From e46128fb50e108e8cfdf6bb6e9ab040e00f5dfdc Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Mon, 19 Jun 2017 12:46:40 +0200
Subject: [PATCH 001/198] Filters: Do not clamp EC set values to 16 bit for
 EC_GENERICs

Thanks to Lennert Buytenhek <buytenh@wantstofly.org> for the patch.
---
 filter/config.Y | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/filter/config.Y b/filter/config.Y
index 5ea83f81..c9f9c753 100644
--- a/filter/config.Y
+++ b/filter/config.Y
@@ -112,7 +112,7 @@ f_new_ec_item(u32 kind, u32 ipv4_used, u32 key, u32 vf, u32 vt)
 {
   u64 fm, to;
 
-  if (ipv4_used || (key >= 0x10000)) {
+  if ((kind != EC_GENERIC) && (ipv4_used || (key >= 0x10000))) {
     check_u16(vf);
     if (vt == EC_ALL)
       vt = 0xFFFF;

From b4a33e21ea3a8d23dee3ed23a515cc7430a38643 Mon Sep 17 00:00:00 2001
From: Martin Mares <mj@ucw.cz>
Date: Tue, 18 Jul 2017 23:48:25 +0200
Subject: [PATCH 002/198] Configure: Fix a typo in checking of backtrace()

---
 configure.ac | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/configure.ac b/configure.ac
index 9e375762..5514fb56 100644
--- a/configure.ac
+++ b/configure.ac
@@ -314,7 +314,7 @@ if test "$enable_debug" = yes ; then
     [
       AC_DEFINE([HAVE_EXECINFO_H], [1], [Define to 1 if you have the <execinfo.h> header file.])
       AC_SEARCH_LIBS([backtrace], [execinfo],
-	[]
+	[],
 	[AC_MSG_ERROR([Function backtrace not available.])]
       )
     ]

From afd9845e2636146a66508639614a93167d5d1728 Mon Sep 17 00:00:00 2001
From: Michal 'vorner' Vaner <michal.vaner@nic.cz>
Date: Fri, 4 Aug 2017 10:52:57 +0200
Subject: [PATCH 003/198] radv: Fix RFC reference in comments

---
 proto/radv/radv.h | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/proto/radv/radv.h b/proto/radv/radv.h
index 735aa218..3e6c1651 100644
--- a/proto/radv/radv.h
+++ b/proto/radv/radv.h
@@ -60,14 +60,14 @@ struct radv_iface_config
   list rdnss_list;		/* Local list of RDNSS configs (struct radv_rdnss_config) */
   list dnssl_list;		/* Local list of DNSSL configs (struct radv_dnssl_config) */
 
-  u32 min_ra_int;		/* Standard options from RFC 4261 */
+  u32 min_ra_int;		/* Standard options from RFC 4861 */
   u32 max_ra_int;
   u32 min_delay;
 
   u8 rdnss_local;		/* Global list is not used for RDNSS */
   u8 dnssl_local;		/* Global list is not used for DNSSL */
 
-  u8 managed;			/* Standard options from RFC 4261 */
+  u8 managed;			/* Standard options from RFC 4861 */
   u8 other_config;
   u32 link_mtu;
   u32 reachable_time;
@@ -84,7 +84,7 @@ struct radv_prefix_config
   net_addr_ip6 prefix;
 
   u8 skip;			/* Do not include this prefix to RA */
-  u8 onlink;			/* Standard options from RFC 4261 */
+  u8 onlink;			/* Standard options from RFC 4861 */
   u8 autonomous;
   u32 valid_lifetime;
   u32 preferred_lifetime;

From 637ed49868c56c0e05467f0e0ddb6aa1231deaa4 Mon Sep 17 00:00:00 2001
From: Michal 'vorner' Vaner <michal.vaner@nic.cz>
Date: Fri, 4 Aug 2017 10:52:57 +0200
Subject: [PATCH 004/198] radv: Fix RFC reference in comments

---
 proto/radv/radv.h | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/proto/radv/radv.h b/proto/radv/radv.h
index e2bf07ba..0591486b 100644
--- a/proto/radv/radv.h
+++ b/proto/radv/radv.h
@@ -62,14 +62,14 @@ struct radv_iface_config
   list rdnss_list;		/* Local list of RDNSS configs (struct radv_rdnss_config) */
   list dnssl_list;		/* Local list of DNSSL configs (struct radv_dnssl_config) */
 
-  u32 min_ra_int;		/* Standard options from RFC 4261 */
+  u32 min_ra_int;		/* Standard options from RFC 4861 */
   u32 max_ra_int;
   u32 min_delay;
 
   u8 rdnss_local;		/* Global list is not used for RDNSS */
   u8 dnssl_local;		/* Global list is not used for DNSSL */
 
-  u8 managed;			/* Standard options from RFC 4261 */
+  u8 managed;			/* Standard options from RFC 4861 */
   u8 other_config;
   u32 link_mtu;
   u32 reachable_time;
@@ -87,7 +87,7 @@ struct radv_prefix_config
   uint pxlen;
 
   u8 skip;			/* Do not include this prefix to RA */
-  u8 onlink;			/* Standard options from RFC 4261 */
+  u8 onlink;			/* Standard options from RFC 4861 */
   u8 autonomous;
   u32 valid_lifetime;
   u32 preferred_lifetime;

From 5a41eed26d1b12861ba0ecddcd7cade335d2a192 Mon Sep 17 00:00:00 2001
From: Michal 'vorner' Vaner <michal.vaner@nic.cz>
Date: Tue, 8 Aug 2017 14:40:51 +0200
Subject: [PATCH 005/198] RAdv: Style updates

Adapt the naming conventions to be a bit closer to the other protocols.

proto_radv -> radv_proto
struct radv_proto *ra -> struct radv_proto *p
struct proto *p -> struct proto *P
---
 proto/radv/packets.c |  22 ++++----
 proto/radv/radv.c    | 126 +++++++++++++++++++++----------------------
 proto/radv/radv.h    |   8 +--
 3 files changed, 78 insertions(+), 78 deletions(-)

diff --git a/proto/radv/packets.c b/proto/radv/packets.c
index f4352155..15ca5738 100644
--- a/proto/radv/packets.c
+++ b/proto/radv/packets.c
@@ -82,8 +82,8 @@ static struct radv_prefix_config default_prefix = {
 static struct radv_prefix_config *
 radv_prefix_match(struct radv_iface *ifa, struct ifa *a)
 {
-  struct proto *p = &ifa->ra->p;
-  struct radv_config *cf = (struct radv_config *) (p->cf);
+  struct radv_proto *p = ifa->ra;
+  struct radv_config *cf = (struct radv_config *) (p->p.cf);
   struct radv_prefix_config *pc;
 
   if (a->scope <= SCOPE_LINK)
@@ -206,7 +206,7 @@ radv_prepare_dnssl(struct radv_iface *ifa, list *dnssl_list, char **buf, char *b
     else
       op->lifetime = htonl(dcf->lifetime);
 
-    while(NODE_VALID(dcf) && 
+    while(NODE_VALID(dcf) &&
 	  (dcf->lifetime == dcf_base->lifetime) &&
 	  (dcf->lifetime_mult == dcf_base->lifetime_mult))
       {
@@ -237,8 +237,8 @@ radv_prepare_dnssl(struct radv_iface *ifa, list *dnssl_list, char **buf, char *b
 static void
 radv_prepare_ra(struct radv_iface *ifa)
 {
-  struct proto_radv *ra = ifa->ra;
-  struct radv_config *cf = (struct radv_config *) (ra->p.cf);
+  struct radv_proto *p = ifa->ra;
+  struct radv_config *cf = (struct radv_config *) (p->p.cf);
   struct radv_iface_config *ic = ifa->cf;
 
   char *buf = ifa->sk->tbuf;
@@ -250,7 +250,7 @@ radv_prepare_ra(struct radv_iface *ifa)
   pkt->code = 0;
   pkt->checksum = 0;
   pkt->current_hop_limit = ic->current_hop_limit;
-  pkt->router_lifetime = (ra->active || !ic->default_lifetime_sensitive) ?
+  pkt->router_lifetime = (p->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) |
@@ -283,7 +283,7 @@ radv_prepare_ra(struct radv_iface *ifa)
 
     if (buf + sizeof(struct radv_opt_prefix) > bufend)
     {
-      log(L_WARN "%s: Too many prefixes on interface %s", ra->p.name, ifa->iface->name);
+      log(L_WARN "%s: Too many prefixes on interface %s", p->p.name, ifa->iface->name);
       goto done;
     }
 
@@ -293,9 +293,9 @@ radv_prepare_ra(struct radv_iface *ifa)
     op->pxlen = net6_pxlen(&addr->prefix);
     op->flags = (pc->onlink ? OPT_PX_ONLINK : 0) |
       (pc->autonomous ? OPT_PX_AUTONOMOUS : 0);
-    op->valid_lifetime = (ra->active || !pc->valid_lifetime_sensitive) ?
+    op->valid_lifetime = (p->active || !pc->valid_lifetime_sensitive) ?
       htonl(pc->valid_lifetime) : 0;
-    op->preferred_lifetime = (ra->active || !pc->preferred_lifetime_sensitive) ?
+    op->preferred_lifetime = (p->active || !pc->preferred_lifetime_sensitive) ?
       htonl(pc->preferred_lifetime) : 0;
     op->reserved = 0;
     op->prefix = ip6_hton(net6_prefix(&addr->prefix));
@@ -324,7 +324,7 @@ radv_prepare_ra(struct radv_iface *ifa)
 void
 radv_send_ra(struct radv_iface *ifa, int shutdown)
 {
-  struct proto_radv *ra = ifa->ra;
+  struct radv_proto *p = ifa->ra;
 
   /* We store prepared RA in tbuf */
   if (!ifa->plen)
@@ -352,7 +352,7 @@ static int
 radv_rx_hook(sock *sk, uint size)
 {
   struct radv_iface *ifa = sk->data;
-  struct proto_radv *ra = ifa->ra;
+  struct radv_proto *p = ifa->ra;
 
   /* We want just packets from sk->iface */
   if (sk->lifindex != sk->iface->index)
diff --git a/proto/radv/radv.c b/proto/radv/radv.c
index d7d975ab..4a6593b8 100644
--- a/proto/radv/radv.c
+++ b/proto/radv/radv.c
@@ -47,7 +47,7 @@ static void
 radv_timer(timer *tm)
 {
   struct radv_iface *ifa = tm->data;
-  struct proto_radv *ra = ifa->ra;
+  struct radv_proto *p = ifa->ra;
 
   RADV_TRACE(D_EVENTS, "Timer fired on %s", ifa->iface->name);
 
@@ -72,7 +72,7 @@ static char* ev_name[] = { NULL, "Init", "Change", "RS" };
 void
 radv_iface_notify(struct radv_iface *ifa, int event)
 {
-  struct proto_radv *ra = ifa->ra;
+  struct radv_proto *p = ifa->ra;
 
   if (!ifa->sk)
     return;
@@ -102,21 +102,21 @@ radv_iface_notify(struct radv_iface *ifa, int event)
 }
 
 static void
-radv_iface_notify_all(struct proto_radv *ra, int event)
+radv_iface_notify_all(struct radv_proto *p, int event)
 {
   struct radv_iface *ifa;
 
-  WALK_LIST(ifa, ra->iface_list)
+  WALK_LIST(ifa, p->iface_list)
     radv_iface_notify(ifa, event);
 }
 
 
 static struct radv_iface *
-radv_iface_find(struct proto_radv *ra, struct iface *what)
+radv_iface_find(struct radv_proto *p, struct iface *what)
 {
   struct radv_iface *ifa;
 
-  WALK_LIST(ifa, ra->iface_list)
+  WALK_LIST(ifa, p->iface_list)
     if (ifa->iface == what)
       return ifa;
 
@@ -127,11 +127,11 @@ static void
 radv_iface_add(struct object_lock *lock)
 {
   struct radv_iface *ifa = lock->data;
-  struct proto_radv *ra = ifa->ra;
+  struct radv_proto *p = ifa->ra;
 
   if (! radv_sk_open(ifa))
   {
-    log(L_ERR "%s: Socket open failed on interface %s", ra->p.name, ifa->iface->name);
+    log(L_ERR "%s: Socket open failed on interface %s", p->p.name, ifa->iface->name);
     return;
   }
 
@@ -150,24 +150,24 @@ find_lladdr(struct iface *iface)
 }
 
 static void
-radv_iface_new(struct proto_radv *ra, struct iface *iface, struct radv_iface_config *cf)
+radv_iface_new(struct radv_proto *p, struct iface *iface, struct radv_iface_config *cf)
 {
-  pool *pool = ra->p.pool;
+  pool *pool = p->p.pool;
   struct radv_iface *ifa;
 
   RADV_TRACE(D_EVENTS, "Adding interface %s", iface->name);
 
   ifa = mb_allocz(pool, sizeof(struct radv_iface));
-  ifa->ra = ra;
+  ifa->ra = p;
   ifa->cf = cf;
   ifa->iface = iface;
 
-  add_tail(&ra->iface_list, NODE ifa);
+  add_tail(&p->iface_list, NODE ifa);
 
   ifa->addr = find_lladdr(iface);
   if (!ifa->addr)
   {
-    log(L_ERR "%s: Cannot find link-locad addr on interface %s", ra->p.name, iface->name);
+    log(L_ERR "%s: Cannot find link-locad addr on interface %s", p->p.name, iface->name);
     return;
   }
 
@@ -193,7 +193,7 @@ radv_iface_new(struct proto_radv *ra, struct iface *iface, struct radv_iface_con
 static void
 radv_iface_remove(struct radv_iface *ifa)
 {
-  struct proto_radv *ra = ifa->ra;
+  struct radv_proto *p = ifa->ra;
   RADV_TRACE(D_EVENTS, "Removing interface %s", ifa->iface->name);
 
   rem_node(NODE ifa);
@@ -206,10 +206,10 @@ radv_iface_remove(struct radv_iface *ifa)
 }
 
 static void
-radv_if_notify(struct proto *p, unsigned flags, struct iface *iface)
+radv_if_notify(struct proto *P, unsigned flags, struct iface *iface)
 {
-  struct proto_radv *ra = (struct proto_radv *) p;
-  struct radv_config *cf = (struct radv_config *) (p->cf);
+  struct radv_proto *p = (struct radv_proto *) P;
+  struct radv_config *cf = (struct radv_config *) (P->cf);
 
   if (iface->flags & IF_IGNORE)
     return;
@@ -220,12 +220,12 @@ radv_if_notify(struct proto *p, unsigned flags, struct iface *iface)
       iface_patt_find(&cf->patt_list, iface, NULL);
 
     if (ic)
-      radv_iface_new(ra, iface, ic);
+      radv_iface_new(p, iface, ic);
 
     return;
   }
 
-  struct radv_iface *ifa = radv_iface_find(ra, iface);
+  struct radv_iface *ifa = radv_iface_find(p, iface);
   if (!ifa)
     return;
 
@@ -240,9 +240,9 @@ radv_if_notify(struct proto *p, unsigned flags, struct iface *iface)
 }
 
 static void
-radv_ifa_notify(struct proto *p, unsigned flags UNUSED, struct ifa *a)
+radv_ifa_notify(struct proto *P, unsigned flags UNUSED, struct ifa *a)
 {
-  struct proto_radv *ra = (struct proto_radv *) p;
+  struct radv_proto *p = (struct radv_proto *) P;
 
   if (a->flags & IA_SECONDARY)
     return;
@@ -250,7 +250,7 @@ radv_ifa_notify(struct proto *p, unsigned flags UNUSED, struct ifa *a)
   if (a->scope <= SCOPE_LINK)
     return;
 
-  struct radv_iface *ifa = radv_iface_find(ra, a->iface);
+  struct radv_iface *ifa = radv_iface_find(p, a->iface);
 
   if (ifa)
     radv_iface_notify(ifa, RA_EV_CHANGE);
@@ -269,10 +269,10 @@ radv_net_match_trigger(struct radv_config *cf, net *n)
 }
 
 int
-radv_import_control(struct proto *p, rte **new, ea_list **attrs UNUSED, struct linpool *pool UNUSED)
+radv_import_control(struct proto *P, rte **new, ea_list **attrs UNUSED, struct linpool *pool UNUSED)
 {
-  // struct proto_radv *ra = (struct proto_radv *) p;
-  struct radv_config *cf = (struct radv_config *) (p->cf);
+  // struct radv_proto *p = (struct radv_proto *) P;
+  struct radv_config *cf = (struct radv_config *) (P->cf);
 
   if (radv_net_match_trigger(cf, (*new)->net))
     return RIC_PROCESS;
@@ -281,38 +281,38 @@ radv_import_control(struct proto *p, rte **new, ea_list **attrs UNUSED, struct l
 }
 
 static void
-radv_rt_notify(struct proto *p, struct channel *ch UNUSED, net *n, rte *new, rte *old UNUSED, ea_list *attrs UNUSED)
+radv_rt_notify(struct proto *P, struct channel *ch UNUSED, net *n, rte *new, rte *old UNUSED, ea_list *attrs UNUSED)
 {
-  struct proto_radv *ra = (struct proto_radv *) p;
-  struct radv_config *cf = (struct radv_config *) (p->cf);
+  struct radv_proto *p = (struct radv_proto *) P;
+  struct radv_config *cf = (struct radv_config *) (P->cf);
 
   if (radv_net_match_trigger(cf, n))
   {
-    u8 old_active = ra->active;
-    ra->active = !!new;
+    u8 old_active = p->active;
+    p->active = !!new;
 
-    if (ra->active == old_active)
+    if (p->active == old_active)
       return;
 
-    if (ra->active)
+    if (p->active)
       RADV_TRACE(D_EVENTS, "Triggered");
     else
       RADV_TRACE(D_EVENTS, "Suppressed");
 
-    radv_iface_notify_all(ra, RA_EV_CHANGE);
+    radv_iface_notify_all(p, RA_EV_CHANGE);
   }
 }
 
 static int
-radv_check_active(struct proto_radv *ra)
+radv_check_active(struct radv_proto *p)
 {
-  struct radv_config *cf = (struct radv_config *) (ra->p.cf);
+  struct radv_config *cf = (struct radv_config *) (p->p.cf);
 
   if (!radv_trigger_valid(cf))
     return 1;
 
-  struct channel *c = ra->p.main_channel;
-  return rt_examine(c->table, &cf->trigger, &ra->p, c->out_filter);
+  struct channel *c = p->p.main_channel;
+  return rt_examine(c->table, &cf->trigger, &p->p, c->out_filter);
 }
 
 static void
@@ -328,26 +328,26 @@ radv_postconfig(struct proto_config *CF)
 static struct proto *
 radv_init(struct proto_config *CF)
 {
-  struct proto *p = proto_new(CF);
+  struct proto *P = proto_new(CF);
 
-  p->main_channel = proto_add_channel(p, proto_cf_main_channel(CF));
+  P->main_channel = proto_add_channel(P, proto_cf_main_channel(CF));
 
-  p->import_control = radv_import_control;
-  p->rt_notify = radv_rt_notify;
-  p->if_notify = radv_if_notify;
-  p->ifa_notify = radv_ifa_notify;
+  P->import_control = radv_import_control;
+  P->rt_notify = radv_rt_notify;
+  P->if_notify = radv_if_notify;
+  P->ifa_notify = radv_ifa_notify;
 
-  return p;
+  return P;
 }
 
 static int
-radv_start(struct proto *p)
+radv_start(struct proto *P)
 {
-  struct proto_radv *ra = (struct proto_radv *) p;
-  struct radv_config *cf = (struct radv_config *) (p->cf);
+  struct radv_proto *p = (struct radv_proto *) P;
+  struct radv_config *cf = (struct radv_config *) (P->cf);
 
-  init_list(&(ra->iface_list));
-  ra->active = !radv_trigger_valid(cf);
+  init_list(&(p->iface_list));
+  p->active = !radv_trigger_valid(cf);
 
   return PS_UP;
 }
@@ -360,21 +360,21 @@ radv_iface_shutdown(struct radv_iface *ifa)
 }
 
 static int
-radv_shutdown(struct proto *p)
+radv_shutdown(struct proto *P)
 {
-  struct proto_radv *ra = (struct proto_radv *) p;
+  struct radv_proto *p = (struct radv_proto *) P;
 
   struct radv_iface *ifa;
-  WALK_LIST(ifa, ra->iface_list)
+  WALK_LIST(ifa, p->iface_list)
     radv_iface_shutdown(ifa);
 
   return PS_DOWN;
 }
 
 static int
-radv_reconfigure(struct proto *p, struct proto_config *CF)
+radv_reconfigure(struct proto *P, struct proto_config *CF)
 {
-  struct proto_radv *ra = (struct proto_radv *) p;
+  struct radv_proto *p = (struct radv_proto *) P;
   // struct radv_config *old = (struct radv_config *) (p->cf);
   struct radv_config *new = (struct radv_config *) CF;
 
@@ -386,16 +386,16 @@ radv_reconfigure(struct proto *p, struct proto_config *CF)
    * causing nodes to temporary remove their default routes.
    */
 
-  if (!proto_configure_channel(p, &p->main_channel, proto_cf_main_channel(CF)))
+  if (!proto_configure_channel(P, &P->main_channel, proto_cf_main_channel(CF)))
     return 0;
 
-  p->cf = CF; /* radv_check_active() requires proper p->cf */
-  ra->active = radv_check_active(ra);
+  P->cf = CF; /* radv_check_active() requires proper p->cf */
+  p->active = radv_check_active(p);
 
   struct iface *iface;
   WALK_LIST(iface, iface_list)
   {
-    struct radv_iface *ifa = radv_iface_find(ra, iface);
+    struct radv_iface *ifa = radv_iface_find(p, iface);
     struct radv_iface_config *ic = (struct radv_iface_config *)
       iface_patt_find(&new->patt_list, iface, NULL);
 
@@ -415,7 +415,7 @@ radv_reconfigure(struct proto *p, struct proto_config *CF)
     }
 
     if (!ifa && ic)
-      radv_iface_new(ra, iface, ic);
+      radv_iface_new(p, iface, ic);
   }
 
   return 1;
@@ -435,11 +435,11 @@ radv_copy_config(struct proto_config *dest, struct proto_config *src)
 }
 
 static void
-radv_get_status(struct proto *p, byte *buf)
+radv_get_status(struct proto *P, byte *buf)
 {
-  struct proto_radv *ra = (struct proto_radv *) p;
+  struct radv_proto *p = (struct radv_proto *) P;
 
-  if (!ra->active)
+  if (!p->active)
     strcpy(buf, "Suppressed");
 }
 
@@ -447,7 +447,7 @@ struct protocol proto_radv = {
   .name =		"RAdv",
   .template =		"radv%d",
   .channel_mask =	NB_IP6,
-  .proto_size =		sizeof(struct proto_radv),
+  .proto_size =		sizeof(struct radv_proto),
   .config_size =	sizeof(struct radv_config),
   .postconfig =		radv_postconfig,
   .init =		radv_init,
diff --git a/proto/radv/radv.h b/proto/radv/radv.h
index 3e6c1651..35eea9ba 100644
--- a/proto/radv/radv.h
+++ b/proto/radv/radv.h
@@ -111,7 +111,7 @@ struct radv_dnssl_config
 };
 
 
-struct proto_radv
+struct radv_proto
 {
   struct proto p;
   list iface_list;		/* List of active ifaces */
@@ -121,7 +121,7 @@ struct proto_radv
 struct radv_iface
 {
   node n;
-  struct proto_radv *ra;
+  struct radv_proto *ra;
   struct radv_iface_config *cf;	/* Related config, must be updated in reconfigure */
   struct iface *iface;
   struct ifa *addr;		/* Link-local address of iface */
@@ -151,8 +151,8 @@ struct radv_iface
 #else
 #define RADV_FORCE_DEBUG 0
 #endif
-#define RADV_TRACE(flags, msg, args...) do { if ((ra->p.debug & flags) || RADV_FORCE_DEBUG) \
-	log(L_TRACE "%s: " msg, ra->p.name , ## args ); } while(0)
+#define RADV_TRACE(flags, msg, args...) do { if ((p->p.debug & flags) || RADV_FORCE_DEBUG) \
+	log(L_TRACE "%s: " msg, p->p.name , ## args ); } while(0)
 
 
 /* radv.c */

From 1f182675c816e4708312f99dd817fb894f3a90b3 Mon Sep 17 00:00:00 2001
From: Michal 'vorner' Vaner <michal.vaner@nic.cz>
Date: Tue, 8 Aug 2017 14:40:51 +0200
Subject: [PATCH 006/198] RAdv: Style updates

Adapt the naming conventions to be a bit closer to the other protocols.

proto_radv -> radv_proto
struct radv_proto *ra -> struct radv_proto *p
struct proto *p -> struct proto *P
---
 proto/radv/packets.c |  22 ++++----
 proto/radv/radv.c    | 123 ++++++++++++++++++++++---------------------
 proto/radv/radv.h    |   8 +--
 3 files changed, 77 insertions(+), 76 deletions(-)

diff --git a/proto/radv/packets.c b/proto/radv/packets.c
index 4bf960c3..1c9837a4 100644
--- a/proto/radv/packets.c
+++ b/proto/radv/packets.c
@@ -82,8 +82,8 @@ static struct radv_prefix_config default_prefix = {
 static struct radv_prefix_config *
 radv_prefix_match(struct radv_iface *ifa, struct ifa *a)
 {
-  struct proto *p = &ifa->ra->p;
-  struct radv_config *cf = (struct radv_config *) (p->cf);
+  struct radv_proto *p = ifa->ra;
+  struct radv_config *cf = (struct radv_config *) (p->p.cf);
   struct radv_prefix_config *pc;
 
   if (a->scope <= SCOPE_LINK)
@@ -207,7 +207,7 @@ radv_prepare_dnssl(struct radv_iface *ifa, list *dnssl_list, char **buf, char *b
     else
       op->lifetime = htonl(dcf->lifetime);
 
-    while(NODE_VALID(dcf) && 
+    while(NODE_VALID(dcf) &&
 	  (dcf->lifetime == dcf_base->lifetime) &&
 	  (dcf->lifetime_mult == dcf_base->lifetime_mult))
       {
@@ -238,8 +238,8 @@ radv_prepare_dnssl(struct radv_iface *ifa, list *dnssl_list, char **buf, char *b
 static void
 radv_prepare_ra(struct radv_iface *ifa)
 {
-  struct proto_radv *ra = ifa->ra;
-  struct radv_config *cf = (struct radv_config *) (ra->p.cf);
+  struct radv_proto *p = ifa->ra;
+  struct radv_config *cf = (struct radv_config *) (p->p.cf);
   struct radv_iface_config *ic = ifa->cf;
 
   char *buf = ifa->sk->tbuf;
@@ -251,7 +251,7 @@ radv_prepare_ra(struct radv_iface *ifa)
   pkt->code = 0;
   pkt->checksum = 0;
   pkt->current_hop_limit = ic->current_hop_limit;
-  pkt->router_lifetime = (ra->active || !ic->default_lifetime_sensitive) ?
+  pkt->router_lifetime = (p->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) |
@@ -281,7 +281,7 @@ radv_prepare_ra(struct radv_iface *ifa)
 
     if (buf + sizeof(struct radv_opt_prefix) > bufend)
     {
-      log(L_WARN "%s: Too many prefixes on interface %s", ra->p.name, ifa->iface->name);
+      log(L_WARN "%s: Too many prefixes on interface %s", p->p.name, ifa->iface->name);
       goto done;
     }
 
@@ -291,9 +291,9 @@ radv_prepare_ra(struct radv_iface *ifa)
     op->pxlen = addr->pxlen;
     op->flags = (pc->onlink ? OPT_PX_ONLINK : 0) |
       (pc->autonomous ? OPT_PX_AUTONOMOUS : 0);
-    op->valid_lifetime = (ra->active || !pc->valid_lifetime_sensitive) ?
+    op->valid_lifetime = (p->active || !pc->valid_lifetime_sensitive) ?
       htonl(pc->valid_lifetime) : 0;
-    op->preferred_lifetime = (ra->active || !pc->preferred_lifetime_sensitive) ?
+    op->preferred_lifetime = (p->active || !pc->preferred_lifetime_sensitive) ?
       htonl(pc->preferred_lifetime) : 0;
     op->reserved = 0;
     op->prefix = addr->prefix;
@@ -323,7 +323,7 @@ radv_prepare_ra(struct radv_iface *ifa)
 void
 radv_send_ra(struct radv_iface *ifa, int shutdown)
 {
-  struct proto_radv *ra = ifa->ra;
+  struct radv_proto *p = ifa->ra;
 
   /* We store prepared RA in tbuf */
   if (!ifa->plen)
@@ -351,7 +351,7 @@ static int
 radv_rx_hook(sock *sk, uint size)
 {
   struct radv_iface *ifa = sk->data;
-  struct proto_radv *ra = ifa->ra;
+  struct radv_proto *p = ifa->ra;
 
   /* We want just packets from sk->iface */
   if (sk->lifindex != sk->iface->index)
diff --git a/proto/radv/radv.c b/proto/radv/radv.c
index 5c52217d..4ea1f5c2 100644
--- a/proto/radv/radv.c
+++ b/proto/radv/radv.c
@@ -47,7 +47,7 @@ static void
 radv_timer(timer *tm)
 {
   struct radv_iface *ifa = tm->data;
-  struct proto_radv *ra = ifa->ra;
+  struct radv_proto *p = ifa->ra;
 
   RADV_TRACE(D_EVENTS, "Timer fired on %s", ifa->iface->name);
 
@@ -72,7 +72,7 @@ static char* ev_name[] = { NULL, "Init", "Change", "RS" };
 void
 radv_iface_notify(struct radv_iface *ifa, int event)
 {
-  struct proto_radv *ra = ifa->ra;
+  struct radv_proto *p = ifa->ra;
 
   if (!ifa->sk)
     return;
@@ -102,21 +102,21 @@ radv_iface_notify(struct radv_iface *ifa, int event)
 }
 
 static void
-radv_iface_notify_all(struct proto_radv *ra, int event)
+radv_iface_notify_all(struct radv_proto *p, int event)
 {
   struct radv_iface *ifa;
 
-  WALK_LIST(ifa, ra->iface_list)
+  WALK_LIST(ifa, p->iface_list)
     radv_iface_notify(ifa, event);
 }
 
 
 static struct radv_iface *
-radv_iface_find(struct proto_radv *ra, struct iface *what)
+radv_iface_find(struct radv_proto *p, struct iface *what)
 {
   struct radv_iface *ifa;
 
-  WALK_LIST(ifa, ra->iface_list)
+  WALK_LIST(ifa, p->iface_list)
     if (ifa->iface == what)
       return ifa;
 
@@ -127,11 +127,11 @@ static void
 radv_iface_add(struct object_lock *lock)
 {
   struct radv_iface *ifa = lock->data;
-  struct proto_radv *ra = ifa->ra;
+  struct radv_proto *p = ifa->ra;
 
   if (! radv_sk_open(ifa))
   {
-    log(L_ERR "%s: Socket open failed on interface %s", ra->p.name, ifa->iface->name);
+    log(L_ERR "%s: Socket open failed on interface %s", p->p.name, ifa->iface->name);
     return;
   }
 
@@ -150,24 +150,24 @@ find_lladdr(struct iface *iface)
 }
 
 static void
-radv_iface_new(struct proto_radv *ra, struct iface *iface, struct radv_iface_config *cf)
+radv_iface_new(struct radv_proto *p, struct iface *iface, struct radv_iface_config *cf)
 {
-  pool *pool = ra->p.pool;
+  pool *pool = p->p.pool;
   struct radv_iface *ifa;
 
   RADV_TRACE(D_EVENTS, "Adding interface %s", iface->name);
 
   ifa = mb_allocz(pool, sizeof(struct radv_iface));
-  ifa->ra = ra;
+  ifa->ra = p;
   ifa->cf = cf;
   ifa->iface = iface;
 
-  add_tail(&ra->iface_list, NODE ifa);
+  add_tail(&p->iface_list, NODE ifa);
 
   ifa->addr = find_lladdr(iface);
   if (!ifa->addr)
   {
-    log(L_ERR "%s: Cannot find link-locad addr on interface %s", ra->p.name, iface->name);
+    log(L_ERR "%s: Cannot find link-locad addr on interface %s", p->p.name, iface->name);
     return;
   }
 
@@ -193,7 +193,7 @@ radv_iface_new(struct proto_radv *ra, struct iface *iface, struct radv_iface_con
 static void
 radv_iface_remove(struct radv_iface *ifa)
 {
-  struct proto_radv *ra = ifa->ra;
+  struct radv_proto *p = ifa->ra;
   RADV_TRACE(D_EVENTS, "Removing interface %s", ifa->iface->name);
 
   rem_node(NODE ifa);
@@ -206,10 +206,10 @@ radv_iface_remove(struct radv_iface *ifa)
 }
 
 static void
-radv_if_notify(struct proto *p, unsigned flags, struct iface *iface)
+radv_if_notify(struct proto *P, unsigned flags, struct iface *iface)
 {
-  struct proto_radv *ra = (struct proto_radv *) p;
-  struct radv_config *cf = (struct radv_config *) (p->cf);
+  struct radv_proto *p = (struct radv_proto *) P;
+  struct radv_config *cf = (struct radv_config *) (P->cf);
 
   if (iface->flags & IF_IGNORE)
     return;
@@ -220,12 +220,12 @@ radv_if_notify(struct proto *p, unsigned flags, struct iface *iface)
       iface_patt_find(&cf->patt_list, iface, NULL);
 
     if (ic)
-      radv_iface_new(ra, iface, ic);
+      radv_iface_new(p, iface, ic);
 
     return;
   }
 
-  struct radv_iface *ifa = radv_iface_find(ra, iface);
+  struct radv_iface *ifa = radv_iface_find(p, iface);
   if (!ifa)
     return;
 
@@ -240,9 +240,9 @@ radv_if_notify(struct proto *p, unsigned flags, struct iface *iface)
 }
 
 static void
-radv_ifa_notify(struct proto *p, unsigned flags UNUSED, struct ifa *a)
+radv_ifa_notify(struct proto *P, unsigned flags UNUSED, struct ifa *a)
 {
-  struct proto_radv *ra = (struct proto_radv *) p;
+  struct radv_proto *p = (struct radv_proto *) P;
 
   if (a->flags & IA_SECONDARY)
     return;
@@ -250,7 +250,7 @@ radv_ifa_notify(struct proto *p, unsigned flags UNUSED, struct ifa *a)
   if (a->scope <= SCOPE_LINK)
     return;
 
-  struct radv_iface *ifa = radv_iface_find(ra, a->iface);
+  struct radv_iface *ifa = radv_iface_find(p, a->iface);
 
   if (ifa)
     radv_iface_notify(ifa, RA_EV_CHANGE);
@@ -264,10 +264,10 @@ static inline int radv_net_match_trigger(struct radv_config *cf, net *n)
 }
 
 int
-radv_import_control(struct proto *p, rte **new, ea_list **attrs UNUSED, struct linpool *pool UNUSED)
+radv_import_control(struct proto *P, rte **new, ea_list **attrs UNUSED, struct linpool *pool UNUSED)
 {
-  // struct proto_radv *ra = (struct proto_radv *) p;
-  struct radv_config *cf = (struct radv_config *) (p->cf);
+  // struct radv_proto *p = (struct radv_proto *) P;
+  struct radv_config *cf = (struct radv_config *) (P->cf);
 
   if (radv_net_match_trigger(cf, (*new)->net))
     return RIC_PROCESS;
@@ -276,61 +276,62 @@ radv_import_control(struct proto *p, rte **new, ea_list **attrs UNUSED, struct l
 }
 
 static void
-radv_rt_notify(struct proto *p, rtable *tbl UNUSED, net *n, rte *new, rte *old UNUSED, ea_list *attrs UNUSED)
+radv_rt_notify(struct proto *P, rtable *tbl UNUSED, net *n, rte *new, rte *old UNUSED, ea_list *attrs UNUSED)
 {
-  struct proto_radv *ra = (struct proto_radv *) p;
-  struct radv_config *cf = (struct radv_config *) (p->cf);
+  struct radv_proto *p = (struct radv_proto *) P;
+  struct radv_config *cf = (struct radv_config *) (P->cf);
 
   if (radv_net_match_trigger(cf, n))
   {
-    u8 old_active = ra->active;
-    ra->active = !!new;
+    u8 old_active = p->active;
+    p->active = !!new;
 
-    if (ra->active == old_active)
+    if (p->active == old_active)
       return;
 
-    if (ra->active)
+    if (p->active)
       RADV_TRACE(D_EVENTS, "Triggered");
     else
       RADV_TRACE(D_EVENTS, "Suppressed");
 
-    radv_iface_notify_all(ra, RA_EV_CHANGE);
+    radv_iface_notify_all(p, RA_EV_CHANGE);
   }
 }
 
 static int
-radv_check_active(struct proto_radv *ra)
+radv_check_active(struct radv_proto *p)
 {
-  struct radv_config *cf = (struct radv_config *) (ra->p.cf);
+  struct radv_config *cf = (struct radv_config *) (p->p.cf);
 
   if (! cf->trigger_valid)
     return 1;
 
-  return rt_examine(ra->p.table, cf->trigger_prefix, cf->trigger_pxlen,
-		    &(ra->p), ra->p.cf->out_filter);
+  return rt_examine(p->p.table, cf->trigger_prefix, cf->trigger_pxlen,
+		    &(p->p), p->p.cf->out_filter);
 }
 
 static struct proto *
 radv_init(struct proto_config *c)
 {
-  struct proto *p = proto_new(c, sizeof(struct proto_radv));
+  struct proto *P = proto_new(c, sizeof(struct radv_proto));
 
-  p->accept_ra_types = RA_OPTIMAL;
-  p->import_control = radv_import_control;
-  p->rt_notify = radv_rt_notify;
-  p->if_notify = radv_if_notify;
-  p->ifa_notify = radv_ifa_notify;
-  return p;
+  P->accept_ra_types = RA_OPTIMAL;
+  P->import_control = radv_import_control;
+  P->rt_notify = radv_rt_notify;
+  P->if_notify = radv_if_notify;
+  P->ifa_notify = radv_ifa_notify;
+
+  return P;
 }
 
 static int
-radv_start(struct proto *p)
+radv_start(struct proto *P)
 {
-  struct proto_radv *ra = (struct proto_radv *) p;
-  struct radv_config *cf = (struct radv_config *) (p->cf);
+  struct radv_proto *p = (struct radv_proto *) P;
+  struct radv_config *cf = (struct radv_config *) (P->cf);
 
-  init_list(&(ra->iface_list));
-  ra->active = !cf->trigger_valid;
+  init_list(&(p->iface_list));
+  p->active = !cf->trigger_valid;
 
   return PS_UP;
 }
@@ -343,21 +344,21 @@ radv_iface_shutdown(struct radv_iface *ifa)
 }
 
 static int
-radv_shutdown(struct proto *p)
+radv_shutdown(struct proto *P)
 {
-  struct proto_radv *ra = (struct proto_radv *) p;
+  struct radv_proto *p = (struct radv_proto *) P;
 
   struct radv_iface *ifa;
-  WALK_LIST(ifa, ra->iface_list)
+  WALK_LIST(ifa, p->iface_list)
     radv_iface_shutdown(ifa);
 
   return PS_DOWN;
 }
 
 static int
-radv_reconfigure(struct proto *p, struct proto_config *c)
+radv_reconfigure(struct proto *P, struct proto_config *c)
 {
-  struct proto_radv *ra = (struct proto_radv *) p;
+  struct radv_proto *p = (struct radv_proto *) P;
   // struct radv_config *old = (struct radv_config *) (p->cf);
   struct radv_config *new = (struct radv_config *) c;
 
@@ -369,13 +370,13 @@ radv_reconfigure(struct proto *p, struct proto_config *c)
    * causing nodes to temporary remove their default routes.
    */
 
-  p->cf = c; /* radv_check_active() requires proper p->cf */
-  ra->active = radv_check_active(ra);
+  P->cf = c; /* radv_check_active() requires proper P->cf */
+  p->active = radv_check_active(p);
 
   struct iface *iface;
   WALK_LIST(iface, iface_list)
   {
-    struct radv_iface *ifa = radv_iface_find(ra, iface);
+    struct radv_iface *ifa = radv_iface_find(p, iface);
     struct radv_iface_config *ic = (struct radv_iface_config *)
       iface_patt_find(&new->patt_list, iface, NULL);
 
@@ -395,7 +396,7 @@ radv_reconfigure(struct proto *p, struct proto_config *c)
     }
 
     if (!ifa && ic)
-      radv_iface_new(ra, iface, ic);
+      radv_iface_new(p, iface, ic);
   }
 
   return 1;
@@ -415,11 +416,11 @@ radv_copy_config(struct proto_config *dest, struct proto_config *src)
 }
 
 static void
-radv_get_status(struct proto *p, byte *buf)
+radv_get_status(struct proto *P, byte *buf)
 {
-  struct proto_radv *ra = (struct proto_radv *) p;
+  struct radv_proto *p = (struct radv_proto *) P;
 
-  if (!ra->active)
+  if (!p->active)
     strcpy(buf, "Suppressed");
 }
 
diff --git a/proto/radv/radv.h b/proto/radv/radv.h
index 0591486b..559a3f3f 100644
--- a/proto/radv/radv.h
+++ b/proto/radv/radv.h
@@ -114,7 +114,7 @@ struct radv_dnssl_config
 };
 
 
-struct proto_radv
+struct radv_proto
 {
   struct proto p;
   list iface_list;		/* List of active ifaces */
@@ -124,7 +124,7 @@ struct proto_radv
 struct radv_iface
 {
   node n;
-  struct proto_radv *ra;
+  struct radv_proto *ra;
   struct radv_iface_config *cf;	/* Related config, must be updated in reconfigure */
   struct iface *iface;
   struct ifa *addr;		/* Link-local address of iface */
@@ -154,8 +154,8 @@ struct radv_iface
 #else
 #define RADV_FORCE_DEBUG 0
 #endif
-#define RADV_TRACE(flags, msg, args...) do { if ((ra->p.debug & flags) || RADV_FORCE_DEBUG) \
-	log(L_TRACE "%s: " msg, ra->p.name , ## args ); } while(0)
+#define RADV_TRACE(flags, msg, args...) do { if ((p->p.debug & flags) || RADV_FORCE_DEBUG) \
+	log(L_TRACE "%s: " msg, p->p.name , ## args ); } while(0)
 
 
 /* radv.c */

From b3fae3a81796b6deea48445402793660fade5eb6 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 9 Aug 2017 12:41:44 +0200
Subject: [PATCH 007/198] RAdv: Fix typo

---
 proto/radv/radv.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/proto/radv/radv.c b/proto/radv/radv.c
index 4ea1f5c2..91f9853d 100644
--- a/proto/radv/radv.c
+++ b/proto/radv/radv.c
@@ -167,7 +167,7 @@ radv_iface_new(struct radv_proto *p, struct iface *iface, struct radv_iface_conf
   ifa->addr = find_lladdr(iface);
   if (!ifa->addr)
   {
-    log(L_ERR "%s: Cannot find link-locad addr on interface %s", p->p.name, iface->name);
+    log(L_ERR "%s: Missing link-local address on interface %s", p->p.name, iface->name);
     return;
   }
 

From da390bb11c8efc70767fc4b6d00bebf558d3b00f Mon Sep 17 00:00:00 2001
From: Michal 'vorner' Vaner <vorner@vorner.cz>
Date: Wed, 9 Aug 2017 13:35:55 +0200
Subject: [PATCH 008/198] Drop stale TODO

A TODO file last updated 5 years ago is useless.
---
 TODO | 45 ---------------------------------------------
 1 file changed, 45 deletions(-)
 delete mode 100644 TODO

diff --git a/TODO b/TODO
deleted file mode 100644
index 23cd1877..00000000
--- a/TODO
+++ /dev/null
@@ -1,45 +0,0 @@
-Core
-~~~~
-- socket open failure should not be fatal
-- &&,||: priorities
-- static: allow specifying a per-route filter program for setting route attributes?
-
-Globals
-~~~~~~~
-- right usage of DBG vs. debug
-- logging and tracing; use appropriate log levels
-- check incoming packets and log errors!!
-- check log calls for trailing newlines and log levels followed by comma
-- check if all protocols set proper packet priorities and TTL's.
-- try compiling with -Wunused
-- does everybody test return value of sk_open?
-- protocols: implement CLI hooks and per-procotol CLI commands
-- protocols: implement reconfigure hook
-- protocols: use locking
-- check use of system includes and sprintf()
-
-Various ideas
-~~~~~~~~~~~~~
-- client: Ctrl-R eats one more enter
-- bgp: timing of updates?
-- netlink: import Linux route attributes to our rta's, so that they can be filtered?
-- config: executable config files
-- filters: user defined attributes?
-- io: use poll if available
-- route recalculation timing and flap dampening [see RFC2439 for algorithms]
-- aggregate engine: standard route aggregation and summarization [RFC2519]
-- aggregate engine: injection of manually configured pseudo-static routes
-- generate default route if any working BGP connection exists (aggregate engine again?)
-- generate default route to IGP's (aggregate engine yet another time?)
-- look at RFC 2386 (QoS-based routing)
-- cli: show tables?
-
-OSPF
-~~~~
-	- check incoming packets using neighbor cache
-	- RFC2328 appendix E: Use a better algorithm
-	- automatic generation of external route tags (RFC1403)
-	- RFC2370 opaque LSA's
-	- Limit export rate of external LSAs (like Gated does)
-	- Bugfix in link state retransmission list (aging)
-	- Graceful OSPF restart - RFC3623

From 9b776458efdfae0c30b97f3670a1f4646221f072 Mon Sep 17 00:00:00 2001
From: Michal 'vorner' Vaner <vorner@vorner.cz>
Date: Wed, 9 Aug 2017 13:39:20 +0200
Subject: [PATCH 009/198] Gitignore: File created by autoreconf

---
 .gitignore | 1 +
 1 file changed, 1 insertion(+)

diff --git a/.gitignore b/.gitignore
index 0b95d3a1..0dcebfd1 100644
--- a/.gitignore
+++ b/.gitignore
@@ -11,3 +11,4 @@
 /config.status
 /configure
 /sysdep/autoconf.h.in
+/sysdep/autoconf.h.in~

From 15a4421f9cb2c077cc484e3cda94e8710a1d68f5 Mon Sep 17 00:00:00 2001
From: Michal 'vorner' Vaner <michal.vaner@nic.cz>
Date: Thu, 10 Aug 2017 12:32:30 +0200
Subject: [PATCH 010/198] Fix bird.conf example

Make it syntactically correct, so it is accepted.
---
 bird.conf | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/bird.conf b/bird.conf
index c2b47378..410f190c 100644
--- a/bird.conf
+++ b/bird.conf
@@ -52,6 +52,6 @@ protocol rip {
 	ipv4;
 }
 
-protocol ripng {
+protocol rip ng {
 	ipv6;
 }

From 5699a2036cfb32cb3a683814f83242eb21e40d02 Mon Sep 17 00:00:00 2001
From: Michal 'vorner' Vaner <michal.vaner@nic.cz>
Date: Thu, 10 Aug 2017 13:06:05 +0200
Subject: [PATCH 011/198] Less confusing log message

A non-primary address isn't necessarily secondary, that's an independent
flag.
---
 nest/iface.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/nest/iface.c b/nest/iface.c
index a23cdf4f..addb1363 100644
--- a/nest/iface.c
+++ b/nest/iface.c
@@ -142,7 +142,7 @@ ifa_send_notify(struct proto *p, unsigned c, struct ifa *a)
     {
       if (p->debug & D_IFACES)
 	log(L_TRACE "%s < %s address %I/%d on interface %s %s",
-	    p->name, (a->flags & IA_PRIMARY) ? "primary" : "secondary",
+	    p->name, (a->flags & IA_PRIMARY) ? "primary" : "non-primary",
 	    a->prefix, a->pxlen, a->iface->name,
 	    (c & IF_CHANGE_UP) ? "added" : "removed");
       p->ifa_notify(p, c, a);

From 080d9e4ce25f8f14e61ba0a81d5385bf8de3e48f Mon Sep 17 00:00:00 2001
From: Michal 'vorner' Vaner <michal.vaner@nic.cz>
Date: Thu, 10 Aug 2017 14:35:14 +0200
Subject: [PATCH 012/198] Nicer log output

non-primary is ugly, just omit it (and use primary in the other case).
---
 nest/iface.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/nest/iface.c b/nest/iface.c
index addb1363..ff362938 100644
--- a/nest/iface.c
+++ b/nest/iface.c
@@ -141,8 +141,8 @@ ifa_send_notify(struct proto *p, unsigned c, struct ifa *a)
   if (p->ifa_notify)
     {
       if (p->debug & D_IFACES)
-	log(L_TRACE "%s < %s address %I/%d on interface %s %s",
-	    p->name, (a->flags & IA_PRIMARY) ? "primary" : "non-primary",
+	log(L_TRACE "%s <%s address %I/%d on interface %s %s",
+	    p->name, (a->flags & IA_PRIMARY) ? " primary" : "",
 	    a->prefix, a->pxlen, a->iface->name,
 	    (c & IF_CHANGE_UP) ? "added" : "removed");
       p->ifa_notify(p, c, a);

From e7ed9ecba77162b4c09f34354378b0e752b9078e Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 22 Aug 2017 14:03:38 +0200
Subject: [PATCH 013/198] Client: Fix include

---
 client/client.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/client/client.c b/client/client.c
index 0d4bdf3e..97cf6639 100644
--- a/client/client.c
+++ b/client/client.c
@@ -25,9 +25,10 @@
 #include <fcntl.h>
 #include <unistd.h>
 #include <errno.h>
+#include <sys/select.h>
 #include <sys/socket.h>
-#include <sys/un.h>
 #include <sys/types.h>
+#include <sys/un.h>
 
 #include "nest/bird.h"
 #include "lib/resource.h"

From 3ac5d1ce4c10719dd48555521a50d4a9c5eadd15 Mon Sep 17 00:00:00 2001
From: Michal 'vorner' Vaner <vorner@vorner.cz>
Date: Wed, 9 Aug 2017 16:00:16 +0200
Subject: [PATCH 014/198] RAdv: Extract prefix option preparation

Put the prefix option preparation into a separate function. We're going
to reuse that bit of code.
---
 proto/radv/packets.c | 54 +++++++++++++++++++++++++++-----------------
 1 file changed, 33 insertions(+), 21 deletions(-)

diff --git a/proto/radv/packets.c b/proto/radv/packets.c
index 1c9837a4..bf051f62 100644
--- a/proto/radv/packets.c
+++ b/proto/radv/packets.c
@@ -123,7 +123,7 @@ radv_prepare_rdnss(struct radv_iface *ifa, list *rdnss_list, char **buf, char *b
     else
       op->lifetime = htonl(rcf->lifetime);
 
-    while(NODE_VALID(rcf) && 
+    while(NODE_VALID(rcf) &&
 	  (rcf->lifetime == rcf_base->lifetime) &&
 	  (rcf->lifetime_mult == rcf_base->lifetime_mult))
       {
@@ -136,7 +136,7 @@ radv_prepare_rdnss(struct radv_iface *ifa, list *rdnss_list, char **buf, char *b
 
 	rcf = NODE_NEXT(rcf);
       }
-  
+
     op->length = 1+2*i;
     *buf += 8 * op->length;
   }
@@ -235,6 +235,36 @@ radv_prepare_dnssl(struct radv_iface *ifa, list *dnssl_list, char **buf, char *b
   return -1;
 }
 
+static int
+radv_prepare_prefix(struct radv_iface *ifa, struct radv_prefix_config *pc,
+  struct ifa *addr, char **buf, char *bufend)
+{
+  struct radv_opt_prefix *op = (void *) *buf;
+
+  if (*buf + sizeof(*op) > bufend)
+  {
+    log(L_WARN "%s: Too many prefixes on interface %s", ifa->ra->p.name,
+      ifa->iface->name);
+    return -1;
+  }
+
+  op->type = OPT_PREFIX;
+  op->length = 4;
+  op->pxlen = addr->pxlen;
+  op->flags = (pc->onlink ? OPT_PX_ONLINK : 0) |
+    (pc->autonomous ? OPT_PX_AUTONOMOUS : 0);
+  op->valid_lifetime = (ifa->ra->active || !pc->valid_lifetime_sensitive) ?
+    htonl(pc->valid_lifetime) : 0;
+  op->preferred_lifetime = (ifa->ra->active || !pc->preferred_lifetime_sensitive) ?
+    htonl(pc->preferred_lifetime) : 0;
+  op->reserved = 0;
+  op->prefix = addr->prefix;
+  ipa_hton(op->prefix);
+  *buf += sizeof(*op);
+
+  return 0;
+}
+
 static void
 radv_prepare_ra(struct radv_iface *ifa)
 {
@@ -279,26 +309,8 @@ radv_prepare_ra(struct radv_iface *ifa)
     if (!pc || pc->skip)
       continue;
 
-    if (buf + sizeof(struct radv_opt_prefix) > bufend)
-    {
-      log(L_WARN "%s: Too many prefixes on interface %s", p->p.name, ifa->iface->name);
+    if (radv_prepare_prefix(ifa, pc, addr, &buf, bufend) < 0)
       goto done;
-    }
-
-    struct radv_opt_prefix *op = (void *) buf;
-    op->type = OPT_PREFIX;
-    op->length = 4;
-    op->pxlen = addr->pxlen;
-    op->flags = (pc->onlink ? OPT_PX_ONLINK : 0) |
-      (pc->autonomous ? OPT_PX_AUTONOMOUS : 0);
-    op->valid_lifetime = (p->active || !pc->valid_lifetime_sensitive) ?
-      htonl(pc->valid_lifetime) : 0;
-    op->preferred_lifetime = (p->active || !pc->preferred_lifetime_sensitive) ?
-      htonl(pc->preferred_lifetime) : 0;
-    op->reserved = 0;
-    op->prefix = addr->prefix;
-    ipa_hton(op->prefix);
-    buf += sizeof(*op);
   }
 
   if (! ic->rdnss_local)

From e2d2b3ef21f082a034e79d0880d4b36428029a59 Mon Sep 17 00:00:00 2001
From: Michal 'vorner' Vaner <michal.vaner@nic.cz>
Date: Fri, 11 Aug 2017 12:25:36 +0200
Subject: [PATCH 015/198] RAdv: Buffer prefixes awhile after they disappear

Keep a cache of all the relevant prefixes we send out. When a prefix
appears, insert it into the cache. If it dies, keep it there for a
while, marked as dead.

