MTRIE Optimisations 2

1) 16-8-8 stride. Reduce trie depth walk traded with increased memory in the top PLY.
2) separate the vector of protocol-independent (PI) fib_table_t with the vector of protocol dependent (PD) FIBs. PD FIBs are large structures, we don't want to burn the memory for ech PD type
3) Go straight to the PD FIB in the data-path thus avoiding an indirection through, e.g., a PLY pool.

Change-Id: I800d1ed0b2049040d5da95213f3ed6b12bdd78b7
Signed-off-by: Neale Ranns <nranns@cisco.com>

1 files changed, 5 insertions, 1 deletions

diff --git a/src/vnet/dpo/load_balance.c b/src/vnet/dpo/load_balance.c
index d5e98e4e32e..6b0eda0e8b1 100644
--- a/src/vnet/dpo/load_balance.c
+++ b/src/vnet/dpo/load_balance.c
@@ -827,14 +827,18 @@ const static char* const * const load_balance_nodes[DPO_PROTO_NUM] =
 void
 load_balance_module_init (void)
 {
+    index_t lbi;
+
     dpo_register(DPO_LOAD_BALANCE, &lb_vft, load_balance_nodes);
 
     /*
      * Special LB with index zero. we need to define this since the v4 mtrie
      * assumes an index of 0 implies the ply is empty. therefore all 'real'
      * adjs need a non-zero index.
+     * This should never be used, but just in case, stack it on a drop.
      */
-    load_balance_create(0, DPO_PROTO_IP4, 0);
+    lbi = load_balance_create(1, DPO_PROTO_IP4, 0);
+    load_balance_set_bucket(lbi, 0, drop_dpo_get(DPO_PROTO_IP4));
 
     load_balance_map_module_init();
 }