->route_map_set.metric < 1)
+			ei->route_map_set.metric = -1;
+		if (ei->route_map_set.metric > OSPF_LS_INFINITY)
+			ei->route_map_set.metric = OSPF_LS_INFINITY;
 	}
 	return RMAP_OKAY;
 }
@@ -370,22 +382,27 @@ static void *route_set_metric_compile(const char *arg)
 	metric = XCALLOC(MTYPE_ROUTE_MAP_COMPILED, sizeof(u_int32_t));
 	metric->used = false;
 
-	/* OSPF doesn't support the +/- in
+	if (all_digit(arg))
-	   set metric <+/-metric> check
+		metric->type = metric_absolute;
-	   Ignore the +/- component */
-	if (!all_digit(arg)) {
-		if ((arg[0] == '+' || arg[0] == '-') && all_digit(arg + 1)) {
-			zlog_warn("OSPF does not support 'set metric +/-'");
-			arg++;
-		} else {
-			if (strmatch(arg, "+rtt") || strmatch(arg, "-rtt"))
-				zlog_warn(
-					"OSPF does not support 'set metric +rtt / -rtt'");
 
+	if (strmatch(arg, "+rtt") || strmatch(arg, "-rtt")) {
+		zlog_warn("OSPF does not support 'set metric +rtt / -rtt'");
 		return metric;
 	}
+
+	if ((arg[0] == '+') && all_digit(arg + 1)) {
+		metric->type = metric_increment;
+		arg++;
 	}
+
+	if ((arg[0] == '-') && all_digit(arg + 1)) {
+		metric->type = metric_decrement;
+		arg++;
+	}
+
 	metric->metric = strtoul(arg, NULL, 10);
+
+	if (metric->metric)
 		metric->used = true;
 
 	return metric;

ospfd/ospf_sr.c

@@ -149,14 +149,6 @@ static struct sr_node *sr_node_new(struct in_addr *rid)
 	new->ext_link->del = del_sr_link;
 	new->ext_prefix->del = del_sr_pref;
 
-	/* Check if list are correctly created */
-	if (new->ext_link == NULL || new->ext_prefix == NULL) {
-		list_delete_original(new->ext_link);
-		list_delete_original(new->ext_prefix);
-		XFREE(MTYPE_OSPF_SR_PARAMS, new);
-		return NULL;
-	}
-
 	IPV4_ADDR_COPY(&new->adv_router, rid);
 	new->neighbor = NULL;
 	new->instance = 0;
@@ -440,7 +432,7 @@ static struct ospf_path *get_nexthop_by_addr(struct ospf *top,
 	struct route_node *rn;
 
 	/* Sanity Check */
-	if ((top == NULL) && (top->new_table))
+	if (top == NULL)
 		return NULL;
 
 	if (IS_DEBUG_OSPF_SR)
@@ -499,6 +491,12 @@ static int compute_link_nhlfe(struct sr_link *srl)
 	srl->nhlfe[0].ifindex = nh->oi->ifp->ifindex;
 	srl->nhlfe[1].ifindex = nh->oi->ifp->ifindex;
 
+	/* Update neighbor address for LAN_ADJ_SID */
+	if (srl->type == LAN_ADJ_SID) {
+		IPV4_ADDR_COPY(&srl->nhlfe[0].nexthop, &nh->src);
+		IPV4_ADDR_COPY(&srl->nhlfe[1].nexthop, &nh->src);
+	}
+
 	/* Set Input & Output Label */
 	if (CHECK_FLAG(srl->flags[0], EXT_SUBTLV_LINK_ADJ_SID_VFLG))
 		srl->nhlfe[0].label_in = srl->sid[0];

ospfd/ospf_sr.h

@@ -49,11 +49,10 @@
 /* Segment Routing TLVs as per draft-ietf-ospf-segment-routing-extensions-19 */
 
 /* Segment ID could be a Label (3 bytes) or an Index (4 bytes) */
-#define SID_BASE_SIZE	4
 #define SID_LABEL	3
-#define SID_LABEL_SIZE	(SID_BASE_SIZE + SID_LABEL)
+#define SID_LABEL_SIZE(U) (U - 1)
 #define SID_INDEX	4
-#define SID_INDEX_SIZE	(SID_BASE_SIZE + SID_INDEX)
+#define SID_INDEX_SIZE(U) (U)
 
 /* SID/Label Sub TLV - section 2.1 */
 #define SUBTLV_SID_LABEL		1

ospfd/ospf_vty.c

@@ -3693,7 +3693,7 @@ static int show_ip_ospf_interface_common(struct vty *vty, struct ospf *ospf,
 {
 	struct interface *ifp;
 	struct vrf *vrf = vrf_lookup_by_id(ospf->vrf_id);
-	json_object *json_vrf = NULL, *json_intf_array = NULL;
+	json_object *json_vrf = NULL;
 	json_object *json_interface_sub = NULL, *json_interface = NULL;
 
 	if (use_json) {
@@ -3701,7 +3701,7 @@ static int show_ip_ospf_interface_common(struct vty *vty, struct ospf *ospf,
 			json_vrf = json_object_new_object();
 		else
 			json_vrf = json;
-		json_intf_array = json_object_new_array();
+		json_interface = json_object_new_object();
 	}
 
 	if (ospf->instance) {
@@ -3715,15 +3715,10 @@ static int show_ip_ospf_interface_common(struct vty *vty, struct ospf *ospf,
 	ospf_show_vrf_name(ospf, vty, json_vrf, use_vrf);
 
 	if (intf_name == NULL) {
-		if (use_json)
-			json_object_object_add(json_vrf, "interfaces",
-				       json_intf_array);
 		/* Show All Interfaces.*/
 		FOR_ALL_INTERFACES (vrf, ifp) {
 			if (ospf_oi_count(ifp)) {
 				if (use_json) {
-					json_interface =
-						json_object_new_object();
 					json_interface_sub =
 						json_object_new_object();
 				}
@@ -3732,14 +3727,15 @@ static int show_ip_ospf_interface_common(struct vty *vty, struct ospf *ospf,
 							   use_json);
 
 				if (use_json) {
-					json_object_array_add(json_intf_array,
-							      json_interface);
 					json_object_object_add(
 						json_interface, ifp->name,
 						json_interface_sub);
 				}
 			}
 		}
+		if (use_json)
+			json_object_object_add(json_vrf, "interfaces",
+				       json_interface);
 	} else {
 		/* Interface name is specified. */
 		ifp = if_lookup_by_name(intf_name, ospf->vrf_id);
@@ -3753,19 +3749,17 @@ static int show_ip_ospf_interface_common(struct vty *vty, struct ospf *ospf,
 			if (use_json) {
 				json_interface_sub = json_object_new_object();
 				json_interface = json_object_new_object();
-				json_object_object_add(json_vrf, "interfaces",
-						       json_intf_array);
 			}
 
 			show_ip_ospf_interface_sub(
 				vty, ospf, ifp, json_interface_sub, use_json);
 
 			if (use_json) {
-				json_object_array_add(json_intf_array,
-						      json_interface);
 				json_object_object_add(json_interface,
 						       ifp->name,
 						       json_interface_sub);
+				json_object_object_add(json_vrf, "interfaces",
+						       json_interface);
 			}
 		}
 	}
@@ -4300,7 +4294,7 @@ static int show_ip_ospf_neighbor_common(struct vty *vty, struct ospf *ospf,
 {
 	struct ospf_interface *oi;
 	struct listnode *node;
-	json_object *json_vrf = NULL, *json_nbr_array = NULL;
+	json_object *json_vrf = NULL;
 	json_object *json_nbr_sub = NULL;
 
 	if (use_json) {
@@ -4308,7 +4302,7 @@ static int show_ip_ospf_neighbor_common(struct vty *vty, struct ospf *ospf,
 			json_vrf = json_object_new_object();
 		else
 			json_vrf = json;
-		json_nbr_array = json_object_new_array();
+		json_nbr_sub = json_object_new_object();
 	}
 
 	if (ospf->instance) {
@@ -4322,21 +4316,16 @@ static int show_ip_ospf_neighbor_common(struct vty *vty, struct ospf *ospf,
 	ospf_show_vrf_name(ospf, vty, json_vrf, use_vrf);
 	if (!use_json)
 		show_ip_ospf_neighbour_header(vty);
-	else
-		json_object_object_add(json_vrf, "neighbors",
-				       json_nbr_array);
 
 	for (ALL_LIST_ELEMENTS_RO(ospf->oiflist, node, oi)) {
 		if (ospf_interface_neighbor_count(oi) == 0)
 			continue;
-		if (use_json) {
-			json_nbr_sub = json_object_new_object();
-			json_object_array_add(json_nbr_array, json_nbr_sub);
-		}
 		show_ip_ospf_neighbor_sub(vty, oi, json_nbr_sub, use_json);
 	}
 
 	if (use_json) {
+		json_object_object_add(json_vrf, "neighbors",
+				       json_nbr_sub);
 		if (use_vrf) {
 			if (ospf->vrf_id == VRF_DEFAULT)
 				json_object_object_add(json, "default",
@@ -8558,6 +8547,10 @@ DEFUN (ospf_distance_ospf,
 	VTY_DECLVAR_INSTANCE_CONTEXT(ospf, ospf);
 	int idx = 0;
 
+	ospf->distance_intra = 0;
+	ospf->distance_inter = 0;
+	ospf->distance_external = 0;
+
 	if (argv_find(argv, argc, "intra-area", &idx))
 		ospf->distance_intra = atoi(argv[idx + 1]->arg);
 	idx = 0;

ospfd/ospfd.c

@@ -308,12 +308,6 @@ static struct ospf *ospf_new(u_short instance, const char *name)
 			 new->lsa_refresh_interval, &new->t_lsa_refresher);
 	new->lsa_refresher_started = monotime(NULL);
 
-	if ((ospf_sock_init(new)) < 0) {
-		zlog_err(
-			"ospf_new: fatal error: ospf_sock_init was unable to open "
-			"a socket");
-		exit(1);
-	}
 	if ((new->ibuf = stream_new(OSPF_MAX_PACKET_SIZE + 1)) == NULL) {
 		zlog_err(
 			"ospf_new: fatal error: stream_new(%u) failed allocating ibuf",
@@ -321,7 +315,6 @@ static struct ospf *ospf_new(u_short instance, const char *name)
 		exit(1);
 	}
 	new->t_read = NULL;
-	thread_add_read(master, ospf_read, new, new->fd, &new->t_read);
 	new->oi_write_q = list_new();
 	new->write_oi_count = OSPF_WRITE_INTERFACE_COUNT_DEFAULT;
 
@@ -332,6 +325,16 @@ static struct ospf *ospf_new(u_short instance, const char *name)
 
 	QOBJ_REG(new, ospf);
 
+	new->fd = -1;
+	if ((ospf_sock_init(new)) < 0) {
+		if (new->vrf_id != VRF_UNKNOWN)
+			zlog_warn(
+				  "%s: ospf_sock_init is unable to open a socket",
+				  __func__);
+		return new;
+	}
+	thread_add_read(master, ospf_read, new, new->fd, &new->t_read);
+
 	return new;
 }
 
@@ -2049,7 +2052,8 @@ static int ospf_vrf_delete(struct vrf *vrf)
 static int ospf_vrf_enable(struct vrf *vrf)
 {
 	struct ospf *ospf = NULL;
-	vrf_id_t old_vrf_id = VRF_DEFAULT;
+	vrf_id_t old_vrf_id;
+	int ret = 0;
 
 	if (IS_DEBUG_OSPF_EVENT)
 		zlog_debug("%s: VRF %s id %u enabled",
@@ -2070,13 +2074,15 @@ static int ospf_vrf_enable(struct vrf *vrf)
 				zlog_err("ospf_sock_init: could not raise privs, %s",
 					 safe_strerror(errno));
 			}
-			if (ospf_bind_vrfdevice(ospf, ospf->fd) < 0)
+			ret = ospf_sock_init(ospf);
-				return 0;
 			if (ospfd_privs.change(ZPRIVS_LOWER)) {
 				zlog_err("ospf_sock_init: could not lower privs, %s",
 					 safe_strerror(errno));
 			}
-
+			if (ret < 0 || ospf->fd <= 0)
+				return 0;
+			thread_add_read(master, ospf_read, ospf,
+					ospf->fd, &ospf->t_read);
 			ospf->oi_running = 1;
 			ospf_zebra_vrf_register(ospf);
 			ospf_router_id_update(ospf);
@@ -2111,6 +2117,9 @@ static int ospf_vrf_disable(struct vrf *vrf)
 		if (IS_DEBUG_OSPF_EVENT)
 			zlog_debug("%s: ospf old_vrf_id %d unlinked",
 				    __PRETTY_FUNCTION__, old_vrf_id);
+		thread_cancel(ospf->t_read);
+		close(ospf->fd);
+		ospf->fd = -1;
 	}
 
 	/* Note: This is a callback, the VRF will be deleted by the caller. */

pimd/.gitignore

@@ -2,6 +2,7 @@
 Makefile.in
 libpim.a
 pimd
+mtracebis
 test_igmpv3_join
 tags
 TAGS

pimd/CAVEATS

@@ -173,4 +173,10 @@ C19 Provision to prevent group mode clash
     flow. There could be some provision to prevent such a behavior,
     but currently there is none.
 
+C20 Multicast traceroute module is based on:
+    draft-ietf-idmr-traceroute-ipm-07
+    It only implements, so far, weak traceroutes. The multicast routing
+    state of the router is not quieried but RPF path is followed along
+    PIM and IGMP enabled interfaces.
+
 -x-

pimd/COMMANDS

@@ -59,6 +59,7 @@ debug commands:
        clear ip pim interfaces		Reset PIM interfaces
        clear ip pim oil			Rescan PIM OIL (output interface list)
        debug igmp			IGMP protocol activity
+       debug mtrace			Mtrace protocol activity
        debug mroute			PIM interaction with kernel MFC cache
        debug pim			PIM protocol activity
        debug pim zebra			ZEBRA protocol activity
@@ -76,4 +77,6 @@ statistics commands:
        pimd:
        show memory pim		PIM memory statistics
 