Send out the dead prefixes with zero lifetime.
---
 proto/radv/packets.c |  53 +++-------------
 proto/radv/radv.c    | 146 ++++++++++++++++++++++++++++++++++++++++---
 proto/radv/radv.h    |  23 ++++++-
 3 files changed, 170 insertions(+), 52 deletions(-)

diff --git a/proto/radv/packets.c b/proto/radv/packets.c
index bf051f62..088d177c 100644
--- a/proto/radv/packets.c
+++ b/proto/radv/packets.c
@@ -70,36 +70,6 @@ struct radv_opt_dnssl
   char domain[];
 };
 
-
-static struct radv_prefix_config default_prefix = {
-  .onlink = 1,
-  .autonomous = 1,
-  .valid_lifetime = DEFAULT_VALID_LIFETIME,
-  .preferred_lifetime = DEFAULT_PREFERRED_LIFETIME
-};
-
-
-static struct radv_prefix_config *
-radv_prefix_match(struct radv_iface *ifa, struct ifa *a)
-{
-  struct radv_proto *p = ifa->ra;
-  struct radv_config *cf = (struct radv_config *) (p->p.cf);
-  struct radv_prefix_config *pc;
-
-  if (a->scope <= SCOPE_LINK)
-    return NULL;
-
-  WALK_LIST(pc, ifa->cf->pref_list)
-    if ((a->pxlen >= pc->pxlen) && ipa_in_net(a->prefix, pc->prefix, pc->pxlen))
-      return pc;
-
-  WALK_LIST(pc, cf->pref_list)
-    if ((a->pxlen >= pc->pxlen) && ipa_in_net(a->prefix, pc->prefix, pc->pxlen))
-      return pc;
-
-  return &default_prefix;
-}
-
 static int
 radv_prepare_rdnss(struct radv_iface *ifa, list *rdnss_list, char **buf, char *bufend)
 {
@@ -236,9 +206,10 @@ radv_prepare_dnssl(struct radv_iface *ifa, list *dnssl_list, char **buf, char *b
 }
 
 static int
-radv_prepare_prefix(struct radv_iface *ifa, struct radv_prefix_config *pc,
-  struct ifa *addr, char **buf, char *bufend)
+radv_prepare_prefix(struct radv_iface *ifa, struct radv_prefix *prefix,
+		    char **buf, char *bufend)
 {
+  struct radv_prefix_config *pc = prefix->config;
   struct radv_opt_prefix *op = (void *) *buf;
 
   if (*buf + sizeof(*op) > bufend)
@@ -250,7 +221,7 @@ radv_prepare_prefix(struct radv_iface *ifa, struct radv_prefix_config *pc,
 
   op->type = OPT_PREFIX;
   op->length = 4;
-  op->pxlen = addr->pxlen;
+  op->pxlen = prefix->len;
   op->flags = (pc->onlink ? OPT_PX_ONLINK : 0) |
     (pc->autonomous ? OPT_PX_AUTONOMOUS : 0);
   op->valid_lifetime = (ifa->ra->active || !pc->valid_lifetime_sensitive) ?
@@ -258,7 +229,7 @@ radv_prepare_prefix(struct radv_iface *ifa, struct radv_prefix_config *pc,
   op->preferred_lifetime = (ifa->ra->active || !pc->preferred_lifetime_sensitive) ?
     htonl(pc->preferred_lifetime) : 0;
   op->reserved = 0;
-  op->prefix = addr->prefix;
+  op->prefix = prefix->prefix;
   ipa_hton(op->prefix);
   *buf += sizeof(*op);
 
@@ -300,16 +271,10 @@ radv_prepare_ra(struct radv_iface *ifa)
     buf += sizeof (*om);
   }
 
-  struct ifa *addr;
-  WALK_LIST(addr, ifa->iface->addrs)
+  struct radv_prefix *prefix;
+  WALK_LIST(prefix, ifa->prefixes)
   {
-    struct radv_prefix_config *pc;
-    pc = radv_prefix_match(ifa, addr);
-
-    if (!pc || pc->skip)
-      continue;
-
-    if (radv_prepare_prefix(ifa, pc, addr, &buf, bufend) < 0)
+    if (radv_prepare_prefix(ifa, prefix, &buf, bufend) < 0)
       goto done;
   }
 
@@ -419,7 +384,7 @@ radv_err_hook(sock *sk, int err)
 int
 radv_sk_open(struct radv_iface *ifa)
 {
-  sock *sk = sk_new(ifa->ra->p.pool);
+  sock *sk = sk_new(ifa->pool);
   sk->type = SK_IP;
   sk->dport = ICMPV6_PROTO;
   sk->saddr = ifa->addr->ip;
diff --git a/proto/radv/radv.c b/proto/radv/radv.c
index 91f9853d..303cf251 100644
--- a/proto/radv/radv.c
+++ b/proto/radv/radv.c
@@ -51,6 +51,15 @@ radv_timer(timer *tm)
 
   RADV_TRACE(D_EVENTS, "Timer fired on %s", ifa->iface->name);
 
+  /* If some dead prefixes expired, regenerate the prefix list and the packet.
+   * We do so by pretending there was a change on the interface.
+   *
+   * This sets the timer, but we replace it just at the end of this function
+   * (replacing a timer is fine).
+   */
+  if (ifa->prefix_expires != 0 && ifa->prefix_expires <= now)
+    radv_iface_notify(ifa, RA_EV_GC);
+
   radv_send_ra(ifa, 0);
 
   /* Update timer */
@@ -67,7 +76,129 @@ radv_timer(timer *tm)
   tm_start(ifa->timer, after);
 }
 
-static char* ev_name[] = { NULL, "Init", "Change", "RS" };
+static struct radv_prefix_config default_prefix = {
+  .onlink = 1,
+  .autonomous = 1,
+  .valid_lifetime = DEFAULT_VALID_LIFETIME,
+  .preferred_lifetime = DEFAULT_PREFERRED_LIFETIME
+};
+
+static struct radv_prefix_config dead_prefix = {
+};
+
+/* Find a corresponding config for the given prefix */
+static struct radv_prefix_config *
+radv_prefix_match(struct radv_iface *ifa, struct ifa *a)
+{
+  struct radv_proto *p = ifa->ra;
+  struct radv_config *cf = (struct radv_config *) (p->p.cf);
+  struct radv_prefix_config *pc;
+
+  if (a->scope <= SCOPE_LINK)
+    return NULL;
+
+  WALK_LIST(pc, ifa->cf->pref_list)
+    if ((a->pxlen >= pc->pxlen) && ipa_in_net(a->prefix, pc->prefix, pc->pxlen))
+      return pc;
+
+  WALK_LIST(pc, cf->pref_list)
+    if ((a->pxlen >= pc->pxlen) && ipa_in_net(a->prefix, pc->prefix, pc->pxlen))
+      return pc;
+
+  return &default_prefix;
+}
+
+/*
+ * Go through the list of prefixes, compare them with configs and decide if we
+ * want them or not. */
+static void
+prefixes_prepare(struct radv_iface *ifa)
+{
+  struct radv_proto *p = ifa->ra;
+  /* First mark all the prefixes as unused */
+  struct radv_prefix *pfx;
+
+  WALK_LIST(pfx, ifa->prefixes)
+    pfx->mark = 0;
+
+  /* Now find all the prefixes we want to use and make sure they are in the
+   * list. */
+  struct ifa *addr;
+  WALK_LIST(addr, ifa->iface->addrs)
+  {
+    struct radv_prefix_config *pc = radv_prefix_match(ifa, addr);
+
+    if (!pc || pc->skip)
+      continue;
+
+    /* Do we have it already? */
+    struct radv_prefix *existing = NULL;
+    WALK_LIST(pfx, ifa->prefixes)
+      if (pfx->len == addr->pxlen &&
+	  memcmp(&pfx->prefix, &addr->prefix, sizeof pfx->prefix) == 0)
+      {
+	existing = pfx;
+	break;
+      }
+
+    if (!existing)
+    {
+      RADV_TRACE(D_EVENTS, "Allocating new prefix %I on %s", addr->prefix,
+		 ifa->iface->name);
+      existing = mb_allocz(ifa->pool, sizeof *existing);
+      existing->prefix = addr->prefix;
+      existing->len = addr->pxlen;
+      add_tail(&ifa->prefixes, NODE existing);
+    }
+    /*
+     * Update the information (it may have changed, or even bring a prefix back
+     * to life).
+     */
+    existing->alive = 1;
+    existing->mark = 1;
+    existing->config = pc;
+  }
+
+  /*
+   * Garbage-collect the prefixes. If something isn't used, it dies (but isn't
+   * dropped just yet). If something is dead and rots there for long enough,
+   * clean it up.
+   */
+  // XXX: Make these 5 minutes it configurable
+  bird_clock_t rotten = now + 300;
+  struct radv_prefix *next;
+  bird_clock_t expires_soonest = 0;
+  WALK_LIST_DELSAFE(pfx, next, ifa->prefixes) {
+    if (pfx->alive && !pfx->mark)
+    {
+      RADV_TRACE(D_EVENTS, "Marking prefix %I on %s as dead", pfx->prefix,
+		 ifa->iface->name);
+      // It just died
+      pfx->alive = 0;
+      pfx->expires = rotten;
+      pfx->config = &dead_prefix;
+    }
+    if (!pfx->alive)
+      if (pfx->expires <= now)
+      {
+	RADV_TRACE(D_EVENTS, "Dropping long dead prefix %I on %s", pfx->prefix,
+		   ifa->iface->name);
+	// It's dead and rotten, clean it up
+	rem_node(NODE pfx);
+	mb_free(pfx);
+      }
+      else
+      {
+	ASSERT(pfx->expires != 0);
+	// Let it rot for a while more, but look when it's ripe.
+	if (expires_soonest == 0 || pfx->expires < expires_soonest)
+	  expires_soonest = pfx->expires;
+      }
+  }
+  ifa->prefix_expires = expires_soonest;
+}
+
+static char* ev_name[] = { NULL, "Init", "Change", "RS", "Garbage collect" };
 
 void
 radv_iface_notify(struct radv_iface *ifa, int event)
@@ -82,6 +213,7 @@ radv_iface_notify(struct radv_iface *ifa, int event)
   switch (event)
   {
   case RA_EV_CHANGE:
+  case RA_EV_GC:
     ifa->plen = 0;
   case RA_EV_INIT:
     ifa->initial = MAX_INITIAL_RTR_ADVERTISEMENTS;
@@ -91,6 +223,8 @@ radv_iface_notify(struct radv_iface *ifa, int event)
     break;
   }
 
+  prefixes_prepare(ifa);
+
   /* Update timer */
   unsigned delta = now - ifa->last;
   unsigned after = 0;
@@ -152,15 +286,17 @@ find_lladdr(struct iface *iface)
 static void
 radv_iface_new(struct radv_proto *p, struct iface *iface, struct radv_iface_config *cf)
 {
-  pool *pool = p->p.pool;
   struct radv_iface *ifa;
 
   RADV_TRACE(D_EVENTS, "Adding interface %s", iface->name);
 
+  pool *pool = rp_new(p->p.pool, iface->name);
   ifa = mb_allocz(pool, sizeof(struct radv_iface));
+  ifa->pool = pool;
   ifa->ra = p;
   ifa->cf = cf;
   ifa->iface = iface;
+  init_list(&ifa->prefixes);
 
   add_tail(&p->iface_list, NODE ifa);
 
@@ -198,11 +334,7 @@ radv_iface_remove(struct radv_iface *ifa)
 
   rem_node(NODE ifa);
 
-  rfree(ifa->sk);
-  rfree(ifa->timer);
-  rfree(ifa->lock);
-
-  mb_free(ifa);
+  rfree(ifa->pool);
 }
 
 static void
diff --git a/proto/radv/radv.h b/proto/radv/radv.h
index 559a3f3f..54d7c6f0 100644
--- a/proto/radv/radv.h
+++ b/proto/radv/radv.h
@@ -121,6 +121,23 @@ struct radv_proto
   u8 active;			/* Whether radv is active w.r.t. triggers */
 };
 
+struct radv_prefix		/* One prefix we advertise */
+{
+  node n;
+  ip_addr prefix;
+  u8 len;
+  /* Is the prefix alive? If not, we advertise it with 0 lifetime, so clients
+   * stop using it. */
+  u8 alive;
+  u8 mark;			/* A temporary mark for processing */
+  /* The (absolute) time when we drop this prefix from advertising. It is valid
+   * only if !alive. */
+  bird_clock_t expires;
+  /* The config tied to this prefix. Always valid (we place a dummy config here
+   * when !alive). */
+  struct radv_prefix_config *config;
+};
+
 struct radv_iface
 {
   node n;
@@ -128,6 +145,9 @@ struct radv_iface
   struct radv_iface_config *cf;	/* Related config, must be updated in reconfigure */
   struct iface *iface;
   struct ifa *addr;		/* Link-local address of iface */
+  struct pool *pool;		/* A pool for interface-specific things */
+  list prefixes;		/* The prefixes we advertise */
+  bird_clock_t prefix_expires;	/* When the soonest prefix expires (0 = none dead) */
 
   timer *timer;
   struct object_lock *lock;
@@ -135,12 +155,13 @@ struct radv_iface
 
   bird_clock_t last;		/* Time of last sending of RA */
   u16 plen;			/* Length of prepared RA in tbuf, or 0 if not valid */
-  byte initial;			/* List of active ifaces */
+  byte initial;			/* How many RAs are still to be sent as initial */
 };
 
 #define RA_EV_INIT 1		/* Switch to initial mode */
 #define RA_EV_CHANGE 2		/* Change of options or prefixes */
 #define RA_EV_RS 3		/* Received RS */
+#define RA_EV_GC 4		/* Internal garbage collection of prefixes */
 
 /* Default Router Preferences (RFC 4191) */
 #define RA_PREF_LOW	0x18

From ec7d6a506ecae89e5019ce7fcb380a713be28bce Mon Sep 17 00:00:00 2001
From: Michal 'vorner' Vaner <michal.vaner@nic.cz>
Date: Thu, 17 Aug 2017 11:34:25 +0200
Subject: [PATCH 016/198] RAdv: Configure how long a dead prefix is advertised

---
 doc/bird.sgml       | 6 ++++++
 proto/radv/config.Y | 4 +++-
 proto/radv/radv.c   | 4 ++--
 proto/radv/radv.h   | 6 +++++-
 4 files changed, 16 insertions(+), 4 deletions(-)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index ffd28964..d8dad431 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -3436,6 +3436,12 @@ dsc-iface
 	as a default router. For <cf/sensitive/ option, see <ref id="radv-trigger" name="trigger">.
 	Default: 3 * <cf/max ra	interval/, <cf/sensitive/ yes.
 
+	<tag><label id="radv-iface-linger-time">linger time <m/expr/</tag>
+	When a prefix disappears, it is advertised for some time with 0 lifetime,
+	to inform clients the prefix is no longer usable. This sets the time for
+	how long it is advertised (in seconds). Maximum is 3600, 0 means
+	disable. Default 300.
+
 	<tag><label id="radv-iface-default-preference-low">default preference low|medium|high</tag>
 	This option specifies the Default Router Preference value to advertise
 	to hosts. Default: medium.
diff --git a/proto/radv/config.Y b/proto/radv/config.Y
index da300667..0ff84aeb 100644
--- a/proto/radv/config.Y
+++ b/proto/radv/config.Y
@@ -27,7 +27,7 @@ static u8 radv_mult_val;	/* Used by radv_mult for second return value */
 CF_DECLS
 
 CF_KEYWORDS(RADV, PREFIX, INTERFACE, MIN, MAX, RA, DELAY, INTERVAL,
-	MANAGED, OTHER, CONFIG, LINK, MTU, REACHABLE, TIME, RETRANS,
+	MANAGED, OTHER, CONFIG, LINGER, LINK, MTU, REACHABLE, TIME, RETRANS,
 	TIMER, CURRENT, HOP, LIMIT, DEFAULT, VALID, PREFERRED, MULT,
 	LIFETIME, SKIP, ONLINK, AUTONOMOUS, RDNSS, DNSSL, NS, DOMAIN,
 	LOCAL, TRIGGER, SENSITIVE, PREFERENCE, LOW, MEDIUM, HIGH)
@@ -82,6 +82,7 @@ radv_iface_start:
   RADV_IFACE->max_ra_int = DEFAULT_MAX_RA_INT;
   RADV_IFACE->min_delay = DEFAULT_MIN_DELAY;
   RADV_IFACE->current_hop_limit = DEFAULT_CURRENT_HOP_LIMIT;
+  RADV_IFACE->linger_time = DEFAULT_LINGER_TIME;
   RADV_IFACE->default_lifetime = -1;
   RADV_IFACE->default_lifetime_sensitive = 1;
   RADV_IFACE->default_preference = RA_PREF_MEDIUM;
@@ -96,6 +97,7 @@ radv_iface_item:
  | LINK MTU expr { RADV_IFACE->link_mtu = $3; if ($3 < 0) cf_error("Link MTU must be 0 or positive"); }
  | REACHABLE TIME expr { RADV_IFACE->reachable_time = $3; if (($3 < 0) || ($3 > 3600000)) cf_error("Reachable time must be in range 0-3600000"); }
  | RETRANS TIMER expr { RADV_IFACE->retrans_timer = $3; if ($3 < 0) cf_error("Retrans timer must be 0 or positive"); }
+ | LINGER TIME expr { RADV_IFACE->linger_time = $3; if (($3 < 0) || ($3 > 3600)) cf_error("Linger time must be in range 0-3600"); }
  | CURRENT HOP LIMIT expr { RADV_IFACE->current_hop_limit = $4; if (($4 < 0) || ($4 > 255))  cf_error("Current hop limit must be in range 0-255"); }
  | DEFAULT LIFETIME expr radv_sensitive {
      RADV_IFACE->default_lifetime = $3;
diff --git a/proto/radv/radv.c b/proto/radv/radv.c
index 303cf251..44040a4b 100644
--- a/proto/radv/radv.c
+++ b/proto/radv/radv.c
@@ -115,6 +115,7 @@ static void
 prefixes_prepare(struct radv_iface *ifa)
 {
   struct radv_proto *p = ifa->ra;
+  struct radv_iface_config *cf = ifa->cf;
   /* First mark all the prefixes as unused */
   struct radv_prefix *pfx;
 
@@ -164,8 +165,7 @@ prefixes_prepare(struct radv_iface *ifa)
    * dropped just yet). If something is dead and rots there for long enough,
    * clean it up.
    */
-  // XXX: Make these 5 minutes it configurable
-  bird_clock_t rotten = now + 300;
+  bird_clock_t rotten = now + cf->linger_time;
   struct radv_prefix *next;
   bird_clock_t expires_soonest = 0;
   WALK_LIST_DELSAFE(pfx, next, ifa->prefixes) {
diff --git a/proto/radv/radv.h b/proto/radv/radv.h
index 54d7c6f0..777c6e16 100644
--- a/proto/radv/radv.h
+++ b/proto/radv/radv.h
@@ -35,6 +35,7 @@
 #define DEFAULT_MAX_RA_INT 600
 #define DEFAULT_MIN_DELAY 3
 #define DEFAULT_CURRENT_HOP_LIMIT 64
+#define DEFAULT_LINGER_TIME 300
 
 #define DEFAULT_VALID_LIFETIME 86400
 #define DEFAULT_PREFERRED_LIFETIME 14400
@@ -66,6 +67,9 @@ struct radv_iface_config
   u32 max_ra_int;
   u32 min_delay;
 
+  u32 linger_time;		/* How long a dead prefix should still be advertised with 0
+				   lifetime */
+
   u8 rdnss_local;		/* Global list is not used for RDNSS */
   u8 dnssl_local;		/* Global list is not used for DNSSL */
 
@@ -77,7 +81,7 @@ 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) */
+  u8 default_preference;	 /* Default Router Preference (RFC 4191) */
 };
 
 struct radv_prefix_config

From 08b6a617e862ffc7b7460abb79fe4c198806517c Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 29 Aug 2017 19:17:35 +0200
Subject: [PATCH 017/198] RAdv: Some style nitpicks

---
 doc/bird.sgml        |  8 +++---
 proto/radv/packets.c | 10 +++----
 proto/radv/radv.c    | 64 ++++++++++++++++++++++++--------------------
 proto/radv/radv.h    | 16 +++++------
 4 files changed, 50 insertions(+), 48 deletions(-)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index d8dad431..516f4c6e 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -3437,10 +3437,10 @@ dsc-iface
 	Default: 3 * <cf/max ra	interval/, <cf/sensitive/ yes.
 
 	<tag><label id="radv-iface-linger-time">linger time <m/expr/</tag>
-	When a prefix disappears, it is advertised for some time with 0 lifetime,
-	to inform clients the prefix is no longer usable. This sets the time for
-	how long it is advertised (in seconds). Maximum is 3600, 0 means
-	disable. Default 300.
+	When a prefix disappears, it is advertised for some time with 0
+	lifetime, to inform clients the prefix is no longer usable. This option
+	sets the time for how long it is advertised (in seconds). Maximum is
+	3600, 0 means disabled. Default: 300.
 
 	<tag><label id="radv-iface-default-preference-low">default preference low|medium|high</tag>
 	This option specifies the Default Router Preference value to advertise
diff --git a/proto/radv/packets.c b/proto/radv/packets.c
index 088d177c..19d71f97 100644
--- a/proto/radv/packets.c
+++ b/proto/radv/packets.c
@@ -209,16 +209,16 @@ static int
 radv_prepare_prefix(struct radv_iface *ifa, struct radv_prefix *prefix,
 		    char **buf, char *bufend)
 {
-  struct radv_prefix_config *pc = prefix->config;
-  struct radv_opt_prefix *op = (void *) *buf;
+  struct radv_prefix_config *pc = prefix->cf;
 
-  if (*buf + sizeof(*op) > bufend)
+  if (*buf + sizeof(struct radv_opt_prefix) > bufend)
   {
-    log(L_WARN "%s: Too many prefixes on interface %s", ifa->ra->p.name,
-      ifa->iface->name);
+    log(L_WARN "%s: Too many prefixes on interface %s",
+	ifa->ra->p.name, ifa->iface->name);
     return -1;
   }
 
+  struct radv_opt_prefix *op = (void *) *buf;
   op->type = OPT_PREFIX;
   op->length = 4;
   op->pxlen = prefix->len;
diff --git a/proto/radv/radv.c b/proto/radv/radv.c
index 44040a4b..227c8ef6 100644
--- a/proto/radv/radv.c
+++ b/proto/radv/radv.c
@@ -51,13 +51,14 @@ radv_timer(timer *tm)
 
   RADV_TRACE(D_EVENTS, "Timer fired on %s", ifa->iface->name);
 
-  /* If some dead prefixes expired, regenerate the prefix list and the packet.
+  /*
+   * If some dead prefixes expired, regenerate the prefix list and the packet.
    * We do so by pretending there was a change on the interface.
    *
    * This sets the timer, but we replace it just at the end of this function
    * (replacing a timer is fine).
    */
-  if (ifa->prefix_expires != 0 && ifa->prefix_expires <= now)
+  if (ifa->prefix_expires && (ifa->prefix_expires <= now))
     radv_iface_notify(ifa, RA_EV_GC);
 
   radv_send_ra(ifa, 0);
@@ -110,20 +111,20 @@ radv_prefix_match(struct radv_iface *ifa, struct ifa *a)
 
 /*
  * Go through the list of prefixes, compare them with configs and decide if we
- * want them or not. */
+ * want them or not.
+ */
 static void
-prefixes_prepare(struct radv_iface *ifa)
+radv_prepare_prefixes(struct radv_iface *ifa)
 {
   struct radv_proto *p = ifa->ra;
   struct radv_iface_config *cf = ifa->cf;
-  /* First mark all the prefixes as unused */
   struct radv_prefix *pfx;
 
+  /* First mark all the prefixes as unused */
   WALK_LIST(pfx, ifa->prefixes)
     pfx->mark = 0;
 
-  /* Now find all the prefixes we want to use and make sure they are in the
-   * list. */
+  /* Find all the prefixes we want to use and make sure they are in the list. */
   struct ifa *addr;
   WALK_LIST(addr, ifa->iface->addrs)
   {
@@ -135,8 +136,7 @@ prefixes_prepare(struct radv_iface *ifa)
     /* Do we have it already? */
     struct radv_prefix *existing = NULL;
     WALK_LIST(pfx, ifa->prefixes)
-      if (pfx->len == addr->pxlen &&
-	  memcmp(&pfx->prefix, &addr->prefix, sizeof pfx->prefix) == 0)
+      if ((pfx->len == addr->pxlen) && ipa_equal(pfx->prefix, addr->prefix))
       {
 	existing = pfx;
 	break;
@@ -144,20 +144,22 @@ prefixes_prepare(struct radv_iface *ifa)
 
     if (!existing)
     {
-      RADV_TRACE(D_EVENTS, "Allocating new prefix %I on %s", addr->prefix,
-		 ifa->iface->name);
+      RADV_TRACE(D_EVENTS, "Adding new prefix %I/%d on %s",
+		 addr->prefix, addr->pxlen, ifa->iface->name);
+
       existing = mb_allocz(ifa->pool, sizeof *existing);
       existing->prefix = addr->prefix;
       existing->len = addr->pxlen;
       add_tail(&ifa->prefixes, NODE existing);
     }
+
     /*
      * Update the information (it may have changed, or even bring a prefix back
      * to life).
      */
     existing->alive = 1;
     existing->mark = 1;
-    existing->config = pc;
+    existing->cf = pc;
   }
 
   /*
@@ -165,37 +167,41 @@ prefixes_prepare(struct radv_iface *ifa)
    * dropped just yet). If something is dead and rots there for long enough,
    * clean it up.
    */
-  bird_clock_t rotten = now + cf->linger_time;
+  bird_clock_t expires = now + cf->linger_time;
+  bird_clock_t expires_min = 0;
   struct radv_prefix *next;
-  bird_clock_t expires_soonest = 0;
-  WALK_LIST_DELSAFE(pfx, next, ifa->prefixes) {
+  WALK_LIST_DELSAFE(pfx, next, ifa->prefixes)
+  {
     if (pfx->alive && !pfx->mark)
     {
-      RADV_TRACE(D_EVENTS, "Marking prefix %I on %s as dead", pfx->prefix,
-		 ifa->iface->name);
-      // It just died
+      RADV_TRACE(D_EVENTS, "Marking prefix %I/$d on %s as dead",
+		 pfx->prefix, pfx->len, ifa->iface->name);
+
       pfx->alive = 0;
-      pfx->expires = rotten;
-      pfx->config = &dead_prefix;
+      pfx->expires = expires;
+      pfx->cf = &dead_prefix;
     }
+
     if (!pfx->alive)
+    {
       if (pfx->expires <= now)
       {
-	RADV_TRACE(D_EVENTS, "Dropping long dead prefix %I on %s", pfx->prefix,
-		   ifa->iface->name);
-	// It's dead and rotten, clean it up
+	RADV_TRACE(D_EVENTS, "Removing prefix %I/%d on %s",
+		   pfx->prefix, pfx->len, ifa->iface->name);
+
 	rem_node(NODE pfx);
 	mb_free(pfx);
       }
       else
       {
-	ASSERT(pfx->expires != 0);
-	// Let it rot for a while more, but look when it's ripe.
-	if (expires_soonest == 0 || pfx->expires < expires_soonest)
-	  expires_soonest = pfx->expires;
+	/* Find minimum expiration time */
+	if (!expires_min || (pfx->expires < expires_min))
+	  expires_min = pfx->expires;
       }
+    }
   }
-  ifa->prefix_expires = expires_soonest;
+
+  ifa->prefix_expires = expires_min;
 }
 
 static char* ev_name[] = { NULL, "Init", "Change", "RS", "Garbage collect" };
@@ -223,7 +229,7 @@ radv_iface_notify(struct radv_iface *ifa, int event)
     break;
   }
 
-  prefixes_prepare(ifa);
+  radv_prepare_prefixes(ifa);
 
   /* Update timer */
   unsigned delta = now - ifa->last;
diff --git a/proto/radv/radv.h b/proto/radv/radv.h
index 777c6e16..60b9980f 100644
--- a/proto/radv/radv.h
+++ b/proto/radv/radv.h
@@ -130,16 +130,12 @@ struct radv_prefix		/* One prefix we advertise */
   node n;
   ip_addr prefix;
   u8 len;
-  /* Is the prefix alive? If not, we advertise it with 0 lifetime, so clients
-   * stop using it. */
-  u8 alive;
+  u8 alive;			/* Is the prefix alive? If not, we advertise it
+				   with 0 lifetime, so clients stop using it */
   u8 mark;			/* A temporary mark for processing */
-  /* The (absolute) time when we drop this prefix from advertising. It is valid
-   * only if !alive. */
-  bird_clock_t expires;
-  /* The config tied to this prefix. Always valid (we place a dummy config here
-   * when !alive). */
-  struct radv_prefix_config *config;
+  bird_clock_t expires;		/* The time when we drop this prefix from
+				   advertising. It is valid only if !alive. */
+  struct radv_prefix_config *cf; /* The config tied to this prefix */
 };
 
 struct radv_iface
@@ -150,7 +146,7 @@ struct radv_iface
   struct iface *iface;
   struct ifa *addr;		/* Link-local address of iface */
   struct pool *pool;		/* A pool for interface-specific things */
-  list prefixes;		/* The prefixes we advertise */
+  list prefixes;		/* The prefixes we advertise (struct radv_prefix) */
   bird_clock_t prefix_expires;	/* When the soonest prefix expires (0 = none dead) */
 
   timer *timer;

From 5c4dfe0c30e5516fac73e6ed8455af2245ebc62a Mon Sep 17 00:00:00 2001
From: Michal 'vorner' Vaner <michal.vaner@nic.cz>
Date: Thu, 17 Aug 2017 15:02:35 +0200
Subject: [PATCH 018/198] Gitlab CI support

Add configuration and docker definitions for tests and builds in Gitlab
CI platform.

Some of them currently fail, which is a known problem.
---
 .gitlab-ci.yml                              | 451 ++++++++++++++++++++
 misc/docker/centos-6-amd64/Dockerfile       |  11 +
 misc/docker/centos-7-amd64/Dockerfile       |  11 +
 misc/docker/debian-7-amd64/Dockerfile       |  12 +
 misc/docker/debian-7-i386/Dockerfile        |  12 +
 misc/docker/debian-8-amd64/Dockerfile       |  12 +
 misc/docker/debian-8-i386/Dockerfile        |  12 +
 misc/docker/debian-9-amd64/Dockerfile       |  12 +
 misc/docker/debian-9-i386/Dockerfile        |  12 +
 misc/docker/debian-testing-amd64/Dockerfile |  12 +
 misc/docker/debian-testing-i386/Dockerfile  |  12 +
 misc/docker/fedora-25-amd64/Dockerfile      |  10 +
 misc/docker/fedora-26-amd64/Dockerfile      |  10 +
 misc/docker/opensuse-42.3-amd64/Dockerfile  |  11 +
 misc/docker/ubuntu-14.04-amd64/Dockerfile   |  12 +
 misc/docker/ubuntu-16.04-amd64/Dockerfile   |  12 +
 16 files changed, 624 insertions(+)
 create mode 100644 .gitlab-ci.yml
 create mode 100644 misc/docker/centos-6-amd64/Dockerfile
 create mode 100644 misc/docker/centos-7-amd64/Dockerfile
 create mode 100644 misc/docker/debian-7-amd64/Dockerfile
 create mode 100644 misc/docker/debian-7-i386/Dockerfile
 create mode 100644 misc/docker/debian-8-amd64/Dockerfile
 create mode 100644 misc/docker/debian-8-i386/Dockerfile
 create mode 100644 misc/docker/debian-9-amd64/Dockerfile
 create mode 100644 misc/docker/debian-9-i386/Dockerfile
 create mode 100644 misc/docker/debian-testing-amd64/Dockerfile
 create mode 100644 misc/docker/debian-testing-i386/Dockerfile
 create mode 100644 misc/docker/fedora-25-amd64/Dockerfile
 create mode 100644 misc/docker/fedora-26-amd64/Dockerfile
 create mode 100644 misc/docker/opensuse-42.3-amd64/Dockerfile
 create mode 100644 misc/docker/ubuntu-14.04-amd64/Dockerfile
 create mode 100644 misc/docker/ubuntu-16.04-amd64/Dockerfile

diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
new file mode 100644
index 00000000..c9863a1c
--- /dev/null
+++ b/.gitlab-ci.yml
@@ -0,0 +1,451 @@
+variables:
+  DEBIAN_FRONTEND: noninteractive
+  LC_ALL: C
+  GIT_STRATEGY: fetch
+  DOCKER_CMD: docker --config="$HOME/.docker/$CI_JOB_ID/"
+  IMG_BASE: registry.labs.nic.cz/labs/bird
+
+stages:
+  - image
+  - build
+
+.docker: &docker_build
+  stage: image
+  allow_failure: true
+  script:
+  - $DOCKER_CMD login -u gitlab-ci-token -p $CI_BUILD_TOKEN registry.labs.nic.cz
+  # Make sure we refresh the base image if it updates (eg. security updates, etc)
+  # If we do just the build, cache is always reused and the freshness of the
+  # base image is never checked. However, pull always asks and updates the
+  # image only if it changed ‒ therefore, the cache is used unless there's a
+  # change.
+  - $DOCKER_CMD pull `sed -ne 's/^FROM //p' "misc/docker/$IMG_NAME/Dockerfile"`
+  - $DOCKER_CMD build -t "bird:$IMG_NAME" "misc/docker/$IMG_NAME"
+  - $DOCKER_CMD tag "bird:$IMG_NAME" "$IMG_BASE:$IMG_NAME"
+  - $DOCKER_CMD push "$IMG_BASE:$IMG_NAME"
+  after_script:
+  - rm -f "$HOME/.docker/$CI_JOB_ID/" # cleanup the credentials
+  tags:
+  # That's Docker in Docker
+  - dind
+
+docker_debian-7-amd64:
+  variables:
+    IMG_NAME: "debian-7-amd64"
+  <<: *docker_build
+
+docker_debian-8-amd64:
+  variables:
+    IMG_NAME: "debian-8-amd64"
+  <<: *docker_build
+
+docker_debian-9-amd64:
+  variables:
+    IMG_NAME: "debian-9-amd64"
+  <<: *docker_build
+
+docker_debian-testing-amd64:
+  variables:
+    IMG_NAME: "debian-testing-amd64"
+  <<: *docker_build
+
+docker_debian-7-i386:
+  variables:
+    IMG_NAME: "debian-7-i386"
+  <<: *docker_build
+
+docker_debian-8-i386:
+  variables:
+    IMG_NAME: "debian-8-i386"
+  <<: *docker_build
+
+docker_debian-9-i386:
+  variables:
+    IMG_NAME: "debian-9-i386"
+  <<: *docker_build
+
+docker_debian-testing-i386:
+  variables:
+    IMG_NAME: "debian-testing-i386"
+  <<: *docker_build
+
+docker_fedora-25-amd64:
+  variables:
+    IMG_NAME: "fedora-25-amd64"
+  <<: *docker_build
+
+docker_fedora-26-amd64:
+  variables:
+    IMG_NAME: "fedora-26-amd64"
+  <<: *docker_build
+
+docker_centos-6-amd64:
+  variables:
+    IMG_NAME: "centos-6-amd64"
+  <<: *docker_build
+
+docker_centos-7-amd64:
+  variables:
+    IMG_NAME: "centos-7-amd64"
+  <<: *docker_build
+
+docker_opensuse-42_3-amd64:
+  variables:
+    IMG_NAME: "opensuse-42.3-amd64"
+  <<: *docker_build
+
+docker_ubuntu-14_04-amd64:
+  variables:
+    IMG_NAME: "ubuntu-14.04-amd64"
+  <<: *docker_build
+
+docker_ubuntu-16_04-amd64:
+  variables:
+    IMG_NAME: "ubuntu-16.04-amd64"
+  <<: *docker_build
+
+.debian-7-i386: &debian-7-i386_env
+  image: registry.labs.nic.cz/labs/bird:debian-7-i386
+  tags:
+  - docker
+  - linux
+  - amd64
+
+.debian-8-i386: &debian-8-i386_env
+  image: registry.labs.nic.cz/labs/bird:debian-8-i386
+  tags:
+  - docker
+  - linux
+  - amd64
+
+.debian-9-i386: &debian-9-i386_env
+  image: registry.labs.nic.cz/labs/bird:debian-9-i386
+  tags:
+  - docker
+  - linux
+  - amd64
+
+.debian-testing-i386: &debian-testing-i386_env
+  image: registry.labs.nic.cz/labs/bird:debian-testing-i386
+  tags:
+  - docker
+  - linux
+  - amd64
+
+.debian-7-amd64: &debian-7-amd64_env
+  image: registry.labs.nic.cz/labs/bird:debian-7-amd64
+  tags:
+  - docker
+  - linux
+  - amd64
+
+.debian-8-amd64: &debian-8-amd64_env
+  image: registry.labs.nic.cz/labs/bird:debian-8-amd64
+  tags:
+  - docker
+  - linux
+  - amd64
+
+.debian-9-amd64: &debian-9-amd64_env
+  image: registry.labs.nic.cz/labs/bird:debian-9-amd64
+  tags:
+  - docker
+  - linux
+  - amd64
+
+.debian-testing-amd64: &debian-testing-amd64_env
+  image: registry.labs.nic.cz/labs/bird:debian-testing-amd64
+  tags:
+  - docker
+  - linux
+  - amd64
+
+.fedora-25-amd64: &fedora-25-amd64_env
+  image: registry.labs.nic.cz/labs/bird:fedora-25-amd64
+  tags:
+  - docker
+  - linux
+  - amd64
+
+.fedora-26-amd64: &fedora-26-amd64_env
+  image: registry.labs.nic.cz/labs/bird:fedora-26-amd64
+  tags:
+  - docker
+  - linux
+  - amd64
+
+.centos-6-amd64: &centos-6-amd64_env
+  image: registry.labs.nic.cz/labs/bird:centos-6-amd64
+  tags:
+  - docker
+  - linux
+  - amd64
+
+.centos-7-amd64: &centos-7-amd64_env
+  image: registry.labs.nic.cz/labs/bird:centos-7-amd64
+  tags:
+  - docker
+  - linux
+  - amd64
+
+.opensuse-42_3-amd64: &opensuse-42_3-amd64_env
+  image: registry.labs.nic.cz/labs/bird:opensuse-42.3-amd64
+  tags:
+  - docker
+  - linux
+  - amd64
+
+.ubuntu-14_04-amd64: &ubuntu-14_04-amd64_env
+  image: registry.labs.nic.cz/labs/bird:ubuntu-14.04-amd64
+  tags:
+  - docker
+  - linux
+  - amd64
+
+.ubuntu-16_04-amd64: &ubuntu-16_04-amd64_env
+  image: registry.labs.nic.cz/labs/bird:ubuntu-16.04-amd64
+  tags:
+  - docker
+  - linux
+  - amd64
+
+# TODO We want to copy these BSDs to our own virtual machines, to make sure someone doesn't update them by accident.
+.freebsd-11-i386: &freebsd-11-i386_env
+  variables:
+    CPPFLAGS: "-I/usr/local/include"
+    LDFLAGS: "-L/usr/local/lib"
+  tags:
+  - freebsd
+  - i386
+  #only:
+  #- master
+  #- triggers
+  #- tags
+
+.freebsd-11-amd64: &freebsd-11-amd64_env
+  variables:
+    CPPFLAGS: "-I/usr/local/include"
+    LDFLAGS: "-L/usr/local/lib"
+  tags:
+  - freebsd
+  - amd64
+  #only:
+  #- master
+  #- triggers
+  #- tags
+
+.build: &build_job
+  stage: build
+  script:
+  - autoreconf
+  - ./configure --enable-ipv6=$IPV6 CPPFLAGS="$CPPFLAGS" LDFLAGS="$LDFLAGS"
+  # Detect which make is available
+  - MAKE=make
+  - which gmake 2>/dev/null >/dev/null && MAKE=gmake
+  - $MAKE
+  # Run tests if they are available (eg. don't fail if "check" isn't a valid make target)
+  - $MAKE check || [ "$?" = 2 ]
+
+build-debian-7-amd64:
+  variables:
+    IPV6: "no"
+  <<: *debian-7-amd64_env
+  <<: *build_job
+
+build-debian-8-amd64:
+  variables:
+    IPV6: "no"
+  <<: *debian-8-amd64_env
+  <<: *build_job
+
+build-debian-9-amd64:
+  variables:
+    IPV6: "no"
+  <<: *debian-9-amd64_env
+  <<: *build_job
+
+build-debian-testing-amd64:
+  variables:
+    IPV6: "no"
+  <<: *debian-testing-amd64_env
+  <<: *build_job
+
+build-debian-7-amd64-v6:
+  variables:
+    IPV6: "yes"
+  <<: *debian-7-amd64_env
+  <<: *build_job
+
+build-debian-8-amd64-v6:
+  variables:
+    IPV6: "yes"
+  <<: *debian-8-amd64_env
+  <<: *build_job
+
+build-debian-9-amd64-v6:
+  variables:
+    IPV6: "yes"
+  <<: *debian-9-amd64_env
+  <<: *build_job
+
+build-debian-testing-amd64-v6:
+  variables:
+    IPV6: "yes"
+  <<: *debian-testing-amd64_env
+  <<: *build_job
+
+build-fedora-25-amd64:
+  variables:
+    IPV6: "no"
+  <<: *fedora-25-amd64_env
+  <<: *build_job
+
+build-fedora-25-amd64-v6:
+  variables:
+    IPV6: "yes"
+  <<: *fedora-25-amd64_env
+  <<: *build_job
+
+build-fedora-26-amd64:
+  variables:
+    IPV6: "no"
+  <<: *fedora-26-amd64_env
+  <<: *build_job
+
+build-fedora-26-amd64-v6:
+  variables:
+    IPV6: "yes"
+  <<: *fedora-26-amd64_env
+  <<: *build_job
+
+build-centos-6-amd64:
+  variables:
+    IPV6: "no"
+  <<: *centos-6-amd64_env
+  <<: *build_job
+
+build-centos-6-amd64-v6:
+  variables:
+    IPV6: "yes"
+  <<: *centos-6-amd64_env
+  <<: *build_job
+
+build-centos-7-amd64:
+  variables:
+    IPV6: "no"
+  <<: *centos-7-amd64_env
+  <<: *build_job
+
+build-centos-7-amd64-v6:
+  variables:
+    IPV6: "yes"
+  <<: *centos-7-amd64_env
+  <<: *build_job
+
+build-opensuse-42_3-amd64:
+  variables:
+    IPV6: "no"
+  <<: *opensuse-42_3-amd64_env
+  <<: *build_job
+
+build-opensuse-42_3-amd64-v6:
+  variables:
+    IPV6: "yes"
+  <<: *opensuse-42_3-amd64_env
+  <<: *build_job
+
+build-ubuntu-14_04-amd64:
+  variables:
+    IPV6: "no"
+  <<: *ubuntu-14_04-amd64_env
+  <<: *build_job
+
+build-ubuntu-14_04-amd64-v6:
+  variables:
+    IPV6: "yes"
+  <<: *ubuntu-14_04-amd64_env
+  <<: *build_job
+
+build-ubuntu-16_04-amd64:
+  variables:
+    IPV6: "no"
+  <<: *ubuntu-16_04-amd64_env
+  <<: *build_job
+
+build-ubuntu-16_04-amd64-v6:
+  variables:
+    IPV6: "yes"
+  <<: *ubuntu-16_04-amd64_env
+  <<: *build_job
+
+build-debian-7-i386:
+  variables:
+    IPV6: "no"
+  <<: *debian-7-i386_env
+  <<: *build_job
+
+build-debian-7-i386-v6:
+  variables:
+    IPV6: "yes"
+  <<: *debian-7-i386_env
+  <<: *build_job
+
+build-debian-8-i386:
+  variables:
+    IPV6: "no"
+  <<: *debian-8-i386_env
+  <<: *build_job
+
+build-debian-8-i386-v6:
+  variables:
+    IPV6: "yes"
+  <<: *debian-8-i386_env
+  <<: *build_job
+
+build-debian-9-i386:
+  variables:
+    IPV6: "no"
+  <<: *debian-9-i386_env
+  <<: *build_job
+
+build-debian-9-i386-v6:
+  variables:
+    IPV6: "yes"
+  <<: *debian-9-i386_env
+  <<: *build_job
+
+build-debian-testing-i386:
+  variables:
+    IPV6: "no"
+  <<: *debian-testing-i386_env
+  <<: *build_job
+
+build-debian-testing-i386-v6:
+  variables:
+    IPV6: "yes"
+  <<: *debian-testing-i386_env
+  <<: *build_job
+
+build-freebsd-11-amd64:
+  variables:
+    IPV6: "no"
+  <<: *freebsd-11-amd64_env
+  <<: *build_job
+
+build-freebsd-11-amd64-v6:
+  variables:
+    IPV6: "yes"
+  <<: *freebsd-11-amd64_env
+  <<: *build_job
+
+build-freebsd-11-i386:
+  variables:
+    IPV6: "no"
+  <<: *freebsd-11-i386_env
+  <<: *build_job
+
+build-freebsd-i386-v6:
+  variables:
+    IPV6: "yes"
+  <<: *freebsd-11-i386_env
+  <<: *build_job
diff --git a/misc/docker/centos-6-amd64/Dockerfile b/misc/docker/centos-6-amd64/Dockerfile
new file mode 100644
index 00000000..d92f64f9
--- /dev/null
+++ b/misc/docker/centos-6-amd64/Dockerfile
@@ -0,0 +1,11 @@
+FROM centos:6
+RUN yum -y upgrade
+RUN yum -y install \
+	autoconf \
+	flex \
+	bison \
+	pkgconfig \
+	'readline-devel' \
+	'pkgconfig(ncurses)' \
+	gcc \
+	make
diff --git a/misc/docker/centos-7-amd64/Dockerfile b/misc/docker/centos-7-amd64/Dockerfile
new file mode 100644
index 00000000..c01731b2
--- /dev/null
+++ b/misc/docker/centos-7-amd64/Dockerfile
@@ -0,0 +1,11 @@
+FROM centos:7
+RUN yum -y upgrade
+RUN yum -y install \
+	autoconf \
+	flex \
+	bison \
+	pkgconfig \
+	'readline-devel' \
+	'pkgconfig(ncurses)' \
+	gcc \
+	make
diff --git a/misc/docker/debian-7-amd64/Dockerfile b/misc/docker/debian-7-amd64/Dockerfile
new file mode 100644
index 00000000..dd5e9236
--- /dev/null
+++ b/misc/docker/debian-7-amd64/Dockerfile
@@ -0,0 +1,12 @@
+FROM debian:wheezy-slim
+ENV DEBIAN_FRONTEND noninteractive
+RUN sed -i 's/deb.debian.org/ftp.cz.debian.org/' /etc/apt/sources.list
+RUN apt-get -y update
+RUN apt-get -y upgrade
+RUN apt-get -y install \
+	autoconf \
+	build-essential \
+	flex \
+	bison \
+	ncurses-dev \
+	libreadline-dev
diff --git a/misc/docker/debian-7-i386/Dockerfile b/misc/docker/debian-7-i386/Dockerfile
new file mode 100644
index 00000000..753ff1cb
--- /dev/null
+++ b/misc/docker/debian-7-i386/Dockerfile
@@ -0,0 +1,12 @@
+FROM i386/debian:wheezy-slim
+ENV DEBIAN_FRONTEND noninteractive
+RUN sed -i 's/deb.debian.org/ftp.cz.debian.org/' /etc/apt/sources.list
+RUN apt-get -y update
+RUN apt-get -y upgrade
+RUN apt-get -y install \
+	autoconf \
+	build-essential \
+	flex \
+	bison \
+	ncurses-dev \
+	libreadline-dev
diff --git a/misc/docker/debian-8-amd64/Dockerfile b/misc/docker/debian-8-amd64/Dockerfile
new file mode 100644
index 00000000..d0e8ed0a
--- /dev/null
+++ b/misc/docker/debian-8-amd64/Dockerfile
@@ -0,0 +1,12 @@
+FROM debian:jessie-slim
+ENV DEBIAN_FRONTEND noninteractive
+RUN sed -i 's/deb.debian.org/ftp.cz.debian.org/' /etc/apt/sources.list
+RUN apt-get -y update
+RUN apt-get -y upgrade
+RUN apt-get -y install \
+	autoconf \
+	build-essential \
+	flex \
+	bison \
+	ncurses-dev \
+	libreadline-dev
diff --git a/misc/docker/debian-8-i386/Dockerfile b/misc/docker/debian-8-i386/Dockerfile
new file mode 100644
index 00000000..9dd6f022
--- /dev/null
+++ b/misc/docker/debian-8-i386/Dockerfile
@@ -0,0 +1,12 @@
+FROM i386/debian:jessie-slim
+ENV DEBIAN_FRONTEND noninteractive
+RUN sed -i 's/deb.debian.org/ftp.cz.debian.org/' /etc/apt/sources.list
+RUN apt-get -y update
+RUN apt-get -y upgrade
+RUN apt-get -y install \
+	autoconf \
+	build-essential \
+	flex \
+	bison \
+	ncurses-dev \
+	libreadline-dev
diff --git a/misc/docker/debian-9-amd64/Dockerfile b/misc/docker/debian-9-amd64/Dockerfile
new file mode 100644
index 00000000..e8a17efc
--- /dev/null
+++ b/misc/docker/debian-9-amd64/Dockerfile
@@ -0,0 +1,12 @@
+FROM debian:stretch-slim
+ENV DEBIAN_FRONTEND noninteractive
+RUN sed -i 's/deb.debian.org/ftp.cz.debian.org/' /etc/apt/sources.list
+RUN apt-get -y update
+RUN apt-get -y upgrade
+RUN apt-get -y install \
+	autoconf \
+	build-essential \
+	flex \
+	bison \
+	ncurses-dev \
+	libreadline-dev
diff --git a/misc/docker/debian-9-i386/Dockerfile b/misc/docker/debian-9-i386/Dockerfile
new file mode 100644
index 00000000..73e70750
--- /dev/null
+++ b/misc/docker/debian-9-i386/Dockerfile
@@ -0,0 +1,12 @@
+FROM i386/debian:stretch-slim
+ENV DEBIAN_FRONTEND noninteractive
+RUN sed -i 's/deb.debian.org/ftp.cz.debian.org/' /etc/apt/sources.list
+RUN apt-get -y update
+RUN apt-get -y upgrade
+RUN apt-get -y install \
+	autoconf \
+	build-essential \
+	flex \
+	bison \
+	ncurses-dev \
+	libreadline-dev
diff --git a/misc/docker/debian-testing-amd64/Dockerfile b/misc/docker/debian-testing-amd64/Dockerfile
new file mode 100644
index 00000000..3b66e3fa
--- /dev/null
+++ b/misc/docker/debian-testing-amd64/Dockerfile
@@ -0,0 +1,12 @@
+FROM debian:testing-slim
+ENV DEBIAN_FRONTEND noninteractive
+RUN sed -i 's/deb.debian.org/ftp.cz.debian.org/' /etc/apt/sources.list
+RUN apt-get -y update
+RUN apt-get -y upgrade
+RUN apt-get -y install \
+	autoconf \
+	build-essential \
+	flex \
+	bison \
+	ncurses-dev \
+	libreadline-dev
diff --git a/misc/docker/debian-testing-i386/Dockerfile b/misc/docker/debian-testing-i386/Dockerfile
new file mode 100644
index 00000000..998ce969
--- /dev/null
+++ b/misc/docker/debian-testing-i386/Dockerfile
@@ -0,0 +1,12 @@
+FROM i386/debian:testing-slim
+ENV DEBIAN_FRONTEND noninteractive
+RUN sed -i 's/deb.debian.org/ftp.cz.debian.org/' /etc/apt/sources.list
+RUN apt-get -y update
+RUN apt-get -y upgrade
+RUN apt-get -y install \
+	autoconf \
+	build-essential \
+	flex \
+	bison \
+	ncurses-dev \
+	libreadline-dev
diff --git a/misc/docker/fedora-25-amd64/Dockerfile b/misc/docker/fedora-25-amd64/Dockerfile
new file mode 100644
index 00000000..9cc4c3f6
--- /dev/null
+++ b/misc/docker/fedora-25-amd64/Dockerfile
@@ -0,0 +1,10 @@
+FROM fedora:25
+RUN dnf -y upgrade
+RUN dnf -y install \
+	autoconf \
+	flex \
+	bison \
+	pkgconfig \
+	'readline-devel' \
+	'pkgconfig(ncurses)' \
+	gcc
diff --git a/misc/docker/fedora-26-amd64/Dockerfile b/misc/docker/fedora-26-amd64/Dockerfile
new file mode 100644
index 00000000..d1e84439
--- /dev/null
+++ b/misc/docker/fedora-26-amd64/Dockerfile
@@ -0,0 +1,10 @@
+FROM fedora:26
+RUN dnf -y upgrade
+RUN dnf -y install \
+	autoconf \
+	flex \
+	bison \
+	pkgconfig \
+	'readline-devel' \
+	'pkgconfig(ncurses)' \
+	gcc
diff --git a/misc/docker/opensuse-42.3-amd64/Dockerfile b/misc/docker/opensuse-42.3-amd64/Dockerfile
new file mode 100644
index 00000000..c9f7127d
--- /dev/null
+++ b/misc/docker/opensuse-42.3-amd64/Dockerfile
@@ -0,0 +1,11 @@
+FROM opensuse:42.3
+RUN zypper -n up
+RUN zypper -n install \
+	autoconf \
+	flex \
+	bison \
+	pkgconfig \
+	readline-devel \
+	ncurses-devel \
+	gcc \
+	gmake
diff --git a/misc/docker/ubuntu-14.04-amd64/Dockerfile b/misc/docker/ubuntu-14.04-amd64/Dockerfile
new file mode 100644
index 00000000..ccac3eb4
--- /dev/null
+++ b/misc/docker/ubuntu-14.04-amd64/Dockerfile
@@ -0,0 +1,12 @@
+FROM ubuntu:14.04
+ENV DEBIAN_FRONTEND noninteractive
+RUN sed -i 's/deb.debian.org/ftp.cz.debian.org/' /etc/apt/sources.list
+RUN apt-get -y update
+RUN apt-get -y upgrade
+RUN apt-get -y install \
+	autoconf \
+	build-essential \
+	flex \
+	bison \
+	ncurses-dev \
+	libreadline-dev
diff --git a/misc/docker/ubuntu-16.04-amd64/Dockerfile b/misc/docker/ubuntu-16.04-amd64/Dockerfile
new file mode 100644
index 00000000..50a4918c
--- /dev/null
+++ b/misc/docker/ubuntu-16.04-amd64/Dockerfile
@@ -0,0 +1,12 @@
+FROM ubuntu:16.04
+ENV DEBIAN_FRONTEND noninteractive
+RUN sed -i 's/deb.debian.org/ftp.cz.debian.org/' /etc/apt/sources.list
+RUN apt-get -y update
+RUN apt-get -y upgrade
+RUN apt-get -y install \
+	autoconf \
+	build-essential \
+	flex \
+	bison \
+	ncurses-dev \
+	libreadline-dev

From 9befc7cc4f26889077ace537019de92903139133 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Mon, 4 Sep 2017 22:32:45 +0200
Subject: [PATCH 019/198] BSD: Fix alignment issue

Incorrect structure alignment breaks kernel routing table updates on
FreeBSD/ARM (and perhaps other platforms).

Thanks to Eugene Sevastyanov for the original patch.
---
 sysdep/bsd/krt-sock.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sysdep/bsd/krt-sock.c b/sysdep/bsd/krt-sock.c
index 9c9df51d..f0cebd11 100644
--- a/sysdep/bsd/krt-sock.c
+++ b/sysdep/bsd/krt-sock.c
@@ -168,7 +168,7 @@ struct ks_msg
 {
   struct rt_msghdr rtm;
   struct sockaddr_storage buf[RTAX_MAX];
-};
+} PACKED;
 
 #define ROUNDUP(a) \
         ((a) > 0 ? (1 + (((a) - 1) | (sizeof(long) - 1))) : sizeof(long))

From 98bb80a243b58c43453e9be69d19d0350286549c Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 5 Sep 2017 00:02:20 +0200
Subject: [PATCH 020/198] KRT: Fix IPv6 ECMP handling with Linux 4.11+

Starting from Linux 4.11, IPv6 ECMP routes are now notified using
RTA_MULTIPATH, like IPv4 ones. The patch adds support for RTA_MULTIPATH
parsing for IPv6 routes. This also enables to parse ECMP alien routes
correctly.

Thanks to Vincent Bernat for the original patch.
---
 sysdep/linux/netlink.c | 55 ++++++++++++++++++++++++++++++++----------
 1 file changed, 42 insertions(+), 13 deletions(-)

diff --git a/sysdep/linux/netlink.c b/sysdep/linux/netlink.c
index 22313f43..3658c46b 100644
--- a/sysdep/linux/netlink.c
+++ b/sysdep/linux/netlink.c
@@ -59,22 +59,26 @@
 /*
  * Structure nl_parse_state keeps state of received route processing. Ideally,
  * we could just independently parse received Netlink messages and immediately
- * propagate received routes to the rest of BIRD, but Linux kernel represents
- * and announces IPv6 ECMP routes not as one route with multiple next hops (like
- * RTA_MULTIPATH in IPv4 ECMP), but as a set of routes with the same prefix.
+ * propagate received routes to the rest of BIRD, but older Linux kernel (before
+ * version 4.11) represents and announces IPv6 ECMP routes not as one route with
+ * multiple next hops (like RTA_MULTIPATH in IPv4 ECMP), but as a sequence of
+ * routes with the same prefix. More recent kernels work as with IPv4.
  *
  * Therefore, BIRD keeps currently processed route in nl_parse_state structure
  * and postpones its propagation until we expect it to be final; i.e., when
  * non-matching route is received or when the scan ends. When another matching
  * route is received, it is merged with the already processed route to form an
  * ECMP route. Note that merging is done only for IPv6 (merge == 1), but the
- * postponing is done in both cases (for simplicity). All IPv4 routes are just
- * considered non-matching.
+ * postponing is done in both cases (for simplicity). All IPv4 routes or IPv6
+ * routes with RTA_MULTIPATH set are just considered non-matching.
  *
  * This is ignored for asynchronous notifications (every notification is handled
  * as a separate route). It is not an issue for our routes, as we ignore such
  * notifications anyways. But importing alien IPv6 ECMP routes does not work
- * properly.
+ * properly with older kernels.
+ *
+ * Whatever the kernel version is, IPv6 ECMP routes are sent as multiple routes
+ * for the same prefix.
  */
 
 struct nl_parse_state
@@ -320,9 +324,15 @@ static struct nl_want_attrs ifa_attr_want6[BIRD_IFA_MAX] = {
 
 #define BIRD_RTA_MAX  (RTA_TABLE+1)
 
+#ifndef IPV6
 static struct nl_want_attrs mpnh_attr_want4[BIRD_RTA_MAX] = {
   [RTA_GATEWAY]	  = { 1, 1, sizeof(ip4_addr) },
 };
+#else
+static struct nl_want_attrs mpnh_attr_want6[BIRD_RTA_MAX] = {
+  [RTA_GATEWAY]	  = { 1, 1, sizeof(ip6_addr) },
+};
+#endif
 
 #ifndef IPV6
 static struct nl_want_attrs rtm_attr_want4[BIRD_RTA_MAX] = {
@@ -345,6 +355,7 @@ static struct nl_want_attrs rtm_attr_want6[BIRD_RTA_MAX] = {
   [RTA_PRIORITY]  = { 1, 1, sizeof(u32) },
   [RTA_PREFSRC]	  = { 1, 1, sizeof(ip6_addr) },
   [RTA_METRICS]	  = { 1, 0, 0 },
+  [RTA_MULTIPATH] = { 1, 0, 0 },
   [RTA_FLOW]	  = { 1, 1, sizeof(u32) },
   [RTA_TABLE]	  = { 1, 1, sizeof(u32) },
 };
@@ -477,7 +488,7 @@ nl_add_multipath(struct nlmsghdr *h, unsigned bufsize, struct mpnh *nh)
 }
 
 static struct mpnh *
-nl_parse_multipath(struct krt_proto *p, struct rtattr *ra)
+nl_parse_multipath(struct krt_proto *p, struct rtattr *ra, int af)
 {
   /* Temporary buffer for multicast nexthops */
   static struct mpnh *nh_buffer;
@@ -515,10 +526,26 @@ nl_parse_multipath(struct krt_proto *p, struct rtattr *ra)
 
       /* Nonexistent RTNH_PAYLOAD ?? */
       nl_attr_len = nh->rtnh_len - RTNH_LENGTH(0);
-      nl_parse_attrs(RTNH_DATA(nh), mpnh_attr_want4, a, sizeof(a));
+      switch (af)
+        {
+#ifndef IPV6
+	case AF_INET:
+	  if (!nl_parse_attrs(RTNH_DATA(nh), mpnh_attr_want4, a, sizeof(a)))
+	    return NULL;
+	  break;
+#else
+	case AF_INET6:
+	  if (!nl_parse_attrs(RTNH_DATA(nh), mpnh_attr_want6, a, sizeof(a)))
+	    return NULL;
+	  break;
+#endif
+	default:
+	  return NULL;
+	}
+
       if (a[RTA_GATEWAY])
 	{
-	  memcpy(&rv->gw, RTA_DATA(a[RTA_GATEWAY]), sizeof(ip_addr));
+	  memcpy(&rv->gw, RTA_DATA(a[RTA_GATEWAY]), sizeof(rv->gw));
 	  ipa_ntoh(rv->gw);
 
 	  neighbor *ng = neigh_find2(&p->p, &rv->gw, rv->iface,
@@ -1240,10 +1267,10 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h)
     {
     case RTN_UNICAST:
 
-      if (a[RTA_MULTIPATH] && (i->rtm_family == AF_INET))
+      if (a[RTA_MULTIPATH])
 	{
 	  ra->dest = RTD_MULTIPATH;
-	  ra->nexthops = nl_parse_multipath(p, a[RTA_MULTIPATH]);
+	  ra->nexthops = nl_parse_multipath(p, a[RTA_MULTIPATH], i->rtm_family);
 	  if (!ra->nexthops)
 	    {
 	      log(L_ERR "KRT: Received strange multipath route %I/%d",
@@ -1385,8 +1412,10 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h)
 
   /*
    * Ideally, now we would send the received route to the rest of kernel code.
-   * But IPv6 ECMP routes are sent as a sequence of routes, so we postpone it
-   * and merge next hops until the end of the sequence.
+   * But IPv6 ECMP routes before 4.11 are sent as a sequence of routes, so we
+   * postpone it and merge next hops until the end of the sequence. Note that
+   * proper multipath updates are rejected by nl_mergable_route(), so it is
+   * always the first case for them.
    */
 
   if (!s->net)

From 943478b00f585725c3e7406909ee867dcfac5f87 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 6 Sep 2017 17:38:48 +0200
Subject: [PATCH 021/198] Basic VRF support

Add basic VRF (virtual routing and forwarding) support. Protocols can be
associated with VRFs, such protocols will be restricted to interfaces
assigned to the VRF (as reported by Linux kernel) and will use sockets
bound to the VRF. E.g., different multihop BGP instances can use diffent
kernel routing tables to handle BGP TCP connections.

The VRF support is preliminary, currently there are several limitations:

- Recent Linux kernels (4.11) do not handle correctly sockets bound
to interaces that are part of VRF, so most protocols other than multihop
BGP do not work. This will be fixed by future kernel versions.

- Neighbor cache ignores VRFs. Breaks config with the same prefix on
local interfaces in different VRFs. Not much problem as single hop
protocols do not work anyways.

- Olock code ignores VRFs. Breaks config with multiple BGP peers with the
same IP address in different VRFs.

- Incoming BGP connections are not dispatched according to VRFs.
Breaks config with multiple BGP peers with the same IP address in
different VRFs. Perhaps we would need some kernel API to read VRF of
incoming connection? Or probably use multiple listening sockets in
int-new branch.

- We should handle master VRF interface up/down events and perhaps
disable associated protocols when VRF goes down. Or at least disable
associated interfaces.

- Also we should check if the master iface is really VRF iface and
not some other kind of master iface.

- BFD session request dispatch should be aware of VRFs.

- Perhaps kernel protocol should read default kernel table ID from VRF
iface so it is not necessary to configure it.

- Perhaps we should have per-VRF default table.
---
 doc/bird.sgml          |  9 +++++++++
 lib/socket.h           |  1 +
 nest/config.Y          |  3 ++-
 nest/iface.c           | 19 ++++++++++++++-----
 nest/iface.h           |  2 ++
 nest/proto.c           |  6 +++++-
 nest/protocol.h        |  2 ++
 proto/babel/packets.c  |  1 +
 proto/bgp/bgp.c        |  1 +
 proto/ospf/iface.c     |  2 ++
 proto/radv/packets.c   |  1 +
 proto/rip/packets.c    |  1 +
 sysdep/linux/netlink.c | 29 ++++++++++++++++++++++++++++-
 sysdep/unix/io.c       | 12 ++++++++++++
 14 files changed, 81 insertions(+), 8 deletions(-)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index 516f4c6e..03d37f46 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -598,6 +598,15 @@ agreement").
 
