Merge branch 'oz-trie-table'

doc/bird.sgml

@@ -252,16 +252,9 @@ The global best route selection algorithm is (roughly) as follows:
 </itemize>
 
 <p><label id="dsc-table-sorted">Usually, a routing table just chooses a selected
-route from a list of entries for one network. But if the <cf/sorted/ option is
+route from a list of entries for one network. Optionally, these lists of entries
-activated, these lists of entries are kept completely sorted (according to
+are kept completely sorted (according to preference or some protocol-dependent
-preference or some protocol-dependent metric). This is needed for some features
+metric). See <ref id="rtable-sorted" name="sorted"> table option for details.
-of some protocols (e.g. <cf/secondary/ option of BGP protocol, which allows to
-accept not just a selected route, but the first route (in the sorted list) that
-is accepted by filters), but it is incompatible with some other features (e.g.
-<cf/deterministic med/ option of BGP protocol, which activates a way of choosing
-selected route that cannot be described using comparison and ordering). Minor
-advantage is that routes are shown sorted in <cf/show route/, minor disadvantage
-is that it is slightly more computationally expensive.
 
 <sect>Routes and network types
 <label id="routes">
@@ -628,18 +621,73 @@ include "tablename.conf";;
 	<cf/protocol/ times, and the <cf/iso long ms/ format for <cf/base/ and
 	<cf/log/ times.
 
-	<tag><label id="opt-table"><m/nettype/ table <m/name/ [sorted]</tag>
+	<tag><label id="opt-table"><m/nettype/ table <m/name/ [ { <m/option/; [<m/.../] } ]</tag>
-	Create a new routing table. The default routing tables <cf/master4/ and
+	Define a new routing table. The default routing tables <cf/master4/ and
-	<cf/master6/ are created implicitly, other routing tables have to be
+	<cf/master6/ are defined implicitly, other routing tables have to be
-	added by this command.  Option <cf/sorted/ can be used to enable sorting
+	defined by this option. See the <ref id="rtable-opts"
-	of routes, see <ref id="dsc-table-sorted" name="sorted table">
+	name="routing table configuration section"> for routing table options.
-	description for details.
 
 	<tag><label id="opt-eval">eval <m/expr/</tag>
 	Evaluates given filter expression. It is used by the developers for testing of filters.
 </descrip>
 
 
+<sect>Routing table options
+<label id="rtable-opts">
+
+<p>Most routing tables do not need any options and are defined without an option
+block, but there are still some options to tweak routing table behavior. Note
+that implicit tables (<cf/master4/ and <cf/master6/) can be redefined in order
+to set options.
+
+<descrip>
+	<tag><label id="rtable-sorted">sorted <m/switch/</tag>
+	Usually, a routing table just chooses the selected (best) route from a
+	list of routes for each network, while keeping remaining routes unsorted.
+	If enabled, these lists of routes are kept completely sorted (according
+	to preference or some protocol-dependent metric).
+
+	This is needed for some protocol features (e.g. <cf/secondary/ option of
+	BGP protocol, which allows to accept not just a selected route, but the
+	first route (in the sorted list) that is accepted by filters), but it is
+	incompatible with some other features (e.g. <cf/deterministic med/
+	option of BGP protocol, which activates a way of choosing selected route
+	that cannot be described using comparison and ordering). Minor advantage
+	is that routes are shown sorted in <cf/show route/, minor disadvantage
+	is that it is slightly more computationally expensive. Default: off.
+
+	<tag><label id="rtable-trie">trie <m/switch/</tag>
+	BIRD routing tables are implemented with hash tables, which is efficient
+	for exact-match lookups, but inconvenient for longest-match lookups or
+	interval lookups (finding superprefix or subprefixes). This option
+	activates additional trie structure that is used to accelerate these
+	lookups, while using the hash table for exact-match lookups.
+
+	This has advantage for <ref id="rpki" name="RPKI"> (on ROA tables),
+	for <ref id="bgp-gateway" name="recursive next-hops"> (on IGP tables),
+	and is required for <ref id="bgp-validate" name="flowspec validation">
+	(on base IP tables). Another advantage is that interval results (like
+	from <cf/show route in .../ command) are lexicographically sorted. The
+	disadvantage is that trie-enabled routing tables require more memory,
+	which may be an issue especially in multi-table setups. Default: off.
+
+	<tag><label id="rtable-min-settle-time">min settle time <m/time/</tag>
+	Specify a minimum value of the settle time. When a ROA table changes,
+	automatic <ref id="proto-rpki-reload" name="RPKI reload"> may be
+	triggered, after a short settle time. Minimum settle time is a delay
+	from the last ROA table change to wait for more updates. Default: 1 s.
+
+
+	<tag><label id="rtable-max-settle-time">max settle time <m/time/</tag>
+	Specify a maximum value of the settle time. When a ROA table changes,
+	automatic <ref id="proto-rpki-reload" name="RPKI reload"> may be
+	triggered, after a short settle time. Maximum settle time is an upper
+	limit to the settle time from the initial ROA table change even if
+	there are consecutive updates gradually renewing the settle time.
+	Default: 20 s.
+</descrip>
+
+
 <sect>Protocol options
 <label id="protocol-opts">
 
@@ -2290,6 +2338,7 @@ avoid routing loops.
 <item> <rfc id="8092"> - BGP Large Communities Attribute
 <item> <rfc id="8203"> - BGP Administrative Shutdown Communication
 <item> <rfc id="8212"> - Default EBGP Route Propagation Behavior without Policies
+<item> <rfc id="9117"> - Revised Validation Procedure for BGP Flow Specifications
 </itemize>
 
 <sect1>Route selection rules
@@ -2674,7 +2723,7 @@ using the following configuration parameters:
 
 	<tag><label id="bgp-error-wait-time">error wait time <m/number/,<m/number/</tag>
 	Minimum and maximum delay in seconds between a protocol failure (either
-	local or reported by the peer) and automatic restart. Doesn't apply
+	local or reported by the peer) and automatic restart. Doesn not apply
 	when <cf/disable after error/ is configured. If consecutive errors
 	happen, the delay is increased exponentially until it reaches the
 	maximum. Default: 60, 300.
@@ -2852,6 +2901,31 @@ be used in explicit configuration.
 	explicitly (to conserve memory). This option requires that the connected
 	routing table is <ref id="dsc-table-sorted" name="sorted">. Default: off.
 
+	<tag><label id="bgp-validate">validate <m/switch/</tag>
+	Apply flowspec validation procedure as described in <rfc id="8955">
+	section 6 and <rfc id="9117">. The Validation procedure enforces that
+	only routers in the forwarding path for a network can originate flowspec
+	rules for that network. The validation procedure should be used for EBGP
+	to prevent injection of malicious flowspec rules from outside, but it
+	should also be used for IBGP to ensure that selected flowspec rules are
+	consistent with selected IP routes. The validation procedure uses an IP
+	routing table (<ref id="bgp-base-table" name="base table">, see below)
+	against which flowspec rules are validated. This option is limited to
+	flowspec channels. Default: off (for compatibility reasons).
+
+	Note that currently the flowspec validation does not work reliably
+	together with <ref id="bgp-import-table" name="import table"> option
+	enabled on flowspec channels.
+
+	<tag><label id="bgp-base-table">base table <m/name/</tag>
+	Specifies an IP table used for the flowspec validation procedure. The
+	table must have enabled <cf/trie/ option, otherwise the validation
+	procedure would not work. The type of the table must be <cf/ipv4/ for
+	<cf/flow4/ channels and <cf/ipv6/ for <cf/flow6/ channels. This option
+	is limited to flowspec channels. Default: the main table of the
+	<cf/ipv4/ / <cf/ipv6/ channel of the same BGP instance, or the
+	<cf/master4/ / <cf/master6/ table if there is no such channel.
+
 	<tag><label id="bgp-extended-next-hop">extended next hop <m/switch/</tag>
 	BGP expects that announced next hops have the same address family as
 	associated network prefixes. This option provides an extension to use

filter/data.h

@@ -140,18 +140,23 @@ struct f_tree {
   void *data;
 };
 
+#define TRIE_STEP		4
+#define TRIE_STACK_LENGTH	33
+
 struct f_trie_node4
 {
   ip4_addr addr, mask, accept;
-  uint plen;
+  u16 plen;
-  struct f_trie_node4 *c[2];
+  u16 local;
+  struct f_trie_node4 *c[1 << TRIE_STEP];
 };
 
 struct f_trie_node6
 {
   ip6_addr addr, mask, accept;
-  uint plen;
+  u16 plen;
-  struct f_trie_node6 *c[2];
+  u16 local;
+  struct f_trie_node6 *c[1 << TRIE_STEP];
 };
 
 struct f_trie_node
@@ -168,9 +173,20 @@ struct f_trie
   u8 zero;
   s8 ipv4;				/* -1 for undefined / empty */
   u16 data_size;			/* Additional data for each trie node */
+  u32 prefix_count;			/* Works only for restricted tries (pxlen == l == h) */
   struct f_trie_node root;		/* Root trie node */
 };
 
+struct f_trie_walk_state
+{
+  u8 ipv4;
+  u8 accept_length;			/* Current inter-node prefix position */
+  u8 start_pos;				/* Initial prefix position in stack[0] */
+  u8 local_pos;				/* Current intra-node prefix position */
+  u8 stack_pos;				/* Current node in stack below */
+  const struct f_trie_node *stack[TRIE_STACK_LENGTH];
+};
+
 struct f_tree *f_new_tree(void);
 struct f_tree *build_tree(struct f_tree *);
 const struct f_tree *find_tree(const struct f_tree *t, const struct f_val *val);
@@ -181,9 +197,70 @@ void tree_walk(const struct f_tree *t, void (*hook)(const struct f_tree *, void
 struct f_trie *f_new_trie(linpool *lp, uint data_size);
 void *trie_add_prefix(struct f_trie *t, const net_addr *n, uint l, uint h);
 int trie_match_net(const struct f_trie *t, const net_addr *n);
+int trie_match_longest_ip4(const struct f_trie *t, const net_addr_ip4 *net, net_addr_ip4 *dst, ip4_addr *found0);
+int trie_match_longest_ip6(const struct f_trie *t, const net_addr_ip6 *net, net_addr_ip6 *dst, ip6_addr *found0);
+void trie_walk_init(struct f_trie_walk_state *s, const struct f_trie *t, const net_addr *from);
+int trie_walk_next(struct f_trie_walk_state *s, net_addr *net);
 int trie_same(const struct f_trie *t1, const struct f_trie *t2);
 void trie_format(const struct f_trie *t, buffer *buf);
 
