summaryrefslogtreecommitdiffstats
path: root/lib/route
diff options
context:
space:
mode:
authorroopa <roopa@cumulusnetworks.com>2012-12-13 05:50:49 (GMT)
committerThomas Graf <tgraf@suug.ch>2012-12-17 14:16:02 (GMT)
commit29b71371e764b04666d903bcbda452426aa1c634 (patch)
tree57da2cd2792fead6f465c03680fffa96a59ccc43 /lib/route
parent52b635f1f7c8509b99f944185c6545d81df750a2 (diff)
downloadlibnl-29b71371e764b04666d903bcbda452426aa1c634.zip
libnl-29b71371e764b04666d903bcbda452426aa1c634.tar.gz
libnl-29b71371e764b04666d903bcbda452426aa1c634.tar.bz2
route cache: Fix handling of ipv6 multipath routes
There are two ways kernel handles ipv6 equal cost multipath routes depending on which kernel version you are looking at. older kernels without ipv6 ECMP support, accept the below ECMP routes, #ip -6 route add 2001::/16 nexthop via fe80:2::2 dev swp1 #ip -6 route add 2001::/16 nexthop via fe80:2::3 dev swp1 store them as separate routes and pick the last one during lookup. Newer kernels, after the support for equal cost multipath routes was added http://patchwork.ozlabs.org/patch/188562/, now accept multipath routes added individually using the above 'ip -6 route' format OR sent using RTA_MULTIPATH with the below command #ip -6 route add 2001::/16 nexthop via fe80:2::2 dev swp1 nexthop via fe80:2::3 dev swp1 And the kernel now stores and treats them as equal cost multipath routes during lookups. However in all cases above, netlink notifications to ipv6 ECMP route adds and deletes come separately for each next hop. Example libnl notification in the above case with both old and new kernels: inet6 2001::/16 table main type unicast scope global priority 0x400 protocol boot nexthop via fe80:2::2 dev swp1 inet6 2001::/16 table main type unicast scope global priority 0x400 protocol boot nexthop via fe80:2::3 dev swp1 Since they are separate route notifications for objects with same key, libnl replaces the existing ones and hence only the last route update sticks. This patch uses the oo_update feature to not replace but update an existing route if its a ipv6 equal cost multipath route. The object after an update looks like the below (similar to ipv4 ECMP routes): inet6 2001::/16 table main type unicast scope global priority 0x400 protocol boot nexthop via fe80:2::2 dev swp1 nexthop via fe80:2::3 dev swp1 Signed-off-by: Shrijeet Mukherjee <shm@cumulusnetworks.com> Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com> Reviewed-by: Nolan Leake <nolan@cumulusnetworks.com> Reviewed-by: Wilson Kok <wkok@cumulusnetworks.com> Signed-off-by: Thomas Graf <tgraf@suug.ch>
Diffstat (limited to 'lib/route')
-rw-r--r--lib/route/route_obj.c96
1 files changed, 96 insertions, 0 deletions
diff --git a/lib/route/route_obj.c b/lib/route/route_obj.c
index cc29746..e092d80 100644
--- a/lib/route/route_obj.c
+++ b/lib/route/route_obj.c
@@ -441,6 +441,101 @@ nh_mismatch:
#undef ROUTE_DIFF
}
+static int route_update(struct nl_object *old_obj, struct nl_object *new_obj)
+{
+ struct rtnl_route *new_route = (struct rtnl_route *) new_obj;
+ struct rtnl_route *old_route = (struct rtnl_route *) old_obj;
+ struct rtnl_nexthop *new_nh;
+ char buf[INET6_ADDRSTRLEN+5];
+ int action = new_obj->ce_msgtype;
+
+ /*
+ * ipv6 ECMP route notifications from the kernel come as
+ * separate notifications, one for every nexthop. This update
+ * function collapses such route msgs into a single
+ * route with multiple nexthops. The resulting object looks
+ * similar to a ipv4 ECMP route
+ */
+ if (new_route->rt_family != AF_INET6 ||
+ new_route->rt_table == RT_TABLE_LOCAL)
+ return -NLE_OPNOTSUPP;
+
+ /*
+ * For routes that are already multipath,
+ * or dont have a nexthop dont do anything
+ */
+ if (rtnl_route_get_nnexthops(new_route) != 1)
+ return -NLE_OPNOTSUPP;
+
+ /*
+ * Get the only nexthop entry from the new route. For
+ * IPv6 we always get a route with a 0th NH
+ * filled or nothing at all
+ */
+ new_nh = rtnl_route_nexthop_n(new_route, 0);
+ if (!new_nh || !rtnl_route_nh_get_gateway(new_nh))
+ return -NLE_OPNOTSUPP;
+
+ switch(action) {
+ case RTM_NEWROUTE : {
+ struct rtnl_nexthop *cloned_nh;
+
+ /*
+ * Add the nexthop to old route
+ */
+ cloned_nh = rtnl_route_nh_clone(new_nh);
+ if (!cloned_nh)
+ return -NLE_NOMEM;
+ rtnl_route_add_nexthop(old_route, cloned_nh);
+
+ NL_DBG(2, "Route obj %p updated. Added "
+ "nexthop %p via %s\n", old_route, cloned_nh,
+ nl_addr2str(cloned_nh->rtnh_gateway, buf,
+ sizeof(buf)));
+ }
+ break;
+ case RTM_DELROUTE : {
+ struct rtnl_nexthop *old_nh;
+
+ /*
+ * Only take care of nexthop deletes and not
+ * route deletes. So, if there is only one nexthop
+ * quite likely we did not update it. So dont do
+ * anything and return
+ */
+ if (rtnl_route_get_nnexthops(old_route) <= 1)
+ return -NLE_OPNOTSUPP;
+
+ /*
+ * Find the next hop in old route and delete it
+ */
+ nl_list_for_each_entry(old_nh, &old_route->rt_nexthops,
+ rtnh_list) {
+ if (!rtnl_route_nh_compare(old_nh, new_nh, ~0, 0)) {
+
+ rtnl_route_remove_nexthop(old_route, old_nh);
+
+ NL_DBG(2, "Route obj %p updated. Removed "
+ "nexthop %p via %s\n", old_route,
+ old_nh,
+ nl_addr2str(old_nh->rtnh_gateway, buf,
+ sizeof(buf)));
+
+ rtnl_route_nh_free(old_nh);
+ break;
+ }
+ }
+ }
+ break;
+ default:
+ NL_DBG(2, "Unknown action associated "
+ "to object %p during route update\n", new_obj);
+ return -NLE_OPNOTSUPP;
+ }
+
+ return NLE_SUCCESS;
+}
+
static const struct trans_tbl route_attrs[] = {
__ADD(ROUTE_ATTR_FAMILY, family)
__ADD(ROUTE_ATTR_TOS, tos)
@@ -1201,6 +1296,7 @@ struct nl_object_ops route_obj_ops = {
},
.oo_compare = route_compare,
.oo_keygen = route_keygen,
+ .oo_update = route_update,
.oo_attrs2str = route_attrs2str,
.oo_id_attrs = (ROUTE_ATTR_FAMILY | ROUTE_ATTR_TOS |
ROUTE_ATTR_TABLE | ROUTE_ATTR_DST),