 	<tag><label id="proto-table">table <m/name/</tag>
 	Connect this protocol to a non-default routing table.
+
+	<tag><label id="proto-vrf">vrf "<m/text/"</tag>
+	Associate the protocol with specific VRF. The protocol will be
+	restricted to interfaces assigned to the VRF and will use sockets bound
+	to the VRF. Appropriate VRF interface must exist on OS level. For kernel
+	protocol, an appropriate table still must be explicitly selected by
+	<cf/table/ option. Note that the VRF support in BIRD and Linux kernel
+	(4.11) is still in development and is currently problematic outside of
+	multihop BGP.
 </descrip>
 
 <p>There are several options that give sense only with certain protocols:
diff --git a/lib/socket.h b/lib/socket.h
index 2da52127..0769489b 100644
--- a/lib/socket.h
+++ b/lib/socket.h
@@ -26,6 +26,7 @@ typedef struct birdsock {
   int ttl;				/* Time To Live, -1 = default */
   u32 flags;
   struct iface *iface;			/* Interface; specify this for broad/multicast sockets */
+  struct iface *vrf;			/* Related VRF instance, NULL if global */
 
   byte *rbuf, *rpos;			/* NULL=allocate automatically */
   uint fast_rx;				/* RX has higher priority in event loop */
diff --git a/nest/config.Y b/nest/config.Y
index 878224fe..025e8969 100644
--- a/nest/config.Y
+++ b/nest/config.Y
@@ -55,7 +55,7 @@ get_passwords(void)
 CF_DECLS
 
 CF_KEYWORDS(ROUTER, ID, PROTOCOL, TEMPLATE, PREFERENCE, DISABLED, DEBUG, ALL, OFF, DIRECT)
-CF_KEYWORDS(INTERFACE, IMPORT, EXPORT, FILTER, NONE, TABLE, STATES, ROUTES, FILTERS)
+CF_KEYWORDS(INTERFACE, IMPORT, EXPORT, FILTER, NONE, VRF, 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(ALGORITHM, KEYED, HMAC, MD5, SHA1, SHA256, SHA384, SHA512)
@@ -227,6 +227,7 @@ proto_item:
  | IMPORT LIMIT limit_spec { this_proto->in_limit = $3; }
  | EXPORT LIMIT limit_spec { this_proto->out_limit = $3; }
  | IMPORT KEEP FILTERED bool { this_proto->in_keep_filtered = $4; }
+ | VRF TEXT { this_proto->vrf = if_get_by_name($2); }
  | TABLE rtable { this_proto->table = $2; }
  | ROUTER ID idval { this_proto->router_id = $3; }
  | DESCRIPTION text { this_proto->dsc = $2; }
diff --git a/nest/iface.c b/nest/iface.c
index ff362938..3dd45065 100644
--- a/nest/iface.c
+++ b/nest/iface.c
@@ -116,7 +116,7 @@ if_what_changed(struct iface *i, struct iface *j)
   unsigned c;
 
   if (((i->flags ^ j->flags) & ~(IF_UP | IF_SHUTDOWN | IF_UPDATED | IF_ADMIN_UP | IF_LINK_UP | IF_TMP_DOWN | IF_JUST_CREATED))
-      || i->index != j->index)
+      || (i->index != j->index) || (i->master != j->master))
     return IF_CHANGE_TOO_MUCH;
   c = 0;
   if ((i->flags ^ j->flags) & IF_UP)
@@ -133,12 +133,14 @@ if_copy(struct iface *to, struct iface *from)
 {
   to->flags = from->flags | (to->flags & IF_TMP_DOWN);
   to->mtu = from->mtu;
+  to->master_index = from->master_index;
+  to->master = from->master;
 }
 
 static inline void
 ifa_send_notify(struct proto *p, unsigned c, struct ifa *a)
 {
-  if (p->ifa_notify)
+  if (p->ifa_notify && (!p->vrf || p->vrf == a->iface->master))
     {
       if (p->debug & D_IFACES)
 	log(L_TRACE "%s <%s address %I/%d on interface %s %s",
@@ -175,7 +177,7 @@ ifa_notify_change(unsigned c, struct ifa *a)
 static inline void
 if_send_notify(struct proto *p, unsigned c, struct iface *i)
 {
-  if (p->if_notify)
+  if (p->if_notify && (!p->vrf || p->vrf == i->master))
     {
       if (p->debug & D_IFACES)
 	log(L_TRACE "%s < interface %s %s", p->name, i->name,
@@ -238,7 +240,8 @@ if_recalc_flags(struct iface *i, unsigned flags)
 {
   if ((flags & (IF_SHUTDOWN | IF_TMP_DOWN)) ||
       !(flags & IF_ADMIN_UP) ||
-      !i->addr)
+      !i->addr ||
+      (i->master_index && !i->master))
     flags &= ~IF_UP;
   else
     flags |= IF_UP;
@@ -771,7 +774,13 @@ if_show(void)
       if (i->flags & IF_SHUTDOWN)
 	continue;
 
-      cli_msg(-1001, "%s %s (index=%d)", i->name, (i->flags & IF_UP) ? "up" : "DOWN", i->index);
+      char mbuf[16 + sizeof(i->name)] = {};
+      if (i->master)
+	bsprintf(mbuf, " master=%s", i->master->name);
+      else if (i->master_index)
+	bsprintf(mbuf, " master=#%u", i->master_index);
+
+      cli_msg(-1001, "%s %s (index=%d%s)", i->name, (i->flags & IF_UP) ? "up" : "DOWN", i->index, mbuf);
       if (!(i->flags & IF_MULTIACCESS))
 	type = "PtP";
       else
diff --git a/nest/iface.h b/nest/iface.h
index 56710e4a..b8e69838 100644
--- a/nest/iface.h
+++ b/nest/iface.h
@@ -34,8 +34,10 @@ struct iface {
   unsigned flags;
   unsigned mtu;
   unsigned index;			/* OS-dependent interface index */
+  unsigned master_index;		/* Interface index of master iface */
   list addrs;				/* Addresses assigned to this interface */
   struct ifa *addr;			/* Primary address */
+  struct iface *master;			/* Master iface (e.g. for VRF) */
   list neighbors;			/* All neighbors on this interface */
 };
 
diff --git a/nest/proto.c b/nest/proto.c
index 1091b321..64422ee3 100644
--- a/nest/proto.c
+++ b/nest/proto.c
@@ -386,6 +386,7 @@ proto_init(struct proto_config *c)
   q->core_state = FS_HUNGRY;
   q->export_state = ES_DOWN;
   q->last_state_change = now;
+  q->vrf = c->vrf;
 
   add_tail(&initial_proto_list, &q->n);
 
@@ -409,6 +410,7 @@ proto_reconfigure(struct proto *p, struct proto_config *oc, struct proto_config
   /* If there is a too big change in core attributes, ... */
   if ((nc->protocol != oc->protocol) ||
       (nc->disabled != p->disabled) ||
+      (nc->vrf != oc->vrf) ||
       (nc->table->table != oc->table->table))
     return 0;
 
@@ -1474,7 +1476,9 @@ proto_show_limit(struct proto_limit *l, const char *dsc)
 void
 proto_show_basic_info(struct proto *p)
 {
-  // cli_msg(-1006, "  Table:          %s", p->table->name);
+  if (p->vrf)
+    cli_msg(-1006, "  VRF:            %s", p->vrf->name);
+
   cli_msg(-1006, "  Preference:     %d", p->preference);
   cli_msg(-1006, "  Input filter:   %s", filter_name(p->cf->in_filter));
   cli_msg(-1006, "  Output filter:  %s", filter_name(p->cf->out_filter));
diff --git a/nest/protocol.h b/nest/protocol.h
index ec787355..3c30a581 100644
--- a/nest/protocol.h
+++ b/nest/protocol.h
@@ -94,6 +94,7 @@ struct proto_config {
   unsigned preference, disabled;	/* Generic parameters */
   int in_keep_filtered;			/* Routes rejected in import filter are kept */
   u32 router_id;			/* Protocol specific router ID */
+  struct iface *vrf;			/* Related VRF instance, NULL if global */
   struct rtable_config *table;		/* Table we're attached to */
   struct filter *in_filter, *out_filter; /* Attached filters */
   struct proto_limit *rx_limit;		/* Limit for receiving routes from protocol
@@ -213,6 +214,7 @@ struct proto {
   void (*rte_insert)(struct network *, struct rte *);
   void (*rte_remove)(struct network *, struct rte *);
 
+  struct iface *vrf;			/* Related VRF instance, NULL if global */
   struct rtable *table;			/* Our primary routing table */
   struct rte_src *main_source;		/* Primary route source */
   struct announce_hook *main_ahook;	/* Primary announcement hook */
diff --git a/proto/babel/packets.c b/proto/babel/packets.c
index 08054832..90421836 100644
--- a/proto/babel/packets.c
+++ b/proto/babel/packets.c
@@ -1080,6 +1080,7 @@ babel_open_socket(struct babel_iface *ifa)
   sk->sport = ifa->cf->port;
   sk->dport = ifa->cf->port;
   sk->iface = ifa->iface;
+  sk->vrf = p->p.vrf;
 
   sk->rx_hook = babel_rx_hook;
   sk->tx_hook = babel_tx_hook;
diff --git a/proto/bgp/bgp.c b/proto/bgp/bgp.c
index f706e76e..913685a4 100644
--- a/proto/bgp/bgp.c
+++ b/proto/bgp/bgp.c
@@ -745,6 +745,7 @@ bgp_connect(struct bgp_proto *p)	/* Enter Connect state and start establishing c
   s->daddr = p->cf->remote_ip;
   s->dport = p->cf->remote_port;
   s->iface = p->neigh ? p->neigh->iface : NULL;
+  s->vrf = p->p.vrf;
   s->ttl = p->cf->ttl_security ? 255 : hops;
   s->rbsize = p->cf->enable_extended_messages ? BGP_RX_BUFFER_EXT_SIZE : BGP_RX_BUFFER_SIZE;
   s->tbsize = p->cf->enable_extended_messages ? BGP_TX_BUFFER_EXT_SIZE : BGP_TX_BUFFER_SIZE;
diff --git a/proto/ospf/iface.c b/proto/ospf/iface.c
index 280fa4c1..1795ec22 100644
--- a/proto/ospf/iface.c
+++ b/proto/ospf/iface.c
@@ -118,6 +118,7 @@ ospf_sk_open(struct ospf_iface *ifa)
   sk->dport = OSPF_PROTO;
   sk->saddr = ifa->addr->ip;
   sk->iface = ifa->iface;
+  sk->vrf = p->p.vrf;
 
   sk->tos = ifa->cf->tx_tos;
   sk->priority = ifa->cf->tx_priority;
@@ -200,6 +201,7 @@ ospf_open_vlink_sk(struct ospf_proto *p)
   sock *sk = sk_new(p->p.pool);
   sk->type = SK_IP;
   sk->dport = OSPF_PROTO;
+  sk->vrf = p->p.vrf;
 
   /* FIXME: configurable tos/priority ? */
   sk->tos = IP_PREC_INTERNET_CONTROL;
diff --git a/proto/radv/packets.c b/proto/radv/packets.c
index 19d71f97..8a301854 100644
--- a/proto/radv/packets.c
+++ b/proto/radv/packets.c
@@ -388,6 +388,7 @@ radv_sk_open(struct radv_iface *ifa)
   sk->type = SK_IP;
   sk->dport = ICMPV6_PROTO;
   sk->saddr = ifa->addr->ip;
+  sk->vrf = ifa->ra->p.vrf;
 
   sk->ttl = 255; /* Mandatory for Neighbor Discovery packets */
   sk->rx_hook = radv_rx_hook;
diff --git a/proto/rip/packets.c b/proto/rip/packets.c
index 468927e6..722a9012 100644
--- a/proto/rip/packets.c
+++ b/proto/rip/packets.c
@@ -739,6 +739,7 @@ rip_open_socket(struct rip_iface *ifa)
   sk->sport = ifa->cf->port;
   sk->dport = ifa->cf->port;
   sk->iface = ifa->iface;
+  sk->vrf = p->p.vrf;
 
   /*
    * For RIPv2, we explicitly choose a primary address, mainly to ensure that
diff --git a/sysdep/linux/netlink.c b/sysdep/linux/netlink.c
index 3658c46b..4802897b 100644
--- a/sysdep/linux/netlink.c
+++ b/sysdep/linux/netlink.c
@@ -300,6 +300,7 @@ struct nl_want_attrs {
 static struct nl_want_attrs ifla_attr_want[BIRD_IFLA_MAX] = {
   [IFLA_IFNAME]	  = { 1, 0, 0 },
   [IFLA_MTU]	  = { 1, 1, sizeof(u32) },
+  [IFLA_MASTER]	  = { 1, 1, sizeof(u32) },
   [IFLA_WIRELESS] = { 1, 0, 0 },
 };
 
@@ -618,7 +619,7 @@ nl_parse_link(struct nlmsghdr *h, int scan)
   struct iface f = {};
   struct iface *ifi;
   char *name;
-  u32 mtu;
+  u32 mtu, master = 0;
   uint fl;
 
   if (!(i = nl_checkin(h, sizeof(*i))) || !nl_parse_attrs(IFLA_RTA(i), ifla_attr_want, a, sizeof(a)))
@@ -641,6 +642,9 @@ nl_parse_link(struct nlmsghdr *h, int scan)
   name = RTA_DATA(a[IFLA_IFNAME]);
   mtu = rta_get_u32(a[IFLA_MTU]);
 
+  if (a[IFLA_MASTER])
+    master = rta_get_u32(a[IFLA_MASTER]);
+
   ifi = if_find_by_index(i->ifi_index);
   if (!new)
     {
@@ -660,6 +664,9 @@ nl_parse_link(struct nlmsghdr *h, int scan)
       f.index = i->ifi_index;
       f.mtu = mtu;
 
+      f.master_index = master;
+      f.master = if_find_by_index(master);
+
       fl = i->ifi_flags;
       if (fl & IFF_UP)
 	f.flags |= IF_ADMIN_UP;
@@ -835,6 +842,26 @@ kif_do_scan(struct kif_proto *p UNUSED)
     else
       log(L_DEBUG "nl_scan_ifaces: Unknown packet received (type=%d)", h->nlmsg_type);
 
+  /* Re-resolve master interface for slaves */
+  struct iface *i;
+  WALK_LIST(i, iface_list)
+    if (i->master_index)
+    {
+      struct iface f = {
+	.flags = i->flags,
+	.mtu = i->mtu,
+	.index = i->index,
+	.master_index = i->master_index,
+	.master = if_find_by_index(i->master_index)
+      };
+
+      if (f.master != i->master)
+      {
+	memcpy(f.name, i->name, sizeof(f.name));
+	if_update(&f);
+      }
+    }
+
   nl_request_dump(BIRD_AF, RTM_GETADDR);
   while (h = nl_get_scan())
     if (h->nlmsg_type == RTM_NEWADDR || h->nlmsg_type == RTM_DELADDR)
diff --git a/sysdep/unix/io.c b/sysdep/unix/io.c
index 561d4dea..6d6a0990 100644
--- a/sysdep/unix/io.c
+++ b/sysdep/unix/io.c
@@ -1211,6 +1211,18 @@ sk_setup(sock *s)
   }
 #endif
 
+  if (s->vrf && !s->iface)
+  {
+    /* Bind socket to associated VRF interface.
+       This is Linux-specific, but so is SO_BINDTODEVICE. */
+#ifdef SO_BINDTODEVICE
+    struct ifreq ifr = {};
+    strcpy(ifr.ifr_name, s->vrf->name);
+    if (setsockopt(s->fd, SOL_SOCKET, SO_BINDTODEVICE, &ifr, sizeof(ifr)) < 0)
+      ERR("SO_BINDTODEVICE");
+#endif
+  }
+
   if (s->iface)
   {
 #ifdef SO_BINDTODEVICE

From 9f4908fe78cb3e5191bca721588ee1acb10876e3 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 12 Sep 2017 15:49:36 +0200
Subject: [PATCH 022/198] Nest: VRF support for neighbor cache and olock code

Actually much simpler than expected.
---
 nest/locks.c      | 1 +
 nest/locks.h      | 1 +
 nest/neighbor.c   | 8 +++++---
 proto/bgp/bgp.c   | 1 +
 proto/radv/radv.c | 1 -
 5 files changed, 8 insertions(+), 4 deletions(-)

diff --git a/nest/locks.c b/nest/locks.c
index 84b8b0ae..86c9ff14 100644
--- a/nest/locks.c
+++ b/nest/locks.c
@@ -45,6 +45,7 @@ olock_same(struct object_lock *x, struct object_lock *y)
   return
     x->type == y->type &&
     x->iface == y->iface &&
+    x->vrf == y->vrf &&
     x->port == y->port &&
     x->inst == y->inst &&
     ipa_equal(x->addr, y->addr);
diff --git a/nest/locks.h b/nest/locks.h
index 3d58c8ed..37026c68 100644
--- a/nest/locks.h
+++ b/nest/locks.h
@@ -30,6 +30,7 @@ struct object_lock {
   uint port;		/* ... port number */
   uint inst;		/* ... instance ID */
   struct iface *iface;	/* ... interface */
+  struct iface *vrf;	/* ... or VRF (if iface is unknown) */
   void (*hook)(struct object_lock *);	/* Called when the lock succeeds */
   void *data;		/* User data */
   /* ... internal to lock manager, don't touch ... */
diff --git a/nest/neighbor.c b/nest/neighbor.c
index d974fa51..f8159d35 100644
--- a/nest/neighbor.c
+++ b/nest/neighbor.c
@@ -30,7 +30,8 @@
  * when the protocol has explicitly requested it via the %NEF_STICKY
  * flag because it wishes to be notified when the node will again become
  * a neighbor. Such entries are enqueued in a special list which is walked
- * whenever an interface changes its state to up.
+ * whenever an interface changes its state to up. Neighbor entry VRF
+ * association is implied by respective protocol.
  *
  * When a neighbor event occurs (a neighbor gets disconnected or a sticky
  * inactive neighbor becomes connected), the protocol hook neigh_notify()
@@ -152,8 +153,9 @@ neigh_find2(struct proto *p, ip_addr *a, struct iface *ifa, unsigned flags)
     }
   else
     WALK_LIST(i, iface_list)
-      if ((scope = if_connected(a, i, &addr)) >= 0)
-	{
+      if ((!p->vrf || p->vrf == i->master) &&
+	  ((scope = if_connected(a, i, &addr)) >= 0))
+        {
 	  ifa = i;
 	  break;
 	}
diff --git a/proto/bgp/bgp.c b/proto/bgp/bgp.c
index 913685a4..8a6b2f02 100644
--- a/proto/bgp/bgp.c
+++ b/proto/bgp/bgp.c
@@ -1181,6 +1181,7 @@ bgp_start(struct proto *P)
   lock->addr = p->cf->remote_ip;
   lock->port = p->cf->remote_port;
   lock->iface = p->cf->iface;
+  lock->vrf = p->cf->iface ? NULL : p->p.vrf;
   lock->type = OBJLOCK_TCP;
   lock->hook = bgp_start_locked;
   lock->data = p;
diff --git a/proto/radv/radv.c b/proto/radv/radv.c
index 227c8ef6..c53a0a95 100644
--- a/proto/radv/radv.c
+++ b/proto/radv/radv.c
@@ -321,7 +321,6 @@ radv_iface_new(struct radv_proto *p, struct iface *iface, struct radv_iface_conf
   ifa->timer = tm;
 
   struct object_lock *lock = olock_new(pool);
-  lock->addr = IPA_NONE;
   lock->type = OBJLOCK_IP;
   lock->port = ICMPV6_PROTO;
   lock->iface = iface;

From f2dd602fef2ecf0a6598b817d71ce2ee8fadd5cc Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 12 Sep 2017 16:33:29 +0200
Subject: [PATCH 023/198] Backport some minor changes from int-new

---
 lib/birdlib.h     |  3 +++
 nest/protocol.h   | 12 ++++++++++++
 nest/route.h      |  5 +++--
 nest/rt-table.c   | 16 ++++------------
 proto/bgp/attrs.c |  2 --
 5 files changed, 22 insertions(+), 16 deletions(-)

diff --git a/lib/birdlib.h b/lib/birdlib.h
index d21cdf1f..aaa7a0a3 100644
--- a/lib/birdlib.h
+++ b/lib/birdlib.h
@@ -37,6 +37,9 @@ struct align_probe { char x; long int y; };
 #define ABS(a)   ((a)>=0 ? (a) : -(a))
 #define DELTA(a,b) (((a)>=(b))?(a)-(b):(b)-(a))
 #define ARRAY_SIZE(a) (sizeof(a)/sizeof(*(a)))
+#define BYTES(n) ((((uint) (n)) + 7) / 8)
+#define CALL(fn, args...) ({ if (fn) fn(args); })
+#define ADVANCE(w, r, l) ({ r -= l; w += l; })
 
 
 /* Bitfield macros */
diff --git a/nest/protocol.h b/nest/protocol.h
index 3c30a581..18dfbd6f 100644
--- a/nest/protocol.h
+++ b/nest/protocol.h
@@ -12,6 +12,7 @@
 #include "lib/lists.h"
 #include "lib/resource.h"
 #include "lib/timer.h"
+#include "nest/route.h"
 #include "conf/conf.h"
 
 struct iface;
@@ -287,6 +288,17 @@ proto_get_router_id(struct proto_config *pc)
   return pc->router_id ? pc->router_id : pc->global->router_id;
 }
 
+static inline struct ea_list *
+rte_make_tmp_attrs(struct rte *rt, struct linpool *pool)
+{
+  struct ea_list *(*mta)(struct rte *rt, struct linpool *pool);
+  mta = rt->attrs->src->proto->make_tmp_attrs;
+  return mta ? mta(rt, pool) : NULL;
+}
+
+/* Moved from route.h to avoid dependency conflicts */
+static inline void rte_update(struct proto *p, net *net, rte *new) { rte_update2(p->main_ahook, net, new, p->main_source); }
+
 extern list active_proto_list;
 
 /*
diff --git a/nest/route.h b/nest/route.h
index 383f4def..2e6ae5bf 100644
--- a/nest/route.h
+++ b/nest/route.h
@@ -12,10 +12,11 @@
 #include "lib/lists.h"
 #include "lib/resource.h"
 #include "lib/timer.h"
-#include "nest/protocol.h"
 
+struct ea_list;
 struct protocol;
 struct proto;
+struct rte_src;
 struct symbol;
 struct filter;
 struct cli;
@@ -273,7 +274,7 @@ static inline net *net_get(rtable *tab, ip_addr addr, unsigned len) { return (ne
 rte *rte_find(net *net, struct rte_src *src);
 rte *rte_get_temp(struct rta *);
 void rte_update2(struct announce_hook *ah, net *net, rte *new, struct rte_src *src);
-static inline void rte_update(struct proto *p, net *net, rte *new) { rte_update2(p->main_ahook, net, new, p->main_source); }
+/* rte_update() moved to protocol.h to avoid dependency conflicts */
 int rt_examine(rtable *t, ip_addr prefix, int pxlen, struct proto *p, struct filter *filter);
 rte *rt_export_merged(struct announce_hook *ah, net *net, rte **rt_free, struct ea_list **tmpa, linpool *pool, int silent);
 void rt_refresh_begin(rtable *t, struct announce_hook *ah);
diff --git a/nest/rt-table.c b/nest/rt-table.c
index c6e48c38..bcb48b53 100644
--- a/nest/rt-table.c
+++ b/nest/rt-table.c
@@ -60,14 +60,6 @@ static inline void rt_schedule_gc(rtable *tab);
 static inline void rt_schedule_prune(rtable *tab);
 
 
-static inline struct ea_list *
-make_tmp_attrs(struct rte *rt, struct linpool *pool)
-{
-  struct ea_list *(*mta)(struct rte *rt, struct linpool *pool);
-  mta = rt->attrs->src->proto->make_tmp_attrs;
-  return mta ? mta(rt, pool) : NULL;
-}
-
 /* Like fib_route(), but skips empty net entries */
 static net *
 net_route(rtable *tab, ip_addr a, int len)
@@ -260,7 +252,7 @@ export_filter_(struct announce_hook *ah, rte *rt0, rte **rt_free, ea_list **tmpa
   if (!tmpa)
     tmpa = &tmpb;
 
-  *tmpa = make_tmp_attrs(rt, pool);
+  *tmpa = rte_make_tmp_attrs(rt, pool);
 
   v = p->import_control ? p->import_control(p, &rt, tmpa, pool) : 0;
   if (v < 0)
@@ -1223,7 +1215,7 @@ rte_update2(struct announce_hook *ah, net *net, rte *new, struct rte_src *src)
 	}
       else
 	{
-	  tmpa = make_tmp_attrs(new, rte_update_pool);
+	  tmpa = rte_make_tmp_attrs(new, rte_update_pool);
 	  if (filter && (filter != FILTER_REJECT))
 	    {
 	      ea_list *old_tmpa = tmpa;
@@ -1303,7 +1295,7 @@ rt_examine(rtable *t, ip_addr prefix, int pxlen, struct proto *p, struct filter
   rte_update_lock();
 
   /* Rest is stripped down export_filter() */
-  ea_list *tmpa = make_tmp_attrs(rt, rte_update_pool);
+  ea_list *tmpa = rte_make_tmp_attrs(rt, rte_update_pool);
   int v = p->import_control ? p->import_control(p, &rt, &tmpa, rte_update_pool) : 0;
   if (v == RIC_PROCESS)
     v = (f_run(filter, &rt, &tmpa, rte_update_pool, FF_FORCE_TMPATTR) <= F_ACCEPT);
@@ -2470,7 +2462,7 @@ rt_show_net(struct cli *c, net *n, struct rt_show_data *d)
 
       ee = e;
       rte_update_lock();		/* We use the update buffer for filtering */
-      tmpa = make_tmp_attrs(e, rte_update_pool);
+      tmpa = rte_make_tmp_attrs(e, rte_update_pool);
 
       /* Special case for merged export */
       if ((d->export_mode == RSEM_EXPORT) && (d->export_protocol->accept_ra_types == RA_MERGED))
diff --git a/proto/bgp/attrs.c b/proto/bgp/attrs.c
index b9e2490d..73eb4040 100644
--- a/proto/bgp/attrs.c
+++ b/proto/bgp/attrs.c
@@ -469,8 +469,6 @@ bgp_get_attr_len(eattr *a)
   return len;
 }
 
-#define ADVANCE(w, r, l) do { r -= l; w += l; } while (0)
-
 /**
  * bgp_encode_attrs - encode BGP attributes
  * @p: BGP instance

From 18ea2ea759963a68a1a9f82d50aa9ed90b128df9 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 13 Sep 2017 13:10:50 +0200
Subject: [PATCH 024/198] Doc: Document 'empty' operator

Thanks to Alexander Zubkov for the notification.
---
 doc/bird.sgml | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index 03d37f46..c549e3c3 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -1196,6 +1196,8 @@ foot).
 
 	<cf><m/P/.len</cf> returns the length of path <m/P/.
 
+	<cf><m/P/.empty</cf> resets path <m/P/ to empty path.
+
 	<cf>prepend(<m/P/,<m/A/)</cf> prepends ASN <m/A/ to path <m/P/ and
 	returns the result.
 
@@ -1234,6 +1236,8 @@ foot).
 
 	<cf><m/C/.len</cf> returns the length of clist <m/C/.
 
+	<cf><m/C/.empty</cf> resets clist <m/C/ to empty clist.
+
 	<cf>add(<m/C/,<m/P/)</cf> adds pair (or quad) <m/P/ to clist <m/C/ and
 	returns the result. If item <m/P/ is already in clist <m/C/, it does
 	nothing. <m/P/ may also be a clist, in that case all its members are

From 7b2c5f3d2826e3175bf31b1c36056c9efc587a2b Mon Sep 17 00:00:00 2001
From: Michal 'vorner' Vaner <michal.vaner@nic.cz>
Date: Thu, 14 Sep 2017 12:31:15 +0200
Subject: [PATCH 025/198] Docs: FIB iteration macros

---
 nest/rt-fib.c | 18 ++++++++++++++++++
 1 file changed, 18 insertions(+)

diff --git a/nest/rt-fib.c b/nest/rt-fib.c
index a73de1fd..9af333c9 100644
--- a/nest/rt-fib.c
+++ b/nest/rt-fib.c
@@ -32,6 +32,24 @@
  * Basic FIB operations are performed by functions defined by this module,
  * enumerating of FIB contents is accomplished by using the FIB_WALK() macro
  * or FIB_ITERATE_START() if you want to do it asynchronously.
+ *
+ * For simple iteration just place the body of the loop between FIB_WALK() and
+ * FIB_WALK_END(). You can't modify the FIB during the iteration (you can modify
+ * data in the node, but not add or remove nodes).
+ *
+ * If you need more freedom, you can use the FIB_ITERATE_*() group of macros.
+ * First, you initialize an iterator with FIB_ITERATE_INIT(). Then you can put
+ * the loop body in between FIB_ITERATE_START() and FIB_ITERATE_END(). In
+ * addition, the iteration can be suspended by calling FIB_ITERATE_PUT().
+ * This'll link the iterator inside the FIB. While suspended, you may modify the
+ * FIB, exit the current function, etc. To resume the iteration, enter the loop
+ * again. You can use FIB_ITERATE_UNLINK() to unlink the iterator (while
+ * iteration is suspended) in cases like premature end of FIB iteration.
+ *
+ * Note that the iterator must not be destroyed when the iteration is suspended,
+ * the FIB would then contain a pointer to invalid memory. Therefore, after each
+ * FIB_ITERATE_INIT() or FIB_ITERATE_PUT() there must be either
+ * FIB_ITERATE_START() or FIB_ITERATE_UNLINK() before the iterator is destroyed.
  */
 
 #undef LOCAL_DEBUG

From cd1d99611e445c9fe2452d05627ccfc624f35c39 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 19 Sep 2017 19:55:37 +0200
Subject: [PATCH 026/198] BGP: Shutdown communication (RFC 8203)

The patch implements BGP Administrative Shutdown Communication (RFC 8203)
allowing BGP operators to pass messages related to BGP session
administrative shutdown/restart. It handles both transmit and receive of
shutdown messages. Messages are logged and may be displayed by show
protocol all command.

Thanks to Job Snijders for the basic patch.
---
 nest/config.Y         | 12 ++++-----
 nest/proto.c          | 63 +++++++++++++++++++++++++++++++++++--------
 nest/protocol.h       | 18 +++++++------
 proto/babel/packets.c |  2 --
 proto/bgp/bgp.c       | 48 +++++++++++++++++++++++----------
 proto/bgp/bgp.h       |  2 +-
 proto/bgp/packets.c   | 58 ++++++++++++++++++++++++++++++---------
 7 files changed, 149 insertions(+), 54 deletions(-)

diff --git a/nest/config.Y b/nest/config.Y
index 025e8969..f51be6ea 100644
--- a/nest/config.Y
+++ b/nest/config.Y
@@ -696,12 +696,12 @@ echo_size:
    }
  ;
 
-CF_CLI(DISABLE, proto_patt, <protocol> | \"<pattern>\" | all, [[Disable protocol]])
-{ proto_apply_cmd($2, proto_cmd_disable, 1, 0); } ;
-CF_CLI(ENABLE, proto_patt, <protocol> | \"<pattern>\" | all, [[Enable protocol]])
-{ proto_apply_cmd($2, proto_cmd_enable, 1, 0); } ;
-CF_CLI(RESTART, proto_patt, <protocol> | \"<pattern>\" | all, [[Restart protocol]])
-{ proto_apply_cmd($2, proto_cmd_restart, 1, 0); } ;
+CF_CLI(DISABLE, proto_patt text_or_none, (<protocol> | \"<pattern>\" | all) [message], [[Disable protocol]])
+{ proto_apply_cmd($2, proto_cmd_disable, 1, (uintptr_t) $3); } ;
+CF_CLI(ENABLE, proto_patt text_or_none, (<protocol> | \"<pattern>\" | all) [message], [[Enable protocol]])
+{ proto_apply_cmd($2, proto_cmd_enable, 1, (uintptr_t) $3); } ;
+CF_CLI(RESTART, proto_patt text_or_none, (<protocol> | \"<pattern>\" | all) [message], [[Restart protocol]])
+{ proto_apply_cmd($2, proto_cmd_restart, 1, (uintptr_t) $3); } ;
 CF_CLI(RELOAD, proto_patt, <protocol> | \"<pattern>\" | all, [[Reload protocol]])
 { proto_apply_cmd($2, proto_cmd_reload, 1, CMD_RELOAD); } ;
 CF_CLI(RELOAD IN, proto_patt, <protocol> | \"<pattern>\" | all, [[Reload protocol (just imported routes)]])
diff --git a/nest/proto.c b/nest/proto.c
index 64422ee3..552d53ae 100644
--- a/nest/proto.c
+++ b/nest/proto.c
@@ -610,6 +610,7 @@ proto_rethink_goal(struct proto *p)
       config_del_obstacle(p->cf->global);
       rem_node(&p->n);
       rem_node(&p->glob_node);
+      mb_free(p->message);
       mb_free(p);
       if (!nc)
 	return;
@@ -1096,6 +1097,39 @@ proto_schedule_down(struct proto *p, byte restart, byte code)
   tm_start_max(proto_shutdown_timer, restart ? 2 : 0);
 }
 
+/**
+ * proto_set_message - set administrative message to protocol
+ * @p: protocol
+ * @msg: message
+ * @len: message length (-1 for NULL-terminated string)
+ *
+ * The function sets administrative message (string) related to protocol state
+ * change. It is called by the nest code for manual enable/disable/restart
+ * commands all routes to the protocol, and by protocol-specific code when the
+ * protocol state change is initiated by the protocol. Using NULL message clears
+ * the last message. The message string may be either NULL-terminated or with an
+ * explicit length.
+ */
+void
+proto_set_message(struct proto *p, char *msg, int len)
+{
+  mb_free(p->message);
+  p->message = NULL;
+
+  if (!msg || !len)
+    return;
+
+  if (len < 0)
+    len = strlen(msg);
+
+  if (!len)
+    return;
+
+  p->message = mb_alloc(proto_pool, len + 1);
+  memcpy(p->message, msg, len);
+  p->message[len] = 0;
+}
+
 
 /**
  * proto_request_feeding - request feeding routes to the protocol
@@ -1497,7 +1531,7 @@ proto_show_basic_info(struct proto *p)
 }
 
 void
-proto_cmd_show(struct proto *p, uint verbose, int cnt)
+proto_cmd_show(struct proto *p, uintptr_t verbose, int cnt)
 {
   byte buf[256], tbuf[TM_DATETIME_BUFFER_SIZE];
 
@@ -1520,6 +1554,10 @@ proto_cmd_show(struct proto *p, uint verbose, int cnt)
     {
       if (p->cf->dsc)
 	cli_msg(-1006, "  Description:    %s", p->cf->dsc);
+
+      if (p->message)
+	cli_msg(-1006, "  Message:        %s", p->message);
+
       if (p->cf->router_id)
 	cli_msg(-1006, "  Router ID:      %R", p->cf->router_id);
 
@@ -1533,7 +1571,7 @@ proto_cmd_show(struct proto *p, uint verbose, int cnt)
 }
 
 void
-proto_cmd_disable(struct proto *p, uint arg UNUSED, int cnt UNUSED)
+proto_cmd_disable(struct proto *p, uintptr_t arg, int cnt UNUSED)
 {
   if (p->disabled)
     {
@@ -1544,12 +1582,13 @@ proto_cmd_disable(struct proto *p, uint arg UNUSED, int cnt UNUSED)
   log(L_INFO "Disabling protocol %s", p->name);
   p->disabled = 1;
   p->down_code = PDC_CMD_DISABLE;
+  proto_set_message(p, (char *) arg, -1);
   proto_rethink_goal(p);
   cli_msg(-9, "%s: disabled", p->name);
 }
 
 void
-proto_cmd_enable(struct proto *p, uint arg UNUSED, int cnt UNUSED)
+proto_cmd_enable(struct proto *p, uintptr_t arg, int cnt UNUSED)
 {
   if (!p->disabled)
     {
@@ -1559,12 +1598,13 @@ proto_cmd_enable(struct proto *p, uint arg UNUSED, int cnt UNUSED)
 
   log(L_INFO "Enabling protocol %s", p->name);
   p->disabled = 0;
+  proto_set_message(p, (char *) arg, -1);
   proto_rethink_goal(p);
   cli_msg(-11, "%s: enabled", p->name);
 }
 
 void
-proto_cmd_restart(struct proto *p, uint arg UNUSED, int cnt UNUSED)
+proto_cmd_restart(struct proto *p, uintptr_t arg, int cnt UNUSED)
 {
   if (p->disabled)
     {
@@ -1575,6 +1615,7 @@ proto_cmd_restart(struct proto *p, uint arg UNUSED, int cnt UNUSED)
   log(L_INFO "Restarting protocol %s", p->name);
   p->disabled = 1;
   p->down_code = PDC_CMD_RESTART;
+  proto_set_message(p, (char *) arg, -1);
   proto_rethink_goal(p);
   p->disabled = 0;
   proto_rethink_goal(p);
@@ -1582,7 +1623,7 @@ proto_cmd_restart(struct proto *p, uint arg UNUSED, int cnt UNUSED)
 }
 
 void
-proto_cmd_reload(struct proto *p, uint dir, int cnt UNUSED)
+proto_cmd_reload(struct proto *p, uintptr_t dir, int cnt UNUSED)
 {
   if (p->disabled)
     {
@@ -1624,19 +1665,19 @@ proto_cmd_reload(struct proto *p, uint dir, int cnt UNUSED)
 }
 
 void
-proto_cmd_debug(struct proto *p, uint mask, int cnt UNUSED)
+proto_cmd_debug(struct proto *p, uintptr_t mask, int cnt UNUSED)
 {
   p->debug = mask;
 }
 
 void
-proto_cmd_mrtdump(struct proto *p, uint mask, int cnt UNUSED)
+proto_cmd_mrtdump(struct proto *p, uintptr_t mask, int cnt UNUSED)
 {
   p->mrtdump = mask;
 }
 
 static void
-proto_apply_cmd_symbol(struct symbol *s, void (* cmd)(struct proto *, uint, int), uint arg)
+proto_apply_cmd_symbol(struct symbol *s, void (* cmd)(struct proto *, uintptr_t, int), uintptr_t arg)
 {
   if (s->class != SYM_PROTO)
     {
@@ -1649,7 +1690,7 @@ proto_apply_cmd_symbol(struct symbol *s, void (* cmd)(struct proto *, uint, int)
 }
 
 static void
-proto_apply_cmd_patt(char *patt, void (* cmd)(struct proto *, uint, int), uint arg)
+proto_apply_cmd_patt(char *patt, void (* cmd)(struct proto *, uintptr_t, int), uintptr_t arg)
 {
   int cnt = 0;
 
@@ -1669,8 +1710,8 @@ proto_apply_cmd_patt(char *patt, void (* cmd)(struct proto *, uint, int), uint a
 }
 
 void
-proto_apply_cmd(struct proto_spec ps, void (* cmd)(struct proto *, uint, int),
-		int restricted, uint arg)
+proto_apply_cmd(struct proto_spec ps, void (* cmd)(struct proto *, uintptr_t, int),
+		int restricted, uintptr_t arg)
 {
   if (restricted && cli_access_restricted())
     return;
diff --git a/nest/protocol.h b/nest/protocol.h
index 18dfbd6f..5aca9a4e 100644
--- a/nest/protocol.h
+++ b/nest/protocol.h
@@ -164,6 +164,7 @@ struct proto {
   u32 hash_key;				/* Random key used for hashing of neighbors */
   bird_clock_t last_state_change;	/* Time of last state transition */
   char *last_state_name_announced;	/* Last state name we've announced to the user */
+  char *message;			/* State-change message, allocated from proto_pool */
   struct proto_stats stats;		/* Current protocol statistics */
 
   /*
@@ -250,6 +251,7 @@ struct proto_spec {
 void *proto_new(struct proto_config *, unsigned size);
 void *proto_config_new(struct protocol *, int class);
 void proto_copy_config(struct proto_config *dest, struct proto_config *src);
+void proto_set_message(struct proto *p, char *msg, int len);
 void proto_request_feeding(struct proto *p);
 
 static inline void
@@ -267,15 +269,15 @@ void proto_graceful_restart_unlock(struct proto *p);
 void proto_show_limit(struct proto_limit *l, const char *dsc);
 void proto_show_basic_info(struct proto *p);
 
-void proto_cmd_show(struct proto *, uint, int);
-void proto_cmd_disable(struct proto *, uint, int);
-void proto_cmd_enable(struct proto *, uint, int);
-void proto_cmd_restart(struct proto *, uint, int);
-void proto_cmd_reload(struct proto *, uint, int);
-void proto_cmd_debug(struct proto *, uint, int);
-void proto_cmd_mrtdump(struct proto *, uint, int);
+void proto_cmd_show(struct proto *, uintptr_t, int);
+void proto_cmd_disable(struct proto *, uintptr_t, int);
+void proto_cmd_enable(struct proto *, uintptr_t, int);
+void proto_cmd_restart(struct proto *, uintptr_t, int);
+void proto_cmd_reload(struct proto *, uintptr_t, int);
+void proto_cmd_debug(struct proto *, uintptr_t, int);
+void proto_cmd_mrtdump(struct proto *, uintptr_t, int);
 
-void proto_apply_cmd(struct proto_spec ps, void (* cmd)(struct proto *, uint, int), int restricted, uint arg);
+void proto_apply_cmd(struct proto_spec ps, void (* cmd)(struct proto *, uintptr_t, int), int restricted, uintptr_t arg);
 struct proto *proto_get_named(struct symbol *, struct protocol *);
 
 #define CMD_RELOAD	0
diff --git a/proto/babel/packets.c b/proto/babel/packets.c
index 90421836..e9c6d51d 100644
--- a/proto/babel/packets.c
+++ b/proto/babel/packets.c
@@ -146,8 +146,6 @@ struct babel_write_state {
 #define TLV_HDR(tlv,t,l) ({ tlv->type = t; tlv->length = l - sizeof(struct babel_tlv); })
 #define TLV_HDR0(tlv,t) TLV_HDR(tlv, t, tlv_data[t].min_length)
 
-#define BYTES(n) ((((uint) n) + 7) / 8)
-
 static inline u16
 get_time16(const void *p)
 {
diff --git a/proto/bgp/bgp.c b/proto/bgp/bgp.c
index 8a6b2f02..b99672f5 100644
--- a/proto/bgp/bgp.c
+++ b/proto/bgp/bgp.c
@@ -290,7 +290,7 @@ bgp_update_startup_delay(struct bgp_proto *p)
 }
 
 static void
-bgp_graceful_close_conn(struct bgp_conn *conn, unsigned subcode)
+bgp_graceful_close_conn(struct bgp_conn *conn, uint subcode, byte *data, uint len)
 {
   switch (conn->state)
     {
@@ -304,7 +304,7 @@ bgp_graceful_close_conn(struct bgp_conn *conn, unsigned subcode)
     case BS_OPENSENT:
     case BS_OPENCONFIRM:
     case BS_ESTABLISHED:
-      bgp_error(conn, 6, subcode, NULL, 0);
+      bgp_error(conn, 6, subcode, data, len);
       return;
     default:
       bug("bgp_graceful_close_conn: Unknown state %d", conn->state);
@@ -340,11 +340,11 @@ bgp_decision(void *vp)
 }
 
 void
-bgp_stop(struct bgp_proto *p, unsigned subcode)
+bgp_stop(struct bgp_proto *p, uint subcode, byte *data, uint len)
 {
   proto_notify_state(&p->p, PS_STOP);
-  bgp_graceful_close_conn(&p->outgoing_conn, subcode);
-  bgp_graceful_close_conn(&p->incoming_conn, subcode);
+  bgp_graceful_close_conn(&p->outgoing_conn, subcode, data, len);
+  bgp_graceful_close_conn(&p->incoming_conn, subcode, data, len);
   ev_schedule(p->event);
 }
 
@@ -420,7 +420,7 @@ bgp_conn_leave_established_state(struct bgp_proto *p)
   bgp_free_bucket_table(p);
 
   if (p->p.proto_state == PS_UP)
-    bgp_stop(p, 0);
+    bgp_stop(p, 0, NULL, 0);
 }
 
 void
@@ -516,7 +516,7 @@ bgp_graceful_restart_timeout(timer *t)
   struct bgp_proto *p = t->data;
 
   BGP_TRACE(D_EVENTS, "Neighbor graceful restart timeout");
-  bgp_stop(p, 0);
+  bgp_stop(p, 0, NULL, 0);
 }
 
 
@@ -973,7 +973,7 @@ bgp_neigh_notify(neighbor *n)
 	  BGP_TRACE(D_EVENTS, "Neighbor lost");
 	  bgp_store_error(p, NULL, BE_MISC, BEM_NEIGHBOR_LOST);
 	  /* Perhaps also run bgp_update_startup_delay(p)? */
-	  bgp_stop(p, 0);
+	  bgp_stop(p, 0, NULL, 0);
 	}
     }
   else if (p->cf->check_link && !(n->iface->flags & IF_LINK_UP))
@@ -984,7 +984,7 @@ bgp_neigh_notify(neighbor *n)
 	  bgp_store_error(p, NULL, BE_MISC, BEM_LINK_DOWN);
 	  if (ps == PS_UP)
 	    bgp_update_startup_delay(p);
-	  bgp_stop(p, 0);
+	  bgp_stop(p, 0, NULL, 0);
 	}
     }
   else
@@ -1009,7 +1009,7 @@ bgp_bfd_notify(struct bfd_request *req)
       bgp_store_error(p, NULL, BE_MISC, BEM_BFD_DOWN);
       if (ps == PS_UP)
 	bgp_update_startup_delay(p);
-      bgp_stop(p, 0);
+      bgp_stop(p, 0, NULL, 0);
     }
 }
 
@@ -1196,7 +1196,11 @@ static int
 bgp_shutdown(struct proto *P)
 {
   struct bgp_proto *p = (struct bgp_proto *) P;
-  unsigned subcode = 0;
+  uint subcode = 0;
+
+  char *message = NULL;
+  byte *data = NULL;
+  uint len = 0;
 
   BGP_TRACE(D_EVENTS, "Shutdown requested");
 
@@ -1214,10 +1218,12 @@ bgp_shutdown(struct proto *P)
     case PDC_CMD_DISABLE:
     case PDC_CMD_SHUTDOWN:
       subcode = 2; // Errcode 6, 2 - administrative shutdown
+      message = P->message;
       break;
 
     case PDC_CMD_RESTART:
       subcode = 4; // Errcode 6, 4 - administrative reset
+      message = P->message;
       break;
 
     case PDC_RX_LIMIT_HIT:
@@ -1242,8 +1248,22 @@ bgp_shutdown(struct proto *P)
   bgp_store_error(p, NULL, BE_MAN_DOWN, 0);
   p->startup_delay = 0;
 
- done:
-  bgp_stop(p, subcode);
+  /* RFC 8203 - shutdown communication */
+  if (message)
+  {
+    uint msg_len = strlen(message);
+    msg_len = MIN(msg_len, 128);
+
+    /* Buffer will be freed automatically by protocol shutdown */
+    data = mb_alloc(p->p.pool, msg_len + 1);
+    len = msg_len + 1;
+
+    data[0] = msg_len;
+    memcpy(data+1, message, msg_len);
+  }
+
+done:
+  bgp_stop(p, subcode, data, len);
   return p->p.proto_state;
 }
 
@@ -1433,7 +1453,7 @@ bgp_error(struct bgp_conn *c, unsigned code, unsigned subcode, byte *data, int l
   if (code != 6)
     {
       bgp_update_startup_delay(p);
-      bgp_stop(p, 0);
+      bgp_stop(p, 0, NULL, 0);
     }
 }
 
diff --git a/proto/bgp/bgp.h b/proto/bgp/bgp.h
index e47a0eb1..22a150ab 100644
--- a/proto/bgp/bgp.h
+++ b/proto/bgp/bgp.h
@@ -212,7 +212,7 @@ void bgp_graceful_restart_done(struct bgp_proto *p);
 void bgp_refresh_begin(struct bgp_proto *p);
 void bgp_refresh_end(struct bgp_proto *p);
 void bgp_store_error(struct bgp_proto *p, struct bgp_conn *c, u8 class, u32 code);
-void bgp_stop(struct bgp_proto *p, unsigned subcode);
+void bgp_stop(struct bgp_proto *p, uint subcode, byte *data, uint len);
 
 struct rte_source *bgp_find_source(struct bgp_proto *p, u32 path_id);
 struct rte_source *bgp_get_source(struct bgp_proto *p, u32 path_id);
diff --git a/proto/bgp/packets.c b/proto/bgp/packets.c
index ab87bdcc..af3b15b5 100644
--- a/proto/bgp/packets.c
+++ b/proto/bgp/packets.c
@@ -1494,38 +1494,72 @@ bgp_error_dsc(unsigned code, unsigned subcode)
   return buff;
 }
 
+/* RFC 8203 - shutdown communication message */
+static int
+bgp_handle_message(struct bgp_proto *p, byte *data, uint len, byte **bp)
+{
+  byte *msg = data + 1;
+  uint msg_len = data[0];
+  uint i;
+
+  /* Handle zero length message */
+  if (msg_len == 0)
+    return 1;
+
+  /* Handle proper message */
+  if ((msg_len > 128) && (msg_len + 1 > len))
+    return 0;
+
+  /* Some elementary cleanup */
+  for (i = 0; i < msg_len; i++)
+    if (msg[i] < ' ')
+      msg[i] = ' ';
+
+  proto_set_message(&p->p, msg, msg_len);
+  *bp += bsprintf(*bp, ": \"%s\"", p->p.message);
+  return 1;
+}
+
 void
 bgp_log_error(struct bgp_proto *p, u8 class, char *msg, unsigned code, unsigned subcode, byte *data, unsigned len)
 {
-  const byte *name;
-  byte *t, argbuf[36];
+  byte argbuf[256], *t = argbuf;
   unsigned i;
 
   /* Don't report Cease messages generated by myself */
   if (code == 6 && class == BE_BGP_TX)
     return;
 
-  name = bgp_error_dsc(code, subcode);
-  t = argbuf;
+  /* Reset shutdown message */
+  if ((code == 6) && ((subcode == 2) || (subcode == 4)))
+    proto_set_message(&p->p, NULL, 0);
+
   if (len)
     {
-      *t++ = ':';
-      *t++ = ' ';
-
+      /* Bad peer AS - we would like to print the AS */
       if ((code == 2) && (subcode == 2) && ((len == 2) || (len == 4)))
 	{
-	  /* Bad peer AS - we would like to print the AS */
-	  t += bsprintf(t, "%d", (len == 2) ? get_u16(data) : get_u32(data));
+	  t += bsprintf(t, ": %u", (len == 2) ? get_u16(data) : get_u32(data));
 	  goto done;
 	}
+
+      /* RFC 8203 - shutdown communication */
+      if (((code == 6) && ((subcode == 2) || (subcode == 4))))
+	if (bgp_handle_message(p, data, len, &t))
+	  goto done;
+
+      *t++ = ':';
+      *t++ = ' ';
       if (len > 16)
 	len = 16;
       for (i=0; i<len; i++)
 	t += bsprintf(t, "%02x", data[i]);
     }
- done:
+
+done:
   *t = 0;
-  log(L_REMOTE "%s: %s: %s%s", p->p.name, msg, name, argbuf);
+  const byte *dsc = bgp_error_dsc(code, subcode);
+  log(L_REMOTE "%s: %s: %s%s", p->p.name, msg, dsc, argbuf);
 }
 
 static void
@@ -1571,7 +1605,7 @@ bgp_rx_notification(struct bgp_conn *conn, byte *pkt, uint len)
   if (err) 
     {
       bgp_update_startup_delay(p);
-      bgp_stop(p, 0);
+      bgp_stop(p, 0, NULL, 0);
     }
 }
 

From 5a8b1fb047d675badc17ab24175d0db06d7cc00c Mon Sep 17 00:00:00 2001
From: Michal 'vorner' Vaner <michal.vaner@nic.cz>
Date: Tue, 5 Sep 2017 15:50:00 +0200
Subject: [PATCH 027/198] filter: Allow assigning enums into extended
 attributes

They are internally ints, but they got refused as a wrong type. This
fixes setting of the BGP origin and is also needed for RA.
---
 filter/filter.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/filter/filter.c b/filter/filter.c
index f18970e0..1e8f7d5a 100644
--- a/filter/filter.c
+++ b/filter/filter.c
@@ -1088,7 +1088,8 @@ interpret(struct f_inst *what)
 
       switch (what->aux & EAF_TYPE_MASK) {
       case EAF_TYPE_INT:
-	if (v1.type != T_INT)
+	// Enums are also ints, so allow them in.
+	if (v1.type != T_INT && (v1.type < T_ENUM_LO || v1.type > T_ENUM_HI))
 	  runtime( "Setting int attribute to non-int value" );
 	l->attrs[0].u.data = v1.val.i;
 	break;

From 2a95e63343a94243745e5d7000bb3e0cb61a4a0f Mon Sep 17 00:00:00 2001
From: Michal 'vorner' Vaner <michal.vaner@nic.cz>
Date: Thu, 31 Aug 2017 15:40:23 +0200
Subject: [PATCH 028/198] RAdv: Support for more specific routes (RFC 4191)

The patch implements Default Router Preferences and More-Specific Routes
(RFC 4191) for RAdv protocol, allowing to announce router preference and
more specific routes in router advertisements. Routes can be exported to
RAdv like to regular routing protocols.

