Nest: Implement prefix trie pruning

When rtable is pruned and network fib nodes are removed, we also need to
prune prefix trie. Unfortunately, rebuilding prefix trie takes long time
(got about 400 ms for 1M networks), so must not be atomic, we have to
rebuild a new trie while current one is still active. That may require
some considerable amount of temporary memory, so we do that only if
we expect significant trie size reduction.

diff --git a/nest/route.h b/nest/route.h
index 102ea0ea46b4282e63e043f05f648351099ba5bd..c4e0d2c60c3a8b9840632dd5d6f519f8fb3464ef 100644 (file)
--- a/nest/route.h
+++ b/nest/route.h
@@ -185,10 +185,12 @@ typedef struct rtable {
   btime gc_time;                       /* Time of last GC */
   int gc_counter;                      /* Number of operations since last GC */
   byte prune_state;                    /* Table prune state, 1 -> scheduled, 2-> running */
+  byte prune_trie;                     /* Prune prefix trie during next table prune */
   byte hcu_scheduled;                  /* Hostcache update is scheduled */
   byte nhu_state;                      /* Next Hop Update state */
   struct fib_iterator prune_fit;       /* Rtable prune FIB iterator */
   struct fib_iterator nhu_fit;         /* Next Hop Update FIB iterator */
+  struct f_trie *trie_new;             /* New prefix trie defined during pruning */
 
   list subscribers;                    /* Subscribers for notifications */
   struct timer *settle_timer;          /* Settle time for notifications */

diff --git a/nest/rt-table.c b/nest/rt-table.c
index ee679898132f8b7b976c0bab97963bb7b3a140b2..ac14aec97402080c2b054e1c3bd66ea2fbf90a03 100644 (file)
--- a/nest/rt-table.c
+++ b/nest/rt-table.c
@@ -134,6 +134,9 @@ net_init_with_trie(struct fib *f, void *N)
 
   if (tab->trie)
     trie_add_prefix(tab->trie, n->n.addr, n->n.addr->pxlen, n->n.addr->pxlen);
+
+  if (tab->trie_new)
+    trie_add_prefix(tab->trie_new, n->n.addr, n->n.addr->pxlen, n->n.addr->pxlen);
 }
 
 static inline net *
@@ -2399,6 +2402,13 @@ rt_prune_table(rtable *tab)
 
     FIB_ITERATE_INIT(fit, &tab->fib);
     tab->prune_state = 2;
+
+    if (tab->prune_trie)
+    {
+      /* Init prefix trie pruning */
+      tab->trie_new = f_new_trie(lp_new_default(tab->rp), 0);
+      tab->trie_new->ipv4 = tab->trie->ipv4;
+    }
   }
 
 again:
@@ -2407,17 +2417,17 @@ again:
       rte *e;
 
     rescan:
+      if (limit <= 0)
+      {
+       FIB_ITERATE_PUT(fit);
+       ev_schedule(tab->rt_event);
+       return;
+      }
+
       for (e=n->routes; e; e=e->next)
       {
        if (e->sender->flush_active || (e->flags & REF_DISCARD))
          {
-           if (limit <= 0)
-             {
-               FIB_ITERATE_PUT(fit);
-               ev_schedule(tab->rt_event);
-               return;
-             }
-
            rte_discard(e);
            limit--;
 
@@ -2426,13 +2436,6 @@ again:
 
        if (e->flags & REF_MODIFY)
          {
-           if (limit <= 0)
-             {
-               FIB_ITERATE_PUT(fit);
-               ev_schedule(tab->rt_event);
-               return;
-             }
-
            rte_modify(e);
            limit--;
 
@@ -2446,6 +2449,12 @@ again:
          fib_delete(&tab->fib, n);
          goto again;
        }
+
+      if (tab->trie_new)
+      {
+       trie_add_prefix(tab->trie_new, n->n.addr, n->n.addr->pxlen, n->n.addr->pxlen);
+       limit--;
+      }
     }
   FIB_ITERATE_END;
 
@@ -2459,6 +2468,25 @@ again:
   /* state change 2->0, 3->1 */
   tab->prune_state &= 1;
 
+  if (tab->trie_new)
+  {
+    /* Finish prefix trie pruning */
+    rfree(tab->trie->lp);
+    tab->trie = tab->trie_new;
+    tab->trie_new = NULL;
+    tab->prune_trie = 0;
+  }
+  else
+  {
+    /* Schedule prefix trie pruning */
+    if (tab->trie && (tab->trie->prefix_count > (2 * tab->fib.entries)))
+    {
+      /* state change 0->1, 2->3 */
+      tab->prune_state |= 1;
+      tab->prune_trie = 1;
+    }
+  }
+
   if (tab->prune_state > 0)
     ev_schedule(tab->rt_event);