+static inline int
+trie_match_next_longest_ip4(net_addr_ip4 *n, ip4_addr *found)
+{
+  while (n->pxlen)
+  {
+    n->pxlen--;
+    ip4_clrbit(&n->prefix, n->pxlen);
+
+    if (ip4_getbit(*found, n->pxlen))
+      return 1;
+  }
+
+  return 0;
+}
+
+static inline int
+trie_match_next_longest_ip6(net_addr_ip6 *n, ip6_addr *found)
+{
+  while (n->pxlen)
+  {
+    n->pxlen--;
+    ip6_clrbit(&n->prefix, n->pxlen);
+
+    if (ip6_getbit(*found, n->pxlen))
+      return 1;
+  }
+
+  return 0;
+}
+
+
+#define TRIE_WALK_TO_ROOT_IP4(trie, net, dst) ({		\
+  net_addr_ip4 dst;						\
+  ip4_addr _found;						\
+  for (int _n = trie_match_longest_ip4(trie, net, &dst, &_found); \
+       _n;							\
+       _n = trie_match_next_longest_ip4(&dst, &_found))
+
+#define TRIE_WALK_TO_ROOT_IP6(trie, net, dst) ({		\
+  net_addr_ip6 dst;						\
+  ip6_addr _found;						\
+  for (int _n = trie_match_longest_ip6(trie, net, &dst, &_found); \
+       _n;							\
+       _n = trie_match_next_longest_ip6(&dst, &_found))
+
+#define TRIE_WALK_TO_ROOT_END })
+
+
+#define TRIE_WALK(trie, net, from) ({				\
+  net_addr net;							\
+  struct f_trie_walk_state tws_;				\
+  trie_walk_init(&tws_, trie, from);				\
+  while (trie_walk_next(&tws_, &net))
+
+#define TRIE_WALK_END })
+
+
 #define F_CMP_ERROR 999
 
 const char *f_type_name(enum f_type t);

filter/test.conf

@@ -499,6 +499,33 @@ prefix set pxs;
 
 	bt_assert(1.2.0.0/16 ~ [ 1.0.0.0/8{ 15 , 17 } ]);
 	bt_assert([ 10.0.0.0/8{ 15 , 17 } ] != [ 11.0.0.0/8{ 15 , 17 } ]);
+
+	/* Formatting of prefix sets, some cases are a bit strange */
+	bt_assert(format([ 0.0.0.0/0 ]) = "[0.0.0.0/0]");
+	bt_assert(format([ 10.10.0.0/32 ]) = "[10.10.0.0/32{0.0.0.1}]");
+	bt_assert(format([ 10.10.0.0/17 ]) = "[10.10.0.0/17{0.0.128.0}]");
+	bt_assert(format([ 10.10.0.0/17{17,19} ]) = "[10.10.0.0/17{0.0.224.0}]"); # 224 = 128+64+32
+	bt_assert(format([ 10.10.128.0/17{18,19} ]) = "[10.10.128.0/18{0.0.96.0}, 10.10.192.0/18{0.0.96.0}]"); # 96 = 64+32
+	bt_assert(format([ 10.10.64.0/18- ]) = "[0.0.0.0/0, 0.0.0.0/1{128.0.0.0}, 0.0.0.0/2{64.0.0.0}, 0.0.0.0/3{32.0.0.0}, 10.10.0.0/16{255.255.0.0}, 10.10.0.0/17{0.0.128.0}, 10.10.64.0/18{0.0.64.0}]");
+	bt_assert(format([ 10.10.64.0/18+ ]) = "[10.10.64.0/18{0.0.96.0}, 10.10.64.0/20{0.0.31.255}, 10.10.80.0/20{0.0.31.255}, 10.10.96.0/20{0.0.31.255}, 10.10.112.0/20{0.0.31.255}]");
+
+	bt_assert(format([ 10.10.160.0/19 ]) = "[10.10.160.0/19{0.0.32.0}]");
+	bt_assert(format([ 10.10.160.0/19{19,22} ]) = "[10.10.160.0/19{0.0.32.0}, 10.10.160.0/20{0.0.28.0}, 10.10.176.0/20{0.0.28.0}]"); # 28 = 16+8+4
+	bt_assert(format([ 10.10.160.0/19+ ]) = "[10.10.160.0/19{0.0.32.0}, 10.10.160.0/20{0.0.31.255}, 10.10.176.0/20{0.0.31.255}]");
+
+	bt_assert(format([ ::/0 ]) = "[::/0]");
+	bt_assert(format([ 11:22:33:44:55:66:77:88/128 ]) = "[11:22:33:44:55:66:77:88/128{::1}]");
+	bt_assert(format([ 11:22:33:44::/64 ]) = "[11:22:33:44::/64{0:0:0:1::}]");
+	bt_assert(format([ 11:22:33:44::/64+ ]) = "[11:22:33:44::/64{::1:ffff:ffff:ffff:ffff}]");
+
+	bt_assert(format([ 11:22:33:44::/65 ]) = "[11:22:33:44::/65{::8000:0:0:0}]");
+	bt_assert(format([ 11:22:33:44::/65{65,67} ]) = "[11:22:33:44::/65{::e000:0:0:0}]"); # e = 8+4+2
+	bt_assert(format([ 11:22:33:44:8000::/65{66,67} ]) = "[11:22:33:44:8000::/66{::6000:0:0:0}, 11:22:33:44:c000::/66{::6000:0:0:0}]"); # 6 = 4+2
+	bt_assert(format([ 11:22:33:44:4000::/66- ]) = "[::/0, ::/1{8000::}, ::/2{4000::}, ::/3{2000::}, 11:22:33:44::/64{ffff:ffff:ffff:ffff::}, 11:22:33:44::/65{::8000:0:0:0}, 11:22:33:44:4000::/66{::4000:0:0:0}]");
+	bt_assert(format([ 11:22:33:44:4000::/66+ ]) = "[11:22:33:44:4000::/66{::6000:0:0:0}, 11:22:33:44:4000::/68{::1fff:ffff:ffff:ffff}, 11:22:33:44:5000::/68{::1fff:ffff:ffff:ffff}, 11:22:33:44:6000::/68{::1fff:ffff:ffff:ffff}, 11:22:33:44:7000::/68{::1fff:ffff:ffff:ffff}]");
+	bt_assert(format([ 11:22:33:44:c000::/67 ]) = "[11:22:33:44:c000::/67{::2000:0:0:0}]");
+	bt_assert(format([ 11:22:33:44:c000::/67{67,71} ]) = "[11:22:33:44:c000::/67{::2000:0:0:0}, 11:22:33:44:c000::/68{::1e00:0:0:0}, 11:22:33:44:d000::/68{::1e00:0:0:0}]");
+	bt_assert(format([ 11:22:33:44:c000::/67+ ]) = "[11:22:33:44:c000::/67{::2000:0:0:0}, 11:22:33:44:c000::/68{::1fff:ffff:ffff:ffff}, 11:22:33:44:d000::/68{::1fff:ffff:ffff:ffff}]");
 }
 
 bt_test_suite(t_prefix_set, "Testing prefix sets");
@@ -578,6 +605,12 @@ prefix set pxs;
 	bt_assert(2000::/29 !~ pxs);
 	bt_assert(1100::/10 !~ pxs);
 	bt_assert(2010::/26 !~ pxs);
+
+	pxs = [ 52E0::/13{13,128} ];
+	bt_assert(52E7:BE81:379B:E6FD:541F:B0D0::/93 ~ pxs);
+
+	pxs = [ 41D8:8718::/30{0,30}, 413A:99A8:6C00::/38{38,128} ];
+	bt_assert(4180::/9 ~ pxs);
 }
 
 bt_test_suite(t_prefix6_set, "Testing prefix IPv6 sets");

filter/trie_test.c

@@ -14,9 +14,12 @@
 #include "conf/conf.h"
 
 #define TESTS_NUM		10
-#define PREFIXES_NUM 		10
+#define PREFIXES_NUM 		32
 #define PREFIX_TESTS_NUM 	10000
+#define PREFIX_BENCH_NUM 	100000000
 
+#define TRIE_BUFFER_SIZE	1024
+#define TEST_BUFFER_SIZE	(1024*1024)
 #define BIG_BUFFER_SIZE		10000
 
 /* Wrapping structure for storing f_prefixes structures in list */
@@ -31,146 +34,860 @@ xrandom(u32 max)
   return (bt_random() % max);
 }
 
+static inline uint
+get_exp_random(void)
+{
+  uint r, n = 0;
+
+  for (r = bt_random(); r & 1; r = r >> 1)
+    n++;
+
+  return n;
+}
+
 static int
-is_prefix_included(list *prefixes, struct f_prefix *needle)
+compare_prefixes(const void *a, const void *b)
+{
+  return net_compare(&((const struct f_prefix *) a)->net,
+		     &((const struct f_prefix *) b)->net);
+}
+
+static inline int
+matching_ip4_nets(const net_addr_ip4 *a, const net_addr_ip4 *b)
+{
+  ip4_addr cmask = ip4_mkmask(MIN(a->pxlen, b->pxlen));
+  return ip4_compare(ip4_and(a->prefix, cmask), ip4_and(b->prefix, cmask)) == 0;
+}
+
+static inline int
+matching_ip6_nets(const net_addr_ip6 *a, const net_addr_ip6 *b)
+{
+  ip6_addr cmask = ip6_mkmask(MIN(a->pxlen, b->pxlen));
+  return ip6_compare(ip6_and(a->prefix, cmask), ip6_and(b->prefix, cmask)) == 0;
+}
+
+static inline int
+matching_nets(const net_addr *a, const net_addr *b)
+{
+  if (a->type != b->type)
+    return 0;
+
+  return (a->type == NET_IP4) ?
+    matching_ip4_nets((const net_addr_ip4 *) a, (const net_addr_ip4 *) b) :
+    matching_ip6_nets((const net_addr_ip6 *) a, (const net_addr_ip6 *) b);
+}
+
+static int
+is_prefix_included(list *prefixes, const net_addr *needle)
 {
   struct f_prefix_node *n;
   WALK_LIST(n, *prefixes)
-  {
+    if (matching_nets(&n->prefix.net, needle) &&
-    ip6_addr cmask = ip6_mkmask(MIN(n->prefix.net.pxlen, needle->net.pxlen));
+	(n->prefix.lo <= needle->pxlen) && (needle->pxlen <= n->prefix.hi))
-
-    ip6_addr ip = net6_prefix(&n->prefix.net);
-    ip6_addr needle_ip = net6_prefix(&needle->net);
-
-    if ((ipa_compare(ipa_and(ip, cmask), ipa_and(needle_ip, cmask)) == 0) &&
-	(n->prefix.lo <= needle->net.pxlen) && (needle->net.pxlen <= n->prefix.hi))
     {
-      bt_debug("FOUND\t" PRIip6 "/%d %d-%d\n", ARGip6(net6_prefix(&n->prefix.net)), n->prefix.net.pxlen, n->prefix.lo, n->prefix.hi);
+      char buf[64];
+      bt_format_net(buf, 64, &n->prefix.net);
+      bt_debug("FOUND %s %d-%d\n", buf, n->prefix.lo, n->prefix.hi);
+
       return 1; /* OK */
     }
-  }
+
   return 0; /* FAIL */
 }
 
-static struct f_prefix
+static void
-get_random_ip6_prefix(void)
+get_random_net(net_addr *net, int v6)
 {
-  struct f_prefix p;
+  if (!v6)
-  u8 pxlen = xrandom(120)+8;
-  ip6_addr ip6 = ip6_build(bt_random(),bt_random(),bt_random(),bt_random());
-  net_addr_ip6 net6 = NET_ADDR_IP6(ip6, pxlen);
-
-  p.net = *((net_addr*) &net6);
-
-  if (bt_random() % 2)
   {
-    p.lo = 0;
+    uint pxlen = xrandom(24)+8;
-    p.hi = p.net.pxlen;
+    ip4_addr ip4 = ip4_from_u32((u32) bt_random());
+    net_fill_ip4(net, ip4_and(ip4, ip4_mkmask(pxlen)), pxlen);
   }
   else
   {
-    p.lo = p.net.pxlen;
+    uint pxlen = xrandom(120)+8;
-    p.hi = net_max_prefix_length[p.net.type];
+    ip6_addr ip6 = ip6_build(bt_random(), bt_random(), bt_random(), bt_random());
+    net_fill_ip6(net, ip6_and(ip6, ip6_mkmask(pxlen)), pxlen);
   }
-
-  return p;
 }
 
 static void
-generate_random_ipv6_prefixes(list *prefixes)
+get_random_prefix(struct f_prefix *px, int v6, int tight)
 {
-  int i;
+  get_random_net(&px->net, v6);
-  for (i = 0; i < PREFIXES_NUM; i++)
+
+  if (tight)
   {
-    struct f_prefix f = get_random_ip6_prefix();
+    px->lo = px->hi = px->net.pxlen;
-
+  }
-    struct f_prefix_node *px = calloc(1, sizeof(struct f_prefix_node));
+  else if (bt_random() % 2)
-    px->prefix = f;
+  {
-
+    px->lo = 0;
-    bt_debug("ADD\t" PRIip6 "/%d %d-%d\n", ARGip6(net6_prefix(&px->prefix.net)), px->prefix.net.pxlen, px->prefix.lo, px->prefix.hi);
+    px->hi = px->net.pxlen;
-    add_tail(prefixes, &px->n);
+  }
+  else
+  {
+    px->lo = px->net.pxlen;
+    px->hi = net_max_prefix_length[px->net.type];
   }
 }
 