Some cleanups, bugfixes and other changes done by Ondrej Zajicek.
---
 filter/filter.h      |   1 +
 nest/route.h         |   3 +-
 proto/radv/config.Y  |  22 ++-
 proto/radv/packets.c |  72 ++++++++--
 proto/radv/radv.c    | 322 +++++++++++++++++++++++++++++++++----------
 proto/radv/radv.h    |  32 ++++-
 6 files changed, 357 insertions(+), 95 deletions(-)

diff --git a/filter/filter.h b/filter/filter.h
index 049ceb76..72b37461 100644
--- a/filter/filter.h
+++ b/filter/filter.h
@@ -168,6 +168,7 @@ void val_format(struct f_val v, buffer *buf);
 #define T_ENUM_RTC 0x33
 #define T_ENUM_RTD 0x34
 #define T_ENUM_ROA 0x35
+#define T_ENUM_RA_PREFERENCE 0x36
 /* new enums go here */
 #define T_ENUM_EMPTY 0x3f	/* Special hack for atomic_aggr */
 
diff --git a/nest/route.h b/nest/route.h
index 2e6ae5bf..0834da45 100644
--- a/nest/route.h
+++ b/nest/route.h
@@ -430,7 +430,8 @@ typedef struct eattr {
 #define EAP_OSPF 3			/* OSPF */
 #define EAP_KRT 4			/* Kernel route attributes */
 #define EAP_BABEL 5			/* Babel attributes */
-#define EAP_MAX 6
+#define EAP_RADV 6			/* Router advertisment attributes */
+#define EAP_MAX 7
 
 #define EA_CODE(proto,id) (((proto) << 8) | (id))
 #define EA_PROTO(ea) ((ea) >> 8)
diff --git a/proto/radv/config.Y b/proto/radv/config.Y
index 0ff84aeb..2fa11ce2 100644
--- a/proto/radv/config.Y
+++ b/proto/radv/config.Y
@@ -30,7 +30,10 @@ CF_KEYWORDS(RADV, PREFIX, INTERFACE, MIN, MAX, RA, DELAY, INTERVAL,
 	MANAGED, OTHER, CONFIG, LINGER, LINK, MTU, REACHABLE, TIME, RETRANS,
 	TIMER, CURRENT, HOP, LIMIT, DEFAULT, VALID, PREFERRED, MULT,
 	LIFETIME, SKIP, ONLINK, AUTONOMOUS, RDNSS, DNSSL, NS, DOMAIN,
-	LOCAL, TRIGGER, SENSITIVE, PREFERENCE, LOW, MEDIUM, HIGH)
+	LOCAL, TRIGGER, SENSITIVE, PREFERENCE, LOW, MEDIUM, HIGH, PROPAGATE,
+	ROUTE, ROUTES, RA_PREFERENCE, RA_LIFETIME)
+
+CF_ENUM(T_ENUM_RA_PREFERENCE, RA_PREF_, LOW, MEDIUM, HIGH)
 
 %type<i> radv_mult radv_sensitive radv_preference
 
@@ -45,6 +48,8 @@ radv_proto_start: proto_start RADV
   init_list(&RADV_CFG->pref_list);
   init_list(&RADV_CFG->rdnss_list);
   init_list(&RADV_CFG->dnssl_list);
+  RADV_CFG->route_lifetime = DEFAULT_VALID_LIFETIME;
+  RADV_CFG->route_linger_time = DEFAULT_LINGER_TIME;
 };
 
 radv_proto_item:
@@ -58,6 +63,16 @@ radv_proto_item:
      RADV_CFG->trigger_pxlen = $2.len;
      RADV_CFG->trigger_valid = 1;
    }
+ | PROPAGATE ROUTES bool { RADV_CFG->propagate_routes = $3; }
+ | ROUTE LIFETIME expr radv_sensitive {
+     RADV_CFG->route_lifetime = $3;
+     if ($4 != -1) RADV_CFG->route_lifetime_sensitive = $4;
+   }
+ | ROUTE LINGER TIME expr {
+     RADV_CFG->route_linger_time = $4;
+     if (($4 < 0) || ($4 > 3600))
+       cf_error("Linger time must be in range 0-3600");
+   }
  ;
 
 radv_proto_opts:
@@ -168,12 +183,10 @@ radv_prefix_item:
  | AUTONOMOUS bool { RADV_PREFIX->autonomous = $2; }
  | VALID LIFETIME expr radv_sensitive {
      RADV_PREFIX->valid_lifetime = $3;
-     if ($3 < 0) cf_error("Valid lifetime must be 0 or positive");
      if ($4 != -1) RADV_PREFIX->valid_lifetime_sensitive = $4;
    }
  | PREFERRED LIFETIME expr radv_sensitive {
      RADV_PREFIX->preferred_lifetime = $3;
-     if ($3 < 0) cf_error("Preferred lifetime must be 0 or positive");
      if ($4 != -1) RADV_PREFIX->preferred_lifetime_sensitive = $4;
    }
  ;
@@ -303,6 +316,9 @@ radv_sensitive:
  | SENSITIVE bool { $$ = $2; }
  ;
 
+CF_ADDTO(dynamic_attr, RA_PREFERENCE { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_ENUM_RA_PREFERENCE, EA_RA_PREFERENCE); })
+CF_ADDTO(dynamic_attr, RA_LIFETIME { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_RA_LIFETIME); })
+
 CF_CODE
 
 CF_END
diff --git a/proto/radv/packets.c b/proto/radv/packets.c
index 8a301854..fbe02060 100644
--- a/proto/radv/packets.c
+++ b/proto/radv/packets.c
@@ -26,6 +26,7 @@ struct radv_ra_packet
 
 #define OPT_PREFIX	3
 #define OPT_MTU		5
+#define OPT_ROUTE	24
 #define OPT_RDNSS	25
 #define OPT_DNSSL	31
 
@@ -52,6 +53,15 @@ struct radv_opt_mtu
   u32 mtu;
 };
 
+struct radv_opt_route {
+  u8 type;
+  u8 length;
+  u8 pxlen;
+  u8 flags;
+  u32 lifetime;
+  u8 prefix[];
+};
+
 struct radv_opt_rdnss
 {
   u8 type;
@@ -70,6 +80,41 @@ struct radv_opt_dnssl
   char domain[];
 };
 
+static int
+radv_prepare_route(struct radv_iface *ifa, struct radv_route *rt,
+		   char **buf, char *bufend)
+{
+  struct radv_proto *p = ifa->ra;
+  struct radv_config *cf = (void *) p->p.cf;
+  u8 px_blocks = (rt->n.pxlen + 63) / 64;
+  u8 opt_len = 8 * (1 + px_blocks);
+
+  if (*buf + opt_len > bufend)
+  {
+    log(L_WARN, "%s: Too many RA options on interface %s",
+	p->p.name, ifa->iface->name);
+    return -1;
+  }
+
+  struct radv_opt_route *opt = (void *) *buf;
+  *buf += opt_len;
+  opt->type = OPT_ROUTE;
+  opt->length = 1 + px_blocks;
+  opt->pxlen = rt->n.pxlen;
+  opt->flags = rt->preference;
+
+  if (p->valid && (p->active || !cf->route_lifetime_sensitive) && rt->alive)
+    opt->lifetime = htonl(rt->lifetime_set ? rt->lifetime : cf->route_lifetime);
+  else
+    opt->lifetime = 0;
+
+  /* Copy the relevant part of the prefix */
+  ip6_addr px_addr = ip6_hton(rt->n.prefix);
+  memcpy(opt->prefix, &px_addr, 8 * px_blocks);
+
+  return 0;
+}
+
 static int
 radv_prepare_rdnss(struct radv_iface *ifa, list *rdnss_list, char **buf, char *bufend)
 {
@@ -252,7 +297,7 @@ radv_prepare_ra(struct radv_iface *ifa)
   pkt->code = 0;
   pkt->checksum = 0;
   pkt->current_hop_limit = ic->current_hop_limit;
-  pkt->router_lifetime = (p->active || !ic->default_lifetime_sensitive) ?
+  pkt->router_lifetime = (p->valid && (p->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) |
@@ -292,13 +337,23 @@ radv_prepare_ra(struct radv_iface *ifa)
   if (radv_prepare_dnssl(ifa, &ic->dnssl_list, &buf, bufend) < 0)
     goto done;
 
+  if (p->fib_up)
+  {
+    FIB_WALK(&p->routes, rt)
+    {
+      if (radv_prepare_route(ifa, (struct radv_route *) rt, &buf, bufend) < 0)
+	goto done;
+    }
+    FIB_WALK_END;
+  }
+
  done:
   ifa->plen = buf - bufstart;
 }
 
 
 void
-radv_send_ra(struct radv_iface *ifa, int shutdown)
+radv_send_ra(struct radv_iface *ifa)
 {
   struct radv_proto *p = ifa->ra;
 
@@ -306,19 +361,6 @@ radv_send_ra(struct radv_iface *ifa, int shutdown)
   if (!ifa->plen)
     radv_prepare_ra(ifa);
 
-  if (shutdown)
-  {
-    /*
-     * 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);
   sk_send_to(ifa->sk, ifa->plen, IP6_ALL_NODES, 0);
 }
diff --git a/proto/radv/radv.c b/proto/radv/radv.c
index c53a0a95..d799dab5 100644
--- a/proto/radv/radv.c
+++ b/proto/radv/radv.c
@@ -12,37 +12,41 @@
 /**
  * DOC: Router Advertisements
  *
- * The RAdv protocol is implemented in two files: |radv.c| containing
- * the interface with BIRD core and the protocol logic and |packets.c|
- * handling low level protocol stuff (RX, TX and packet formats).
- * The protocol does not export any routes.
+ * The RAdv protocol is implemented in two files: |radv.c| containing the
+ * interface with BIRD core and the protocol logic and |packets.c| handling low
+ * level protocol stuff (RX, TX and packet formats). The protocol does not
+ * export any routes.
  *
- * The RAdv is structured in the usual way - for each handled interface
- * there is a structure &radv_iface that contains a state related to
- * that interface together with its resources (a socket, a timer).
- * There is also a prepared RA stored in a TX buffer of the socket
- * associated with an iface. These iface structures are created
- * and removed according to iface events from BIRD core handled by
- * radv_if_notify() callback.
+ * The RAdv is structured in the usual way - for each handled interface there is
+ * a structure &radv_iface that contains a state related to that interface
+ * together with its resources (a socket, a timer). There is also a prepared RA
+ * stored in a TX buffer of the socket associated with an iface. These iface
+ * structures are created and removed according to iface events from BIRD core
+ * handled by radv_if_notify() callback.
  *
- * The main logic of RAdv consists of two functions:
- * radv_iface_notify(), which processes asynchronous events (specified
- * by RA_EV_* codes), and radv_timer(), which triggers sending RAs and
- * computes the next timeout.
+ * The main logic of RAdv consists of two functions: radv_iface_notify(), which
+ * processes asynchronous events (specified by RA_EV_* codes), and radv_timer(),
+ * which triggers sending RAs and computes the next timeout.
  *
- * The RAdv protocol could receive routes (through
- * radv_import_control() and radv_rt_notify()), but only the
- * configured trigger route is tracked (in &active var).  When a radv
- * protocol is reconfigured, the connected routing table is examined
- * (in radv_check_active()) to have proper &active value in case of
- * the specified trigger prefix was changed.
+ * The RAdv protocol could receive routes (through radv_import_control() and
+ * radv_rt_notify()), but only the configured trigger route is tracked (in
+ * &active var).  When a radv protocol is reconfigured, the connected routing
+ * table is examined (in radv_check_active()) to have proper &active value in
+ * case of the specified trigger prefix was changed.
  *
  * Supported standards:
  * - RFC 4861 - main RA standard
+ * - RFC 4191 - Default Router Preferences and More-Specific Routes
  * - RFC 6106 - DNS extensions (RDDNS, DNSSL)
- * - RFC 4191 (partial) - Default Router Preference
  */
 
+static void radv_prune_prefixes(struct radv_iface *ifa);
+static void radv_prune_routes(struct radv_proto *p);
+
+/* Invalidate cached RA packet */
+static inline void radv_invalidate(struct radv_iface *ifa)
+{ ifa->plen = 0; }
+
 static void
 radv_timer(timer *tm)
 {
@@ -51,17 +55,13 @@ radv_timer(timer *tm)
 
   RADV_TRACE(D_EVENTS, "Timer fired on %s", ifa->iface->name);
 
-  /*
-   * If some dead prefixes expired, regenerate the prefix list and the packet.
-   * We do so by pretending there was a change on the interface.
-   *
-   * This sets the timer, but we replace it just at the end of this function
-   * (replacing a timer is fine).
-   */
-  if (ifa->prefix_expires && (ifa->prefix_expires <= now))
-    radv_iface_notify(ifa, RA_EV_GC);
+  if (ifa->prune_time <= now)
+    radv_prune_prefixes(ifa);
 
-  radv_send_ra(ifa, 0);
+  if (p->prune_time <= now)
+    radv_prune_routes(p);
+
+  radv_send_ra(ifa);
 
   /* Update timer */
   ifa->last = now;
@@ -118,7 +118,7 @@ radv_prepare_prefixes(struct radv_iface *ifa)
 {
   struct radv_proto *p = ifa->ra;
   struct radv_iface_config *cf = ifa->cf;
-  struct radv_prefix *pfx;
+  struct radv_prefix *pfx, *next;
 
   /* First mark all the prefixes as unused */
   WALK_LIST(pfx, ifa->prefixes)
@@ -162,49 +162,53 @@ radv_prepare_prefixes(struct radv_iface *ifa)
     existing->cf = pc;
   }
 
-  /*
-   * Garbage-collect the prefixes. If something isn't used, it dies (but isn't
-   * dropped just yet). If something is dead and rots there for long enough,
-   * clean it up.
-   */
-  bird_clock_t expires = now + cf->linger_time;
-  bird_clock_t expires_min = 0;
-  struct radv_prefix *next;
   WALK_LIST_DELSAFE(pfx, next, ifa->prefixes)
   {
     if (pfx->alive && !pfx->mark)
     {
-      RADV_TRACE(D_EVENTS, "Marking prefix %I/$d on %s as dead",
+      RADV_TRACE(D_EVENTS, "Invalidating prefix %I/$d on %s",
 		 pfx->prefix, pfx->len, ifa->iface->name);
 
       pfx->alive = 0;
-      pfx->expires = expires;
+      pfx->expires = now + cf->linger_time;
       pfx->cf = &dead_prefix;
     }
+  }
+}
 
-    if (!pfx->alive)
+static void
+radv_prune_prefixes(struct radv_iface *ifa)
+{
+  struct radv_proto *p = ifa->ra;
+  bird_clock_t next = TIME_INFINITY;
+  int changed = 0;
+
+  struct radv_prefix *px, *pxn;
+  WALK_LIST_DELSAFE(px, pxn, ifa->prefixes)
+  {
+    if (!px->alive)
     {
-      if (pfx->expires <= now)
+      if (px->expires <= now)
       {
 	RADV_TRACE(D_EVENTS, "Removing prefix %I/%d on %s",
-		   pfx->prefix, pfx->len, ifa->iface->name);
+		   px->prefix, px->len, ifa->iface->name);
 
-	rem_node(NODE pfx);
-	mb_free(pfx);
+	rem_node(NODE px);
+	mb_free(px);
+	changed = 1;
       }
       else
-      {
-	/* Find minimum expiration time */
-	if (!expires_min || (pfx->expires < expires_min))
-	  expires_min = pfx->expires;
-      }
+	next = MIN(next, px->expires);
     }
   }
 
-  ifa->prefix_expires = expires_min;
+  if (changed)
+    radv_invalidate(ifa);
+
+  ifa->prune_time = next;
 }
 
-static char* ev_name[] = { NULL, "Init", "Change", "RS", "Garbage collect" };
+static char* ev_name[] = { NULL, "Init", "Change", "RS" };
 
 void
 radv_iface_notify(struct radv_iface *ifa, int event)
@@ -219,18 +223,17 @@ radv_iface_notify(struct radv_iface *ifa, int event)
   switch (event)
   {
   case RA_EV_CHANGE:
-  case RA_EV_GC:
-    ifa->plen = 0;
+    radv_invalidate(ifa);
   case RA_EV_INIT:
     ifa->initial = MAX_INITIAL_RTR_ADVERTISEMENTS;
+    radv_prepare_prefixes(ifa);
+    radv_prune_prefixes(ifa);
     break;
 
   case RA_EV_RS:
     break;
   }
 
-  radv_prepare_prefixes(ifa);
-
   /* Update timer */
   unsigned delta = now - ifa->last;
   unsigned after = 0;
@@ -250,7 +253,6 @@ radv_iface_notify_all(struct radv_proto *p, int event)
     radv_iface_notify(ifa, event);
 }
 
-
 static struct radv_iface *
 radv_iface_find(struct radv_proto *p, struct iface *what)
 {
@@ -303,6 +305,7 @@ radv_iface_new(struct radv_proto *p, struct iface *iface, struct radv_iface_conf
   ifa->cf = cf;
   ifa->iface = iface;
   init_list(&ifa->prefixes);
+  ifa->prune_time = TIME_INFINITY;
 
   add_tail(&p->iface_list, NODE ifa);
 
@@ -409,14 +412,18 @@ radv_import_control(struct proto *P, rte **new, ea_list **attrs UNUSED, struct l
   if (radv_net_match_trigger(cf, (*new)->net))
     return RIC_PROCESS;
 
-  return RIC_DROP;
+  if (cf->propagate_routes)
+    return RIC_PROCESS;
+  else
+    return RIC_DROP;
 }
 
 static void
-radv_rt_notify(struct proto *P, rtable *tbl UNUSED, net *n, rte *new, rte *old UNUSED, ea_list *attrs UNUSED)
+radv_rt_notify(struct proto *P, rtable *tbl UNUSED, net *n, rte *new, rte *old UNUSED, ea_list *attrs)
 {
   struct radv_proto *p = (struct radv_proto *) P;
   struct radv_config *cf = (struct radv_config *) (P->cf);
+  struct radv_route *rt;
 
   if (radv_net_match_trigger(cf, n))
   {
@@ -432,7 +439,117 @@ radv_rt_notify(struct proto *P, rtable *tbl UNUSED, net *n, rte *new, rte *old U
       RADV_TRACE(D_EVENTS, "Suppressed");
 
     radv_iface_notify_all(p, RA_EV_CHANGE);
+    return;
   }
+
+  if (!cf->propagate_routes)
+    return;
+
+  /*
+   * Some other route we want to send (or stop sending). Update the cache,
+   * with marking a removed one as dead or creating a new one as needed.
+   *
+   * And yes, we exclude the trigger route on purpose.
+   */
+
+  if (new)
+  {
+    /* Update */
+    uint preference = ea_get_int(attrs, EA_RA_PREFERENCE, RA_PREF_MEDIUM);
+    uint lifetime = ea_get_int(attrs, EA_RA_LIFETIME, 0);
+    uint lifetime_set = !!ea_find(attrs, EA_RA_LIFETIME);
+
+    if ((preference != RA_PREF_LOW) &&
+	(preference != RA_PREF_MEDIUM) &&
+	(preference != RA_PREF_HIGH))
+    {
+      log(L_WARN "%s: Invalid ra_preference value %u on route %I/%d",
+	  p->p.name, preference, n->n.prefix, n->n.pxlen);
+      preference = RA_PREF_MEDIUM;
+      lifetime = 0;
+      lifetime_set = 1;
+    }
+
+    rt = fib_get(&p->routes, &n->n.prefix, n->n.pxlen);
+
+    /* Ignore update if nothing changed */
+    if (rt->alive &&
+	(rt->preference == preference) &&
+	(rt->lifetime == lifetime) &&
+	(rt->lifetime_set == lifetime_set))
+      return;
+
+    if (p->routes.entries == 18)
+      log(L_WARN "%s: More than 17 routes exported to RAdv", p->p.name);
+
+    rt->alive = 1;
+    rt->preference = preference;
+    rt->lifetime = lifetime;
+    rt->lifetime_set = lifetime_set;
+  }
+  else
+  {
+    /* Withdraw */
+    rt = fib_find(&p->routes, &n->n.prefix, n->n.pxlen);
+
+    if (!rt || !rt->alive)
+      return;
+
+    /* Invalidate the route */
+    rt->alive = 0;
+    rt->expires = now + cf->route_linger_time;
+    p->prune_time = MIN(p->prune_time, rt->expires);
+  }
+
+  radv_iface_notify_all(p, RA_EV_CHANGE);
+}
+
+/*
+ * Cleans up all the dead routes that expired and schedules itself to be run
+ * again if there are more routes waiting for expiration.
+ */
+static void
+radv_prune_routes(struct radv_proto *p)
+{
+  bird_clock_t next = TIME_INFINITY;
+  int changed = 0;
+
+  /* Should not happen */
+  if (!p->fib_up)
+    return;
+
+  struct fib_iterator fit;
+  FIB_ITERATE_INIT(&fit, &p->routes);
+
+again:
+  FIB_ITERATE_START(&p->routes, &fit, node)
+  {
+    struct radv_route *rt = (void *) node;
+
+    if (!rt->alive)
+    {
+      /* Delete expired nodes */
+      if (rt->expires <= now)
+      {
+	FIB_ITERATE_PUT(&fit, node);
+	fib_delete(&p->routes, node);
+	changed = 1;
+	goto again;
+      }
+      else
+	next = MIN(next, rt->expires);
+    }
+  }
+  FIB_ITERATE_END(node);
+
+  if (changed)
+  {
+    struct radv_iface *ifa;
+    WALK_LIST(ifa, p->iface_list)
+      radv_invalidate(ifa);
+  }
+
+  p->prune_time = next;
 }
 
 static int
@@ -461,6 +578,21 @@ radv_init(struct proto_config *c)
   return P;
 }
 
+static void
+radv_set_fib(struct radv_proto *p, int up)
+{
+  if (up == p->fib_up)
+    return;
+
+  if (up)
+    fib_init(&p->routes, p->p.pool, sizeof(struct radv_route), 4, NULL);
+  else
+    fib_free(&p->routes);
+
+  p->fib_up = up;
+  p->prune_time = TIME_INFINITY;
+}
+
 static int
 radv_start(struct proto *P)
 {
@@ -468,8 +600,13 @@ radv_start(struct proto *P)
   struct radv_config *cf = (struct radv_config *) (P->cf);
 
   init_list(&(p->iface_list));
+  p->valid = 1;
   p->active = !cf->trigger_valid;
 
+  p->fib_up = 0;
+  radv_set_fib(p, cf->propagate_routes);
+  p->prune_time = TIME_INFINITY;
+
   return PS_UP;
 }
 
@@ -477,7 +614,10 @@ static inline void
 radv_iface_shutdown(struct radv_iface *ifa)
 {
   if (ifa->sk)
-    radv_send_ra(ifa, 1);
+  {
+    radv_invalidate(ifa);
+    radv_send_ra(ifa);
+  }
 }
 
 static int
@@ -485,6 +625,8 @@ radv_shutdown(struct proto *P)
 {
   struct radv_proto *p = (struct radv_proto *) P;
 
+  p->valid = 0;
+
   struct radv_iface *ifa;
   WALK_LIST(ifa, p->iface_list)
     radv_iface_shutdown(ifa);
@@ -496,20 +638,19 @@ static int
 radv_reconfigure(struct proto *P, struct proto_config *c)
 {
   struct radv_proto *p = (struct radv_proto *) P;
-  // struct radv_config *old = (struct radv_config *) (p->cf);
+  struct radv_config *old = (struct radv_config *) (P->cf);
   struct radv_config *new = (struct radv_config *) c;
 
-  /*
-   * The question is why there is a reconfigure function for RAdv if
-   * it has almost none internal state so restarting the protocol
-   * would probably suffice. One small reason is that restarting the
-   * protocol would lead to sending a RA with Router Lifetime 0
-   * causing nodes to temporary remove their default routes.
-   */
-
   P->cf = c; /* radv_check_active() requires proper P->cf */
   p->active = radv_check_active(p);
 
+  /* Allocate or free FIB */
+  radv_set_fib(p, new->propagate_routes);
+
+  /* We started to accept routes so we need to refeed them */
+  if (!old->propagate_routes && new->propagate_routes)
+    proto_request_feeding(&p->p);
+
   struct iface *iface;
   WALK_LIST(iface, iface_list)
   {
@@ -561,14 +702,49 @@ radv_get_status(struct proto *P, byte *buf)
     strcpy(buf, "Suppressed");
 }
 
+static const char *
+radv_pref_str(u32 pref)
+{
+  switch (pref)
+  {
+    case RA_PREF_LOW:
+      return "low";
+    case RA_PREF_MEDIUM:
+      return "medium";
+    case RA_PREF_HIGH:
+      return "high";
+    default:
+      return "??";
+  }
+}
+
+/* The buffer has some minimal size */
+static int
+radv_get_attr(eattr *a, byte *buf, int buflen UNUSED)
+{
+  switch (a->id)
+  {
+  case EA_RA_PREFERENCE:
+    bsprintf(buf, "preference: %s", radv_pref_str(a->u.data));
+    return GA_FULL;
+  case EA_RA_LIFETIME:
+    bsprintf(buf, "lifetime");
+    return GA_NAME;
+  default:
+    return GA_UNKNOWN;
+  }
+}
+
 struct protocol proto_radv = {
   .name =		"RAdv",
   .template =		"radv%d",
+  .attr_class =		EAP_RADV,
   .config_size =	sizeof(struct radv_config),
   .init =		radv_init,
   .start =		radv_start,
   .shutdown =		radv_shutdown,
   .reconfigure =	radv_reconfigure,
   .copy_config =	radv_copy_config,
-  .get_status =		radv_get_status
+  .get_status =		radv_get_status,
+  .get_attr =		radv_get_attr
 };
diff --git a/proto/radv/radv.h b/proto/radv/radv.h
index 60b9980f..3bc18249 100644
--- a/proto/radv/radv.h
+++ b/proto/radv/radv.h
@@ -54,6 +54,10 @@ struct radv_config
   ip_addr trigger_prefix;	/* Prefix of a trigger route, if defined */
   u8 trigger_pxlen;		/* Pxlen of a trigger route, if defined */
   u8 trigger_valid;		/* Whether a trigger route is defined */
+  u8 propagate_routes;		/* Do we propagate more specific routes (RFC 4191)? */
+  u32 route_lifetime;		/* Lifetime for the RFC 4191 routes */
+  u32 route_lifetime_sensitive; /* Whether route_lifetime depends on trigger */
+  u32 route_linger_time;	/* For how long we advertise dead routes with lifetime = 0 */
 };
 
 struct radv_iface_config
@@ -117,12 +121,32 @@ struct radv_dnssl_config
   char *domain;			/* Domain for DNS search list, in processed form */
 };
 
+/*
+ * One more specific route as per RFC 4191.
+ *
+ * Note that it does *not* contain the next hop field. The next hop is always
+ * the router sending the advertisment and the more specific route only allows
+ * overriding the preference of the route.
+ */
+struct radv_route
+{
+  struct fib_node n;
+  u32 lifetime;			/* Lifetime from an attribute */
+  u8 lifetime_set;		/* Is the lifetime set by an attribute? */
+  u8 preference;		/* Preference of the route, RA_PREF_* */
+  u8 alive;
+  bird_clock_t expires;		/* Time to remove when !alive */
+};
 
 struct radv_proto
 {
   struct proto p;
   list iface_list;		/* List of active ifaces */
+  u8 valid;			/* Router is valid for forwarding, used for shutdown */
   u8 active;			/* Whether radv is active w.r.t. triggers */
+  u8 fib_up;			/* FIB table (routes) is initialized */
+  struct fib routes;		/* FIB table of specific routes (struct radv_route) */
+  bird_clock_t prune_time;	/* Next time of route table pruning */
 };
 
 struct radv_prefix		/* One prefix we advertise */
@@ -147,7 +171,7 @@ struct radv_iface
   struct ifa *addr;		/* Link-local address of iface */
   struct pool *pool;		/* A pool for interface-specific things */
   list prefixes;		/* The prefixes we advertise (struct radv_prefix) */
-  bird_clock_t prefix_expires;	/* When the soonest prefix expires (0 = none dead) */
+  bird_clock_t prune_time;	/* Next time of prefix list pruning */
 
   timer *timer;
   struct object_lock *lock;
@@ -161,7 +185,6 @@ struct radv_iface
 #define RA_EV_INIT 1		/* Switch to initial mode */
 #define RA_EV_CHANGE 2		/* Change of options or prefixes */
 #define RA_EV_RS 3		/* Received RS */
-#define RA_EV_GC 4		/* Internal garbage collection of prefixes */
 
 /* Default Router Preferences (RFC 4191) */
 #define RA_PREF_LOW	0x18
@@ -169,6 +192,9 @@ struct radv_iface
 #define RA_PREF_HIGH	0x08
 #define RA_PREF_MASK	0x18
 
+/* Attributes */
+#define EA_RA_PREFERENCE	EA_CODE(EAP_RADV, 0)
+#define EA_RA_LIFETIME		EA_CODE(EAP_RADV, 1)
 
 #ifdef LOCAL_DEBUG
 #define RADV_FORCE_DEBUG 1
@@ -184,7 +210,7 @@ void radv_iface_notify(struct radv_iface *ifa, int event);
 
 /* packets.c */
 int radv_process_domain(struct radv_dnssl_config *cf);
-void radv_send_ra(struct radv_iface *ifa, int shutdown);
+void radv_send_ra(struct radv_iface *ifa);
 int radv_sk_open(struct radv_iface *ifa);
 
 

From 7c0bab3a3987b42bc699c4417c7b2e838f189158 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Fri, 6 Oct 2017 12:22:18 +0200
Subject: [PATCH 029/198] RAdv: Change specific route options to be
 per-interface

And change default values of specific route options to be consistent with
values of default router options.
---
 proto/radv/config.Y  | 45 ++++++++++++++++++--------
 proto/radv/packets.c | 44 ++++++++++++++++++-------
 proto/radv/radv.c    | 77 +++++++++++++++++++++++++-------------------
 proto/radv/radv.h    | 27 ++++++++--------
 4 files changed, 121 insertions(+), 72 deletions(-)

diff --git a/proto/radv/config.Y b/proto/radv/config.Y
index 2fa11ce2..84a2de0e 100644
--- a/proto/radv/config.Y
+++ b/proto/radv/config.Y
@@ -48,8 +48,6 @@ radv_proto_start: proto_start RADV
   init_list(&RADV_CFG->pref_list);
   init_list(&RADV_CFG->rdnss_list);
   init_list(&RADV_CFG->dnssl_list);
-  RADV_CFG->route_lifetime = DEFAULT_VALID_LIFETIME;
-  RADV_CFG->route_linger_time = DEFAULT_LINGER_TIME;
 };
 
 radv_proto_item:
@@ -64,15 +62,6 @@ radv_proto_item:
      RADV_CFG->trigger_valid = 1;
    }
  | PROPAGATE ROUTES bool { RADV_CFG->propagate_routes = $3; }
- | ROUTE LIFETIME expr radv_sensitive {
-     RADV_CFG->route_lifetime = $3;
-     if ($4 != -1) RADV_CFG->route_lifetime_sensitive = $4;
-   }
- | ROUTE LINGER TIME expr {
-     RADV_CFG->route_linger_time = $4;
-     if (($4 < 0) || ($4 > 3600))
-       cf_error("Linger time must be in range 0-3600");
-   }
  ;
 
 radv_proto_opts:
@@ -96,11 +85,15 @@ radv_iface_start:
   RADV_IFACE->min_ra_int = -1; /* undefined */
   RADV_IFACE->max_ra_int = DEFAULT_MAX_RA_INT;
   RADV_IFACE->min_delay = DEFAULT_MIN_DELAY;
+  RADV_IFACE->prefix_linger_time = -1;
+  RADV_IFACE->route_linger_time = -1;
   RADV_IFACE->current_hop_limit = DEFAULT_CURRENT_HOP_LIMIT;
-  RADV_IFACE->linger_time = DEFAULT_LINGER_TIME;
   RADV_IFACE->default_lifetime = -1;
   RADV_IFACE->default_lifetime_sensitive = 1;
   RADV_IFACE->default_preference = RA_PREF_MEDIUM;
+  RADV_IFACE->route_lifetime = -1;
+  RADV_IFACE->route_lifetime_sensitive = 0;
+  RADV_IFACE->route_preference = RA_PREF_MEDIUM;
 };
 
 radv_iface_item:
@@ -112,14 +105,20 @@ radv_iface_item:
  | LINK MTU expr { RADV_IFACE->link_mtu = $3; if ($3 < 0) cf_error("Link MTU must be 0 or positive"); }
  | REACHABLE TIME expr { RADV_IFACE->reachable_time = $3; if (($3 < 0) || ($3 > 3600000)) cf_error("Reachable time must be in range 0-3600000"); }
  | RETRANS TIMER expr { RADV_IFACE->retrans_timer = $3; if ($3 < 0) cf_error("Retrans timer must be 0 or positive"); }
- | LINGER TIME expr { RADV_IFACE->linger_time = $3; if (($3 < 0) || ($3 > 3600)) cf_error("Linger time must be in range 0-3600"); }
  | CURRENT HOP LIMIT expr { RADV_IFACE->current_hop_limit = $4; if (($4 < 0) || ($4 > 255))  cf_error("Current hop limit must be in range 0-255"); }
  | DEFAULT LIFETIME expr radv_sensitive {
      RADV_IFACE->default_lifetime = $3;
      if (($3 < 0) || ($3 > 9000))  cf_error("Default lifetime must be in range 0-9000");
      if ($4 != -1) RADV_IFACE->default_lifetime_sensitive = $4;
    }
+ | ROUTE LIFETIME expr radv_sensitive {
+     RADV_IFACE->route_lifetime = $3;
+     if ($4 != -1) RADV_IFACE->route_lifetime_sensitive = $4;
+   }
  | DEFAULT PREFERENCE radv_preference { RADV_IFACE->default_preference = $3; }
+ | ROUTE PREFERENCE radv_preference { RADV_IFACE->route_preference = $3; }
+ | PREFIX LINGER TIME expr { RADV_IFACE->prefix_linger_time = $4; }
+ | ROUTE LINGER TIME expr { RADV_IFACE->route_linger_time = $4; }
  | 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); }
@@ -142,12 +141,32 @@ radv_iface_finish:
   if (ic->default_lifetime == (u32) -1)
     ic->default_lifetime = 3 * ic->max_ra_int;
 
+  if (ic->route_lifetime == (u32) -1)
+    ic->route_lifetime = 3 * ic->max_ra_int;
+
+  if (ic->prefix_linger_time == (u32) -1)
+    ic->prefix_linger_time = 3 * ic->max_ra_int;
+
+  if (ic->route_linger_time == (u32) -1)
+    ic->route_linger_time = 3 * ic->max_ra_int;
+
   if ((ic->min_ra_int > 3) &&
       (ic->min_ra_int > (ic->max_ra_int * 3 / 4)))
     cf_error("Min RA interval must be at most 3/4 * Max RA interval %d %d", ic->min_ra_int, ic->max_ra_int);
 
   if ((ic->default_lifetime > 0) && (ic->default_lifetime < ic->max_ra_int))
     cf_error("Default lifetime must be either 0 or at least Max RA interval");
+
+  if ((ic->route_lifetime > 0) && (ic->route_lifetime < ic->max_ra_int))
+    cf_error("Route lifetime must be either 0 or at least Max RA interval");
+
+  if ((ic->prefix_linger_time > 0) && (ic->prefix_linger_time < ic->max_ra_int))
+    cf_error("Prefix linger time must be either 0 or at least Max RA interval");
+
+  if ((ic->route_linger_time > 0) && (ic->route_linger_time < ic->max_ra_int))
+    cf_error("Route linger time must be either 0 or at least Max RA interval");
+
+  RADV_CFG->max_linger_time = MAX_(RADV_CFG->max_linger_time, ic->route_linger_time);
 };
 
 
diff --git a/proto/radv/packets.c b/proto/radv/packets.c
index fbe02060..7d54a827 100644
--- a/proto/radv/packets.c
+++ b/proto/radv/packets.c
@@ -85,7 +85,6 @@ radv_prepare_route(struct radv_iface *ifa, struct radv_route *rt,
 		   char **buf, char *bufend)
 {
   struct radv_proto *p = ifa->ra;
-  struct radv_config *cf = (void *) p->p.cf;
   u8 px_blocks = (rt->n.pxlen + 63) / 64;
   u8 opt_len = 8 * (1 + px_blocks);
 
@@ -96,22 +95,26 @@ radv_prepare_route(struct radv_iface *ifa, struct radv_route *rt,
     return -1;
   }
 
+  uint preference = rt->preference_set ? rt->preference : ifa->cf->route_preference;
+  uint lifetime = rt->lifetime_set ? rt->lifetime : ifa->cf->route_lifetime;
+  uint valid = rt->valid && p->valid && (p->active || !ifa->cf->route_lifetime_sensitive);
+
   struct radv_opt_route *opt = (void *) *buf;
   *buf += opt_len;
   opt->type = OPT_ROUTE;
   opt->length = 1 + px_blocks;
   opt->pxlen = rt->n.pxlen;
-  opt->flags = rt->preference;
-
-  if (p->valid && (p->active || !cf->route_lifetime_sensitive) && rt->alive)
-    opt->lifetime = htonl(rt->lifetime_set ? rt->lifetime : cf->route_lifetime);
-  else
-    opt->lifetime = 0;
+  opt->flags = preference;
+  opt->lifetime = valid ? htonl(lifetime) : 0;
 
   /* Copy the relevant part of the prefix */
   ip6_addr px_addr = ip6_hton(rt->n.prefix);
   memcpy(opt->prefix, &px_addr, 8 * px_blocks);
 
+  /* Keeping track of first linger timeout */
+  if (!rt->valid)
+    ifa->valid_time = MIN(ifa->valid_time, rt->changed + ifa->cf->route_linger_time);
+
   return 0;
 }
 
@@ -278,6 +281,10 @@ radv_prepare_prefix(struct radv_iface *ifa, struct radv_prefix *prefix,
   ipa_hton(op->prefix);
   *buf += sizeof(*op);
 
+  /* Keeping track of first linger timeout */
+  if (!prefix->valid)
+    ifa->valid_time = MIN(ifa->valid_time, prefix->changed + ifa->cf->prefix_linger_time);
+
   return 0;
 }
 
@@ -316,10 +323,17 @@ radv_prepare_ra(struct radv_iface *ifa)
     buf += sizeof (*om);
   }
 
-  struct radv_prefix *prefix;
-  WALK_LIST(prefix, ifa->prefixes)
+  /* Keeping track of first linger timeout */
+  ifa->valid_time = TIME_INFINITY;
+
+  struct radv_prefix *px;
+  WALK_LIST(px, ifa->prefixes)
   {
-    if (radv_prepare_prefix(ifa, prefix, &buf, bufend) < 0)
+    /* Skip invalid prefixes that are past linger timeout but still not pruned */
+    if (!px->valid && (px->changed + ic->prefix_linger_time <= now))
+	continue;
+
+    if (radv_prepare_prefix(ifa, px, &buf, bufend) < 0)
       goto done;
   }
 
@@ -339,9 +353,15 @@ radv_prepare_ra(struct radv_iface *ifa)
 
   if (p->fib_up)
   {
-    FIB_WALK(&p->routes, rt)
+    FIB_WALK(&p->routes, n)
     {
-      if (radv_prepare_route(ifa, (struct radv_route *) rt, &buf, bufend) < 0)
+      struct radv_route *rt = (void *) n;
+
+      /* Skip invalid routes that are past linger timeout but still not pruned */
+      if (!rt->valid && (rt->changed + ic->route_linger_time <= now))
+	continue;
+
+      if (radv_prepare_route(ifa, rt, &buf, bufend) < 0)
 	goto done;
     }
     FIB_WALK_END;
diff --git a/proto/radv/radv.c b/proto/radv/radv.c
index d799dab5..d53b65e3 100644
--- a/proto/radv/radv.c
+++ b/proto/radv/radv.c
@@ -55,6 +55,9 @@ radv_timer(timer *tm)
 
   RADV_TRACE(D_EVENTS, "Timer fired on %s", ifa->iface->name);
 
+  if (ifa->valid_time <= now)
+    radv_invalidate(ifa);
+
   if (ifa->prune_time <= now)
     radv_prune_prefixes(ifa);
 
@@ -117,7 +120,6 @@ static void
 radv_prepare_prefixes(struct radv_iface *ifa)
 {
   struct radv_proto *p = ifa->ra;
-  struct radv_iface_config *cf = ifa->cf;
   struct radv_prefix *pfx, *next;
 
   /* First mark all the prefixes as unused */
@@ -157,20 +159,21 @@ radv_prepare_prefixes(struct radv_iface *ifa)
      * Update the information (it may have changed, or even bring a prefix back
      * to life).
      */
-    existing->alive = 1;
+    existing->valid = 1;
+    existing->changed = now;
     existing->mark = 1;
     existing->cf = pc;
   }
 
   WALK_LIST_DELSAFE(pfx, next, ifa->prefixes)
   {
-    if (pfx->alive && !pfx->mark)
+    if (pfx->valid && !pfx->mark)
     {
       RADV_TRACE(D_EVENTS, "Invalidating prefix %I/$d on %s",
 		 pfx->prefix, pfx->len, ifa->iface->name);
 
-      pfx->alive = 0;
-      pfx->expires = now + cf->linger_time;
+      pfx->valid = 0;
+      pfx->changed = now;
       pfx->cf = &dead_prefix;
     }
   }
@@ -181,30 +184,28 @@ radv_prune_prefixes(struct radv_iface *ifa)
 {
   struct radv_proto *p = ifa->ra;
   bird_clock_t next = TIME_INFINITY;
-  int changed = 0;
+  bird_clock_t expires = 0;
 
   struct radv_prefix *px, *pxn;
   WALK_LIST_DELSAFE(px, pxn, ifa->prefixes)
   {
-    if (!px->alive)
+    if (!px->valid)
     {
-      if (px->expires <= now)
+      expires = px->changed + ifa->cf->prefix_linger_time;
+
+      if (expires <= now)
       {
 	RADV_TRACE(D_EVENTS, "Removing prefix %I/%d on %s",
 		   px->prefix, px->len, ifa->iface->name);
 
 	rem_node(NODE px);
 	mb_free(px);
-	changed = 1;
       }
       else
-	next = MIN(next, px->expires);
+	next = MIN(next, expires);
     }
   }
 
-  if (changed)
-    radv_invalidate(ifa);
-
   ifa->prune_time = next;
 }
 
@@ -424,6 +425,7 @@ radv_rt_notify(struct proto *P, rtable *tbl UNUSED, net *n, rte *new, rte *old U
   struct radv_proto *p = (struct radv_proto *) P;
   struct radv_config *cf = (struct radv_config *) (P->cf);
   struct radv_route *rt;
+  eattr *ea;
 
   if (radv_net_match_trigger(cf, n))
   {
@@ -455,9 +457,14 @@ radv_rt_notify(struct proto *P, rtable *tbl UNUSED, net *n, rte *new, rte *old U
   if (new)
   {
     /* Update */
-    uint preference = ea_get_int(attrs, EA_RA_PREFERENCE, RA_PREF_MEDIUM);
-    uint lifetime = ea_get_int(attrs, EA_RA_LIFETIME, 0);
-    uint lifetime_set = !!ea_find(attrs, EA_RA_LIFETIME);
+
+    ea = ea_find(attrs, EA_RA_PREFERENCE);
+    uint preference = ea ? ea->u.data : RA_PREF_MEDIUM;
+    uint preference_set = !!ea;
+
+    ea = ea_find(attrs, EA_RA_LIFETIME);
+    uint lifetime = ea ? ea->u.data : 0;
+    uint lifetime_set = !!ea;
 
     if ((preference != RA_PREF_LOW) &&
 	(preference != RA_PREF_MEDIUM) &&
@@ -466,6 +473,7 @@ radv_rt_notify(struct proto *P, rtable *tbl UNUSED, net *n, rte *new, rte *old U
       log(L_WARN "%s: Invalid ra_preference value %u on route %I/%d",
 	  p->p.name, preference, n->n.prefix, n->n.pxlen);
       preference = RA_PREF_MEDIUM;
+      preference_set = 1;
       lifetime = 0;
       lifetime_set = 1;
     }
@@ -473,8 +481,9 @@ radv_rt_notify(struct proto *P, rtable *tbl UNUSED, net *n, rte *new, rte *old U
     rt = fib_get(&p->routes, &n->n.prefix, n->n.pxlen);
 
     /* Ignore update if nothing changed */
-    if (rt->alive &&
+    if (rt->valid &&
 	(rt->preference == preference) &&
+	(rt->preference_set == preference_set) &&
 	(rt->lifetime == lifetime) &&
 	(rt->lifetime_set == lifetime_set))
       return;
@@ -482,8 +491,10 @@ radv_rt_notify(struct proto *P, rtable *tbl UNUSED, net *n, rte *new, rte *old U
     if (p->routes.entries == 18)
       log(L_WARN "%s: More than 17 routes exported to RAdv", p->p.name);
 
-    rt->alive = 1;
+    rt->valid = 1;
+    rt->changed = now;
     rt->preference = preference;
+    rt->preference_set = preference_set;
     rt->lifetime = lifetime;
     rt->lifetime_set = lifetime_set;
   }
@@ -492,13 +503,16 @@ radv_rt_notify(struct proto *P, rtable *tbl UNUSED, net *n, rte *new, rte *old U
     /* Withdraw */
     rt = fib_find(&p->routes, &n->n.prefix, n->n.pxlen);
 
-    if (!rt || !rt->alive)
+    if (!rt || !rt->valid)
       return;
 
     /* Invalidate the route */
-    rt->alive = 0;
-    rt->expires = now + cf->route_linger_time;
-    p->prune_time = MIN(p->prune_time, rt->expires);
+    rt->valid = 0;
+    rt->changed = now;
+
+    /* Invalidated route will be pruned eventually */
+    bird_clock_t expires = rt->changed + cf->max_linger_time;
+    p->prune_time = MIN(p->prune_time, expires);
   }
 
   radv_iface_notify_all(p, RA_EV_CHANGE);
@@ -511,8 +525,9 @@ radv_rt_notify(struct proto *P, rtable *tbl UNUSED, net *n, rte *new, rte *old U
 static void
 radv_prune_routes(struct radv_proto *p)
 {
+  struct radv_config *cf = (struct radv_config *) (p->p.cf);
   bird_clock_t next = TIME_INFINITY;
-  int changed = 0;
+  bird_clock_t expires = 0;
 
   /* Should not happen */
   if (!p->fib_up)
@@ -526,29 +541,23 @@ again:
   {
     struct radv_route *rt = (void *) node;
 
-    if (!rt->alive)
+    if (!rt->valid)
     {
+      expires = rt->changed + cf->max_linger_time;
+
       /* Delete expired nodes */
-      if (rt->expires <= now)
+      if (expires <= now)
       {
 	FIB_ITERATE_PUT(&fit, node);
 	fib_delete(&p->routes, node);
-	changed = 1;
 	goto again;
       }
       else
-	next = MIN(next, rt->expires);
+	next = MIN(next, expires);
     }
   }
   FIB_ITERATE_END(node);
 
-  if (changed)
-  {
-    struct radv_iface *ifa;
-    WALK_LIST(ifa, p->iface_list)
-      radv_invalidate(ifa);
-  }
-
   p->prune_time = next;
 }
 
diff --git a/proto/radv/radv.h b/proto/radv/radv.h
index 3bc18249..ab081397 100644
--- a/proto/radv/radv.h
+++ b/proto/radv/radv.h
@@ -35,7 +35,6 @@
 #define DEFAULT_MAX_RA_INT 600
 #define DEFAULT_MIN_DELAY 3
 #define DEFAULT_CURRENT_HOP_LIMIT 64
-#define DEFAULT_LINGER_TIME 300
 
 #define DEFAULT_VALID_LIFETIME 86400
 #define DEFAULT_PREFERRED_LIFETIME 14400
@@ -55,9 +54,7 @@ struct radv_config
   u8 trigger_pxlen;		/* Pxlen of a trigger route, if defined */
   u8 trigger_valid;		/* Whether a trigger route is defined */
   u8 propagate_routes;		/* Do we propagate more specific routes (RFC 4191)? */
-  u32 route_lifetime;		/* Lifetime for the RFC 4191 routes */
-  u32 route_lifetime_sensitive; /* Whether route_lifetime depends on trigger */
-  u32 route_linger_time;	/* For how long we advertise dead routes with lifetime = 0 */
+  u32 max_linger_time;		/* Maximum of interface route_linger_time */
 };
 
 struct radv_iface_config
@@ -71,8 +68,8 @@ struct radv_iface_config
   u32 max_ra_int;
   u32 min_delay;
 
-  u32 linger_time;		/* How long a dead prefix should still be advertised with 0
-				   lifetime */
+  u32 prefix_linger_time;	/* How long we advertise dead prefixes with lifetime 0 */
+  u32 route_linger_time;	/* How long we advertise dead routes with lifetime 0 */
 
   u8 rdnss_local;		/* Global list is not used for RDNSS */
   u8 dnssl_local;		/* Global list is not used for DNSSL */
@@ -84,8 +81,11 @@ struct radv_iface_config
   u32 retrans_timer;
   u32 current_hop_limit;
   u32 default_lifetime;
+  u32 route_lifetime;		/* Lifetime for the RFC 4191 routes */
   u8 default_lifetime_sensitive; /* Whether default_lifetime depends on trigger */
-  u8 default_preference;	 /* Default Router Preference (RFC 4191) */
+  u8 route_lifetime_sensitive;	/* Whether route_lifetime depends on trigger */
+  u8 default_preference;	/* Default Router Preference (RFC 4191) */
+  u8 route_preference;		/* Specific Route Preference (RFC 4191) */
 };
 
 struct radv_prefix_config
@@ -132,10 +132,11 @@ struct radv_route
 {
   struct fib_node n;
   u32 lifetime;			/* Lifetime from an attribute */
-  u8 lifetime_set;		/* Is the lifetime set by an attribute? */
+  u8 lifetime_set;		/* Whether lifetime is defined */
   u8 preference;		/* Preference of the route, RA_PREF_* */
-  u8 alive;
-  bird_clock_t expires;		/* Time to remove when !alive */
+  u8 preference_set;		/* Whether preference is defined */
+  u8 valid;			/* Whethe route is valid or withdrawn */
+  bird_clock_t changed;		/* Last time when the route changed */
 };
 
 struct radv_proto
@@ -154,11 +155,10 @@ struct radv_prefix		/* One prefix we advertise */
   node n;
   ip_addr prefix;
   u8 len;
-  u8 alive;			/* Is the prefix alive? If not, we advertise it
+  u8 valid;			/* Is the prefix valid? If not, we advertise it
 				   with 0 lifetime, so clients stop using it */
   u8 mark;			/* A temporary mark for processing */
-  bird_clock_t expires;		/* The time when we drop this prefix from
-				   advertising. It is valid only if !alive. */
+  bird_clock_t changed;		/* Last time when the prefix changed */
   struct radv_prefix_config *cf; /* The config tied to this prefix */
 };
 
@@ -172,6 +172,7 @@ struct radv_iface
   struct pool *pool;		/* A pool for interface-specific things */
   list prefixes;		/* The prefixes we advertise (struct radv_prefix) */
   bird_clock_t prune_time;	/* Next time of prefix list pruning */
+  bird_clock_t valid_time;	/* Cached packet is valid until first linger timeout */
 
   timer *timer;
   struct object_lock *lock;

From 18352188ed7415ddb3aa7b0d7a2fc16d57a2d5d2 Mon Sep 17 00:00:00 2001
From: Michal 'vorner' Vaner <michal.vaner@nic.cz>
Date: Tue, 19 Sep 2017 17:23:31 +0200
Subject: [PATCH 030/198] RAdv: Documentation for more specific routes

---
 doc/bird.sgml | 94 +++++++++++++++++++++++++++++++++++++++++++++------
 1 file changed, 83 insertions(+), 11 deletions(-)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index c549e3c3..129e9c08 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -3336,7 +3336,8 @@ time intervals or as an answer to a request) advertisement packets to connected
 networks. These packets contain basic information about a local network (e.g. a
 list of network prefixes), which allows network hosts to autoconfigure network
 addresses and choose a default route. BIRD implements router behavior as defined
-in <rfc id="4861"> and also the DNS extensions from <rfc id="6106">.
+in <rfc id="4861">, router preferences and specific routes (<rfc id="4191">),
+and DNS extensions (<rfc id="6106">).
 
 <sect1>Configuration
 <label id="radv-config">
@@ -3373,7 +3374,7 @@ definitions, prefix definitions and DNS definitions:
 	definitions may also be interface-specific when used inside interface
 	options. By default, interface uses both global and interface-specific
 	options, but that can be changed by <cf/rdnss local/ option.
-dsc-iface
+
 	<tag><label id="radv-dnssl">dnssl { <m/options/ }</tag>
 	DNSSL definitions allow to specify a list of advertised DNS search
 	domains together with their options. Like <cf/rdnss/ above, multiple
@@ -3398,6 +3399,18 @@ dsc-iface
 	default router. <cf/preferred lifetime/ and <cf/valid lifetime/ could
 	also be configured as <cf/sensitive/ for a prefix, which would cause
 	autoconfigured IPs to be deprecated or even removed.
+
+	<tag><label id="radv-propagate-routes">propagate routes <m/switch/</tag>
+	This option controls propagation of more specific routes, as defined in
+	<rfc id="4191">. If enabled, all routes exported to the RAdv protocol,
+	with the exception of the trigger prefix, are added to advertisments as
+	additional options. The lifetime and preference of advertised routes can
+	be set individually by <cf/ra_lifetime/ and <cf/ra_preference/ route
+	attributes, or per interface by <cf/route lifetime/ and
+	<cf/route preference/ options. Default: disabled.
+
+	Note that the RFC discourages from sending more than 17 routes and
+	recommends the routes to be configured manually.
 </descrip>
 
 <p>Interface specific options:
@@ -3444,21 +3457,45 @@ dsc-iface
 	hosts. Valid values are 0-255, 0 means unspecified. Default: 64
 
 	<tag><label id="radv-iface-default-lifetime">default lifetime <m/expr/ [sensitive <m/switch/]</tag>
-	This option specifies the time (in seconds) how long (after the receipt
+	This option specifies the time (in seconds) how long (since the receipt
 	of RA) hosts may use the router as a default router. 0 means do not use
 	as a default router. For <cf/sensitive/ option, see <ref id="radv-trigger" name="trigger">.
 	Default: 3 * <cf/max ra	interval/, <cf/sensitive/ yes.
 
-	<tag><label id="radv-iface-linger-time">linger time <m/expr/</tag>
-	When a prefix disappears, it is advertised for some time with 0
-	lifetime, to inform clients the prefix is no longer usable. This option
-	sets the time for how long it is advertised (in seconds). Maximum is
-	3600, 0 means disabled. Default: 300.
-
-	<tag><label id="radv-iface-default-preference-low">default preference low|medium|high</tag>
+	<tag><label id="radv-iface-default-preference">default preference low|medium|high</tag>
 	This option specifies the Default Router Preference value to advertise
 	to hosts. Default: medium.
 
+	<tag><label id="radv-iface-route-lifetime">route lifetime <m/expr/ [sensitive <m/switch/]</tag>
+	This option specifies the default value of advertised lifetime for
+	specific routes; i.e., the time (in seconds) for how long (since the
+	receipt of RA) hosts should consider these routes valid. A special value
+	0xffffffff represents infinity. The lifetime can be overriden on a per
+	route basis by the <ref id="rta-ra-lifetime" name="ra_lifetime"> route
+	attribute. Default: 3 * <cf/max ra interval/, <cf/sensitive/ no.
+
+	For the <cf/sensitive/ option, see <ref id="radv-trigger" name="trigger">.
+	If <cf/sensitive/ is enabled, even the routes with the <cf/ra_lifetime/
+	attribute become sensitive to the trigger.
+
+	<tag><label id="radv-iface-route-preference">route preference low|medium|high</tag>
+	This option specifies the default value of advertised route preference
+	for specific routes. The value can be overriden on a per route basis by
+	the <ref id="rta-ra-preference" name="ra_preference"> route attribute.
+	Default: medium.
+
+	<tag><label id="radv-prefix-linger-time">prefix linger time <m/expr/</tag>
+	When a prefix or a route disappears, it is advertised for some time with
+	zero lifetime, to inform clients it is no longer valid. This option
+	specifies the time (in seconds) for how long prefixes are advertised
+	that way. Default: 3 * <cf/max ra interval/.
+
+	<tag><label id="radv-route-linger-time">route linger time <m/expr/</tag>
+	When a prefix or a route disappears, it is advertised for some time with
+	zero lifetime, to inform clients it is no longer valid. This option
+	specifies the time (in seconds) for how long routes are advertised
+	that way. Default: 3 * <cf/max ra interval/.
+
 	<tag><label id="radv-iface-rdnss-local">rdnss local <m/switch/</tag>
 	Use only local (interface-specific) RDNSS definitions for this
 	interface. Otherwise, both global and local definitions are used. Could
@@ -3505,7 +3542,6 @@ dsc-iface
 	<cf/sensitive/ no.
 </descrip>
 
-
 <p>RDNSS specific options:
 
 <descrip>
@@ -3538,12 +3574,48 @@ dsc-iface
 	RDNSS <cf/lifetime/ option above. Default: 3 * <cf/max ra interval/.
 </descrip>
 
+<sect1>Attributes
+<label id="radv-attr">
+
+<p>RAdv defines two route attributes:
+
+<descrip>
+	<tag><label id="rta-ra-preference">enum ra_preference/</tag>
+	The preference of the route. The value can be <it/RA_PREF_LOW/,
+	<it/RA_PREF_MEDIUM/ or <it/RA_PREF_HIGH/. If the attribute is not set,
+	the <ref id="radv-iface-route-preference" name="route preference">
+	option is used.
+
+	<tag><label id="rta-ra-lifetime">int ra_lifetime/</tag>
+	The advertised lifetime of the route, in seconds. The special value of
+	0xffffffff represents infinity. If the attribute is not set, the
+	<ref id="radv-iface-route-lifetime" name="route lifetime">
+	option is used.
+</descrip>
 
 <sect1>Example
 <label id="radv-exam">
 
 <p><code>
+table radv_routes;			# Manually configured routes go here
+
+protocol static {
+	table radv_routes;
+
+	route 2001:0DB8:4000::/48 unreachable;
+	route 2001:0DB8:4010::/48 unreachable;
+
+	route 2001:0DB8:4020::/48 unreachable {
+		ra_preference = RA_PREF_HIGH;
+		ra_lifetime = 3600;
+	};
+}
+
 protocol radv {
+	propagate routes yes;		# Propagate the routes from the radv_routes table
+	table radv_routes;
+	export all;
+
 	interface "eth2" {
 		max ra interval 5;	# Fast failover with more routers
 		managed yes;		# Using DHCPv6 on eth2

From 7e8d6116df68e96ef4abd64f6b9cbe7406c5985d Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 10 Oct 2017 13:27:03 +0200
Subject: [PATCH 031/198] Doc: Fix reference

---
 doc/bird.sgml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index 129e9c08..d9200f26 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -1767,7 +1767,7 @@ protocol bfd [&lt;name&gt;] {
 	computation.
 
 	<tag>password "<M>text</M>"</tag>
-	Specifies a password used for authentication. See <ref id="dsc-pass"
+	Specifies a password used for authentication. See <ref id="proto-pass"
 	name="password"> common option for detailed description. Note that
 	password option <cf/algorithm/ is not available in BFD protocol. The
 	algorithm is selected by <cf/authentication/ option for all passwords.

From d9573a40ecaf9758690e4482782bebaf7847ba9b Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 10 Oct 2017 15:06:39 +0200
Subject: [PATCH 032/198] Doc: Update sgml2* tools

The old ones do not work on current Debian.
---
 doc/sgml2html  | 8 +++++---
 doc/sgml2latex | 8 +++++---
 doc/sgml2txt   | 8 +++++---
 3 files changed, 15 insertions(+), 9 deletions(-)

diff --git a/doc/sgml2html b/doc/sgml2html
index 774a03d5..a5bbee9e 100755
--- a/doc/sgml2html
+++ b/doc/sgml2html
@@ -17,8 +17,10 @@ use strict;
 
 use vars qw($prefix $DataDir $BinDir $progs);
 
+use FindBin;
+
 $prefix = "/usr";
-$DataDir = "sbase";
+$DataDir = "$FindBin::Bin/sbase";
 $BinDir = "/usr/bin";
 
 use lib "/usr/share/linuxdoc-tools";
@@ -32,9 +34,9 @@ $progs = {
   "GROFFMACRO" => "-ms",
   "AWK" => "/usr/share/linuxdoc-tools/awkwhich"
 };
-$ENV{"SGML_CATALOG_FILES"} = "sbase/dtd/catalog";
+$ENV{"SGML_CATALOG_FILES"} = "$DataDir/dtd/catalog";
 
-require "./LinuxDocTools.pm";
+require "$FindBin::Bin/LinuxDocTools.pm";
 &LinuxDocTools::init;
 
 my @FileList = LinuxDocTools::process_options ("html", @ARGV);
diff --git a/doc/sgml2latex b/doc/sgml2latex
index 27aae4c8..02b60d94 100755
--- a/doc/sgml2latex
+++ b/doc/sgml2latex
@@ -17,8 +17,10 @@ use strict;
 
 use vars qw($prefix $DataDir $BinDir $progs);
 
+use FindBin;
+
 $prefix = "/usr";
-$DataDir = "sbase";
+$DataDir = "$FindBin::Bin/sbase";
 $BinDir = "/usr/bin";
 
 use lib "/usr/share/linuxdoc-tools";
@@ -32,9 +34,9 @@ $progs = {
   "GROFFMACRO" => "-ms",
   "AWK" => "/usr/share/linuxdoc-tools/awkwhich"
 };
-$ENV{"SGML_CATALOG_FILES"} = "sbase/dtd/catalog";
+$ENV{"SGML_CATALOG_FILES"} = "$DataDir/dtd/catalog";
 
-require "./LinuxDocTools.pm";
+require "$FindBin::Bin/LinuxDocTools.pm";
 &LinuxDocTools::init;
 
 my @FileList = LinuxDocTools::process_options ("latex", @ARGV);
diff --git a/doc/sgml2txt b/doc/sgml2txt
index 90dc4855..dfc017de 100755
--- a/doc/sgml2txt
+++ b/doc/sgml2txt
@@ -17,8 +17,10 @@ use strict;
 
 use vars qw($prefix $DataDir $BinDir $progs);
 
+use FindBin;
+
 $prefix = "/usr";
-$DataDir = "sbase";
+$DataDir = "$FindBin::Bin/sbase";
 $BinDir = "/usr/bin";
 
 use lib "/usr/share/linuxdoc-tools";
@@ -32,9 +34,9 @@ $progs = {
   "GROFFMACRO" => "-ms",
   "AWK" => "/usr/share/linuxdoc-tools/awkwhich"
 };
-$ENV{"SGML_CATALOG_FILES"} = "sbase/dtd/catalog";
+$ENV{"SGML_CATALOG_FILES"} = "$DataDir/dtd/catalog";
 
-require "./LinuxDocTools.pm";
+require "$FindBin::Bin/LinuxDocTools.pm";
 &LinuxDocTools::init;
 
 my @FileList = LinuxDocTools::process_options ("txt", @ARGV);

From d3f4f92b0ece0ce4031087a25735e6cbf0d741e2 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Mon, 9 Oct 2017 01:16:29 +0200
Subject: [PATCH 033/198] OSPF: Support of address families in OSPFv3

OSPFv3-AF can handle multiple topologies of diferent address families
(IPv4, IPv6, both unicast and multicast) using separate instances
distinguished by instance ID ranges.
---
 proto/ospf/config.Y   |  89 +++++++++++++++++++++++++----
 proto/ospf/dbdes.c    |   2 +-
 proto/ospf/hello.c    |  23 ++++----
 proto/ospf/iface.c    |   9 ++-
 proto/ospf/lsalib.c   |  10 ++--
 proto/ospf/lsalib.h   |   4 +-
 proto/ospf/ospf.c     |  62 ++++++++++++---------
 proto/ospf/ospf.h     | 127 +++++++++++++++++++++++++++---------------
 proto/ospf/rt.c       |  28 ++++++----
 proto/ospf/topology.c |  44 +++++++++------
 10 files changed, 262 insertions(+), 136 deletions(-)

diff --git a/proto/ospf/config.Y b/proto/ospf/config.Y
index 1f379bf4..85149ea1 100644
--- a/proto/ospf/config.Y
+++ b/proto/ospf/config.Y
@@ -78,18 +78,66 @@ static void
 ospf_proto_finish(void)
 {
   struct ospf_config *cf = OSPF_CFG;
-
-  if (EMPTY_LIST(cf->area_list))
-    cf_error( "No configured areas in OSPF");
+  struct ospf_area_config *ac;
+  struct ospf_iface_patt *ic;
 
   /* Define default channel */
   if (EMPTY_LIST(this_proto->channels))
+  {
+    this_proto->net_type = ospf_cfg_is_v2() ? NET_IP4 : NET_IP6;
     channel_config_new(NULL, this_proto->net_type, this_proto);
+  }
+
+  /* Propagate global instance ID to interfaces */
+  if (cf->instance_id_set)
+  {
+    WALK_LIST(ac, cf->area_list)
+      WALK_LIST(ic, ac->patt_list)
+	if (!ic->instance_id_set)
+	{ ic->instance_id = cf->instance_id; ic->instance_id_set = 1; }
+
+    WALK_LIST(ic, cf->vlink_list)
+      if (!ic->instance_id_set)
+      { ic->instance_id = cf->instance_id; ic->instance_id_set = 1; }
+  }
+
+  if (ospf_cfg_is_v3())
+  {
+    uint ipv4 = (this_proto->net_type == NET_IP4);
+    uint base = (ipv4 ? 64 : 0) + (cf->af_mc ? 32 : 0);
+
+    /* RFC 5838 - OSPFv3-AF */
+    if (cf->af_ext)
+    {
+      /* RFC 5838 2.1 - instance IDs based on AFs */
+      WALK_LIST(ac, cf->area_list)
+	WALK_LIST(ic, ac->patt_list)
+	{
+	  if (!ic->instance_id_set)
+	    ic->instance_id = base;
+	  else if (ic->instance_id >= 128)
+	    log(L_WARN "Instance ID %d from unassigned/private range", ic->instance_id);
+	  else if ((ic->instance_id < base) || (ic->instance_id >= (base + 32)))
+	    cf_error("Instance ID %d invalid for given channel type", ic->instance_id);
+	}
+
+      /* RFC 5838 2.8 - vlinks limited to IPv6 unicast */
+      if ((ipv4 || cf->af_mc) && !EMPTY_LIST(cf->vlink_list))
+	cf_error("Vlinks not supported in AFs other than IPv6 unicast");
+    }
+    else
+    {
+      if (ipv4 || cf->af_mc)
+	cf_error("Different channel type");
+    }
+  }
+
+  if (EMPTY_LIST(cf->area_list))
+    cf_error("No configured areas in OSPF");
 
   int areano = 0;
   int backbone = 0;
   int nssa = 0;
-  struct ospf_area_config *ac;
   WALK_LIST(ac, cf->area_list)
   {
     areano++;
@@ -148,10 +196,11 @@ CF_KEYWORDS(ELIGIBLE, POLL, NETWORKS, HIDDEN, VIRTUAL, CHECK, LINK, ONLY, BFD)
 CF_KEYWORDS(RX, BUFFER, LARGE, NORMAL, STUBNET, HIDDEN, SUMMARY, TAG, EXTERNAL)
 CF_KEYWORDS(WAIT, DELAY, LSADB, ECMP, LIMIT, WEIGHT, NSSA, TRANSLATOR, STABILITY)
 CF_KEYWORDS(GLOBAL, LSID, ROUTER, SELF, INSTANCE, REAL, NETMASK, TX, PRIORITY, LENGTH)
-CF_KEYWORDS(SECONDARY, MERGE, LSA, SUPPRESSION)
+CF_KEYWORDS(SECONDARY, MERGE, LSA, SUPPRESSION, MULTICAST)
 
 %type <ld> lsadb_args
-%type <i> ospf_variant nbma_eligible
+%type <i> ospf_variant ospf_af_mc nbma_eligible
+%type <cc> ospf_channel_start ospf_channel
 
 CF_GRAMMAR
 
@@ -166,12 +215,13 @@ ospf_variant:
 ospf_proto_start: proto_start ospf_variant
 {
   this_proto = proto_config_new(&proto_ospf, $1);
-  this_proto->net_type = $2 ? NET_IP4 : NET_IP6;
+  this_proto->net_type = $2 ? NET_IP4 : 0;
 
   init_list(&OSPF_CFG->area_list);
   init_list(&OSPF_CFG->vlink_list);
   OSPF_CFG->tick = OSPF_DEFAULT_TICK;
   OSPF_CFG->ospf2 = $2;
+  OSPF_CFG->af_ext = !$2;
 };
 
 ospf_proto:
@@ -179,16 +229,33 @@ ospf_proto:
  | ospf_proto ospf_proto_item ';'
  ;
 
