ovn-controller: Support VIF-based local encap IPs selection.

Commit dd527a2 partially supported multiple encap IPs. It supported remote encap IP selection based on the destination VIF's encap_ip configuration. This patch adds the support for selecting local encap IP based on the source VIF's encap_ip configuration. Co-authored-by: Lei Huang <leih@nvidia.com> Signed-off-by: Lei Huang <leih@nvidia.com> Signed-off-by: Han Zhou <hzhou@ovn.org> Acked-by: Ales Musil <amusil@redhat.com>
ovn-org · Jan 30, 2024 · 17b6a12 · 17b6a12
1 parent 41eefcb
commit 17b6a12
Show file tree

Hide file tree

Showing 11 changed files with 240 additions and 54 deletions.
diff --git a/NEWS b/NEWS
@@ -15,6 +15,9 @@ Post v23.09.0
   - A new LSP option "enable_router_port_acl" has been added to enable
     conntrack for the router port whose peer is l3dgw_port if set it true.
   - Enable PMTU discovery on geneve/vxlan tunnels for E/W traffic.
+  - Support selecting encapsulation IP based on the source/destination VIF's
+    settting. See ovn-controller(8) 'external_ids:ovn-encap-ip' for more
+    details.
 
 OVN v23.09.0 - 15 Sep 2023
 --------------------------

diff --git a/controller/chassis.c b/controller/chassis.c
@@ -61,7 +61,7 @@ struct ovs_chassis_cfg {
 
     /* Set of encap types parsed from the 'ovn-encap-type' external-id. */
     struct sset encap_type_set;
-    /* Set of encap IPs parsed from the 'ovn-encap-type' external-id. */
+    /* Set of encap IPs parsed from the 'ovn-encap-ip' external-id. */
     struct sset encap_ip_set;
     /* Interface type list formatted in the OVN-SB Chassis required format. */
     struct ds iface_types;

diff --git a/controller/local_data.c b/controller/local_data.c
@@ -528,7 +528,7 @@ chassis_tunnels_destroy(struct hmap *chassis_tunnels)
  */
 struct chassis_tunnel *
 chassis_tunnel_find(const struct hmap *chassis_tunnels, const char *chassis_id,
-                    char *remote_encap_ip, char *local_encap_ip)
+                    char *remote_encap_ip, const char *local_encap_ip)
 {
     /*
      * If the specific encap_ip is given, look for the chassisid_ip entry,

diff --git a/controller/local_data.h b/controller/local_data.h
@@ -154,7 +154,7 @@ bool local_nonvif_data_handle_ovs_iface_changes(
 struct chassis_tunnel *chassis_tunnel_find(const struct hmap *chassis_tunnels,
                                            const char *chassis_id,
                                            char *remote_encap_ip,
-                                           char *local_encap_ip);
+                                           const char *local_encap_ip);
 
 bool get_chassis_tunnel_ofport(const struct hmap *chassis_tunnels,
                                const char *chassis_name,

diff --git a/controller/ovn-controller.8.xml b/controller/ovn-controller.8.xml
@@ -176,10 +176,32 @@
 
       <dt><code>external_ids:ovn-encap-ip</code></dt>
       <dd>
-        The IP address that a chassis should use to connect to this node
-        using encapsulation types specified by
-        <code>external_ids:ovn-encap-type</code>. Multiple encapsulation IPs
-        may be specified with a comma-separated list.
+        <p>
+          The IP address that a chassis should use to connect to this node
+          using encapsulation types specified by
+          <code>external_ids:ovn-encap-type</code>. Multiple encapsulation IPs
+          may be specified with a comma-separated list.
+        </p>
+        <p>
+          In scenarios where multiple encapsulation IPs are present, distinct
+          tunnels are established for each remote chassis. These tunnels are
+          differentiated by setting unique <code>options:local_ip</code> and
+          <code>options:remote_ip</code> values in the tunnel interface. When
+          transmitting a packet to a remote chassis, the selection of local_ip
+          is guided by the <code>Interface:external_ids:encap-ip</code> from
+          the local OVSDB, corresponding to the VIF originating the packet, if
+          specified. The <code>Interface:external_ids:encap-ip</code> setting
+          of the VIF is also populated to the <code>Port_Binding</code>
+          table in the OVN SB database via the <code>encap</code> column.
+          Consequently, when a remote chassis needs to send a packet to a
+          port-binding associated with this VIF, it utilizes the tunnel with
+          the appropriate <code>options:remote_ip</code> that matches the
+          <code>ip</code> in <code>Port_Binding:encap</code>. This mechanism
+          is particularly beneficial for chassis with multiple physical
+          interfaces designated for tunneling, where each interface is
+          optimized for handling specific traffic associated with particular
+          VIFs.
+        </p>
       </dd>
 
       <dt><code>external_ids:ovn-encap-df_default</code></dt>

diff --git a/controller/ovn-controller.c b/controller/ovn-controller.c
@@ -4544,6 +4544,30 @@ struct ed_type_pflow_output {
     struct physical_debug debug;
 };
 