+static void
+get_random_ip4_subnet(net_addr_ip4 *net, const net_addr_ip4 *src, int pxlen)
+{
+  *net = NET_ADDR_IP4(ip4_and(src->prefix, ip4_mkmask(pxlen)), pxlen);
+
+  if (pxlen > src->pxlen)
+  {
+    ip4_addr rnd = ip4_from_u32((u32) bt_random());
+    ip4_addr mask = ip4_xor(ip4_mkmask(src->pxlen), ip4_mkmask(pxlen));
+    net->prefix = ip4_or(net->prefix, ip4_and(rnd, mask));
+  }
+}
+
+static void
+get_random_ip6_subnet(net_addr_ip6 *net, const net_addr_ip6 *src, int pxlen)
+{
+  *net = NET_ADDR_IP6(ip6_and(src->prefix, ip6_mkmask(pxlen)), pxlen);
+
+  if (pxlen > src->pxlen)
+  {
+    ip6_addr rnd = ip6_build(bt_random(), bt_random(), bt_random(), bt_random());
+    ip6_addr mask = ip6_xor(ip6_mkmask(src->pxlen), ip6_mkmask(pxlen));
+    net->prefix = ip6_or(net->prefix, ip6_and(rnd, mask));
+  }
+}
+
+static void
+get_random_subnet(net_addr *net, const net_addr *src, int pxlen)
+{
+  if (src->type == NET_IP4)
+    get_random_ip4_subnet((net_addr_ip4 *) net, (const net_addr_ip4 *) src, pxlen);
+  else
+    get_random_ip6_subnet((net_addr_ip6 *) net, (const net_addr_ip6 *) src, pxlen);
+}
+
+static void
+get_inner_net(net_addr *net, const struct f_prefix *src)
+{
+  int pxlen, step;
+
+  if (bt_random() % 2)
+  {
+    step = get_exp_random();
+    step = MIN(step, src->hi - src->lo);
+    pxlen = (bt_random() % 2) ? (src->lo + step) : (src->hi - step);
+  }
+  else
+    pxlen = src->lo + bt_random() % (src->hi - src->lo + 1);
+
+  get_random_subnet(net, &src->net, pxlen);
+}
+
+static void
+swap_random_bits_ip4(net_addr_ip4 *net, int num)
+{
+  for (int i = 0; i < num; i++)
+  {
+    ip4_addr swap = IP4_NONE;
+    ip4_setbit(&swap, bt_random() % net->pxlen);
+    net->prefix = ip4_xor(net->prefix, swap);
+  }
+}
+
+static void
+swap_random_bits_ip6(net_addr_ip6 *net, int num)
+{
+  for (int i = 0; i < num; i++)
+  {
+    ip6_addr swap = IP6_NONE;
+    ip6_setbit(&swap, bt_random() % net->pxlen);
+    net->prefix = ip6_xor(net->prefix, swap);
+  }
+}
+
+static void
+swap_random_bits(net_addr *net, int num)
+{
+  if (net->type == NET_IP4)
+    swap_random_bits_ip4((net_addr_ip4 *) net, num);
+  else
+    swap_random_bits_ip6((net_addr_ip6 *) net, num);
+}
+
+static void
+get_outer_net(net_addr *net, const struct f_prefix *src)
+{
+  int pxlen, step;
+  int inside = 0;
+  int max = net_max_prefix_length[src->net.type];
+
+  if ((src->lo > 0) && (bt_random() % 3))
+  {
+    step = 1 + get_exp_random();
+    step = MIN(step, src->lo);
+    pxlen = src->lo - step;
+  }
+  else if ((src->hi < max) && (bt_random() % 2))
+  {
+    step = 1 + get_exp_random();
+    step = MIN(step, max - src->hi);
+    pxlen = src->hi + step;
+  }
+  else
+  {
+    pxlen = src->lo + bt_random() % (src->hi - src->lo + 1);
+    inside = 1;
+  }
+
+  get_random_subnet(net, &src->net, pxlen);
+
+  /* Perhaps swap some bits in prefix */
+  if ((net->pxlen > 0) && (inside || (bt_random() % 4)))
+    swap_random_bits(net, 1 + get_exp_random());
+}
+
+static list *
+make_random_prefix_list(linpool *lp, int num, int v6, int tight)
+{
+  list *prefixes = lp_allocz(lp, sizeof(struct f_prefix_node));
+  init_list(prefixes);
+
+  for (int i = 0; i < num; i++)
+  {
+    struct f_prefix_node *px = lp_allocz(lp, sizeof(struct f_prefix_node));
+    get_random_prefix(&px->prefix, v6, tight);
+    add_tail(prefixes, &px->n);
+
+    char buf[64];
+    bt_format_net(buf, 64, &px->prefix.net);
+    bt_debug("ADD %s{%d,%d}\n", buf, px->prefix.lo, px->prefix.hi);
+  }
+
+  return prefixes;
+}
+
+static struct f_trie *
+make_trie_from_prefix_list(linpool *lp, list *prefixes)
+{
+  struct f_trie *trie = f_new_trie(lp, 0);
+
+  struct f_prefix_node *n;
+  WALK_LIST(n, *prefixes)
+    trie_add_prefix(trie, &n->prefix.net, n->prefix.lo, n->prefix.hi);
+
+  return trie;
+}
+
+/*
+ * Read sequence of prefixes from file handle and return prefix list.
+ * Each prefix is on one line, sequence terminated by empty line or eof.
+ * Arg @plus means prefix should include all longer ones.
+ */
+static list *
+read_prefix_list(linpool *lp, FILE *f, int v6, int plus)
+{
+  ASSERT(!v6);
+
+  uint a0, a1, a2, a3, pl;
+  char s[32];
+  int n;
+
+  list *pxlist = lp_allocz(lp, sizeof(struct f_prefix_node));
+  init_list(pxlist);
+
+  errno = 0;
+  while (fgets(s, 32, f))
+  {
+    if (s[0] == '\n')
+      return pxlist;
+
+    n = sscanf(s, "%u.%u.%u.%u/%u", &a0, &a1, &a2, &a3, &pl);
+
+    if (n != 5)
+      bt_abort_msg("Invalid content of trie_data");
+
+    struct f_prefix_node *px = lp_allocz(lp, sizeof(struct f_prefix_node));
+    net_fill_ip4(&px->prefix.net, ip4_build(a0, a1, a2, a3), pl);
+    px->prefix.lo = pl;
+    px->prefix.hi = plus ? IP4_MAX_PREFIX_LENGTH : pl;
+    add_tail(pxlist, &px->n);
+
+    char buf[64];
+    bt_format_net(buf, 64, &px->prefix.net);
+    bt_debug("ADD %s{%d,%d}\n", buf, px->prefix.lo, px->prefix.hi);
+  }
+
+  bt_syscall(errno, "fgets()");
+  return EMPTY_LIST(*pxlist) ? NULL : pxlist;
+}
+
+/*
+ * Open file, read multiple sequences of prefixes from it. Fill @data with
+ * prefix lists and @trie with generated tries. Return number of sequences /
+ * tries. Use separate linpool @lp0 for prefix lists and @lp1 for tries.
+ * Arg @plus means prefix should include all longer ones.
+ */
 static int
-t_match_net(void)
+read_prefix_file(const char *filename, int plus,
+		 linpool *lp0, linpool *lp1,
+		 list *data[], struct f_trie *trie[])
+{
+  FILE *f = fopen(filename, "r");
+  bt_syscall(!f, "fopen(%s)", filename);
+
+  int n = 0;
+  list *pxlist;
+  while (pxlist = read_prefix_list(lp0, f, 0, plus))
+  {
+    data[n] = pxlist;
+    trie[n] = make_trie_from_prefix_list(lp1, pxlist);
+    bt_debug("NEXT\n");
+    n++;
+  }
+
+  fclose(f);
+  bt_debug("DONE reading %d tries\n", n);
+
+  return n;
+}
+
+/*
+ * Select random subset of @dn prefixes from prefix list @src of length @sn,
+ * and store them to buffer @dst (of size @dn). Prefixes may be chosen multiple
+ * times. Randomize order of prefixes in @dst buffer.
+ */
+static void
+select_random_prefix_subset(list *src[], net_addr dst[], int sn, int dn)
+{
+  int pn = 0;
+
+  if (!dn)
+    return;
+
+  /* Compute total prefix number */
+  for (int i = 0; i < sn; i++)
+    pn += list_length(src[i]);
+
+  /* Change of selecting a prefix */
+  int rnd = (pn / dn) + 10;
+  int n = 0;
+
+  /* Iterate indefinitely over src array */
+  for (int i = 0; 1; i++, i = (i < sn) ? i : 0)
+  {
+    struct f_prefix_node *px;
+    WALK_LIST(px, *src[i])
+    {
+      if (xrandom(rnd) != 0)
+	continue;
+
+      net_copy(&dst[n], &px->prefix.net);
+      n++;
+
+      /* We have enough */
+      if (n == dn)
+	goto done;
+    }
+  }
+
+done:
+  /* Shuffle networks */
+  for (int i = 0; i < dn; i++)
+  {
+    int j = xrandom(dn);
+
+    if (i == j)
+      continue;
+
+    net_addr tmp;
+    net_copy(&tmp, &dst[i]);
+    net_copy(&dst[i], &dst[j]);
+    net_copy(&dst[j], &tmp);
+  }
+}
+
+/* Fill @dst buffer with @dn randomly generated /32 prefixes */
+static void
+make_random_addresses(net_addr dst[], int dn)
+{
+  for (int i = 0; i < dn; i++)
+    net_fill_ip4(&dst[i], ip4_from_u32((u32) bt_random()), IP4_MAX_PREFIX_LENGTH);
+}
+
+static void
+test_match_net(list *prefixes, struct f_trie *trie, const net_addr *net)
+{
+  char buf[64];
+  bt_format_net(buf, 64, net);
+  bt_debug("TEST %s\n", buf);
+
+  int should_be = is_prefix_included(prefixes, net);
+  int is_there  = trie_match_net(trie, net);
+
+  bt_assert_msg(should_be == is_there, "Prefix %s %s match", buf,
+		(should_be ? "should" : "should not"));
+}
+
+static int
+t_match_random_net(void)
 {
   bt_bird_init();
   bt_config_parse(BT_CONFIG_SIMPLE);
 
-  uint round;
+  int v6 = 0;
-  for (round = 0; round < TESTS_NUM; round++)
+  linpool *lp = lp_new_default(&root_pool);
+  for (int round = 0; round < TESTS_NUM; round++)
   {
-    list prefixes; /* of structs f_extended_prefix */
+    list *prefixes = make_random_prefix_list(lp, PREFIXES_NUM, v6, 0);
-    init_list(&prefixes);
+    struct f_trie *trie = make_trie_from_prefix_list(lp, prefixes);
-    struct f_trie *trie = f_new_trie(config->mem, 0);
 
-    generate_random_ipv6_prefixes(&prefixes);
+    for (int i = 0; i < PREFIX_TESTS_NUM; i++)
-    struct f_prefix_node *n;
-    WALK_LIST(n, prefixes)
     {
-      trie_add_prefix(trie, &n->prefix.net, n->prefix.lo, n->prefix.hi);
+      net_addr net;
+      get_random_net(&net, v6);
+      test_match_net(prefixes, trie, &net);
     }
 