+vtysh:
+	mtrace			Multicast traceroute
 -x-

pimd/TODO

@@ -1,375 +1,21 @@
-T1 DONE Implement debug command
+T1  Consider reliable pim solution (refresh reduction)
-   test pim receive join
-
-T2 DONE Implement debug command
-   test pim receive prune
-
-T3 DONE Per-interface Downstream (S,G) state machine
-   (RFC 4601 4.5.3. Receiving (S,G) Join/Prune Messages)
-
-T4 DONE Upstream (S,G) state machine
-   (RFC 4601 4.5.7. Sending (S,G) Join/Prune Messages)
-
-T5 DONE Verify Data Packet Forwarding Rules
-   RFC 4601 4.2.  Data Packet Forwarding Rules
-   RFC 4601 4.8.2.  PIM-SSM-Only Routers
-
-   Additionally, the Packet forwarding rules of Section 4.2 can be
-   simplified in a PIM-SSM-only router:
-
-     iif is the incoming interface of the packet.
-     oiflist = NULL
-     if (iif == RPF_interface(S) AND UpstreamJPState(S,G) == Joined) {
-       oiflist = inherited_olist(S,G)
-     } else if (iif is in inherited_olist(S,G)) {
-       send Assert(S,G) on iif
-     }
-     oiflist = oiflist (-) iif
-     forward packet on all interfaces in oiflist
-
-   Macro:
-     inherited_olist(S,G) =
-       joins(S,G) (+) pim_include(S,G) (-) lost_assert(S,G)
-
-T6 DONE Implement (S,G) Assert state machine (RFC 4601, section 4.6.1).
-   Changes in pim_ifchannel.ifassert_winner should trigger
-   pim_upstream_update_join_desired().
-   Depends on TODO T27.
-   Depends on TODO T33.
-   See also CAVEAT C7.
-   See also: RFC 4601 4.5.7. Sending (S,G) Join/Prune Messages
-    Transitions from Joined State
-     RPF'(S,G) changes due to an Assert
-
-   http://www.hep.ucl.ac.uk/~ytl/multi-cast/pim-dm_01.html:
-
-     The PIM Assert mechanism is used to shutoff duplicate flows onto
-     the same multiaccess network. Routers detect this condiction when
-     they receive an (S,G) packet via a multi-access interface that is
-     in the (S,G) OIL. This causes the routers to send Assert
-     Messages.
-
-   Note that neighbors will not accept Join/Prune or Assert messages
-   from a router unless they have first heard a Hello message from that
-   router.  Thus, if a router needs to send a Join/Prune or Assert
-   message on an interface on which it has not yet sent a Hello message
-   with the currently configured IP address, then it MUST immediately
-   send the relevant Hello message without waiting for the Hello Timer
-   to expire, followed by the Join/Prune or Assert message.
-
-T7 DONE Implement hello option: LAN Prune Delay
-
-T8 DONE Implement J/P_Override_Interval(I)
-   Depends on TODO T7.
-   See pim_ifchannel.c, pim_ifchannel_prune(), jp_override_interval.
-
-T9 DONE Detect change in IGMPv3 RPF interface/next-hop for S and update.
-   channel_oil vif index accordingly ?
-   Beware accidentaly adding looped MFC entries (IIF=OIF).
-
-T10 DONE React to (S,G) join directed to another upstream address. See
-    also:
-    
-    RFC 4601: 4.5.7.  Sending (S,G) Join/Prune Messages
-
-    If a router wishes to propagate a Join(S,G) upstream, it must also
-    watch for messages on its upstream interface from other routers on
-    that subnet, and these may modify its behavior.  If it sees a
-    Join(S,G) to the correct upstream neighbor, it should suppress its
-    own Join(S,G).  If it sees a Prune(S,G), Prune(S,G,rpt), or
-    Prune(*,G) to the correct upstream neighbor towards S, it should
-    be prepared to override that prune by scheduling a Join(S,G) to be
-    sent almost immediately.
-
-T11 DONE Review protocol modifications for SSM
-    (RFC 4601 4.8.1.  Protocol Modifications for SSM Destination
-    Addresses)
-
-T12 DONE Review updates of RPF entries.
-    FIXME pim_upstream.c send_join():
-    Currently only one upstream state is affected by detection of RPF change.
-    RPF change should affect all upstream states sharing the RPF cache.
-
-T13 DONE Check that RFC macros using S,G,RPF_interface(S) are actually
-    implemented with this strategy:
-    rpf_ifch=find_ifch(up->rpf->interface).
-    See pim_rpf.c pim_rpf_find_rpf_addr() for a correct example.
-
-    $ grep -i macro pimd/*.c
-    pimd/pim_iface.c:  RFC 4601: 4.1.6.  State Summarization Macros
-    pimd/pim_ifchannel.c:    RFC 4601: 4.6.5.  Assert State Macros
-    pimd/pim_ifchannel.c:  RFC 4601: 4.1.6.  State Summarization Macros
-    pimd/pim_ifchannel.c:  RFC 4601: 4.1.6.  State Summarization Macros
-    pimd/pim_ifchannel.c:  RFC 4601: 4.6.5.  Assert State Macros
-    pimd/pim_ifchannel.c:  Macro:
-    pimd/pim_rpf.c:  RFC 4601: 4.1.6.  State Summarization Macros
-
-T14 DONE Send Assert(S,G) on iif as response to WRONGVIF kernel upcall.
-    See pim_mroute.c mroute_msg().
-
-T15 DONE Interface command to statically join (S,G).
-    interface eth0
-     ip igmp join-group 239.1.1.1 source 1.1.1.1
-
-T16 DONE RPF'(S,G) lookup is not working for S reachable with default route.
-    See "RPF'(S,G) not found" in pim_rpf_update() from pim_rpf.c.
-    Zebra daemon RIB is not reflecting changes in kernel routes
-    accurately?
-
-T17 DONE Prevent CLI from creating bogus interfaces.
-    Example:
-    conf t
-     interface xxx
-
-T18 Consider reliable pim solution (refresh reduction)
     A Reliable Transport Mechanism for PIM
     http://tools.ietf.org/wg/pim/draft-ietf-pim-port/
     PORT=PIM-Over-Reliable-Transport
 
-T19 DONE Fix self as neighbor 
+T2  If an interface changes one of its secondary IP addresses, a Hello
-    See mailing list post:
-    http://lists.gnu.org/archive/html/qpimd-users/2009-04/msg00000.html
-
-T20 DONE Fix debug message: "pim_neighbor_update: internal error:
-    trying to replace same prefix list"
-    See mailing list post:
-    http://lists.gnu.org/archive/html/qpimd-users/2009-04/msg00000.html
-
-T21 DONE Clean-up PIM/IGMP interface mismatch debugging
-    See option PIM_CHECK_RECV_IFINDEX_SANITY in pimd/Makefile.am
-    See mailing list post:
-    http://lists.nongnu.org/archive/html/qpimd-users/2009-04/msg00003.html
-
-T22 DONE IGMP must be protected against adding looped MFC entries
-    created by both source and receiver attached to the same
-    interface.
-
-T23 DONE libfrr crash after zclient_lookup_nexthop.
-    See mailing list post:
-    http://lists.nongnu.org/archive/html/qpimd-users/2009-04/msg00008.html
-
-T24 DONE zserv may return recursive routes:
-     - nexthop type is set to ZEBRA_NEXTHOP_IPV4
-     - ifindex is not reported
-     - calls expecting ifindex (fib_lookup_if_vif_index) are disrupted
-    See also this mailing list post:
-    [PATCH 21/21] Link detect and recursive routes
-    http://www.gossamer-threads.com/lists/quagga/dev/17564
-
-T25 DONE Zclient nexthop lookup missing OSPF route to 1.1.1.1/32
-    See also:
-    pim_zlookup.c zclient_lookup_nexthop misses OSPF 1.1.1.1/32
-    zebra/zebra_vty.c show_ip_route_addr_cmd hits OSPF 1.1.1.1/32
-
-T26 DONE Zebra daemon is marking recursive static route as inactive.
-
-    FIXED: zebra daemon was incorrectly marking recursive routes
-    pointing to kernel routes as inactive:
-      zebra/zebra_rib.c nexthop_active_ipv4:
-        -- Original:
-	  else if (CHECK_FLAG (rib->flags, ZEBRA_FLAG_INTERNAL))
-        -- Fixed:
-	  else if (CHECK_FLAG (rib->flags, ZEBRA_FLAG_INTERNAL) ||
-		   match->type == ZEBRA_ROUTE_KERNEL)
-
-    Old problem description:
-
-    This prevents rib_match_ipv4 from returning its nexthop:
-    client: pim_zlookup.c zclient_read_nexthop
-    server: zebra/zserv.c zsend_ipv4_nexthop_lookup_v2 -> rib_match_ipv4
-
-    Kernel route is injected into zebra in zebra_rib.c rib_add_ipv4
-    Examples:
-    rt_netlink.c:726: rib_add_ipv4 (ZEBRA_ROUTE_KERNEL, flags, &p, gate, src, index, table, metric, 0);
-    rt_netlink.c:864: rib_add_ipv4 (ZEBRA_ROUTE_KERNEL, 0, &p, gate, src, index, table, 0, 0);
-
-    This patch didn't fix the issue:
-    [PATCH 21/21] Link detect and recursive routes
-    http://www.gossamer-threads.com/lists/quagga/dev/17564
-
-    See the example below for the route 2.2.2.2.
-
-bash# route add -host 1.1.1.1 gw 127.0.0.1
-bash# route add -host 2.2.2.2 gw 1.1.1.1
-bash# netstat -nvr
-Kernel IP routing table
-Destination     Gateway         Genmask         Flags   MSS Window  irtt Iface
-2.2.2.2         1.1.1.1         255.255.255.255 UGH       0 0          0 lo
-1.1.1.1         127.0.0.1       255.255.255.255 UGH       0 0          0 lo
-192.168.0.0     0.0.0.0         255.255.255.0   U         0 0          0 eth0
-0.0.0.0         192.168.0.2     0.0.0.0         UG        0 0          0 eth0
-bash# 
-
-zebra# sh ip route         
-Codes: K - kernel route, C - connected, S - static, R - RIP, O - OSPF,
-       I - ISIS, B - BGP, > - selected route, * - FIB route
-
-K>* 0.0.0.0/0 via 192.168.0.2, eth0
-K>* 1.1.1.1/32 via 127.0.0.1, lo
-K * 2.2.2.2/32 via 1.1.1.1, lo inactive
-C>* 127.0.0.0/8 is directly connected, lo
-C>* 192.168.0.0/24 is directly connected, eth0
-
-quagga-pimd-router# sh ip route 1.1.1.1
-Address         NextHop         Interface Metric Preference
-1.1.1.1         127.0.0.1       lo             0          0
-quagga-pimd-router# 
-quagga-pimd-router# sh ip route 2.2.2.2
-Address         NextHop         Interface Metric Preference
-2.2.2.2         192.168.0.2     eth0           0          0
-quagga-pimd-router# 
-
-T27 DONE Implement debug command
-    test pim receive assert
-    See also TODO T6: (S,G) Assert state machine.
-
-T28 DONE Bad IPv4 address family=02 in Join/Prune dump
-    Reported by Andrew Lunn <andrew.lunn@ascom.ch>
-    
-    # 58-byte pim v2 Join/Prune dump
-    # ------------------------------
-    # IPv4 address family=02 is wrong, correct IPv4 address family is 01
-    # See http://www.iana.org/assignments/address-family-numbers
-    #
-    c8XX YY03 : ip src 200.xx.yy.3
-    e000 000d : ip dst 224.0.0.13
-    9404 0000 : ip router alert option 148.4.0.0
-    2300 ab13 : pimv2,type=3 res=00 checksum=ab13
-    0200      : upstream family=02, encoding=00
-    c8XX YY08 : upstream 200.xx.yy.8
-    0001 00d2 : res=00 groups=01 holdtime=00d2
-    0200 0020 : group family=02, encoding=00, res=00, mask_len=20
-    ef01 0101 : group address 239.1.1.1
-    0001 0000 : joined=0001 pruned=0000
-    0200 0020 : source family=02, encoding=00, res=00, mask_len=20
-    0101 0101 : source address 1.1.1.1
-
-T29 DONE Reset interface PIM-hello-sent counter when primary address changes
-    See pim_ifp->pim_ifstat_hello_sent
-
-    RFC 4601: 4.3.1.  Sending Hello Messages
-
-    Thus, if a router needs to send a Join/Prune or Assert message on
-    an interface on which it has not yet sent a Hello message with the
-    currently configured IP address, then it MUST immediately send the
-    relevant Hello message without waiting for the Hello Timer to
-    expire, followed by the Join/Prune or Assert message.
-
-T30 DONE Run interface DR election when primary address changes
-    Reported by Andrew Lunn <andrew.lunn@ascom.ch>
-    See pim_if_dr_election().
-
-T31 If an interface changes one of its secondary IP addresses, a Hello
     message with an updated Address_List option and a non-zero
     HoldTime should be sent immediately.
     See also detect_secondary_address_change
     See also CAVEAT C15.
     See also RFC 4601: 4.3.1.  Sending Hello Messages
 
-T32 FIXED Detection of interface primary address changes may fail when
+T3  Lightweight MLDv2
-    there are multiple addresses.
-    See also CAVEAT C14.
-
-    pim_find_primary_addr() should return interface primary address
-    from connected list. Currently it returns the first address.
-
-    Zebra daemon "show int" is able to keep the primary address as
-    first address.
-
-T33 DONE Implement debug command: test pim receive upcall
-    See also TODO T6: (S,G) Assert state machine.
-
-T34 DONE assert_action_a1
-
-T35 DONE Review macros depending on interface I.
-
-    See also: grep ,I\) pimd/*.c
-
-    For the case (S,G,I) check if I is either
-    1) interface attached to this per-interface S,G state (don't think so)
-    or
-    2) an arbitrary interface (most probably)
-
-    For the arbitrary interface case (2), consider representing
-    interface ifp as its primary address (struct in_addr ifaddr).  The
-    benefit is in_addr does not need to be dereferenced, so it does
-    not demand protection against crashes.
-
-T36 DONE React to zebra daemon link-detect up/down notification.
-    pim_ifp->primary_address is managed by detect_primary_address_change()
-    depending on to ifp->connected (managed by zebra_interface_address_read()).
-
-T37 DONE Review list of variables which may affect pim_upstream.c
-    pim_upstream_evaluate_join_desired().
-    Call pim_upstream_update_join_desired() accordingly.
-
-    See the order of invokation:
-      pim_if_dr_election(ifp);
-      pim_if_update_join_desired(pim_ifp); /* depends on DR */
-      pim_if_update_could_assert(ifp); /* depends on DR */
-      pim_if_update_my_assert_metric(ifp); /* depends on could_assert */
-
-    join_desired depends on:
-      pim_ifp->primary_address
-      pim_ifp->pim_dr_addr
-      ch->ifassert_winner_metric
-      ch->ifassert_winner
-      ch->local_ifmembership 
-      ch->ifjoin_state
-      ch->upstream->rpf.source_nexthop.mrib_metric_preference
-      ch->upstream->rpf.source_nexthop.mrib_route_metric
-      ch->upstream->rpf.source_nexthop.interface
-
-T38 DONE Detect change in AssertTrackingDesired(S,G,I)
-
-    See the order of invokation:
-      dr_election: none
-      update_join_desired: depends on DR
-      update_tracking_desired: depends on DR, join_desired
-
-    AssertTrackingDesired(S,G,I) depends on:
-      pim_ifp->primary_address
-      pim_ifp->pim_dr_addr
-      ch->local_ifmembership
-      ch->ifassert_winner
-      ch->ifjoin_state
-      ch->upstream->rpf.source_nexthop.interface
-      PIM_UPSTREAM_FLAG_TEST_DR_JOIN_DESIRED(ch->upstream->flags)
-
-T39 DONE AssertTrackingDesired: flags is not matching evaluation
-    
-    # show ip pim assert-internal 
-    CA:   CouldAssert
-    ECA:  Evaluate CouldAssert
-    ATD:  AssertTrackingDesired
-    eATD: Evaluate AssertTrackingDesired
-
-    Interface Address         Source          Group           CA  eCA ATD eATD
-    eth0      192.168.1.100   1.1.1.1         239.1.1.1       no  no  no  yes 
-    # 
-
-T40 Lightweight MLDv2
     http://tools.ietf.org/html/draft-ietf-mboned-lightweight-igmpv3-mldv2-05
     http://www.ietf.org/internet-drafts/draft-ietf-mboned-lightweight-igmpv3-mldv2-05.txt
     http://www.ietf.org/html.charters/mboned-charter.html
 
-T41 DONE ssmping support
+T4  Static igmp join fails when loading config at boot time
-
-    See also:
-      http://www.venaas.no/multicast/ssmping/
-      draft-ietf-mboned-ssmping-07
-      http://tools.ietf.org/html/draft-ietf-mboned-ssmping-07
-
-    Example:
-
-    debug ssmpingd
-    
-    conf t
-     ip ssmpingd 1.1.1.1
-    
-    show ip ssmpingd
-
-T42 Static igmp join fails when loading config at boot time
 
     ! Wrong behavior seen at boot time:
     !
@@ -396,7 +42,7 @@ T42 Static igmp join fails when loading config at boot time
     eth0      200.202.112.3     2   2       0       0          0          0
     lo        127.0.0.1         1   1       0       0          0          0
 
-T43 PIM Neighbor Reduction
+T5  PIM Neighbor Reduction
     https://datatracker.ietf.org/doc/draft-wijnands-pim-neighbor-reduction/
     
     "In a transit LAN (no directly connected source or receiver), many
@@ -404,19 +50,19 @@ T43 PIM Neighbor Reduction
     a procedure to reduce the amount of neighbors established over a
     transit LAN."
 