+ospf_af_mc:
+             { $$ = 0; }
+ | MULTICAST { $$ = 1; }
+ ;
+
+/* We redefine proto_channel to add multicast flag */
+ospf_channel_start: net_type ospf_af_mc
+{
+  $$ = this_channel = channel_config_new(NULL, $1, this_proto);
+
+  /* Save the multicast flag */
+  if (this_channel == proto_cf_main_channel(this_proto))
+    OSPF_CFG->af_mc = $2;
+};
+
+ospf_channel: ospf_channel_start channel_opt_list channel_end;
+
 ospf_proto_item:
    proto_item
- | proto_channel
+ | ospf_channel { this_proto->net_type = $1->net_type; }
  | RFC1583COMPAT bool { OSPF_CFG->rfc1583 = $2; }
  | STUB ROUTER bool { OSPF_CFG->stub_router = $3; }
  | ECMP bool { OSPF_CFG->ecmp = $2 ? OSPF_DEFAULT_ECMP_LIMIT : 0; }
  | ECMP bool LIMIT expr { OSPF_CFG->ecmp = $2 ? $4 : 0; }
  | MERGE EXTERNAL bool { OSPF_CFG->merge_external = $3; }
  | TICK expr { OSPF_CFG->tick = $2; if($2 <= 0) cf_error("Tick must be greater than zero"); }
- | INSTANCE ID expr { OSPF_CFG->instance_id = $3; if ($3 > 255) cf_error("Instance ID must be in range 0-255"); }
+ | INSTANCE ID expr { OSPF_CFG->instance_id = $3; OSPF_CFG->instance_id_set = 1; if ($3 > 255) cf_error("Instance ID must be in range 0-255"); }
  | ospf_area
  ;
 
@@ -293,7 +360,6 @@ ospf_vlink_start: VIRTUAL LINK idval
   OSPF_PATT->inftransdelay = INFTRANSDELAY_D;
   OSPF_PATT->deadc = DEADC_D;
   OSPF_PATT->type = OSPF_IT_VLINK;
-  OSPF_PATT->instance_id = OSPF_CFG->instance_id;
   init_list(&OSPF_PATT->nbma_list);
   reset_passwords();
  }
@@ -393,7 +459,6 @@ ospf_iface_start:
   OSPF_PATT->priority = PRIORITY_D;
   OSPF_PATT->deadc = DEADC_D;
   OSPF_PATT->type = OSPF_IT_UNDEF;
-  OSPF_PATT->instance_id = OSPF_CFG->instance_id;
   init_list(&OSPF_PATT->nbma_list);
   OSPF_PATT->ptp_netmask = 2; /* not specified */
   OSPF_PATT->tx_tos = IP_PREC_INTERNET_CONTROL;
@@ -404,7 +469,7 @@ ospf_iface_start:
 
 ospf_instance_id:
    /* empty */
- | INSTANCE expr { OSPF_PATT->instance_id = $2; if ($2 > 255) cf_error("Instance ID must be in range 0-255"); }
+ | INSTANCE expr { OSPF_PATT->instance_id = $2; OSPF_PATT->instance_id_set = 1; if ($2 > 255) cf_error("Instance ID must be in range 0-255"); }
  ;
 
 ospf_iface_patt_list:
diff --git a/proto/ospf/dbdes.c b/proto/ospf/dbdes.c
index 195b03ad..a4452cc8 100644
--- a/proto/ospf/dbdes.c
+++ b/proto/ospf/dbdes.c
@@ -356,7 +356,7 @@ ospf_receive_dbdes(struct ospf_packet *pkt, struct ospf_iface *ifa,
       LOG_PKT_WARN("MTU mismatch with nbr %R on %s (remote %d, local %d)",
 		   n->rid, ifa->ifname, rcv_iface_mtu, ifa->iface->mtu);
 
-    if ((rcv_imms == DBDES_IMMS) &&
+    if (((rcv_imms & DBDES_IMMS) == DBDES_IMMS) &&
 	(n->rid > p->router_id) &&
 	(plen == ospf_dbdes_hdrlen(p)))
     {
diff --git a/proto/ospf/hello.c b/proto/ospf/hello.c
index 2c55155d..e706ea0f 100644
--- a/proto/ospf/hello.c
+++ b/proto/ospf/hello.c
@@ -32,10 +32,7 @@ struct ospf_hello3_packet
   struct ospf_packet hdr;
 
   u32 iface_id;
-  u8 priority;
-  u8 options3;
-  u8 options2;
-  u8 options;
+  u32 options;
   u16 helloint;
   u16 deadint;
   u32 dr;
@@ -91,10 +88,7 @@ ospf_send_hello(struct ospf_iface *ifa, int kind, struct ospf_neighbor *dirn)
     struct ospf_hello3_packet *ps = (void *) pkt;
 
     ps->iface_id = htonl(ifa->iface_id);
-    ps->priority = ifa->priority;
-    ps->options3 = ifa->oa->options >> 16;
-    ps->options2 = ifa->oa->options >> 8;
-    ps->options = ifa->oa->options;
+    ps->options = ntohl(ifa->oa->options | (ifa->priority << 24));
     ps->helloint = ntohs(ifa->helloint);
     ps->deadint = htons(ifa->deadint);
     ps->dr = htonl(ifa->drid);
@@ -190,7 +184,8 @@ ospf_receive_hello(struct ospf_packet *pkt, struct ospf_iface *ifa,
   struct ospf_proto *p = ifa->oa->po;
   const char *err_dsc = NULL;
   u32 rcv_iface_id, rcv_helloint, rcv_deadint, rcv_dr, rcv_bdr;
-  u8 rcv_options, rcv_priority;
+  uint rcv_options, rcv_priority;
+  uint loc_options = ifa->oa->options;
   u32 *neighbors;
   u32 neigh_count;
   uint plen, i, err_val = 0;
@@ -245,8 +240,8 @@ ospf_receive_hello(struct ospf_packet *pkt, struct ospf_iface *ifa,
     rcv_deadint = ntohs(ps->deadint);
     rcv_dr = ntohl(ps->dr);
     rcv_bdr = ntohl(ps->bdr);
-    rcv_options = ps->options;
-    rcv_priority = ps->priority;
+    rcv_options = ntohl(ps->options) & 0x00FFFFFF;
+    rcv_priority = ntohl(ps->options) >> 24;
 
     neighbors = ps->neighbors;
     neigh_count = (plen - sizeof(struct ospf_hello3_packet)) / sizeof(u32);
@@ -259,9 +254,13 @@ ospf_receive_hello(struct ospf_packet *pkt, struct ospf_iface *ifa,
     DROP("dead interval mismatch", rcv_deadint);
 
   /* Check whether bits E, N match */
-  if ((rcv_options ^ ifa->oa->options) & (OPT_E | OPT_N))
+  if ((rcv_options ^ loc_options) & (OPT_E | OPT_N))
     DROP("area type mismatch", rcv_options);
 
+  /* RFC 5838 2.4 - AF-bit check unless on IPv6 unicast */
+  if ((loc_options & OPT_AF) && !(loc_options & OPT_V6) && !(rcv_options & OPT_AF))
+    DROP("AF-bit mismatch", rcv_options);
+
   /* Check consistency of existing neighbor entry */
   if (n)
   {
diff --git a/proto/ospf/iface.c b/proto/ospf/iface.c
index 675cf76d..98d48aa1 100644
--- a/proto/ospf/iface.c
+++ b/proto/ospf/iface.c
@@ -1113,9 +1113,6 @@ ospf_ifa_notify3(struct proto *P, uint flags, struct ifa *a)
 {
   struct ospf_proto *p = (struct ospf_proto *) P;
 
-  if (a->prefix.type != NET_IP6)
-    return;
-
   if (a->flags & IA_SECONDARY)
     return;
 
@@ -1126,6 +1123,9 @@ ospf_ifa_notify3(struct proto *P, uint flags, struct ifa *a)
      other addresses are used for link-LSA. */
   if (a->scope == SCOPE_LINK)
   {
+    if (a->prefix.type != NET_IP6)
+      return;
+
     if (flags & IF_CHANGE_UP)
     {
       struct ospf_mip_walk s = { .iface = a->iface };
@@ -1143,6 +1143,9 @@ ospf_ifa_notify3(struct proto *P, uint flags, struct ifa *a)
   }
   else
   {
+    if (a->prefix.type != ospf_get_af(p))
+      return;
+
     struct ospf_iface *ifa;
     WALK_LIST(ifa, p->iface_list)
       if (ifa->iface == a->iface)
diff --git a/proto/ospf/lsalib.c b/proto/ospf/lsalib.c
index b88a114d..fbfd8d29 100644
--- a/proto/ospf/lsalib.c
+++ b/proto/ospf/lsalib.c
@@ -280,7 +280,7 @@ lsa_walk_rt(struct ospf_lsa_rt_walk *rt)
 
 
 void
-lsa_parse_sum_net(struct top_hash_entry *en, int ospf2, net_addr *net, u8 *pxopts, u32 *metric)
+lsa_parse_sum_net(struct top_hash_entry *en, int ospf2, int af, net_addr *net, u8 *pxopts, u32 *metric)
 {
   if (ospf2)
   {
@@ -292,7 +292,7 @@ lsa_parse_sum_net(struct top_hash_entry *en, int ospf2, net_addr *net, u8 *pxopt
   else
   {
     struct ospf_lsa_sum3_net *ls = en->lsa_body;
-    ospf_get_ipv6_prefix(ls->prefix, net, pxopts, NULL);
+    ospf3_get_prefix(ls->prefix, af, net, pxopts, NULL);
     *metric = ls->metric & LSA_METRIC_MASK;
   }
 }
@@ -317,7 +317,7 @@ lsa_parse_sum_rt(struct top_hash_entry *en, int ospf2, u32 *drid, u32 *metric, u
 }
 
 void
-lsa_parse_ext(struct top_hash_entry *en, int ospf2, struct ospf_lsa_ext_local *rt)
+lsa_parse_ext(struct top_hash_entry *en, int ospf2, int af, struct ospf_lsa_ext_local *rt)
 {
   if (ospf2)
   {
@@ -338,13 +338,13 @@ lsa_parse_ext(struct top_hash_entry *en, int ospf2, struct ospf_lsa_ext_local *r
   else
   {
     struct ospf_lsa_ext3 *ext = en->lsa_body;
-    u32 *buf = ospf_get_ipv6_prefix(ext->rest, &rt->net, &rt->pxopts, NULL);
+    u32 *buf = ospf3_get_prefix(ext->rest, af, &rt->net, &rt->pxopts, NULL);
     rt->metric = ext->metric & LSA_METRIC_MASK;
     rt->ebit = ext->metric & LSA_EXT3_EBIT;
 
     rt->fbit = ext->metric & LSA_EXT3_FBIT;
     if (rt->fbit)
-      buf = ospf_get_ipv6_addr(buf, &rt->fwaddr);
+      buf = ospf3_get_addr(buf, af, &rt->fwaddr);
     else
       rt->fwaddr = IPA_NONE;
 
diff --git a/proto/ospf/lsalib.h b/proto/ospf/lsalib.h
index c93f0295..0d477f58 100644
--- a/proto/ospf/lsalib.h
+++ b/proto/ospf/lsalib.h
@@ -55,9 +55,9 @@ u16 lsa_verify_checksum(const void *lsa_n, int lsa_len);
 int lsa_comp(struct ospf_lsa_header *l1, struct ospf_lsa_header *l2);
 void lsa_walk_rt_init(struct ospf_proto *po, struct top_hash_entry *act, struct ospf_lsa_rt_walk *rt);
 int lsa_walk_rt(struct ospf_lsa_rt_walk *rt);
-void lsa_parse_sum_net(struct top_hash_entry *en, int ospf2, net_addr *net, u8 *pxopts, u32 *metric);
+void lsa_parse_sum_net(struct top_hash_entry *en, int ospf2, int af, net_addr *net, u8 *pxopts, u32 *metric);
 void lsa_parse_sum_rt(struct top_hash_entry *en, int ospf2, u32 *drid, u32 *metric, u32 *options);
-void lsa_parse_ext(struct top_hash_entry *en, int ospf2, struct ospf_lsa_ext_local *rt);
+void lsa_parse_ext(struct top_hash_entry *en, int ospf2, int af, struct ospf_lsa_ext_local *rt);
 int lsa_validate(struct ospf_lsa_header *lsa, u32 lsa_type, int ospf2, void *body);
 
 #endif /* _BIRD_OSPF_LSALIB_H_ */
diff --git a/proto/ospf/ospf.c b/proto/ospf/ospf.c
index daf76ff2..7ce6698e 100644
--- a/proto/ospf/ospf.c
+++ b/proto/ospf/ospf.c
@@ -92,8 +92,10 @@
  * - RFC 2328 - main OSPFv2 standard
  * - RFC 5340 - main OSPFv3 standard
  * - RFC 3101 - OSPFv2 NSSA areas
- * - RFC 6549 - OSPFv2 multi-instance extensions
- * - RFC 6987 - OSPF stub router advertisement
+ * - RFC 5709 - OSPFv2 HMAC-SHA Cryptographic Authentication
+ * - RFC 5838 - OSPFv3 Support of Address Families
+ * - RFC 6549 - OSPFv2 Multi-Instance Extensions
+ * - RFC 6987 - OSPF Stub Router Advertisement
  */
 
 #include <stdlib.h>
@@ -115,9 +117,9 @@ add_area_nets(struct ospf_area *oa, struct ospf_area_config *ac)
   struct area_net_config *anc;
   struct area_net *an;
 
-  fib_init(&oa->net_fib,  p->p.pool, ospf_is_v2(p) ? NET_IP4 : NET_IP6,
+  fib_init(&oa->net_fib,  p->p.pool, ospf_get_af(p),
 	   sizeof(struct area_net), OFFSETOF(struct area_net, fn), 0, NULL);
-  fib_init(&oa->enet_fib, p->p.pool, ospf_is_v2(p) ? NET_IP4 : NET_IP6,
+  fib_init(&oa->enet_fib, p->p.pool, ospf_get_af(p),
 	   sizeof(struct area_net), OFFSETOF(struct area_net, fn), 0, NULL);
 
   WALK_LIST(anc, ac->net_list)
@@ -134,6 +136,16 @@ add_area_nets(struct ospf_area *oa, struct ospf_area_config *ac)
   }
 }
 
+static inline uint
+ospf_opts(struct ospf_proto *p)
+{
+  if (ospf_is_v2(p))
+    return 0;
+
+  return ((ospf_is_ip6(p) && !p->af_mc) ? OPT_V6 : 0) |
+    (!p->stub_router ? OPT_R : 0) | (p->af_ext ? OPT_AF : 0);
+}
+
 static void
 ospf_area_add(struct ospf_proto *p, struct ospf_area_config *ac)
 {
@@ -155,10 +167,7 @@ ospf_area_add(struct ospf_proto *p, struct ospf_area_config *ac)
   if (oa->areaid == 0)
     p->backbone = oa;
 
-  if (ospf_is_v2(p))
-    oa->options = ac->type;
-  else
-    oa->options = ac->type | OPT_V6 | (p->stub_router ? 0 : OPT_R);
+  oa->options = ac->type | ospf_opts(p);
 
   ospf_notify_rt_lsa(oa);
 }
@@ -224,6 +233,8 @@ ospf_start(struct proto *P)
 
   p->router_id = proto_get_router_id(P->cf);
   p->ospf2 = c->ospf2;
+  p->af_ext = c->af_ext;
+  p->af_mc = c->af_mc;
   p->rfc1583 = c->rfc1583;
   p->stub_router = c->stub_router;
   p->merge_external = c->merge_external;
@@ -238,8 +249,7 @@ ospf_start(struct proto *P)
   p->nhpool = lp_new(P->pool, 12*sizeof(struct nexthop));
   init_list(&(p->iface_list));
   init_list(&(p->area_list));
-  fib_init(&p->rtf, P->pool, p->ospf2 ? NET_IP4 : NET_IP6,
-	   sizeof(ort), OFFSETOF(ort, fn), 0, NULL);
+  fib_init(&p->rtf, P->pool, ospf_get_af(p), sizeof(ort), OFFSETOF(ort, fn), 0, NULL);
   if (ospf_is_v3(p))
     idm_init(&p->idm, P->pool, 16);
   p->areano = 0;
@@ -601,11 +611,7 @@ ospf_area_reconfigure(struct ospf_area *oa, struct ospf_area_config *nac)
   struct ospf_iface *ifa;
 
   oa->ac = nac;
-
-  if (ospf_is_v2(p))
-    oa->options = nac->type;
-  else
-    oa->options = nac->type | OPT_V6 | (p->stub_router ? 0 : OPT_R);
+  oa->options = nac->type | ospf_opts(p);
 
   if (nac->type != oac->type)
   {
@@ -659,6 +665,9 @@ ospf_reconfigure(struct proto *P, struct proto_config *CF)
   if (old->abr != new->abr)
     return 0;
 
+  if ((p->af_ext != new->af_ext) || (p->af_mc != new->af_mc))
+    return 0;
+
   if (!proto_configure_channel(P, &P->main_channel, proto_cf_main_channel(CF)))
     return 0;
 
@@ -1073,13 +1082,13 @@ show_lsa_network(struct top_hash_entry *he, int ospf2)
 }
 
 static inline void
-show_lsa_sum_net(struct top_hash_entry *he, int ospf2)
+show_lsa_sum_net(struct top_hash_entry *he, int ospf2, int af)
 {
   net_addr net;
   u8 pxopts;
   u32 metric;
 
-  lsa_parse_sum_net(he, ospf2, &net, &pxopts, &metric);
+  lsa_parse_sum_net(he, ospf2, af, &net, &pxopts, &metric);
   cli_msg(-1016, "\t\txnetwork %N metric %u", &net, metric);
 }
 
@@ -1096,7 +1105,7 @@ show_lsa_sum_rt(struct top_hash_entry *he, int ospf2)
 
 
 static inline void
-show_lsa_external(struct top_hash_entry *he, int ospf2)
+show_lsa_external(struct top_hash_entry *he, int ospf2, int af)
 {
   struct ospf_lsa_ext_local rt;
   char str_via[IPA_MAX_TEXT_LENGTH + 8] = "";
@@ -1105,7 +1114,7 @@ show_lsa_external(struct top_hash_entry *he, int ospf2)
   if (he->lsa_type == LSA_T_EXT)
     he->domain = 0; /* Unmark the LSA */
 
-  lsa_parse_ext(he, ospf2, &rt);
+  lsa_parse_ext(he, ospf2, af, &rt);
 
   if (rt.fbit)
     bsprintf(str_via, " via %I", rt.fwaddr);
@@ -1119,7 +1128,7 @@ show_lsa_external(struct top_hash_entry *he, int ospf2)
 }
 
 static inline void
-show_lsa_prefix(struct top_hash_entry *he, struct top_hash_entry *cnode)
+show_lsa_prefix(struct top_hash_entry *he, struct top_hash_entry *cnode, int af)
 {
   struct ospf_lsa_prefix *px = he->lsa_body;
   u32 *buf;
@@ -1142,7 +1151,7 @@ show_lsa_prefix(struct top_hash_entry *he, struct top_hash_entry *cnode)
     u8 pxopts;
     u16 metric;
 
-    buf = ospf_get_ipv6_prefix(buf, &net, &pxopts, &metric);
+    buf = ospf3_get_prefix(buf, af, &net, &pxopts, &metric);
 
     if (px->ref_type == LSA_T_RT)
       cli_msg(-1016, "\t\tstubnet %N metric %u", &net, metric);
@@ -1156,6 +1165,7 @@ ospf_sh_state(struct proto *P, int verbose, int reachable)
 {
   struct ospf_proto *p = (struct ospf_proto *) P;
   int ospf2 = ospf_is_v2(p);
+  int af = ospf_get_af(p);
   uint i, ix, j1, jx;
   u32 last_area = 0xFFFFFFFF;
 
@@ -1276,7 +1286,7 @@ ospf_sh_state(struct proto *P, int verbose, int reachable)
 
     case LSA_T_SUM_NET:
       if (cnode->lsa_type == LSA_T_RT)
-	show_lsa_sum_net(he, ospf2);
+	show_lsa_sum_net(he, ospf2, af);
       break;
 
     case LSA_T_SUM_RT:
@@ -1286,11 +1296,11 @@ ospf_sh_state(struct proto *P, int verbose, int reachable)
 
     case LSA_T_EXT:
     case LSA_T_NSSA:
-      show_lsa_external(he, ospf2);
+      show_lsa_external(he, ospf2, af);
       break;
 
     case LSA_T_PREFIX:
-      show_lsa_prefix(he, cnode);
+      show_lsa_prefix(he, cnode, af);
       break;
     }
 
@@ -1304,7 +1314,7 @@ ospf_sh_state(struct proto *P, int verbose, int reachable)
 	ix++;
 
       while ((ix < jx) && (hex[ix]->lsa.rt == cnode->lsa.rt))
-	show_lsa_external(hex[ix++], ospf2);
+	show_lsa_external(hex[ix++], ospf2, af);
 
       cnode = NULL;
     }
@@ -1338,7 +1348,7 @@ ospf_sh_state(struct proto *P, int verbose, int reachable)
 	last_rt = he->lsa.rt;
       }
 
-      show_lsa_external(he, ospf2);
+      show_lsa_external(he, ospf2, af);
     }
   }
 
diff --git a/proto/ospf/ospf.h b/proto/ospf/ospf.h
index e3eae2b5..1530b26a 100644
--- a/proto/ospf/ospf.h
+++ b/proto/ospf/ospf.h
@@ -84,10 +84,13 @@ struct ospf_config
   struct proto_config c;
   uint tick;
   u8 ospf2;
+  u8 af_ext;
+  u8 af_mc;
   u8 rfc1583;
   u8 stub_router;
   u8 merge_external;
   u8 instance_id;
+  u8 instance_id_set;
   u8 abr;
   u8 asbr;
   int ecmp;
@@ -168,9 +171,9 @@ struct ospf_iface_patt
   int tx_priority;
   u16 tx_length;
   u16 rx_buffer;
-
 #define OSPF_RXBUF_MINSIZE 256	/* Minimal allowed size */
   u8 instance_id;
+  u8 instance_id_set;
   u8 autype;			/* OSPF_AUTH_*, not really used in OSPFv3 */
   u8 strictnbma;
   u8 check_link;
@@ -211,12 +214,14 @@ struct ospf_proto
   int padj;			/* Number of neighbors in Exchange or Loading state */
   struct fib rtf;		/* Routing table */
   struct idm idm;		/* OSPFv3 LSA ID map */
-  byte ospf2;			/* OSPF v2 or v3 */
-  byte rfc1583;			/* RFC1583 compatibility */
-  byte stub_router;		/* Do not forward transit traffic */
-  byte merge_external;		/* Should i merge external routes? */
-  byte asbr;			/* May i originate any ext/NSSA lsa? */
-  byte ecmp;			/* Maximal number of nexthops in ECMP route, or 0 */
+  u8 ospf2;			/* OSPF v2 or v3 */
+  u8 af_ext;			/* OSPFv3-AF extension */
+  u8 af_mc;			/* OSPFv3-AF multicast */
+  u8 rfc1583;			/* RFC1583 compatibility */
+  u8 stub_router;		/* Do not forward transit traffic */
+  u8 merge_external;		/* Should i merge external routes? */
+  u8 asbr;			/* May i originate any ext/NSSA lsa? */
+  u8 ecmp;			/* Maximal number of nexthops in ECMP route, or 0 */
   struct ospf_area *backbone;	/* If exists */
   event *flood_event;		/* Event for flooding LS updates */
   void *lsab;			/* LSA buffer used when originating router LSAs */
@@ -449,14 +454,15 @@ struct ospf_neighbor
 
 
 /* Generic option flags */
-#define OPT_V6		0x01	/* OSPFv3, LSA relevant for IPv6 routing calculation */
-#define OPT_E		0x02	/* Related to AS-external LSAs */
-#define OPT_MC		0x04	/* Related to MOSPF, not used and obsolete */
-#define OPT_N		0x08	/* Related to NSSA */
-#define OPT_P		0x08	/* OSPFv2, flags P and N share position, see NSSA RFC */
-#define OPT_EA		0x10	/* OSPFv2, external attributes, not used and obsolete */
-#define OPT_R		0x10	/* OSPFv3, originator is active router */
-#define OPT_DC		0x20	/* Related to demand circuits, not used */
+#define OPT_V6		0x0001	/* OSPFv3, LSA relevant for IPv6 routing calculation */
+#define OPT_E		0x0002	/* Related to AS-external LSAs */
+#define OPT_MC		0x0004	/* Related to MOSPF, not used and obsolete */
+#define OPT_N		0x0008	/* Related to NSSA */
+#define OPT_P		0x0008	/* OSPFv2, flags P and N share position, see NSSA RFC */
+#define OPT_EA		0x0010	/* OSPFv2, external attributes, not used and obsolete */
+#define OPT_R		0x0010	/* OSPFv3, originator is active router */
+#define OPT_DC		0x0020	/* Related to demand circuits, not used */
+#define OPT_AF		0x0100	/* OSPFv3 Address Families (RFC 5838) */
 
 /* Router-LSA VEB flags are are stored together with links (OSPFv2) or options (OSPFv3) */
 #define OPT_RT_B	(0x01 << 24)
@@ -718,74 +724,96 @@ lsa_net_count(struct ospf_lsa_header *lsa)
 #define IPV6_PREFIX_SPACE(x) ((((x) + 63) / 32) * 4)
 #define IPV6_PREFIX_WORDS(x) (((x) + 63) / 32)
 
-/* FIXME: these functions should be significantly redesigned w.r.t. integration,
-   also should be named as ospf3_* instead of *_ipv6_* */
 
 static inline int
 ospf_valid_prefix(net_addr *n)
 {
-  /* In OSPFv2, prefix is stored as netmask; ip4_masklen() returns 255 for invalid one */
-  return n->pxlen <= IP6_MAX_PREFIX_LENGTH;
+  /*
+   * In OSPFv2, prefix is stored as netmask; ip4_masklen() returns 255 for
+   * invalid one. But OSPFv3-AF may receive IPv4 net with 32 < pxlen < 128.
+   */
+  uint max = (n->type == NET_IP4) ? IP4_MAX_PREFIX_LENGTH : IP6_MAX_PREFIX_LENGTH;
+  return n->pxlen <= max;
 }
 
+/*
+ * In OSPFv3-AF (RFC 5835), IPv4 address is encoded by just placing it in the
+ * first 32 bits of IPv6 address and setting remaining bits to zero. Likewise
+ * for IPv4 prefix, where remaining bits do not matter. We use following
+ * functions to convert between IPv4 and IPv4-in-IPv6 representations:
+ */
+
+static inline ip4_addr ospf3_6to4(ip6_addr a)
+{ return _MI4(_I0(a)); }
+
+static inline ip6_addr ospf3_4to6(ip4_addr a)
+{ return _MI6(_I(a), 0, 0, 0); }
+
+
 static inline u32 *
-ospf_get_ipv6_prefix(u32 *buf, net_addr *N, u8 *pxopts, u16 *rest)
+ospf3_get_prefix(u32 *buf, int af, net_addr *n, u8 *pxopts, u16 *rest)
 {
-  net_addr_ip6 *net = (void *) N;
-  u8 pxlen = (*buf >> 24);
+  ip6_addr px = IP6_NONE;
+  uint pxlen = (*buf >> 24);
   *pxopts = (*buf >> 16) & 0xff;
   if (rest) *rest = *buf & 0xffff;
   buf++;
 
-  *net = NET_ADDR_IP6(IP6_NONE, pxlen);
-
   if (pxlen > 0)
-    _I0(net->prefix) = *buf++;
+    _I0(px) = *buf++;
   if (pxlen > 32)
-    _I1(net->prefix) = *buf++;
+    _I1(px) = *buf++;
   if (pxlen > 64)
-    _I2(net->prefix) = *buf++;
+    _I2(px) = *buf++;
   if (pxlen > 96)
-    _I3(net->prefix) = *buf++;
+    _I3(px) = *buf++;
 
   /* Clean up remaining bits */
   if (pxlen < 128)
-    net->prefix.addr[pxlen / 32] &= u32_mkmask(pxlen % 32);
+    px.addr[pxlen / 32] &= u32_mkmask(pxlen % 32);
+
+  if (af == NET_IP4)
+    net_fill_ip4(n, ospf3_6to4(px), pxlen);
+  else
+    net_fill_ip6(n, px, pxlen);
 
   return buf;
 }
 
 static inline u32 *
-ospf_get_ipv6_addr(u32 *buf, ip_addr *addr)
+ospf3_put_prefix(u32 *buf, net_addr *n, u8 pxopts, u16 rest)
 {
-  *addr = ipa_from_ip6(*(ip6_addr *) buf);
-  return buf + 4;
-}
-
-static inline u32 *
-ospf_put_ipv6_prefix(u32 *buf, net_addr *N, u8 pxopts, u16 rest)
-{
-  net_addr_ip6 *net = (void *) N;
-  u32 pxlen = net->pxlen;
+  ip6_addr px = (n->type == NET_IP4) ? ospf3_4to6(net4_prefix(n)) : net6_prefix(n);
+  uint pxlen = n->pxlen;
 
   *buf++ = ((pxlen << 24) | (pxopts << 16) | rest);
 
   if (pxlen > 0)
-    *buf++ = _I0(net->prefix);
+    *buf++ = _I0(px);
   if (pxlen > 32)
-    *buf++ = _I1(net->prefix);
+    *buf++ = _I1(px);
   if (pxlen > 64)
-    *buf++ = _I2(net->prefix);
+    *buf++ = _I2(px);
   if (pxlen > 96)
-    *buf++ = _I3(net->prefix);
+    *buf++ = _I3(px);
 
   return buf;
 }
 
 static inline u32 *
-ospf_put_ipv6_addr(u32 *buf, ip_addr addr)
+ospf3_get_addr(u32 *buf, int af, ip_addr *addr)
 {
-  *(ip6_addr *) buf = ipa_to_ip6(addr);
+  ip6_addr a;
+  memcpy(&a, buf, 16);
+  *addr = (af == NET_IP4) ? ipa_from_ip4(ospf3_6to4(a)) : ipa_from_ip6(a);
+  return buf + 4;
+}
+
+static inline u32 *
+ospf3_put_addr(u32 *buf, ip_addr addr)
+{
+  ip6_addr a = ipa_is_ip4(addr) ? ospf3_4to6(ipa_to_ip4(addr)) : ipa_to_ip6(addr);
+  memcpy(buf, &a, 16);
   return buf + 4;
 }
 
@@ -838,6 +866,15 @@ static inline int ospf_is_v3(struct ospf_proto *p)
 static inline int ospf_get_version(struct ospf_proto *p)
 { return ospf_is_v2(p) ? 2 : 3; }
 
+static inline int ospf_is_ip4(struct ospf_proto *p)
+{ return p->p.net_type == NET_IP4; }
+
+static inline int ospf_is_ip6(struct ospf_proto *p)
+{ return p->p.net_type == NET_IP6; }
+
+static inline int ospf_get_af(struct ospf_proto *p)
+{ return p->p.net_type; }
+
 struct ospf_area *ospf_find_area(struct ospf_proto *p, u32 aid);
 
 static inline struct ospf_area *ospf_main_area(struct ospf_proto *p)
diff --git a/proto/ospf/rt.c b/proto/ospf/rt.c
index df9eb75b..f57925c3 100644
--- a/proto/ospf/rt.c
+++ b/proto/ospf/rt.c
@@ -576,20 +576,20 @@ spfa_process_prefixes(struct ospf_proto *p, struct ospf_area *oa)
     buf = px->rest;
     for (i = 0; i < px->pxcount; i++)
     {
-      net_addr_ip6 net;
+      net_addr net;
       u8 pxopts;
       u16 metric;
 
-      buf = ospf_get_ipv6_prefix(buf, (net_addr *) &net, &pxopts, &metric);
+      buf = ospf3_get_prefix(buf, ospf_get_af(p), &net, &pxopts, &metric);
 
       if (pxopts & OPT_PX_NU)
 	continue;
 
       /* Store the first global address to use it later as a vlink endpoint */
-      if ((pxopts & OPT_PX_LA) && ipa_zero(src->lb))
-	src->lb = ipa_from_ip6(net.prefix);
+      if ((pxopts & OPT_PX_LA) && (net.type == NET_IP6) && ipa_zero(src->lb))
+	src->lb = ipa_from_ip6(net6_prefix(&net));
 
-      add_network(oa, (net_addr *) &net, src->dist + metric, src, i);
+      add_network(oa, &net, src->dist + metric, src, i);
     }
   }
 }
@@ -761,7 +761,7 @@ ospf_rt_sum(struct ospf_area *oa)
 
     if (en->lsa_type == LSA_T_SUM_NET)
     {
-      lsa_parse_sum_net(en, ospf_is_v2(p), &net, &pxopts, &metric);
+      lsa_parse_sum_net(en, ospf_is_v2(p), ospf_get_af(p), &net, &pxopts, &metric);
 
       if (!ospf_valid_prefix(&net))
       {
@@ -858,7 +858,7 @@ ospf_rt_sum_tr(struct ospf_area *oa)
       net_addr net;
       u8 pxopts;
 
-      lsa_parse_sum_net(en, ospf_is_v2(p), &net, &pxopts, &metric);
+      lsa_parse_sum_net(en, ospf_is_v2(p), ospf_get_af(p), &net, &pxopts, &metric);
 
       if (!ospf_valid_prefix(&net))
       {
@@ -1058,7 +1058,7 @@ decide_nssa_lsa(struct ospf_proto *p, ort *nf, struct ospf_lsa_ext_local *rt)
     return 0;
 
   /* We do not store needed data in struct orta, we have to parse the LSA */
-  lsa_parse_ext(en, ospf_is_v2(p), rt);
+  lsa_parse_ext(en, ospf_is_v2(p), ospf_get_af(p), rt);
 
   if (rt->pxopts & OPT_PX_NU)
     return 0;
@@ -1450,7 +1450,7 @@ ospf_ext_spf(struct ospf_proto *p)
     DBG("%s: Working on LSA. ID: %R, RT: %R, Type: %u\n",
 	p->p.name, en->lsa.id, en->lsa.rt, en->lsa_type);
 
-    lsa_parse_ext(en, ospf_is_v2(p), &rt);
+    lsa_parse_ext(en, ospf_is_v2(p), ospf_get_af(p), &rt);
 
     if (!ospf_valid_prefix(&rt.net))
     {
@@ -1765,7 +1765,11 @@ calc_next_hop(struct ospf_area *oa, struct top_hash_entry *en,
       if (ip6_zero(llsa->lladdr))
 	return NULL;
 
-      return new_nexthop(p, ipa_from_ip6(llsa->lladdr), pn->iface, pn->weight);
+      ip_addr nh = ospf_is_ip4(p) ?
+	ipa_from_ip4(ospf3_6to4(llsa->lladdr)) :
+	ipa_from_ip6(llsa->lladdr);
+
+      return new_nexthop(p, nh, pn->iface, pn->weight);
     }
   }
 
@@ -1792,9 +1796,9 @@ add_cand(list * l, struct top_hash_entry *en, struct top_hash_entry *par,
   if (en->lsa.age == LSA_MAXAGE)
     return;
 
-  if (ospf_is_v3(p) && (en->lsa_type == LSA_T_RT))
+  if (ospf_is_v3(p) && (oa->options & OPT_V6) && (en->lsa_type == LSA_T_RT))
   {
-    /* In OSPFv3, check V6 flag */
+    /* In OSPFv3 IPv6 unicast, check V6 flag */
     struct ospf_lsa_rt *rt = en->lsa_body;
     if (!(rt->options & OPT_V6))
       return;
diff --git a/proto/ospf/topology.c b/proto/ospf/topology.c
index ce77f57a..8dd91a40 100644
--- a/proto/ospf/topology.c
+++ b/proto/ospf/topology.c
@@ -1008,7 +1008,7 @@ prepare_sum3_net_lsa_body(struct ospf_proto *p, ort *nf, u32 metric)
   sum = lsab_allocz(p, sizeof(struct ospf_lsa_sum3_net) +
 		    IPV6_PREFIX_SPACE(nf->fn.addr->pxlen));
   sum->metric = metric;
-  ospf_put_ipv6_prefix(sum->prefix, nf->fn.addr, 0, 0);
+  ospf3_put_prefix(sum->prefix, nf->fn.addr, 0, 0);
 }
 
 static inline void
@@ -1097,7 +1097,7 @@ prepare_ext3_lsa_body(struct ospf_proto *p, ort *nf,
   ext->metric = metric & LSA_METRIC_MASK;
   u32 *buf = ext->rest;
 
-  buf = ospf_put_ipv6_prefix(buf, nf->fn.addr, pbit ? OPT_PX_P : 0, 0);
+  buf = ospf3_put_prefix(buf, nf->fn.addr, pbit ? OPT_PX_P : 0, 0);
 
   if (ebit)
     ext->metric |= LSA_EXT3_EBIT;
@@ -1105,7 +1105,7 @@ prepare_ext3_lsa_body(struct ospf_proto *p, ort *nf,
   if (ipa_nonzero(fwaddr))
   {
     ext->metric |= LSA_EXT3_FBIT;
-    buf = ospf_put_ipv6_addr(buf, fwaddr);
+    buf = ospf3_put_addr(buf, fwaddr);
   }
 
   if (tag)
@@ -1222,7 +1222,7 @@ find_surrogate_fwaddr(struct ospf_proto *p, struct ospf_area *oa)
     {
       WALK_LIST(a, ifa->iface->addrs)
       {
-	if ((a->prefix.type != NET_IP6) ||
+	if ((a->prefix.type != ospf_get_af(p)) ||
 	    (a->flags & IA_SECONDARY) ||
 	    (a->flags & IA_PEER) ||
 	    (a->scope <= SCOPE_LINK))
@@ -1316,39 +1316,47 @@ ospf_rt_notify(struct proto *P, struct channel *ch UNUSED, net *n, rte *new, rte
  */
 
 static inline void
-lsab_put_prefix(struct ospf_proto *p, net_addr *net, u32 cost)
+lsab_put_prefix(struct ospf_proto *p, net_addr *n, u32 cost)
 {
-  void *buf = lsab_alloc(p, IPV6_PREFIX_SPACE(net6_pxlen(net)));
-  u8 flags = (net6_pxlen(net) < IP6_MAX_PREFIX_LENGTH) ? 0 : OPT_PX_LA;
-  ospf_put_ipv6_prefix(buf, net, flags, cost);
+  void *buf = lsab_alloc(p, IPV6_PREFIX_SPACE(net_pxlen(n)));
+  uint max = (n->type == NET_IP4) ? IP4_MAX_PREFIX_LENGTH : IP6_MAX_PREFIX_LENGTH;
+  u8 flags = (net_pxlen(n) < max) ? 0 : OPT_PX_LA;
+  ospf3_put_prefix(buf, n, flags, cost);
 }
 
 static void
 prepare_link_lsa_body(struct ospf_proto *p, struct ospf_iface *ifa)
 {
-  struct ospf_lsa_link *ll;
+  ip_addr nh = ospf_is_ip4(p) ? IPA_NONE : ifa->addr->ip;
   int i = 0;
 
+  /* Preallocating space for header */
   ASSERT(p->lsab_used == 0);
-  ll = lsab_allocz(p, sizeof(struct ospf_lsa_link));
-  ll->options = ifa->oa->options | (ifa->priority << 24);
-  ll->lladdr = ipa_to_ip6(ifa->addr->ip);
-  ll = NULL; /* buffer might be reallocated later */
+  lsab_allocz(p, sizeof(struct ospf_lsa_link));
 
   struct ifa *a;
   WALK_LIST(a, ifa->iface->addrs)
   {
-    if ((a->prefix.type != NET_IP6) ||
+    if ((a->prefix.type != ospf_get_af(p)) ||
 	(a->flags & IA_SECONDARY) ||
 	(a->scope <= SCOPE_LINK))
       continue;
 
+    if (ospf_is_ip4(p) && ipa_zero(nh))
+      nh = a->ip;
+
     lsab_put_prefix(p, &a->prefix, 0);
     i++;
   }
 
-  ll = p->lsab;
+  /* Filling the preallocated header */
+  struct ospf_lsa_link *ll = p->lsab;
+  ll->options = ifa->oa->options | (ifa->priority << 24);
+  ll->lladdr = ospf_is_ip4(p) ? ospf3_4to6(ipa_to_ip4(nh)) : ipa_to_ip6(nh);
   ll->pxcount = i;
+
+  if (ipa_zero(nh))
+    log(L_ERR "%s: Cannot find next hop address for %s", p->p.name, ifa->ifname);
 }
 
 static void
@@ -1410,7 +1418,7 @@ prepare_prefix_rt_lsa_body(struct ospf_proto *p, struct ospf_area *oa)
     struct ifa *a;
     WALK_LIST(a, ifa->iface->addrs)
     {
-      if ((a->prefix.type != NET_IP6) ||
+      if ((a->prefix.type != ospf_get_af(p)) ||
 	  (a->flags & IA_SECONDARY) ||
 	  (a->flags & IA_PEER) ||
 	  (a->scope <= SCOPE_LINK))
@@ -1448,7 +1456,7 @@ prepare_prefix_rt_lsa_body(struct ospf_proto *p, struct ospf_area *oa)
 
   /* If there are some configured vlinks, find some global address
      (even from another area), which will be used as a vlink endpoint. */
-  if (!EMPTY_LIST(cf->vlink_list) && !host_addr)
+  if (!EMPTY_LIST(cf->vlink_list) && !host_addr && ospf_is_ip6(p))
   {
     WALK_LIST(ifa, p->iface_list)
     {
@@ -1571,7 +1579,7 @@ add_link_lsa(struct ospf_proto *p, struct ospf_lsa_link *ll, int offset, int *px
       continue;
 
     /* Skip link-local prefixes */
-    if ((pxlen >= 10) && ((pxb[1] & 0xffc00000) == 0xfe800000))
+    if (ospf_is_ip6(p) && (pxlen >= 10) && ((pxb[1] & 0xffc00000) == 0xfe800000))
       continue;
 
     add_prefix(p, pxb, offset, pxc);

From f097f7659c7ff226a53c51673158e32fb69a6d21 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Mon, 9 Oct 2017 20:36:14 +0200
Subject: [PATCH 034/198] OSPF: Fix next hop calculation for PtP links in IPv4
 OSPFv3-AF

In such case, next hop has to be taken from Link-LSA like in broadcast
case, not from neighbor source address like in other PtP cases.

Also add some checks, comments and code cleanup.
---
 proto/ospf/rt.c | 115 ++++++++++++++++++++++++++++++++++--------------
 1 file changed, 82 insertions(+), 33 deletions(-)

diff --git a/proto/ospf/rt.c b/proto/ospf/rt.c
index f57925c3..b289d767 100644
--- a/proto/ospf/rt.c
+++ b/proto/ospf/rt.c
@@ -10,9 +10,7 @@
 
 #include "ospf.h"
 
-static void add_cand(list * l, struct top_hash_entry *en,
-		     struct top_hash_entry *par, u32 dist,
-		     struct ospf_area *oa, int i);
+static void add_cand(struct ospf_area *oa, struct top_hash_entry *en, struct top_hash_entry *par, u32 dist, int i, uint lif, uint nif);
 static void rt_sync(struct ospf_proto *p);
 
 
@@ -509,7 +507,7 @@ spfa_process_rt(struct ospf_proto *p, struct ospf_area *oa, struct top_hash_entr
       break;
     }
 
-    add_cand(&oa->cand, tmp, act, act->dist + rtl.metric, oa, i);
+    add_cand(oa, tmp, act, act->dist + rtl.metric, i, rtl.lif, rtl.nif);
   }
 }
 
@@ -532,7 +530,7 @@ spfa_process_net(struct ospf_proto *p, struct ospf_area *oa, struct top_hash_ent
   for (i = 0; i < cnt; i++)
   {
     tmp = ospf_hash_find_rt(p->gr, oa->areaid, ln->routers[i]);
-    add_cand(&oa->cand, tmp, act, act->dist, oa, -1);
+    add_cand(oa, tmp, act, act->dist, -1, 0, 0);
   }
 }
 
@@ -651,7 +649,8 @@ ospf_rt_spfa(struct ospf_area *oa)
 }
 
 static int
-link_back(struct ospf_area *oa, struct top_hash_entry *en, struct top_hash_entry *par)
+link_back(struct ospf_area *oa, struct top_hash_entry *en,
+	  struct top_hash_entry *par, uint lif, uint nif)
 {
   struct ospf_proto *p = oa->po;
   struct ospf_lsa_rt_walk rtl;
@@ -689,6 +688,10 @@ link_back(struct ospf_area *oa, struct top_hash_entry *en, struct top_hash_entry
 	tmp = ospf_hash_find_net(p->gr, oa->areaid, rtl.id, rtl.nif);
 	if (tmp == par)
 	{
+	  /*
+	   * Note that there may be multiple matching Rt-fields if router 'en'
+	   * have multiple interfaces to net 'par'. Perhaps we should do ECMP.
+	   */
 	  if (ospf_is_v2(p))
 	    en->lb = ipa_from_u32(rtl.data);
 	  else
@@ -700,7 +703,13 @@ link_back(struct ospf_area *oa, struct top_hash_entry *en, struct top_hash_entry
 
       case LSART_VLNK:
       case LSART_PTP:
-	/* Not necessary the same link, see RFC 2328 [23] */
+	/*
+	 * For OSPFv2, not necessary the same link, see RFC 2328 [23].
+	 * For OSPFv3, we verify that by comparing nif and lif fields.
+	 */
+	if (ospf_is_v3(p) && ((rtl.lif != nif) || (rtl.nif != lif)))
+	  break;
+
 	tmp = ospf_hash_find_rt(p->gr, oa->areaid, rtl.id);
 	if (tmp == par)
 	  return 1;
@@ -1679,13 +1688,27 @@ inherit_nexthops(struct nexthop *pn)
   return pn && (ipa_nonzero(pn->gw) || !pn->iface);
 }
 
+static inline ip_addr
+link_lsa_lladdr(struct ospf_proto *p, struct top_hash_entry *en)
+{
+  struct ospf_lsa_link *link_lsa = en->lsa_body;
+  ip6_addr ll = link_lsa->lladdr;
+
+  if (ip6_zero(ll))
+    return IPA_NONE;
+
+  return ospf_is_ip4(p) ? ipa_from_ip4(ospf3_6to4(ll)) : ipa_from_ip6(ll);
+}
+
 static struct nexthop *
 calc_next_hop(struct ospf_area *oa, struct top_hash_entry *en,
-	      struct top_hash_entry *par, int pos)
+	      struct top_hash_entry *par, int pos, uint lif, uint nif)
 {
   struct ospf_proto *p = oa->po;
   struct nexthop *pn = par->nhs;
-  struct ospf_iface *ifa;
+  struct top_hash_entry *link = NULL;
+  struct ospf_iface *ifa = NULL;
+  ip_addr nh = IPA_NONE;
   u32 rid = en->lsa.rt;
 
   /* 16.1.1. The next hop calculation */
@@ -1710,6 +1733,9 @@ calc_next_hop(struct ospf_area *oa, struct top_hash_entry *en,
     if (!ifa)
       return NULL;
 
+    if (ospf_is_v3(p) && (ifa->iface_id != lif))
+      log(L_WARN "%s: Inconsistent interface ID %u/%u", p->p.name, ifa->iface_id, lif);
+
     return new_nexthop(p, IPA_NONE, ifa->iface, ifa->ecmp_weight);
   }
 
@@ -1720,14 +1746,44 @@ calc_next_hop(struct ospf_area *oa, struct top_hash_entry *en,
     if (!ifa)
       return NULL;
 
+    if (ospf_is_v3(p) && (ifa->iface_id != lif))
+      log(L_WARN "%s: Inconsistent interface ID %u/%u", p->p.name, ifa->iface_id, lif);
+
     if (ifa->type == OSPF_IT_VLINK)
       return new_nexthop(p, IPA_NONE, NULL, 0);
 
-    struct ospf_neighbor *m = find_neigh(ifa, rid);
-    if (!m || (m->state != NEIGHBOR_FULL))
-      return NULL;
+    /* FIXME: On physical PtP links we may skip next-hop altogether */
 
-    return new_nexthop(p, m->ip, ifa->iface, ifa->ecmp_weight);
+    if (ospf_is_v2(p) || ospf_is_ip6(p))
+    {
+      /*
+       * In this case, next-hop is a source address from neighbor's packets.
+       * That is necessary for OSPFv2 and practical for OSPFv3 (as it works even
+       * if neighbor uses LinkLSASuppression), but does not work with OSPFv3-AF
+       * on IPv4 topology, where src is IPv6 but next-hop should be IPv4.
+       */
+      struct ospf_neighbor *m = find_neigh(ifa, rid);
+      if (!m || (m->state != NEIGHBOR_FULL))
+	return NULL;
+
+      nh = m->ip;
+    }
+    else
+    {
+      /*
+       * Next-hop is taken from lladdr field of Link-LSA, based on Neighbor
+       * Iface ID (nif) field in our Router-LSA, which is just nbr->iface_id.
+       */
+      link = ospf_hash_find(p->gr, ifa->iface_id, nif, rid, LSA_T_LINK);
+      if (!link)
+	return NULL;
+
+      nh = link_lsa_lladdr(p, link);
+      if (ipa_zero(nh))
+	return NULL;
+    }
+
+    return new_nexthop(p, nh, ifa->iface, ifa->ecmp_weight);
   }
 
   /* The third case - bcast or nbma neighbor */
@@ -1754,21 +1810,14 @@ calc_next_hop(struct ospf_area *oa, struct top_hash_entry *en,
        * Next-hop is taken from lladdr field of Link-LSA, en->lb_id
        * is computed in link_back().
        */
-      struct top_hash_entry *lhe;
-      lhe = ospf_hash_find(p->gr, pn->iface->index, en->lb_id, rid, LSA_T_LINK);
-
-      if (!lhe)
+      link = ospf_hash_find(p->gr, pn->iface->index, en->lb_id, rid, LSA_T_LINK);
+      if (!link)
 	return NULL;
 
-      struct ospf_lsa_link *llsa = lhe->lsa_body;
-
-      if (ip6_zero(llsa->lladdr))
+      nh = link_lsa_lladdr(p, link);
+      if (ipa_zero(nh))
 	return NULL;
 
-      ip_addr nh = ospf_is_ip4(p) ?
-	ipa_from_ip4(ospf3_6to4(llsa->lladdr)) :
-	ipa_from_ip6(llsa->lladdr);
-
       return new_nexthop(p, nh, pn->iface, pn->weight);
     }
   }
@@ -1782,8 +1831,8 @@ calc_next_hop(struct ospf_area *oa, struct top_hash_entry *en,
 
 /* Add LSA into list of candidates in Dijkstra's algorithm */
 static void
-add_cand(list * l, struct top_hash_entry *en, struct top_hash_entry *par,
-	 u32 dist, struct ospf_area *oa, int pos)
+add_cand(struct ospf_area *oa, struct top_hash_entry *en, struct top_hash_entry *par,
+	 u32 dist, int pos, uint lif, uint nif)
 {
   struct ospf_proto *p = oa->po;
   node *prev, *n;
@@ -1813,10 +1862,10 @@ add_cand(list * l, struct top_hash_entry *en, struct top_hash_entry *par,
     return;
 
   /* We should check whether there is a reverse link from en to par, */
-  if (!link_back(oa, en, par))
+  if (!link_back(oa, en, par, lif, nif))
     return;
 
-  struct nexthop *nhs = calc_next_hop(oa, en, par, pos);
+  struct nexthop *nhs = calc_next_hop(oa, en, par, pos, lif, nif);
   if (!nhs)
   {
     log(L_WARN "%s: Cannot find next hop for LSA (Type: %04x, Id: %R, Rt: %R)",
@@ -1873,20 +1922,20 @@ add_cand(list * l, struct top_hash_entry *en, struct top_hash_entry *par,
 
   prev = NULL;
 
-  if (EMPTY_LIST(*l))
+  if (EMPTY_LIST(oa->cand))
   {
-    add_head(l, &en->cn);
+    add_head(&oa->cand, &en->cn);
   }
   else
   {
-    WALK_LIST(n, *l)
+    WALK_LIST(n, oa->cand)
     {
       act = SKIP_BACK(struct top_hash_entry, cn, n);
       if ((act->dist > dist) ||
 	  ((act->dist == dist) && (act->lsa_type == LSA_T_RT)))
       {
 	if (prev == NULL)
-	  add_head(l, &en->cn);
+	  add_head(&oa->cand, &en->cn);
 	else
 	  insert_node(&en->cn, prev);
 	added = 1;
@@ -1897,7 +1946,7 @@ add_cand(list * l, struct top_hash_entry *en, struct top_hash_entry *par,
 
     if (!added)
     {
-      add_tail(l, &en->cn);
+      add_tail(&oa->cand, &en->cn);
     }
   }
 }

From 2549300b54f262932dd14e6d465926627e6dc8ef Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Mon, 9 Oct 2017 21:11:53 +0200
Subject: [PATCH 035/198] OSPF: Fix minor issue in TTL check

The TTL check must be done after instance ID dispatch to avoid warnings
when a physical iface is shared by multiple instances and some use TTL
security and some not.
---
 proto/ospf/packet.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/proto/ospf/packet.c b/proto/ospf/packet.c
index b84780d3..7f9f3b39 100644
--- a/proto/ospf/packet.c
+++ b/proto/ospf/packet.c
@@ -270,9 +270,6 @@ ospf_rx_hook(sock *sk, uint len)
   if (pkt == NULL)
     DROP("bad IP header", len);
 
-  if (ifa->check_ttl && (sk->rcv_ttl < 255))
-    DROP("wrong TTL", sk->rcv_ttl);
-
   if (len < sizeof(struct ospf_packet))
     DROP("too short", len);
 
@@ -379,6 +376,10 @@ found:
   if (ipa_equal(sk->laddr, ifa->des_routers) && (ifa->sk_dr == 0))
     return 1;
 
+  /* TTL check must be done after instance dispatch */
+  if (ifa->check_ttl && (sk->rcv_ttl < 255))
+    DROP("wrong TTL", sk->rcv_ttl);
+
   if (rid == p->router_id)
     DROP1("my own router ID");
 

From 2d7fb19c3c6eadfbd9b994f5c306e79b7c4ee994 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 10 Oct 2017 16:07:54 +0200
Subject: [PATCH 036/198] OSPF: Add option to disable OSPFv3-AF

---
 doc/bird.sgml       | 15 ++++++++++++---
 proto/ospf/config.Y |  3 ++-
 2 files changed, 14 insertions(+), 4 deletions(-)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index 964895e3..1b4b3d67 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -2866,6 +2866,7 @@ constant definitions and list of neighbors on nonbroadcast networks.
 <code>
 protocol ospf [v2|v3] &lt;name&gt; {
 	rfc1583compat &lt;switch&gt;;
+	rfc5838 &lt;switch&gt;;
 	instance id &lt;num&gt;;
 	stub router &lt;switch&gt;;
 	tick &lt;num&gt;;
@@ -2966,15 +2967,23 @@ protocol ospf [v2|v3] &lt;name&gt; {
 	This option controls compatibility of routing table calculation with
 	<rfc id="1583">. Default value is no.
 
+	<tag><label id="ospf-rfc5838">rfc5838 <m/switch/</tag>
+	Basic OSPFv3 is limited to IPv6 unicast routing. The <rfc id="5838">
+	extension defines support for more address families (IPv4, IPv6, both
+	unicast and multicast). The extension is enabled by default, but can be
+	disabled if necessary, as it restricts the range of available instance
+	IDs. Default value is yes.
+
 	<tag><label id="ospf-instance-id">instance id <m/num/</tag>
 	When multiple OSPF protocol instances are active on the same links, they
 	should use different instance IDs to distinguish their packets. Although
 	it could be done on per-interface basis, it is often preferred to set
 	one instance ID to whole OSPF domain/topology (e.g., when multiple
 	instances are used to represent separate logical topologies on the same
-	physical network). This option specifies the default instance ID for all
-	interfaces of the OSPF instance. Note that this option, if used, must
-	precede interface definitions. Default value is 0.
+	physical network). This option specifies the instance ID for all
+	interfaces of the OSPF instance, but can be overridden by
+	<cf/interface/ option. Default value is 0 unless OSPFv3-AF extended
+	address families are used, see <rfc id="5838"> for that case.
 
 	<tag><label id="ospf-stub-router">stub router <M>switch</M></tag>
 	This option configures the router to be a stub router, i.e., a router
diff --git a/proto/ospf/config.Y b/proto/ospf/config.Y
index 85149ea1..98ddf5d0 100644
--- a/proto/ospf/config.Y
+++ b/proto/ospf/config.Y
@@ -196,7 +196,7 @@ CF_KEYWORDS(ELIGIBLE, POLL, NETWORKS, HIDDEN, VIRTUAL, CHECK, LINK, ONLY, BFD)
 CF_KEYWORDS(RX, BUFFER, LARGE, NORMAL, STUBNET, HIDDEN, SUMMARY, TAG, EXTERNAL)
 CF_KEYWORDS(WAIT, DELAY, LSADB, ECMP, LIMIT, WEIGHT, NSSA, TRANSLATOR, STABILITY)
 CF_KEYWORDS(GLOBAL, LSID, ROUTER, SELF, INSTANCE, REAL, NETMASK, TX, PRIORITY, LENGTH)
-CF_KEYWORDS(SECONDARY, MERGE, LSA, SUPPRESSION, MULTICAST)
+CF_KEYWORDS(SECONDARY, MERGE, LSA, SUPPRESSION, MULTICAST, RFC5838)
 
 %type <ld> lsadb_args
 %type <i> ospf_variant ospf_af_mc nbma_eligible
@@ -250,6 +250,7 @@ ospf_proto_item:
    proto_item
  | ospf_channel { this_proto->net_type = $1->net_type; }
  | RFC1583COMPAT bool { OSPF_CFG->rfc1583 = $2; }
+ | RFC5838 bool { OSPF_CFG->af_ext = $2; if (!ospf_cfg_is_v3()) cf_error("RFC5838 option requires OSPFv3"); }
  | STUB ROUTER bool { OSPF_CFG->stub_router = $3; }
  | ECMP bool { OSPF_CFG->ecmp = $2 ? OSPF_DEFAULT_ECMP_LIMIT : 0; }
  | ECMP bool LIMIT expr { OSPF_CFG->ecmp = $2 ? $4 : 0; }

From 289c1a7968c5a3c0a76124a89eb45de010f1c640 Mon Sep 17 00:00:00 2001
From: Jan Moskyto Matejka <mq@ucw.cz>
Date: Mon, 21 Aug 2017 14:14:07 +0200
Subject: [PATCH 037/198] Iface address debug dump fix

---
 nest/iface.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/nest/iface.c b/nest/iface.c
index 00af5052..f14e70a0 100644
--- a/nest/iface.c
+++ b/nest/iface.c
@@ -46,10 +46,11 @@ list iface_list;
 void
 ifa_dump(struct ifa *a)
 {
-  debug("\t%I, net %N bc %I -> %I%s%s%s\n", a->ip, &a->prefix, a->brd, a->opposite,
-	(a->flags & IF_UP) ? "" : " DOWN",
-	(a->flags & IA_PRIMARY) ? "" : " SEC",
-	(a->flags & IA_PEER) ? "PEER" : "");
+  debug("\t%I, net %N bc %I -> %I%s%s%s%s\n", a->ip, &a->prefix, a->brd, a->opposite,
+	(a->flags & IA_PRIMARY) ? " PRIMARY" : "",
+	(a->flags & IA_SECONDARY) ? " SEC" : "",
+	(a->flags & IA_HOST) ? " HOST" : "",
+	(a->flags & IA_PEER) ? " PEER" : "");
 }
 
 /**

From 4ae3ee1200b386219673c2168eae996c6207b077 Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Tue, 22 Aug 2017 13:47:01 +0200
Subject: [PATCH 038/198] Babel: Interface address irrelevant for interface
 pattern matching.

---
 proto/babel/babel.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/proto/babel/babel.c b/proto/babel/babel.c
index 67a59df6..a0ec3629 100644
--- a/proto/babel/babel.c
+++ b/proto/babel/babel.c
@@ -1570,7 +1570,7 @@ babel_if_notify(struct proto *P, unsigned flags, struct iface *iface)
 
   if (flags & IF_CHANGE_UP)
   {
-    struct babel_iface_config *ic = (void *) iface_patt_find(&cf->iface_list, iface, iface->addr);
+    struct babel_iface_config *ic = (void *) iface_patt_find(&cf->iface_list, iface, NULL);
 
     /* we only speak multicast */
     if (!(iface->flags & IF_MULTICAST))

From 9ba4b4a63d27943586b76574b7f310ee532509a3 Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Thu, 9 Nov 2017 15:04:05 +0100
Subject: [PATCH 039/198] Filter test: typo fix

---
 filter/test.conf | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/filter/test.conf b/filter/test.conf
index 676b47d1..4dc67c49 100644
--- a/filter/test.conf
+++ b/filter/test.conf
@@ -437,7 +437,7 @@ string st;
 	print "1.2.3.4 = ", onetwo;
 
 	i = 4200000000;
-	print "4200000000 = ", i, "   false: ", i = 4200000000, " ", i > 4100000000, "   false: ", i > 4250000000;
+	print "4200000000 = ", i, "   true: ", i = 4200000000, " ", i > 4100000000, "   false: ", i > 4250000000;
 
 	test_undef(2);
 	test_undef(3);

From 1e8721e2aeccfbc3f533e8b8abc07582cee77e9a Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 28 Nov 2017 19:33:33 +0100
Subject: [PATCH 040/198] Babel: Parse flags in Hello TLV

RFC6126bis introduces a flags field for the Hello TLV, and adds a unicast flag
that is used to signify that a hello was sent as unicast. This adds parsing of
the flags field and ignores such unicast hellos, which preserves compatibility
until we can add a proper implementation of the unicast hello mechanism.