-    int i;
+    v6 = !v6;
-    for (i = 0; i < PREFIX_TESTS_NUM; i++)
+    lp_flush(lp);
-    {
-      struct f_prefix f = get_random_ip6_prefix();
-      bt_debug("TEST\t" PRIip6 "/%d\n", ARGip6(net6_prefix(&f.net)), f.net.pxlen);
-
-      int should_be = is_prefix_included(&prefixes, &f);
-      int is_there  = trie_match_net(trie, &f.net);
-      bt_assert_msg(should_be == is_there, "Prefix " PRIip6 "/%d %s", ARGip6(net6_prefix(&f.net)), f.net.pxlen, (should_be ? "should be found in trie" : "should not be found in trie"));
-    }
-
-    struct f_prefix_node *nxt;
-    WALK_LIST_DELSAFE(n, nxt, prefixes)
-    {
-      free(n);
-    }
   }
 
   bt_bird_cleanup();
   return 1;
 }
 
+static int
+t_match_inner_net(void)
+{
+  bt_bird_init();
+  bt_config_parse(BT_CONFIG_SIMPLE);
+
+  int v6 = 0;
+  linpool *lp = lp_new_default(&root_pool);
+  for (int round = 0; round < TESTS_NUM; round++)
+  {
+    list *prefixes = make_random_prefix_list(lp, PREFIXES_NUM, v6, 0);
+    struct f_trie *trie = make_trie_from_prefix_list(lp, prefixes);
+
+    struct f_prefix_node *n = HEAD(*prefixes);
+    for (int i = 0; i < PREFIX_TESTS_NUM; i++)
+    {
+      net_addr net;
+      get_inner_net(&net, &n->prefix);
+      test_match_net(prefixes, trie, &net);
+
+      n = NODE_VALID(NODE_NEXT(n)) ? NODE_NEXT(n) : HEAD(*prefixes);
+    }
+
+    v6 = !v6;
+    lp_flush(lp);
+  }
+
+  bt_bird_cleanup();
+  return 1;
+}
+
+static int
+t_match_outer_net(void)
+{
+  bt_bird_init();
+  bt_config_parse(BT_CONFIG_SIMPLE);
+
+  int v6 = 0;
+  linpool *lp = lp_new_default(&root_pool);
+  for (int round = 0; round < TESTS_NUM; round++)
+  {
+    list *prefixes = make_random_prefix_list(lp, PREFIXES_NUM, v6, 0);
+    struct f_trie *trie = make_trie_from_prefix_list(lp, prefixes);
+
+    struct f_prefix_node *n = HEAD(*prefixes);
+    for (int i = 0; i < PREFIX_TESTS_NUM; i++)
+    {
+      net_addr net;
+      get_outer_net(&net, &n->prefix);
+      test_match_net(prefixes, trie, &net);
+
+      n = NODE_VALID(NODE_NEXT(n)) ? NODE_NEXT(n) : HEAD(*prefixes);
+    }
+
+    v6 = !v6;
+    lp_flush(lp);
+  }
+
+  v6 = !v6;
+  bt_bird_cleanup();
+  return 1;
+}
+
+/*
+ * Read prefixes from @filename, build set of tries, prepare test data and do
+ * PREFIX_BENCH_NUM trie lookups. With @plus = 0, use random subset of known
+ * prefixes as test data, with @plus = 1, use randomly generated /32 prefixes
+ * as test data.
+ */
+static int
+benchmark_trie_dataset(const char *filename, int plus)
+{
+  int n = 0;
+  linpool *lp0 = lp_new_default(&root_pool);
+  linpool *lp1 = lp_new_default(&root_pool);
+  list *data[TRIE_BUFFER_SIZE];
+  struct f_trie *trie[TRIE_BUFFER_SIZE];
+  net_addr *nets;
+
+  bt_reset_suite_case_timer();
+  bt_log_suite_case_result(1, "Reading %s", filename, n);
+  n = read_prefix_file(filename, plus, lp0, lp1, data, trie);
+  bt_log_suite_case_result(1, "Read prefix data, %d lists, ", n);
+
+  size_t trie_size = rmemsize(lp1).effective * 1000 / (1024*1024);
+  bt_log_suite_case_result(1, "Trie size %u.%03u MB",
+			   (uint) (trie_size / 1000), (uint) (trie_size % 1000));
+
+  int t = PREFIX_BENCH_NUM / n;
+  int tb = MIN(t, TEST_BUFFER_SIZE);
+  nets = lp_alloc(lp0, tb * sizeof(net_addr));
+
+  if (!plus)
+    select_random_prefix_subset(data, nets, n, tb);
+  else
+    make_random_addresses(nets, tb);
+
+  bt_log_suite_case_result(1, "Make test data, %d (%d) tests", t, tb);
+  bt_reset_suite_case_timer();
+
+  /*
+  int match = 0;
+  for (int i = 0; i < t; i++)
+    for (int j = 0; j < n; j++)
+      test_match_net(data[j], trie[j], &nets[i]);
+  */
+
+  int match = 0;
+  for (int i = 0; i < t; i++)
+    for (int j = 0; j < n; j++)
+      if (trie_match_net(trie[j], &nets[i % TEST_BUFFER_SIZE]))
+	match++;
+
+  bt_log_suite_case_result(1, "Matching done, %d / %d matches", match, t * n);
+
+  rfree(lp0);
+  rfree(lp1);
+
+  return 1;
+}
+
+static int UNUSED
+t_bench_trie_datasets_subset(void)
+{
+  bt_bird_init();
+  bt_config_parse(BT_CONFIG_SIMPLE);
+
+  /* Specific datasets, not included */
+  benchmark_trie_dataset("trie-data-bgp-1", 0);
+  benchmark_trie_dataset("trie-data-bgp-10", 0);
+  benchmark_trie_dataset("trie-data-bgp-100", 0);
+  benchmark_trie_dataset("trie-data-bgp-1000", 0);
+
+  bt_bird_cleanup();
+
+  return 1;
+}
+
+static int UNUSED
+t_bench_trie_datasets_random(void)
+{
+  bt_bird_init();
+  bt_config_parse(BT_CONFIG_SIMPLE);
+
+  /* Specific datasets, not included */
+  benchmark_trie_dataset("trie-data-bgp-1", 1);
+  benchmark_trie_dataset("trie-data-bgp-10", 1);
+  benchmark_trie_dataset("trie-data-bgp-100", 1);
+  benchmark_trie_dataset("trie-data-bgp-1000", 1);
+
+  bt_bird_cleanup();
+
+  return 1;
+}
+
+
 static int
 t_trie_same(void)
 {
   bt_bird_init();
   bt_config_parse(BT_CONFIG_SIMPLE);
 
-  int round;
+  int v6 = 0;
-  for (round = 0; round < TESTS_NUM*4; round++)
+  linpool *lp = lp_new_default(&root_pool);
+  for (int round = 0; round < TESTS_NUM*4; round++)
   {
-    struct f_trie * trie1 = f_new_trie(config->mem, 0);
+    list *prefixes = make_random_prefix_list(lp, 100 * PREFIXES_NUM, v6, 0);
-    struct f_trie * trie2 = f_new_trie(config->mem, 0);
+    struct f_trie *trie1 = f_new_trie(lp, 0);
-
+    struct f_trie *trie2 = f_new_trie(lp, 0);
-    list prefixes; /* a list of f_extended_prefix structures */
-    init_list(&prefixes);
-    int i;
-    for (i = 0; i < 100; i++)
-      generate_random_ipv6_prefixes(&prefixes);
 
     struct f_prefix_node *n;
-    WALK_LIST(n, prefixes)
+    WALK_LIST(n, *prefixes)
-    {
       trie_add_prefix(trie1, &n->prefix.net, n->prefix.lo, n->prefix.hi);
-    }
+
-    WALK_LIST_BACKWARDS(n, prefixes)
+    WALK_LIST_BACKWARDS(n, *prefixes)
-    {
       trie_add_prefix(trie2, &n->prefix.net, n->prefix.lo, n->prefix.hi);
-    }
 
     bt_assert(trie_same(trie1, trie2));
 
-    struct f_prefix_node *nxt;
+    v6 = !v6;
-    WALK_LIST_DELSAFE(n, nxt, prefixes)
+    lp_flush(lp);
-    {
-      free(n);
-    }
   }
 
+  bt_bird_cleanup();
+  return 1;
+}
+
+static inline void
+log_networks(const net_addr *a, const net_addr *b)
+{
+  if (bt_verbose >= BT_VERBOSE_ABSOLUTELY_ALL)
+  {
+    char buf0[64];
+    char buf1[64];
+    bt_format_net(buf0, 64, a);
+    bt_format_net(buf1, 64, b);
+    bt_debug("Found %s expected %s\n", buf0, buf1);
+  }
+}
+
+static int
+t_trie_walk(void)
+{
+  bt_bird_init();
+  bt_config_parse(BT_CONFIG_SIMPLE);
+
+  linpool *lp = lp_new_default(&root_pool);
+  for (int round = 0; round < TESTS_NUM*8; round++)
+  {
+    int level = round / TESTS_NUM;
+    int v6 = level % 2;
+    int num = PREFIXES_NUM * (int[]){1, 10, 100, 1000}[level / 2];
+    int pos = 0, end = 0;
+    list *prefixes = make_random_prefix_list(lp, num, v6, 1);
+    struct f_trie *trie = make_trie_from_prefix_list(lp, prefixes);
+    struct f_prefix *pxset = malloc((num + 1) * sizeof(struct f_prefix));
+
+    struct f_prefix_node *n;
+    WALK_LIST(n, *prefixes)
+      pxset[pos++] = n->prefix;
+    memset(&pxset[pos], 0, sizeof (struct f_prefix));
+
+    qsort(pxset, num, sizeof(struct f_prefix), compare_prefixes);
+
+
+    /* Full walk */
+    bt_debug("Full walk (round %d, %d nets)\n", round, num);
+
+    pos = 0;
+    uint pxc = 0;
+    TRIE_WALK(trie, net, NULL)
+    {
+      log_networks(&net, &pxset[pos].net);
+      bt_assert(net_equal(&net, &pxset[pos].net));
+
+      /* Skip possible duplicates */
+      while (net_equal(&pxset[pos].net, &pxset[pos + 1].net))
+	pos++;
+
+      pos++;
+      pxc++;
+    }
+    TRIE_WALK_END;
+
+    bt_assert(pos == num);
+    bt_assert(pxc == trie->prefix_count);
+    bt_debug("Full walk done\n");
+
+
+    /* Prepare net for subnet walk - start with random prefix */
+    pos = bt_random() % num;
+    end = pos + (int[]){2, 2, 3, 4}[level / 2];
+    end = MIN(end, num);
+
+    struct f_prefix from = pxset[pos];
+
+    /* Find a common superprefix to several subsequent prefixes */
+    for (; pos < end; pos++)
+    {
+      if (net_equal(&from.net, &pxset[pos].net))
+	continue;
+
+      int common = !v6 ?
+	ip4_pxlen(net4_prefix(&from.net), net4_prefix(&pxset[pos].net)) :
+	ip6_pxlen(net6_prefix(&from.net), net6_prefix(&pxset[pos].net));
+      from.net.pxlen = MIN(from.net.pxlen, common);
+
+      if (!v6)
+	((net_addr_ip4 *) &from.net)->prefix =
+	  ip4_and(net4_prefix(&from.net), net4_prefix(&pxset[pos].net));
+      else
+	((net_addr_ip6 *) &from.net)->prefix =
+	  ip6_and(net6_prefix(&from.net), net6_prefix(&pxset[pos].net));
+    }
+
+    /* Fix irrelevant bits */
+    if (!v6)
+      ((net_addr_ip4 *) &from.net)->prefix =
+	ip4_and(net4_prefix(&from.net), ip4_mkmask(net4_pxlen(&from.net)));
+    else
+      ((net_addr_ip6 *) &from.net)->prefix =
+	ip6_and(net6_prefix(&from.net), ip6_mkmask(net6_pxlen(&from.net)));
+
+
+    /* Find initial position for final prefix */
+    for (pos = 0; pos < num; pos++)
+      if (compare_prefixes(&pxset[pos], &from) >= 0)
+	break;
+
+    int p0 = pos;
+    char buf0[64];
+    bt_format_net(buf0, 64, &from.net);
+    bt_debug("Subnet walk for %s (round %d, %d nets)\n", buf0, round, num);
+
+    /* Subnet walk */
+    TRIE_WALK(trie, net, &from.net)
+    {
+      log_networks(&net, &pxset[pos].net);
+      bt_assert(net_equal(&net, &pxset[pos].net));
+      bt_assert(net_in_netX(&net, &from.net));
+
+      /* Skip possible duplicates */
+      while (net_equal(&pxset[pos].net, &pxset[pos + 1].net))
+	pos++;
+
+      pos++;
+    }
+    TRIE_WALK_END;
+
+    bt_assert((pos == num) || !net_in_netX(&pxset[pos].net, &from.net));
+    bt_debug("Subnet walk done for %s (found %d nets)\n", buf0, pos - p0);
+
+    lp_flush(lp);
+  }
+
+  bt_bird_cleanup();
+  return 1;
+}
+
+static int
+find_covering_nets(struct f_prefix *prefixes, int num, const net_addr *net, net_addr *found)
+{
+  struct f_prefix key;
+  net_addr *n = &key.net;
+  int found_num = 0;
+
+  net_copy(n, net);
+
+  while (1)
+  {
+    struct f_prefix *px =
+      bsearch(&key, prefixes, num, sizeof(struct f_prefix), compare_prefixes);
+
+    if (px)
+    {
+      net_copy(&found[found_num], n);
+      found_num++;
+    }
+
+    if (n->pxlen == 0)
+      return found_num;
+
+    n->pxlen--;
+
+    if (n->type == NET_IP4)
+      ip4_clrbit(&((net_addr_ip4 *) n)->prefix, n->pxlen);
+    else
+      ip6_clrbit(&((net_addr_ip6 *) n)->prefix, n->pxlen);
+  }
+}
+
+static int
+t_trie_walk_to_root(void)
+{
+  bt_bird_init();
+  bt_config_parse(BT_CONFIG_SIMPLE);
+
+  linpool *lp = lp_new_default(&root_pool);
+  for (int round = 0; round < TESTS_NUM * 4; round++)
+  {
+    int level = round / TESTS_NUM;
+    int v6 = level % 2;
+    int num = PREFIXES_NUM  * (int[]){32, 512}[level / 2];
+    int pos = 0;
+    int st = 0, sn = 0, sm = 0;
+
+    list *prefixes = make_random_prefix_list(lp, num, v6, 1);
+    struct f_trie *trie = make_trie_from_prefix_list(lp, prefixes);
+    struct f_prefix *pxset = malloc((num + 1) * sizeof(struct f_prefix));
+
+    struct f_prefix_node *pxn;
+    WALK_LIST(pxn, *prefixes)
+      pxset[pos++] = pxn->prefix;
+    memset(&pxset[pos], 0, sizeof (struct f_prefix));
+
+    qsort(pxset, num, sizeof(struct f_prefix), compare_prefixes);
+
+    int i;
+    for (i = 0; i < (PREFIX_TESTS_NUM / 10); i++)
+    {
+      net_addr from;
+      get_random_net(&from, v6);
+
+      net_addr found[129];
+      int found_num = find_covering_nets(pxset, num, &from, found);
+      int n = 0;
+
+      if (bt_verbose >= BT_VERBOSE_ABSOLUTELY_ALL)
+      {
+	char buf[64];
+	bt_format_net(buf, 64, &from);
+	bt_debug("Lookup for %s (expect %d)\n", buf, found_num);
+      }
+
+      /* Walk to root, separate for IPv4 and IPv6 */
+      if (!v6)
+      {
+	TRIE_WALK_TO_ROOT_IP4(trie, (net_addr_ip4 *) &from, net)
+	{
+	  log_networks((net_addr *) &net, &found[n]);
+	  bt_assert((n < found_num) && net_equal((net_addr *) &net, &found[n]));
+	  n++;
+	}
+	TRIE_WALK_TO_ROOT_END;
+      }
+      else
+      {
+	TRIE_WALK_TO_ROOT_IP6(trie, (net_addr_ip6 *) &from, net)
+	{
+	  log_networks((net_addr *) &net, &found[n]);
+	  bt_assert((n < found_num) && net_equal((net_addr *) &net, &found[n]));
+	  n++;
+	}
+	TRIE_WALK_TO_ROOT_END;
+      }
+
+      bt_assert(n == found_num);
+
+      /* Stats */
+      st += n;
+      sn += !!n;
+      sm = MAX(sm, n);
+    }
+
+    bt_debug("Success in %d / %d, sum %d, max %d\n", sn, i, st, sm);
+
+    lp_flush(lp);
+  }
+
+  bt_bird_cleanup();
   return 1;
 }
 
