Merge remote-tracking branch 'origin/master' into soft-int

NEWS

@@ -1,3 +1,17 @@
+Version 1.4.5 (2014-10-06)
+  o New 'show route noexport' command option.
+  o Port option for BGP sessions.
+  o Better constant handling in set literals.
+  o Better rate filtering of log messages.
+  o Several minor bugfixes.
+
+Version 1.4.4 (2014-07-09)
+  o Extended OSPF multipath support.
+  o Default router preference for RAdv.
+  o Significant changes in socket layer.
+  o Important bugfix in BGP.
+  o Several minor bugfixes.
+
 Version 1.4.3 (2014-04-14)
   o Important bugfix in IPv6 BGP.
 

conf/cf-lex.l

@@ -646,7 +646,7 @@ cf_walk_symbols(struct config *cf, struct symbol *sym, int *pos)
 char *
 cf_symbol_class_name(struct symbol *sym)
 {
-  if ((sym->class & 0xff00) == SYM_CONSTANT)
+  if (cf_symbol_is_constant(sym))
     return "constant";
 
   switch (sym->class)

conf/conf.h

@@ -149,6 +149,10 @@ void cf_pop_scope(void);
 struct symbol *cf_walk_symbols(struct config *cf, struct symbol *sym, int *pos);
 char *cf_symbol_class_name(struct symbol *sym);
 
+static inline int cf_symbol_is_constant(struct symbol *sym)
+{ return (sym->class & 0xff00) == SYM_CONSTANT; }
+
+
 /* Parser */
 
 int cf_parse(void);

conf/confbase.Y

@@ -72,7 +72,7 @@ CF_DECLS
 %token <t> TEXT
 %type <iface> ipa_scope
 
-%type <i> expr bool pxlen
+%type <i> expr bool pxlen ipa_port
 %type <i32> expr_us
 %type <time> datetime
 %type <a> ipa
@@ -88,7 +88,7 @@ CF_DECLS
 %left '!'
 %nonassoc '.'
 
-CF_KEYWORDS(DEFINE, ON, OFF, YES, NO, S, MS, US)
+CF_KEYWORDS(DEFINE, ON, OFF, YES, NO, S, MS, US, PORT)
 
 CF_GRAMMAR
 
@@ -161,6 +161,14 @@ ipa_scope:
  | '%' SYM { $$ = if_get_by_name($2->name); }
  ;
 
+ipa_port:
+   /* empty */ { $$ = 0; }
+ | PORT expr {
+     if (($2 < 1) || ($2 > 65535)) cf_error("Invalid port number");
+     $$ = $2;
+   }
+ ;
+
 prefix:
    ipa pxlen {
      if (!ip_is_prefix($1, $2)) cf_error("Invalid prefix");

doc/bird.sgml

@@ -318,7 +318,7 @@ protocol rip {
 <p><descrip>
 	<tag>include "<m/filename/"</tag> 
 	This statement causes inclusion of a new file. <m/Filename/ could also
-	be a wildcard. The maximal depth is 5. Note that this statement could be
+	be a wildcard. The maximal depth is 8. Note that this statement could be
 	used anywhere in the config file, not just as a top-level option.
 
 	<tag><label id="dsc-log">log "<m/filename/"|syslog [name <m/name/]|stderr all|{ <m/list of classes/ }</tag> 
@@ -735,7 +735,7 @@ This argument can be omitted if there exists only a single instance.
 	Show the list of symbols defined in the configuration (names of
 	protocols, routing tables etc.).
 
-	<tag>show route [[for] <m/prefix/|<m/IP/] [table <m/sym/] [filter <m/f/|where <m/c/] [(export|preexport) <m/p/] [protocol <m/p/] [<m/options/]</tag>
+	<tag>show route [[for] <m/prefix/|<m/IP/] [table <m/sym/] [filter <m/f/|where <m/c/] [(export|preexport|noexport) <m/p/] [protocol <m/p/] [<m/options/]</tag>
 	Show contents of a routing table (by default of the main one or the
 	table attached to a respective protocol), that is routes, their metrics
 	and (in case the <cf/all/ switch is given) all their attributes.
@@ -750,9 +750,14 @@ This argument can be omitted if there exists only a single instance.
 	<p>You can also ask for printing only routes processed and accepted by
 	a given filter (<cf>filter <m/name/</cf> or <cf>filter { <m/filter/ }
 	</cf> or matching a given condition (<cf>where <m/condition/</cf>).
-	The <cf/export/ and <cf/preexport/ switches ask for printing of entries
+
-	that are exported to the specified protocol. With <cf/preexport/, the
+	The <cf/export/, <cf/preexport/ and <cf/noexport/ switches ask for
-	export filter of the protocol is skipped.
+	printing of routes that are exported to the specified protocol.
+	With <cf/preexport/, the export filter of the protocol is skipped.
+	With <cf/noexport/, routes rejected by the export filter are printed
+	instead. Note that routes not exported to the protocol for other reasons
+	(e.g. secondary routes or routes imported from that protocol) are not
+	printed even with <cf/noexport/.
 
 	<p>You can also select just routes added by a specific protocol.
 	<cf>protocol <m/p/</cf>.
@@ -1004,7 +1009,6 @@ foot).
 	So <cf>1.2.0.0/16.pxlen = 16</cf> is true.
 
 	<tag/ec/
-
 	This is a specialized type used to represent BGP extended community
 	values. It is essentially a 64bit value, literals of this type are
 	usually written as <cf>(<m/kind/, <m/key/, <m/value/)</cf>, where
@@ -1616,7 +1620,7 @@ using the following configuration parameters:
 	address, equivalent to the <cf/source address/ option (see below). This
 	parameter is mandatory.
 