Thanks to Toke Hoiland-Jorgensen for the patch.
---
 proto/babel/packets.c | 19 ++++++++++++++-----
 1 file changed, 14 insertions(+), 5 deletions(-)

diff --git a/proto/babel/packets.c b/proto/babel/packets.c
index e9c6d51d..768858d0 100644
--- a/proto/babel/packets.c
+++ b/proto/babel/packets.c
@@ -40,7 +40,7 @@ struct babel_tlv_ack {
 struct babel_tlv_hello {
   u8 type;
   u8 length;
-  u16 reserved;
+  u16 flags;
   u16 seqno;
   u16 interval;
 } PACKED;
@@ -104,8 +104,12 @@ struct babel_tlv_seqno_request {
 } PACKED;
 
 
-#define BABEL_FLAG_DEF_PREFIX	0x80
-#define BABEL_FLAG_ROUTER_ID	0x40
+/* Hello flags */
+#define BABEL_HF_UNICAST	0x8000
+
+/* Update flags */
+#define BABEL_UF_DEF_PREFIX	0x80
+#define BABEL_UF_ROUTER_ID	0x40
 
 
 struct babel_parse_state {
@@ -310,6 +314,11 @@ babel_read_hello(struct babel_tlv *hdr, union babel_msg *m,
   struct babel_tlv_hello *tlv = (void *) hdr;
   struct babel_msg_hello *msg = &m->hello;
 
+  /* We currently don't support unicast Hello */
+  u16 flags = get_u16(&tlv->flags);
+  if (flags & BABEL_HF_UNICAST)
+    return PARSE_IGNORE;
+
   msg->type = BABEL_TLV_HELLO;
   msg->seqno = get_u16(&tlv->seqno);
   msg->interval = get_time16(&tlv->interval);
@@ -500,13 +509,13 @@ babel_read_update(struct babel_tlv *hdr, union babel_msg *m,
     msg->plen = tlv->plen;
     msg->prefix = ipa_from_ip6(get_ip6(buf));
 
-    if (tlv->flags & BABEL_FLAG_DEF_PREFIX)
+    if (tlv->flags & BABEL_UF_DEF_PREFIX)
     {
       put_ip6(state->def_ip6_prefix, msg->prefix);
       state->def_ip6_prefix_seen = 1;
     }
 
-    if (tlv->flags & BABEL_FLAG_ROUTER_ID)
+    if (tlv->flags & BABEL_UF_ROUTER_ID)
     {
       state->router_id = ((u64) _I2(msg->prefix)) << 32 | _I3(msg->prefix);
       state->router_id_seen = 1;

From 153f02da3bce1f3f1a99295648679c71327e8319 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Thu, 7 Dec 2017 13:06:01 +0100
Subject: [PATCH 041/198] Nest: Maintain separate IPv4, IPv6 and LLv6 preferred
 addresses

Also redesign preferred address selection and update protocols to use
appropriate preferred address.

Based on a previous work by Jan Maria Matejka.
---
 nest/iface.c           | 226 ++++++++++++++++++++++++++++-------------
 nest/iface.h           |  20 +++-
 nest/rt-dev.c          |   2 +-
 proto/babel/babel.c    |  47 ++++-----
 proto/babel/packets.c  |   3 +-
 proto/bgp/bgp.c        |  15 +--
 proto/radv/packets.c   |   2 +-
 proto/radv/radv.c      |  41 ++++----
 proto/rip/packets.c    |  12 +--
 proto/rip/rip.c        |  18 +++-
 sysdep/bsd/krt-sock.c  |  41 +++-----
 sysdep/bsd/sysio.h     |   4 +-
 sysdep/linux/netlink.c |   6 ++
 sysdep/unix/krt.Y      |  44 ++++++--
 sysdep/unix/krt.c      |  74 +++-----------
 sysdep/unix/krt.h      |  20 ++--
 sysdep/unix/unix.h     |   3 +
 17 files changed, 325 insertions(+), 253 deletions(-)

diff --git a/nest/iface.c b/nest/iface.c
index f14e70a0..01b1aa48 100644
--- a/nest/iface.c
+++ b/nest/iface.c
@@ -32,11 +32,14 @@
 #include "lib/resource.h"
 #include "lib/string.h"
 #include "conf/conf.h"
+#include "sysdep/unix/krt.h"
 
 static pool *if_pool;
 
 list iface_list;
 
+static void if_recalc_preferred(struct iface *i);
+
 /**
  * ifa_dump - dump interface address
  * @a: interface address descriptor
@@ -183,6 +186,7 @@ if_send_notify(struct proto *p, unsigned c, struct iface *i)
 	    (c & IF_CHANGE_DOWN) ? "goes down" :
 	    (c & IF_CHANGE_MTU) ? "changes MTU" :
 	    (c & IF_CHANGE_LINK) ? "changes link" :
+	    (c & IF_CHANGE_PREFERRED) ? "changes preferred address" :
 	    (c & IF_CHANGE_CREATE) ? "created" :
 	    "sends unknown event");
       p->if_notify(p, c, i);
@@ -211,20 +215,14 @@ if_notify_change(unsigned c, struct iface *i)
 
   if (c & IF_CHANGE_DOWN)
     WALK_LIST(a, i->addrs)
-      {
-	a->flags = (i->flags & ~IA_FLAGS) | (a->flags & IA_FLAGS);
-	ifa_notify_change_(IF_CHANGE_DOWN, a);
-      }
+      ifa_notify_change_(IF_CHANGE_DOWN, a);
 
   WALK_LIST(p, proto_list)
     if_send_notify(p, c, i);
 
   if (c & IF_CHANGE_UP)
     WALK_LIST(a, i->addrs)
-      {
-	a->flags = (i->flags & ~IA_FLAGS) | (a->flags & IA_FLAGS);
-	ifa_notify_change_(IF_CHANGE_UP, a);
-      }
+      ifa_notify_change_(IF_CHANGE_UP, a);
 
   if (c & IF_CHANGE_UP)
     neigh_if_up(i);
@@ -233,24 +231,23 @@ if_notify_change(unsigned c, struct iface *i)
     neigh_if_link(i);
 }
 
-static unsigned
-if_recalc_flags(struct iface *i, unsigned flags)
+static uint
+if_recalc_flags(struct iface *i UNUSED, uint flags)
 {
-  if ((flags & (IF_SHUTDOWN | IF_TMP_DOWN)) ||
-      !(flags & IF_ADMIN_UP) ||
-      !i->addr)
-    flags &= ~IF_UP;
-  else
+  if ((flags & IF_ADMIN_UP) && !(flags & (IF_SHUTDOWN | IF_TMP_DOWN)))
     flags |= IF_UP;
+  else
+    flags &= ~IF_UP;
+
   return flags;
 }
 
 static void
-if_change_flags(struct iface *i, unsigned flags)
+if_change_flags(struct iface *i, uint flags)
 {
-  unsigned of = i->flags;
-
+  uint of = i->flags;
   i->flags = if_recalc_flags(i, flags);
+
   if ((i->flags ^ of) & IF_UP)
     if_notify_change((i->flags & IF_UP) ? IF_CHANGE_UP : IF_CHANGE_DOWN, i);
 }
@@ -298,7 +295,6 @@ if_update(struct iface *new)
   WALK_LIST(i, iface_list)
     if (!strcmp(new->name, i->name))
       {
-	new->addr = i->addr;
 	new->flags = if_recalc_flags(new, new->flags);
 	c = if_what_changed(i, new);
 	if (c & IF_CHANGE_TOO_MUCH)	/* Changed a lot, convert it to down/up */
@@ -306,10 +302,13 @@ if_update(struct iface *new)
 	    DBG("Interface %s changed too much -- forcing down/up transition\n", i->name);
 	    if_change_flags(i, i->flags | IF_TMP_DOWN);
 	    rem_node(&i->n);
-	    new->addr = i->addr;
+	    new->addr4 = i->addr4;
+	    new->addr6 = i->addr6;
+	    new->llv6 = i->llv6;
+	    new->sysdep = i->sysdep;
 	    memcpy(&new->addrs, &i->addrs, sizeof(i->addrs));
 	    memcpy(i, new, sizeof(*i));
-	    i->flags &= ~IF_UP; /* IF_TMP_DOWN will be added later */
+	    i->flags &= ~IF_UP; 	/* IF_TMP_DOWN will be added later */
 	    goto newif;
 	  }
 
@@ -340,13 +339,16 @@ if_start_update(void)
     {
       i->flags &= ~IF_UPDATED;
       WALK_LIST(a, i->addrs)
-	a->flags &= ~IF_UPDATED;
+	a->flags &= ~IA_UPDATED;
     }
 }
 
 void
 if_end_partial_update(struct iface *i)
 {
+  if (i->flags & IF_NEEDS_RECALC)
+    if_recalc_preferred(i);
+
   if (i->flags & IF_TMP_DOWN)
     if_change_flags(i, i->flags & ~IF_TMP_DOWN);
 }
@@ -364,7 +366,7 @@ if_end_update(void)
       else
 	{
 	  WALK_LIST_DELSAFE(a, b, i->addrs)
-	    if (!(a->flags & IF_UPDATED))
+	    if (!(a->flags & IA_UPDATED))
 	      ifa_delete(a);
 	  if_end_partial_update(i);
 	}
@@ -461,40 +463,99 @@ if_get_by_name(char *name)
   return i;
 }
 
-struct ifa *kif_choose_primary(struct iface *i);
-
-static int
-ifa_recalc_primary(struct iface *i)
+static inline void
+if_set_preferred(struct ifa **pos, struct ifa *new)
 {
-  struct ifa *a = kif_choose_primary(i);
+  if (*pos)
+    (*pos)->flags &= ~IA_PRIMARY;
+  if (new)
+    new->flags |= IA_PRIMARY;
 
-  if (a == i->addr)
-    return 0;
+  *pos = new;
+}
 
-  if (i->addr)
-    i->addr->flags &= ~IA_PRIMARY;
+static void
+if_recalc_preferred(struct iface *i)
+{
+  /*
+   * Preferred address selection priority:
+   * 1) Address configured in Device protocol
+   * 2) Sysdep IPv4 address (BSD)
+   * 3) Old preferred address
+   * 4) First address in list
+   */
 
-  if (a)
+  struct kif_iface_config *ic = kif_get_iface_config(i);
+  struct ifa *a4 = i->addr4, *a6 = i->addr6, *ll = i->llv6;
+  ip_addr pref_v4 = ic->pref_v4;
+  uint change = 0;
+
+  if (kif_update_sysdep_addr(i))
+    change |= IF_CHANGE_SYSDEP;
+
+  /* BSD sysdep address */
+  if (ipa_zero(pref_v4) && ip4_nonzero(i->sysdep))
+    pref_v4 = ipa_from_ip4(i->sysdep);
+
+  struct ifa *a;
+  WALK_LIST(a, i->addrs)
     {
-      a->flags |= IA_PRIMARY;
-      rem_node(&a->n);
-      add_head(&i->addrs, &a->n);
+      /* Secondary address is never selected */
+      if (a->flags & IA_SECONDARY)
+	continue;
+
+      if (ipa_is_ip4(a->ip)) {
+	if (!a4 || ipa_equal(a->ip, pref_v4))
+	  a4 = a;
+      } else if (!ipa_is_link_local(a->ip)) {
+	if (!a6 || ipa_equal(a->ip, ic->pref_v6))
+	  a6 = a;
+      } else {
+	if (!ll || ipa_equal(a->ip, ic->pref_ll))
+	  ll = a;
+      }
     }
 
-  i->addr = a;
-  return 1;
+  if (a4 != i->addr4)
+  {
+    if_set_preferred(&i->addr4, a4);
+    change |= IF_CHANGE_ADDR4;
+  }
+
+  if (a6 != i->addr6)
+  {
+    if_set_preferred(&i->addr6, a6);
+    change |= IF_CHANGE_ADDR6;
+  }
+
+  if (ll != i->llv6)
+  {
+    if_set_preferred(&i->llv6, ll);
+    change |= IF_CHANGE_LLV6;
+  }
+
+  i->flags &= ~IF_NEEDS_RECALC;
+
+  /*
+   * FIXME: There should be proper notification instead of iface restart:
+   * if_notify_change(change, i)
+   */
+  if (change)
+    if_change_flags(i, i->flags | IF_TMP_DOWN);
 }
 
 void
-ifa_recalc_all_primary_addresses(void)
+if_recalc_all_preferred_addresses(void)
 {
   struct iface *i;
 
   WALK_LIST(i, iface_list)
-    {
-      if (ifa_recalc_primary(i))
-	if_change_flags(i, i->flags | IF_TMP_DOWN);
-    }
+  {
+    if_recalc_preferred(i);
+
+    if (i->flags & IF_TMP_DOWN)
+      if_change_flags(i, i->flags & ~IF_TMP_DOWN);
+  }
 }
 
 static inline int
@@ -526,7 +587,7 @@ ifa_update(struct ifa *a)
 	    b->scope == a->scope &&
 	    !((b->flags ^ a->flags) & IA_PEER))
 	  {
-	    b->flags |= IF_UPDATED;
+	    b->flags |= IA_UPDATED;
 	    return b;
 	  }
 	ifa_delete(b);
@@ -534,15 +595,15 @@ ifa_update(struct ifa *a)
       }
 
   if ((a->prefix.type == NET_IP4) && (i->flags & IF_BROADCAST) && ipa_zero(a->brd))
-    log(L_ERR "Missing broadcast address for interface %s", i->name);
+    log(L_WARN "Missing broadcast address for interface %s", i->name);
 
   b = mb_alloc(if_pool, sizeof(struct ifa));
   memcpy(b, a, sizeof(struct ifa));
   add_tail(&i->addrs, &b->n);
-  b->flags = (i->flags & ~IA_FLAGS) | (a->flags & IA_FLAGS);
-  if (ifa_recalc_primary(i))
-    if_change_flags(i, i->flags | IF_TMP_DOWN);
-  if (b->flags & IF_UP)
+  b->flags |= IA_UPDATED;
+
+  i->flags |= IF_NEEDS_RECALC;
+  if (i->flags & IF_UP)
     ifa_notify_change(IF_CHANGE_CREATE | IF_CHANGE_UP, b);
   return b;
 }
@@ -565,16 +626,24 @@ ifa_delete(struct ifa *a)
     if (ifa_same(b, a))
       {
 	rem_node(&b->n);
-	if (b->flags & IF_UP)
-	  {
-	    b->flags &= ~IF_UP;
-	    ifa_notify_change(IF_CHANGE_DOWN, b);
-	  }
+
 	if (b->flags & IA_PRIMARY)
 	  {
-	    if_change_flags(i, i->flags | IF_TMP_DOWN);
-	    ifa_recalc_primary(i);
+	    /*
+	     * We unlink deleted preferred address and mark for recalculation.
+	     * FIXME: This could break if we make iface scan non-atomic, as
+	     * protocols still could use the freed address until they get
+	     * if_notify from preferred route recalculation.
+	     */
+	    if (b == i->addr4) i->addr4 = NULL;
+	    if (b == i->addr6) i->addr6 = NULL;
+	    if (b == i->llv6) i->llv6 = NULL;
+	    i->flags |= IF_NEEDS_RECALC;
 	  }
+
+	if (i->flags & IF_UP)
+	  ifa_notify_change(IF_CHANGE_DOWN, b);
+
 	mb_free(b);
 	return;
       }
@@ -741,16 +810,17 @@ iface_patts_equal(list *a, list *b, int (*comp)(struct iface_patt *, struct ifac
 static void
 if_show_addr(struct ifa *a)
 {
-  byte opp[IPA_MAX_TEXT_LENGTH + 16];
+  byte *flg, opp[IPA_MAX_TEXT_LENGTH + 16];
+
+  flg = (a->flags & IA_PRIMARY) ? "Preferred, " : (a->flags & IA_SECONDARY) ? "Secondary, " : "";
 
   if (ipa_nonzero(a->opposite))
-    bsprintf(opp, ", opposite %I", a->opposite);
+    bsprintf(opp, "opposite %I, ", a->opposite);
   else
     opp[0] = 0;
-  cli_msg(-1003, "\t%I/%d (%s%s, scope %s)",
-	  a->ip, a->prefix.pxlen,
-	  (a->flags & IA_PRIMARY) ? "Primary" : (a->flags & IA_SECONDARY) ? "Secondary" : "Unselected",
-	  opp, ip_scope_text(a->scope));
+
+  cli_msg(-1003, "\t%I/%d (%s%sscope %s)",
+	  a->ip, a->prefix.pxlen, flg, opp, ip_scope_text(a->scope));
 }
 
 void
@@ -765,7 +835,7 @@ if_show(void)
       if (i->flags & IF_SHUTDOWN)
 	continue;
 
-      cli_msg(-1001, "%s %s (index=%d)", i->name, (i->flags & IF_UP) ? "up" : "DOWN", i->index);
+      cli_msg(-1001, "%s %s (index=%d)", i->name, (i->flags & IF_UP) ? "Up" : "Down", i->index);
       if (!(i->flags & IF_MULTIACCESS))
 	type = "PtP";
       else
@@ -779,10 +849,13 @@ if_show(void)
 	      (i->flags & IF_LOOPBACK) ? " Loopback" : "",
 	      (i->flags & IF_IGNORE) ? " Ignored" : "",
 	      i->mtu);
-      if (i->addr)
-	if_show_addr(i->addr);
+
       WALK_LIST(a, i->addrs)
-	if (a != i->addr)
+	if (a->prefix.type == NET_IP4)
+	  if_show_addr(a);
+
+      WALK_LIST(a, i->addrs)
+	if (a->prefix.type == NET_IP6)
 	  if_show_addr(a);
     }
   cli_msg(0, "");
@@ -792,16 +865,25 @@ void
 if_show_summary(void)
 {
   struct iface *i;
-  byte addr[IPA_MAX_TEXT_LENGTH + 16];
 
-  cli_msg(-2005, "interface state address");
+  cli_msg(-2005,  "%-10s %-6s %-18s %s", "Interface", "State", "IPv4 address", "IPv6 address");
   WALK_LIST(i, iface_list)
     {
-      if (i->addr)
-	bsprintf(addr, "%I/%d", i->addr->ip, i->addr->prefix.pxlen);
+      byte a4[IPA_MAX_TEXT_LENGTH + 17];
+      byte a6[IPA_MAX_TEXT_LENGTH + 17];
+
+      if (i->addr4)
+	bsprintf(a4, "%I/%d", i->addr4->ip, i->addr4->prefix.pxlen);
       else
-	addr[0] = 0;
-      cli_msg(-1005, "%-9s %-5s %s", i->name, (i->flags & IF_UP) ? "up" : "DOWN", addr);
+	a4[0] = 0;
+
+      if (i->addr6)
+	bsprintf(a6, "%I/%d", i->addr6->ip, i->addr6->prefix.pxlen);
+      else
+	a6[0] = 0;
+
+      cli_msg(-1005, "%-10s %-6s %-18s %s",
+	      i->name, (i->flags & IF_UP) ? "Up" : "Down", a4, a6);
     }
   cli_msg(0, "");
 }
diff --git a/nest/iface.h b/nest/iface.h
index de5070d6..59b1253c 100644
--- a/nest/iface.h
+++ b/nest/iface.h
@@ -35,11 +35,14 @@ struct iface {
   unsigned mtu;
   unsigned index;			/* OS-dependent interface index */
   list addrs;				/* Addresses assigned to this interface */
-  struct ifa *addr;			/* Primary address */
+  struct ifa *addr4;			/* Primary address for IPv4 */
+  struct ifa *addr6;			/* Primary address for IPv6 */
+  struct ifa *llv6;			/* Primary link-local address for IPv6 */
+  ip4_addr sysdep;			/* Arbitrary IPv4 address for internal sysdep use */
   list neighbors;			/* All neighbors on this interface */
 };
 
-#define IF_UP 1				/* IF_ADMIN_UP and IP address known */
+#define IF_UP 1				/* Currently just IF_ADMIN_UP */
 #define IF_MULTIACCESS 2
 #define IF_BROADCAST 4
 #define IF_MULTICAST 8
@@ -70,7 +73,10 @@ struct iface {
 
 #define IF_JUST_CREATED 0x10000000	/* Send creation event as soon as possible */
 #define IF_TMP_DOWN 0x20000000		/* Temporary shutdown due to interface reconfiguration */
-#define IF_UPDATED 0x40000000		/* Touched in last scan */
+#define IF_UPDATED 0x40000000		/* Iface touched in last scan */
+#define IF_NEEDS_RECALC	0x80000000	/* Preferred address recalculation is needed */
+
+#define IA_UPDATED IF_UPDATED		/* Address touched in last scan */
 
 /* Interface change events */
 
@@ -79,8 +85,14 @@ struct iface {
 #define IF_CHANGE_MTU 4
 #define IF_CHANGE_CREATE 8		/* Seen this interface for the first time */
 #define IF_CHANGE_LINK 0x10
+#define IF_CHANGE_ADDR4	0x100		/* Change of iface->addr4 */
+#define IF_CHANGE_ADDR6	0x200		/* ... */
+#define IF_CHANGE_LLV6 0x400
+#define IF_CHANGE_SYSDEP 0x800
 #define IF_CHANGE_TOO_MUCH 0x40000000	/* Used internally */
 
+#define IF_CHANGE_PREFERRED (IF_CHANGE_ADDR4 | IF_CHANGE_ADDR6 | IF_CHANGE_LLV6)
+
 void if_init(void);
 void if_dump(struct iface *);
 void if_dump_all(void);
@@ -99,7 +111,7 @@ void if_feed_baby(struct proto *);
 struct iface *if_find_by_index(unsigned);
 struct iface *if_find_by_name(char *);
 struct iface *if_get_by_name(char *);
-void ifa_recalc_all_primary_addresses(void);
+void if_recalc_all_preferred_addresses(void);
 
 
 /* The Neighbor Cache */
diff --git a/nest/rt-dev.c b/nest/rt-dev.c
index 9993da24..b3d5bf97 100644
--- a/nest/rt-dev.c
+++ b/nest/rt-dev.c
@@ -33,7 +33,7 @@ dev_ifa_notify(struct proto *P, uint flags, struct ifa *ad)
   struct channel *c;
 
   if (!EMPTY_LIST(cf->iface_list) &&
-      !iface_patt_find(&cf->iface_list, ad->iface, ad->iface->addr))
+      !iface_patt_find(&cf->iface_list, ad->iface, ad))
     /* Empty list is automatically treated as "*" */
     return;
 
diff --git a/proto/babel/babel.c b/proto/babel/babel.c
index a0ec3629..23746155 100644
--- a/proto/babel/babel.c
+++ b/proto/babel/babel.c
@@ -1501,26 +1501,14 @@ babel_add_iface(struct babel_proto *p, struct iface *new, struct babel_iface_con
   ifa->cf = ic;
   ifa->pool = pool;
   ifa->ifname = new->name;
+  ifa->addr = new->llv6->ip;
 
   add_tail(&p->interfaces, NODE ifa);
 
-  ip_addr addr4 = IPA_NONE;
-  struct ifa *addr;
-  WALK_LIST(addr, new->addrs)
-  {
-    if (ipa_is_link_local(addr->ip))
-      ifa->addr = addr->ip;
-
-    if (ipa_zero(addr4) && ipa_is_ip4(addr->ip))
-      addr4 = addr->ip;
-  }
-
+  ip_addr addr4 = new->addr4 ? new->addr4->ip : IPA_NONE;
   ifa->next_hop_ip4 = ipa_nonzero(ic->next_hop_ip4) ? ic->next_hop_ip4 : addr4;
   ifa->next_hop_ip6 = ipa_nonzero(ic->next_hop_ip6) ? ic->next_hop_ip6 : ifa->addr;
 
-  if (ipa_zero(ifa->addr))
-    log(L_WARN "%s: Cannot find link-local addr on %s", p->p.name, new->name);
-
   if (ipa_zero(ifa->next_hop_ip4) && p->ip4_channel)
     log(L_WARN "%s: Cannot find IPv4 next hop addr on %s", p->p.name, new->name);
 
@@ -1576,6 +1564,10 @@ babel_if_notify(struct proto *P, unsigned flags, struct iface *iface)
     if (!(iface->flags & IF_MULTICAST))
       return;
 
+    /* Ignore ifaces without link-local address */
+    if (!iface->llv6)
+      return;
+
     if (ic)
       babel_add_iface(p, iface, ic);
 
@@ -1615,21 +1607,8 @@ babel_reconfigure_iface(struct babel_proto *p, struct babel_iface *ifa, struct b
 
   ifa->cf = new;
 
-  if (ipa_nonzero(new->next_hop_ip4))
-    ifa->next_hop_ip4 = new->next_hop_ip4;
-  else
-  {
-    ifa->next_hop_ip4 = IPA_NONE;
-
-    struct ifa *addr;
-    WALK_LIST(addr, ifa->iface->addrs)
-      if (ipa_is_ip4(addr->ip))
-      {
-	ifa->next_hop_ip4 = addr->ip;
-	break;
-      }
-  }
-
+  ip_addr addr4 = ifa->iface->addr4 ? ifa->iface->addr4->ip : IPA_NONE;
+  ifa->next_hop_ip4 = ipa_nonzero(new->next_hop_ip4) ? new->next_hop_ip4 : addr4;
   ifa->next_hop_ip6 = ipa_nonzero(new->next_hop_ip6) ? new->next_hop_ip6 : ifa->addr;
 
   if (ipa_zero(ifa->next_hop_ip4) && p->ip4_channel)
@@ -1660,7 +1639,15 @@ babel_reconfigure_ifaces(struct babel_proto *p, struct babel_config *cf)
 
   WALK_LIST(iface, iface_list)
   {
-    if (! (iface->flags & IF_UP))
+    if (!(iface->flags & IF_UP))
+      continue;
+
+    /* Ignore non-multicast ifaces */
+    if (!(iface->flags & IF_MULTICAST))
+      continue;
+
+    /* Ignore ifaces without link-local address */
+    if (!iface->llv6)
       continue;
 
     struct babel_iface *ifa = babel_find_iface(p, iface);
diff --git a/proto/babel/packets.c b/proto/babel/packets.c
index 78c133e0..d136efe8 100644
--- a/proto/babel/packets.c
+++ b/proto/babel/packets.c
@@ -1294,7 +1294,7 @@ babel_rx_hook(sock *sk, uint len)
       sk->iface->name, sk->faddr, sk->laddr);
 
   /* Silently ignore my own packets */
-  if (ipa_equal(ifa->iface->addr->ip, sk->faddr))
+  if (ipa_equal(sk->faddr, sk->saddr))
     return 1;
 
   if (!ipa_is_link_local(sk->faddr))
@@ -1329,6 +1329,7 @@ babel_open_socket(struct babel_iface *ifa)
   sk->sport = ifa->cf->port;
   sk->dport = ifa->cf->port;
   sk->iface = ifa->iface;
+  sk->saddr = ifa->addr;
 
   sk->rx_hook = babel_rx_hook;
   sk->tx_hook = babel_tx_hook;
diff --git a/proto/bgp/bgp.c b/proto/bgp/bgp.c
index cccced57..a8d5cf9d 100644
--- a/proto/bgp/bgp.c
+++ b/proto/bgp/bgp.c
@@ -1094,19 +1094,8 @@ bgp_start_neighbor(struct bgp_proto *p)
 
   if (ipa_is_link_local(p->source_addr))
     p->link_addr = p->source_addr;
-  else
-  {
-    /* Find some link-local address for given iface */
-    struct ifa *a;
-    WALK_LIST(a, p->neigh->iface->addrs)
-      if (a->scope == SCOPE_LINK)
-      {
-	p->link_addr = a->ip;
-	break;
-      }
-
-    DBG("%s: Selected link-local address %I\n", p->p.name, p->link_addr);
-  }
+  else if (p->neigh->iface->llv6)
+    p->link_addr = p->neigh->iface->llv6->ip;
 
   bgp_initiate(p);
 }
diff --git a/proto/radv/packets.c b/proto/radv/packets.c
index 15ca5738..9ea8feee 100644
--- a/proto/radv/packets.c
+++ b/proto/radv/packets.c
@@ -358,7 +358,7 @@ radv_rx_hook(sock *sk, uint size)
   if (sk->lifindex != sk->iface->index)
     return 1;
 
-  if (ipa_equal(sk->faddr, ifa->addr->ip))
+  if (ipa_equal(sk->faddr, sk->saddr))
     return 1;
 
   if (size < 8)
diff --git a/proto/radv/radv.c b/proto/radv/radv.c
index b4b1bc58..2c0a23ad 100644
--- a/proto/radv/radv.c
+++ b/proto/radv/radv.c
@@ -138,17 +138,6 @@ radv_iface_add(struct object_lock *lock)
   radv_iface_notify(ifa, RA_EV_INIT);
 }
 
-static inline struct ifa *
-find_lladdr(struct iface *iface)
-{
-  struct ifa *a;
-  WALK_LIST(a, iface->addrs)
-    if ((a->prefix.type == NET_IP6) && (a->scope == SCOPE_LINK))
-      return a;
-
-  return NULL;
-}
-
 static void
 radv_iface_new(struct radv_proto *p, struct iface *iface, struct radv_iface_config *cf)
 {
@@ -161,16 +150,10 @@ radv_iface_new(struct radv_proto *p, struct iface *iface, struct radv_iface_conf
   ifa->ra = p;
   ifa->cf = cf;
   ifa->iface = iface;
+  ifa->addr = iface->llv6;
 
   add_tail(&p->iface_list, NODE ifa);
 
-  ifa->addr = find_lladdr(iface);
-  if (!ifa->addr)
-  {
-    log(L_ERR "%s: Missing link-local address on interface %s", p->p.name, iface->name);
-    return;
-  }
-
   timer *tm = tm_new(pool);
   tm->hook = radv_timer;
   tm->data = ifa;
@@ -216,8 +199,15 @@ radv_if_notify(struct proto *P, unsigned flags, struct iface *iface)
 
   if (flags & IF_CHANGE_UP)
   {
-    struct radv_iface_config *ic = (struct radv_iface_config *)
-      iface_patt_find(&cf->patt_list, iface, NULL);
+    struct radv_iface_config *ic = (void *) iface_patt_find(&cf->patt_list, iface, NULL);
+
+    /* Ignore non-multicast ifaces */
+    if (!(iface->flags & IF_MULTICAST))
+      return;
+
+    /* Ignore ifaces without link-local address */
+    if (!iface->llv6)
+      return;
 
     if (ic)
       radv_iface_new(p, iface, ic);
@@ -395,6 +385,17 @@ radv_reconfigure(struct proto *P, struct proto_config *CF)
   struct iface *iface;
   WALK_LIST(iface, iface_list)
   {
+    if (!(iface->flags & IF_UP))
+      continue;
+
+    /* Ignore non-multicast ifaces */
+    if (!(iface->flags & IF_MULTICAST))
+      continue;
+
+    /* Ignore ifaces without link-local address */
+    if (!iface->llv6)
+      continue;
+
     struct radv_iface *ifa = radv_iface_find(p, iface);
     struct radv_iface_config *ic = (struct radv_iface_config *)
       iface_patt_find(&new->patt_list, iface, NULL);
diff --git a/proto/rip/packets.c b/proto/rip/packets.c
index e97809c8..de48fd0a 100644
--- a/proto/rip/packets.c
+++ b/proto/rip/packets.c
@@ -669,8 +669,7 @@ rip_rx_hook(sock *sk, uint len)
       sk->iface->name, sk->faddr, sk->laddr);
 
   /* Silently ignore my own packets */
-  /* FIXME: Better local address check */
-  if (ipa_equal(ifa->iface->addr->ip, sk->faddr))
+  if (ipa_equal(sk->faddr, sk->saddr))
     return 1;
 
   if (rip_is_ng(p) && !ipa_is_link_local(sk->faddr))
@@ -742,14 +741,7 @@ rip_open_socket(struct rip_iface *ifa)
   sk->sport = ifa->cf->port;
   sk->dport = ifa->cf->port;
   sk->iface = ifa->iface;
-
-  /*
-   * For RIPv2, we explicitly choose a primary address, mainly to ensure that
-   * RIP and BFD uses the same one. For RIPng, we left it to kernel, which
-   * should choose some link-local address based on the same scope rule.
-   */
-  if (rip_is_v2(p))
-    sk->saddr = ifa->iface->addr->ip;
+  sk->saddr = rip_is_v2(p) ? ifa->iface->addr4->ip : ifa->iface->llv6->ip;
 
   sk->rx_hook = rip_rx_hook;
   sk->tx_hook = rip_tx_hook;
diff --git a/proto/rip/rip.c b/proto/rip/rip.c
index 55fb47c5..820c5117 100644
--- a/proto/rip/rip.c
+++ b/proto/rip/rip.c
@@ -630,7 +630,13 @@ rip_add_iface(struct rip_proto *p, struct iface *iface, struct rip_iface_config
   else if (ic->mode == RIP_IM_MULTICAST)
     ifa->addr = rip_is_v2(p) ? IP4_RIP_ROUTERS : IP6_RIP_ROUTERS;
   else /* Broadcast */
-    ifa->addr = iface->addr->brd;
+    ifa->addr = iface->addr4->brd;
+  /*
+   * The above is just a workaround for BSD as it can't send broadcasts
+   * to 255.255.255.255. BSD systems need the network broadcast address instead.
+   *
+   * TODO: move this to sysdep code
+   */
 
   init_list(&ifa->neigh_list);
 
@@ -706,7 +712,11 @@ rip_reconfigure_ifaces(struct rip_proto *p, struct rip_config *cf)
 
   WALK_LIST(iface, iface_list)
   {
-    if (! (iface->flags & IF_UP))
+    if (!(iface->flags & IF_UP))
+      continue;
+
+    /* Ignore ifaces without appropriate address */
+    if (rip_is_v2(p) ? !iface->addr4 : !iface->llv6)
       continue;
 
     struct rip_iface *ifa = rip_find_iface(p, iface);
@@ -744,6 +754,10 @@ rip_if_notify(struct proto *P, unsigned flags, struct iface *iface)
   {
     struct rip_iface_config *ic = (void *) iface_patt_find(&cf->patt_list, iface, NULL);
 
+    /* Ignore ifaces without appropriate address */
+    if (rip_is_v2(p) ? !iface->addr4 : !iface->llv6)
+      return;
+
     if (ic)
       rip_add_iface(p, iface, ic);
 
diff --git a/sysdep/bsd/krt-sock.c b/sysdep/bsd/krt-sock.c
index df639816..be8b50d6 100644
--- a/sysdep/bsd/krt-sock.c
+++ b/sysdep/bsd/krt-sock.c
@@ -287,18 +287,21 @@ krt_send_route(struct krt_proto *p, int cmd, rte *e)
 #endif
   {
     /* Fallback for all other valid cases */
-    if (!i->addr)
-    {
-      log(L_ERR "KRT: interface %s has no IP addess", i->name);
-      return -1;
-    }
 
 #ifdef RTF_CLONING
     if (cmd == RTM_ADD && (i->flags & IF_MULTIACCESS) != IF_MULTIACCESS)	/* PTP */
       msg.rtm.rtm_flags |= RTF_CLONING;
 #endif
 
-    sockaddr_fill(&gate, ipa_is_ip4(i->addr->ip) ? AF_INET : AF_INET6, i->addr->ip, NULL, 0);
+    struct ifa *addr = (net->n.addr->type == NET_IP4) ? i->addr4 : (i->addr6 ?: i->llv6);
+
+    if (!addr)
+    {
+      log(L_ERR "KRT: interface %s has no IP addess", i->name);
+      return -1;
+    }
+
+    sockaddr_fill(&gate, af, addr->ip, i, 0);
     msg.rtm.rtm_addrs |= RTA_GATEWAY;
     break;
   }
@@ -1124,13 +1127,11 @@ kif_sys_shutdown(struct kif_proto *p)
   krt_buffer_release(&p->p);
 }
 
-
-struct ifa *
-kif_get_primary_ip(struct iface *i UNUSED)
+int
+kif_update_sysdep_addr(struct iface *i)
 {
-#if 0
   static int fd = -1;
-  
+
   if (fd < 0)
     fd = socket(AF_INET, SOCK_DGRAM, 0);
 
@@ -1140,20 +1141,10 @@ kif_get_primary_ip(struct iface *i UNUSED)
 
   int rv = ioctl(fd, SIOCGIFADDR, (char *) &ifr);
   if (rv < 0)
-    return NULL;
+    return 0;
 
-  ip_addr addr;
-  struct sockaddr_in *sin = (struct sockaddr_in *) &ifr.ifr_addr;
-  memcpy(&addr, &sin->sin_addr.s_addr, sizeof(ip_addr));
-  ipa_ntoh(addr);
+  ip4_addr old = i->sysdep;
+  i->sysdep = ip4_from_ipa(ipa_from_sa4(&ifr.ifr_addr);
 
-  struct ifa *a;
-  WALK_LIST(a, i->addrs)
-  {
-    if (ipa_equal(a->ip, addr))
-      return a;
-  }
-#endif
-
-  return NULL;
+  return !ip4_equal(i->sysdep, addr);
 }
diff --git a/sysdep/bsd/sysio.h b/sysdep/bsd/sysio.h
index 0e895e20..68296e65 100644
--- a/sysdep/bsd/sysio.h
+++ b/sysdep/bsd/sysio.h
@@ -38,12 +38,12 @@
  */
 
 #define INIT_MREQ4(maddr,ifa) \
-  { .imr_multiaddr = ipa_to_in4(maddr), .imr_interface = ipa_to_in4(ifa->addr->ip) }
+  { .imr_multiaddr = ipa_to_in4(maddr), .imr_interface = ip4_to_in4(ifa->sysdep) }
 
 static inline int
 sk_setup_multicast4(sock *s)
 {
-  struct in_addr ifa = ipa_to_in4(s->iface->addr->ip);
+  struct in_addr ifa = ip4_to_in4(s->iface->sysdep);
   u8 ttl = s->ttl;
   u8 n = 0;
 
diff --git a/sysdep/linux/netlink.c b/sysdep/linux/netlink.c
index 2b7b13fb..c9d5cdec 100644
--- a/sysdep/linux/netlink.c
+++ b/sysdep/linux/netlink.c
@@ -2030,3 +2030,9 @@ void
 kif_sys_shutdown(struct kif_proto *p UNUSED)
 {
 }
+
+int
+kif_update_sysdep_addr(struct iface *i UNUSED)
+{
+  return 0;
+}
diff --git a/sysdep/unix/krt.Y b/sysdep/unix/krt.Y
index b261c91e..3bf7da65 100644
--- a/sysdep/unix/krt.Y
+++ b/sysdep/unix/krt.Y
@@ -14,6 +14,7 @@ CF_DEFINES
 
 #define THIS_KRT ((struct krt_config *) this_proto)
 #define THIS_KIF ((struct kif_config *) this_proto)
+#define KIF_IFACE ((struct kif_iface_config *) this_ipatt)
 
 static void
 krt_set_merge_paths(struct channel_config *cc, uint merge, uint limit)
@@ -25,6 +26,17 @@ krt_set_merge_paths(struct channel_config *cc, uint merge, uint limit)
   cc->merge_limit = limit;
 }
 
+static void
+kif_set_preferred(ip_addr ip)
+{
+  if (ipa_is_ip4(ip))
+    KIF_IFACE->pref_v4 = ip;
+  else if (!ipa_is_link_local(ip))
+    KIF_IFACE->pref_v6 = ip;
+  else
+    KIF_IFACE->pref_ll = ip;
+}
+
 CF_DECLS
 
 CF_KEYWORDS(KERNEL, PERSIST, SCAN, TIME, LEARN, DEVICE, ROUTES, GRACEFUL, RESTART, KRT_SOURCE, KRT_METRIC, MERGE, PATHS)
@@ -88,18 +100,38 @@ CF_ADDTO(kif_proto, kif_proto kif_item ';')
 
 kif_item:
    proto_item
+ | INTERFACE kif_iface
  | SCAN TIME expr {
       /* Scan time of 0 means scan on startup only */
       THIS_KIF->scan_time = $3;
    }
- | PRIMARY opttext net_or_ipa {
-     struct kif_primary_item *kpi = cfg_alloc(sizeof (struct kif_primary_item));
-     kpi->pattern = $2;
-     kpi->addr = $3;
-     add_tail(&THIS_KIF->primary, &kpi->n);
-   }
  ;
 
+kif_iface_start:
+{
+  this_ipatt = cfg_allocz(sizeof(struct kif_iface_config));
+  add_tail(&THIS_KIF->iface_list, NODE this_ipatt);
+  init_list(&this_ipatt->ipn_list);
+}
+
+kif_iface_item:
+   PREFERRED ipa { kif_set_preferred($2); }
+ ;
+
+kif_iface_opts:
+   /* empty */
+ | kif_iface_opts kif_iface_item ';'
+ ;
+
+kif_iface_opt_list:
+   /* empty */
+ | '{' kif_iface_opts '}'
+ ;
+
+kif_iface:
+  kif_iface_start iface_patt_list_nopx kif_iface_opt_list;
+
+
 CF_ADDTO(dynamic_attr, KRT_SOURCE { $$ = f_new_dynamic_attr(EAF_TYPE_INT | EAF_TEMP, T_INT, EA_KRT_SOURCE); })
 CF_ADDTO(dynamic_attr, KRT_METRIC { $$ = f_new_dynamic_attr(EAF_TYPE_INT | EAF_TEMP, T_INT, EA_KRT_METRIC); })
 
diff --git a/sysdep/unix/krt.c b/sysdep/unix/krt.c
index f0241777..0349a09f 100644
--- a/sysdep/unix/krt.c
+++ b/sysdep/unix/krt.c
@@ -89,6 +89,16 @@ static struct kif_config *kif_cf;
 static timer *kif_scan_timer;
 static bird_clock_t kif_last_shot;
 
+static struct kif_iface_config kif_default_iface = {};
+
+struct kif_iface_config *
+kif_get_iface_config(struct iface *iface)
+{
+  struct kif_config *cf = (void *) (kif_proto->p.cf);
+  struct kif_iface_config *ic = (void *) iface_patt_find(&cf->iface_list, iface, NULL);
+  return ic ?: &kif_default_iface;
+}
+
 static void
 kif_scan(timer *t)
 {
@@ -116,57 +126,6 @@ kif_request_scan(void)
     tm_start(kif_scan_timer, 1);
 }
 
-static inline int
-prefer_addr(struct ifa *a, struct ifa *b)
-{
-  int sa = a->scope > SCOPE_LINK;
-  int sb = b->scope > SCOPE_LINK;
-
-  if (sa < sb)
-    return 0;
-  else if (sa > sb)
-    return 1;
-  else
-    return ipa_compare(a->ip, b->ip) < 0;
-}
-
-static inline struct ifa *
-find_preferred_ifa(struct iface *i, const net_addr *n)
-{
-  struct ifa *a, *b = NULL;
-
-  WALK_LIST(a, i->addrs)
-    {
-      if (!(a->flags & IA_SECONDARY) &&
-	  (!n || ipa_in_netX(a->ip, n)) &&
-	  (!b || prefer_addr(a, b)))
-	b = a;
-    }
-
-  return b;
-}
-
-struct ifa *
-kif_choose_primary(struct iface *i)
-{
-  struct kif_config *cf = (struct kif_config *) (kif_proto->p.cf);
-  struct kif_primary_item *it;
-  struct ifa *a;
-
-  WALK_LIST(it, cf->primary)
-    {
-      if (!it->pattern || patmatch(it->pattern, i->name))
-	if (a = find_preferred_ifa(i, &it->addr))
-	  return a;
-    }
-
-  if (a = kif_get_primary_ip(i))
-    return a;
-
-  return find_preferred_ifa(i, NULL);
-}
-
-
 static struct proto *
 kif_init(struct proto_config *c)
 {
@@ -224,15 +183,15 @@ kif_reconfigure(struct proto *p, struct proto_config *new)
       tm_start(kif_scan_timer, n->scan_time);
     }
 
-  if (!EMPTY_LIST(o->primary) || !EMPTY_LIST(n->primary))
+  if (!EMPTY_LIST(o->iface_list) || !EMPTY_LIST(n->iface_list))
     {
       /* This is hack, we have to update a configuration
        * to the new value just now, because it is used
-       * for recalculation of primary addresses.
+       * for recalculation of preferred addresses.
        */
       p->cf = new;
 
-      ifa_recalc_all_primary_addresses();
+      if_recalc_all_preferred_addresses();
     }
 
   return 1;
@@ -254,7 +213,7 @@ kif_init_config(int class)
 
   kif_cf = (struct kif_config *) proto_config_new(&proto_unix_iface, class);
   kif_cf->scan_time = 60;
-  init_list(&kif_cf->primary);
+  init_list(&kif_cf->iface_list);
 
   kif_sys_init_config(kif_cf);
   return (struct proto_config *) kif_cf;
@@ -266,14 +225,13 @@ kif_copy_config(struct proto_config *dest, struct proto_config *src)
   struct kif_config *d = (struct kif_config *) dest;
   struct kif_config *s = (struct kif_config *) src;
 
-  /* Copy primary addr list */
-  cfg_copy_list(&d->primary, &s->primary, sizeof(struct kif_primary_item));
+  /* Copy interface config list */
+  cfg_copy_list(&d->iface_list, &s->iface_list, sizeof(struct kif_iface_config));
 
   /* Fix sysdep parts */
   kif_sys_copy_config(d, s);
 }
 
-
 struct protocol proto_unix_iface = {
   .name = 		"Device",
   .template = 		"device%d",
diff --git a/sysdep/unix/krt.h b/sysdep/unix/krt.h
index cb404de3..089c97ad 100644
--- a/sysdep/unix/krt.h
+++ b/sysdep/unix/krt.h
@@ -94,17 +94,20 @@ void krt_got_route_async(struct krt_proto *p, struct rte *e, int new);
 
 extern struct protocol proto_unix_iface;
 
-struct kif_primary_item {
-  node n;
-  byte *pattern;
-  net_addr addr;
-};
-
 struct kif_config {
   struct proto_config c;
   struct kif_params sys;	/* Sysdep params */
+
+  list iface_list;		/* List of iface configs (struct kif_iface_config) */
   int scan_time;		/* How often we re-scan interfaces */
-  list primary;			/* Preferences for primary addresses (struct kif_primary_item) */
+};
+
+struct kif_iface_config {
+  struct iface_patt i;
+
+  ip_addr pref_v4;
+  ip_addr pref_v6;
+  ip_addr pref_ll;
 };
 
 struct kif_proto {
@@ -116,6 +119,7 @@ extern struct kif_proto *kif_proto;
 
 #define KIF_CF ((struct kif_config *)p->p.cf)
 
+struct kif_iface_config * kif_get_iface_config(struct iface *iface);
 struct proto_config * krt_init_config(int class);
 
 
@@ -150,6 +154,6 @@ void kif_sys_copy_config(struct kif_config *, struct kif_config *);
 
 void kif_do_scan(struct kif_proto *);
 
-struct ifa *kif_get_primary_ip(struct iface *i);
+int kif_update_sysdep_addr(struct iface *i);
 
 #endif
diff --git a/sysdep/unix/unix.h b/sysdep/unix/unix.h
index dcaab729..4b0fb005 100644
--- a/sysdep/unix/unix.h
+++ b/sysdep/unix/unix.h
@@ -80,6 +80,9 @@ static inline ip_addr ipa_from_sa(sockaddr *sa)
 static inline struct in_addr ipa_to_in4(ip_addr a)
 { return (struct in_addr) { htonl(ipa_to_u32(a)) }; }
 
+static inline struct in_addr ip4_to_in4(ip4_addr a)
+{ return (struct in_addr) { htonl(ip4_to_u32(a)) }; }
+
 static inline struct in6_addr ipa_to_in6(ip_addr a)
 { return (struct in6_addr) { .s6_addr32 = { htonl(_I0(a)), htonl(_I1(a)), htonl(_I2(a)), htonl(_I3(a)) } }; }
 

From abae1cc58b7c4ccf683651a6fcd51cfaecd03566 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Thu, 7 Dec 2017 13:28:24 +0100
Subject: [PATCH 042/198] KRT: Minor fix of last commit

---
 sysdep/unix/krt.h | 1 +
 1 file changed, 1 insertion(+)

diff --git a/sysdep/unix/krt.h b/sysdep/unix/krt.h
index 089c97ad..39e16a2a 100644
--- a/sysdep/unix/krt.h
+++ b/sysdep/unix/krt.h
@@ -15,6 +15,7 @@ struct krt_proto;
 struct kif_config;
 struct kif_proto;
 
+#include "nest/iface.h"
 #include "sysdep/config.h"
 #include CONFIG_INCLUDE_KRTSYS_H
 

From 7c454d918682c072a6ae6ad8e0cd8d35b9edd2aa Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Thu, 7 Dec 2017 13:44:00 +0100
Subject: [PATCH 043/198] BSD: Minor fix of penultimate commit

---
 sysdep/bsd/krt-sock.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/sysdep/bsd/krt-sock.c b/sysdep/bsd/krt-sock.c
index be8b50d6..efbfdd5e 100644
--- a/sysdep/bsd/krt-sock.c
+++ b/sysdep/bsd/krt-sock.c
@@ -1144,7 +1144,7 @@ kif_update_sysdep_addr(struct iface *i)
     return 0;
 
   ip4_addr old = i->sysdep;
-  i->sysdep = ip4_from_ipa(ipa_from_sa4(&ifr.ifr_addr);
+  i->sysdep = ipa_to_ip4(ipa_from_sa4(&ifr.ifr_addr));
 
-  return !ip4_equal(i->sysdep, addr);
+  return !ip4_equal(i->sysdep, old);
 }

From 534215a18fb3fb7ce5b26c9e6ec1fdb32bf22ae6 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 30 May 2017 19:12:35 +0200
Subject: [PATCH 044/198] Timers: Split microsecond timers from BFD code to lib

---
 lib/Makefile       |   2 +-
 lib/timer.c        | 207 +++++++++++++++++++++++++++++++++
 lib/timer.h        | 104 +++++++++++++++++
 proto/bfd/io.c     | 277 ++++-----------------------------------------
 proto/bfd/io.h     |  67 +----------
 sysdep/unix/io.c   |  42 +++++++
 sysdep/unix/main.c |   2 +
 7 files changed, 379 insertions(+), 322 deletions(-)
 create mode 100644 lib/timer.c
 create mode 100644 lib/timer.h

diff --git a/lib/Makefile b/lib/Makefile
index 0c352869..01f3114d 100644
--- a/lib/Makefile
+++ b/lib/Makefile
@@ -1,4 +1,4 @@
-src := bitops.c checksum.c event.c flowspec.c idm.c ip.c lists.c mac.c md5.c mempool.c net.c patmatch.c printf.c resource.c sha1.c sha256.c sha512.c slab.c slists.c tbf.c xmalloc.c
+src := bitops.c checksum.c event.c flowspec.c idm.c ip.c lists.c mac.c md5.c mempool.c net.c patmatch.c printf.c resource.c sha1.c sha256.c sha512.c slab.c slists.c tbf.c timer.c xmalloc.c
 obj := $(src-o-files)
 $(all-daemon)
 
diff --git a/lib/timer.c b/lib/timer.c
new file mode 100644
index 00000000..00ac4b03
--- /dev/null
+++ b/lib/timer.c
@@ -0,0 +1,207 @@
+/*
+ *	BIRD -- Timers
+ *
+ *	(c) 2013--2017 Ondrej Zajicek <santiago@crfreenet.org>
+ *	(c) 2013--2017 CZ.NIC z.s.p.o.
+ *
+ *	Can be freely distributed and used under the terms of the GNU GPL.
+ */
+
+
+#include <stdlib.h>
+
+#include "nest/bird.h"
+
+#include "lib/heap.h"
+#include "lib/resource.h"
+#include "lib/timer.h"
+
+
+struct timeloop main_timeloop;
+
+
+#ifdef USE_PTHREADS
+
+#include <pthread.h>
+
+/* Data accessed and modified from proto/bfd/io.c */
+pthread_key_t current_time_key;
+
+static inline struct timeloop *
+timeloop_current(void)
+{
+  return pthread_getspecific(current_time_key);
+}
+
+static inline void
+timeloop_init_current(void)
+{
+  pthread_key_create(&current_time_key, NULL);
+  pthread_setspecific(current_time_key, &main_timeloop);
+}
+
+void wakeup_kick_current(void);
+
+#else
+
+/* Just use main timelooop */
+static inline struct timeloop * timeloop_current(void) { return &main_timeloop; }
+static inline void timeloop_init_current(void) { }
+
+#endif
+
+btime
+current_time(void)
+{
+  return timeloop_current()->last_time;
+}
+
+
+#define TIMER_LESS(a,b)		((a)->expires < (b)->expires)
+#define TIMER_SWAP(heap,a,b,t)	(t = heap[a], heap[a] = heap[b], heap[b] = t, \
+				   heap[a]->index = (a), heap[b]->index = (b))
+
+
+static void
+tm2_free(resource *r)
+{
+  timer2 *t = (timer2 *) r;
+
+  tm2_stop(t);
+}
+
+static void
+tm2_dump(resource *r)
+{
+  timer2 *t = (timer2 *) r;
+
+  debug("(code %p, data %p, ", t->hook, t->data);
+  if (t->randomize)
+    debug("rand %d, ", t->randomize);
+  if (t->recurrent)
+    debug("recur %d, ", t->recurrent);
+  if (t->expires)
+    debug("expires in %d ms)\n", (t->expires - current_time()) TO_MS);
+  else
+    debug("inactive)\n");
+}
+
+
+static struct resclass tm2_class = {
+  "Timer",
+  sizeof(timer2),
+  tm2_free,
+  tm2_dump,
+  NULL,
+  NULL
+};
+
+timer2 *
+tm2_new(pool *p)
+{
+  timer2 *t = ralloc(p, &tm2_class);
+  t->index = -1;
+  return t;
+}
+
+void
+tm2_set(timer2 *t, btime when)
+{
+  struct timeloop *loop = timeloop_current();
+  uint tc = timers_count(loop);
+
+  if (!t->expires)
+  {
+    t->index = ++tc;
+    t->expires = when;
+    BUFFER_PUSH(loop->timers) = t;
+    HEAP_INSERT(loop->timers.data, tc, timer2 *, TIMER_LESS, TIMER_SWAP);
+  }
+  else if (t->expires < when)
+  {
+    t->expires = when;
+    HEAP_INCREASE(loop->timers.data, tc, timer2 *, TIMER_LESS, TIMER_SWAP, t->index);
+  }
+  else if (t->expires > when)
+  {
+    t->expires = when;
+    HEAP_DECREASE(loop->timers.data, tc, timer2 *, TIMER_LESS, TIMER_SWAP, t->index);
+  }
+
+#ifdef CONFIG_BFD
+  /* Hack to notify BFD loops */
+  if ((loop != &main_timeloop) && (t->index == 1))
+    wakeup_kick_current();
+#endif
+}
+
+void
+tm2_start(timer2 *t, btime after)
+{
+  tm2_set(t, current_time() + MAX(after, 0));
+}
+
+void
+tm2_stop(timer2 *t)
+{
+  if (!t->expires)
+    return;
+
+  struct timeloop *loop = timeloop_current();
+  uint tc = timers_count(loop);
+
+  HEAP_DELETE(loop->timers.data, tc, timer2 *, TIMER_LESS, TIMER_SWAP, t->index);
+  BUFFER_POP(loop->timers);
+
+  t->index = -1;
+  t->expires = 0;
+}
+
+void
+timers_init(struct timeloop *loop, pool *p)
+{
+  times_init(loop);
+
+  BUFFER_INIT(loop->timers, p, 4);
+  BUFFER_PUSH(loop->timers) = NULL;
+}
+
+void
+timers_fire(struct timeloop *loop)
+{
+  btime base_time;
+  timer2 *t;
+
+  times_update(loop);
+  base_time = loop->last_time;
+
+  while (t = timers_first(loop))
+  {
+    if (t->expires > base_time)
+      return;
+
+    if (t->recurrent)
+    {
+      btime when = t->expires + t->recurrent;
+
+      if (when <= loop->last_time)
+	when = loop->last_time + t->recurrent;
+
+      if (t->randomize)
+	when += random() % (t->randomize + 1);
+
+      tm2_set(t, when);
+    }
+    else
+      tm2_stop(t);
+
+    t->hook(t);
+  }
+}
+
+void
+timer_init(void)
+{
+  timers_init(&main_timeloop, &root_pool);
+  timeloop_init_current();
+}
diff --git a/lib/timer.h b/lib/timer.h
new file mode 100644
index 00000000..88c53547
--- /dev/null
+++ b/lib/timer.h
@@ -0,0 +1,104 @@
+/*
+ *	BIRD -- Timers
+ *
+ *	(c) 2013--2017 Ondrej Zajicek <santiago@crfreenet.org>
+ *	(c) 2013--2017 CZ.NIC z.s.p.o.
+ *
+ *	Can be freely distributed and used under the terms of the GNU GPL.
+ */
+
+#ifndef _BIRD_TIMER2_H_
+#define _BIRD_TIMER2_H_
+
+#include "nest/bird.h"
+#include "lib/buffer.h"
+#include "lib/resource.h"
+
+
+typedef struct timer2
+{
+  resource r;
+  void (*hook)(struct timer2 *);
+  void *data;
+
+  btime expires;			/* 0=inactive */
+  uint randomize;			/* Amount of randomization */
+  uint recurrent;			/* Timer recurrence */
+
+  int index;
+} timer2;
+
+struct timeloop
+{
+  BUFFER(timer2 *) timers;
+  btime last_time;
+  btime real_time;
+};
+
+static inline uint timers_count(struct timeloop *loop)
+{ return loop->timers.used - 1; }
+
+static inline timer2 *timers_first(struct timeloop *loop)
+{ return (loop->timers.used > 1) ? loop->timers.data[1] : NULL; }
+
+extern struct timeloop main_timeloop;
+
+btime current_time(void);
+
+timer2 *tm2_new(pool *p);
+void tm2_set(timer2 *t, btime when);
+void tm2_start(timer2 *t, btime after);
+void tm2_stop(timer2 *t);
+
+static inline int
+tm2_active(timer2 *t)
+{
+  return t->expires != 0;
+}
+
+static inline btime
+tm2_remains(timer2 *t)
+{
+  btime now = current_time();
+  return (t->expires > now) ? (t->expires - now) : 0;
+}
+
+static inline timer2 *
+tm2_new_init(pool *p, void (*hook)(struct timer2 *), void *data, uint rec, uint rand)
+{
+  timer2 *t = tm2_new(p);
+  t->hook = hook;
+  t->data = data;
+  t->recurrent = rec;
+  t->randomize = rand;
+  return t;
+}
+
+static inline void
+tm2_set_max(timer2 *t, btime when)
+{
+  if (when > t->expires)
+    tm2_set(t, when);
+}
+
+/*
+static inline void
+tm2_start_max(timer2 *t, btime after)
+{
+  btime rem = tm2_remains(t);
+  tm2_start(t, MAX_(rem, after));
+}
+*/
+
+/* In sysdep code */
+void times_init(struct timeloop *loop);
+void times_update(struct timeloop *loop);
+
+/* For I/O loop */
+void timers_init(struct timeloop *loop, pool *p);
+void timers_fire(struct timeloop *loop);
+
+void timer_init(void);
+
+
+#endif
diff --git a/proto/bfd/io.c b/proto/bfd/io.c
index 8f4f5007..ab846113 100644
--- a/proto/bfd/io.c
+++ b/proto/bfd/io.c
@@ -18,10 +18,10 @@
 #include "proto/bfd/io.h"
 
 #include "lib/buffer.h"
-#include "lib/heap.h"
 #include "lib/lists.h"
 #include "lib/resource.h"
 #include "lib/event.h"
+#include "lib/timer.h"
 #include "lib/socket.h"
 
 
@@ -31,16 +31,12 @@ struct birdloop
   pthread_t thread;
   pthread_mutex_t mutex;
 
-  btime last_time;
-  btime real_time;
-  u8 use_monotonic_clock;
-
   u8 stop_called;
   u8 poll_active;
   u8 wakeup_masked;
   int wakeup_fds[2];
 