-T44 Single Stream Multicast Fast Reroute (SMFR) Method
+T6  Single Stream Multicast Fast Reroute (SMFR) Method
     https://datatracker.ietf.org/doc/draft-liu-pim-single-stream-multicast-frr/
 
     "This document proposes an IP multicast fast convergence method
     based on differentiating primary and backup PIM join."
 
-T45 RFC5384 - The Join Attribute Format
+T7  RFC5384 - The Join Attribute Format
     "This document describes a modification of the Join message that
     allows a node to associate attributes with a particular tree."
 
-T46 PIM Multi-Topology ID (MT-ID) Join-Attribute
+T8  PIM Multi-Topology ID (MT-ID) Join-Attribute
     http://tools.ietf.org/html/draft-cai-pim-mtid-00
-    Depends on T45.
+    Depends on T7.
 
     "This draft introduces a new type of PIM Join Attribute used to
     encode the identity of the topology PIM uses for RPF."

pimd/mtracebis.c

@@ -0,0 +1,373 @@
+/*
+ * Multicast Traceroute for FRRouting
+ * Copyright (C) 2018  Mladen Sablic
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; see the file COPYING; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#ifdef __linux__
+
+#include "pim_igmp_mtrace.h"
+
+#include "checksum.h"
+#include "mtracebis_routeget.h"
+
+#include <sys/select.h>
+#include <netinet/in.h>
+#include <arpa/inet.h>
+#include <unistd.h>
+#include <sys/types.h>
+#include <sys/time.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <time.h>
+#include <net/if.h>
+#include <unistd.h>
+#include <getopt.h>
+
+#define MTRACEBIS_VERSION "0.1"
+#define MTRACE_TIMEOUT (5)
+
+#define IP_HDR_LEN (sizeof(struct ip))
+#define IP_RA_LEN (4)
+#define MTRACE_BUF_LEN (MTRACE_HDR_SIZE + (MTRACE_MAX_HOPS * MTRACE_RSP_SIZE))
+#define IP_AND_MTRACE_BUF_LEN (IP_HDR_LEN + IP_RA_LEN + MTRACE_BUF_LEN)
+
+static const char *progname;
+static void usage(void)
+{
+	fprintf(stderr, "Usage : %s <multicast source>\n", progname);
+}
+static void version(void)
+{
+	fprintf(stderr, "%s %s\n", progname, MTRACEBIS_VERSION);
+}
+
+static int send_query(int fd, struct in_addr to_addr,
+		      struct igmp_mtrace *mtrace)
+{
+	struct sockaddr_in to;
+	socklen_t tolen;
+	int sent;
+
+	memset(&to, 0, sizeof(to));
+	to.sin_family = AF_INET;
+	to.sin_addr = to_addr;
+	tolen = sizeof(to);
+
+	sent = sendto(fd, (char *)mtrace, sizeof(*mtrace), MSG_DONTWAIT,
+		      (struct sockaddr *)&to, tolen);
+
+	if (sent < 1)
+		return -1;
+	return 0;
+}
+
+static void print_query(struct igmp_mtrace *mtrace)
+{
+	char src_str[INET_ADDRSTRLEN];
+	char dst_str[INET_ADDRSTRLEN];
+	char grp_str[INET_ADDRSTRLEN];
+
+	printf("* Mtrace from %s to %s via group %s\n",
+	       inet_ntop(AF_INET, &mtrace->src_addr, src_str, sizeof(src_str)),
+	       inet_ntop(AF_INET, &mtrace->dst_addr, dst_str, sizeof(dst_str)),
+	       inet_ntop(AF_INET, &mtrace->grp_addr, grp_str, sizeof(grp_str)));
+}
+
+static int recv_response(int fd, long msec, int *hops)
+{
+	int recvd;
+	char mtrace_buf[IP_AND_MTRACE_BUF_LEN];
+	struct ip *ip;
+	struct igmp_mtrace *mtrace;
+	int mtrace_len;
+	int responses;
+	int i;
+	u_short sum;
+
+	recvd = recvfrom(fd, mtrace_buf, IP_AND_MTRACE_BUF_LEN, 0, NULL, 0);
+
+	if (recvd < 1) {
+		fprintf(stderr, "recvfrom error: %s\n", strerror(errno));
+		return -1;
+	}
+
+	if (recvd < (int)sizeof(struct ip)) {
+		fprintf(stderr, "no ip header\n");
+		return -1;
+	}
+
+	ip = (struct ip *)mtrace_buf;
+
+	if (ip->ip_v != 4) {
+		fprintf(stderr, "IP not version 4\n");
+		return -1;
+	}
+
+	sum = ip->ip_sum;
+	ip->ip_sum = 0;
+
+	if (sum != in_cksum(ip, ip->ip_hl * 4))
+		return -1;
+
+	mtrace = (struct igmp_mtrace *)(mtrace_buf + (4 * ip->ip_hl));
+
+	mtrace_len = ntohs(ip->ip_len) - ip->ip_hl * 4;
+
+	if (mtrace_len < (int)MTRACE_HDR_SIZE)
+		return -1;
+
+	sum = mtrace->checksum;
+	mtrace->checksum = 0;
+	if (sum != in_cksum(mtrace, mtrace_len)) {
+		fprintf(stderr, "mtrace checksum wrong\n");
+		return -1;
+	}
+
+	if (mtrace->type != PIM_IGMP_MTRACE_RESPONSE)
+		return -1;
+
+
+	responses = mtrace_len - sizeof(struct igmp_mtrace);
+	responses /= sizeof(struct igmp_mtrace_rsp);
+
+	printf("%ld ms received responses from %d hops.\n", msec, responses);
+
+	if (hops)
+		*hops = responses;
+
+	for (i = 0; i < responses; i++) {
+		struct igmp_mtrace_rsp *rsp = &mtrace->rsp[i];
+
+		if (rsp->fwd_code != 0)
+			printf("-%d fwd. code 0x%2x.\n", i, rsp->fwd_code);
+	}
+
+	return 0;
+}
+
+static int wait_for_response(int fd, int *hops)
+{
+	fd_set readfds;
+	struct timeval timeout;
+	int ret = -1;
+	long msec, rmsec, tmsec;
+
+	FD_ZERO(&readfds);
+	FD_SET(fd, &readfds);
+
+	memset(&timeout, 0, sizeof(timeout));
+
+	timeout.tv_sec = MTRACE_TIMEOUT;
+
+	tmsec = timeout.tv_sec * 1000 + timeout.tv_usec / 1000;
+	do {
+		ret = select(fd + 1, &readfds, NULL, NULL, &timeout);
+		if (ret <= 0)
+			return ret;
+		rmsec = timeout.tv_sec * 1000 + timeout.tv_usec / 1000;
+		msec = tmsec - rmsec;
+	} while (recv_response(fd, msec, hops) != 0);
+
+	return ret;
+}
+
+int main(int argc, char *const argv[])
+{
+	struct in_addr mc_source;
+	struct in_addr iface_addr;
+	struct in_addr gw_addr;
+	struct in_addr mtrace_addr;
+	struct igmp_mtrace mtrace;
+	int hops = 255;
+	int rhops;
+	int maxhops = 255;
+	int perhop = 3;
+	int ifindex;
+	int unicast = 1;
+	int ttl = 64;
+	int fd = -1;
+	int ret = -1;
+	int c;
+	int i, j;
+	char ifname[IF_NAMESIZE];
+	char ip_str[INET_ADDRSTRLEN];
+
+	mtrace_addr.s_addr = inet_addr("224.0.1.32");
+
+	uid_t uid = getuid();
+
+	if (uid != 0) {
+		printf("must run as root\n");
+		exit(EXIT_FAILURE);
+	}
+
+	if (argc <= 0)
+		progname = "mtracebis";
+	else
+		progname = argv[0];
+
+	if (argc != 2) {
+		usage();
+		exit(EXIT_FAILURE);
+	}
+
+	while (1) {
+		static struct option long_options[] = {
+			{"help", no_argument, 0, 'h'},
+			{"version", no_argument, 0, 'v'},
+			{0, 0, 0, 0} };
+		int option_index = 0;
+
+		c = getopt_long(argc, argv, "vh", long_options, &option_index);
+
+		if (c == -1)
+			break;
+
+		switch (c) {
+		case 'h':
+			usage();
+			exit(0);
+		case 'v':
+			version();
+			exit(0);
+		default:
+			usage();
+			exit(EXIT_FAILURE);
+		}
+	}
+	if (inet_pton(AF_INET, argv[1], &mc_source) != 1) {
+		usage();
+		fprintf(stderr, "%s: %s not a valid IPv4 address\n", argv[0],
+			argv[1]);
+		exit(EXIT_FAILURE);
+	}
+
+	ifindex = routeget(mc_source, &iface_addr, &gw_addr);
+	if (ifindex < 0) {
+		fprintf(stderr, "%s: failed to get route to source %s\n",
+			argv[0], argv[1]);
+		exit(EXIT_FAILURE);
+	}
+
+	if (if_indextoname(ifindex, ifname) == NULL) {
+		fprintf(stderr, "%s: if_indextoname error: %s\n", argv[0],
+			strerror(errno));
+		exit(EXIT_FAILURE);
+	}
+
+	/* zero mtrace struct */
+	memset((char *)&mtrace, 0, sizeof(mtrace));
+
+	/* set up query */
+	mtrace.type = PIM_IGMP_MTRACE_QUERY_REQUEST;
+	mtrace.hops = hops;
+	mtrace.checksum = 0;
+	mtrace.grp_addr.s_addr = 0;
+	mtrace.src_addr = mc_source;
+	mtrace.dst_addr = iface_addr;
+	mtrace.rsp_addr = unicast ? iface_addr : mtrace_addr;
+	mtrace.rsp_ttl = ttl;
+	mtrace.qry_id = 0xffffff & time(NULL);
+
+	mtrace.checksum = in_cksum(&mtrace, sizeof(mtrace));
+
+	fd = socket(AF_INET, SOCK_RAW, IPPROTO_IGMP);
+
+	if (fd < 1) {
+		fprintf(stderr, "%s: socket error: %s\n", argv[0],
+			strerror(errno));
+		exit(EXIT_FAILURE);
+	}
+
+	ret = setsockopt(fd, SOL_SOCKET, SO_BINDTODEVICE, ifname,
+			 strlen(ifname));
+
+	if (ret < 0) {
+		fprintf(stderr, "%s: setsockopt error: %s\n", argv[0],
+			strerror(errno));
+		ret = EXIT_FAILURE;
+		goto close_fd;
+	}
+
+	print_query(&mtrace);
+	if (send_query(fd, gw_addr, &mtrace) < 0) {
+		fprintf(stderr, "%s: sendto error: %s\n", argv[0],
+			strerror(errno));
+		ret = EXIT_FAILURE;
+		goto close_fd;
+	}
+	printf("Querying full reverse path...\n");
+	ret = wait_for_response(fd, NULL);
+	if (ret > 0) {
+		ret = 0;
+		goto close_fd;
+	}
+	if (ret < 0) {
+		fprintf(stderr, "%s: select error: %s\n", argv[0],
+			strerror(errno));
+		ret = EXIT_FAILURE;
+		goto close_fd;
+	}
+	printf(" * ");
+	printf("switching to hop-by-hop:\n");
+	printf("%3d  ? (%s)\n", 0,
+	       inet_ntop(AF_INET, &mtrace.dst_addr, ip_str, sizeof(ip_str)));
+	for (i = 1; i < maxhops; i++) {
+		printf("%3d ", -i);
+		mtrace.hops = i;
+		for (j = 0; j < perhop; j++) {
+			mtrace.qry_id++;
+			mtrace.checksum = 0;
+			mtrace.checksum = in_cksum(&mtrace, sizeof(mtrace));
+			if (send_query(fd, gw_addr, &mtrace) < 0) {
+				fprintf(stderr, "%s: sendto error: %s\n",
+					argv[0], strerror(errno));
+				ret = EXIT_FAILURE;
+				goto close_fd;
+			}
+			ret = wait_for_response(fd, &rhops);
+			if (ret > 0) {
+				if (i > rhops) {
+					ret = 0;
+					goto close_fd;
+				}
+				break;
+			}
+			printf(" *");
+		}
+		if (ret <= 0)
+			printf("\n");
+	}
+	ret = 0;
+close_fd:
+	close(fd);
+	exit(ret);
+}
+
+#else /* __linux__ */
+
+#include <stdio.h>
+#include <stdlib.h>
+
+int main(int argc, char *argv[])
+{
+	printf("%s implemented only for GNU/Linux\n", argv[0]);
+	exit(0);
+}
+
+#endif /* __linux__ */