-	<tag>neighbor <m/ip/ as <m/number/</tag>
+	<tag>neighbor <m/ip/ [port <m/number/] as <m/number/</tag>
 	Define neighboring router this instance will be talking to and what AS
 	it's located in. In case the neighbor is in the same AS as we are, we
 	automatically switch to iBGP. This parameter is mandatory.
@@ -3036,6 +3040,10 @@ definitions, prefix definitions and DNS definitions:
 	as a default router. For <cf/sensitive/ option, see <ref id="dsc-trigger" name="trigger">.
 	Default: 3 * <cf/max ra	interval/, <cf/sensitive/ yes.
 
+	<tag>default preference low|medium|high</tag>
+	This option specifies the Default Router Preference value to advertise
+	to hosts. Default: medium.
+
 	<tag>rdnss local <m/switch/</tag>
 	Use only local (interface-specific) RDNSS definitions for this
 	interface. Otherwise, both global and local definitions are used. Could

filter/config.Y

@@ -25,6 +25,23 @@ static inline u32 pair_b(u32 p) { return p & 0xFFFF; }
  * to the last item in a list (right ptr). For convenience, even items
  * are handled as one-item lists. Lists are merged by f_merge_items().
  */
+static int
+f_valid_set_type(int type)
+{
+  switch (type)
+  {
+  case T_INT:
+  case T_PAIR:
+  case T_QUAD:
+  case T_ENUM:
+  case T_IP:
+  case T_EC:
+    return 1;
+
+  default:
+    return 0;
+  }
+}
 
 static inline struct f_tree *
 f_new_item(struct f_val from, struct f_val to)
@@ -473,10 +490,19 @@ fipa:
  */
 
 set_atom:
-   expr  { $$.type = T_INT; $$.val.i = $1; }
+   NUM   { $$.type = T_INT; $$.val.i = $1; }
  | RTRID { $$.type = T_QUAD; $$.val.i = $1; }
  | fipa  { $$ = $1; }
  | ENUM  { $$.type = pair_a($1); $$.val.i = pair_b($1); }
+ | '(' term ')' {
+     $$ = f_eval($2, cfg_mem);
+     if (!f_valid_set_type($$.type)) cf_error("Set-incompatible type");
+   }
+ | SYM {
+     if (!cf_symbol_is_constant($1)) cf_error("%s: constant expected", $1->name);
+     if (!f_valid_set_type(SYM_TYPE($1))) cf_error("%s: set-incompatible type", $1->name);
+     $$ = *(struct f_val *)($1->def);
+   }
  ;
 
 switch_atom:

filter/filter.c

@@ -493,7 +493,7 @@ f_rta_cow(void)
   }
 }
 
-static struct rate_limit rl_runtime_err;
+static struct tbf rl_runtime_err = TBF_DEFAULT_LOG_LIMITS;
 
 #define runtime(x) do { \
     log_rl(&rl_runtime_err, L_ERR "filters, line %d: %s", what->lineno, x); \
@@ -1492,7 +1492,7 @@ f_run(struct filter *filter, struct rte **rte, struct ea_list **tmp_attrs, struc
 
 
   if (res.type != T_RETURN) {
-    log( L_ERR "Filter %s did not return accept nor reject. Make up your mind", filter->name); 
+    log_rl(&rl_runtime_err, L_ERR "Filter %s did not return accept nor reject. Make up your mind", filter->name);
     return F_ERROR;
   }
   DBG( "done (%u)\n", res.val.i );

filter/test.conf

@@ -13,6 +13,9 @@ define '1a-a1' = (20+10);
 define one = 1;
 define ten = 10;
 
+define p23 = (2, 3);
+define ip1222 = 1.2.2.2;
+
 function onef(int a)
 {
 	return 1;
@@ -78,7 +81,7 @@ function test_roa()
 	      " ", roa_check(rl, 10.130.130.0/24, 3000) = ROA_VALID;
 }
 
-function paths()
+function path_test()
 bgpmask pm1;
 bgpmask pm2;
 bgppath p2;
@@ -98,7 +101,7 @@ eclist el2;
 	print "Should be true: ", p2 ~ pm1, " ", p2 ~ pm2, " ", 3 ~ p2, " ", p2 ~ [2, 10..20], " ", p2 ~ [4, 10..20];
 	print "4 = ", p2.len;
 	p2 = prepend( p2, 5 );
-	print "Should be false: ", p2 ~ pm1, " ", p2 ~ pm2, " ", 10 ~ p2, " ", p2 ~ [8, 10..20],;
+	print "Should be false: ", p2 ~ pm1, " ", p2 ~ pm2, " ", 10 ~ p2, " ", p2 ~ [8, ten..(2*ten)];
 	print "Should be true: ", p2 ~  / ? 4 3 2 1 /,  " ", p2, " ",  / ? 4 3 2 1 /;
 	print "Should be true: ", p2 ~ [= * 4 3 * 1 =], " ", p2, " ", [= * 4 3 * 1 =];
 	print "Should be true: ", p2 ~ [= (3+2) (2*2) 3 2 1 =], " ", p2 ~ mkpath(5, 4);
@@ -124,7 +127,7 @@ eclist el2;
 	print "Should be always true: ", l ~ [(*,*)];
 	l = add( l, (2,one+2) );
 	print "Community list (1,2) (2,3) ", l;
-	print "Should be true: ", (2,3) ~ l, " ", l ~ [(1,*)], " ", l ~ [(2,3)]," ", l ~ [(2,2..3)], " ", l ~ [(1,1..2)], " ", l ~ [(1,1)..(1,2)];
+	print "Should be true: ", (2,3) ~ l, " ", l ~ [(1,*)], " ", l ~ [p23]," ", l ~ [(2,2..3)], " ", l ~ [(1,1..2)], " ", l ~ [(1,1)..(1,2)];
 	l = add( l, (2,5) );
 	l = add( l, (5,one) );
 	l = add( l, (6,one) );
@@ -361,7 +364,7 @@ string st;
         if ( b = true ) then print "Testing bool comparison b = true: ", b;
 	else { print "*** FAIL: TRUE test failed" ; quitbird; }
 	