@@ -179,8 +896,15 @@ main(int argc, char *argv[])
 {
   bt_init(argc, argv);
 
-  bt_test_suite(t_match_net, "Testing random prefix matching");
+  bt_test_suite(t_match_random_net, "Testing random prefix matching");
+  bt_test_suite(t_match_inner_net, "Testing random inner prefix matching");
+  bt_test_suite(t_match_outer_net, "Testing random outer prefix matching");
   bt_test_suite(t_trie_same, "A trie filled forward should be same with a trie filled backward.");
+  bt_test_suite(t_trie_walk, "Testing TRIE_WALK() on random tries");
+  bt_test_suite(t_trie_walk_to_root, "Testing TRIE_WALK_TO_ROOT() on random tries");
+
+  // bt_test_suite(t_bench_trie_datasets_subset, "Benchmark tries from datasets by random subset of nets");
+  // bt_test_suite(t_bench_trie_datasets_random, "Benchmark tries from datasets by generated addresses");
 
   return bt_exit_value();
 }

lib/birdlib.h

@@ -32,6 +32,9 @@ struct align_probe { char x; long int y; };
 #define MAX(a,b) MAX_(a,b)
 #endif
 
+#define ROUND_DOWN_POW2(a,b)  ((a) & ~((b)-1))
+#define ROUND_UP_POW2(a,b)  (((a)+((b)-1)) & ~((b)-1))
+
 #define U64(c) UINT64_C(c)
 #define ABS(a)   ((a)>=0 ? (a) : -(a))
 #define DELTA(a,b) (((a)>=(b))?(a)-(b):(b)-(a))

lib/ip.h

@@ -279,11 +279,35 @@ static inline uint ip6_pxlen(ip6_addr a, ip6_addr b)
   return 32 * i + 31 - u32_log2(a.addr[i] ^ b.addr[i]);
 }
 
+static inline int ip4_prefix_equal(ip4_addr a, ip4_addr b, uint n)
+{
+  return (_I(a) ^ _I(b)) < ((u64) 1 << (32 - n));
+}
+
+static inline int ip6_prefix_equal(ip6_addr a, ip6_addr b, uint n)
+{
+  uint n0 = n / 32;
+  uint n1 = n % 32;
+
+  return
+    ((n0 <= 0) || (_I0(a) == _I0(b))) &&
+    ((n0 <= 1) || (_I1(a) == _I1(b))) &&
+    ((n0 <= 2) || (_I2(a) == _I2(b))) &&
+    ((n0 <= 3) || (_I3(a) == _I3(b))) &&
+    (!n1 || ((a.addr[n0] ^ b.addr[n0]) < (1u << (32 - n1))));
+}
+
 static inline u32 ip4_getbit(ip4_addr a, uint pos)
-{ return _I(a) & (0x80000000 >> pos); }
+{ return (_I(a) >> (31 - pos)) & 1; }
+
+static inline u32 ip4_getbits(ip4_addr a, uint pos, uint n)
+{ return (_I(a) >> ((32 - n) - pos)) & ((1u << n) - 1); }
 
 static inline u32 ip6_getbit(ip6_addr a, uint pos)
-{ return a.addr[pos / 32] & (0x80000000 >> (pos % 32)); }
+{ return (a.addr[pos / 32] >> (31 - (pos % 32))) & 0x1; }
+
+static inline u32 ip6_getbits(ip6_addr a, uint pos, uint n)
+{ return (a.addr[pos / 32] >> ((32 - n) - (pos % 32))) & ((1u << n) - 1); }
 
 static inline u32 ip4_setbit(ip4_addr *a, uint pos)
 { return _I(*a) |= (0x80000000 >> pos); }
@@ -297,6 +321,13 @@ static inline u32 ip4_clrbit(ip4_addr *a, uint pos)
 static inline u32 ip6_clrbit(ip6_addr *a, uint pos)
 { return a->addr[pos / 32] &= ~(0x80000000 >> (pos % 32)); }
 
+static inline ip4_addr ip4_setbits(ip4_addr a, uint pos, uint val)
+{ _I(a) |= val << (31 - pos); return a; }
+
+static inline ip6_addr ip6_setbits(ip6_addr a, uint pos, uint val)
+{ a.addr[pos / 32] |= val << (31 - pos % 32); return a; }
+
+
 static inline ip4_addr ip4_opposite_m1(ip4_addr a)
 { return _MI4(_I(a) ^ 1); }
 

lib/ip_test.c

@@ -167,6 +167,70 @@ t_ip6_ntop(void)
   return bt_assert_batch(test_vectors, test_ipa_ntop, bt_fmt_ipa, bt_fmt_str);
 }
 
+static int
+t_ip4_prefix_equal(void)
+{
+  bt_assert( ip4_prefix_equal(ip4_from_u32(0x12345678), ip4_from_u32(0x1234ffff), 16));
+  bt_assert(!ip4_prefix_equal(ip4_from_u32(0x12345678), ip4_from_u32(0x1234ffff), 17));
+  bt_assert( ip4_prefix_equal(ip4_from_u32(0x12345678), ip4_from_u32(0x12345000), 21));
+  bt_assert(!ip4_prefix_equal(ip4_from_u32(0x12345678), ip4_from_u32(0x12345000), 22));
+
+  bt_assert( ip4_prefix_equal(ip4_from_u32(0x00000000), ip4_from_u32(0xffffffff),  0));
+  bt_assert( ip4_prefix_equal(ip4_from_u32(0x12345678), ip4_from_u32(0x12345678),  0));
+
+  bt_assert( ip4_prefix_equal(ip4_from_u32(0x12345678), ip4_from_u32(0x12345678),  32));
+  bt_assert(!ip4_prefix_equal(ip4_from_u32(0x12345678), ip4_from_u32(0x12345679),  32));
+  bt_assert(!ip4_prefix_equal(ip4_from_u32(0x12345678), ip4_from_u32(0x92345678),  32));
+
+  return 1;
+}
+
+static int
+t_ip6_prefix_equal(void)
+{
+  bt_assert( ip6_prefix_equal(ip6_build(0x20010db8, 0x12345678, 0x10101010, 0x20202020),
+			      ip6_build(0x20010db8, 0x1234ffff, 0xfefefefe, 0xdcdcdcdc),
+			      48));
+
+  bt_assert(!ip6_prefix_equal(ip6_build(0x20010db8, 0x12345678, 0x10101010, 0x20202020),
+			      ip6_build(0x20010db8, 0x1234ffff, 0xfefefefe, 0xdcdcdcdc),
+			      49));
+
+  bt_assert(!ip6_prefix_equal(ip6_build(0x20010db8, 0x12345678, 0x10101010, 0x20202020),
+			      ip6_build(0x20020db8, 0x12345678, 0xfefefefe, 0xdcdcdcdc),
+			      48));
+
+  bt_assert( ip6_prefix_equal(ip6_build(0x20010db8, 0x12345678, 0x10101010, 0x20202020),
+			      ip6_build(0x20010db8, 0x12345678, 0xfefefefe, 0xdcdcdcdc),
+			      64));
+
+  bt_assert(!ip6_prefix_equal(ip6_build(0x20010db8, 0x12345678, 0x10101010, 0x20202020),
+			      ip6_build(0x20010db8, 0x1234567e, 0xfefefefe, 0xdcdcdcdc),
+			      64));
+
+  bt_assert( ip6_prefix_equal(ip6_build(0x20010db8, 0x12345678, 0x10101010, 0x20002020),
+			      ip6_build(0x20010db8, 0x12345678, 0x10101010, 0x20202020),
+			      106));
+
+  bt_assert(!ip6_prefix_equal(ip6_build(0x20010db8, 0x12345678, 0x10101010, 0x20002020),
+			      ip6_build(0x20010db8, 0x12345678, 0x10101010, 0x20202020),
+			      107));
+
+  bt_assert( ip6_prefix_equal(ip6_build(0xfeef0db8, 0x87654321, 0x10101010, 0x20202020),
+			      ip6_build(0x20010db8, 0x12345678, 0xfefefefe, 0xdcdcdcdc),
+			      0));
+
+  bt_assert( ip6_prefix_equal(ip6_build(0x20010db8, 0x12345678, 0x10101010, 0x20202020),
+			      ip6_build(0x20010db8, 0x12345678, 0x10101010, 0x20202020),
+			      128));
+
+  bt_assert(!ip6_prefix_equal(ip6_build(0x20010db8, 0x12345678, 0x10101010, 0x20202020),
+			      ip6_build(0x20010db8, 0x12345678, 0x10101010, 0x20202021),
+			      128));
+
+  return 1;
+}
+
 int
 main(int argc, char *argv[])
 {
@@ -176,6 +240,8 @@ main(int argc, char *argv[])
   bt_test_suite(t_ip6_pton, "Converting IPv6 string to ip6_addr struct");
   bt_test_suite(t_ip4_ntop, "Converting ip4_addr struct to IPv4 string");
   bt_test_suite(t_ip6_ntop, "Converting ip6_addr struct to IPv6 string");
+  bt_test_suite(t_ip4_prefix_equal, "Testing ip4_prefix_equal()");
+  bt_test_suite(t_ip6_prefix_equal, "Testing ip6_prefix_equal()");
 