pimd/mtracebis_netlink.c

@@ -0,0 +1,700 @@
+/*
+ * libnetlink.c	RTnetlink service routines.
+ *
+ *		This program is free software; you can redistribute it and/or
+ *		modify it under the terms of the GNU General Public License
+ *		as published by the Free Software Foundation; either version
+ *		2 of the License, or (at your option) any later version.
+ *
+ * Authors:	Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
+ *
+ */
+
+#ifdef __linux__
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <unistd.h>
+#include <syslog.h>
+#include <fcntl.h>
+#include <net/if_arp.h>
+#include <sys/socket.h>
+#include <netinet/in.h>
+#include <string.h>
+#include <errno.h>
+#include <time.h>
+#include <sys/uio.h>
+#include <assert.h>
+
+#include "mtracebis_netlink.h"
+
+int rcvbuf = 1024 * 1024;
+
+void rtnl_close(struct rtnl_handle *rth)
+{
+	if (rth->fd >= 0) {
+		close(rth->fd);
+		rth->fd = -1;
+	}
+}
+
+int rtnl_open_byproto(struct rtnl_handle *rth, unsigned subscriptions,
+		      int protocol)
+{
+	socklen_t addr_len;
+	int sndbuf = 32768;
+
+	memset(rth, 0, sizeof(*rth));
+
+	rth->fd = socket(AF_NETLINK, SOCK_RAW, protocol);
+	if (rth->fd < 0) {
+		perror("Cannot open netlink socket");
+		return -1;
+	}
+
+	if (setsockopt(rth->fd,SOL_SOCKET,SO_SNDBUF,&sndbuf,sizeof(sndbuf)) < 0) {
+		perror("SO_SNDBUF");
+		return -1;
+	}
+
+	if (setsockopt(rth->fd,SOL_SOCKET,SO_RCVBUF,&rcvbuf,sizeof(rcvbuf)) < 0) {
+		perror("SO_RCVBUF");
+		return -1;
+	}
+
+	memset(&rth->local, 0, sizeof(rth->local));
+	rth->local.nl_family = AF_NETLINK;
+	rth->local.nl_groups = subscriptions;
+
+	if (bind(rth->fd, (struct sockaddr*)&rth->local, sizeof(rth->local)) < 0) {
+		perror("Cannot bind netlink socket");
+		return -1;
+	}
+	addr_len = sizeof(rth->local);
+	if (getsockname(rth->fd, (struct sockaddr*)&rth->local, &addr_len) < 0) {
+		perror("Cannot getsockname");
+		return -1;
+	}
+	if (addr_len != sizeof(rth->local)) {
+		fprintf(stderr, "Wrong address length %d\n", addr_len);
+		return -1;
+	}
+	if (rth->local.nl_family != AF_NETLINK) {
+		fprintf(stderr, "Wrong address family %d\n", rth->local.nl_family);
+		return -1;
+	}
+	rth->seq = time(NULL);
+	return 0;
+}
+
+int rtnl_open(struct rtnl_handle *rth, unsigned subscriptions)
+{
+	return rtnl_open_byproto(rth, subscriptions, NETLINK_ROUTE);
+}
+
+int rtnl_wilddump_request(struct rtnl_handle *rth, int family, int type)
+{
+	struct {
+		struct nlmsghdr nlh;
+		struct rtgenmsg g;
+	} req;
+
+	memset(&req, 0, sizeof(req));
+	req.nlh.nlmsg_len = sizeof(req);
+	req.nlh.nlmsg_type = type;
+	req.nlh.nlmsg_flags = NLM_F_ROOT|NLM_F_MATCH|NLM_F_REQUEST;
+	req.nlh.nlmsg_pid = 0;
+	req.nlh.nlmsg_seq = rth->dump = ++rth->seq;
+	req.g.rtgen_family = family;
+
+	return send(rth->fd, (void*)&req, sizeof(req), 0);
+}
+
+int rtnl_send(struct rtnl_handle *rth, const char *buf, int len)
+{
+	return send(rth->fd, buf, len, 0);
+}
+
+int rtnl_send_check(struct rtnl_handle *rth, const char *buf, int len)
+{
+	struct nlmsghdr *h;
+	int status;
+	char resp[1024];
+
+	status = send(rth->fd, buf, len, 0);
+	if (status < 0)
+		return status;
+
+	/* Check for immediate errors */
+	status = recv(rth->fd, resp, sizeof(resp), MSG_DONTWAIT|MSG_PEEK);
+	if (status < 0) {
+		if (errno == EAGAIN)
+			return 0;
+		return -1;
+	}
+
+	for (h = (struct nlmsghdr *)resp; NLMSG_OK(h, (uint32_t)status);
+	     h = NLMSG_NEXT(h, status)) {
+		if (h->nlmsg_type == NLMSG_ERROR) {
+			struct nlmsgerr *err = (struct nlmsgerr*)NLMSG_DATA(h);
+			if (h->nlmsg_len < NLMSG_LENGTH(sizeof(struct nlmsgerr)))
+				fprintf(stderr, "ERROR truncated\n");
+			else
+				errno = -err->error;
+			return -1;
+		}
+	}
+
+	return 0;
+}
+
+int rtnl_dump_request(struct rtnl_handle *rth, int type, void *req, int len)
+{
+	struct nlmsghdr nlh;
+	struct sockaddr_nl nladdr;
+	struct iovec iov[2] = {
+		{ .iov_base = &nlh, .iov_len = sizeof(nlh) },
+		{ .iov_base = req, .iov_len = len }
+	};
+	struct msghdr msg = {
+		.msg_name = &nladdr,
+		.msg_namelen = 	sizeof(nladdr),
+		.msg_iov = iov,
+		.msg_iovlen = 2,
+	};
+
+	memset(&nladdr, 0, sizeof(nladdr));
+	nladdr.nl_family = AF_NETLINK;
+
+	nlh.nlmsg_len = NLMSG_LENGTH(len);
+	nlh.nlmsg_type = type;
+	nlh.nlmsg_flags = NLM_F_ROOT|NLM_F_MATCH|NLM_F_REQUEST;
+	nlh.nlmsg_pid = 0;
+	nlh.nlmsg_seq = rth->dump = ++rth->seq;
+
+	return sendmsg(rth->fd, &msg, 0);
+}
+
+int rtnl_dump_filter_l(struct rtnl_handle *rth,
+		       const struct rtnl_dump_filter_arg *arg)
+{
+	struct sockaddr_nl nladdr;
+	struct iovec iov;
+	struct msghdr msg = {
+		.msg_name = &nladdr,
+		.msg_namelen = sizeof(nladdr),
+		.msg_iov = &iov,
+		.msg_iovlen = 1,
+	};
+	char buf[16384];
+
+	iov.iov_base = buf;
+	while (1) {
+		int status;
+		const struct rtnl_dump_filter_arg *a;
+		int found_done = 0;
+		int msglen = 0;
+
+		iov.iov_len = sizeof(buf);
+		status = recvmsg(rth->fd, &msg, 0);
+
+		if (status < 0) {
+			if (errno == EINTR || errno == EAGAIN)
+				continue;
+			fprintf(stderr, "netlink receive error %s (%d)\n",
+				strerror(errno), errno);
+			return -1;
+		}
+
+		if (status == 0) {
+			fprintf(stderr, "EOF on netlink\n");
+			return -1;
+		}
+
+		for (a = arg; a->filter; a++) {
+			struct nlmsghdr *h = (struct nlmsghdr*)buf;
+			msglen = status;
+
+			while (NLMSG_OK(h, (uint32_t)msglen)) {
+				int err;
+
+				if (nladdr.nl_pid != 0 ||
+				    h->nlmsg_pid != rth->local.nl_pid ||
+				    h->nlmsg_seq != rth->dump) {
+					if (a->junk) {
+						err = a->junk(&nladdr, h,
+							      a->arg2);
+						if (err < 0)
+							return err;
+					}
+					goto skip_it;
+				}
+
+				if (h->nlmsg_type == NLMSG_DONE) {
+					found_done = 1;
+					break; /* process next filter */
+				}
+				if (h->nlmsg_type == NLMSG_ERROR) {
+					struct nlmsgerr *err = (struct nlmsgerr*)NLMSG_DATA(h);
+					if (h->nlmsg_len < NLMSG_LENGTH(sizeof(struct nlmsgerr))) {
+						fprintf(stderr,
+							"ERROR truncated\n");
+					} else {
+						errno = -err->error;
+						perror("RTNETLINK answers");
+					}
+					return -1;
+				}
+				err = a->filter(&nladdr, h, a->arg1);
+				if (err < 0)
+					return err;
+
+skip_it:
+				h = NLMSG_NEXT(h, msglen);
+			}
+		}
+
+		if (found_done)
+			return 0;
+
+		if (msg.msg_flags & MSG_TRUNC) {
+			fprintf(stderr, "Message truncated\n");
+			continue;
+		}
+		if (msglen) {
+			fprintf(stderr, "!!!Remnant of size %d\n", msglen);
+			exit(1);
+		}
+	}
+}
+
+int rtnl_dump_filter(struct rtnl_handle *rth,
+		     rtnl_filter_t filter,
+		     void *arg1,
+		     rtnl_filter_t junk,
+		     void *arg2)
+{
+	const struct rtnl_dump_filter_arg a[2] = {
+		{ .filter = filter, .arg1 = arg1, .junk = junk, .arg2 = arg2 },
+		{ .filter = NULL,   .arg1 = NULL, .junk = NULL, .arg2 = NULL }
+	};
+
+	return rtnl_dump_filter_l(rth, a);
+}
+
+int rtnl_talk(struct rtnl_handle *rtnl, struct nlmsghdr *n, pid_t peer,
+	      unsigned groups, struct nlmsghdr *answer,
+	      rtnl_filter_t junk,
+	      void *jarg)
+{
+	int status;
+	unsigned seq;
+	struct nlmsghdr *h;
+	struct sockaddr_nl nladdr;
+	struct iovec iov = {
+		.iov_base = (void*) n,
+		.iov_len = n->nlmsg_len
+	};
+	struct msghdr msg = {
+		.msg_name = &nladdr,
+		.msg_namelen = sizeof(nladdr),
+		.msg_iov = &iov,
+		.msg_iovlen = 1,
+	};
+	char   buf[16384];
+
+	memset(&nladdr, 0, sizeof(nladdr));
+	nladdr.nl_family = AF_NETLINK;
+	nladdr.nl_pid = peer;
+	nladdr.nl_groups = groups;
+
+	n->nlmsg_seq = seq = ++rtnl->seq;
+
+	if (answer == NULL)
+		n->nlmsg_flags |= NLM_F_ACK;
+
+	status = sendmsg(rtnl->fd, &msg, 0);
+
+	if (status < 0) {
+		perror("Cannot talk to rtnetlink");
+		return -1;
+	}
+
+	memset(buf,0,sizeof(buf));
+
+	iov.iov_base = buf;
+
+	while (1) {
+		iov.iov_len = sizeof(buf);
+		status = recvmsg(rtnl->fd, &msg, 0);
+
+		if (status < 0) {
+			if (errno == EINTR || errno == EAGAIN)
+				continue;
+			fprintf(stderr, "netlink receive error %s (%d)\n",
+				strerror(errno), errno);
+			return -1;
+		}
+		if (status == 0) {
+			fprintf(stderr, "EOF on netlink\n");
+			return -1;
+		}
+		if (msg.msg_namelen != sizeof(nladdr)) {
+			fprintf(stderr, "sender address length == %d\n", msg.msg_namelen);
+			exit(1);
+		}
+		for (h = (struct nlmsghdr*)buf; status >= (int)sizeof(*h); ) {
+			int err;
+			int len = h->nlmsg_len;
+			int l = len - sizeof(*h);
+
+			if (l<0 || len>status) {
+				if (msg.msg_flags & MSG_TRUNC) {
+					fprintf(stderr, "Truncated message\n");
+					return -1;
+				}
+				fprintf(stderr, "!!!malformed message: len=%d\n", len);
+				exit(1);
+			}
+
+			if ((int)nladdr.nl_pid != peer ||
+			    h->nlmsg_pid != rtnl->local.nl_pid ||
+			    h->nlmsg_seq != seq) {
+				if (junk) {
+					err = junk(&nladdr, h, jarg);
+					if (err < 0)
+						return err;
+				}
+				/* Don't forget to skip that message. */
+				status -= NLMSG_ALIGN(len);
+				h = (struct nlmsghdr*)((char*)h + NLMSG_ALIGN(len));
+				continue;
+			}
+
+			if (h->nlmsg_type == NLMSG_ERROR) {
+				struct nlmsgerr *err = (struct nlmsgerr*)NLMSG_DATA(h);
+				if (l < (int)sizeof(struct nlmsgerr)) {
+					fprintf(stderr, "ERROR truncated\n");
+				} else {
+					errno = -err->error;
+					if (errno == 0) {
+						if (answer)
+							memcpy(answer, h, h->nlmsg_len);
+						return 0;
+					}
+					perror("RTNETLINK answers");
+				}
+				return -1;
+			}
+			if (answer) {
+				memcpy(answer, h, h->nlmsg_len);
+				return 0;
+			}
+
+			fprintf(stderr, "Unexpected reply!!!\n");
+
+			status -= NLMSG_ALIGN(len);
+			h = (struct nlmsghdr*)((char*)h + NLMSG_ALIGN(len));
+		}
+		if (msg.msg_flags & MSG_TRUNC) {
+			fprintf(stderr, "Message truncated\n");
+			continue;
+		}
+		if (status) {
+			fprintf(stderr, "!!!Remnant of size %d\n", status);
+			exit(1);
+		}
+	}
+}
+
+int rtnl_listen(struct rtnl_handle *rtnl,
+		rtnl_filter_t handler,
+		void *jarg)
+{
+	int status;
+	struct nlmsghdr *h;
+	struct sockaddr_nl nladdr;
+	struct iovec iov;
+	struct msghdr msg = {
+		.msg_name = &nladdr,
+		.msg_namelen = sizeof(nladdr),
+		.msg_iov = &iov,
+		.msg_iovlen = 1,
+	};
+	char   buf[8192];
+
+	memset(&nladdr, 0, sizeof(nladdr));
+	nladdr.nl_family = AF_NETLINK;
+	nladdr.nl_pid = 0;
+	nladdr.nl_groups = 0;
+
+	iov.iov_base = buf;
+	while (1) {
+		iov.iov_len = sizeof(buf);
+		status = recvmsg(rtnl->fd, &msg, 0);
+
+		if (status < 0) {
+			if (errno == EINTR || errno == EAGAIN)
+				continue;
+			fprintf(stderr, "netlink receive error %s (%d)\n",
+				strerror(errno), errno);
+			if (errno == ENOBUFS)
+				continue;
+			return -1;
+		}
+		if (status == 0) {
+			fprintf(stderr, "EOF on netlink\n");
+			return -1;
+		}
+		if (msg.msg_namelen != sizeof(nladdr)) {
+			fprintf(stderr, "Sender address length == %d\n", msg.msg_namelen);
+			exit(1);
+		}
+		for (h =(struct nlmsghdr*)buf; status >= (int)sizeof(*h); ) {
+			int err;
+			int len = h->nlmsg_len;
+			int l = len - sizeof(*h);
+
+			if (l<0 || len>status) {
+				if (msg.msg_flags & MSG_TRUNC) {
+					fprintf(stderr, "Truncated message\n");
+					return -1;
+				}
+				fprintf(stderr, "!!!malformed message: len=%d\n", len);
+				exit(1);
+			}
+
+			err = handler(&nladdr, h, jarg);
+			if (err < 0)
+				return err;
+
+			status -= NLMSG_ALIGN(len);
+			h = (struct nlmsghdr*)((char*)h + NLMSG_ALIGN(len));
+		}
+		if (msg.msg_flags & MSG_TRUNC) {
+			fprintf(stderr, "Message truncated\n");
+			continue;
+		}
+		if (status) {
+			fprintf(stderr, "!!!Remnant of size %d\n", status);
+			exit(1);
+		}
+	}
+}
+
+int rtnl_from_file(FILE *rtnl, rtnl_filter_t handler,
+		   void *jarg)
+{
+	int status;
+	struct sockaddr_nl nladdr;
+	char   buf[8192];
+	struct nlmsghdr *h = (void*)buf;
+
+	memset(&nladdr, 0, sizeof(nladdr));
+	nladdr.nl_family = AF_NETLINK;
+	nladdr.nl_pid = 0;
+	nladdr.nl_groups = 0;
+
+	while (1) {
+		int err, len;
+		int l;
+
+		status = fread(&buf, 1, sizeof(*h), rtnl);
+
+		if (status < 0) {
+			if (errno == EINTR)
+				continue;
+			perror("rtnl_from_file: fread");
+			return -1;
+		}
+		if (status == 0)
+			return 0;
+
+		len = h->nlmsg_len;
+		l = len - sizeof(*h);
+
+		if (l<0 || len>(int)sizeof(buf)) {
+			fprintf(stderr, "!!!malformed message: len=%d @%lu\n",
+				len, ftell(rtnl));
+			return -1;
+		}
+
+		status = fread(NLMSG_DATA(h), 1, NLMSG_ALIGN(l), rtnl);
+
+		if (status < 0) {
+			perror("rtnl_from_file: fread");
+			return -1;
+		}
+		if (status < l) {
+			fprintf(stderr, "rtnl-from_file: truncated message\n");
+			return -1;
+		}
+
+		err = handler(&nladdr, h, jarg);
+		if (err < 0)
+			return err;
+	}
+}
+
+int addattr32(struct nlmsghdr *n, int maxlen, int type, __u32 data)
+{
+	int len = RTA_LENGTH(4);
+	struct rtattr *rta;
+	if ((int)(NLMSG_ALIGN(n->nlmsg_len) + len) > maxlen) {
+		fprintf(stderr,"addattr32: Error! max allowed bound %d exceeded\n",maxlen);
+		return -1;
+	}
+	rta = NLMSG_TAIL(n);
+	rta->rta_type = type;
+	rta->rta_len = len;
+	memcpy(RTA_DATA(rta), &data, 4);
+	n->nlmsg_len = NLMSG_ALIGN(n->nlmsg_len) + len;
+	return 0;
+}
+
+int addattr_l(struct nlmsghdr *n, int maxlen, int type, const void *data,
+	      int alen)
+{
+	int len = RTA_LENGTH(alen);
+	struct rtattr *rta;
+
+	if ((int)(NLMSG_ALIGN(n->nlmsg_len) + RTA_ALIGN(len)) > maxlen) {
+		fprintf(stderr, "addattr_l ERROR: message exceeded bound of %d\n",maxlen);
+		return -1;
+	}
+	rta = NLMSG_TAIL(n);
+	rta->rta_type = type;
+	rta->rta_len = len;
+
+	if (data)
+		memcpy(RTA_DATA(rta), data, alen);
+	else
+		assert(alen == 0);
+
+	n->nlmsg_len = NLMSG_ALIGN(n->nlmsg_len) + RTA_ALIGN(len);
+	return 0;
+}
+
+int addraw_l(struct nlmsghdr *n, int maxlen, const void *data, int len)
+{
+	if ((int)(NLMSG_ALIGN(n->nlmsg_len) + NLMSG_ALIGN(len)) > maxlen) {
+		fprintf(stderr, "addraw_l ERROR: message exceeded bound of %d\n",maxlen);
+		return -1;
+	}
+
+	memcpy(NLMSG_TAIL(n), data, len);
+	memset((uint8_t *) NLMSG_TAIL(n) + len, 0, NLMSG_ALIGN(len) - len);
+	n->nlmsg_len = NLMSG_ALIGN(n->nlmsg_len) + NLMSG_ALIGN(len);
+	return 0;
+}
+
+struct rtattr *addattr_nest(struct nlmsghdr *n, int maxlen, int type)
+{
+	struct rtattr *nest = NLMSG_TAIL(n);
+
+	addattr_l(n, maxlen, type, NULL, 0);
+	return nest;
+}
+
+int addattr_nest_end(struct nlmsghdr *n, struct rtattr *nest)
+{
+	nest->rta_len = (uint8_t *)NLMSG_TAIL(n) - (uint8_t *)nest;
+	return n->nlmsg_len;
+}
+
+struct rtattr *addattr_nest_compat(struct nlmsghdr *n, int maxlen, int type,
+				   const void *data, int len)
+{
+	struct rtattr *start = NLMSG_TAIL(n);
+
+	addattr_l(n, maxlen, type, data, len);
+	addattr_nest(n, maxlen, type);
+	return start;
+}
+
+int addattr_nest_compat_end(struct nlmsghdr *n, struct rtattr *start)
+{
+	struct rtattr *nest = start + NLMSG_ALIGN(start->rta_len);
+
+	start->rta_len = (uint8_t *)NLMSG_TAIL(n) - (uint8_t *)start;
+	addattr_nest_end(n, nest);
+	return n->nlmsg_len;
+}
+
+int rta_addattr32(struct rtattr *rta, int maxlen, int type, __u32 data)
+{
+	int len = RTA_LENGTH(4);
+	struct rtattr *subrta;
+
+	if ((int)(RTA_ALIGN(rta->rta_len) + len) > maxlen) {
+		fprintf(stderr,"rta_addattr32: Error! max allowed bound %d exceeded\n",maxlen);
+		return -1;
+	}
+	subrta = (struct rtattr*)(((char*)rta) + RTA_ALIGN(rta->rta_len));
+	subrta->rta_type = type;
+	subrta->rta_len = len;
+	memcpy(RTA_DATA(subrta), &data, 4);
+	rta->rta_len = NLMSG_ALIGN(rta->rta_len) + len;
+	return 0;
+}
+
+int rta_addattr_l(struct rtattr *rta, int maxlen, int type,
+		  const void *data, int alen)
+{
+	struct rtattr *subrta;
+	int len = RTA_LENGTH(alen);
+
+	if ((int)(RTA_ALIGN(rta->rta_len) + RTA_ALIGN(len)) > maxlen) {
+		fprintf(stderr,"rta_addattr_l: Error! max allowed bound %d exceeded\n",maxlen);
+		return -1;
+	}
+	subrta = (struct rtattr*)(((char*)rta) + RTA_ALIGN(rta->rta_len));
+	subrta->rta_type = type;
+	subrta->rta_len = len;
+	memcpy(RTA_DATA(subrta), data, alen);
+	rta->rta_len = NLMSG_ALIGN(rta->rta_len) + RTA_ALIGN(len);
+	return 0;
+}
+
+int parse_rtattr(struct rtattr *tb[], int max, struct rtattr *rta, int len)
+{
+	memset(tb, 0, sizeof(struct rtattr *) * (max + 1));
+	while (RTA_OK(rta, len)) {
+		if ((rta->rta_type <= max) && (!tb[rta->rta_type]))
+			tb[rta->rta_type] = rta;
+		rta = RTA_NEXT(rta,len);
+	}
+	if (len)
+		fprintf(stderr, "!!!Deficit %d, rta_len=%d\n", len, rta->rta_len);
+	return 0;
+}
+
+int parse_rtattr_byindex(struct rtattr *tb[], int max, struct rtattr *rta, int len)
+{
+	int i = 0;
+
+	memset(tb, 0, sizeof(struct rtattr *) * max);
+	while (RTA_OK(rta, len)) {
+		if (rta->rta_type <= max && i < max)
+			tb[i++] = rta;
+		rta = RTA_NEXT(rta,len);
+	}
+	if (len)
+		fprintf(stderr, "!!!Deficit %d, rta_len=%d\n", len, rta->rta_len);
+	return i;
+}
+
+int __parse_rtattr_nested_compat(struct rtattr *tb[], int max, struct rtattr *rta,
+			         int len)
+{
+	if ((int)RTA_PAYLOAD(rta) < len)
+		return -1;
+	if (RTA_PAYLOAD(rta) >= RTA_ALIGN(len) + sizeof(struct rtattr)) {
+		rta = (struct rtattr *)(uint8_t *)RTA_DATA(rta)+RTA_ALIGN(len);
+		return parse_rtattr_nested(tb, max, rta);
+	}
+	memset(tb, 0, sizeof(struct rtattr *) * (max + 1));
+	return 0;
+}
+
+#endif /* __linux__ */