-	ips = [ 1.1.1.0 .. 1.1.1.255, 1.2.2.2];
+	ips = [ 1.1.1.0 .. 1.1.1.255, ip1222];
 	print "Testing IP sets: ";
 	print ips;
 	print "  must be true:  ",	1.1.1.0 ~ ips, ",", 1.1.1.100 ~ ips, ",", 1.2.2.2 ~ ips;
@@ -389,7 +392,7 @@ string st;
 	i = fifteen();
 	print "Testing function calls: 15 = ", i;
 
-	paths();
+	path_test();
 
 	print "1.2.3.4 = ", onetwo;
 

lib/Doc

@@ -1,7 +1,7 @@
 H Library functions
 S ip.c ipv4.c ipv6.c
 S lists.c
-S checksum.c bitops.c patmatch.c printf.c xmalloc.c
+S checksum.c bitops.c patmatch.c printf.c xmalloc.c tbf.c
 D resource.sgml
 S resource.c
 S mempool.c

lib/Modules

@@ -19,6 +19,7 @@ resource.c
 resource.h
 slab.c
 socket.h
+tbf.c
 unaligned.h
 xmalloc.c
 printf.c

lib/birdlib.h

@@ -74,6 +74,38 @@ typedef s64 btime;
 #endif
 
 
+/* Rate limiting */
+
+struct tbf {
+  bird_clock_t timestamp;		/* Last update */
+  u16 count;				/* Available tokens */
+  u16 burst;				/* Max number of tokens */
+  u16 rate;				/* Rate of replenishment */
+  u16 mark;				/* Whether last op was limited */
+};
+
+/* Default TBF values for rate limiting log messages */
+#define TBF_DEFAULT_LOG_LIMITS { .rate = 1, .burst = 5 }
+
+void tbf_update(struct tbf *f);
+
+static inline int
+tbf_limit(struct tbf *f)
+{
+  tbf_update(f);
+
+  if (!f->count)
+  {
+    f->mark = 1;
+    return 1;
+  }
+
+  f->count--;
+  f->mark = 0;
+  return 0;
+}
+
+
 /* Logging and dying */
 
 typedef struct buffer {
@@ -94,16 +126,10 @@ typedef struct buffer {
 
 #define LOG_BUFFER_SIZE 1024
 
-
-struct rate_limit {
-  bird_clock_t timestamp;
-  int count;
-};
-
 #define log log_msg
 void log_commit(int class, buffer *buf);
 void log_msg(char *msg, ...);
-void log_rl(struct rate_limit *rl, char *msg, ...);
+void log_rl(struct tbf *rl, char *msg, ...);
 void die(char *msg, ...) NORET;
 void bug(char *msg, ...) NORET;
 

lib/slab.c

@@ -123,7 +123,7 @@ static size_t
 slab_memsize(resource *r)
 {
   slab *s = (slab *) r;
-  int cnt = 0;
+  size_t cnt = 0;
   struct sl_obj *o;
 
   WALK_LIST(o, s->objs)
@@ -346,7 +346,7 @@ static size_t
 slab_memsize(resource *r)
 {
   slab *s = (slab *) r;
-  int heads = 0;
+  size_t heads = 0;
   struct sl_head *h;
 
   WALK_LIST(h, s->empty_heads)

lib/tbf.c

@@ -0,0 +1,29 @@
+/*
+ *	BIRD Library -- Token Bucket Filter
+ *
+ *	(c) 2014 Ondrej Zajicek <santiago@crfreenet.org>
+ *	(c) 2014 CZ.NIC z.s.p.o.
+ *
+ *	Can be freely distributed and used under the terms of the GNU GPL.
+ */
+
+#include "nest/bird.h"
+
+void
+tbf_update(struct tbf *f)
+{
+  bird_clock_t delta = now - f->timestamp;
+
+  if (delta == 0)
+    return;
+
+  f->timestamp = now;
+
+  if ((0 < delta) && (delta < f->burst))
+  {
+    u32 next = f->count + delta * f->rate;
+    f->count = MIN(next, f->burst);
+  }
+  else
+    f->count = f->burst;
+}

misc/bird.spec

@@ -1,6 +1,6 @@
 Summary: BIRD Internet Routing Daemon
 Name: bird
-Version: 1.4.3
+Version: 1.4.5
 Release: 1
 Copyright: GPL
 Group: Networking/Daemons

