fib: Table Replace

Type: feature

from the API doc, a table replace is:

"
    The use-case is that, for some unspecified reason, the control plane
    has a very different set of entries it wants in the table than VPP
    currently has. The CP would thus like to 'replace' VPP's current table
    only by specifying what the new set of entries shall be, i.e. it is not
    going to delete anything that already eixts.
    the CP delcartes the start of this procedure with this begin_replace
    API Call, and when it has populated all the entries it wants, it calls
    the below end_replace API. From this point on it is of coursce free
    to add and delete entries as usual.
    The underlying mechanism by which VPP implements this replace is
    purposefully left unspecified.
"

In the FIB, the algorithm is implemented using mark and sweep.

Algorithm goes:
1) replace_begin: this marks all the entries in that table as 'stale'
2) download all the entries that should be in this table
 - this clears the stale flag on those entries
3) signal the table converged: ip_table_replace_end
 - this removes all entries that are still stale

this procedure can be used when an agent first connects to VPP,
as an alternative to dump and diff state reconciliation.

Change-Id: I168edec10cf7670866076b129ebfe6149ea8222e
Signed-off-by: Neale Ranns <nranns@cisco.com>

2 files changed, 126 insertions, 131 deletions

diff --git a/src/vnet/ip/ip.api b/src/vnet/ip/ip.api
index 69c7b88f2e6..b37530ef955 100644
--- a/src/vnet/ip/ip.api
+++ b/src/vnet/ip/ip.api
@@ -65,6 +65,59 @@ define ip_table_dump
   u32 context;
 };
 
+/** \brief IP table replace being
+
+    The use-case is that, for some unspecified reason, the control plane
+    has a very different set of entries it wants in the table than VPP
+    currently has. The CP would thus like to 'replace' VPP's current table
+    only by specifying what the new set of entries shall be, i.e. it is not
+    going to delete anything that already eixts.
+    the CP delcartes the start of this procedure with this begin_replace
+    API Call, and when it has populated all the entries it wants, it calls
+    the below end_replace API. From this point on it is of coursce free
+    to add and delete entries as usual.
+    The underlying mechanism by which VPP implements this replace is
+    purposefully left unspecified.
+
+    @param client_index - opaque cookie to identify the sender
+    @param context - sender context, to match reply w/ request
+    @param table - The table to resync
+*/
+autoreply define ip_table_replace_begin
+{
+  u32 client_index;
+  u32 context;
+  vl_api_ip_table_t table;
+};
+
+/** \brief IP table replace end
+
+    see replace start/
+
+    @param client_index - opaque cookie to identify the sender
+    @param context - sender context, to match reply w/ request
+    @param table - The table that has converged
+*/
+autoreply define ip_table_replace_end
+{
+  u32 client_index;
+  u32 context;
+  vl_api_ip_table_t table;
+};
+
+/** \brief IP table flush
+    Flush a table of all routes
+    @param client_index - opaque cookie to identify the sender
+    @param context - sender context, to match reply w/ request
+    @param table - The table to flush
+*/
+autoreply define ip_table_flush
+{
+  u32 client_index;
+  u32 context;
+  vl_api_ip_table_t table;
+};
+
 /** \brief IP FIB table response
     @param context - sender context
     @param table - description of the table
@@ -957,20 +1010,6 @@ define proxy_arp_intfc_details
   u32 sw_if_index;
 };
 
-/** \brief Reset fib table request
-    @param client_index - opaque cookie to identify the sender
-    @param context - sender context, to match reply w/ request
-    @param vrf_id - vrf/table id of the fib table to reset
-    @param is_ipv6 - an ipv6 fib to reset if non-zero, else ipv4
-*/
-autoreply define reset_fib
-{
-  u32 client_index;
-  u32 context;
-  u32 vrf_id;
-  u8 is_ipv6;
-};
-
 /** \brief Set max allowed ARP or ip6 neighbor entries request
     @param client_index - opaque cookie to identify the sender
     @param context - sender context, to match reply w/ request
diff --git a/src/vnet/ip/ip_api.c b/src/vnet/ip/ip_api.c
index 322a3ed115a..0f9e3709643 100644
--- a/src/vnet/ip/ip_api.c
+++ b/src/vnet/ip/ip_api.c
@@ -92,7 +92,9 @@ _(PROXY_ARP_ADD_DEL, proxy_arp_add_del)                                 \
 _(PROXY_ARP_DUMP, proxy_arp_dump)                                       \
 _(PROXY_ARP_INTFC_ENABLE_DISABLE, proxy_arp_intfc_enable_disable)       \
  _(PROXY_ARP_INTFC_DUMP, proxy_arp_intfc_dump)                          \
-_(RESET_FIB, reset_fib)							\
+_(IP_TABLE_REPLACE_BEGIN, ip_table_replace_begin)                       \
+_(IP_TABLE_REPLACE_END, ip_table_replace_end)                           \
+_(IP_TABLE_FLUSH, ip_table_flush)                                       \
 _(IP_ROUTE_ADD_DEL, ip_route_add_del)                                   \
 _(IP_TABLE_ADD_DEL, ip_table_add_del)                                   \
 _(IP_PUNT_POLICE, ip_punt_police)                                       \
@@ -313,6 +315,7 @@ send_ip_route_details (vpe_api_main_t * am,
   }
 
   vl_api_send_msg (reg, (u8 *) mp);
+  vec_free (rpaths);
 }
 
 typedef struct apt_ip6_fib_show_ctx_t_
@@ -399,14 +402,14 @@ typedef struct vl_api_ip_mfib_dump_ctx_t_
   fib_node_index_t *entries;
 } vl_api_ip_mfib_dump_ctx_t;
 