   return bt_exit_value();
 }

lib/net.h

@@ -38,6 +38,7 @@
 
 #define NB_IP		(NB_IP4 | NB_IP6)
 #define NB_VPN		(NB_VPN4 | NB_VPN6)
+#define NB_ROA		(NB_ROA4 | NB_ROA6)
 #define NB_FLOW		(NB_FLOW4 | NB_FLOW6)
 #define NB_DEST		(NB_IP | NB_IP6_SADR | NB_VPN | NB_MPLS)
 #define NB_ANY		0xffffffff

nest/config.Y

@@ -17,6 +17,7 @@ CF_HDR
 
 CF_DEFINES
 
+static struct rtable_config *this_table;
 static struct proto_config *this_proto;
 static struct channel_config *this_channel;
 static struct iface_patt *this_ipatt;
@@ -117,13 +118,14 @@ CF_KEYWORDS(IPV4, IPV6, VPN4, VPN6, ROA4, ROA6, FLOW4, FLOW6, SADR, MPLS)
 CF_KEYWORDS(RECEIVE, LIMIT, ACTION, WARN, BLOCK, RESTART, DISABLE, KEEP, FILTERED, RPKI)
 CF_KEYWORDS(PASSWORD, KEY, FROM, PASSIVE, TO, ID, EVENTS, PACKETS, PROTOCOLS, CHANNELS, INTERFACES)
 CF_KEYWORDS(ALGORITHM, KEYED, HMAC, MD5, SHA1, SHA256, SHA384, SHA512, BLAKE2S128, BLAKE2S256, BLAKE2B256, BLAKE2B512)
-CF_KEYWORDS(PRIMARY, STATS, COUNT, BY, FOR, COMMANDS, PREEXPORT, NOEXPORT, EXPORTED, GENERATE)
+CF_KEYWORDS(PRIMARY, STATS, COUNT, BY, FOR, IN, COMMANDS, PREEXPORT, NOEXPORT, EXPORTED, GENERATE)
-CF_KEYWORDS(BGP, PASSWORDS, DESCRIPTION, SORTED)
+CF_KEYWORDS(BGP, PASSWORDS, DESCRIPTION)
 CF_KEYWORDS(RELOAD, IN, OUT, MRTDUMP, MESSAGES, RESTRICT, MEMORY, IGP_METRIC, CLASS, DSCP)
 CF_KEYWORDS(TIMEFORMAT, ISO, SHORT, LONG, ROUTE, PROTOCOL, BASE, LOG, S, MS, US)
 CF_KEYWORDS(GRACEFUL, RESTART, WAIT, MAX, FLUSH, AS)
 CF_KEYWORDS(MIN, IDLE, RX, TX, INTERVAL, MULTIPLIER, PASSIVE)
 CF_KEYWORDS(CHECK, LINK)
+CF_KEYWORDS(SORTED, TRIE, MIN, MAX, SETTLE, TIME)
 
 /* For r_args_channel */
 CF_KEYWORDS(IPV4, IPV4_MC, IPV4_MPLS, IPV6, IPV6_MC, IPV6_MPLS, IPV6_SADR, VPN4, VPN4_MC, VPN4_MPLS, VPN6, VPN6_MC, VPN6_MPLS, ROA4, ROA6, FLOW4, FLOW6, MPLS, PRI, SEC)
@@ -141,7 +143,7 @@ CF_ENUM_PX(T_ENUM_AF, AF_, AFI_, IPV4, IPV6)
 %type <s> optproto
 %type <ra> r_args
 %type <sd> sym_args
-%type <i> proto_start echo_mask echo_size debug_mask debug_list debug_flag mrtdump_mask mrtdump_list mrtdump_flag export_mode limit_action net_type table_sorted tos password_algorithm
+%type <i> proto_start echo_mask echo_size debug_mask debug_list debug_flag mrtdump_mask mrtdump_list mrtdump_flag export_mode limit_action net_type tos password_algorithm
 %type <ps> proto_patt proto_patt2
 %type <cc> channel_start proto_channel
 %type <cl> limit_spec
@@ -206,16 +208,37 @@ CF_ENUM(T_ENUM_NETTYPE, NET_, IP4, IP6, VPN4, VPN6, ROA4, ROA6, FLOW4, FLOW6, IP
 
 conf: table ;
 
-table_sorted:
+table: table_start table_sorted table_opt_list ;
-	  { $$ = 0; }
+
- | SORTED { $$ = 1; }
+table_start: net_type TABLE symbol {
+   this_table = rt_new_table($3, $1);
+   }
  ;
 
-table: net_type TABLE symbol table_sorted {
+table_sorted:
-   struct rtable_config *cf;
+   /* empty */
-   cf = rt_new_table($3, $1);
+ | SORTED { this_table->sorted = 1; }
-   cf->sorted = $4;
+ ;
+
+table_opt:
+   SORTED bool { this_table->sorted = $2; }
+ | TRIE bool {
+     if (!net_val_match(this_table->addr_type, NB_IP | NB_VPN | NB_ROA | NB_IP6_SADR))
+       cf_error("Trie option not supported for %s table", net_label[this_table->addr_type]);
+     this_table->trie_used = $2;
    }
+ | MIN SETTLE TIME expr_us { this_table->min_settle_time = $4; }
+ | MAX SETTLE TIME expr_us { this_table->max_settle_time = $4; }
+ ;
+
+table_opts:
+   /* empty */
+ | table_opts table_opt ';'
+ ;
+
+table_opt_list:
+   /* empty */
+ | '{' table_opts '}'
  ;
 
 
@@ -621,14 +644,22 @@ r_args:
      $$ = $1;
      if ($$->addr) cf_error("Only one prefix expected");
      $$->addr = $2;
+     $$->addr_mode = RSD_ADDR_EQUAL;
    }
  | r_args FOR r_args_for {
      $$ = $1;
      if ($$->addr) cf_error("Only one prefix expected");
-     $$->show_for = 1;
      $$->addr = $3;
+     $$->addr_mode = RSD_ADDR_FOR;
    }
- | r_args TABLE CF_SYM_KNOWN {
+ | r_args IN net_any {
+     $$ = $1;
+     if ($$->addr) cf_error("Only one prefix expected");
+     if (!net_type_match($3, NB_IP)) cf_error("Only IP networks accepted for 'in' argument");
+     $$->addr = $3;
+     $$->addr_mode = RSD_ADDR_IN;
+   }
+| r_args TABLE CF_SYM_KNOWN {
      cf_assert_symbol($3, SYM_TABLE);
      $$ = $1;
      rt_show_add_table($$, $3->table->table);

nest/route.h

@@ -20,7 +20,10 @@ struct proto;
 struct rte_src;
 struct symbol;
 struct timer;
+struct fib;
 struct filter;
+struct f_trie;
+struct f_trie_walk_state;
 struct cli;
 
 /*
@@ -49,7 +52,7 @@ struct fib_iterator {			/* See lib/slists.h for an explanation */
   uint hash;
 };
 
-typedef void (*fib_init_fn)(void *);
+typedef void (*fib_init_fn)(struct fib *, void *);
 
 struct fib {
   pool *fib_pool;			/* Pool holding all our data */
@@ -149,6 +152,7 @@ struct rtable_config {
   int gc_min_time;			/* Minimum time between two consecutive GC runs */
   byte sorted;				/* Routes of network are sorted according to rte_better() */
   byte internal;			/* Internal table of a protocol */
+  byte trie_used;			/* Rtable has attached trie */
   btime min_settle_time;		/* Minimum settle time for notifications */
   btime max_settle_time;		/* Maximum settle time for notifications */
 };
@@ -158,6 +162,7 @@ typedef struct rtable {
   node n;				/* Node in list of all tables */
   pool *rp;				/* Resource pool to allocate everything from, including itself */
   struct fib fib;
+  struct f_trie *trie;			/* Trie of prefixes defined in fib */
   char *name;				/* Name of this table */
   list channels;			/* List of attached channels (struct channel) */
   uint addr_type;			/* Type of address data stored in table (NET_*) */
@@ -180,13 +185,20 @@ 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 */
+  struct f_trie *trie_old;		/* Old prefix trie waiting to be freed */
+  u32 trie_lock_count;			/* Prefix trie locked by walks */
+  u32 trie_old_lock_count;		/* Old prefix trie locked by walks */
 
   list subscribers;			/* Subscribers for notifications */
   struct timer *settle_timer;		/* Settle time for notifications */
+  list flowspec_links;			/* List of flowspec links, src for NET_IPx and dst for NET_FLOWx */
+  struct f_trie *flowspec_trie;		/* Trie for evaluation of flowspec notifications */
 } rtable;
 
 struct rt_subscription {
@@ -196,6 +208,13 @@ struct rt_subscription {
   void *data;
 };
 
+struct rt_flowspec_link {
+  node n;
+  rtable *src;
+  rtable *dst;
+  u32 uc;
+};
+
 #define NHU_CLEAN	0
 #define NHU_SCHEDULED	1
 #define NHU_RUNNING	2
@@ -262,6 +281,7 @@ typedef struct rte {
     struct {
       u8 suppressed;			/* Used for deterministic MED comparison */
       s8 stale;				/* Route is LLGR_STALE, -1 if unknown */
+      struct rtable *base_table;	/* Base table for Flowspec validation */
     } bgp;
 #endif
 #ifdef CONFIG_BABEL
@@ -315,8 +335,12 @@ void rt_preconfig(struct config *);
 void rt_commit(struct config *new, struct config *old);
 void rt_lock_table(rtable *);
 void rt_unlock_table(rtable *);
+struct f_trie * rt_lock_trie(rtable *tab);
+void rt_unlock_trie(rtable *tab, struct f_trie *trie);
 void rt_subscribe(rtable *tab, struct rt_subscription *s);
 void rt_unsubscribe(struct rt_subscription *s);
+void rt_flowspec_link(rtable *src, rtable *dst);
+void rt_flowspec_unlink(rtable *src, rtable *dst);
 rtable *rt_setup(pool *, struct rtable_config *);
 static inline void rt_shutdown(rtable *r) { rfree(r->rp); }
 
@@ -324,7 +348,8 @@ static inline net *net_find(rtable *tab, const net_addr *addr) { return (net *)
 static inline net *net_find_valid(rtable *tab, const net_addr *addr)
 { net *n = net_find(tab, addr); return (n && rte_is_valid(n->routes)) ? n : NULL; }
 static inline net *net_get(rtable *tab, const net_addr *addr) { return (net *) fib_get(&tab->fib, addr); }
-void *net_route(rtable *tab, const net_addr *n);
+net *net_get(rtable *tab, const net_addr *addr);
+net *net_route(rtable *tab, const net_addr *n);
 int net_roa_check(rtable *tab, const net_addr *n, u32 asn);
 rte *rte_find(net *net, struct rte_src *src);
 rte *rte_get_temp(struct rta *);
@@ -356,6 +381,18 @@ void rt_prune_sync(rtable *t, int all);
 int rte_update_out(struct channel *c, const net_addr *n, rte *new, rte *old0, int refeed);
 struct rtable_config *rt_new_table(struct symbol *s, uint addr_type);
 
+static inline int rt_is_ip(rtable *tab)
+{ return (tab->addr_type == NET_IP4) || (tab->addr_type == NET_IP6); }
+
+static inline int rt_is_vpn(rtable *tab)
+{ return (tab->addr_type == NET_VPN4) || (tab->addr_type == NET_VPN6); }
+
+static inline int rt_is_roa(rtable *tab)
+{ return (tab->addr_type == NET_ROA4) || (tab->addr_type == NET_ROA6); }
+
+static inline int rt_is_flow(rtable *tab)
+{ return (tab->addr_type == NET_FLOW4) || (tab->addr_type == NET_FLOW6); }
+
 