nest/config.Y

@@ -57,10 +57,10 @@ CF_KEYWORDS(ROUTER, ID, PROTOCOL, TEMPLATE, PREFERENCE, DISABLED, DEBUG, ALL, OF
 CF_KEYWORDS(INTERFACE, IMPORT, EXPORT, FILTER, NONE, TABLE, STATES, ROUTES, FILTERS)
 CF_KEYWORDS(RECEIVE, LIMIT, ACTION, WARN, BLOCK, RESTART, DISABLE, KEEP, FILTERED)
 CF_KEYWORDS(PASSWORD, FROM, PASSIVE, TO, ID, EVENTS, PACKETS, PROTOCOLS, INTERFACES)
-CF_KEYWORDS(PRIMARY, STATS, COUNT, FOR, COMMANDS, PREEXPORT, GENERATE, ROA, MAX, FLUSH, AS)
+CF_KEYWORDS(PRIMARY, STATS, COUNT, FOR, COMMANDS, PREEXPORT, NOEXPORT, GENERATE, ROA)
 CF_KEYWORDS(LISTEN, BGP, V6ONLY, DUAL, ADDRESS, PORT, PASSWORDS, DESCRIPTION, SORTED)
 CF_KEYWORDS(RELOAD, IN, OUT, MRTDUMP, MESSAGES, RESTRICT, MEMORY, IGP_METRIC, CLASS, DSCP)
-CF_KEYWORDS(GRACEFUL, RESTART, WAIT)
+CF_KEYWORDS(GRACEFUL, RESTART, WAIT, MAX, FLUSH, AS)
 
 CF_ENUM(T_ENUM_RTS, RTS_, DUMMY, STATIC, INHERIT, DEVICE, STATIC_DEVICE, REDIRECT,
 	RIP, OSPF, OSPF_IA, OSPF_EXT1, OSPF_EXT2, BGP, PIPE)
@@ -77,7 +77,7 @@ CF_ENUM(T_ENUM_ROA, ROA_, UNKNOWN, VALID, INVALID)
 %type <ro> roa_args
 %type <rot> roa_table_arg
 %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_or_preexport roa_mode limit_action tab_sorted tos
+%type <i> proto_start echo_mask echo_size debug_mask debug_list debug_flag mrtdump_mask mrtdump_list mrtdump_flag export_mode roa_mode limit_action tab_sorted tos
 %type <ps> proto_patt proto_patt2
 %type <g> limit_spec
 
@@ -443,7 +443,7 @@ CF_CLI(SHOW INTERFACES SUMMARY,,, [[Show summary of network interfaces]])
 { if_show_summary(); } ;
 
 CF_CLI_HELP(SHOW ROUTE, ..., [[Show routing table]])
-CF_CLI(SHOW ROUTE, r_args, [[[<prefix>|for <prefix>|for <ip>] [table <t>] [filter <f>|where <cond>] [all] [primary] [filtered] [(export|preexport) <p>] [protocol <p>] [stats|count]]], [[Show routing table]])
+CF_CLI(SHOW ROUTE, r_args, [[[<prefix>|for <prefix>|for <ip>] [table <t>] [filter <f>|where <cond>] [all] [primary] [filtered] [(export|preexport|noexport) <p>] [protocol <p>] [stats|count]]], [[Show routing table]])
 { rt_show($3); } ;
 
 r_args:
@@ -492,7 +492,7 @@ r_args:
      $$ = $1;
      $$->filtered = 1;
    }
- | r_args export_or_preexport SYM {
+ | r_args export_mode SYM {
      struct proto_config *c = (struct proto_config *) $3->def;
      $$ = $1;
      if ($$->export_mode) cf_error("Protocol specified twice");
@@ -519,9 +519,10 @@ r_args:
    }
  ;
 
-export_or_preexport:
+export_mode:
-   PREEXPORT { $$ = 1; }
+   PREEXPORT	{ $$ = RSEM_PREEXPORT; }
- | EXPORT { $$ = 2; }
+ | EXPORT	{ $$ = RSEM_EXPORT; }
+ | NOEXPORT	{ $$ = RSEM_NOEXPORT; }
  ;
 
 

nest/route.h

@@ -301,6 +301,12 @@ struct rt_show_data {
 };
 void rt_show(struct rt_show_data *);
 
+/* 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 */
+#define RSEM_EXPORT	2		/* Routes accepted by export filter */
+#define RSEM_NOEXPORT	3		/* Routes rejected by export filter */
+
 /*
  *	Route Attributes
  *

nest/rt-attr.c

@@ -815,7 +815,7 @@ rta_alloc_hash(void)
 static inline unsigned int
 rta_hash(rta *a)
 {
-  return (((unsigned) a->src) ^ ipa_hash(a->gw) ^
+  return (((uint) (uintptr_t) a->src) ^ ipa_hash(a->gw) ^
 	  mpnh_hash(a->nexthops) ^ ea_hash(a->eattrs)) & 0xffff;
 }
 

nest/rt-table.c

@@ -645,7 +645,7 @@ rte_recalculate(struct announce_hook *ah, net *net, rte *new, ea_list *tmpa, str
   struct proto *p = ah->proto;
   struct rtable *table = ah->table;
   struct proto_stats *stats = ah->stats;
-  static struct rate_limit rl_pipe;
+  static struct tbf rl_pipe = TBF_DEFAULT_LOG_LIMITS;
   rte *before_old = NULL;
   rte *old_best = net->routes;
   rte *old = NULL;
@@ -1367,7 +1367,7 @@ rt_init(void)
 static int
 rt_prune_step(rtable *tab, int step, int *limit)
 {
-  static struct rate_limit rl_flush;
+  static struct tbf rl_flush = TBF_DEFAULT_LOG_LIMITS;
   struct fib_iterator *fit = &tab->prune_fit;
 
   DBG("Pruning route table %s\n", tab->name);
@@ -2255,6 +2255,9 @@ rt_show_net(struct cli *c, net *n, struct rt_show_data *d)
 
   if (d->export_mode)
     {
+      if (! d->export_protocol->rt_notify)
+	return;
+
       a = proto_find_announce_hook(d->export_protocol, d->table);
       if (!a)
 	return;
@@ -2287,18 +2290,20 @@ rt_show_net(struct cli *c, net *n, struct rt_show_data *d)
 	  if (ic < 0)
 	    goto skip;
 