-static int
+static walk_rc_t
 mfib_route_dump_walk (fib_node_index_t fei, void *arg)
 {
   vl_api_ip_mfib_dump_ctx_t *ctx = arg;
 
   vec_add1 (ctx->entries, fei);
 
-  return (0);
+  return (WALK_CONTINUE);
 }
 
 static void
@@ -2572,137 +2575,90 @@ static void
   REPLY_MACRO (VL_API_IP_SCAN_NEIGHBOR_ENABLE_DISABLE_REPLY);
 }
 
-static int
-ip4_reset_fib_t_handler (vl_api_reset_fib_t * mp)
+static void
+vl_api_ip_table_replace_begin_t_handler (vl_api_ip_table_replace_begin_t * mp)
 {
-  vnet_main_t *vnm = vnet_get_main ();
-  vnet_interface_main_t *im = &vnm->interface_main;
-  ip4_main_t *im4 = &ip4_main;
-  static u32 *sw_if_indices_to_shut;
-  fib_table_t *fib_table;
-  ip4_fib_t *fib;
-  u32 sw_if_index;
-  int i;
-  int rv = VNET_API_ERROR_NO_SUCH_FIB;
-  u32 target_fib_id = ntohl (mp->vrf_id);
-
-  /* *INDENT-OFF* */
-  pool_foreach (fib_table, im4->fibs,
-  ({
-    vnet_sw_interface_t * si;
-
-    fib = pool_elt_at_index (im4->v4_fibs, fib_table->ft_index);
-
-    if (fib->table_id != target_fib_id)
-      continue;
-
-    /* remove any mpls encap/decap labels */
-    mpls_fib_reset_labels (fib->table_id);
-
-    /* remove any proxy arps in this fib */
-    vnet_proxy_arp_fib_reset (fib->table_id);
-
-    /* Set the flow hash for this fib to the default */
-    vnet_set_ip4_flow_hash (fib->table_id, IP_FLOW_HASH_DEFAULT);
-
-    vec_reset_length (sw_if_indices_to_shut);
-
-    /* Shut down interfaces in this FIB / clean out intfc routes */
-    pool_foreach (si, im->sw_interfaces,
-    ({
-      u32 sw_if_index = si->sw_if_index;
-
-      if (sw_if_index < vec_len (im4->fib_index_by_sw_if_index)
-          && (im4->fib_index_by_sw_if_index[si->sw_if_index] ==
-              fib->index))
-        vec_add1 (sw_if_indices_to_shut, si->sw_if_index);
-    }));
+  vl_api_ip_table_replace_begin_reply_t *rmp;
+  fib_protocol_t fproto;
+  u32 fib_index;
+  int rv = 0;
 
-    for (i = 0; i < vec_len (sw_if_indices_to_shut); i++) {
-      sw_if_index = sw_if_indices_to_shut[i];
+  fproto = (mp->table.is_ip6 ? FIB_PROTOCOL_IP6 : FIB_PROTOCOL_IP4);
+  fib_index = fib_table_find (fproto, ntohl (mp->table.table_id));
 
-      u32 flags = vnet_sw_interface_get_flags (vnm, sw_if_index);
-      flags &= ~(VNET_SW_INTERFACE_FLAG_ADMIN_UP);
-      vnet_sw_interface_set_flags (vnm, sw_if_index, flags);
+  if (INDEX_INVALID == fib_index)
+    rv = VNET_API_ERROR_NO_SUCH_FIB;
+  else
+    {
+      fib_table_mark (fib_index, fproto, FIB_SOURCE_API);
+      mfib_table_mark (mfib_table_find (fproto, ntohl (mp->table.table_id)),
+		       fproto, MFIB_SOURCE_API);
     }
-
-    fib_table_flush(fib->index, FIB_PROTOCOL_IP4, FIB_SOURCE_API);
-
-    rv = 0;
-    break;
-    })); /* pool_foreach (fib) */
-    /* *INDENT-ON* */
-
-  return rv;
+  REPLY_MACRO (VL_API_IP_TABLE_REPLACE_BEGIN_REPLY);
 }
 