+static void
+parse_encap_ips(const struct ovsrec_open_vswitch_table *ovs_table,
+                size_t *n_encap_ips, const char * **encap_ips)
+{
+    const struct ovsrec_open_vswitch *cfg =
+        ovsrec_open_vswitch_table_first(ovs_table);
+    const char *chassis_id = get_ovs_chassis_id(ovs_table);
+    const char *encap_ips_str =
+        get_chassis_external_id_value(&cfg->external_ids, chassis_id,
+                                      "ovn-encap-ip", NULL);
+    struct sset encap_ip_set;
+    sset_from_delimited_string(&encap_ip_set, encap_ips_str,  ",");
+
+    /* Sort the ips so that their index is deterministic. */
+    *encap_ips = sset_sort(&encap_ip_set);
+
+    /* Copy the strings so that we can destroy the sset. */
+    for (size_t i = 0; (*encap_ips)[i]; i++) {
+        (*encap_ips)[i] = xstrdup((*encap_ips)[i]);
+    }
+    *n_encap_ips = sset_count(&encap_ip_set);
+    sset_destroy(&encap_ip_set);
+}
+
 static void init_physical_ctx(struct engine_node *node,
                               struct ed_type_runtime_data *rt_data,
                               struct ed_type_non_vif_data *non_vif_data,
@@ -4595,6 +4619,7 @@ static void init_physical_ctx(struct engine_node *node,
         engine_get_input_data("ct_zones", node);
     struct simap *ct_zones = &ct_zones_data->current;
 
+    parse_encap_ips(ovs_table, &p_ctx->n_encap_ips, &p_ctx->encap_ips);
     p_ctx->sbrec_port_binding_by_name = sbrec_port_binding_by_name;
     p_ctx->sbrec_port_binding_by_datapath = sbrec_port_binding_by_datapath;
     p_ctx->port_binding_table = port_binding_table;
@@ -4618,6 +4643,15 @@ static void init_physical_ctx(struct engine_node *node,
     pflow_output_get_debug(node, &p_ctx->debug);
 }
 
+static void
+destroy_physical_ctx(struct physical_ctx *p_ctx)
+{
+    for (size_t i = 0; i < p_ctx->n_encap_ips; i++) {
+        free((char *)(p_ctx->encap_ips[i]));
+    }
+    free(p_ctx->encap_ips);
+}
+
 static void *
 en_pflow_output_init(struct engine_node *node OVS_UNUSED,
                              struct engine_arg *arg OVS_UNUSED)
@@ -4654,6 +4688,7 @@ en_pflow_output_run(struct engine_node *node, void *data)
     struct physical_ctx p_ctx;
     init_physical_ctx(node, rt_data, non_vif_data, &p_ctx);
     physical_run(&p_ctx, pflow_table);
+    destroy_physical_ctx(&p_ctx);
 
     engine_set_node_state(node, EN_UPDATED);
 }
@@ -4698,6 +4733,7 @@ pflow_output_if_status_mgr_handler(struct engine_node *node,
                 bool removed = sbrec_port_binding_is_deleted(binding);
                 if (!physical_handle_flows_for_lport(binding, removed, &p_ctx,
                                                      &pfo->flow_table)) {
+                    destroy_physical_ctx(&p_ctx);
                     return false;
                 }
             }
@@ -4707,11 +4743,13 @@ pflow_output_if_status_mgr_handler(struct engine_node *node,
             bool removed = sbrec_port_binding_is_deleted(pb);
             if (!physical_handle_flows_for_lport(pb, removed, &p_ctx,
                                                  &pfo->flow_table)) {
+                destroy_physical_ctx(&p_ctx);
                 return false;
             }
         }
         engine_set_node_state(node, EN_UPDATED);
     }
+    destroy_physical_ctx(&p_ctx);
     return true;
 }
 
@@ -4738,11 +4776,13 @@ pflow_output_sb_port_binding_handler(struct engine_node *node,
         bool removed = sbrec_port_binding_is_deleted(pb);
         if (!physical_handle_flows_for_lport(pb, removed, &p_ctx,
                                              &pfo->flow_table)) {
+            destroy_physical_ctx(&p_ctx);
             return false;
         }
     }
 
     engine_set_node_state(node, EN_UPDATED);
+    destroy_physical_ctx(&p_ctx);
     return true;
 }
 
@@ -4762,6 +4802,7 @@ pflow_output_sb_multicast_group_handler(struct engine_node *node, void *data)
     physical_handle_mc_group_changes(&p_ctx, &pfo->flow_table);
 
     engine_set_node_state(node, EN_UPDATED);
+    destroy_physical_ctx(&p_ctx);
     return true;
 }
 
@@ -4794,6 +4835,7 @@ pflow_output_runtime_data_handler(struct engine_node *node, void *data)
         if (tdp->tracked_type != TRACKED_RESOURCE_UPDATED) {
             /* Fall back to full recompute when a local datapath
              * is added or deleted. */
+            destroy_physical_ctx(&p_ctx);
             return false;
         }
 
@@ -4804,12 +4846,14 @@ pflow_output_runtime_data_handler(struct engine_node *node, void *data)
                 lport->tracked_type == TRACKED_RESOURCE_REMOVED ? true: false;
             if (!physical_handle_flows_for_lport(lport->pb, removed, &p_ctx,
                                                  &pfo->flow_table)) {
+                destroy_physical_ctx(&p_ctx);
                 return false;
             }
         }
     }
 
     engine_set_node_state(node, EN_UPDATED);
+    destroy_physical_ctx(&p_ctx);
     return true;
 }
 
@@ -4866,12 +4910,14 @@ pflow_output_activated_ports_handler(struct engine_node *node, void *data)
         if (pb) {
             if (!physical_handle_flows_for_lport(pb, false, &p_ctx,
                                                  &pfo->flow_table)) {
+                destroy_physical_ctx(&p_ctx);
                 return false;
             }
             tag_port_as_activated_in_engine(pp);
         }
     }
     engine_set_node_state(node, EN_UPDATED);
+    destroy_physical_ctx(&p_ctx);
     return true;
 }