-	  if (d->export_mode > 1)
+	  if (d->export_mode > RSEM_PREEXPORT)
 	    {
 	      /*
 	       * FIXME - This shows what should be exported according to current
 	       * filters, but not what was really exported. 'configure soft'
 	       * command may change the export filter and do not update routes.
 	       */
+	      int do_export = (ic > 0) ||
+		(f_run(a->out_filter, &e, &tmpa, rte_update_pool, FF_FORCE_TMPATTR) <= F_ACCEPT);
 
-	      if (!ic && (f_run(a->out_filter, &e, &tmpa, rte_update_pool, FF_FORCE_TMPATTR) > F_ACCEPT))
+	      if (do_export != (d->export_mode == RSEM_EXPORT))
 		goto skip;
 
-	      if (ep->accept_ra_types == RA_ACCEPTED)
+	      if ((d->export_mode == RSEM_EXPORT) && (ep->accept_ra_types == RA_ACCEPTED))
 		pass = 1;
 	    }
 	}

proto/bfd/bfd.c

@@ -1071,7 +1071,7 @@ bfd_show_sessions(struct proto *P)
 {
   byte tbuf[TM_DATETIME_BUFFER_SIZE];
   struct bfd_proto *p = (struct bfd_proto *) P;
-  uint state, diag;
+  uint state, diag UNUSED;
   u32 tx_int, timeout;
   const char *ifname;
 

proto/bgp/bgp.c

@@ -680,8 +680,8 @@ bgp_connect(struct bgp_proto *p)	/* Enter Connect state and start establishing c
   s->type = SK_TCP_ACTIVE;
   s->saddr = p->source_addr;
   s->daddr = p->cf->remote_ip;
+  s->dport = p->cf->remote_port;
   s->iface = p->neigh ? p->neigh->iface : NULL;
-  s->dport = BGP_PORT;
   s->ttl = p->cf->ttl_security ? 255 : hops;
   s->rbsize = BGP_RX_BUFFER_SIZE;
   s->tbsize = BGP_TX_BUFFER_SIZE;
@@ -1016,9 +1016,9 @@ bgp_start(struct proto *P)
 
   lock = p->lock = olock_new(P->pool);
   lock->addr = p->cf->remote_ip;
+  lock->port = p->cf->remote_port;
   lock->iface = p->cf->iface;
   lock->type = OBJLOCK_TCP;
-  lock->port = BGP_PORT;
   lock->hook = bgp_start_locked;
   lock->data = p;
   olock_acquire(lock);

proto/bgp/bgp.h

@@ -23,6 +23,7 @@ struct bgp_config {
   ip_addr remote_ip;
   ip_addr source_addr;			/* Source address to use */
   struct iface *iface;			/* Interface for link-local addresses */
+  u16 remote_port; 			/* Neighbor destination port */
   int multihop;				/* Number of hops if multihop */
   int ttl_security;			/* Enable TTL security [RFC5082] */
   int next_hop_self;			/* Always set next hop to local IP address */

proto/bgp/config.Y

@@ -60,7 +60,7 @@ bgp_proto:
  | bgp_proto proto_item ';'
  | bgp_proto LOCAL AS expr ';' { BGP_CFG->local_as = $4; }
  | bgp_proto LOCAL ipa AS expr ';' { BGP_CFG->source_addr = $3; BGP_CFG->local_as = $5; }
- | bgp_proto NEIGHBOR ipa ipa_scope AS expr ';' {
+ | bgp_proto NEIGHBOR ipa ipa_scope ipa_port AS expr ';' {
      if (ipa_nonzero(BGP_CFG->remote_ip))
        cf_error("Only one neighbor per BGP instance is allowed");
      if (!ipa_has_link_scope($3) != !$4)
@@ -68,7 +68,8 @@ bgp_proto:
 
      BGP_CFG->remote_ip = $3;
      BGP_CFG->iface = $4;
-     BGP_CFG->remote_as = $6;
+     BGP_CFG->remote_port = ($5 > 0) ? $5 : BGP_PORT;
+     BGP_CFG->remote_as = $7;
    }
  | bgp_proto RR CLUSTER ID idval ';' { BGP_CFG->rr_cluster_id = $5; }
  | bgp_proto RR CLIENT ';' { BGP_CFG->rr_client = 1; }

proto/bgp/packets.c

@@ -22,7 +22,8 @@
 
 #include "bgp.h"
 
-static struct rate_limit rl_rcv_update,  rl_snd_update;
+static struct tbf rl_rcv_update = TBF_DEFAULT_LOG_LIMITS;
+static struct tbf rl_snd_update = TBF_DEFAULT_LOG_LIMITS;
 
 /* Table for state -> RFC 6608 FSM error subcodes */
 static byte fsm_err_subcode[BS_MAX] = {

proto/radv/config.Y

@@ -30,9 +30,9 @@ CF_KEYWORDS(RADV, PREFIX, INTERFACE, MIN, MAX, RA, DELAY, INTERVAL,
 	MANAGED, OTHER, CONFIG, LINK, MTU, REACHABLE, TIME, RETRANS,
 	TIMER, CURRENT, HOP, LIMIT, DEFAULT, VALID, PREFERRED, MULT,
 	LIFETIME, SKIP, ONLINK, AUTONOMOUS, RDNSS, DNSSL, NS, DOMAIN,
-	LOCAL, TRIGGER, SENSITIVE)
+	LOCAL, TRIGGER, SENSITIVE, PREFERENCE, LOW, MEDIUM, HIGH)
 