-  BUFFER(timer2 *) timers;
+  struct timeloop time;
   list event_list;
   list sock_list;
   uint sock_num;
@@ -57,6 +53,7 @@ struct birdloop
  */
 
 static pthread_key_t current_loop_key;
+extern pthread_key_t current_time_key;
 
 static inline struct birdloop *
 birdloop_current(void)
@@ -68,6 +65,7 @@ static inline void
 birdloop_set_current(struct birdloop *loop)
 {
   pthread_setspecific(current_loop_key, loop);
+  pthread_setspecific(current_time_key, loop ? &loop->time : &main_timeloop);
 }
 
 static inline void
@@ -77,98 +75,6 @@ birdloop_init_current(void)
 }
 
 
-/*
- *	Time clock
- */
-
-static void times_update_alt(struct birdloop *loop);
-
-static void
-times_init(struct birdloop *loop)
-{
-  struct timespec ts;
-  int rv;
-
-  rv = clock_gettime(CLOCK_MONOTONIC, &ts);
-  if (rv < 0)
-  {
-    log(L_WARN "Monotonic clock is missing");
-
-    loop->use_monotonic_clock = 0;
-    loop->last_time = 0;
-    loop->real_time = 0;
-    times_update_alt(loop);
-    return;
-  }
-
-  if ((ts.tv_sec < 0) || (((s64) ts.tv_sec) > ((s64) 1 << 40)))
-    log(L_WARN "Monotonic clock is crazy");
-
-  loop->use_monotonic_clock = 1;
-  loop->last_time = ((s64) ts.tv_sec S) + (ts.tv_nsec / 1000);
-  loop->real_time = 0;
-}
-
-static void
-times_update_pri(struct birdloop *loop)
-{
-  struct timespec ts;
-  int rv;
-
-  rv = clock_gettime(CLOCK_MONOTONIC, &ts);
-  if (rv < 0)
-    die("clock_gettime: %m");
-
-  btime new_time = ((s64) ts.tv_sec S) + (ts.tv_nsec / 1000);
-
-  if (new_time < loop->last_time)
-    log(L_ERR "Monotonic clock is broken");
-
-  loop->last_time = new_time;
-  loop->real_time = 0;
-}
-
-static void
-times_update_alt(struct birdloop *loop)
-{
-  struct timeval tv;
-  int rv;
-
-  rv = gettimeofday(&tv, NULL);
-  if (rv < 0)
-    die("gettimeofday: %m");
-
-  btime new_time = ((s64) tv.tv_sec S) + tv.tv_usec;
-  btime delta = new_time - loop->real_time;
-
-  if ((delta < 0) || (delta > (60 S)))
-  {
-    if (loop->real_time)
-      log(L_WARN "Time jump, delta %d us", (int) delta);
-
-    delta = 100 MS;
-  }
-
-  loop->last_time += delta;
-  loop->real_time = new_time;
-}
-
-static void
-times_update(struct birdloop *loop)
-{
-  if (loop->use_monotonic_clock)
-    times_update_pri(loop);
-  else
-    times_update_alt(loop);
-}
-
-btime
-current_time(void)
-{
-  return birdloop_current()->last_time;
-}
-
-
 /*
  *	Wakeup code for birdloop
  */
@@ -238,7 +144,7 @@ wakeup_drain(struct birdloop *loop)
 }
 
 static inline void
-wakeup_do_kick(struct birdloop *loop) 
+wakeup_do_kick(struct birdloop *loop)
 {
   pipe_kick(loop->wakeup_fds[1]);
 }
@@ -252,6 +158,16 @@ wakeup_kick(struct birdloop *loop)
     loop->wakeup_masked = 2;
 }
 
+/* For notifications from outside */
+void
+wakeup_kick_current(void)
+{
+  struct birdloop *loop = birdloop_current();
+
+  if (loop && loop->poll_active)
+    wakeup_kick(loop);
+}
+
 
 /*
  *	Events
@@ -272,7 +188,7 @@ events_init(struct birdloop *loop)
 static void
 events_fire(struct birdloop *loop)
 {
-  times_update(loop);
+  times_update(&loop->time);
   ev_run_list(&loop->event_list);
 }
 
@@ -291,154 +207,6 @@ ev2_schedule(event *e)
 }
 
 
-/*
- *	Timers
- */
-
-#define TIMER_LESS(a,b)		((a)->expires < (b)->expires)
-#define TIMER_SWAP(heap,a,b,t)	(t = heap[a], heap[a] = heap[b], heap[b] = t, \
-				   heap[a]->index = (a), heap[b]->index = (b))
-
-static inline uint timers_count(struct birdloop *loop)
-{ return loop->timers.used - 1; }
-
-static inline timer2 *timers_first(struct birdloop *loop)
-{ return (loop->timers.used > 1) ? loop->timers.data[1] : NULL; }
-
-
-static void
-tm2_free(resource *r)
-{
-  timer2 *t = (timer2 *) r;
-
-  tm2_stop(t);
-}
-
-static void
-tm2_dump(resource *r)
-{
-  timer2 *t = (timer2 *) r;
-
-  debug("(code %p, data %p, ", t->hook, t->data);
-  if (t->randomize)
-    debug("rand %d, ", t->randomize);
-  if (t->recurrent)
-    debug("recur %d, ", t->recurrent);
-  if (t->expires)
-    debug("expires in %d ms)\n", (t->expires - current_time()) TO_MS);
-  else
-    debug("inactive)\n");
-}
-
-
-static struct resclass tm2_class = {
-  "Timer",
-  sizeof(timer2),
-  tm2_free,
-  tm2_dump,
-  NULL,
-  NULL
-};
-
-timer2 *
-tm2_new(pool *p)
-{
-  timer2 *t = ralloc(p, &tm2_class);
-  t->index = -1;
-  return t;
-}
-
-void
-tm2_set(timer2 *t, btime when)
-{
-  struct birdloop *loop = birdloop_current();
-  uint tc = timers_count(loop);
-
-  if (!t->expires)
-  {
-    t->index = ++tc;
-    t->expires = when;
-    BUFFER_PUSH(loop->timers) = t;
-    HEAP_INSERT(loop->timers.data, tc, timer2 *, TIMER_LESS, TIMER_SWAP);
-  }
-  else if (t->expires < when)
-  {
-    t->expires = when;
-    HEAP_INCREASE(loop->timers.data, tc, timer2 *, TIMER_LESS, TIMER_SWAP, t->index);
-  }
-  else if (t->expires > when)
-  {
-    t->expires = when;
-    HEAP_DECREASE(loop->timers.data, tc, timer2 *, TIMER_LESS, TIMER_SWAP, t->index);
-  }
-
-  if (loop->poll_active && (t->index == 1))
-    wakeup_kick(loop);
-}
-
-void
-tm2_start(timer2 *t, btime after)
-{
-  tm2_set(t, current_time() + MAX(after, 0));
-}
-
-void
-tm2_stop(timer2 *t)
-{
-  if (!t->expires)
-    return;
-
-  struct birdloop *loop = birdloop_current();
-  uint tc = timers_count(loop);
-
-  HEAP_DELETE(loop->timers.data, tc, timer2 *, TIMER_LESS, TIMER_SWAP, t->index);
-  BUFFER_POP(loop->timers);
-
-  t->index = -1;
-  t->expires = 0;
-}
-
-static void
-timers_init(struct birdloop *loop)
-{
-  BUFFER_INIT(loop->timers, loop->pool, 4);
-  BUFFER_PUSH(loop->timers) = NULL;
-}
-
-static void
-timers_fire(struct birdloop *loop)
-{
-  btime base_time;
-  timer2 *t;
-
-  times_update(loop);
-  base_time = loop->last_time;
-
-  while (t = timers_first(loop))
-  {
-    if (t->expires > base_time)
-      return;
-
-    if (t->recurrent)
-    {
-      btime when = t->expires + t->recurrent;
-      
-      if (when <= loop->last_time)
-	when = loop->last_time + t->recurrent;
-
-      if (t->randomize)
-	when += random() % (t->randomize + 1);
-
-      tm2_set(t, when);
-    }
-    else
-      tm2_stop(t);
-
-    t->hook(t);
-  }
-}
-
-
 /*
  *	Sockets
  */
@@ -586,7 +354,7 @@ sockets_fire(struct birdloop *loop)
   sock **psk = loop->poll_sk.data;
   int poll_num = loop->poll_fd.used - 1;
 
-  times_update(loop);
+  times_update(&loop->time);
 
   /* Last fd is internal wakeup fd */
   if (pfd[poll_num].revents & POLLIN)
@@ -634,11 +402,10 @@ birdloop_new(void)
   loop->pool = p;
   pthread_mutex_init(&loop->mutex, NULL);
 
-  times_init(loop);
   wakeup_init(loop);
 
   events_init(loop);
-  timers_init(loop);
+  timers_init(&loop->time, p);
   sockets_init(loop);
 
   return loop;
@@ -719,12 +486,12 @@ birdloop_main(void *arg)
   while (1)
   {
     events_fire(loop);
-    timers_fire(loop);
+    timers_fire(&loop->time);
 
-    times_update(loop);
+    times_update(&loop->time);
     if (events_waiting(loop))
       timeout = 0;
-    else if (t = timers_first(loop))
+    else if (t = timers_first(&loop->time))
       timeout = (tm2_remains(t) TO_MS) + 1;
     else
       timeout = -1;
@@ -756,7 +523,7 @@ birdloop_main(void *arg)
     if (rv)
       sockets_fire(loop);
 
-    timers_fire(loop);
+    timers_fire(&loop->time);
   }
 
   loop->stop_called = 0;
diff --git a/proto/bfd/io.h b/proto/bfd/io.h
index 45836f84..ec706e9a 100644
--- a/proto/bfd/io.h
+++ b/proto/bfd/io.h
@@ -11,80 +11,15 @@
 #include "lib/lists.h"
 #include "lib/resource.h"
 #include "lib/event.h"
+#include "lib/timer.h"
 #include "lib/socket.h"
-// #include "sysdep/unix/timer.h"
 
 
-typedef struct timer2
-{
-  resource r;
-  void (*hook)(struct timer2 *);
-  void *data;
-
-  btime expires;			/* 0=inactive */
-  uint randomize;			/* Amount of randomization */
-  uint recurrent;			/* Timer recurrence */
-
-  int index;
-} timer2;
-
-
-btime current_time(void);
-
 void ev2_schedule(event *e);
 
-
-timer2 *tm2_new(pool *p);
-void tm2_set(timer2 *t, btime when);
-void tm2_start(timer2 *t, btime after);
-void tm2_stop(timer2 *t);
-
-static inline int
-tm2_active(timer2 *t)
-{
-  return t->expires != 0;
-}
-
-static inline btime 
-tm2_remains(timer2 *t)
-{
-  btime now = current_time();
-  return (t->expires > now) ? (t->expires - now) : 0;
-}
-
-static inline timer2 *
-tm2_new_init(pool *p, void (*hook)(struct timer2 *), void *data, uint rec, uint rand)
-{
-  timer2 *t = tm2_new(p);
-  t->hook = hook;
-  t->data = data;
-  t->recurrent = rec;
-  t->randomize = rand;
-  return t;
-}
-
-static inline void
-tm2_set_max(timer2 *t, btime when)
-{
-  if (when > t->expires)
-    tm2_set(t, when);
-}
-
-/*
-static inline void
-tm2_start_max(timer2 *t, btime after)
-{
-  btime rem = tm2_remains(t);
-  tm2_start(t, MAX_(rem, after));
-}
-*/
-
-
 void sk_start(sock *s);
 void sk_stop(sock *s);
 
-
-
 struct birdloop *birdloop_new(void);
 void birdloop_start(struct birdloop *loop);
 void birdloop_stop(struct birdloop *loop);
diff --git a/sysdep/unix/io.c b/sysdep/unix/io.c
index 0cf48c9d..ebd380ba 100644
--- a/sysdep/unix/io.c
+++ b/sysdep/unix/io.c
@@ -37,6 +37,7 @@
 #include "sysdep/unix/timer.h"
 #include "lib/socket.h"
 #include "lib/event.h"
+#include "lib/timer.h"
 #include "lib/string.h"
 #include "nest/iface.h"
 
@@ -479,6 +480,47 @@ tm_format_datetime(char *x, struct timeformat *fmt_spec, bird_clock_t t)
 }
 
 
+/*
+ *	Time clock
+ */
+
+void
+times_init(struct timeloop *loop)
+{
+  struct timespec ts;
+  int rv;
+
+  rv = clock_gettime(CLOCK_MONOTONIC, &ts);
+  if (rv < 0)
+    die("Monotonic clock is missing");
+
+  if ((ts.tv_sec < 0) || (((s64) ts.tv_sec) > ((s64) 1 << 40)))
+    log(L_WARN "Monotonic clock is crazy");
+
+  loop->last_time = ((s64) ts.tv_sec S) + (ts.tv_nsec / 1000);
+  loop->real_time = 0;
+}
+
+void
+times_update(struct timeloop *loop)
+{
+  struct timespec ts;
+  int rv;
+
+  rv = clock_gettime(CLOCK_MONOTONIC, &ts);
+  if (rv < 0)
+    die("clock_gettime: %m");
+
+  btime new_time = ((s64) ts.tv_sec S) + (ts.tv_nsec / 1000);
+
+  if (new_time < loop->last_time)
+    log(L_ERR "Monotonic clock is broken");
+
+  loop->last_time = new_time;
+  loop->real_time = 0;
+}
+
+
 /**
  * DOC: Sockets
  *
diff --git a/sysdep/unix/main.c b/sysdep/unix/main.c
index c1b92b7e..396310fd 100644
--- a/sysdep/unix/main.c
+++ b/sysdep/unix/main.c
@@ -27,6 +27,7 @@
 #include "lib/resource.h"
 #include "lib/socket.h"
 #include "lib/event.h"
+#include "lib/timer.h"
 #include "lib/string.h"
 #include "nest/route.h"
 #include "nest/protocol.h"
@@ -820,6 +821,7 @@ main(int argc, char **argv)
   log_switch(debug_flag, NULL, NULL);
 
   resource_init();
+  timer_init();
   olock_init();
   io_init();
   rt_init();

From 28a7d3943ef915c405b3552ae06f639a86f4dc1e Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 31 May 2017 15:46:04 +0200
Subject: [PATCH 045/198] Timers: Integrate microsecond timers to the main loop

---
 lib/timer.c      |  6 ++++++
 sysdep/unix/io.c | 17 ++++++++++++++---
 2 files changed, 20 insertions(+), 3 deletions(-)

diff --git a/lib/timer.c b/lib/timer.c
index 00ac4b03..7ca0bdaa 100644
--- a/lib/timer.c
+++ b/lib/timer.c
@@ -166,6 +166,8 @@ timers_init(struct timeloop *loop, pool *p)
   BUFFER_PUSH(loop->timers) = NULL;
 }
 
+void io_log_event(void *hook, void *data);
+
 void
 timers_fire(struct timeloop *loop)
 {
@@ -195,6 +197,10 @@ timers_fire(struct timeloop *loop)
     else
       tm2_stop(t);
 
+    /* This is ugly hack, we want to log just timers executed from the main I/O loop */
+    if (loop == &main_timeloop)
+      io_log_event(t->hook, t->data);
+
     t->hook(t);
   }
 }
diff --git a/sysdep/unix/io.c b/sysdep/unix/io.c
index ebd380ba..13635586 100644
--- a/sysdep/unix/io.c
+++ b/sysdep/unix/io.c
@@ -2507,9 +2507,10 @@ static int short_loops = 0;
 void
 io_loop(void)
 {
-  int poll_tout;
+  int poll_tout, timeout;
   time_t tout;
   int nfds, events, pout;
+  timer2 *t;
   sock *s;
   node *n;
   int fdmax = 256;
@@ -2518,7 +2519,9 @@ io_loop(void)
   watchdog_start1();
   for(;;)
     {
+      times_update(&main_timeloop);
       events = ev_run_list(&global_event_list);
+      timers_fire(&main_timeloop);
     timers:
       update_times();
       tout = tm_first_shot();
@@ -2527,10 +2530,16 @@ io_loop(void)
 	  tm_shot();
 	  goto timers;
 	}
-      poll_tout = (events ? 0 : MIN(tout - now, 3)) * 1000; /* Time in milliseconds */
-
       io_close_event();
 
+      poll_tout = (events ? 0 : MIN(tout - now, 3)) * 1000; /* Time in milliseconds */
+      if (t = timers_first(&main_timeloop))
+      {
+	times_update(&main_timeloop);
+	timeout = (tm2_remains(t) TO_MS) + 1;
+	poll_tout = MIN(poll_tout, timeout);
+      }
+
       nfds = 0;
       WALK_LIST(n, sock_list)
 	{
@@ -2601,6 +2610,8 @@ io_loop(void)
 	}
       if (pout)
 	{
+	  times_update(&main_timeloop);
+
 	  /* guaranteed to be non-empty */
 	  current_sock = SKIP_BACK(sock, n, HEAD(sock_list));
 

From 025525266f6861437ca54aca2a86eb505a486baf Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Thu, 1 Jun 2017 12:33:20 +0200
Subject: [PATCH 046/198] Timers: Replace old timers with microsecond timers

The old timer interface is still kept, but implemented by new timers. The
plan is to switch from the old inteface to the new interface, then clean
it up.
---
 conf/conf.c           |  2 +-
 lib/birdlib.h         |  2 +-
 lib/tbf.c             |  1 +
 lib/timer.c           | 11 +++++++
 lib/timer.h           | 12 ++++---
 nest/cmds.c           |  2 +-
 nest/proto.c          |  8 ++---
 proto/babel/babel.c   |  4 +--
 proto/bgp/bgp.c       | 19 ++++++-----
 proto/bgp/bgp.h       | 16 ++++-----
 proto/ospf/iface.c    |  4 +--
 proto/ospf/neighbor.c |  2 +-
 proto/ospf/ospf.c     |  2 +-
 proto/radv/radv.c     |  7 +---
 proto/rip/rip.c       |  4 +--
 proto/rpki/rpki.c     | 10 +++---
 sysdep/unix/io.c      | 42 +++++++++++++-----------
 sysdep/unix/krt.c     |  6 ++--
 sysdep/unix/main.c    |  2 +-
 sysdep/unix/timer.h   | 75 +++++++++++++++----------------------------
 20 files changed, 112 insertions(+), 119 deletions(-)

diff --git a/conf/conf.c b/conf/conf.c
index e8c0dc67..c4933122 100644
--- a/conf/conf.c
+++ b/conf/conf.c
@@ -452,7 +452,7 @@ config_undo(void)
 extern void cmd_reconfig_undo_notify(void);
 
 static void
-config_timeout(struct timer *t UNUSED)
+config_timeout(timer *t UNUSED)
 {
   log(L_INFO "Config timeout expired, starting undo");
   cmd_reconfig_undo_notify();
diff --git a/lib/birdlib.h b/lib/birdlib.h
index 317b5202..a0826aab 100644
--- a/lib/birdlib.h
+++ b/lib/birdlib.h
@@ -9,7 +9,6 @@
 #ifndef _BIRD_BIRDLIB_H_
 #define _BIRD_BIRDLIB_H_
 
-#include "sysdep/unix/timer.h"
 #include "lib/alloca.h"
 
 /* Ugly structure offset handling macros */
@@ -70,6 +69,7 @@ static inline int u64_cmp(u64 i1, u64 i2)
 /* Microsecond time */
 
 typedef s64 btime;
+typedef s64 bird_clock_t;
 
 #define S_	*1000000
 #define MS_	*1000
diff --git a/lib/tbf.c b/lib/tbf.c
index 39e18e57..c1dafee8 100644
--- a/lib/tbf.c
+++ b/lib/tbf.c
@@ -8,6 +8,7 @@
  */
 
 #include "nest/bird.h"
+#include "lib/timer.h"
 
 void
 tbf_update(struct tbf *f)
diff --git a/lib/timer.c b/lib/timer.c
index 7ca0bdaa..2c08b353 100644
--- a/lib/timer.c
+++ b/lib/timer.c
@@ -56,6 +56,17 @@ current_time(void)
   return timeloop_current()->last_time;
 }
 
+btime
+current_real_time(void)
+{
+  struct timeloop *loop = timeloop_current();
+
+  if (!loop->real_time)
+    times_update_real_time(loop);
+
+  return loop->real_time;
+}
+
 
 #define TIMER_LESS(a,b)		((a)->expires < (b)->expires)
 #define TIMER_SWAP(heap,a,b,t)	(t = heap[a], heap[a] = heap[b], heap[b] = t, \
diff --git a/lib/timer.h b/lib/timer.h
index 88c53547..b70ac48d 100644
--- a/lib/timer.h
+++ b/lib/timer.h
@@ -44,6 +44,11 @@ static inline timer2 *timers_first(struct timeloop *loop)
 extern struct timeloop main_timeloop;
 
 btime current_time(void);
+btime current_real_time(void);
+
+#define now (current_time() TO_S)
+#define now_real (current_real_time() TO_S)
+extern btime boot_time;
 
 timer2 *tm2_new(pool *p);
 void tm2_set(timer2 *t, btime when);
@@ -59,8 +64,8 @@ tm2_active(timer2 *t)
 static inline btime
 tm2_remains(timer2 *t)
 {
-  btime now = current_time();
-  return (t->expires > now) ? (t->expires - now) : 0;
+  btime now_ = current_time();
+  return (t->expires > now_) ? (t->expires - now_) : 0;
 }
 
 static inline timer2 *
@@ -81,18 +86,17 @@ tm2_set_max(timer2 *t, btime when)
     tm2_set(t, when);
 }
 
-/*
 static inline void
 tm2_start_max(timer2 *t, btime after)
 {
   btime rem = tm2_remains(t);
   tm2_start(t, MAX_(rem, after));
 }
-*/
 
 /* In sysdep code */
 void times_init(struct timeloop *loop);
 void times_update(struct timeloop *loop);
+void times_update_real_time(struct timeloop *loop);
 
 /* For I/O loop */
 void timers_init(struct timeloop *loop, pool *p);
diff --git a/nest/cmds.c b/nest/cmds.c
index 2bc672ea..371e8877 100644
--- a/nest/cmds.c
+++ b/nest/cmds.c
@@ -28,7 +28,7 @@ cmd_show_status(void)
   tm_format_datetime(tim, &config->tf_base, now);
   cli_msg(-1011, "Router ID is %R", config->router_id);
   cli_msg(-1011, "Current server time is %s", tim);
-  tm_format_datetime(tim, &config->tf_base, boot_time);
+  tm_format_datetime(tim, &config->tf_base, boot_time TO_S);
   cli_msg(-1011, "Last reboot on %s", tim);
   tm_format_datetime(tim, &config->tf_base, config->load_time);
   cli_msg(-1011, "Last reconfiguration on %s", tim);
diff --git a/nest/proto.c b/nest/proto.c
index 361bb225..65375c35 100644
--- a/nest/proto.c
+++ b/nest/proto.c
@@ -43,7 +43,7 @@ static char *c_states[] = { "DOWN", "START", "UP", "FLUSHING" };
 
 extern struct protocol proto_unix_iface;
 
-static void proto_shutdown_loop(struct timer *);
+static void proto_shutdown_loop(timer *);
 static void proto_rethink_goal(struct proto *p);
 static char *proto_state_name(struct proto *p);
 static void channel_verify_limits(struct channel *c);
@@ -1046,7 +1046,7 @@ proto_rethink_goal(struct proto *p)
  *
  */
 
-static void graceful_restart_done(struct timer *t);
+static void graceful_restart_done(timer *t);
 
 /**
  * graceful_restart_recovery - request initial graceful restart recovery
@@ -1099,7 +1099,7 @@ graceful_restart_init(void)
  * restart wait timer fires (but there are still some locks).
  */
 static void
-graceful_restart_done(struct timer *t UNUSED)
+graceful_restart_done(timer *t UNUSED)
 {
   log(L_INFO "Graceful restart done");
   graceful_restart_state = GRS_DONE;
@@ -1298,7 +1298,7 @@ protos_build(void)
 int proto_restart;
 
 static void
-proto_shutdown_loop(struct timer *t UNUSED)
+proto_shutdown_loop(timer *t UNUSED)
 {
   struct proto *p, *p_next;
 
diff --git a/proto/babel/babel.c b/proto/babel/babel.c
index 23746155..bab2e8b3 100644
--- a/proto/babel/babel.c
+++ b/proto/babel/babel.c
@@ -1367,7 +1367,7 @@ babel_iface_timer(timer *t)
 static inline void
 babel_iface_kick_timer(struct babel_iface *ifa)
 {
-  if (ifa->timer->expires > (now + 1))
+  if (ifa->timer->expires TO_S > (now + 1))
     tm_start(ifa->timer, 1);
 }
 
@@ -1948,7 +1948,7 @@ babel_timer(timer *t)
 static inline void
 babel_kick_timer(struct babel_proto *p)
 {
-  if (p->timer->expires > (now + 1))
+  if (p->timer->expires TO_S > (now + 1))
     tm_start(p->timer, 1);
 }
 
diff --git a/proto/bgp/bgp.c b/proto/bgp/bgp.c
index a8d5cf9d..dba7c875 100644
--- a/proto/bgp/bgp.c
+++ b/proto/bgp/bgp.c
@@ -103,6 +103,8 @@
 
 #undef LOCAL_DEBUG
 
+#include <stdlib.h>
+
 #include "nest/bird.h"
 #include "nest/iface.h"
 #include "nest/protocol.h"
@@ -324,8 +326,8 @@ bgp_start_timer(timer *t, int value)
   if (value)
   {
     /* The randomization procedure is specified in RFC 1771: 9.2.3.3 */
-    t->randomize = value / 4;
-    tm_start(t, value - t->randomize);
+    int randomize = random() % ((value / 4) + 1);
+    tm_start(t, value - randomize);
   }
   else
     tm_stop(t);
@@ -2006,17 +2008,18 @@ bgp_show_proto_info(struct proto *P)
     struct bgp_conn *oc = &p->outgoing_conn;
 
     if ((p->start_state < BSS_CONNECT) &&
-	(p->startup_timer->expires))
+	(tm_active(p->startup_timer)))
       cli_msg(-1006, "    Error wait:       %d/%d",
-	      p->startup_timer->expires - now, p->startup_delay);
+	      (int) tm_remains(p->startup_timer), p->startup_delay);
 
     if ((oc->state == BS_ACTIVE) &&
-	(oc->connect_timer->expires))
+	(tm_active(oc->connect_timer)))
       cli_msg(-1006, "    Connect delay:    %d/%d",
-	      oc->connect_timer->expires - now, p->cf->connect_delay_time);
+	      (int) tm_remains(oc->connect_timer), p->cf->connect_delay_time);
 
-    if (p->gr_active_num && p->gr_timer->expires)
-      cli_msg(-1006, "    Restart timer:    %d/-", p->gr_timer->expires - now);
+    if (p->gr_active_num && tm_active(p->gr_timer))
+      cli_msg(-1006, "    Restart timer:    %d/-",
+	      (int) tm_remains(p->gr_timer));
   }
   else if (P->proto_state == PS_UP)
   {
diff --git a/proto/bgp/bgp.h b/proto/bgp/bgp.h
index 7ffcb68a..3b38c05f 100644
--- a/proto/bgp/bgp.h
+++ b/proto/bgp/bgp.h
@@ -210,10 +210,10 @@ struct bgp_conn {
 
   struct bgp_caps *local_caps;
   struct bgp_caps *remote_caps;
-  struct timer *connect_timer;
-  struct timer *hold_timer;
-  struct timer *keepalive_timer;
-  struct event *tx_ev;
+  timer *connect_timer;
+  timer *hold_timer;
+  timer *keepalive_timer;
+  event *tx_ev;
   u32 packets_to_send;			/* Bitmap of packet types to be sent */
   u32 channels_to_send;			/* Bitmap of channels with packets to be sent */
   u8 last_channel;			/* Channel used last time for TX */
@@ -254,9 +254,9 @@ struct bgp_proto {
   struct bfd_request *bfd_req;		/* BFD request, if BFD is used */
   ip_addr source_addr;			/* Local address used as an advertised next hop */
   ip_addr link_addr;			/* Link-local version of source_addr */
-  struct event *event;			/* Event for respawning and shutting process */
-  struct timer *startup_timer;		/* Timer used to delay protocol startup due to previous errors (startup_delay) */
-  struct timer *gr_timer;		/* Timer waiting for reestablishment after graceful restart */
+  event *event;				/* Event for respawning and shutting process */
+  timer *startup_timer;			/* Timer used to delay protocol startup due to previous errors (startup_delay) */
+  timer *gr_timer;			/* Timer waiting for reestablishment after graceful restart */
   unsigned startup_delay;		/* Time to delay protocol startup by due to errors */
   bird_clock_t last_proto_error;	/* Time of last error that leads to protocol stop */
   u8 last_error_class; 			/* Error class of last error */
@@ -422,7 +422,7 @@ extern struct linpool *bgp_linpool;
 extern struct linpool *bgp_linpool2;
 
 
-void bgp_start_timer(struct timer *t, int value);
+void bgp_start_timer(timer *t, int value);
 void bgp_check_config(struct bgp_config *c);
 void bgp_error(struct bgp_conn *c, unsigned code, unsigned subcode, byte *data, int len);
 void bgp_close_conn(struct bgp_conn *c);
diff --git a/proto/ospf/iface.c b/proto/ospf/iface.c
index 98d48aa1..d1f9365e 100644
--- a/proto/ospf/iface.c
+++ b/proto/ospf/iface.c
@@ -715,9 +715,9 @@ ospf_iface_change_timer(timer *tm, uint val)
   if (!tm)
     return;
 
-  tm->recurrent = val;
+  tm->recurrent = val S;
 
-  if (tm->expires)
+  if (tm_active(tm))
     tm_start(tm, val);
 }
 
diff --git a/proto/ospf/neighbor.c b/proto/ospf/neighbor.c
index b68ba6f4..64165e46 100644
--- a/proto/ospf/neighbor.c
+++ b/proto/ospf/neighbor.c
@@ -654,7 +654,7 @@ ospf_sh_neigh_info(struct ospf_neighbor *n)
   char etime[6];
   int exp, sec, min;
 
-  exp = n->inactim->expires - now;
+  exp = tm_remains(n->inactim);
   sec = exp % 60;
   min = exp / 60;
   if (min > 59)
diff --git a/proto/ospf/ospf.c b/proto/ospf/ospf.c
index 7ce6698e..a9081ed0 100644
--- a/proto/ospf/ospf.c
+++ b/proto/ospf/ospf.c
@@ -676,7 +676,7 @@ ospf_reconfigure(struct proto *P, struct proto_config *CF)
   p->asbr = new->asbr;
   p->ecmp = new->ecmp;
   p->tick = new->tick;
-  p->disp_timer->recurrent = p->tick;
+  p->disp_timer->recurrent = p->tick S;
   tm_start(p->disp_timer, 1);
 
   /* Mark all areas and ifaces */
diff --git a/proto/radv/radv.c b/proto/radv/radv.c
index 2c0a23ad..34a3f473 100644
--- a/proto/radv/radv.c
+++ b/proto/radv/radv.c
@@ -154,12 +154,7 @@ radv_iface_new(struct radv_proto *p, struct iface *iface, struct radv_iface_conf
 
   add_tail(&p->iface_list, NODE ifa);
 
-  timer *tm = tm_new(pool);
-  tm->hook = radv_timer;
-  tm->data = ifa;
-  tm->randomize = 0;
-  tm->recurrent = 0;
-  ifa->timer = tm;
+  ifa->timer = tm_new_set(pool, radv_timer, ifa, 0, 0);
 
   struct object_lock *lock = olock_new(pool);
   lock->addr = IPA_NONE;
diff --git a/proto/rip/rip.c b/proto/rip/rip.c
index 820c5117..2583b49b 100644
--- a/proto/rip/rip.c
+++ b/proto/rip/rip.c
@@ -902,7 +902,7 @@ rip_timer(timer *t)
 static inline void
 rip_kick_timer(struct rip_proto *p)
 {
-  if (p->timer->expires > (now + 1))
+  if (p->timer->expires TO_S > (now + 1))
     tm_start(p->timer, 1);	/* Or 100 ms */
 }
 
@@ -962,7 +962,7 @@ rip_iface_timer(timer *t)
 static inline void
 rip_iface_kick_timer(struct rip_iface *ifa)
 {
-  if (ifa->timer->expires > (now + 1))
+  if (ifa->timer->expires TO_S > (now + 1))
     tm_start(ifa->timer, 1);	/* Or 100 ms */
 }
 
diff --git a/proto/rpki/rpki.c b/proto/rpki/rpki.c
index 349d2f70..a331e45e 100644
--- a/proto/rpki/rpki.c
+++ b/proto/rpki/rpki.c
@@ -382,7 +382,7 @@ rpki_do_we_recv_prefix_pdu_in_last_seconds(struct rpki_cache *cache)
  * |End of Data| PDU and has run by some &ERROR is occurred.
  */
 static void
-rpki_refresh_hook(struct timer *tm)
+rpki_refresh_hook(timer *tm)
 {
   struct rpki_cache *cache = tm->data;
 
@@ -428,7 +428,7 @@ rpki_refresh_hook(struct timer *tm)
  * ends by reaching of &ESTABLISHED state again.
  */
 static void
-rpki_retry_hook(struct timer *tm)
+rpki_retry_hook(timer *tm)
 {
   struct rpki_cache *cache = tm->data;
 
@@ -473,7 +473,7 @@ rpki_retry_hook(struct timer *tm)
  * of the protocol.
  */
 static void
-rpki_expire_hook(struct timer *tm)
+rpki_expire_hook(timer *tm)
 {
   struct rpki_cache *cache = tm->data;
 
@@ -789,8 +789,8 @@ rpki_get_status(struct proto *P, byte *buf)
 static void
 rpki_show_proto_info_timer(const char *name, uint num, timer *t)
 {
-  if (t->expires)
-    cli_msg(-1006, "  %-17s %us (remains %us)", name, num, tm_remains(t));
+  if (tm_active(t))
+    cli_msg(-1006, "  %-17s %us (remains %ds)", name, num, (int) tm_remains(t));
   else
     cli_msg(-1006, "  %-17s ---", name);
 }
diff --git a/sysdep/unix/io.c b/sysdep/unix/io.c
index 13635586..d3506d75 100644
--- a/sysdep/unix/io.c
+++ b/sysdep/unix/io.c
@@ -122,11 +122,13 @@ tracked_fopen(pool *p, char *name, char *mode)
  * for the other fields see |timer.h|.
  */
 
+#if 0
 #define NEAR_TIMER_LIMIT 4
 
 static list near_timers, far_timers;
 static bird_clock_t first_far_timer = TIME_INFINITY;
 
+
 /* now must be different from 0, because 0 is a special value in timer->expires */
 bird_clock_t now = 1, now_real, boot_time;
 
@@ -183,7 +185,6 @@ init_times(void)
    log(L_WARN "Monotonic timer is missing");
 }
 
-
 static void
 tm_free(resource *r)
 {
@@ -382,6 +383,7 @@ tm_shot(void)
       t->hook(t);
     }
 }
+#endif
 
 /**
  * tm_parse_datetime - parse a date and time
@@ -484,6 +486,8 @@ tm_format_datetime(char *x, struct timeformat *fmt_spec, bird_clock_t t)
  *	Time clock
  */
 
+btime boot_time;
+
 void
 times_init(struct timeloop *loop)
 {
@@ -520,6 +524,19 @@ times_update(struct timeloop *loop)
   loop->real_time = 0;
 }
 
+void
+times_update_real_time(struct timeloop *loop)
+{
+  struct timespec ts;
+  int rv;
+
+  rv = clock_gettime(CLOCK_REALTIME, &ts);
+  if (rv < 0)
+    die("clock_gettime: %m");
+
+  loop->real_time = ((s64) ts.tv_sec S) + (ts.tv_nsec / 1000);
+}
+
 
 /**
  * DOC: Sockets
@@ -2349,9 +2366,6 @@ io_update_time(void)
   struct timespec ts;
   int rv;
 
-  if (!clock_monotonic_available)
-    return;
-
   /*
    * This is third time-tracking procedure (after update_times() above and
    * times_update() in BFD), dedicated to internal event log and latency
@@ -2490,14 +2504,12 @@ volatile int async_shutdown_flag;
 void
 io_init(void)
 {
-  init_list(&near_timers);
-  init_list(&far_timers);
   init_list(&sock_list);
   init_list(&global_event_list);
   krt_io_init();
-  init_times();
-  update_times();
-  boot_time = now;
+  // XXX init_times();
+  // XXX update_times();
+  boot_time = current_time();
   srandom((int) now_real);
 }
 
@@ -2508,7 +2520,6 @@ void
 io_loop(void)
 {
   int poll_tout, timeout;
-  time_t tout;
   int nfds, events, pout;
   timer2 *t;
   sock *s;
@@ -2522,17 +2533,10 @@ io_loop(void)
       times_update(&main_timeloop);
       events = ev_run_list(&global_event_list);
       timers_fire(&main_timeloop);
-    timers:
-      update_times();
-      tout = tm_first_shot();
-      if (tout <= now)
-	{
-	  tm_shot();
-	  goto timers;
-	}
       io_close_event();
 
-      poll_tout = (events ? 0 : MIN(tout - now, 3)) * 1000; /* Time in milliseconds */
+      // FIXME
+      poll_tout = (events ? 0 : 3000); /* Time in milliseconds */
       if (t = timers_first(&main_timeloop))
       {
 	times_update(&main_timeloop);
diff --git a/sysdep/unix/krt.c b/sysdep/unix/krt.c
index 0349a09f..a5ec8a39 100644
--- a/sysdep/unix/krt.c
+++ b/sysdep/unix/krt.c
@@ -122,7 +122,7 @@ kif_force_scan(void)
 void
 kif_request_scan(void)
 {
-  if (kif_proto && kif_scan_timer->expires > now)
+  if (kif_proto && (kif_scan_timer->expires TO_S > (now + 1)))
     tm_start(kif_scan_timer, 1);
 }
 
@@ -147,7 +147,7 @@ kif_start(struct proto *P)
   kif_scan_timer = tm_new(P->pool);
   kif_scan_timer->hook = kif_scan;
   kif_scan_timer->data = p;
-  kif_scan_timer->recurrent = KIF_CF->scan_time;
+  kif_scan_timer->recurrent = KIF_CF->scan_time S;
   kif_scan(kif_scan_timer);
   tm_start(kif_scan_timer, KIF_CF->scan_time);
 
@@ -178,7 +178,7 @@ kif_reconfigure(struct proto *p, struct proto_config *new)
   if (o->scan_time != n->scan_time)
     {
       tm_stop(kif_scan_timer);
-      kif_scan_timer->recurrent = n->scan_time;
+      kif_scan_timer->recurrent = n->scan_time S;
       kif_scan(kif_scan_timer);
       tm_start(kif_scan_timer, n->scan_time);
     }
diff --git a/sysdep/unix/main.c b/sysdep/unix/main.c
index 396310fd..6cab21ac 100644
--- a/sysdep/unix/main.c
+++ b/sysdep/unix/main.c
@@ -57,7 +57,7 @@ async_dump(void)
 
   rdump(&root_pool);
   sk_dump_all();
-  tm_dump_all();
+  // XXXX tm_dump_all();
   if_dump_all();
   neigh_dump_all();
   rta_dump_all();
diff --git a/sysdep/unix/timer.h b/sysdep/unix/timer.h
index aa3ed143..1c4f6e3b 100644
--- a/sysdep/unix/timer.h
+++ b/sysdep/unix/timer.h
@@ -11,58 +11,38 @@
 
 #include <time.h>
 
-#include "lib/resource.h"
+#include "lib/birdlib.h"
+#include "lib/timer.h"
 
-typedef time_t bird_clock_t;		/* Use instead of time_t */
 
-typedef struct timer {
-  resource r;
-  void (*hook)(struct timer *);
-  void *data;
-  uint randomize;			/* Amount of randomization */
-  uint recurrent;			/* Timer recurrence */
-  node n;				/* Internal link */
-  bird_clock_t expires;			/* 0=inactive */
-} timer;
+typedef struct timer2 timer;
 
-timer *tm_new(pool *);
-void tm_start(timer *, uint after);
-void tm_stop(timer *);
-void tm_dump_all(void);
+static inline timer *tm_new(pool *p)
+{ return (void *) tm2_new(p); }
 
-extern bird_clock_t now; 		/* Relative, monotonic time in seconds */
-extern bird_clock_t now_real;		/* Time in seconds since fixed known epoch */
-extern bird_clock_t boot_time;
+static inline void tm_start(timer *t, bird_clock_t after)
+{ tm2_start(t, after S_); }
 
-static inline int
-tm_active(timer *t)
-{
-  return t->expires != 0;
-}
+static inline void tm_stop(timer *t)
+{ tm2_stop(t); }
 
-static inline bird_clock_t
-tm_remains(timer *t)
-{
-  return t->expires ? t->expires - now : 0;
-}
+// void tm_dump_all(void);
 
-static inline void
-tm_start_max(timer *t, bird_clock_t after)
-{
-  bird_clock_t rem = tm_remains(t);
-  tm_start(t, (rem > after) ? rem : after);
-}
+//extern bird_clock_t now; 		/* Relative, monotonic time in seconds */
+//extern bird_clock_t now_real;		/* Time in seconds since fixed known epoch */
+//extern bird_clock_t boot_time;
 
-static inline timer *
-tm_new_set(pool *p, void (*hook)(struct timer *), void *data, uint rand, uint rec)
-{
-  timer *t = tm_new(p);
-  t->hook = hook;
-  t->data = data;
-  t->randomize = rand;
-  t->recurrent = rec;
-  return t;
-}
+static inline int tm_active(timer *t)
+{ return tm2_active(t); }
+
+static inline bird_clock_t tm_remains(timer *t)
+{ return tm2_remains(t) TO_S; }
+
+static inline void tm_start_max(timer *t, bird_clock_t after)
+{ tm2_start_max(t, after S_); }
+
+static inline timer * tm_new_set(pool *p, void (*hook)(timer *), void *data, uint rand, uint rec)
+{ return tm2_new_init(p, hook, data, rec S_, rand S_); }
 
 
 struct timeformat {
@@ -77,12 +57,7 @@ bird_clock_t tm_parse_datetime(char *);	/* Convert date to bird_clock_t */
 void
 tm_format_datetime(char *x, struct timeformat *fmt_spec, bird_clock_t t);
 
-#define TIME_T_IS_64BIT (sizeof(time_t) == 8)
-#define TIME_T_IS_SIGNED ((time_t) -1 < 0)
+#define TIME_INFINITY ((s64) 0x7fffffffffffffff)
 
-#define TIME_INFINITY							\
-  ((time_t) (TIME_T_IS_SIGNED ?						\
-	     (TIME_T_IS_64BIT ? 0x7fffffffffffffff : 0x7fffffff):	\
-	     (TIME_T_IS_64BIT ? 0xffffffffffffffff : 0xffffffff)))
 
 #endif

From f047271cb963c62663687d63b2f7cf8dd5edfbb7 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 6 Jun 2017 16:47:30 +0200
Subject: [PATCH 047/198] Timers: Parse and format functions for microsecond
 times

Date/time output (e.g. in logs, show commands) can use %f to specify
subsecond time. By default, millisecond precision is used in output.
---
 conf/conf.c          |   6 +-
 conf/conf.h          |   2 +-
 conf/confbase.Y      |  12 ++--
 doc/bird.sgml        |  28 ++++-----
 lib/timer.c          | 135 +++++++++++++++++++++++++++++++++++++++++++
 lib/timer.h          |  19 ++++++
 nest/cmds.c          |   6 +-
 nest/config.Y        |  39 +++++++++++--
 nest/password.c      |   9 ++-
 nest/password.h      |   4 +-
 nest/proto.c         |  10 ++--
 nest/protocol.h      |   4 +-
 nest/route.h         |   4 +-
 nest/rt-show.c       |   2 +-
 nest/rt-table.c      |  10 ++--
 proto/bfd/bfd.c      |   8 +--
 proto/bfd/bfd.h      |   2 +-
 sysdep/unix/config.Y |  26 +--------
 sysdep/unix/io.c     |  97 +------------------------------
 sysdep/unix/log.c    |   2 +-
 sysdep/unix/timer.h  |  12 ----
 21 files changed, 244 insertions(+), 193 deletions(-)

diff --git a/conf/conf.c b/conf/conf.c
index c4933122..a570fad5 100644
--- a/conf/conf.c
+++ b/conf/conf.c
@@ -102,9 +102,9 @@ config_alloc(const char *name)
   c->pool = p;
   c->mem = l;
   c->file_name = ndup;
-  c->load_time = now;
-  c->tf_route = c->tf_proto = (struct timeformat){"%T", "%F", 20*3600};
-  c->tf_base = c->tf_log = (struct timeformat){"%F %T", NULL, 0};
+  c->load_time = current_time();
+  c->tf_route = c->tf_proto = TM_ISO_SHORT_MS;
+  c->tf_base = c->tf_log = TM_ISO_LONG_MS;
   c->gr_wait = DEFAULT_GR_WAIT;
 
   return c;
diff --git a/conf/conf.h b/conf/conf.h
index af92f056..552d0120 100644
--- a/conf/conf.h
+++ b/conf/conf.h
@@ -57,7 +57,7 @@ struct config {
   struct config *fallback;		/* Link to regular config for CLI parsing */
   int obstacle_count;			/* Number of items blocking freeing of this config */
   int shutdown;				/* This is a pseudo-config for daemon shutdown */
-  bird_clock_t load_time;		/* When we've got this configuration */
+  btime load_time;			/* When we've got this configuration */
 };
 
 /* Please don't use these variables in protocols. Use proto_config->global instead. */
diff --git a/conf/confbase.Y b/conf/confbase.Y
index 901ca2b2..390041c4 100644
--- a/conf/confbase.Y
+++ b/conf/confbase.Y
@@ -14,7 +14,7 @@ CF_HDR
 #include "conf/conf.h"
 #include "lib/resource.h"
 #include "lib/socket.h"
-#include "sysdep/unix/timer.h"
+#include "lib/timer.h"
 #include "lib/string.h"
 #include "nest/protocol.h"
 #include "nest/iface.h"
@@ -60,7 +60,7 @@ CF_DECLS
   struct lsadb_show_data *ld;
   struct iface *iface;
   void *g;
-  bird_clock_t time;
+  btime time;
   struct f_prefix px;
   struct proto_spec ps;
   struct channel_limit cl;
@@ -81,7 +81,7 @@ CF_DECLS
 
 %type <i> expr bool pxlen4
 %type <i32> expr_us
-%type <time> datetime
+%type <time> time
 %type <a> ipa
 %type <net> net_ip4_ net_ip6_ net_ip6 net_ip_ net_ip net_or_ipa
 %type <net_ptr> net_ net_any net_vpn4_ net_vpn6_ net_vpn_ net_roa4_ net_roa6_ net_roa_
@@ -308,11 +308,11 @@ label_stack:
   }
 ;
 
-datetime:
+time:
    TEXT {
-     $$ = tm_parse_datetime($1);
+     $$ = tm_parse_time($1);
      if (!$$)
-       cf_error("Invalid date and time");
+       cf_error("Invalid date/time");
    }
  ;
 
diff --git a/doc/bird.sgml b/doc/bird.sgml
index 1b4b3d67..65efebb8 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -456,24 +456,26 @@ protocol rip {
 	used for other commands and <cf/log/ is used in a log file.
 
 	"<m/format1/" is a format string using <it/strftime(3)/ notation (see
-	<it/man strftime/ for details). <m/limit> and "<m/format2/" allow to
-	specify the second format string for times in past deeper than <m/limit/
- 	seconds. There are few shorthands: <cf/iso long/ is a ISO 8601 date/time
-	format (YYYY-MM-DD hh:mm:ss) that can be also specified using <cf/"%F %T"/.
+	<it/man strftime/ for details). It is extended to support sub-second
+	time part with variable precision (up to microseconds) using "%f"
+	conversion code (e.g., "%T.%3f" is hh:mm:ss.sss time). <m/limit/ and
+	"<m/format2/" allow to specify the second format string for times in
+	past deeper than <m/limit/ seconds.
+
+	There are several shorthands: <cf/iso long/ is a ISO 8601 date/time
+	format (YYYY-MM-DD hh:mm:ss) that can be also specified using <cf/"%F
+	%T"/. Similarly, <cf/iso long ms/ and <cf/iso long us/ are ISO 8601
+	date/time formats with millisecond or microsecond precision.
 	<cf/iso short/ is a variant of ISO 8601 that uses just the time format
 	(hh:mm:ss) for near times (up to 20 hours in the past) and the date
-	format (YYYY-MM-DD) for far times. This is a shorthand for
-	<cf/"%T" 72000 "%F"/.
+	format (YYYY-MM-DD) for far times. This is a shorthand for <cf/"%T"
+	72000 "%F"/. And there are also <cf/iso short ms/ and <cf/iso short us/
+	high-precision variants of that.
 
-	By default, BIRD uses the <cf/iso short/ format for <cf/route/ and
-	<cf/protocol/ times, and the <cf/iso long/ format for <cf/base/ and
+	By default, BIRD uses the <cf/iso short ms/ format for <cf/route/ and
+	<cf/protocol/ times, and the <cf/iso long ms/ format for <cf/base/ and
 	<cf/log/ times.
 
-	In pre-1.4.0 versions, BIRD used an short, ad-hoc format for <cf/route/
-	and <cf/protocol/ times, and a <cf/iso long/ similar format (DD-MM-YYYY
-	hh:mm:ss) for <cf/base/ and <cf/log/. These timeformats could be set by
-	<cf/old short/ and <cf/old long/ compatibility shorthands.
-
 	<tag><label id="opt-table">table <m/name/ [sorted]</tag>
 	Create a new routing table. The default routing table is created
 	implicitly, other routing tables have to be added by this command.
diff --git a/lib/timer.c b/lib/timer.c
index 2c08b353..3136a56b 100644
--- a/lib/timer.c
+++ b/lib/timer.c
@@ -8,6 +8,7 @@
  */
 
 
+#include <stdio.h>
 #include <stdlib.h>
 
 #include "nest/bird.h"
@@ -222,3 +223,137 @@ timer_init(void)
   timers_init(&main_timeloop, &root_pool);
   timeloop_init_current();
 }
+
+
+/**
+ * tm_parse_time - parse a date and time
+ * @x: time string
+ *
+ * tm_parse_time() takes a textual representation of a date and time
+ * (yyyy-mm-dd[ hh:mm:ss[.sss]]) and converts it to the corresponding value of
+ * type &btime.
+ */
+btime
+tm_parse_time(char *x)
+{
+  struct tm tm;
+  int usec, n1, n2, n3, r;
+
+  r = sscanf(x, "%d-%d-%d%n %d:%d:%d%n.%d%n",
+	     &tm.tm_year, &tm.tm_mon, &tm.tm_mday, &n1,
+	     &tm.tm_hour, &tm.tm_min, &tm.tm_sec, &n2,
+	     &usec, &n3);
+
+  if ((r == 3) && !x[n1])
+    tm.tm_hour = tm.tm_min = tm.tm_sec = usec = 0;
+  else if ((r == 6) && !x[n2])
+    usec = 0;
+  else if ((r == 7) && !x[n3])
+  {
+    /* Convert subsecond digits to proper precision */
+    int digits = n3 - n2 - 1;
+    if ((usec < 0) || (usec > 999999) || (digits < 1) || (digits > 6))
+      return 0;
+
+    while (digits++ < 6)
+      usec *= 10;
+  }
+  else
+    return 0;
+
+  tm.tm_mon--;
+  tm.tm_year -= 1900;
+  s64 ts = mktime(&tm);
+  if ((ts == (s64) (time_t) -1) || (ts < 0) || (ts > ((s64) 1 << 40)))
+    return 0;
+
+  return ts S + usec;
+}
+
+/**
+ * tm_format_time - convert date and time to textual representation
+ * @x: destination buffer of size %TM_DATETIME_BUFFER_SIZE
+ * @fmt: specification of resulting textual representation of the time
+ * @t: time
+ *
+ * This function formats the given relative time value @t to a textual
+ * date/time representation (dd-mm-yyyy hh:mm:ss) in real time.
+ */
+void
+tm_format_time(char *x, struct timeformat *fmt, btime t)
+{
+  btime dt = current_time() - t;
+  btime rt = current_real_time() - dt;
+  int v1 = !fmt->limit || (dt < fmt->limit);
+
+  tm_format_real_time(x, v1 ? fmt->fmt1 : fmt->fmt2, rt);
+}
+
+/* Replace %f in format string with usec scaled to requested precision */
+static int
+strfusec(char *buf, int size, const char *fmt, uint usec)
+{
+  char *str = buf;
+  int parity = 0;
+
+  while (*fmt)
+  {
+    if (!size)
+      return 0;
+
+    if ((fmt[0] == '%') && (!parity) &&
+	((fmt[1] == 'f') || (fmt[1] >= '1') && (fmt[1] <= '6') && (fmt[2] == 'f')))
+    {
+      int digits = (fmt[1] == 'f') ? 6 : (fmt[1] - '0');
+      uint d = digits, u = usec;
+
+      /* Convert microseconds to requested precision */
+      while (d++ < 6)
+	u /= 10;
+
+      int num = bsnprintf(str, size, "%0*u", digits, u);
+      if (num < 0)
+	return 0;
+
+      fmt += (fmt[1] == 'f') ? 2 : 3;
+      ADVANCE(str, size, num);
+    }
+    else
+    {
+      /* Handle '%%' expression */
+      parity = (*fmt == '%') ? !parity : 0;
+      *str++ = *fmt++;
+      size--;
+    }
+  }
+
+  if (!size)
+    return 0;
+
+  *str = 0;
+  return str - buf;
+}
+
+void
+tm_format_real_time(char *x, const char *fmt, btime t)
+{
+  s64 t1 = t TO_S;
+  s64 t2 = t - t1 S;
+
+  time_t ts = t1;
+  struct tm tm;
+  if (!localtime_r(&ts, &tm))
+    goto err;
+
+  byte tbuf[TM_DATETIME_BUFFER_SIZE];
+  if (!strfusec(tbuf, TM_DATETIME_BUFFER_SIZE, fmt, t2))
+    goto err;
+
+  if (!strftime(x, TM_DATETIME_BUFFER_SIZE, tbuf, &tm))
+    goto err;
+
+  return;
+
+err:
+  strcpy(x, "<error>");
+}
diff --git a/lib/timer.h b/lib/timer.h
index b70ac48d..61a2aa94 100644
--- a/lib/timer.h
+++ b/lib/timer.h
@@ -105,4 +105,23 @@ void timers_fire(struct timeloop *loop);
 void timer_init(void);
 
 
+struct timeformat {
+  char *fmt1, *fmt2;
+  btime limit;
+};
+
+#define TM_ISO_SHORT_S	(struct timeformat){"%T",     "%F", (s64) (20*3600) S_}
+#define TM_ISO_SHORT_MS	(struct timeformat){"%T.%3f", "%F", (s64) (20*3600) S_}
+#define TM_ISO_SHORT_US	(struct timeformat){"%T.%6f", "%F", (s64) (20*3600) S_}
+
+#define TM_ISO_LONG_S	(struct timeformat){"%F %T",     NULL, 0}
+#define TM_ISO_LONG_MS	(struct timeformat){"%F %T.%3f", NULL, 0}
+#define TM_ISO_LONG_US	(struct timeformat){"%F %T.%6f", NULL, 0}
+
+#define TM_DATETIME_BUFFER_SIZE 32	/* Buffer size required by tm_format_time() */
+
+btime tm_parse_time(char *x);
+void tm_format_time(char *x, struct timeformat *fmt, btime t);
+void tm_format_real_time(char *x, const char *fmt, btime t);
+
 #endif
diff --git a/nest/cmds.c b/nest/cmds.c
index 371e8877..ca601ef2 100644
--- a/nest/cmds.c
+++ b/nest/cmds.c
@@ -25,12 +25,12 @@ cmd_show_status(void)
   byte tim[TM_DATETIME_BUFFER_SIZE];
 
   cli_msg(-1000, "BIRD " BIRD_VERSION);
-  tm_format_datetime(tim, &config->tf_base, now);
+  tm_format_time(tim, &config->tf_base, current_time());
   cli_msg(-1011, "Router ID is %R", config->router_id);
   cli_msg(-1011, "Current server time is %s", tim);
-  tm_format_datetime(tim, &config->tf_base, boot_time TO_S);
+  tm_format_time(tim, &config->tf_base, boot_time);
   cli_msg(-1011, "Last reboot on %s", tim);
-  tm_format_datetime(tim, &config->tf_base, config->load_time);
+  tm_format_time(tim, &config->tf_base, config->load_time);
   cli_msg(-1011, "Last reconfiguration on %s", tim);
 
   graceful_restart_show_status();
diff --git a/nest/config.Y b/nest/config.Y
index 220fa8b0..ef29fb96 100644
--- a/nest/config.Y
+++ b/nest/config.Y
@@ -73,6 +73,7 @@ CF_KEYWORDS(ALGORITHM, KEYED, HMAC, MD5, SHA1, SHA256, SHA384, SHA512)
 CF_KEYWORDS(PRIMARY, STATS, COUNT, BY, FOR, COMMANDS, PREEXPORT, NOEXPORT, GENERATE)
 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(TIMEFORMAT, ISO, SHORT, LONG, ROUTE, PROTOCOL, BASE, LOG, S, MS, US)
 CF_KEYWORDS(GRACEFUL, RESTART, WAIT, MAX, FLUSH, AS)
 
 /* For r_args_channel */
@@ -295,6 +296,7 @@ limit_spec:
  | OFF { $$ = (struct channel_limit){}; }
  ;
 
+
 CF_ADDTO(conf, debug_default)
 
 debug_default:
@@ -304,6 +306,31 @@ debug_default:
 
 /* MRTDUMP PROTOCOLS is in systep/unix/config.Y */
 
+CF_ADDTO(conf, timeformat_base)
+
+timeformat_which:
+   ROUTE { $$ = &new_config->tf_route; }
+ | PROTOCOL { $$ = &new_config->tf_proto; }
+ | BASE { $$ = &new_config->tf_base; }
+ | LOG { $$ = &new_config->tf_log; }
+ ;
+
+timeformat_spec:
+   timeformat_which TEXT { *$1 = (struct timeformat){$2, NULL, 0}; }
+ | timeformat_which TEXT expr TEXT { *$1 = (struct timeformat){$2, $4, (s64) $3 S_}; }
+ | timeformat_which ISO SHORT    { *$1 = TM_ISO_SHORT_S; }
+ | timeformat_which ISO SHORT MS { *$1 = TM_ISO_SHORT_MS; }
+ | timeformat_which ISO SHORT US { *$1 = TM_ISO_SHORT_US; }
+ | timeformat_which ISO LONG    { *$1 = TM_ISO_LONG_S; }
+ | timeformat_which ISO LONG MS { *$1 = TM_ISO_LONG_MS; }
+ | timeformat_which ISO LONG US { *$1 = TM_ISO_LONG_US; }
+ ;
+
+timeformat_base:
+   TIMEFORMAT timeformat_spec ';'
+ ;
+
+
 /* Interface patterns */
 
 iface_patt_node_init:
@@ -462,12 +489,12 @@ password_item_begin:
 
 password_item_params:
    /* empty */ { }
- | GENERATE FROM datetime ';' password_item_params { this_p_item->genfrom = $3; }
- | GENERATE TO datetime ';' password_item_params { this_p_item->gento = $3; }
- | ACCEPT FROM datetime ';' password_item_params { this_p_item->accfrom = $3; }
- | ACCEPT TO datetime ';' password_item_params { this_p_item->accto = $3; }
- | FROM datetime ';' password_item_params { this_p_item->genfrom = this_p_item->accfrom = $2; }
- | TO datetime ';' password_item_params { this_p_item->gento = this_p_item->accto = $2; }
+ | GENERATE FROM time ';' password_item_params { this_p_item->genfrom = $3; }
+ | GENERATE TO time ';' password_item_params { this_p_item->gento = $3; }
+ | ACCEPT FROM time ';' password_item_params { this_p_item->accfrom = $3; }
+ | ACCEPT TO time ';' password_item_params { this_p_item->accto = $3; }
+ | FROM time ';' password_item_params { this_p_item->genfrom = this_p_item->accfrom = $2; }
+ | TO time ';' password_item_params { this_p_item->gento = this_p_item->accto = $2; }
  | ID expr ';' password_item_params { this_p_item->id = $2; if ($2 <= 0) cf_error("Password ID has to be greated than zero."); }
  | ALGORITHM password_algorithm ';' password_item_params { this_p_item->alg = $2; }
  ;
diff --git a/nest/password.c b/nest/password.c
index e4813741..72f23f5d 100644
--- a/nest/password.c
+++ b/nest/password.c
@@ -19,12 +19,13 @@ password_find(list *l, int first_fit)
 {
   struct password_item *pi;
   struct password_item *pf = NULL;
+  btime now_ = current_real_time();
 
   if (l)
   {
     WALK_LIST(pi, *l)
     {
-      if ((pi->genfrom < now_real) && (pi->gento > now_real))
+      if ((pi->genfrom < now_) && (pi->gento > now_))
       {
 	if (first_fit)
 	  return pi;
@@ -41,12 +42,13 @@ struct password_item *
 password_find_by_id(list *l, uint id)
 {
   struct password_item *pi;
+  btime now_ = current_real_time();
 
   if (!l)
     return NULL;
 