 /* Default limit for ECMP next hops, defined in sysdep code */
 extern const int rt_default_ecmp;
@@ -372,6 +409,8 @@ struct rt_show_data {
   struct rt_show_data_rtable *tab;	/* Iterator over table list */
   struct rt_show_data_rtable *last_table; /* Last table in output */
   struct fib_iterator fit;		/* Iterator over networks in table */
+  struct f_trie_walk_state *walk_state;	/* Iterator over networks in trie */
+  struct f_trie *walk_lock;		/* Locked trie for walking */
   int verbose, tables_defined_by;
   const struct filter *filter;
   struct proto *show_protocol;
@@ -379,9 +418,10 @@ struct rt_show_data {
   struct channel *export_channel;
   struct config *running_on_config;
   struct krt_proto *kernel;
-  int export_mode, primary_only, filtered, stats, show_for;
+  int export_mode, addr_mode, primary_only, filtered, stats;
 
   int table_open;			/* Iteration (fit) is open */
+  int trie_walk;			/* Current table is iterated using trie */
   int net_counter, rt_counter, show_counter, table_counter;
   int net_counter_last, rt_counter_last, show_counter_last;
 };
@@ -398,6 +438,11 @@ struct rt_show_data_rtable * rt_show_add_table(struct rt_show_data *d, rtable *t
 #define RSD_TDB_SET	  0x1		/* internal: show empty tables */
 #define RSD_TDB_NMN	  0x2		/* internal: need matching net */
 
+/* Value of addr_mode */
+#define RSD_ADDR_EQUAL	1		/* Exact query - show route <addr> */
+#define RSD_ADDR_FOR	2		/* Longest prefix match - show route for <addr> */
+#define RSD_ADDR_IN	3		/* Interval query - show route in <addr> */
+
 /* Value of export_mode in struct rt_show_data */
 #define RSEM_NONE	0		/* Export mode not used */
 #define RSEM_PREEXPORT	1		/* Routes ready for export, before filtering */
@@ -718,6 +763,9 @@ rta_set_recursive_next_hop(rtable *dep, rta *a, rtable *tab, ip_addr gw, ip_addr
 static inline void rt_lock_hostentry(struct hostentry *he) { if (he) he->uc++; }
 static inline void rt_unlock_hostentry(struct hostentry *he) { if (he) he->uc--; }
 
+int rt_flowspec_check(rtable *tab_ip, rtable *tab_flow, const net_addr *n, rta *a, int interior);
+
+
 /*
  *	Default protocol preferences
  */

nest/rt-fib.c

@@ -331,7 +331,7 @@ fib_get(struct fib *f, const net_addr *a)
 
   memset(b, 0, f->node_offset);
   if (f->init)
-    f->init(b);
+    f->init(f, b);
 
   if (f->entries++ > f->entries_max)
     fib_rehash(f, HASH_HI_STEP);

nest/rt-show.c

@@ -15,6 +15,7 @@
 #include "nest/cli.h"
 #include "nest/iface.h"
 #include "filter/filter.h"
+#include "filter/data.h"
 #include "sysdep/unix/krt.h"
 
 static void
@@ -110,10 +111,9 @@ rt_show_net(struct cli *c, net *n, struct rt_show_data *d)
   ASSUME(!d->export_mode || ec);
 
   int first = 1;
+  int first_show = 1;
   int pass = 0;
 
-  bsnprintf(ia, sizeof(ia), "%N", n->n.addr);
-
   for (e = n->routes; e; e = e->next)
     {
       if (rte_is_filtered(e) != d->filtered)
@@ -187,10 +187,17 @@ rt_show_net(struct cli *c, net *n, struct rt_show_data *d)
 	goto skip;
 
       if (d->stats < 2)
+      {
+	if (first_show)
+	  net_format(n->n.addr, ia, sizeof(ia));
+	else
+	  ia[0] = 0;
+
 	rt_show_rte(c, ia, e, d, (e->net->routes == ee));
+	first_show = 0;
+      }
 
       d->show_counter++;
-      ia[0] = 0;
 
     skip:
       if (e != ee)
@@ -212,9 +219,12 @@ rt_show_cleanup(struct cli *c)
   struct rt_show_data_rtable *tab;
 
   /* Unlink the iterator */
-  if (d->table_open)
+  if (d->table_open && !d->trie_walk)
     fit_get(&d->tab->table->fib, &d->fit);
 
+  if (d->walk_lock)
+    rt_unlock_trie(d->tab->table, d->walk_lock);
+
   /* Unlock referenced tables */
   WALK_LIST(tab, d->tables)
     rt_unlock_table(tab->table);
@@ -224,12 +234,13 @@ static void
 rt_show_cont(struct cli *c)
 {
   struct rt_show_data *d = c->rover;
+  struct rtable *tab = d->tab->table;
 #ifdef DEBUGGING
   unsigned max = 4;
 #else
   unsigned max = 64;
 #endif
-  struct fib *fib = &d->tab->table->fib;
+  struct fib *fib = &tab->fib;
   struct fib_iterator *it = &d->fit;
 
   if (d->running_on_config && (d->running_on_config != config))
@@ -240,7 +251,22 @@ rt_show_cont(struct cli *c)
 
   if (!d->table_open)
   {
-    FIB_ITERATE_INIT(&d->fit, &d->tab->table->fib);
+    /* We use either trie-based walk or fib-based walk */
+    d->trie_walk = tab->trie &&
+      (d->addr_mode == RSD_ADDR_IN) &&
+      net_val_match(tab->addr_type, NB_IP);
+
+    if (d->trie_walk && !d->walk_state)
+      d->walk_state = lp_allocz(c->parser_pool, sizeof (struct f_trie_walk_state));
+
+    if (d->trie_walk)
+    {
+      d->walk_lock = rt_lock_trie(tab);
+      trie_walk_init(d->walk_state, tab->trie, d->addr);
+    }
+    else
+      FIB_ITERATE_INIT(&d->fit, &tab->fib);
+
     d->table_open = 1;
     d->table_counter++;
     d->kernel = rt_show_get_kernel(d);
@@ -253,16 +279,44 @@ rt_show_cont(struct cli *c)
       rt_show_table(c, d);
   }
 
-  FIB_ITERATE_START(fib, it, net, n)
+  if (d->trie_walk)
   {
-    if (!max--)
+    /* Trie-based walk */
+    net_addr addr;
+    while (trie_walk_next(d->walk_state, &addr))
     {
-      FIB_ITERATE_PUT(it);
+      net *n = net_find(tab, &addr);
-      return;
+      if (!n)
+	continue;
+
+      rt_show_net(c, n, d);
+
+      if (!--max)
+	return;
     }
-    rt_show_net(c, n, d);
+
+    rt_unlock_trie(tab, d->walk_lock);
+    d->walk_lock = NULL;
+  }
+  else
+  {
+    /* fib-based walk */
+    FIB_ITERATE_START(fib, it, net, n)
+    {
+      if ((d->addr_mode == RSD_ADDR_IN) && (!net_in_netX(n->n.addr, d->addr)))
+	goto next;
+
+      if (!max--)
+      {
+	FIB_ITERATE_PUT(it);
+	return;
+      }
+      rt_show_net(c, n, d);
+
+    next:;
+    }
+    FIB_ITERATE_END;
   }
-  FIB_ITERATE_END;
 
   if (d->stats)
   {
@@ -271,7 +325,7 @@ rt_show_cont(struct cli *c)
 
     cli_printf(c, -1007, "%d of %d routes for %d networks in table %s",
 	       d->show_counter - d->show_counter_last, d->rt_counter - d->rt_counter_last,
-	       d->net_counter - d->net_counter_last, d->tab->table->name);
+	       d->net_counter - d->net_counter_last, tab->name);
   }
 
   d->kernel = NULL;
@@ -402,7 +456,7 @@ rt_show(struct rt_show_data *d)
 
   rt_show_prepare_tables(d);
 
-  if (!d->addr)
+  if (!d->addr || (d->addr_mode == RSD_ADDR_IN))
   {
     WALK_LIST(tab, d->tables)
       rt_lock_table(tab->table);
@@ -420,7 +474,7 @@ rt_show(struct rt_show_data *d)
       d->tab = tab;
       d->kernel = rt_show_get_kernel(d);
 
-      if (d->show_for)
+      if (d->addr_mode == RSD_ADDR_FOR)
 	n = net_route(tab->table, d->addr);
       else
 	n = net_find(tab->table, d->addr);

proto/babel/babel.c

@@ -63,7 +63,7 @@ static inline void babel_iface_kick_timer(struct babel_iface *ifa);
  */
 
 static void
-babel_init_entry(void *E)
+babel_init_entry(struct fib *f UNUSED, void *E)
 {
   struct babel_entry *e = E;
 

proto/bgp/attrs.c

@@ -1683,6 +1683,10 @@ bgp_preexport(struct proto *P, rte **new, struct linpool *pool UNUSED)
   if (src == NULL)
     return 0;
 
+  /* Reject flowspec that failed validation */
+  if ((e->attrs->dest == RTD_UNREACHABLE) && net_is_flow(e->net->n.addr))
+      return -1;
+
   /* IBGP route reflection, RFC 4456 */
   if (p->is_internal && src->is_internal && (p->local_as == src->local_as))
   {

proto/bgp/bgp.c

@@ -101,6 +101,7 @@
  * RFC 8203 - BGP Administrative Shutdown Communication
  * RFC 8212 - Default EBGP Route Propagation Behavior without Policies
  * RFC 8654 - Extended Message Support for BGP
+ * RFC 9117 - Revised Validation Procedure for BGP Flow Specifications
  * draft-ietf-idr-ext-opt-param-07
  * draft-uttaro-idr-bgp-persistence-04
  * draft-walton-bgp-hostname-capability-02
@@ -1740,6 +1741,9 @@ bgp_channel_init(struct channel *C, struct channel_config *CF)
 
   if (cf->igp_table_ip6)
     c->igp_table_ip6 = cf->igp_table_ip6->table;
+
+  if (cf->base_table)
+    c->base_table = cf->base_table->table;
 }
 
 static int
@@ -1755,6 +1759,12 @@ bgp_channel_start(struct channel *C)
   if (c->igp_table_ip6)
     rt_lock_table(c->igp_table_ip6);
 
+  if (c->base_table)
+  {
+    rt_lock_table(c->base_table);
+    rt_flowspec_link(c->base_table, c->c.table);
+  }
+
   c->pool = p->p.pool; // XXXX
   bgp_init_bucket_table(c);
   bgp_init_prefix_table(c);
@@ -1839,6 +1849,12 @@ bgp_channel_cleanup(struct channel *C)
   if (c->igp_table_ip6)
     rt_unlock_table(c->igp_table_ip6);
 
+  if (c->base_table)
+  {
+    rt_flowspec_unlink(c->base_table, c->c.table);
+    rt_unlock_table(c->base_table);
+  }
+
   c->index = 0;
 
   /* Cleanup rest of bgp_channel starting at pool field */
@@ -1886,6 +1902,25 @@ bgp_default_igp_table(struct bgp_config *cf, struct bgp_channel_config *cc, u32
   cf_error("Undefined IGP table");
 }
 
+static struct rtable_config *
+bgp_default_base_table(struct bgp_config *cf, struct bgp_channel_config *cc)
+{
+  /* Expected table type */
+  u32 type = (cc->afi == BGP_AF_FLOW4) ? NET_IP4 : NET_IP6;
+
+  /* First, try appropriate IP channel */
+  u32 afi2 = BGP_AF(BGP_AFI(cc->afi), BGP_SAFI_UNICAST);
+  struct bgp_channel_config *cc2 = bgp_find_channel_config(cf, afi2);
+  if (cc2 && (cc2->c.table->addr_type == type))
+    return cc2->c.table;
+
+  /* Last, try default table of given type */
+  struct rtable_config *tab = cf->c.global->def_tables[type];
+  if (tab)
+    return tab;
+
+  cf_error("Undefined base table");
+}
 
 void
 bgp_postconfig(struct proto_config *CF)
@@ -2030,6 +2065,14 @@ bgp_postconfig(struct proto_config *CF)
 	cf_error("Mismatched IGP table type");
     }
 