-%type<i> radv_mult radv_sensitive
+%type<i> radv_mult radv_sensitive radv_preference
 
 CF_GRAMMAR
 
@@ -84,6 +84,7 @@ radv_iface_start:
   RADV_IFACE->current_hop_limit = DEFAULT_CURRENT_HOP_LIMIT;
   RADV_IFACE->default_lifetime = -1;
   RADV_IFACE->default_lifetime_sensitive = 1;
+  RADV_IFACE->default_preference = RA_PREF_MEDIUM;
 };
 
 radv_iface_item:
@@ -101,6 +102,7 @@ radv_iface_item:
      if (($3 < 0) || ($3 > 9000))  cf_error("Default lifetime must be in range 0-9000");
      if ($4 != -1) RADV_IFACE->default_lifetime_sensitive = $4;
    }
+ | DEFAULT PREFERENCE radv_preference { RADV_IFACE->default_preference = $3; }
  | PREFIX radv_prefix { add_tail(&RADV_IFACE->pref_list, NODE this_radv_prefix); }
  | RDNSS { init_list(&radv_dns_list); } radv_rdnss { add_tail_list(&RADV_IFACE->rdnss_list, &radv_dns_list); }
  | DNSSL { init_list(&radv_dns_list); } radv_dnssl { add_tail_list(&RADV_IFACE->dnssl_list, &radv_dns_list); }
@@ -108,6 +110,11 @@ radv_iface_item:
  | DNSSL LOCAL bool { RADV_IFACE->dnssl_local = $3; }
  ;
 
+radv_preference:
+   LOW { $$ = RA_PREF_LOW; }
+ | MEDIUM { $$ = RA_PREF_MEDIUM; }
+ | HIGH { $$ = RA_PREF_HIGH; }
+
 radv_iface_finish:
 {
   struct radv_iface_config *ic = RADV_IFACE;

proto/radv/packets.c

@@ -251,10 +251,11 @@ radv_prepare_ra(struct radv_iface *ifa)
   pkt->code = 0;
   pkt->checksum = 0;
   pkt->current_hop_limit = ic->current_hop_limit;
-  pkt->flags = (ic->managed ? OPT_RA_MANAGED : 0) |
-    (ic->other_config ? OPT_RA_OTHER_CFG : 0);
   pkt->router_lifetime = (ra->active || !ic->default_lifetime_sensitive) ?
     htons(ic->default_lifetime) : 0;
+  pkt->flags = (ic->managed ? OPT_RA_MANAGED : 0) |
+    (ic->other_config ? OPT_RA_OTHER_CFG : 0) |
+    (pkt->router_lifetime ? ic->default_preference : 0);
   pkt->reachable_time = htonl(ic->reachable_time);
   pkt->retrans_timer = htonl(ic->retrans_timer);
   buf += sizeof(*pkt);
@@ -330,10 +331,15 @@ radv_send_ra(struct radv_iface *ifa, int shutdown)
 
   if (shutdown)
   {
-    /* Modify router lifetime to 0, it is not restored because
+    /*
-       we suppose that the iface will be removed */
+     * Modify router lifetime to 0, it is not restored because we suppose that
+     * the iface will be removed. The preference value also has to be zeroed.
+     * (RFC 4191 2.2: If router lifetime is 0, the preference value must be 0.)
+     */
+
     struct radv_ra_packet *pkt = (void *) ifa->sk->tbuf;
     pkt->router_lifetime = 0;
+    pkt->flags &= ~RA_PREF_MASK;
   }
 
   RADV_TRACE(D_PACKETS, "Sending RA via %s", ifa->iface->name);

proto/radv/radv.c

@@ -40,6 +40,7 @@
  * Supported standards:
  * - RFC 4861 - main RA standard
  * - RFC 6106 - DNS extensions (RDDNS, DNSSL)
+ * - RFC 4191 (partial) - Default Router Preference
  */
 
 static void

proto/radv/radv.h

@@ -80,6 +80,7 @@ struct radv_iface_config
   u32 current_hop_limit;
   u32 default_lifetime;
   u8 default_lifetime_sensitive; /* Whether default_lifetime depends on trigger */
+  u8 default_preference;	/* Default Router Preference (RFC 4191) */
 };
 
 struct radv_prefix_config
@@ -144,6 +145,11 @@ struct radv_iface
 #define RA_EV_CHANGE 2		/* Change of options or prefixes */
 #define RA_EV_RS 3		/* Received RS */
 
+/* Default Router Preferences (RFC 4191) */
+#define RA_PREF_LOW	0x18
+#define RA_PREF_MEDIUM	0x00
+#define RA_PREF_HIGH	0x08
+#define RA_PREF_MASK	0x18
 
 
 #ifdef LOCAL_DEBUG

sysdep/bsd/krt-sock.c

@@ -261,6 +261,7 @@ krt_send_route(struct krt_proto *p, int cmd, rte *e)
       msg.rtm.rtm_flags |= RTF_GATEWAY;
       msg.rtm.rtm_addrs |= RTA_GATEWAY;
       break;
+
 #ifdef RTF_REJECT
     case RTD_UNREACHABLE:
 #endif
@@ -280,7 +281,7 @@ krt_send_route(struct krt_proto *p, int cmd, rte *e)
           return -1;
         }
 