pimd/mtracebis_netlink.h

@@ -0,0 +1,130 @@
+/*
+ * libnetlink.c	RTnetlink service routines.
+ *
+ *		This program is free software; you can redistribute it and/or
+ *		modify it under the terms of the GNU General Public License
+ *		as published by the Free Software Foundation; either version
+ *		2 of the License, or (at your option) any later version.
+ *
+ * Authors:	Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
+ *
+ */
+
+#ifdef __linux__
+
+#ifndef __LIBNETLINK_H__
+#define __LIBNETLINK_H__ 1
+
+#include <asm/types.h>
+#include <linux/netlink.h>
+#include <linux/rtnetlink.h>
+#include <linux/if_link.h>
+#include <linux/if_addr.h>
+#include <linux/neighbour.h>
+
+struct rtnl_handle
+{
+	int			fd;
+	struct sockaddr_nl	local;
+	struct sockaddr_nl	peer;
+	__u32			seq;
+	__u32			dump;
+};
+
+extern int rcvbuf;
+
+extern int rtnl_open(struct rtnl_handle *rth, unsigned subscriptions);
+extern int rtnl_open_byproto(struct rtnl_handle *rth, unsigned subscriptions, int protocol);
+extern void rtnl_close(struct rtnl_handle *rth);
+extern int rtnl_wilddump_request(struct rtnl_handle *rth, int fam, int type);
+extern int rtnl_dump_request(struct rtnl_handle *rth, int type, void *req, int len);
+
+typedef int (*rtnl_filter_t)(const struct sockaddr_nl *,
+			     struct nlmsghdr *n, void *);
+
+struct rtnl_dump_filter_arg
+{
+	rtnl_filter_t filter;
+	void *arg1;
+	rtnl_filter_t junk;
+	void *arg2;
+};
+
+extern int rtnl_dump_filter_l(struct rtnl_handle *rth,
+			      const struct rtnl_dump_filter_arg *arg);
+extern int rtnl_dump_filter(struct rtnl_handle *rth, rtnl_filter_t filter,
+			    void *arg1,
+			    rtnl_filter_t junk,
+			    void *arg2);
+
+extern int rtnl_talk(struct rtnl_handle *rtnl, struct nlmsghdr *n, pid_t peer,
+		     unsigned groups, struct nlmsghdr *answer,
+		     rtnl_filter_t junk,
+		     void *jarg);
+extern int rtnl_send(struct rtnl_handle *rth, const char *buf, int);
+extern int rtnl_send_check(struct rtnl_handle *rth, const char *buf, int);
+
+extern int addattr32(struct nlmsghdr *n, int maxlen, int type, __u32 data);
+extern int addattr_l(struct nlmsghdr *n, int maxlen, int type, const void *data, int alen);
+extern int addraw_l(struct nlmsghdr *n, int maxlen, const void *data, int len);
+extern struct rtattr *addattr_nest(struct nlmsghdr *n, int maxlen, int type);
+extern int addattr_nest_end(struct nlmsghdr *n, struct rtattr *nest);
+extern struct rtattr *addattr_nest_compat(struct nlmsghdr *n, int maxlen, int type, const void *data, int len);
+extern int addattr_nest_compat_end(struct nlmsghdr *n, struct rtattr *nest);
+extern int rta_addattr32(struct rtattr *rta, int maxlen, int type, __u32 data);
+extern int rta_addattr_l(struct rtattr *rta, int maxlen, int type, const void *data, int alen);
+
+extern int parse_rtattr(struct rtattr *tb[], int max, struct rtattr *rta, int len);
+extern int parse_rtattr_byindex(struct rtattr *tb[], int max, struct rtattr *rta, int len);
+extern int __parse_rtattr_nested_compat(struct rtattr *tb[], int max, struct rtattr *rta, int len);
+
+#define parse_rtattr_nested(tb, max, rta) \
+	(parse_rtattr((tb), (max), RTA_DATA(rta), RTA_PAYLOAD(rta)))
+
+#define parse_rtattr_nested_compat(tb, max, rta, data, len) \
+({	data = RTA_PAYLOAD(rta) >= len ? RTA_DATA(rta) : NULL; \
+	__parse_rtattr_nested_compat(tb, max, rta, len); })
+
+extern int rtnl_listen(struct rtnl_handle *, rtnl_filter_t handler,
+		       void *jarg);
+extern int rtnl_from_file(FILE *, rtnl_filter_t handler,
+		       void *jarg);
+
+#define NLMSG_TAIL(nmsg) \
+	((struct rtattr *) (((uint8_t *) (nmsg)) + NLMSG_ALIGN((nmsg)->nlmsg_len)))
+
+#ifndef IFA_RTA
+#define IFA_RTA(r) \
+	((struct rtattr*)(((char*)(r)) + NLMSG_ALIGN(sizeof(struct ifaddrmsg))))
+#endif
+#ifndef IFA_PAYLOAD
+#define IFA_PAYLOAD(n)	NLMSG_PAYLOAD(n,sizeof(struct ifaddrmsg))
+#endif
+
+#ifndef IFLA_RTA
+#define IFLA_RTA(r) \
+	((struct rtattr*)(((char*)(r)) + NLMSG_ALIGN(sizeof(struct ifinfomsg))))
+#endif
+#ifndef IFLA_PAYLOAD
+#define IFLA_PAYLOAD(n)	NLMSG_PAYLOAD(n,sizeof(struct ifinfomsg))
+#endif
+
+#ifndef NDA_RTA
+#define NDA_RTA(r) \
+	((struct rtattr*)(((char*)(r)) + NLMSG_ALIGN(sizeof(struct ndmsg))))
+#endif
+#ifndef NDA_PAYLOAD
+#define NDA_PAYLOAD(n)	NLMSG_PAYLOAD(n,sizeof(struct ndmsg))
+#endif
+
+#ifndef NDTA_RTA
+#define NDTA_RTA(r) \
+	((struct rtattr*)(((char*)(r)) + NLMSG_ALIGN(sizeof(struct ndtmsg))))
+#endif
+#ifndef NDTA_PAYLOAD
+#define NDTA_PAYLOAD(n) NLMSG_PAYLOAD(n,sizeof(struct ndtmsg))
+#endif
+
+#endif /* __LIBNETLINK_H__ */
+
+#endif /* __linux__ */

pimd/mtracebis_routeget.c

@@ -0,0 +1,98 @@
+/*
+ * Multicast Traceroute for FRRouting
+ * Copyright (C) 2018  Mladen Sablic
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; see the file COPYING; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#ifdef __linux__
+
+#include <asm/types.h>
+#include <linux/netlink.h>
+#include <linux/rtnetlink.h>
+#include <sys/types.h>
+#include <stdio.h>
+#include <arpa/inet.h>
+#include <string.h>
+
+#include "mtracebis_netlink.h"
+#include "mtracebis_routeget.h"
+
+static int find_dst(struct nlmsghdr *n, struct in_addr *src, struct in_addr *gw)
+{
+	struct rtmsg *r = NLMSG_DATA(n);
+	int len = n->nlmsg_len;
+	struct rtattr *tb[RTA_MAX + 1];
+
+	len -= NLMSG_LENGTH(sizeof(*r));
+	if (len < 0) {
+		fprintf(stderr, "BUG: wrong nlmsg len %d\n", len);
+		return -1;
+	}
+
+	parse_rtattr(tb, RTA_MAX, RTM_RTA(r), len);
+	if (tb[RTA_PREFSRC])
+		src->s_addr = *(uint32_t *)RTA_DATA(tb[RTA_PREFSRC]);
+	if (tb[RTA_GATEWAY])
+		gw->s_addr = *(uint32_t *)RTA_DATA(tb[RTA_GATEWAY]);
+	if (tb[RTA_OIF])
+		return *(int *)RTA_DATA(tb[RTA_OIF]);
+	return 0;
+}
+
+int routeget(struct in_addr dst, struct in_addr *src, struct in_addr *gw)
+{
+	struct {
+		struct nlmsghdr n;
+		struct rtmsg r;
+		char buf[1024];
+	} req;
+	int ret;
+	struct rtnl_handle rth = {.fd = -1};
+
+	memset(&req, 0, sizeof(req));
+
+	req.n.nlmsg_len = NLMSG_LENGTH(sizeof(struct rtmsg));
+	req.n.nlmsg_flags = NLM_F_REQUEST;
+	req.n.nlmsg_type = RTM_GETROUTE;
+	req.r.rtm_family = AF_INET;
+	req.r.rtm_table = 0;
+	req.r.rtm_protocol = 0;
+	req.r.rtm_scope = 0;
+	req.r.rtm_type = 0;
+	req.r.rtm_src_len = 0;
+	req.r.rtm_dst_len = 0;
+	req.r.rtm_tos = 0;
+
+	addattr_l(&req.n, sizeof(req), RTA_DST, &dst.s_addr, 4);
+	req.r.rtm_dst_len = 32;
+
+	ret = rtnl_open(&rth, 0);
+
+	if (ret < 0)
+		return ret;
+
+	if (rtnl_talk(&rth, &req.n, 0, 0, &req.n, NULL, NULL) < 0) {
+		ret = -1;
+		goto close_rth;
+	}
+
+	ret = find_dst(&req.n, src, gw);
+close_rth:
+	rtnl_close(&rth);
+	return ret;
+}
+
+#endif /* __linux__ */

pimd/mtracebis_routeget.h

@@ -0,0 +1,31 @@
+/*
+ * Multicast Traceroute for FRRouting
+ * Copyright (C) 2018  Mladen Sablic
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; see the file COPYING; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#ifdef __linux__
+
+#ifndef ROUTEGET_H
+#define ROUTEGET_H
+
+#include <netinet/in.h>
+
+int routeget(struct in_addr dst, struct in_addr *src, struct in_addr *gw);
+
+#endif /* ROUTEGET */
+
+#endif /* __linux__ */

pimd/pim_cmd.c

@@ -7311,6 +7311,27 @@ DEFUN (no_debug_msdp_packets,
 ALIAS(no_debug_msdp_packets, undebug_msdp_packets_cmd, "undebug msdp packets",
       UNDEBUG_STR DEBUG_MSDP_STR DEBUG_MSDP_PACKETS_STR)
 