+    /* Default value of base table */
+    if ((BGP_SAFI(cc->afi) == BGP_SAFI_FLOW) && cc->validate && !cc->base_table)
+      cc->base_table = bgp_default_base_table(cf, cc);
+
+    if (cc->base_table && !cc->base_table->trie_used)
+      cf_error("Flowspec validation requires base table (%s) with trie",
+	       cc->base_table->name);
+
     if (cf->multihop && (cc->gw_mode == GW_DIRECT))
       cf_error("Multihop BGP cannot use direct gateway mode");
 
@@ -2098,7 +2141,7 @@ bgp_reconfigure(struct proto *P, struct proto_config *CF)
   return same;
 }
 
-#define IGP_TABLE(cf, sym) ((cf)->igp_table_##sym ? (cf)->igp_table_##sym ->table : NULL )
+#define TABLE(cf, NAME) ((cf)->NAME ? (cf)->NAME->table : NULL )
 
 static int
 bgp_channel_reconfigure(struct channel *C, struct channel_config *CC, int *import_changed, int *export_changed)
@@ -2109,6 +2152,7 @@ bgp_channel_reconfigure(struct channel *C, struct channel_config *CC, int *impor
   struct bgp_channel_config *old = c->cf;
 
   if ((new->secondary != old->secondary) ||
+      (new->validate != old->validate) ||
       (new->gr_able != old->gr_able) ||
       (new->llgr_able != old->llgr_able) ||
       (new->llgr_time != old->llgr_time) ||
@@ -2116,8 +2160,9 @@ bgp_channel_reconfigure(struct channel *C, struct channel_config *CC, int *impor
       (new->add_path != old->add_path) ||
       (new->import_table != old->import_table) ||
       (new->export_table != old->export_table) ||
-      (IGP_TABLE(new, ip4) != IGP_TABLE(old, ip4)) ||
+      (TABLE(new, igp_table_ip4) != TABLE(old, igp_table_ip4)) ||
-      (IGP_TABLE(new, ip6) != IGP_TABLE(old, ip6)))
+      (TABLE(new, igp_table_ip6) != TABLE(old, igp_table_ip6)) ||
+      (TABLE(new, base_table) != TABLE(old, base_table)))
     return 0;
 
   if (new->mandatory && !old->mandatory && (C->channel_state != CS_UP))
@@ -2531,6 +2576,9 @@ bgp_show_proto_info(struct proto *P)
 
       if (c->igp_table_ip6)
 	cli_msg(-1006, "    IGP IPv6 table: %s", c->igp_table_ip6->name);
+
+      if (c->base_table)
+	cli_msg(-1006, "    Base table:     %s", c->base_table->name);
     }
   }
 }

proto/bgp/bgp.h

@@ -147,6 +147,7 @@ struct bgp_channel_config {
   u8 mandatory;				/* Channel is mandatory in capability negotiation */
   u8 gw_mode;				/* How we compute route gateway from next_hop attr, see GW_* */
   u8 secondary;				/* Accept also non-best routes (i.e. RA_ACCEPTED) */
+  u8 validate;				/* Validate Flowspec per RFC 8955 (6) */
   u8 gr_able;				/* Allow full graceful restart for the channel */
   u8 llgr_able;				/* Allow full long-lived GR for the channel */
   uint llgr_time;			/* Long-lived graceful restart stale time */
@@ -160,6 +161,7 @@ struct bgp_channel_config {
 
   struct rtable_config *igp_table_ip4;	/* Table for recursive IPv4 next hop lookups */
   struct rtable_config *igp_table_ip6;	/* Table for recursive IPv6 next hop lookups */
+  struct rtable_config *base_table;	/* Base table for Flowspec validation */
 };
 
 #define BGP_PT_INTERNAL		1
@@ -341,6 +343,7 @@ struct bgp_channel {
 
   rtable *igp_table_ip4;		/* Table for recursive IPv4 next hop lookups */
   rtable *igp_table_ip6;		/* Table for recursive IPv6 next hop lookups */
+  rtable *base_table;			/* Base table for Flowspec validation */
 
   /* Rest are zeroed when down */
   pool *pool;
@@ -451,6 +454,7 @@ struct bgp_parse_state {
   jmp_buf err_jmpbuf;
 
   struct hostentry *hostentry;
+  struct rtable *base_table;
   adata *mpls_labels;
 
   /* Cached state for bgp_rte_update() */
@@ -517,7 +521,7 @@ struct rte_source *bgp_get_source(struct bgp_proto *p, u32 path_id);
 static inline int
 rte_resolvable(rte *rt)
 {
-  return rt->attrs->dest == RTD_UNICAST;
+  return rt->attrs->dest != RTD_UNREACHABLE;
 }
 
 

proto/bgp/config.Y

@@ -31,7 +31,7 @@ CF_KEYWORDS(BGP, LOCAL, NEIGHBOR, AS, HOLD, TIME, CONNECT, RETRY, KEEPALIVE,
 	STRICT, BIND, CONFEDERATION, MEMBER, MULTICAST, FLOW4, FLOW6, LONG,
 	LIVED, STALE, IMPORT, IBGP, EBGP, MANDATORY, INTERNAL, EXTERNAL, SETS,
 	DYNAMIC, RANGE, NAME, DIGITS, BGP_AIGP, AIGP, ORIGINATE, COST, ENFORCE,
-	FIRST, FREE)
+	FIRST, FREE, VALIDATE, BASE)
 
 %type <i> bgp_nh
 %type <i32> bgp_afi
@@ -256,6 +256,11 @@ bgp_channel_item:
  | GATEWAY DIRECT { BGP_CC->gw_mode = GW_DIRECT; }
  | GATEWAY RECURSIVE { BGP_CC->gw_mode = GW_RECURSIVE; }
  | SECONDARY bool { BGP_CC->secondary = $2; }
+ | VALIDATE bool {
+    BGP_CC->validate = $2;
+    if (BGP_SAFI(BGP_CC->afi) != BGP_SAFI_FLOW)
+      cf_error("Validate option limited to flowspec channels");
+   }
  | GRACEFUL RESTART bool { BGP_CC->gr_able = $3; }
  | LONG LIVED GRACEFUL RESTART bool { BGP_CC->llgr_able = $5; }
  | LONG LIVED STALE TIME expr { BGP_CC->llgr_time = $5; }
@@ -279,6 +284,16 @@ bgp_channel_item:
     else
       cf_error("Mismatched IGP table type");
    }
+ | BASE TABLE rtable {
+    if (BGP_SAFI(BGP_CC->afi) != BGP_SAFI_FLOW)
+      cf_error("Base table option limited to flowspec channels");
+
+    if (((BGP_CC->afi == BGP_AF_FLOW4) && ($3->addr_type == NET_IP4)) ||
+        ((BGP_CC->afi == BGP_AF_FLOW6) && ($3->addr_type == NET_IP6)))
+      BGP_CC->base_table = $3;
+    else
+      cf_error("Mismatched base table type");
+   }
  ;
 
 bgp_channel_opts:

proto/bgp/packets.c

@@ -1018,6 +1018,23 @@ bgp_apply_mpls_labels(struct bgp_parse_state *s, rta *a, u32 *labels, uint lnum)
   }
 }
 
+static void
+bgp_apply_flow_validation(struct bgp_parse_state *s, const net_addr *n, rta *a)
+{
+  struct bgp_channel *c = s->channel;
+  int valid = rt_flowspec_check(c->base_table, c->c.table, n, a, s->proto->is_interior);
+  a->dest = valid ? RTD_NONE : RTD_UNREACHABLE;
+
+  /* Set rte.bgp.base_table later from this state variable */
+  s->base_table = c->base_table;
+
+  /* Invalidate cached rta if dest changes */
+  if (s->cached_rta && (s->cached_rta->dest != a->dest))
+  {
+    rta_free(s->cached_rta);
+    s->cached_rta = NULL;
+  }
+}
 
 static int
 bgp_match_src(struct bgp_export_state *s, int mode)
@@ -1383,6 +1400,7 @@ bgp_rte_update(struct bgp_parse_state *s, const net_addr *n, u32 path_id, rta *a
   e->pflags = 0;
   e->u.bgp.suppressed = 0;
   e->u.bgp.stale = -1;
+  e->u.bgp.base_table = s->base_table;
   rte_update3(&s->channel->c, n, e, s->last_src);
 }
 
@@ -1897,6 +1915,10 @@ bgp_decode_nlri_flow4(struct bgp_parse_state *s, byte *pos, uint len, rta *a)
     net_fill_flow4(n, px, pxlen, pos, flen);
     ADVANCE(pos, len, flen);
 
+    /* Apply validation procedure per RFC 8955 (6) */
+    if (a && s->channel->cf->validate)
+      bgp_apply_flow_validation(s, n, a);
+
     bgp_rte_update(s, n, path_id, a);
   }
 }
@@ -1985,6 +2007,10 @@ bgp_decode_nlri_flow6(struct bgp_parse_state *s, byte *pos, uint len, rta *a)
     net_fill_flow6(n, px, pxlen, pos, flen);
     ADVANCE(pos, len, flen);
 
+    /* Apply validation procedure per RFC 8955 (6) */
+    if (a && s->channel->cf->validate)
+      bgp_apply_flow_validation(s, n, a);
+
     bgp_rte_update(s, n, path_id, a);
   }
 }
@@ -2438,6 +2464,8 @@ bgp_decode_nlri(struct bgp_parse_state *s, u32 afi, byte *nlri, uint len, ea_lis
   s->last_id = 0;
   s->last_src = s->proto->p.main_source;
 
+  s->base_table = NULL;
+
   /*
    * IPv4 BGP and MP-BGP may be used together in one update, therefore we do not
    * add BA_NEXT_HOP in bgp_decode_attrs(), but we add it here independently for

proto/pipe/pipe.c

@@ -81,7 +81,10 @@ pipe_rt_notify(struct proto *P, struct channel *src_ch, net *n, rte *new, rte *o
 #ifdef CONFIG_BGP
       /* Hack to cleanup cached value */
       if (e->attrs->src->proto->proto == &proto_bgp)
+      {
 	e->u.bgp.stale = -1;
+	e->u.bgp.base_table = NULL;
+      }
 #endif
 
       src = a->src;

test/birdtest.c

@@ -309,6 +309,12 @@ bt_log_suite_case_result(int result, const char *fmt, ...)
   }
 }
 
+void
+bt_reset_suite_case_timer(void)
+{
+  clock_gettime(CLOCK_MONOTONIC, &bt_suite_case_begin);
+}
+
 int
 bt_test_suite_base(int (*fn)(const void *), const char *id, const void *fn_arg, int forked, int timeout, const char *dsc, ...)
 {
@@ -501,6 +507,15 @@ bt_fmt_ipa(char *buf, size_t size, const void *data)
     bsnprintf(buf, size, "(null)");
 }
 
+void
+bt_format_net(char *buf, size_t size, const void *data)
+{
+  if (data)
+    bsnprintf(buf, size, "%N", (const net_addr *) data);
+  else
+    bsnprintf(buf, size, "(null)");
+}
+
 int
 bt_is_char(byte c)
 {

test/birdtest.h

@@ -32,6 +32,7 @@ extern const char *bt_test_id;
 
 void bt_init(int argc, char *argv[]);
 int  bt_exit_value(void);
+void bt_reset_suite_case_timer(void);
 int bt_test_suite_base(int (*test_fn)(const void *), const char *test_id, const void *test_fn_argument, int forked, int timeout, const char *dsc, ...);
 static inline u64 bt_random(void)
 { return ((u64) random() & 0xffffffff) | ((u64) random() << 32); }
@@ -165,6 +166,8 @@ struct bt_batch {
 void bt_fmt_str(char *buf, size_t size, const void *data);
 void bt_fmt_unsigned(char *buf, size_t size, const void *data);
 void bt_fmt_ipa(char *buf, size_t size, const void *data);
+void bt_format_net(char *buf, size_t size, const void *data);
+
 int bt_assert_batch__(struct bt_batch *opts);
 int bt_is_char(byte c);