-	sockaddr_fill(&dst, BIRD_AF, i->addr->ip, NULL, 0);
+	sockaddr_fill(&gate, BIRD_AF, i->addr->ip, NULL, 0);
         msg.rtm.rtm_addrs |= RTA_GATEWAY;
       }
       break;

sysdep/bsd/sysio.h

@@ -141,6 +141,7 @@ sk_prepare_cmsgs4(sock *s, struct msghdr *msg, void *cbuf, size_t cbuflen)
 #ifdef IP_SENDSRCADDR
   struct cmsghdr *cm;
   struct in_addr *sa;
+  int controllen = 0;
 
   msg->msg_control = cbuf;
   msg->msg_controllen = cbuflen;
@@ -149,11 +150,12 @@ sk_prepare_cmsgs4(sock *s, struct msghdr *msg, void *cbuf, size_t cbuflen)
   cm->cmsg_level = IPPROTO_IP;
   cm->cmsg_type = IP_SENDSRCADDR;
   cm->cmsg_len = CMSG_LEN(sizeof(*sa));
+  controllen += CMSG_SPACE(sizeof(*sa));
 
   sa = (struct in_addr *) CMSG_DATA(cm);
   *sa = ipa_to_in4(s->saddr);
 
-  msg->msg_controllen = cm->cmsg_len;
+  msg->msg_controllen = controllen;
 #endif
 }
 

sysdep/config.h

@@ -7,7 +7,7 @@
 #define _BIRD_CONFIG_H_
 
 /* BIRD version */
-#define BIRD_VERSION "1.4.3"
+#define BIRD_VERSION "1.4.5"
 
 /* Include parameters determined by configure script */
 #include "sysdep/autoconf.h"

sysdep/linux/netlink.c

@@ -151,7 +151,7 @@ nl_get_reply(struct nl_sock *nl)
     }
 }
 
-static struct rate_limit rl_netlink_err;
+static struct tbf rl_netlink_err = TBF_DEFAULT_LOG_LIMITS;
 
 static int
 nl_error(struct nlmsghdr *h)

sysdep/linux/sysio.h

@@ -154,6 +154,7 @@ sk_prepare_cmsgs4(sock *s, struct msghdr *msg, void *cbuf, size_t cbuflen)
 {
   struct cmsghdr *cm;
   struct in_pktinfo *pi;
+  int controllen = 0;
 
   msg->msg_control = cbuf;
   msg->msg_controllen = cbuflen;
@@ -162,13 +163,14 @@ sk_prepare_cmsgs4(sock *s, struct msghdr *msg, void *cbuf, size_t cbuflen)
   cm->cmsg_level = SOL_IP;
   cm->cmsg_type = IP_PKTINFO;
   cm->cmsg_len = CMSG_LEN(sizeof(*pi));
+  controllen += CMSG_SPACE(sizeof(*pi));
 
   pi = (struct in_pktinfo *) CMSG_DATA(cm);
   pi->ipi_ifindex = s->iface ? s->iface->index : 0;
   pi->ipi_spec_dst = ipa_to_in4(s->saddr);
   pi->ipi_addr = ipa_to_in4(IPA_NONE);
 
-  msg->msg_controllen = cm->cmsg_len;
+  msg->msg_controllen = controllen;
 }
 
 

sysdep/unix/io.c

@@ -705,6 +705,7 @@ sk_prepare_cmsgs6(sock *s, struct msghdr *msg, void *cbuf, size_t cbuflen)
 {
   struct cmsghdr *cm;
   struct in6_pktinfo *pi;
+  int controllen = 0;
 
   msg->msg_control = cbuf;
   msg->msg_controllen = cbuflen;
@@ -713,12 +714,13 @@ sk_prepare_cmsgs6(sock *s, struct msghdr *msg, void *cbuf, size_t cbuflen)
   cm->cmsg_level = SOL_IPV6;
   cm->cmsg_type = IPV6_PKTINFO;
   cm->cmsg_len = CMSG_LEN(sizeof(*pi));
+  controllen += CMSG_SPACE(sizeof(*pi));
 
   pi = (struct in6_pktinfo *) CMSG_DATA(cm);
   pi->ipi6_ifindex = s->iface ? s->iface->index : 0;
   pi->ipi6_addr = ipa_to_in6(s->saddr);
 
-  msg->msg_controllen = cm->cmsg_len;
+  msg->msg_controllen = controllen;
 }
 
 
@@ -1965,7 +1967,7 @@ io_loop(void)
 	  while (current_sock && count < MAX_RX_STEPS)
 	    {
 	      sock *s = current_sock;
-	      int e;
+	      int e UNUSED;
 
 	      if ((s->type < SK_MAGIC) && FD_ISSET(s->fd, &rd) && s->rx_hook)
 		{

sysdep/unix/krt.c

@@ -300,10 +300,10 @@ krt_trace_in(struct krt_proto *p, rte *e, char *msg)
 }
 
 static inline void
-krt_trace_in_rl(struct rate_limit *rl, struct krt_proto *p, rte *e, char *msg)
+krt_trace_in_rl(struct tbf *f, struct krt_proto *p, rte *e, char *msg)
 {
   if (p->p.debug & D_PACKETS)
-    log_rl(rl, L_TRACE "%s: %I/%d: %s", p->p.name, e->net->n.prefix, e->net->n.pxlen, msg);
+    log_rl(f, L_TRACE "%s: %I/%d: %s", p->p.name, e->net->n.prefix, e->net->n.pxlen, msg);
 }
 
 /*
@@ -312,7 +312,7 @@ krt_trace_in_rl(struct rate_limit *rl, struct krt_proto *p, rte *e, char *msg)
 
 #ifdef KRT_ALLOW_LEARN
 
-static struct rate_limit rl_alien_seen, rl_alien_updated, rl_alien_created, rl_alien_ignored;
+static struct tbf rl_alien = TBF_DEFAULT_LOG_LIMITS;
 