+DEFUN (debug_mtrace,
+       debug_mtrace_cmd,
+       "debug mtrace",
+       DEBUG_STR
+       DEBUG_MTRACE_STR)
+{
+	PIM_DO_DEBUG_MTRACE;
+	return CMD_SUCCESS;
+}
+
+DEFUN (no_debug_mtrace,
+       no_debug_mtrace_cmd,
+       "no debug mtrace",
+       NO_STR
+       DEBUG_STR
+       DEBUG_MTRACE_STR)
+{
+	PIM_DONT_DEBUG_MTRACE;
+	return CMD_SUCCESS;
+}
+
 DEFUN_NOSH (show_debugging_pim,
 	    show_debugging_pim_cmd,
 	    "show debugging [pim]",
@@ -8721,6 +8742,8 @@ void pim_cmd_init(void)
 	install_element(ENABLE_NODE, &debug_msdp_packets_cmd);
 	install_element(ENABLE_NODE, &no_debug_msdp_packets_cmd);
 	install_element(ENABLE_NODE, &undebug_msdp_packets_cmd);
+	install_element(ENABLE_NODE, &debug_mtrace_cmd);
+	install_element(ENABLE_NODE, &no_debug_mtrace_cmd);
 
 	install_element(CONFIG_NODE, &debug_igmp_cmd);
 	install_element(CONFIG_NODE, &no_debug_igmp_cmd);
@@ -8763,6 +8786,8 @@ void pim_cmd_init(void)
 	install_element(CONFIG_NODE, &debug_msdp_packets_cmd);
 	install_element(CONFIG_NODE, &no_debug_msdp_packets_cmd);
 	install_element(CONFIG_NODE, &undebug_msdp_packets_cmd);
+	install_element(CONFIG_NODE, &debug_mtrace_cmd);
+	install_element(CONFIG_NODE, &no_debug_mtrace_cmd);
 
 	install_element(CONFIG_NODE, &ip_msdp_mesh_group_member_cmd);
 	install_element(VRF_NODE, &ip_msdp_mesh_group_member_cmd);

pimd/pim_cmd.h

@@ -63,6 +63,7 @@
 #define DEBUG_MSDP_EVENTS_STR                       "MSDP protocol events\n"
 #define DEBUG_MSDP_INTERNAL_STR                     "MSDP protocol internal\n"
 #define DEBUG_MSDP_PACKETS_STR                      "MSDP protocol packets\n"
+#define DEBUG_MTRACE_STR                            "Mtrace protocol activity\n"
 
 void pim_cmd_init(void);
 

pimd/pim_iface.c

@@ -85,9 +85,11 @@ static void *if_list_clean(struct pim_interface *pim_ifp)
 	if (pim_ifp->sec_addr_list)
 		list_delete_and_null(&pim_ifp->sec_addr_list);
 
-	while ((ch = RB_ROOT(pim_ifchannel_rb,
+	while (!RB_EMPTY(pim_ifchannel_rb, &pim_ifp->ifchannel_rb)) {
-			     &pim_ifp->ifchannel_rb)) != NULL)
+		ch = RB_ROOT(pim_ifchannel_rb, &pim_ifp->ifchannel_rb);
+
 		pim_ifchannel_delete(ch);
+	}
 
 	XFREE(MTYPE_PIM_INTERFACE, pim_ifp);
 
@@ -250,9 +252,11 @@ void pim_if_delete(struct interface *ifp)
 	if (pim_ifp->boundary_oil_plist)
 		XFREE(MTYPE_PIM_INTERFACE, pim_ifp->boundary_oil_plist);
 
-	while ((ch = RB_ROOT(pim_ifchannel_rb,
+	while (!RB_EMPTY(pim_ifchannel_rb, &pim_ifp->ifchannel_rb)) {
-			     &pim_ifp->ifchannel_rb)) != NULL)
+		ch = RB_ROOT(pim_ifchannel_rb, &pim_ifp->ifchannel_rb);
+
 		pim_ifchannel_delete(ch);
+	}
 
 	XFREE(MTYPE_PIM_INTERFACE, pim_ifp);
 

pimd/pim_ifchannel.c

@@ -211,8 +211,9 @@ void pim_ifchannel_delete_all(struct interface *ifp)
 	if (!pim_ifp)
 		return;
 
-	while ((ch = RB_ROOT(pim_ifchannel_rb,
+	while (!RB_EMPTY(pim_ifchannel_rb, &pim_ifp->ifchannel_rb)) {
-			     &pim_ifp->ifchannel_rb)) != NULL) {
+		ch = RB_ROOT(pim_ifchannel_rb, &pim_ifp->ifchannel_rb);
+
 		pim_ifchannel_delete(ch);
 	}
 }

pimd/pim_igmp.c

@@ -29,6 +29,7 @@
 #include "pim_igmp.h"
 #include "pim_igmpv2.h"
 #include "pim_igmpv3.h"
+#include "pim_igmp_mtrace.h"
 #include "pim_iface.h"
 #include "pim_sock.h"
 #include "pim_mroute.h"
@@ -504,6 +505,16 @@ int pim_igmp_packet(struct igmp_sock *igmp, char *buf, size_t len)
 	case PIM_IGMP_V2_LEAVE_GROUP:
 		return igmp_v2_recv_leave(igmp, ip_hdr->ip_src, from_str,
 					  igmp_msg, igmp_msg_len);
+
+	case PIM_IGMP_MTRACE_RESPONSE:
+		return igmp_mtrace_recv_response(igmp, ip_hdr, ip_hdr->ip_src,
+						 from_str, igmp_msg,
+						 igmp_msg_len);
+		break;
+	case PIM_IGMP_MTRACE_QUERY_REQUEST:
+		return igmp_mtrace_recv_qry_req(igmp, ip_hdr, ip_hdr->ip_src,
+						from_str, igmp_msg,
+						igmp_msg_len);
 	}
 
 	zlog_warn("Ignoring unsupported IGMP message type: %d", msg_type);

pimd/pim_igmp.h

@@ -37,6 +37,8 @@
 #define PIM_IGMP_V1_MEMBERSHIP_REPORT (0x12)
 #define PIM_IGMP_V2_MEMBERSHIP_REPORT (0x16)
 #define PIM_IGMP_V2_LEAVE_GROUP       (0x17)
+#define PIM_IGMP_MTRACE_RESPONSE      (0x1E)
+#define PIM_IGMP_MTRACE_QUERY_REQUEST (0x1F)
 #define PIM_IGMP_V3_MEMBERSHIP_REPORT (0x22)
 
 #define IGMP_V3_REPORT_HEADER_SIZE    (8)