-static int
-ip6_reset_fib_t_handler (vl_api_reset_fib_t * mp)
+static void
+vl_api_ip_table_replace_end_t_handler (vl_api_ip_table_replace_end_t * mp)
 {
-  vnet_main_t *vnm = vnet_get_main ();
-  vnet_interface_main_t *im = &vnm->interface_main;
-  ip6_main_t *im6 = &ip6_main;
-  static u32 *sw_if_indices_to_shut;
-  fib_table_t *fib_table;
-  ip6_fib_t *fib;
-  u32 sw_if_index;
-  int i;
-  int rv = VNET_API_ERROR_NO_SUCH_FIB;
-  u32 target_fib_id = ntohl (mp->vrf_id);
-
-  /* *INDENT-OFF* */
-  pool_foreach (fib_table, im6->fibs,
-  ({
-    vnet_sw_interface_t * si;
-
-    fib = pool_elt_at_index (im6->v6_fibs, fib_table->ft_index);
-
-    if (fib->table_id != target_fib_id)
-      continue;
-
-    vec_reset_length (sw_if_indices_to_shut);
-
-    /* Set the flow hash for this fib to the default */
-    vnet_set_ip6_flow_hash (fib->table_id, IP_FLOW_HASH_DEFAULT);
-
-    /* Shut down interfaces in this FIB / clean out intfc routes */
-    pool_foreach (si, im->sw_interfaces,
-    ({
-      if (im6->fib_index_by_sw_if_index[si->sw_if_index] ==
-          fib->index)
-        vec_add1 (sw_if_indices_to_shut, si->sw_if_index);
-    }));
+  vl_api_ip_table_replace_end_reply_t *rmp;
+  fib_protocol_t fproto;
+  u32 fib_index;
+  int rv = 0;
 
-    for (i = 0; i < vec_len (sw_if_indices_to_shut); i++) {
-      sw_if_index = sw_if_indices_to_shut[i];
+  fproto = (mp->table.is_ip6 ? FIB_PROTOCOL_IP6 : FIB_PROTOCOL_IP4);
+  fib_index = fib_table_find (fproto, ntohl (mp->table.table_id));
 
-      u32 flags = vnet_sw_interface_get_flags (vnm, sw_if_index);
-      flags &= ~(VNET_SW_INTERFACE_FLAG_ADMIN_UP);
-      vnet_sw_interface_set_flags (vnm, sw_if_index, flags);
+  if (INDEX_INVALID == fib_index)
+    rv = VNET_API_ERROR_NO_SUCH_FIB;
+  else
+    {
+      fib_table_sweep (fib_index, fproto, FIB_SOURCE_API);
+      mfib_table_sweep (mfib_table_find
+			(fproto, ntohl (mp->table.table_id)), fproto,
+			MFIB_SOURCE_API);
     }
-
-    fib_table_flush(fib->index, FIB_PROTOCOL_IP6, FIB_SOURCE_API);
-
-    rv = 0;
-    break;
-  })); /* pool_foreach (fib) */
-  /* *INDENT-ON* */
-
-  return rv;
+  REPLY_MACRO (VL_API_IP_TABLE_REPLACE_END_REPLY);
 }
 
 static void
-vl_api_reset_fib_t_handler (vl_api_reset_fib_t * mp)
+vl_api_ip_table_flush_t_handler (vl_api_ip_table_flush_t * mp)
 {
-  int rv;
-  vl_api_reset_fib_reply_t *rmp;
+  vl_api_ip_table_flush_reply_t *rmp;
+  fib_protocol_t fproto;
+  u32 fib_index;
+  int rv = 0;
 
-  if (mp->is_ipv6)
-    rv = ip6_reset_fib_t_handler (mp);
+  fproto = (mp->table.is_ip6 ? FIB_PROTOCOL_IP6 : FIB_PROTOCOL_IP4);
+  fib_index = fib_table_find (fproto, ntohl (mp->table.table_id));
+
+  if (INDEX_INVALID == fib_index)
+    rv = VNET_API_ERROR_NO_SUCH_FIB;
   else
-    rv = ip4_reset_fib_t_handler (mp);
+    {
+      vnet_main_t *vnm = vnet_get_main ();
+      vnet_interface_main_t *im = &vnm->interface_main;
+      vnet_sw_interface_t *si;
+
+      /* Shut down interfaces in this FIB / clean out intfc routes */
+      /* *INDENT-OFF* */
+      pool_foreach (si, im->sw_interfaces,
+      ({
+        if (fib_index == fib_table_get_index_for_sw_if_index (fproto,
+                                                              si->sw_if_index))
+          {
+            u32 flags = si->flags;
+            flags &= ~VNET_SW_INTERFACE_FLAG_ADMIN_UP;
+            vnet_sw_interface_set_flags (vnm, si->sw_if_index, flags);
+          }
+      }));
+      /* *INDENT-ON* */
+
+      fib_table_flush (fib_index, fproto, FIB_SOURCE_API);
+      mfib_table_flush (mfib_table_find (fproto, ntohl (mp->table.table_id)),
+			fproto, MFIB_SOURCE_API);
+    }
 
-  REPLY_MACRO (VL_API_RESET_FIB_REPLY);
+  REPLY_MACRO (VL_API_IP_TABLE_FLUSH_REPLY);
 }
 
 static void