 /*
  * krt_same_key() specifies what (aside from the net) is the key in
@@ -378,20 +378,20 @@ krt_learn_scan(struct krt_proto *p, rte *e)
     {
       if (krt_uptodate(m, e))
 	{
-	  krt_trace_in_rl(&rl_alien_seen, p, e, "[alien] seen");
+	  krt_trace_in_rl(&rl_alien, p, e, "[alien] seen");
 	  rte_free(e);
 	  m->u.krt.seen = 1;
 	}
       else
 	{
-	  krt_trace_in_rl(&rl_alien_updated, p, e, "[alien] updated");
+	  krt_trace_in(p, e, "[alien] updated");
 	  *mm = m->next;
 	  rte_free(m);
 	  m = NULL;
 	}
     }
   else
-    krt_trace_in_rl(&rl_alien_created, p, e, "[alien] created");
+    krt_trace_in(p, e, "[alien] created");
   if (!m)
     {
       e->next = n->routes;
@@ -637,7 +637,7 @@ krt_got_route(struct krt_proto *p, rte *e)
 	krt_learn_scan(p, e);
       else
 	{
-	  krt_trace_in_rl(&rl_alien_ignored, p, e, "[alien] ignored");
+	  krt_trace_in_rl(&rl_alien, p, e, "[alien] ignored");
 	  rte_free(e);
 	}
       return;
@@ -910,7 +910,7 @@ krt_scan_timer_stop(struct krt_proto *p)
 }
 
 static void
-krt_scan_timer_kick(struct krt_proto *p UNUSED)
+krt_scan_timer_kick(struct krt_proto *p)
 {
   tm_start(p->scan_timer, 0);
 }
@@ -974,7 +974,7 @@ krt_import_control(struct proto *P, rte **new, ea_list **attrs, struct linpool *
 }
 
 static void
-krt_notify(struct proto *P, struct rtable *table UNUSED, net *net,
+krt_rt_notify(struct proto *P, struct rtable *table UNUSED, net *net,
 	      rte *new, rte *old, struct ea_list *eattrs)
 {
   struct krt_proto *p = (struct krt_proto *) P;
@@ -991,6 +991,36 @@ krt_notify(struct proto *P, struct rtable *table UNUSED, net *net,
     krt_replace_rte(p, net, new, old, eattrs);
 }
 
+static void
+krt_if_notify(struct proto *P, uint flags, struct iface *iface UNUSED)
+{
+  struct krt_proto *p = (struct krt_proto *) P;
+
+  /*
+   * When interface went down, we should remove routes to it. In the ideal world,
+   * OS kernel would send us route removal notifications in such cases, but we
+   * cannot rely on it as it is often not true. E.g. Linux kernel removes related
+   * routes when an interface went down, but it does not notify userspace about
+   * that. To be sure, we just schedule a scan to ensure synchronization.
+   */
+
+  if ((flags & IF_CHANGE_DOWN) && KRT_CF->learn)
+    krt_scan_timer_kick(p);
+}
+
+static int
+krt_reload_routes(struct proto *P)
+{
+  struct krt_proto *p = (struct krt_proto *) P;
+
+  /* Although we keep learned routes in krt_table, we rather schedule a scan */
+
+  if (KRT_CF->learn)
+    krt_scan_timer_kick(p);
+
+  return 1;
+}
+
 static void
 krt_feed_done(struct proto *P)
 {
@@ -1022,7 +1052,9 @@ krt_init(struct proto_config *c)
 
   p->p.accept_ra_types = RA_OPTIMAL;
   p->p.import_control = krt_import_control;
-  p->p.rt_notify = krt_notify;
+  p->p.rt_notify = krt_rt_notify;
+  p->p.if_notify = krt_if_notify;
+  p->p.reload_routes = krt_reload_routes;
   p->p.feed_done = krt_feed_done;
   p->p.make_tmp_attrs = krt_make_tmp_attrs;
   p->p.store_tmp_attrs = krt_store_tmp_attrs;

sysdep/unix/log.c

@@ -32,9 +32,6 @@ static FILE *dbgf;
 static list *current_log_list;
 static char *current_syslog_name; /* NULL -> syslog closed */
 
-static const bird_clock_t rate_limit_time = 5;
-static const int rate_limit_count = 5;
-
 
 #ifdef USE_PTHREADS
 
@@ -154,7 +151,6 @@ vlog(int class, const char *msg, va_list args)
 }
 
 
-
 /**
  * log - log a message
  * @msg: printf-like formatting string with message class information
@@ -180,31 +176,21 @@ log_msg(char *msg, ...)
 }
 
 void
-log_rl(struct rate_limit *rl, char *msg, ...)
+log_rl(struct tbf *f, char *msg, ...)
 {
+  int last_hit = f->mark;
   int class = 1;
   va_list args;
 
-  bird_clock_t delta = now - rl->timestamp;
+  /* Rate limiting is a bit tricky here as it also logs '...' during the first hit */
-  if ((0 <= delta) && (delta < rate_limit_time))
+  if (tbf_limit(f) && last_hit)
-    {
-      rl->count++;
-    }
-  else
-    {
-      rl->timestamp = now;
-      rl->count = 1;
-    }
-
-  if (rl->count > rate_limit_count)
     return;
 
-  va_start(args, msg);
   if (*msg >= 1 && *msg <= 8)
     class = *msg++;
-  vlog(class, msg, args);
+
-  if (rl->count == rate_limit_count)
+  va_start(args, msg);
-    vlog(class, "...", args);
+  vlog(class, (f->mark ? "..." : msg), args);
   va_end(args);
 }