pimd/pim_igmp_mtrace.c

@@ -0,0 +1,727 @@
+/*
+ * Multicast traceroute for FRRouting
+ * Copyright (C) 2017  Mladen Sablic
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; see the file COPYING; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include <zebra.h>
+
+#include "pimd.h"
+#include "pim_util.h"
+#include "pim_sock.h"
+#include "pim_rp.h"
+#include "pim_oil.h"
+#include "pim_ifchannel.h"
+#include "pim_macro.h"
+#include "pim_igmp_mtrace.h"
+
+static void mtrace_rsp_init(struct igmp_mtrace_rsp *mtrace_rspp)
+{
+	mtrace_rspp->arrival = 0;
+	mtrace_rspp->incoming.s_addr = 0;
+	mtrace_rspp->outgoing.s_addr = 0;
+	mtrace_rspp->prev_hop.s_addr = 0;
+	mtrace_rspp->in_count = MTRACE_UNKNOWN_COUNT;
+	mtrace_rspp->out_count = MTRACE_UNKNOWN_COUNT;
+	mtrace_rspp->total = MTRACE_UNKNOWN_COUNT;
+	mtrace_rspp->rtg_proto = 0;
+	mtrace_rspp->fwd_ttl = 0;
+	mtrace_rspp->mbz = 0;
+	mtrace_rspp->s = 0;
+	mtrace_rspp->src_mask = 0;
+	mtrace_rspp->fwd_code = MTRACE_FWD_CODE_NO_ERROR;
+}
+
+static void mtrace_rsp_debug(uint32_t qry_id, int rsp,
+			     struct igmp_mtrace_rsp *mrspp)
+{
+	char inc_str[INET_ADDRSTRLEN];
+	char out_str[INET_ADDRSTRLEN];
+	char prv_str[INET_ADDRSTRLEN];
+
+	zlog_debug(
+		"Rx mt(%d) qid=%ud arr=%x in=%s out=%s prev=%s proto=%d fwd=%d",
+		rsp, ntohl(qry_id), mrspp->arrival,
+		inet_ntop(AF_INET, &(mrspp->incoming), inc_str,
+			  sizeof(inc_str)),
+		inet_ntop(AF_INET, &(mrspp->outgoing), out_str,
+			  sizeof(out_str)),
+		inet_ntop(AF_INET, &(mrspp->prev_hop), prv_str,
+			  sizeof(prv_str)),
+		mrspp->rtg_proto, mrspp->fwd_code);
+}
+
+static void mtrace_debug(struct pim_interface *pim_ifp,
+			 struct igmp_mtrace *mtracep, int mtrace_len)
+{
+	char inc_str[INET_ADDRSTRLEN];
+	char grp_str[INET_ADDRSTRLEN];
+	char src_str[INET_ADDRSTRLEN];
+	char dst_str[INET_ADDRSTRLEN];
+	char rsp_str[INET_ADDRSTRLEN];
+
+	zlog_debug(
+		"Rx mtrace packet incoming on %s: "
+		"hops=%d type=%d size=%d, grp=%s, src=%s,"
+		" dst=%s rsp=%s ttl=%d qid=%ud",
+		inet_ntop(AF_INET, &(pim_ifp->primary_address), inc_str,
+			  sizeof(inc_str)),
+		mtracep->hops, mtracep->type, mtrace_len,
+		inet_ntop(AF_INET, &(mtracep->grp_addr), grp_str,
+			  sizeof(grp_str)),
+		inet_ntop(AF_INET, &(mtracep->src_addr), src_str,
+			  sizeof(src_str)),
+		inet_ntop(AF_INET, &(mtracep->dst_addr), dst_str,
+			  sizeof(dst_str)),
+		inet_ntop(AF_INET, &(mtracep->rsp_addr), rsp_str,
+			  sizeof(rsp_str)),
+		mtracep->rsp_ttl, ntohl(mtracep->qry_id));
+	if (mtrace_len > (int)sizeof(struct igmp_mtrace)) {
+
+		int i;
+
+		int responses = mtrace_len - sizeof(struct igmp_mtrace);
+
+		if ((responses % sizeof(struct igmp_mtrace_rsp)) != 0)
+			if (PIM_DEBUG_MTRACE)
+				zlog_debug(
+					"Mtrace response block of wrong"
+					 " length");
+
+		responses = responses / sizeof(struct igmp_mtrace_rsp);
+
+		for (i = 0; i < responses; i++)
+			mtrace_rsp_debug(mtracep->qry_id, i, &mtracep->rsp[i]);
+	}
+}
+
+/* 5.1 Query Arrival Time */
+static uint32_t query_arrival_time(void)
+{
+	struct timeval tv;
+	uint32_t qat;
+
+	char m_qat[] = "Query arrival time lookup failed: errno=%d: %s";
+
+	if (gettimeofday(&tv, NULL) < 0) {
+		if (PIM_DEBUG_MTRACE)
+			zlog_warn(m_qat, errno, safe_strerror(errno));
+		return 0;
+	}
+	/* not sure second offset correct, as I get different value */
+	qat = ((tv.tv_sec + 32384) << 16) + ((tv.tv_usec << 10) / 15625);
+
+	return qat;
+}
+
+static int mtrace_send_packet(struct interface *ifp,
+			      struct igmp_mtrace *mtracep,
+			      size_t mtrace_buf_len, struct in_addr dst_addr,
+			      struct in_addr group_addr)
+{
+	struct sockaddr_in to;
+	struct pim_interface *pim_ifp;
+	socklen_t tolen;
+	ssize_t sent;
+	int ret;
+	int fd;
+	char pim_str[INET_ADDRSTRLEN];
+	char rsp_str[INET_ADDRSTRLEN];
+	u_char ttl;
+
+	pim_ifp = ifp->info;
+
+	memset(&to, 0, sizeof(to));
+	to.sin_family = AF_INET;
+	to.sin_addr = dst_addr;
+	tolen = sizeof(to);
+
+	if (PIM_DEBUG_MTRACE)
+		zlog_debug("Sending mtrace packet to %s on %s",
+			   inet_ntop(AF_INET, &mtracep->rsp_addr, rsp_str,
+				     sizeof(rsp_str)),
+			   inet_ntop(AF_INET, &pim_ifp->primary_address,
+				     pim_str, sizeof(pim_str)));
+
+	fd = pim_socket_raw(IPPROTO_IGMP);
+
+	if (fd < 0)
+		return -1;
+
+	ret = pim_socket_bind(fd, ifp);
+
+	if (ret < 0) {
+		ret = -1;
+		goto close_fd;
+	}
+
+	if (IPV4_CLASS_DE(ntohl(dst_addr.s_addr))) {
+		if (IPV4_MC_LINKLOCAL(ntohl(dst_addr.s_addr))) {
+			ttl = 1;
+		} else {
+			if (mtracep->type == PIM_IGMP_MTRACE_RESPONSE)
+				ttl = mtracep->rsp_ttl;
+			else
+				ttl = 64;
+		}
+		ret = setsockopt(fd, IPPROTO_IP, IP_MULTICAST_TTL, &ttl,
+				 sizeof(ttl));
+
+		if (ret < 0) {
+			if (PIM_DEBUG_MTRACE)
+				zlog_warn("Failed to set socket multicast TTL");
+			ret = -1;
+			goto close_fd;
+		}
+	}
+
+	sent = sendto(fd, (char *)mtracep, mtrace_buf_len, MSG_DONTWAIT,
+		      (struct sockaddr *)&to, tolen);
+
+	if (sent != (ssize_t)mtrace_buf_len) {
+		char dst_str[INET_ADDRSTRLEN];
+		char group_str[INET_ADDRSTRLEN];
+
+		pim_inet4_dump("<dst?>", dst_addr, dst_str, sizeof(dst_str));
+		pim_inet4_dump("<group?>", group_addr, group_str,
+			       sizeof(group_str));
+		if (sent < 0) {
+			if (PIM_DEBUG_MTRACE)
+				zlog_warn(
+					"Send mtrace request failed for %s on"
+					"%s: group=%s msg_size=%zd: errno=%d: "
+					" %s",
+					dst_str, ifp->name, group_str,
+					mtrace_buf_len, errno,
+					safe_strerror(errno));
+		} else {
+			if (PIM_DEBUG_MTRACE)
+				zlog_warn(
+					"Send mtrace request failed for %s on"
+					" %s: group=%s msg_size=%zd: sent=%zd",
+					dst_str, ifp->name, group_str,
+					mtrace_buf_len, sent);
+		}
+		ret = -1;
+		goto close_fd;
+	}
+	ret = 0;
+close_fd:
+	close(fd);
+	return ret;
+}
+
+static int mtrace_un_forward_packet(struct pim_instance *pim, struct ip *ip_hdr,
+				    struct interface *interface)
+{
+	struct pim_nexthop nexthop;
+	struct sockaddr_in to;
+	struct interface *if_out;
+	socklen_t tolen;
+	int ret;
+	int fd;
+	int sent;
+	uint16_t checksum;
+
+	checksum = ip_hdr->ip_sum;
+
+	ip_hdr->ip_sum = 0;
+
+	if (checksum != in_cksum(ip_hdr, ip_hdr->ip_hl * 4))
+		return -1;
+
+	if (ip_hdr->ip_ttl-- <= 1)
+		return -1;
+
+	ip_hdr->ip_sum = in_cksum(ip_hdr, ip_hdr->ip_hl * 4);
+
+	fd = pim_socket_raw(IPPROTO_RAW);
+
+	if (fd < 0)
+		return -1;
+
+	pim_socket_ip_hdr(fd);
+
+	if (interface == NULL) {
+		ret = pim_nexthop_lookup(pim, &nexthop, ip_hdr->ip_dst, 0);
+
+		if (ret != 0) {
+			if (PIM_DEBUG_MTRACE)
+				zlog_warn(
+					"Dropping mtrace packet, "
+					"no route to destination");
+			return -1;
+		}
+
+		if_out = nexthop.interface;
+	} else {
+		if_out = interface;
+	}
+
+	ret = pim_socket_bind(fd, if_out);
+
+	if (ret < 0) {
+		close(fd);
+		return -1;
+	}
+
+	memset(&to, 0, sizeof(to));
+	to.sin_family = AF_INET;
+	to.sin_addr = ip_hdr->ip_dst;
+	tolen = sizeof(to);
+
+	sent = sendto(fd, ip_hdr, ntohs(ip_hdr->ip_len), 0,
+		      (struct sockaddr *)&to, tolen);
+
+	close(fd);
+
+	if (sent < 0) {
+		if (PIM_DEBUG_MTRACE)
+			zlog_warn(
+				"Failed to forward mtrace packet:"
+				" sendto errno=%d, %s",
+				errno, safe_strerror(errno));
+		return -1;
+	}
+
+	if (PIM_DEBUG_MTRACE) {
+		zlog_debug("Fwd mtrace packet len=%u to %s ttl=%u",
+			   ntohs(ip_hdr->ip_len), inet_ntoa(ip_hdr->ip_dst),
+			   ip_hdr->ip_ttl);
+	}
+
+	return 0;
+}
+
+static int mtrace_mc_forward_packet(struct pim_instance *pim, struct ip *ip_hdr)
+{
+	struct prefix_sg sg;
+	struct channel_oil *c_oil;
+	struct listnode *chnode;
+	struct listnode *chnextnode;
+	struct pim_ifchannel *ch = NULL;
+	int ret = -1;
+
+	memset(&sg, 0, sizeof(struct prefix_sg));
+	sg.grp = ip_hdr->ip_dst;
+
+	c_oil = pim_find_channel_oil(pim, &sg);
+
+	if (c_oil == NULL) {
+		if (PIM_DEBUG_MTRACE) {
+			zlog_debug(
+				"Dropping mtrace multicast packet "
+				"len=%u to %s ttl=%u",
+				ntohs(ip_hdr->ip_len),
+				inet_ntoa(ip_hdr->ip_dst), ip_hdr->ip_ttl);
+		}
+		return -1;
+	}
+	if (c_oil->up == NULL)
+		return -1;
+	if (c_oil->up->ifchannels == NULL)
+		return -1;
+	for (ALL_LIST_ELEMENTS(c_oil->up->ifchannels, chnode, chnextnode, ch)) {
+		if (pim_macro_chisin_oiflist(ch)) {
+			int r;
+
+			r = mtrace_un_forward_packet(pim, ip_hdr,
+						     ch->interface);
+			if (r == 0)
+				ret = 0;
+		}
+	}
+	return ret;
+}
+
+
+static int mtrace_forward_packet(struct pim_instance *pim, struct ip *ip_hdr)
+{
+	if (IPV4_CLASS_DE(ntohl(ip_hdr->ip_dst.s_addr)))
+		return mtrace_mc_forward_packet(pim, ip_hdr);
+	else
+		return mtrace_un_forward_packet(pim, ip_hdr, NULL);
+}
+
+/* 6.5 Sending Traceroute Responses */
+static int mtrace_send_mc_response(struct pim_instance *pim,
+				   struct igmp_mtrace *mtracep,
+				   size_t mtrace_len)
+{
+	struct prefix_sg sg;
+	struct channel_oil *c_oil;
+	struct listnode *chnode;
+	struct listnode *chnextnode;
+	struct pim_ifchannel *ch = NULL;
+	int ret = -1;
+
+	memset(&sg, 0, sizeof(struct prefix_sg));
+	sg.grp = mtracep->rsp_addr;
+
+	c_oil = pim_find_channel_oil(pim, &sg);
+
+	if (c_oil == NULL) {
+		if (PIM_DEBUG_MTRACE) {
+			zlog_debug(
+				"Dropping mtrace multicast response packet "
+				"len=%u to %s",
+				(unsigned int)mtrace_len,
+				inet_ntoa(mtracep->rsp_addr));
+		}
+		return -1;
+	}
+	if (c_oil->up == NULL)
+		return -1;
+	if (c_oil->up->ifchannels == NULL)
+		return -1;
+	for (ALL_LIST_ELEMENTS(c_oil->up->ifchannels, chnode, chnextnode, ch)) {
+		if (pim_macro_chisin_oiflist(ch)) {
+			int r;
+
+			r = mtrace_send_packet(ch->interface, mtracep,
+					       mtrace_len, mtracep->rsp_addr,
+					       mtracep->grp_addr);
+			if (r == 0)
+				ret = 0;
+		}
+	}
+	return ret;
+}
+
+static int mtrace_send_response(struct pim_instance *pim,
+				struct igmp_mtrace *mtracep, size_t mtrace_len)
+{
+	struct pim_nexthop nexthop;
+	int ret;
+
+	mtracep->type = PIM_IGMP_MTRACE_RESPONSE;
+
+	mtracep->checksum = 0;
+	mtracep->checksum = in_cksum((char *)mtracep, mtrace_len);
+
+	if (IPV4_CLASS_DE(ntohl(mtracep->rsp_addr.s_addr))) {
+		struct pim_rpf *p_rpf;
+		char grp_str[INET_ADDRSTRLEN];
+
+		if (pim_rp_i_am_rp(pim, mtracep->rsp_addr))
+			return mtrace_send_mc_response(pim, mtracep,
+						       mtrace_len);
+
+		p_rpf = pim_rp_g(pim, mtracep->rsp_addr);
+
+		if (p_rpf == NULL) {
+			if (PIM_DEBUG_MTRACE)
+				zlog_warn("mtrace no RP for %s",
+					  inet_ntop(AF_INET,
+						    &(mtracep->rsp_addr),
+						    grp_str, sizeof(grp_str)));
+			return -1;
+		}
+		nexthop = p_rpf->source_nexthop;
+		if (PIM_DEBUG_MTRACE)
+			zlog_debug("mtrace response to RP");
+	} else {
+		/* TODO: should use unicast rib lookup */
+		ret = pim_nexthop_lookup(pim, &nexthop, mtracep->rsp_addr, 1);
+
+		if (ret != 0) {
+			if (PIM_DEBUG_MTRACE)
+				zlog_warn(
+					"Dropped response qid=%ud, no route to "
+					"response address",
+					mtracep->qry_id);
+			return -1;
+		}
+	}
+
+	return mtrace_send_packet(nexthop.interface, mtracep, mtrace_len,
+				  mtracep->rsp_addr, mtracep->grp_addr);
+}
+
+int igmp_mtrace_recv_qry_req(struct igmp_sock *igmp, struct ip *ip_hdr,
+			     struct in_addr from, const char *from_str,
+			     char *igmp_msg, int igmp_msg_len)
+{
+	static uint32_t qry_id, qry_src;
+	char mtrace_buf[MTRACE_HDR_SIZE + MTRACE_MAX_HOPS * MTRACE_RSP_SIZE];
+	struct pim_nexthop nexthop;
+	struct interface *ifp;
+	struct interface *out_ifp;
+	struct pim_interface *pim_ifp;
+	struct pim_interface *pim_out_ifp;
+	struct pim_instance *pim;
+	struct igmp_mtrace *mtracep;
+	struct igmp_mtrace_rsp *rspp;
+	struct in_addr nh_addr;
+	enum mtrace_fwd_code fwd_code = MTRACE_FWD_CODE_NO_ERROR;
+	int ret;
+	size_t r_len;
+	int last_rsp_ind = 0;
+	size_t mtrace_len;
+	uint16_t recv_checksum;
+	uint16_t checksum;
+
+	ifp = igmp->interface;
+	pim_ifp = ifp->info;
+	pim = pim_ifp->pim;
+
+	/*
+	 * 6. Router Behaviour
+	 * Check if mtrace packet is addressed elsewhere and forward,
+	 * if applicable
+	 */
+	if (!IPV4_CLASS_DE(ntohl(ip_hdr->ip_dst.s_addr)))
+		if (!if_lookup_exact_address(&ip_hdr->ip_dst, AF_INET,
+					     pim->vrf_id))
+			return mtrace_forward_packet(pim, ip_hdr);
+
+	if (igmp_msg_len < (int)sizeof(struct igmp_mtrace)) {
+		if (PIM_DEBUG_MTRACE)
+			zlog_warn(
+				"Recv mtrace packet from %s on %s: too short,"
+				" len=%d, min=%zu",
+				from_str, ifp->name, igmp_msg_len,
+				sizeof(struct igmp_mtrace));
+		return -1;
+	}
+
+	mtracep = (struct igmp_mtrace *)igmp_msg;
+
+	recv_checksum = mtracep->checksum;
+
+	mtracep->checksum = 0;
+
+	checksum = in_cksum(igmp_msg, igmp_msg_len);
+
+	if (recv_checksum != checksum) {
+		if (PIM_DEBUG_MTRACE)
+			zlog_warn(
+				"Recv mtrace packet from %s on %s: checksum"
+				" mismatch: received=%x computed=%x",
+				from_str, ifp->name, recv_checksum, checksum);
+		return -1;
+	}
+
+	if (PIM_DEBUG_MTRACE)
+		mtrace_debug(pim_ifp, mtracep, igmp_msg_len);
+
+	/* subtract header from message length */
+	r_len = igmp_msg_len - sizeof(struct igmp_mtrace);
+
+	/* Classify mtrace packet, check if it is a query */
+	if (!r_len) {
+		if (PIM_DEBUG_MTRACE)
+			zlog_debug("Received IGMP multicast traceroute query");
+
+		/* 6.1.1  Packet verification */
+		if (!pim_if_connected_to_source(ifp, mtracep->dst_addr)) {
+			if (IPV4_CLASS_DE(ntohl(ip_hdr->ip_dst.s_addr))) {
+				if (PIM_DEBUG_MTRACE)
+					zlog_debug(
+						"Dropping multicast query "
+						"on wrong interface");
+				return -1;
+			}
+			/* Unicast query on wrong interface */
+			fwd_code = MTRACE_FWD_CODE_WRONG_IF;
+		}
+		if (qry_id == mtracep->qry_id && qry_src == from.s_addr) {
+			if (PIM_DEBUG_MTRACE)
+				zlog_debug(
+					"Dropping multicast query with "
+					"duplicate source and id");
+			return -1;
+		}
+		qry_id = mtracep->qry_id;
+		qry_src = from.s_addr;
+	}
+	/* if response fields length is equal to a whole number of responses */
+	else if ((r_len % sizeof(struct igmp_mtrace_rsp)) == 0) {
+		r_len = igmp_msg_len - sizeof(struct igmp_mtrace);
+
+		if (r_len != 0)
+			last_rsp_ind = r_len / sizeof(struct igmp_mtrace_rsp);
+		if (last_rsp_ind > MTRACE_MAX_HOPS) {
+			if (PIM_DEBUG_MTRACE)
+				zlog_warn("Mtrace request of excessive size");
+			return -1;
+		}
+	} else {
+		if (PIM_DEBUG_MTRACE)
+			zlog_warn(
+				"Recv mtrace packet from %s on %s: "
+				"invalid length %d",
+				from_str, ifp->name, igmp_msg_len);
+		return -1;
+	}
+
+	/* 6.2.1 Packet Verification - drop not link-local multicast */
+	if (IPV4_CLASS_DE(ntohl(ip_hdr->ip_dst.s_addr))
+	    && !IPV4_MC_LINKLOCAL(ntohl(ip_hdr->ip_dst.s_addr))) {
+		if (PIM_DEBUG_MTRACE)
+			zlog_warn(
+				"Recv mtrace packet from %s on %s:"
+				" not link-local multicast %s",
+				from_str, ifp->name, inet_ntoa(ip_hdr->ip_dst));
+		return -1;
+	}
+
+	/* 6.2.2. Normal Processing */
+
+	/* 6.2.2. 1. */
+
+	if (last_rsp_ind == MTRACE_MAX_HOPS) {
+		mtracep->rsp[MTRACE_MAX_HOPS - 1].fwd_code =
+			MTRACE_FWD_CODE_NO_SPACE;
+		return mtrace_send_response(pim_ifp->pim, mtracep,
+					    igmp_msg_len);
+	}
+
+	/* calculate new mtrace mtrace lenght with extra response */
+	mtrace_len = igmp_msg_len + sizeof(struct igmp_mtrace_rsp);
+
+	/* copy received query/request */
+	memcpy(mtrace_buf, igmp_msg, igmp_msg_len);
+
+	/* repoint mtracep pointer to copy */
+	mtracep = (struct igmp_mtrace *)mtrace_buf;
+
+	/* pointer for extra response field to be filled in */
+	rspp = &mtracep->rsp[last_rsp_ind];
+
+	/* initialize extra response field */
+	mtrace_rsp_init(rspp);
+
+	rspp->arrival = htonl(query_arrival_time());
+	rspp->outgoing = pim_ifp->primary_address;
+	rspp->out_count = htonl(MTRACE_UNKNOWN_COUNT);
+
+	/* 6.2.2. 2. Attempt to determine forwarding information */
+
+	nh_addr.s_addr = 0;
+
+	ret = pim_nexthop_lookup(pim, &nexthop, mtracep->src_addr, 1);
+
+	if (ret == 0) {
+		char nexthop_str[INET_ADDRSTRLEN];
+
+		if (PIM_DEBUG_MTRACE)
+			zlog_debug("mtrace pim_nexthop_lookup OK");
+
+		if (PIM_DEBUG_MTRACE)
+			zlog_warn("mtrace next_hop=%s",
+				  inet_ntop(nexthop.mrib_nexthop_addr.family,
+					    &nexthop.mrib_nexthop_addr.u.prefix,
+					    nexthop_str, sizeof(nexthop_str)));
+
+		if (nexthop.mrib_nexthop_addr.family == AF_INET)
+			nh_addr = nexthop.mrib_nexthop_addr.u.prefix4;
+	}
+	/* 6.4 Forwarding Traceroute Requests: ... Otherwise, ... */
+	else {
+		if (PIM_DEBUG_MTRACE)
+			zlog_debug("mtrace not found neighbor");
+		if (!fwd_code)
+			rspp->fwd_code = MTRACE_FWD_CODE_NO_ROUTE;
+		else
+			rspp->fwd_code = fwd_code;
+		/* 6.5 Sending Traceroute Responses */
+		return mtrace_send_response(pim, mtracep, mtrace_len);
+	}
+
+	out_ifp = nexthop.interface;
+	pim_out_ifp = out_ifp->info;
+
+	rspp->incoming = pim_out_ifp->primary_address;
+	rspp->prev_hop = nh_addr;
+	rspp->in_count = htonl(MTRACE_UNKNOWN_COUNT);
+	rspp->total = htonl(MTRACE_UNKNOWN_COUNT);
+	rspp->rtg_proto = MTRACE_RTG_PROTO_PIM;
+	rspp->s = 1;
+	rspp->src_mask = 32;
+
+	if (nh_addr.s_addr == 0) {
+		/* reached source? */
+		if (pim_if_connected_to_source(out_ifp, mtracep->src_addr))
+			return mtrace_send_response(pim, mtracep, mtrace_len);
+		/*
+		 * 6.4 Forwarding Traceroute Requests:
+		 * Previous-hop router not known
+		 */
+		inet_aton(MCAST_ALL_ROUTERS, &nh_addr);
+	}
+
+	if (mtracep->hops <= (last_rsp_ind + 1))
+		return mtrace_send_response(pim, mtracep, mtrace_len);
+
+	mtracep->checksum = 0;
+
+	mtracep->checksum = in_cksum(mtrace_buf, mtrace_len);
+
+	return mtrace_send_packet(out_ifp, mtracep, mtrace_len, nh_addr,
+				  mtracep->grp_addr);
+}
+
+int igmp_mtrace_recv_response(struct igmp_sock *igmp, struct ip *ip_hdr,
+			      struct in_addr from, const char *from_str,
+			      char *igmp_msg, int igmp_msg_len)
+{
+	static uint32_t qry_id, rsp_dst;
+	struct interface *ifp;
+	struct pim_interface *pim_ifp;
+	struct pim_instance *pim;
+	struct igmp_mtrace *mtracep;
+	uint16_t recv_checksum;
+	uint16_t checksum;
+
+	ifp = igmp->interface;
+	pim_ifp = ifp->info;
+	pim = pim_ifp->pim;
+
+	mtracep = (struct igmp_mtrace *)igmp_msg;
+
+	recv_checksum = mtracep->checksum;
+
+	mtracep->checksum = 0;
+
+	checksum = in_cksum(igmp_msg, igmp_msg_len);
+
+	if (recv_checksum != checksum) {
+		if (PIM_DEBUG_MTRACE)
+			zlog_warn(
+				"Recv mtrace response from %s on %s: checksum"
+				" mismatch: received=%x computed=%x",
+				from_str, ifp->name, recv_checksum, checksum);
+		return -1;
+	}
+
+	mtracep->checksum = checksum;
+
+	if (PIM_DEBUG_MTRACE)
+		mtrace_debug(pim_ifp, mtracep, igmp_msg_len);
+
+	/* Drop duplicate packets */
+	if (qry_id == mtracep->qry_id && rsp_dst == ip_hdr->ip_dst.s_addr) {
+		if (PIM_DEBUG_MTRACE)
+			zlog_debug("duplicate mtrace response packet dropped");
+		return -1;
+	}
+
+	qry_id = mtracep->qry_id;
+	rsp_dst = ip_hdr->ip_dst.s_addr;
+
+	return mtrace_forward_packet(pim, ip_hdr);
+}

pimd/pim_igmp_mtrace.h

@@ -0,0 +1,103 @@
+/*
+ * Multicast traceroute for FRRouting
+ * Copyright (C) 2017  Mladen Sablic
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; see the file COPYING; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#ifndef PIM_IGMP_MTRACE_H
+#define PIM_IGMP_MTRACE_H
+
+#include <zebra.h>
+
+#include "pim_igmp.h"
+
+#define MTRACE_MAX_HOPS (255)
+#define MTRACE_UNKNOWN_COUNT (0xffffffff)
+
+enum mtrace_fwd_code {
+	MTRACE_FWD_CODE_NO_ERROR        = 0x00,
+	MTRACE_FWD_CODE_WRONG_IF        = 0x01,
+	MTRACE_FWD_CODE_PRUNE_SENT      = 0x02,
+	MTRACE_FWD_CODE_PRUNE_RCVD      = 0x03,
+	MTRACE_FWD_CODE_SCOPED          = 0x04,
+	MTRACE_FWD_CODE_NO_ROUTE        = 0x05,
+	MTRACE_FWD_CODE_WRONG_LAST_HOP  = 0x06,
+	MTRACE_FWD_CODE_NOT_FORWARDING  = 0x07,
+	MTRACE_FWD_CODE_REACHED_RP      = 0x08,
+	MTRACE_FWD_CODE_RPF_IF          = 0x09,
+	MTRACE_FWD_CODE_NO_MULTICAST    = 0x0A,
+	MTRACE_FWD_CODE_INFO_HIDDEN     = 0x0B,
+	MTRACE_FWD_CODE_NO_SPACE        = 0x81,
+	MTRACE_FWD_CODE_OLD_ROUTER      = 0x82,
+	MTRACE_FWD_CODE_ADMIN_PROHIB    = 0x83
+};
+
+enum mtrace_rtg_proto {
+	MTRACE_RTG_PROTO_DVMRP          = 1,
+	MTRACE_RTG_PROTO_MOSPF          = 2,
+	MTRACE_RTG_PROTO_PIM            = 3,
+	MTRACE_RTG_PROTO_CBT            = 4,
+	MTRACE_RTG_PROTO_PIM_SPECIAL    = 5,
+	MTRACE_RTG_PROTO_PIM_STATIC     = 6,
+	MTRACE_RTG_PROTO_DVMRP_STATIC   = 7,
+	MTRACE_RTG_PROTO_PIM_MBGP       = 8,
+	MTRACE_RTG_PROTO_CBT_SPECIAL    = 9,
+	MTRACE_RTG_PROTO_CBT_STATIC     = 10,
+	MTRACE_RTG_PROTO_PIM_ASSERT     = 11,
+};
+
+struct igmp_mtrace_rsp {
+	uint32_t arrival;
+	struct in_addr incoming;
+	struct in_addr outgoing;
+	struct in_addr prev_hop;
+	uint32_t in_count;
+	uint32_t out_count;
+	uint32_t total;
+	uint32_t rtg_proto : 8;
+	uint32_t fwd_ttl : 8;
+	/* little endian order for next three fields */
+	uint32_t src_mask : 6;
+	uint32_t s : 1;
+	uint32_t mbz : 1;
+	uint32_t fwd_code : 8;
+} __attribute__((packed));
+
+struct igmp_mtrace {
+	uint8_t type;
+	uint8_t hops;
+	uint16_t checksum;
+	struct in_addr grp_addr;
+	struct in_addr src_addr;
+	struct in_addr dst_addr;
+	struct in_addr rsp_addr;
+	uint32_t rsp_ttl : 8;
+	uint32_t qry_id : 24;
+	struct igmp_mtrace_rsp rsp[0];
+} __attribute__((packed));
+
+#define MTRACE_HDR_SIZE (sizeof(struct igmp_mtrace))
+#define MTRACE_RSP_SIZE (sizeof(struct igmp_mtrace_rsp))
+
+int igmp_mtrace_recv_qry_req(struct igmp_sock *igmp, struct ip *ip_hdr,
+			     struct in_addr from, const char *from_str,
+			     char *igmp_msg, int igmp_msg_len);
+
+int igmp_mtrace_recv_response(struct igmp_sock *igmp, struct ip *ip_hdr,
+			      struct in_addr from, const char *from_str,
+			      char *igmp_msg, int igmp_msg_len);
+
+#endif /* PIM_IGMP_MTRACE_H */

pimd/pim_oil.c

@@ -135,7 +135,7 @@ void pim_channel_oil_free(struct channel_oil *c_oil)
 	XFREE(MTYPE_PIM_CHANNEL_OIL, c_oil);
 }
 
-static struct channel_oil *pim_find_channel_oil(struct pim_instance *pim,
+struct channel_oil *pim_find_channel_oil(struct pim_instance *pim,
 					 struct prefix_sg *sg)
 {
 	struct channel_oil *c_oil = NULL;

pimd/pim_oil.h

@@ -86,6 +86,8 @@ void pim_oil_init(struct pim_instance *pim);
 void pim_oil_terminate(struct pim_instance *pim);
 
 void pim_channel_oil_free(struct channel_oil *c_oil);
+struct channel_oil *pim_find_channel_oil(struct pim_instance *pim,
+					 struct prefix_sg *sg);
 struct channel_oil *pim_channel_oil_add(struct pim_instance *pim,
 					struct prefix_sg *sg,
 					int input_vif_index);

pimd/pim_vty.c

@@ -78,6 +78,11 @@ int pim_debug_config_write(struct vty *vty)
 		++writes;
 	}
 
+	if (PIM_DEBUG_MTRACE) {
+		vty_out(vty, "debug mtrace\n");
+		++writes;
+	}
+
 	if (PIM_DEBUG_MROUTE_DETAIL) {
 		vty_out(vty, "debug mroute detail\n");
 		++writes;

pimd/pimd.h

@@ -112,6 +112,7 @@
 #define PIM_MASK_PIM_NHT             (1 << 22)
 #define PIM_MASK_PIM_NHT_DETAIL      (1 << 23)
 #define PIM_MASK_PIM_NHT_RP          (1 << 24)
+#define PIM_MASK_MTRACE              (1 << 25)
 /* Remember 32 bits!!! */
 
 /* PIM error codes */
@@ -188,6 +189,7 @@ extern int32_t qpim_register_probe_time;
 #define PIM_DEBUG_PIM_NHT             (qpim_debugs & PIM_MASK_PIM_NHT)
 #define PIM_DEBUG_PIM_NHT_DETAIL      (qpim_debugs & PIM_MASK_PIM_NHT_DETAIL)
 #define PIM_DEBUG_PIM_NHT_RP          (qpim_debugs & PIM_MASK_PIM_NHT_RP)
+#define PIM_DEBUG_MTRACE              (qpim_debugs & PIM_MASK_MTRACE)
 
 #define PIM_DEBUG_EVENTS       (qpim_debugs & (PIM_MASK_PIM_EVENTS | PIM_MASK_IGMP_EVENTS | PIM_MASK_MSDP_EVENTS))
 #define PIM_DEBUG_PACKETS      (qpim_debugs & (PIM_MASK_PIM_PACKETS | PIM_MASK_IGMP_PACKETS | PIM_MASK_MSDP_PACKETS))
@@ -216,6 +218,7 @@ extern int32_t qpim_register_probe_time;
 #define PIM_DO_DEBUG_MSDP_INTERNAL       (qpim_debugs |= PIM_MASK_MSDP_INTERNAL)
 #define PIM_DO_DEBUG_PIM_NHT             (qpim_debugs |= PIM_MASK_PIM_NHT)
 #define PIM_DO_DEBUG_PIM_NHT_RP          (qpim_debugs |= PIM_MASK_PIM_NHT_RP)
+#define PIM_DO_DEBUG_MTRACE              (qpim_debugs |= PIM_MASK_MTRACE)
 
 #define PIM_DONT_DEBUG_PIM_EVENTS          (qpim_debugs &= ~PIM_MASK_PIM_EVENTS)
 #define PIM_DONT_DEBUG_PIM_PACKETS         (qpim_debugs &= ~PIM_MASK_PIM_PACKETS)
@@ -240,6 +243,7 @@ extern int32_t qpim_register_probe_time;
 #define PIM_DONT_DEBUG_MSDP_INTERNAL       (qpim_debugs &= ~PIM_MASK_MSDP_INTERNAL)
 #define PIM_DONT_DEBUG_PIM_NHT             (qpim_debugs &= ~PIM_MASK_PIM_NHT)
 #define PIM_DONT_DEBUG_PIM_NHT_RP          (qpim_debugs &= ~PIM_MASK_PIM_NHT_RP)
+#define PIM_DONT_DEBUG_MTRACE              (qpim_debugs &= ~PIM_MASK_MTRACE)
 
 void pim_init(void);
 void pim_terminate(void);

pimd/subdir.am

@@ -5,6 +5,7 @@
 if PIMD
 noinst_LIBRARIES += pimd/libpim.a
 sbin_PROGRAMS += pimd/pimd
+bin_PROGRAMS += pimd/mtracebis
 noinst_PROGRAMS += pimd/test_igmpv3_join
 dist_examples_DATA += pimd/pimd.conf.sample
 endif
@@ -18,6 +19,7 @@ pimd_libpim_a_SOURCES = \
 	pimd/pim_iface.c \
 	pimd/pim_ifchannel.c \
 	pimd/pim_igmp.c \
+	pimd/pim_igmp_mtrace.c \
 	pimd/pim_igmpv2.c \
 	pimd/pim_igmpv3.c \
 	pimd/pim_instance.c \
@@ -66,6 +68,7 @@ noinst_HEADERS += \
 	pimd/pim_ifchannel.h \
 	pimd/pim_igmp.h \
 	pimd/pim_igmp_join.h \
+	pimd/pim_igmp_mtrace.h \
 	pimd/pim_igmpv2.h \
 	pimd/pim_igmpv3.h \
 	pimd/pim_instance.h \
@@ -101,6 +104,8 @@ noinst_HEADERS += \
 	pimd/pim_zebra.h \
 	pimd/pim_zlookup.h \
 	pimd/pimd.h \
+	pimd/mtracebis_netlink.h \
+	pimd/mtracebis_routeget.h \
 	# end
 
 pimd_pimd_LDADD = pimd/libpim.a lib/libfrr.la @LIBCAP@
@@ -108,3 +113,9 @@ pimd_pimd_SOURCES = pimd/pim_main.c
 
 pimd_test_igmpv3_join_LDADD = lib/libfrr.la
 pimd_test_igmpv3_join_SOURCES = pimd/test_igmpv3_join.c
+
+pimd_mtracebis_LDADD =  lib/libfrr.la
+pimd_mtracebis_SOURCES = pimd/mtracebis.c \
+			 pimd/mtracebis_netlink.c \
+			 pimd/mtracebis_routeget.c \
+			 # end

sharpd/sharp_vty.c

@@ -80,14 +80,17 @@ DEFPY (install_routes,
 }
 
 DEFPY(vrf_label, vrf_label_cmd,
-      "sharp label vrf NAME$name label (0-100000)$label",
+      "sharp label <ip$ipv4|ipv6$ipv6> vrf NAME$name label (0-100000)$label",
       "Sharp Routing Protocol\n"
       "Give a vrf a label\n"
+      "Pop and forward for IPv4\n"
+      "Pop and forward for IPv6\n"
       VRF_CMD_HELP_STR
       "The label to use, 0 specifies remove the label installed from previous\n"
       "Specified range to use\n")
 {
 	struct vrf *vrf;
+	afi_t afi = (ipv4) ? AFI_IP : AFI_IP6;
 
 	if (strcmp(name, "default") == 0)
 		vrf = vrf_lookup_by_id(VRF_DEFAULT);
@@ -102,7 +105,7 @@ DEFPY(vrf_label, vrf_label_cmd,
 	if (label == 0)
 		label = MPLS_LABEL_NONE;
 
-	vrf_label_add(vrf->vrf_id, label);
+	vrf_label_add(vrf->vrf_id, afi, label);
 	return CMD_SUCCESS;
 }
 

sharpd/sharp_zebra.c

@@ -131,13 +131,14 @@ static int interface_state_down(int command, struct zclient *zclient,
 extern uint32_t total_routes;
 extern uint32_t installed_routes;
 
-static int notify_owner(int command, struct zclient *zclient,
+static int route_notify_owner(int command, struct zclient *zclient,
 			      zebra_size_t length, vrf_id_t vrf_id)
 {
 	struct prefix p;
 	enum zapi_route_notify_owner note;
+	uint32_t table;
 
-	if (!zapi_route_notify_decode(zclient->ibuf, &p, &note))
+	if (!zapi_route_notify_decode(zclient->ibuf, &p, &table, &note))
 		return -1;
 
 	installed_routes++;
@@ -152,9 +153,9 @@ static void zebra_connected(struct zclient *zclient)
 	zclient_send_reg_requests(zclient, VRF_DEFAULT);
 }
 
-void vrf_label_add(vrf_id_t vrf_id, mpls_label_t label)
+void vrf_label_add(vrf_id_t vrf_id, afi_t afi, mpls_label_t label)
 {
-	zclient_send_vrf_label(zclient, vrf_id, label, ZEBRA_LSP_SHARP);
+	zclient_send_vrf_label(zclient, vrf_id, afi, label, ZEBRA_LSP_SHARP);
 }
 
 void route_add(struct prefix *p, struct nexthop *nh)
@@ -168,6 +169,7 @@ void route_add(struct prefix *p, struct nexthop *nh)
 	api.safi = SAFI_UNICAST;
 	memcpy(&api.prefix, p, sizeof(*p));
 
+	SET_FLAG(api.flags, ZEBRA_FLAG_ALLOW_RECURSION);
 	SET_FLAG(api.message, ZAPI_MESSAGE_NEXTHOP);
 
 	api_nh = &api.nexthops[0];
@@ -210,5 +212,5 @@ void sharp_zebra_init(void)
 	zclient->interface_down = interface_state_down;
 	zclient->interface_address_add = interface_address_add;
 	zclient->interface_address_delete = interface_address_delete;
-	zclient->notify_owner = notify_owner;
+	zclient->route_notify_owner = route_notify_owner;
 }