   WALK_LIST(pi, *l)
-    if ((pi->id == id) && (pi->accfrom <= now_real) && (now_real < pi->accto))
+    if ((pi->id == id) && (pi->accfrom <= now_) && (now_ < pi->accto))
       return pi;
 
   return NULL;
@@ -56,12 +58,13 @@ struct password_item *
 password_find_by_value(list *l, char *pass, uint size)
 {
   struct password_item *pi;
+  btime now_ = current_real_time();
 
   if (!l)
     return NULL;
 
   WALK_LIST(pi, *l)
-    if (password_verify(pi, pass, size) && (pi->accfrom <= now_real) && (now_real < pi->accto))
+    if (password_verify(pi, pass, size) && (pi->accfrom <= now_) && (now_ < pi->accto))
       return pi;
 
   return NULL;
diff --git a/nest/password.h b/nest/password.h
index 78244985..c4017848 100644
--- a/nest/password.h
+++ b/nest/password.h
@@ -10,15 +10,13 @@
 #ifndef PASSWORD_H
 #define PASSWORD_H
 
-#include "sysdep/unix/timer.h"
-
 struct password_item {
   node n;
   char *password;			/* Key data, null terminated */
   uint length;				/* Key length, without null */
   uint id;				/* Key ID */
   uint alg;				/* MAC algorithm */
-  bird_clock_t accfrom, accto, genfrom, gento;
+  btime accfrom, accto, genfrom, gento;
 };
 
 extern struct password_item *last_password_item;
diff --git a/nest/proto.c b/nest/proto.c
index 65375c35..3043b648 100644
--- a/nest/proto.c
+++ b/nest/proto.c
@@ -162,7 +162,7 @@ proto_add_channel(struct proto *p, struct channel_config *cf)
 
   c->channel_state = CS_DOWN;
   c->export_state = ES_DOWN;
-  c->last_state_change = now;
+  c->last_state_change = current_time();
   c->reloadable = 1;
 
   CALL(c->channel->init, c, cf);
@@ -341,7 +341,7 @@ channel_set_state(struct channel *c, uint state)
     return;
 
   c->channel_state = state;
-  c->last_state_change = now;
+  c->last_state_change = current_time();
 
   switch (state)
   {
@@ -672,7 +672,7 @@ proto_init(struct proto_config *c, node *n)
   struct proto *p = pr->init(c);
 
   p->proto_state = PS_DOWN;
-  p->last_state_change = now;
+  p->last_state_change = current_time();
   insert_node(&p->n, n);
 
   p->event = ev_new(proto_pool);
@@ -1500,7 +1500,7 @@ proto_notify_state(struct proto *p, uint state)
     return;
 
   p->proto_state = state;
-  p->last_state_change = now;
+  p->last_state_change = current_time();
 
   switch (state)
   {
@@ -1631,7 +1631,7 @@ proto_cmd_show(struct proto *p, uint verbose, int cnt)
   buf[0] = 0;
   if (p->proto->get_status)
     p->proto->get_status(p, buf);
-  tm_format_datetime(tbuf, &config->tf_proto, p->last_state_change);
+  tm_format_time(tbuf, &config->tf_proto, p->last_state_change);
   cli_msg(-1002, "%-8s %-8s %-8s %-5s  %-10s  %s",
 	  p->name,
 	  p->proto->name,
diff --git a/nest/protocol.h b/nest/protocol.h
index eed0a291..abd11aff 100644
--- a/nest/protocol.h
+++ b/nest/protocol.h
@@ -159,7 +159,7 @@ struct proto {
   byte down_sched;			/* Shutdown is scheduled for later (PDS_*) */
   byte down_code;			/* Reason for shutdown (PDC_* codes) */
   u32 hash_key;				/* Random key used for hashing of neighbors */
-  bird_clock_t last_state_change;	/* Time of last state transition */
+  btime last_state_change;		/* Time of last state transition */
   char *last_state_name_announced;	/* Last state name we've announced to the user */
 
   /*
@@ -508,7 +508,7 @@ struct channel {
   u8 gr_lock;				/* Graceful restart mechanism should wait for this channel */
   u8 gr_wait;				/* Route export to channel is postponed until graceful restart */
 
-  bird_clock_t last_state_change;	/* Time of last state transition */
+  btime last_state_change;		/* Time of last state transition */
 };
 
 
diff --git a/nest/route.h b/nest/route.h
index c1a60ccc..2aa2bcd8 100644
--- a/nest/route.h
+++ b/nest/route.h
@@ -159,8 +159,8 @@ typedef struct rtable {
 					 * obstacle from this routing table.
 					 */
   struct event *rt_event;		/* Routing table event */
+  btime gc_time;			/* Time of last GC */
   int gc_counter;			/* Number of operations since last GC */
-  bird_clock_t gc_time;			/* Time of last GC */
   byte prune_state;			/* Table prune state, 1 -> scheduled, 2-> running */
   byte hcu_scheduled;			/* Hostcache update is scheduled */
   byte nhu_state;			/* Next Hop Update state */
@@ -213,7 +213,7 @@ typedef struct rte {
   byte flags;				/* Flags (REF_...) */
   byte pflags;				/* Protocol-specific flags */
   word pref;				/* Route preference */
-  bird_clock_t lastmod;			/* Last modified */
+  btime lastmod;			/* Last modified */
   union {				/* Protocol-dependent data (metrics etc.) */
 #ifdef CONFIG_RIP
     struct {
diff --git a/nest/rt-show.c b/nest/rt-show.c
index afde2810..9989afa4 100644
--- a/nest/rt-show.c
+++ b/nest/rt-show.c
@@ -39,7 +39,7 @@ rt_show_rte(struct cli *c, byte *ia, rte *e, struct rt_show_data *d, ea_list *tm
   void (*get_route_info)(struct rte *, byte *buf, struct ea_list *attrs);
   struct nexthop *nh;
 
-  tm_format_datetime(tm, &config->tf_route, e->lastmod);
+  tm_format_time(tm, &config->tf_route, e->lastmod);
   if (ipa_nonzero(a->from) && !ipa_equal(a->from, a->nh.gw))
     bsprintf(from, " from %I", a->from);
   else
diff --git a/nest/rt-table.c b/nest/rt-table.c
index 2bb78cf2..c42d3a97 100644
--- a/nest/rt-table.c
+++ b/nest/rt-table.c
@@ -1173,7 +1173,7 @@ rte_recalculate(struct channel *c, net *net, rte *new, struct rte_src *src)
     }
 
   if (new)
-    new->lastmod = now;
+    new->lastmod = current_time();
 
   /* Log the route change */
   if (p->debug & D_ROUTES)
@@ -1201,7 +1201,7 @@ rte_recalculate(struct channel *c, net *net, rte *new, struct rte_src *src)
 
   if (!net->routes &&
       (table->gc_counter++ >= table->config->gc_max_ops) &&
-      (table->gc_time + table->config->gc_min_time <= now))
+      (table->gc_time + table->config->gc_min_time <= current_time()))
     rt_schedule_prune(table);
 
   if (old_ok && p->rte_remove)
@@ -1497,7 +1497,7 @@ rte_dump(rte *e)
 {
   net *n = e->net;
   debug("%-1N ", n->n.addr);
-  debug("KF=%02x PF=%02x pref=%d lm=%d ", n->n.flags, e->pflags, e->pref, now-e->lastmod);
+  debug("KF=%02x PF=%02x pref=%d ", n->n.flags, e->pflags, e->pref);
   rta_dump(e->attrs);
   if (e->attrs->src->proto->proto->dump_attrs)
     e->attrs->src->proto->proto->dump_attrs(e);
@@ -1609,7 +1609,7 @@ rt_setup(pool *p, rtable *t, char *name, struct rtable_config *cf)
       t->rt_event = ev_new(p);
       t->rt_event->hook = rt_event;
       t->rt_event->data = t;
-      t->gc_time = now;
+      t->gc_time = current_time();
     }
 }
 
@@ -1708,7 +1708,7 @@ again:
 #endif
 
   tab->gc_counter = 0;
-  tab->gc_time = now;
+  tab->gc_time = current_time();
 
   /* state change 2->0, 3->1 */
   tab->prune_state &= 1;
diff --git a/proto/bfd/bfd.c b/proto/bfd/bfd.c
index e2c6050b..2374cfe8 100644
--- a/proto/bfd/bfd.c
+++ b/proto/bfd/bfd.c
@@ -145,6 +145,7 @@ bfd_session_update_state(struct bfd_session *s, uint state, uint diag)
   bfd_lock_sessions(p);
   s->loc_state = state;
   s->loc_diag = diag;
+  s->last_state_change = current_time();
 
   notify = !NODE_VALID(&s->n);
   if (notify)
@@ -438,7 +439,7 @@ bfd_add_session(struct bfd_proto *p, ip_addr addr, ip_addr local, struct iface *
   bfd_session_control_tx_timer(s, 1);
 
   init_list(&s->request_list);
-  s->last_state_change = now;
+  s->last_state_change = current_time();
 
   TRACE(D_EVENTS, "Session to %I added", s->addr);
 
@@ -879,9 +880,6 @@ bfd_notify_hook(sock *sk, uint len UNUSED)
     diag = s->loc_diag;
     bfd_unlock_sessions(p);
 
-    /* FIXME: convert to btime and move to bfd_session_update_state() */
-    s->last_state_change = now;
-
     s->notify_running = 1;
     WALK_LIST_DELSAFE(n, nn, s->request_list)
       bfd_request_notify(SKIP_BACK(struct bfd_request, n, n), state, diag);
@@ -1105,7 +1103,7 @@ bfd_show_sessions(struct proto *P)
     timeout = (MAX(s->req_min_rx_int, s->rem_min_tx_int) TO_MS) * s->rem_detect_mult;
 
     state = (state < 4) ? state : 0;
-    tm_format_datetime(tbuf, &config->tf_proto, s->last_state_change);
+    tm_format_time(tbuf, &config->tf_proto, s->last_state_change);
 
     cli_msg(-1020, "%-25I %-10s %-10s %-10s  %3u.%03u  %3u.%03u",
 	    s->addr, ifname, bfd_state_names[state], tbuf,
diff --git a/proto/bfd/bfd.h b/proto/bfd/bfd.h
index e2e75753..203da437 100644
--- a/proto/bfd/bfd.h
+++ b/proto/bfd/bfd.h
@@ -144,7 +144,7 @@ struct bfd_session
   timer2 *hold_timer;			/* Timer for session down detection time */
 
   list request_list;			/* List of client requests (struct bfd_request) */
-  bird_clock_t last_state_change;	/* Time of last state change */
+  btime last_state_change;		/* Time of last state change */
   u8 notify_running;			/* 1 if notify hooks are running */
 
   u8 rx_csn_known;			/* Received crypto sequence number is known */
diff --git a/sysdep/unix/config.Y b/sysdep/unix/config.Y
index ebadd454..ccca4a62 100644
--- a/sysdep/unix/config.Y
+++ b/sysdep/unix/config.Y
@@ -14,8 +14,7 @@ CF_HDR
 CF_DECLS
 
 CF_KEYWORDS(LOG, SYSLOG, ALL, DEBUG, TRACE, INFO, REMOTE, WARNING, ERROR, AUTH, FATAL, BUG, STDERR, SOFT)
-CF_KEYWORDS(TIMEFORMAT, ISO, OLD, SHORT, LONG, BASE, NAME, CONFIRM, UNDO, CHECK, TIMEOUT)
-CF_KEYWORDS(DEBUG, LATENCY, LIMIT, WATCHDOG, WARNING, TIMEOUT)
+CF_KEYWORDS(NAME, CONFIRM, UNDO, CHECK, TIMEOUT, DEBUG, LATENCY, LIMIT, WATCHDOG, WARNING)
 
 %type <i> log_mask log_mask_list log_cat cfg_timeout
 %type <g> log_file
@@ -85,29 +84,6 @@ mrtdump_base:
  ;
 
 
-CF_ADDTO(conf, timeformat_base)
-
-timeformat_which:
-   ROUTE { $$ = &new_config->tf_route; }
- | PROTOCOL { $$ = &new_config->tf_proto; }
- | BASE { $$ = &new_config->tf_base; }
- | LOG { $$ = &new_config->tf_log; }
- ;
-
-timeformat_spec:
-   timeformat_which TEXT { *$1 = (struct timeformat){$2, NULL, 0}; }
- | timeformat_which TEXT expr TEXT { *$1 = (struct timeformat){$2, $4, $3}; }
- | timeformat_which ISO SHORT { *$1 = (struct timeformat){"%T", "%F", 20*3600}; }
- | timeformat_which ISO LONG  { *$1 = (struct timeformat){"%F %T", NULL, 0}; }
- | timeformat_which OLD SHORT { *$1 = (struct timeformat){NULL, NULL, 0}; }
- | timeformat_which OLD LONG  { *$1 = (struct timeformat){"%d-%m-%Y %T", NULL, 0}; }
- ;
-
-timeformat_base:
-   TIMEFORMAT timeformat_spec ';'
- ;
-
-
 CF_ADDTO(conf, debug_unix)
 
 debug_unix:
diff --git a/sysdep/unix/io.c b/sysdep/unix/io.c
index d3506d75..99f52aa9 100644
--- a/sysdep/unix/io.c
+++ b/sysdep/unix/io.c
@@ -40,6 +40,7 @@
 #include "lib/timer.h"
 #include "lib/string.h"
 #include "nest/iface.h"
+#include "conf/conf.h"
 
 #include "sysdep/unix/unix.h"
 #include CONFIG_INCLUDE_SYSIO_H
@@ -385,102 +386,6 @@ tm_shot(void)
 }
 #endif
 
-/**
- * tm_parse_datetime - parse a date and time
- * @x: datetime string
- *
- * tm_parse_datetime() takes a textual representation of
- * a date and time (dd-mm-yyyy hh:mm:ss)
- * and converts it to the corresponding value of type &bird_clock_t.
- */
-bird_clock_t
-tm_parse_datetime(char *x)
-{
-  struct tm tm;
-  int n;
-  time_t t;
-
-  if (sscanf(x, "%d-%d-%d %d:%d:%d%n", &tm.tm_mday, &tm.tm_mon, &tm.tm_year, &tm.tm_hour, &tm.tm_min, &tm.tm_sec, &n) != 6 || x[n])
-    return tm_parse_date(x);
-  tm.tm_mon--;
-  tm.tm_year -= 1900;
-  t = mktime(&tm);
-  if (t == (time_t) -1)
-    return 0;
-  return t;
-}
-/**
- * tm_parse_date - parse a date
- * @x: date string
- *
- * tm_parse_date() takes a textual representation of a date (dd-mm-yyyy)
- * and converts it to the corresponding value of type &bird_clock_t.
- */
-bird_clock_t
-tm_parse_date(char *x)
-{
-  struct tm tm;
-  int n;
-  time_t t;
-
-  if (sscanf(x, "%d-%d-%d%n", &tm.tm_mday, &tm.tm_mon, &tm.tm_year, &n) != 3 || x[n])
-    return 0;
-  tm.tm_mon--;
-  tm.tm_year -= 1900;
-  tm.tm_hour = tm.tm_min = tm.tm_sec = 0;
-  t = mktime(&tm);
-  if (t == (time_t) -1)
-    return 0;
-  return t;
-}
-
-static void
-tm_format_reltime(char *x, struct tm *tm, bird_clock_t delta)
-{
-  static char *month_names[12] = { "Jan", "Feb", "Mar", "Apr", "May", "Jun",
-				   "Jul", "Aug", "Sep", "Oct", "Nov", "Dec" };
-
-  if (delta < 20*3600)
-    bsprintf(x, "%02d:%02d", tm->tm_hour, tm->tm_min);
-  else if (delta < 360*86400)
-    bsprintf(x, "%s%02d", month_names[tm->tm_mon], tm->tm_mday);
-  else
-    bsprintf(x, "%d", tm->tm_year+1900);
-}
-
-#include "conf/conf.h"
-
-/**
- * tm_format_datetime - convert date and time to textual representation
- * @x: destination buffer of size %TM_DATETIME_BUFFER_SIZE
- * @fmt_spec: specification of resulting textual representation of the time
- * @t: time
- *
- * This function formats the given relative time value @t to a textual
- * date/time representation (dd-mm-yyyy hh:mm:ss) in real time.
- */
-void
-tm_format_datetime(char *x, struct timeformat *fmt_spec, bird_clock_t t)
-{
-  const char *fmt_used;
-  struct tm *tm;
-  bird_clock_t delta = now - t;
-  t = now_real - delta;
-  tm = localtime(&t);
-
-  if (fmt_spec->fmt1 == NULL)
-    return tm_format_reltime(x, tm, delta);
-
-  if ((fmt_spec->limit == 0) || (delta < fmt_spec->limit))
-    fmt_used = fmt_spec->fmt1;
-  else
-    fmt_used = fmt_spec->fmt2;
-
-  int rv = strftime(x, TM_DATETIME_BUFFER_SIZE, fmt_used, tm);
-  if (((rv == 0) && fmt_used[0]) || (rv == TM_DATETIME_BUFFER_SIZE))
-    strcpy(x, "<too-long>");
-}
-
 
 /*
  *	Time clock
diff --git a/sysdep/unix/log.c b/sysdep/unix/log.c
index 06d3b09e..e564f8f2 100644
--- a/sysdep/unix/log.c
+++ b/sysdep/unix/log.c
@@ -120,7 +120,7 @@ log_commit(int class, buffer *buf)
 	  else
 	    {
 	      byte tbuf[TM_DATETIME_BUFFER_SIZE];
-	      tm_format_datetime(tbuf, &config->tf_log, now);
+	      tm_format_real_time(tbuf, config->tf_log.fmt1, current_real_time());
 	      fprintf(l->fh, "%s <%s> ", tbuf, class_names[class]);
 	    }
 	  fputs(buf->start, l->fh);
diff --git a/sysdep/unix/timer.h b/sysdep/unix/timer.h
index 1c4f6e3b..495d10c7 100644
--- a/sysdep/unix/timer.h
+++ b/sysdep/unix/timer.h
@@ -45,18 +45,6 @@ static inline timer * tm_new_set(pool *p, void (*hook)(timer *), void *data, uin
 { return tm2_new_init(p, hook, data, rec S_, rand S_); }
 
 
-struct timeformat {
-  char *fmt1, *fmt2;
-  bird_clock_t limit;
-};
-
-bird_clock_t tm_parse_date(char *);	/* Convert date to bird_clock_t */
-bird_clock_t tm_parse_datetime(char *);	/* Convert date to bird_clock_t */
-
-#define TM_DATETIME_BUFFER_SIZE 32	/* Buffer size required by tm_format_datetime */
-void
-tm_format_datetime(char *x, struct timeformat *fmt_spec, bird_clock_t t);
-
 #define TIME_INFINITY ((s64) 0x7fffffffffffffff)
 
 

From 49fc021337eba2e9cea228e80e1f95ef21b30cd2 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 13 Jun 2017 15:41:49 +0200
Subject: [PATCH 048/198] Printf: Add support for microsecond times

Use '%t' in bsnprintf() for microsecond times (in btime) with variable
sub-second precision.
---
 lib/printf.c      | 55 +++++++++++++++++++++++++++++++++++++++++++----
 lib/printf_test.c |  9 ++++++++
 2 files changed, 60 insertions(+), 4 deletions(-)

diff --git a/lib/printf.c b/lib/printf.c
index 8e3cbbcf..533a1300 100644
--- a/lib/printf.c
+++ b/lib/printf.c
@@ -125,9 +125,10 @@ static char * number(char * str, long num, int base, int size, int precision,
  * or |%I6| can be used for explicit ip4_addr / ip6_addr arguments, |%N| for
  * generic network addresses (net_addr *), |%R| for Router / Network ID (u32
  * value printed as IPv4 address), |%lR| for 64bit Router / Network ID (u64
- * value printed as eight :-separated octets) and |%m| resp. |%M| for error
- * messages (uses strerror() to translate @errno code to message text). On the
- * other hand, it doesn't support floating point numbers.
+ * value printed as eight :-separated octets), |%t| for time values (btime) with
+ * specified subsecond precision, and |%m| resp. |%M| for error messages (uses
+ * strerror() to translate @errno code to message text). On the other hand, it
+ * doesn't support floating point numbers.
  *
  * Result: number of characters of the output string or -1 if
  * the buffer space was insufficient.
@@ -139,6 +140,8 @@ int bvsnprintf(char *buf, int size, const char *fmt, va_list args)
 	int i, base;
 	u32 x;
 	u64 X;
+	btime t;
+	s64 t1, t2;
 	char *str, *start;
 	const char *s;
 	char ipbuf[NET_MAX_TEXT_LENGTH+1];
@@ -279,7 +282,6 @@ int bvsnprintf(char *buf, int size, const char *fmt, va_list args)
 				return -1;
 			continue;
 
-
 		case 'n':
 			if (qualifier == 'l') {
 				long * ip = va_arg(args, long *);
@@ -360,6 +362,50 @@ int bvsnprintf(char *buf, int size, const char *fmt, va_list args)
 			s = ipbuf;
 			goto str;
 
+		case 't':
+			t = va_arg(args, btime);
+			t1 = t TO_S;
+			t2 = t - t1 S;
+
+			if (precision < 0)
+			  precision = 3;
+
+			if (precision > 6)
+			  precision = 6;
+
+			/* Compute field_width for second part */
+			if ((precision > 0) && (field_width > 0))
+			  field_width -= (1 + precision);
+
+			if (field_width < 0)
+			  field_width = 0;
+
+			/* Print seconds */
+			flags |= SIGN;
+			str = number(str, t1, 10, field_width, 0, flags, size);
+			if (!str)
+			  return -1;
+
+			if (precision > 0)
+			{
+			  size -= (str-start);
+			  start = str;
+
+			  if ((1 + precision) > size)
+			    return -1;
+
+			  /* Convert microseconds to requested precision */
+			  for (i = precision; i < 6; i++)
+			    t2 /= 10;
+
+			  /* Print sub-seconds */
+			  *str++ = '.';
+			  str = number(str, t2, 10, precision, 0, ZEROPAD, size - 1);
+			  if (!str)
+			    return -1;
+			}
+			goto done;
+
 		/* integer number formats - set up the flags and "break" */
 		case 'o':
 			base = 8;
@@ -401,6 +447,7 @@ int bvsnprintf(char *buf, int size, const char *fmt, va_list args)
 		str = number(str, num, base, field_width, precision, flags, size);
 		if (!str)
 			return -1;
+	done:	;
 	}
 	if (!size)
 		return -1;
diff --git a/lib/printf_test.c b/lib/printf_test.c
index 3981d3da..a2683d93 100644
--- a/lib/printf_test.c
+++ b/lib/printf_test.c
@@ -56,6 +56,15 @@ t_simple(void)
   BSPRINTF(2, "-1", buf, "%d", -1);
   BSPRINTF(11, "-2147483648", buf, "%d", -2147483648);
 
+  BSPRINTF(7, "123.456", buf, "%t", (btime) 123456789);
+  BSPRINTF(7, "123.456", buf, "%2t", (btime) 123456789);
+  BSPRINTF(8, " 123.456", buf, "%8t", (btime) 123456789);
+  BSPRINTF(4, " 123", buf, "%4.0t", (btime) 123456789);
+  BSPRINTF(8, "123.4567", buf, "%8.4t", (btime) 123456789);
+  BSPRINTF(9, "0123.4567", buf, "%09.4t", (btime) 123456789);
+  BSPRINTF(12, "  123.456789", buf, "%12.10t", (btime) 123456789);
+  BSPRINTF(8, " 123.004", buf, "%8t", (btime) 123004 MS);
+
   return 1;
 }
 

From 212eda07c4e481e3341ede37b0877fa22bc042a4 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 13 Jun 2017 15:55:13 +0200
Subject: [PATCH 049/198] Timers: Fix tests after timer change

---
 lib/event_test.c | 1 +
 test/bt-utils.c  | 1 +
 2 files changed, 2 insertions(+)

diff --git a/lib/event_test.c b/lib/event_test.c
index 92172389..e1215bba 100644
--- a/lib/event_test.c
+++ b/lib/event_test.c
@@ -55,6 +55,7 @@ t_ev_run_list(void)
 
   resource_init();
   olock_init();
+  timer_init();
   io_init();
   rt_init();
   if_init();
diff --git a/test/bt-utils.c b/test/bt-utils.c
index aaeb77b8..571ef2fa 100644
--- a/test/bt-utils.c
+++ b/test/bt-utils.c
@@ -62,6 +62,7 @@ bt_bird_init(void)
 
   resource_init();
   olock_init();
+  timer_init();
   io_init();
   rt_init();
   if_init();

From d3fa9e84e98d7b8c726f5e35d6a359971eb98f94 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 13 Jun 2017 16:52:21 +0200
Subject: [PATCH 050/198] Timers: Show sub-second times in some protocol
 outputs

---
 nest/proto.c          |  2 +-
 proto/bfd/bfd.c       | 11 +++++------
 proto/bgp/bgp.c       | 20 ++++++++++----------
 proto/ospf/neighbor.c | 19 +++----------------
 proto/rpki/rpki.c     | 10 +++++-----
 5 files changed, 24 insertions(+), 38 deletions(-)

diff --git a/nest/proto.c b/nest/proto.c
index 3043b648..8d5edda3 100644
--- a/nest/proto.c
+++ b/nest/proto.c
@@ -1136,7 +1136,7 @@ graceful_restart_show_status(void)
 
   cli_msg(-24, "Graceful restart recovery in progress");
   cli_msg(-24, "  Waiting for %d channels to recover", graceful_restart_locks);
-  cli_msg(-24, "  Wait timer is %d/%d", tm_remains(gr_wait_timer), config->gr_wait);
+  cli_msg(-24, "  Wait timer is %t/%u", tm2_remains(gr_wait_timer), config->gr_wait);
 }
 
 /**
diff --git a/proto/bfd/bfd.c b/proto/bfd/bfd.c
index 2374cfe8..8915140d 100644
--- a/proto/bfd/bfd.c
+++ b/proto/bfd/bfd.c
@@ -1078,7 +1078,7 @@ bfd_show_sessions(struct proto *P)
   byte tbuf[TM_DATETIME_BUFFER_SIZE];
   struct bfd_proto *p = (struct bfd_proto *) P;
   uint state, diag UNUSED;
-  u32 tx_int, timeout;
+  btime tx_int, timeout;
   const char *ifname;
 
   if (p->p.proto_state != PS_UP)
@@ -1099,15 +1099,14 @@ bfd_show_sessions(struct proto *P)
     state = s->loc_state;
     diag = s->loc_diag;
     ifname = (s->ifa && s->ifa->iface) ? s->ifa->iface->name : "---";
-    tx_int = s->last_tx ? (MAX(s->des_min_tx_int, s->rem_min_rx_int) TO_MS) : 0;
-    timeout = (MAX(s->req_min_rx_int, s->rem_min_tx_int) TO_MS) * s->rem_detect_mult;
+    tx_int = s->last_tx ? MAX(s->des_min_tx_int, s->rem_min_rx_int) : 0;
+    timeout = (btime) MAX(s->req_min_rx_int, s->rem_min_tx_int) * s->rem_detect_mult;
 
     state = (state < 4) ? state : 0;
     tm_format_time(tbuf, &config->tf_proto, s->last_state_change);
 
-    cli_msg(-1020, "%-25I %-10s %-10s %-10s  %3u.%03u  %3u.%03u",
-	    s->addr, ifname, bfd_state_names[state], tbuf,
-	    tx_int / 1000, tx_int % 1000, timeout / 1000, timeout % 1000);
+    cli_msg(-1020, "%-25I %-10s %-10s %-10s  %7t  %7t",
+	    s->addr, ifname, bfd_state_names[state], tbuf, tx_int, timeout);
   }
   HASH_WALK_END;
 
diff --git a/proto/bgp/bgp.c b/proto/bgp/bgp.c
index dba7c875..9823be6b 100644
--- a/proto/bgp/bgp.c
+++ b/proto/bgp/bgp.c
@@ -2009,17 +2009,17 @@ bgp_show_proto_info(struct proto *P)
 
     if ((p->start_state < BSS_CONNECT) &&
 	(tm_active(p->startup_timer)))
-      cli_msg(-1006, "    Error wait:       %d/%d",
-	      (int) tm_remains(p->startup_timer), p->startup_delay);
+      cli_msg(-1006, "    Error wait:       %t/%u",
+	      tm2_remains(p->startup_timer), p->startup_delay);
 
     if ((oc->state == BS_ACTIVE) &&
 	(tm_active(oc->connect_timer)))
-      cli_msg(-1006, "    Connect delay:    %d/%d",
-	      (int) tm_remains(oc->connect_timer), p->cf->connect_delay_time);
+      cli_msg(-1006, "    Connect delay:    %t/%u",
+	      tm2_remains(oc->connect_timer), p->cf->connect_delay_time);
 
     if (p->gr_active_num && tm_active(p->gr_timer))
-      cli_msg(-1006, "    Restart timer:    %d/-",
-	      (int) tm_remains(p->gr_timer));
+      cli_msg(-1006, "    Restart timer:    %t/-",
+	      tm2_remains(p->gr_timer));
   }
   else if (P->proto_state == PS_UP)
   {
@@ -2040,10 +2040,10 @@ bgp_show_proto_info(struct proto *P)
 	      p->ext_messages ? " ext-messages" : "");
 */
     cli_msg(-1006, "    Source address:   %I", p->source_addr);
-    cli_msg(-1006, "    Hold timer:       %d/%d",
-	    tm_remains(p->conn->hold_timer), p->conn->hold_time);
-    cli_msg(-1006, "    Keepalive timer:  %d/%d",
-	    tm_remains(p->conn->keepalive_timer), p->conn->keepalive_time);
+    cli_msg(-1006, "    Hold timer:       %t/%u",
+	    tm2_remains(p->conn->hold_timer), p->conn->hold_time);
+    cli_msg(-1006, "    Keepalive timer:  %t/%u",
+	    tm2_remains(p->conn->keepalive_timer), p->conn->keepalive_time);
   }
 
   if ((p->last_error_class != BE_NONE) &&
diff --git a/proto/ospf/neighbor.c b/proto/ospf/neighbor.c
index 64165e46..74ea63d0 100644
--- a/proto/ospf/neighbor.c
+++ b/proto/ospf/neighbor.c
@@ -651,20 +651,6 @@ ospf_sh_neigh_info(struct ospf_neighbor *n)
 {
   struct ospf_iface *ifa = n->ifa;
   char *pos = "PtP  ";
-  char etime[6];
-  int exp, sec, min;
-
-  exp = tm_remains(n->inactim);
-  sec = exp % 60;
-  min = exp / 60;
-  if (min > 59)
-  {
-    bsprintf(etime, "-Inf-");
-  }
-  else
-  {
-    bsprintf(etime, "%02u:%02u", min, sec);
-  }
 
   if ((ifa->type == OSPF_IT_BCAST) || (ifa->type == OSPF_IT_NBMA))
   {
@@ -676,6 +662,7 @@ ospf_sh_neigh_info(struct ospf_neighbor *n)
       pos = "Other";
   }
 
-  cli_msg(-1013, "%-1R\t%3u\t%s/%s\t%-5s\t%-10s %-1I", n->rid, n->priority,
-	  ospf_ns_names[n->state], pos, etime, ifa->ifname, n->ip);
+  cli_msg(-1013, "%-1R\t%3u\t%s/%s\t%7t\t%-10s %-1I",
+	  n->rid, n->priority, ospf_ns_names[n->state], pos,
+	  tm2_remains(n->inactim), ifa->ifname, n->ip);
 }
diff --git a/proto/rpki/rpki.c b/proto/rpki/rpki.c
index a331e45e..7930363f 100644
--- a/proto/rpki/rpki.c
+++ b/proto/rpki/rpki.c
@@ -790,9 +790,9 @@ static void
 rpki_show_proto_info_timer(const char *name, uint num, timer *t)
 {
   if (tm_active(t))
-    cli_msg(-1006, "  %-17s %us (remains %ds)", name, num, (int) tm_remains(t));
+    cli_msg(-1006, "  %-16s: %t/%u", name, tm2_remains(t), num);
   else
-    cli_msg(-1006, "  %-17s ---", name);
+    cli_msg(-1006, "  %-16s: ---", name);
 }
 
 static void
@@ -836,9 +836,9 @@ rpki_show_proto_info(struct proto *P)
       cli_msg(-1006, "  Last update:      ---");
     }
 
-    rpki_show_proto_info_timer("Refresh interval:", cache->refresh_interval, cache->refresh_timer);
-    rpki_show_proto_info_timer("Retry interval:", cache->retry_interval, cache->retry_timer);
-    rpki_show_proto_info_timer("Expire interval:", cache->expire_interval, cache->expire_timer);
+    rpki_show_proto_info_timer("Refresh timer", cache->refresh_interval, cache->refresh_timer);
+    rpki_show_proto_info_timer("Retry timer", cache->retry_interval, cache->retry_timer);
+    rpki_show_proto_info_timer("Expire timer", cache->expire_interval, cache->expire_timer);
 
     if (p->roa4_channel)
       channel_show_info(p->roa4_channel);

From 21f4f0f4b0785e30ce2af4741ffa6f2ebdd7d714 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 14 Jun 2017 13:15:35 +0200
Subject: [PATCH 051/198] Kernel: Update to new timers

---
 sysdep/unix/krt.Y |  4 ++--
 sysdep/unix/krt.c | 45 +++++++++++++++++++++------------------------
 sysdep/unix/krt.h |  4 ++--
 3 files changed, 25 insertions(+), 28 deletions(-)

diff --git a/sysdep/unix/krt.Y b/sysdep/unix/krt.Y
index 3bf7da65..6d90db64 100644
--- a/sysdep/unix/krt.Y
+++ b/sysdep/unix/krt.Y
@@ -68,7 +68,7 @@ kern_item:
  | PERSIST bool { THIS_KRT->persist = $2; }
  | SCAN TIME expr {
       /* Scan time of 0 means scan on startup only */
-      THIS_KRT->scan_time = $3;
+      THIS_KRT->scan_time = $3 S_;
    }
  | LEARN bool {
       THIS_KRT->learn = $2;
@@ -103,7 +103,7 @@ kif_item:
  | INTERFACE kif_iface
  | SCAN TIME expr {
       /* Scan time of 0 means scan on startup only */
-      THIS_KIF->scan_time = $3;
+      THIS_KIF->scan_time = $3 S_;
    }
  ;
 
diff --git a/sysdep/unix/krt.c b/sysdep/unix/krt.c
index a5ec8a39..5b1c40f0 100644
--- a/sysdep/unix/krt.c
+++ b/sysdep/unix/krt.c
@@ -87,7 +87,7 @@ krt_io_init(void)
 struct kif_proto *kif_proto;
 static struct kif_config *kif_cf;
 static timer *kif_scan_timer;
-static bird_clock_t kif_last_shot;
+static btime kif_last_shot;
 
 static struct kif_iface_config kif_default_iface = {};
 
@@ -105,25 +105,25 @@ kif_scan(timer *t)
   struct kif_proto *p = t->data;
 
   KRT_TRACE(p, D_EVENTS, "Scanning interfaces");
-  kif_last_shot = now;
+  kif_last_shot = current_time();
   kif_do_scan(p);
 }
 
 static void
 kif_force_scan(void)
 {
-  if (kif_proto && kif_last_shot + 2 < now)
+  if (kif_proto && ((kif_last_shot + 2 S) < current_time()))
     {
       kif_scan(kif_scan_timer);
-      tm_start(kif_scan_timer, ((struct kif_config *) kif_proto->p.cf)->scan_time);
+      tm2_start(kif_scan_timer, ((struct kif_config *) kif_proto->p.cf)->scan_time);
     }
 }
 
 void
 kif_request_scan(void)
 {
-  if (kif_proto && (kif_scan_timer->expires TO_S > (now + 1)))
-    tm_start(kif_scan_timer, 1);
+  if (kif_proto && (kif_scan_timer->expires > (current_time() + 1 S)))
+    tm2_start(kif_scan_timer, 1 S);
 }
 
 static struct proto *
@@ -144,12 +144,9 @@ kif_start(struct proto *P)
   kif_sys_start(p);
 
   /* Start periodic interface scanning */
-  kif_scan_timer = tm_new(P->pool);
-  kif_scan_timer->hook = kif_scan;
-  kif_scan_timer->data = p;
-  kif_scan_timer->recurrent = KIF_CF->scan_time S;
+  kif_scan_timer = tm2_new_init(P->pool, kif_scan, p, KIF_CF->scan_time, 0);
   kif_scan(kif_scan_timer);
-  tm_start(kif_scan_timer, KIF_CF->scan_time);
+  tm2_start(kif_scan_timer, KIF_CF->scan_time);
 
   return PS_UP;
 }
@@ -159,7 +156,7 @@ kif_shutdown(struct proto *P)
 {
   struct kif_proto *p = (struct kif_proto *) P;
 
-  tm_stop(kif_scan_timer);
+  tm2_stop(kif_scan_timer);
   kif_sys_shutdown(p);
   kif_proto = NULL;
 
@@ -177,10 +174,10 @@ kif_reconfigure(struct proto *p, struct proto_config *new)
 
   if (o->scan_time != n->scan_time)
     {
-      tm_stop(kif_scan_timer);
-      kif_scan_timer->recurrent = n->scan_time S;
+      tm2_stop(kif_scan_timer);
+      kif_scan_timer->recurrent = n->scan_time;
       kif_scan(kif_scan_timer);
-      tm_start(kif_scan_timer, n->scan_time);
+      tm2_start(kif_scan_timer, n->scan_time);
     }
 
   if (!EMPTY_LIST(o->iface_list) || !EMPTY_LIST(n->iface_list))
@@ -212,7 +209,7 @@ kif_init_config(int class)
     cf_error("Kernel device protocol already defined");
 
   kif_cf = (struct kif_config *) proto_config_new(&proto_unix_iface, class);
-  kif_cf->scan_time = 60;
+  kif_cf->scan_time = 60 S;
   init_list(&kif_cf->iface_list);
 
   kif_sys_init_config(kif_cf);
@@ -843,11 +840,11 @@ static void
 krt_scan_timer_start(struct krt_proto *p)
 {
   if (!krt_scan_count)
-    krt_scan_timer = tm_new_set(krt_pool, krt_scan, NULL, 0, KRT_CF->scan_time);
+    krt_scan_timer = tm2_new_init(krt_pool, krt_scan, NULL, KRT_CF->scan_time, 0);
 
   krt_scan_count++;
 
-  tm_start(krt_scan_timer, 1);
+  tm2_start(krt_scan_timer, 1 S);
 }
 
 static void
@@ -865,7 +862,7 @@ krt_scan_timer_stop(struct krt_proto *p UNUSED)
 static void
 krt_scan_timer_kick(struct krt_proto *p UNUSED)
 {
-  tm_start(krt_scan_timer, 0);
+  tm2_start(krt_scan_timer, 0);
 }
 
 #else
@@ -885,20 +882,20 @@ krt_scan(timer *t)
 static void
 krt_scan_timer_start(struct krt_proto *p)
 {
-  p->scan_timer = tm_new_set(p->p.pool, krt_scan, p, 0, KRT_CF->scan_time);
-  tm_start(p->scan_timer, 1);
+  p->scan_timer = tm2_new_init(p->p.pool, krt_scan, p, KRT_CF->scan_time, 0);
+  tm2_start(p->scan_timer, 1 S);
 }
 
 static void
 krt_scan_timer_stop(struct krt_proto *p)
 {
-  tm_stop(p->scan_timer);
+  tm2_stop(p->scan_timer);
 }
 
 static void
 krt_scan_timer_kick(struct krt_proto *p)
 {
-  tm_start(p->scan_timer, 0);
+  tm2_start(p->scan_timer, 0);
 }
 
 #endif
@@ -1174,7 +1171,7 @@ krt_init_config(int class)
 #endif
 
   krt_cf = (struct krt_config *) proto_config_new(&proto_unix_kernel, class);
-  krt_cf->scan_time = 60;
+  krt_cf->scan_time = 60 S;
 
   krt_sys_init_config(krt_cf);
   return (struct proto_config *) krt_cf;
diff --git a/sysdep/unix/krt.h b/sysdep/unix/krt.h
index 39e16a2a..eda83b91 100644
--- a/sysdep/unix/krt.h
+++ b/sysdep/unix/krt.h
@@ -46,8 +46,8 @@ extern struct protocol proto_unix_kernel;
 struct krt_config {
   struct proto_config c;
   struct krt_params sys;	/* Sysdep params */
+  btime scan_time;		/* How often we re-scan routes */
   int persist;			/* Keep routes when we exit */
-  int scan_time;		/* How often we re-scan routes */
   int learn;			/* Learn routes from other sources */
   int devroutes;		/* Allow export of device routes */
   int graceful_restart;		/* Regard graceful restart recovery */
@@ -100,7 +100,7 @@ struct kif_config {
   struct kif_params sys;	/* Sysdep params */
 
   list iface_list;		/* List of iface configs (struct kif_iface_config) */
-  int scan_time;		/* How often we re-scan interfaces */
+  btime scan_time;		/* How often we re-scan interfaces */
 };
 
 struct kif_iface_config {

From d59c1a295834aa5cc63aceb6769c8413fa0639fe Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 14 Jun 2017 16:32:15 +0200
Subject: [PATCH 052/198] RPKI: Update to new timers

---
 proto/rpki/packets.c | 12 ++++++------
 proto/rpki/rpki.c    | 46 +++++++++++++++++++++++---------------------
 proto/rpki/rpki.h    |  6 +++---
 3 files changed, 33 insertions(+), 31 deletions(-)

diff --git a/proto/rpki/packets.c b/proto/rpki/packets.c
index 60ca3936..59a5efaf 100644
--- a/proto/rpki/packets.c
+++ b/proto/rpki/packets.c
@@ -554,9 +554,9 @@ rpki_check_receive_packet(struct rpki_cache *cache, const struct pdu_header *pdu
 	 * (https://tools.ietf.org/html/draft-ietf-sidr-rpki-rtr-rfc6810-bis-07#section-7)
 	 */
       }
-      else if (cache->last_update == 0
-		&& pdu->ver <= RPKI_MAX_VERSION
-		&& pdu->ver < cache->version)
+      else if (!cache->last_update &&
+	       (pdu->ver <= RPKI_MAX_VERSION) &&
+	       (pdu->ver < cache->version))
       {
         CACHE_TRACE(D_EVENTS, cache, "Downgrade session to %s from %u to %u version", rpki_get_cache_ident(cache), cache->version, pdu->ver);
         cache->version = pdu->ver;
@@ -652,7 +652,7 @@ rpki_handle_cache_response_pdu(struct rpki_cache *cache, const struct pdu_cache_
 {
   if (cache->request_session_id)
   {
-    if (cache->last_update != 0)
+    if (cache->last_update)
     {
       /*
        * This isn't the first sync and we already received records. This point
@@ -748,7 +748,7 @@ rpki_handle_prefix_pdu(struct rpki_cache *cache, const struct pdu_header *pdu)
     return RPKI_ERROR;
   }
 
-  cache->last_rx_prefix = now;
+  cache->last_rx_prefix = current_time();
 
   /* A place for 'flags' is same for both data structures pdu_ipv4 or pdu_ipv6  */
   struct pdu_ipv4 *pfx = (void *) pdu;
@@ -814,7 +814,7 @@ rpki_handle_end_of_data_pdu(struct rpki_cache *cache, const struct pdu_end_of_da
       rt_refresh_end(cache->p->roa6_channel->table, cache->p->roa6_channel);
   }
 
-  cache->last_update = now;
+  cache->last_update = current_time();
   cache->serial_num = pdu->serial_num;
   rpki_cache_change_state(cache, RPKI_CS_ESTABLISHED);
 }
diff --git a/proto/rpki/rpki.c b/proto/rpki/rpki.c
index 7930363f..cddbf017 100644
--- a/proto/rpki/rpki.c
+++ b/proto/rpki/rpki.c
@@ -318,59 +318,60 @@ rpki_cache_change_state(struct rpki_cache *cache, const enum rpki_cache_state ne
 static void
 rpki_schedule_next_refresh(struct rpki_cache *cache)
 {
-  uint time_to_wait = cache->refresh_interval;
+  btime t = (btime) cache->refresh_interval S;
 
-  CACHE_DBG(cache, "after %u seconds", time_to_wait);
-  tm_start(cache->refresh_timer, time_to_wait);
+  CACHE_DBG(cache, "after %t s", t);
+  tm2_start(cache->refresh_timer, t);
 }
 
 static void
 rpki_schedule_next_retry(struct rpki_cache *cache)
 {
-  uint time_to_wait = cache->retry_interval;
+  btime t = (btime) cache->retry_interval S;
 
-  CACHE_DBG(cache, "after %u seconds", time_to_wait);
-  tm_start(cache->retry_timer, time_to_wait);
+  CACHE_DBG(cache, "after %t s", t);
+  tm2_start(cache->retry_timer, t);
 }
 
 static void
 rpki_schedule_next_expire_check(struct rpki_cache *cache)
 {
   /* A minimum time to wait is 1 second */
-  uint time_to_wait = MAX(((int)cache->expire_interval - (int)(now - cache->last_update)), 1);
+  btime t = cache->last_update + (btime) cache->expire_interval S - current_time();
+  t = MAX(t, 1 S);
 
-  CACHE_DBG(cache, "after %u seconds", time_to_wait);
-  tm_start(cache->expire_timer, time_to_wait);
+  CACHE_DBG(cache, "after %t s", t);
+  tm2_start(cache->expire_timer, t);
 }
 
 static void
 rpki_stop_refresh_timer_event(struct rpki_cache *cache)
 {
   CACHE_DBG(cache, "Stop");
-  tm_stop(cache->refresh_timer);
+  tm2_stop(cache->refresh_timer);
 }
 
 static void
 rpki_stop_retry_timer_event(struct rpki_cache *cache)
 {
   CACHE_DBG(cache, "Stop");
-  tm_stop(cache->retry_timer);
+  tm2_stop(cache->retry_timer);
 }
 
 static void UNUSED
 rpki_stop_expire_timer_event(struct rpki_cache *cache)
 {
   CACHE_DBG(cache, "Stop");
-  tm_stop(cache->expire_timer);
+  tm2_stop(cache->expire_timer);
 }
 
 static int
 rpki_do_we_recv_prefix_pdu_in_last_seconds(struct rpki_cache *cache)
 {
-  if (cache->last_rx_prefix == 0)
+  if (!cache->last_rx_prefix)
     return 0;
 
-  return ((now - cache->last_rx_prefix) <= 2);
+  return ((current_time() - cache->last_rx_prefix) <= 2 S);
 }
 
 /**
@@ -477,19 +478,20 @@ rpki_expire_hook(timer *tm)
 {
   struct rpki_cache *cache = tm->data;
 
-  if (cache->last_update == 0)
+  if (!cache->last_update)
     return;
 
   CACHE_DBG(cache, "%s", rpki_cache_state_to_str(cache->state));
 
-  if ((cache->last_update + cache->expire_interval) < now)
+  btime t = cache->last_update + (btime) cache->expire_interval S - current_time();
+  if (t <= 0)
   {
     CACHE_TRACE(D_EVENTS, cache, "All ROAs expired");
     rpki_force_restart_proto(cache->p);
   }
   else
   {
-    CACHE_DBG(cache, "Remains %d seconds to become ROAs obsolete", (int)cache->expire_interval - (int)(now - cache->last_update));
+    CACHE_DBG(cache, "Remains %t seconds to become ROAs obsolete", t);
     rpki_schedule_next_expire_check(cache);
   }
 }
@@ -567,9 +569,9 @@ rpki_init_cache(struct rpki_proto *p, struct rpki_config *cf)
   cache->refresh_interval = cf->refresh_interval;
   cache->retry_interval = cf->retry_interval;
   cache->expire_interval = cf->expire_interval;
-  cache->refresh_timer = tm_new_set(pool, &rpki_refresh_hook, cache, 0, 0);
-  cache->retry_timer = tm_new_set(pool, &rpki_retry_hook, cache, 0, 0);
-  cache->expire_timer = tm_new_set(pool, &rpki_expire_hook, cache, 0, 0);
+  cache->refresh_timer = tm2_new_init(pool, &rpki_refresh_hook, cache, 0, 0);
+  cache->retry_timer = tm2_new_init(pool, &rpki_retry_hook, cache, 0, 0);
+  cache->expire_timer = tm2_new_init(pool, &rpki_expire_hook, cache, 0, 0);
 
   cache->tr_sock = mb_allocz(pool, sizeof(struct rpki_tr_sock));
   cache->tr_sock->cache = cache;
@@ -789,7 +791,7 @@ rpki_get_status(struct proto *P, byte *buf)
 static void
 rpki_show_proto_info_timer(const char *name, uint num, timer *t)
 {
-  if (tm_active(t))
+  if (tm2_active(t))
     cli_msg(-1006, "  %-16s: %t/%u", name, tm2_remains(t), num);
   else
     cli_msg(-1006, "  %-16s: ---", name);
@@ -828,7 +830,7 @@ rpki_show_proto_info(struct proto *P)
     if (cache->last_update)
     {
       cli_msg(-1006, "  Serial number:    %u", cache->serial_num);
-      cli_msg(-1006, "  Last update:      before %us", now - cache->last_update);
+      cli_msg(-1006, "  Last update:      before %t s", current_time() - cache->last_update);
     }
     else
     {
diff --git a/proto/rpki/rpki.h b/proto/rpki/rpki.h
index 80ef83df..8972b33a 100644
--- a/proto/rpki/rpki.h
+++ b/proto/rpki/rpki.h
@@ -60,11 +60,11 @@ struct rpki_cache {
   u8 request_session_id;		/* 1: have to request new session id; 0: we have already received session id */
   u32 serial_num;			/* Serial number denotes the logical version of data from cache server */
   u8 version;				/* Protocol version */
-  bird_clock_t last_update;		/* Last successful synchronization with cache server */
-  bird_clock_t last_rx_prefix;		/* Last received prefix PDU */
+  btime last_update;			/* Last successful synchronization with cache server */
+  btime last_rx_prefix;			/* Last received prefix PDU */
 
   /* Intervals can be changed by cache server on the fly */
-  u32 refresh_interval;			/* Actual refresh interval */
+  u32 refresh_interval;			/* Actual refresh interval (in seconds) */
   u32 retry_interval;
   u32 expire_interval;
   timer *retry_timer;			/* Retry timer event */

From c521b3ac324e1c97b22581a06172dcb08fd740f6 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 14 Jun 2017 17:02:11 +0200
Subject: [PATCH 053/198] RAdv: Update to new timers

---
 proto/radv/radv.c | 26 +++++++++++---------------
 proto/radv/radv.h |  4 ++--
 2 files changed, 13 insertions(+), 17 deletions(-)

diff --git a/proto/radv/radv.c b/proto/radv/radv.c
index 34a3f473..978d2091 100644
--- a/proto/radv/radv.c
+++ b/proto/radv/radv.c
@@ -54,17 +54,18 @@ radv_timer(timer *tm)
   radv_send_ra(ifa, 0);
 
   /* Update timer */
-  ifa->last = now;
-  unsigned after = ifa->cf->min_ra_int;
-  after += random() % (ifa->cf->max_ra_int - ifa->cf->min_ra_int + 1);
+  ifa->last = current_time();
+  btime t = (btime) ifa->cf->min_ra_int S;
+  btime r = (btime) (ifa->cf->max_ra_int - ifa->cf->min_ra_int) S;
+  t += random() % (r + 1);
 
   if (ifa->initial)
+  {
+    t = MIN(t, MAX_INITIAL_RTR_ADVERT_INTERVAL);
     ifa->initial--;
+  }
 
-  if (ifa->initial)
-    after = MIN(after, MAX_INITIAL_RTR_ADVERT_INTERVAL);
-
-  tm_start(ifa->timer, after);
+  tm2_start(ifa->timer, t);
 }
 
 static char* ev_name[] = { NULL, "Init", "Change", "RS" };
@@ -92,13 +93,8 @@ radv_iface_notify(struct radv_iface *ifa, int event)
   }
 
   /* Update timer */
-  unsigned delta = now - ifa->last;
-  unsigned after = 0;
-
-  if (delta < ifa->cf->min_delay)
-    after = ifa->cf->min_delay - delta;
-
-  tm_start(ifa->timer, after);
+  btime t = ifa->last + (btime) ifa->cf->min_delay S - current_time();
+  tm2_start(ifa->timer, t);
 }
 
 static void
@@ -154,7 +150,7 @@ radv_iface_new(struct radv_proto *p, struct iface *iface, struct radv_iface_conf
 
   add_tail(&p->iface_list, NODE ifa);
 
-  ifa->timer = tm_new_set(pool, radv_timer, ifa, 0, 0);
+  ifa->timer = tm2_new_init(pool, radv_timer, ifa, 0, 0);
 
   struct object_lock *lock = olock_new(pool);
   lock->addr = IPA_NONE;
diff --git a/proto/radv/radv.h b/proto/radv/radv.h
index 35eea9ba..a4429c60 100644
--- a/proto/radv/radv.h
+++ b/proto/radv/radv.h
@@ -30,7 +30,7 @@
 #define ICMPV6_RA 134
 
 #define MAX_INITIAL_RTR_ADVERTISEMENTS 3
-#define MAX_INITIAL_RTR_ADVERT_INTERVAL 16
+#define MAX_INITIAL_RTR_ADVERT_INTERVAL (16 S_)
 
 #define DEFAULT_MAX_RA_INT 600
 #define DEFAULT_MIN_DELAY 3
@@ -130,7 +130,7 @@ struct radv_iface
   struct object_lock *lock;
   sock *sk;
 
-  bird_clock_t last;		/* Time of last sending of RA */
+  btime last;			/* Time of last sending of RA */
   u16 plen;			/* Length of prepared RA in tbuf, or 0 if not valid */
   byte initial;			/* List of active ifaces */
 };

From 92cc1e745758893a57a2432a0e11e4cd3ad289b7 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 20 Jun 2017 14:30:44 +0200
Subject: [PATCH 054/198] RIP: Update to new timers

---
 proto/rip/config.Y  | 10 +++---
 proto/rip/packets.c | 10 +++---
 proto/rip/rip.c     | 79 +++++++++++++++++++++++----------------------
 proto/rip/rip.h     | 32 +++++++++---------
 4 files changed, 68 insertions(+), 63 deletions(-)

diff --git a/proto/rip/config.Y b/proto/rip/config.Y
index 46f6c177..ba9ac20b 100644
--- a/proto/rip/config.Y
+++ b/proto/rip/config.Y
@@ -57,8 +57,8 @@ rip_proto_start: proto_start rip_variant
   init_list(&RIP_CFG->patt_list);
   RIP_CFG->rip2 = $2;
   RIP_CFG->infinity = RIP_DEFAULT_INFINITY;
-  RIP_CFG->min_timeout_time = 60;
-  RIP_CFG->max_garbage_time = 60;
+  RIP_CFG->min_timeout_time = 60 S_;
+  RIP_CFG->max_garbage_time = 60 S_;
 };
 
 rip_proto_item:
@@ -147,9 +147,9 @@ rip_iface_item:
  | SPLIT HORIZON bool	{ RIP_IFACE->split_horizon = $3; }
  | POISON REVERSE bool	{ RIP_IFACE->poison_reverse = $3; }
  | CHECK ZERO bool	{ RIP_IFACE->check_zero = $3; }
- | UPDATE TIME expr	{ RIP_IFACE->update_time = $3; if ($3<=0) cf_error("Update time must be positive"); }
- | TIMEOUT TIME expr	{ RIP_IFACE->timeout_time = $3; if ($3<=0) cf_error("Timeout time must be positive"); }
- | GARBAGE TIME expr	{ RIP_IFACE->garbage_time = $3; if ($3<=0) cf_error("Garbage time must be positive"); }
+ | UPDATE TIME expr	{ RIP_IFACE->update_time = (btime) $3 S_; if ($3<=0) cf_error("Update time must be positive"); }
+ | TIMEOUT TIME expr	{ RIP_IFACE->timeout_time = (btime) $3 S_; if ($3<=0) cf_error("Timeout time must be positive"); }
+ | GARBAGE TIME expr	{ RIP_IFACE->garbage_time = (btime) $3 S_; if ($3<=0) cf_error("Garbage time must be positive"); }
  | ECMP WEIGHT expr	{ RIP_IFACE->ecmp_weight = $3 - 1; if (($3<1) || ($3>256)) cf_error("ECMP weight must be in range 1-256"); }
  | RX BUFFER expr	{ RIP_IFACE->rx_buffer = $3; if (($3<256) || ($3>65535)) cf_error("RX length must be in range 256-65535"); }
  | TX LENGTH expr	{ RIP_IFACE->tx_length = $3; if (($3<256) || ($3>65535)) cf_error("TX length must be in range 256-65535"); }
diff --git a/proto/rip/packets.c b/proto/rip/packets.c
index de48fd0a..1518dd3f 100644
--- a/proto/rip/packets.c
+++ b/proto/rip/packets.c
@@ -434,6 +434,7 @@ rip_send_response(struct rip_proto *p, struct rip_iface *ifa)
   byte *max = rip_tx_buffer(ifa) + ifa->tx_plen -
     (rip_is_v2(p) ? RIP_BLOCK_LENGTH : 2*RIP_BLOCK_LENGTH);
   ip_addr last_next_hop = IPA_NONE;
+  btime now_ = current_time();
   int send = 0;
 
   struct rip_packet *pkt = (void *) pos;
@@ -450,7 +451,7 @@ rip_send_response(struct rip_proto *p, struct rip_iface *ifa)
 
     /* Stale entries that should be removed */
     if ((en->valid == RIP_ENTRY_STALE) &&
-	((en->changed + (bird_clock_t) ifa->cf->garbage_time) <= now))
+	((en->changed + ifa->cf->garbage_time) <= now_))
       goto next_entry;
 
     /* Triggered updates */
@@ -540,7 +541,7 @@ break_loop:
  * activating the new one.
  */
 void
-rip_send_table(struct rip_proto *p, struct rip_iface *ifa, ip_addr addr, bird_clock_t changed)
+rip_send_table(struct rip_proto *p, struct rip_iface *ifa, ip_addr addr, btime changed)
 {
   DBG("RIP: Opening TX session to %I on %s\n", addr, ifa->iface->name);
 
@@ -591,6 +592,7 @@ rip_receive_response(struct rip_proto *p, struct rip_iface *ifa, struct rip_pack
 
   byte *pos = (byte *) pkt + sizeof(struct rip_packet);
   byte *end = (byte *) pkt + plen;
+  btime now_ = current_time();
 
   for (; pos < end; pos += RIP_BLOCK_LENGTH)
   {
@@ -638,7 +640,7 @@ rip_receive_response(struct rip_proto *p, struct rip_iface *ifa, struct rip_pack
 	.next_hop = ipa_nonzero(rte.next_hop) ? rte.next_hop : from->nbr->addr,
 	.metric = rte.metric,
 	.tag = rte.tag,
-	.expires = now + ifa->cf->timeout_time
+	.expires = now_ + ifa->cf->timeout_time
       };
 
       rip_update_rte(p, &rte.net, &new);
@@ -705,7 +707,7 @@ rip_rx_hook(sock *sk, uint len)
   if ((plen - sizeof(struct rip_packet)) % RIP_BLOCK_LENGTH)
     DROP("invalid length", plen);
 
-  n->last_seen = now;
+  n->last_seen = current_time();
   rip_update_bfd(p, n);
 
   switch (pkt->command)
diff --git a/proto/rip/rip.c b/proto/rip/rip.c
index 2583b49b..782e6f5c 100644
--- a/proto/rip/rip.c
+++ b/proto/rip/rip.c
@@ -364,7 +364,7 @@ rip_rt_notify(struct proto *P, struct channel *ch UNUSED, struct network *net, s
   /* Activate triggered updates */
   if (en->metric != old_metric)
   {
-    en->changed = now;
+    en->changed = current_time();
     rip_trigger_update(p);
   }
 }
@@ -506,10 +506,10 @@ rip_iface_start(struct rip_iface *ifa)
 
   TRACE(D_EVENTS, "Starting interface %s", ifa->iface->name);
 
-  ifa->next_regular = now + (random() % ifa->cf->update_time) + 1;
-  ifa->next_triggered = now;	/* Available immediately */
-  ifa->want_triggered = 1;	/* All routes in triggered update */
-  tm_start(ifa->timer, 1);	/* Or 100 ms */
+  ifa->next_regular = current_time() + (random() % ifa->cf->update_time) + 100 MS;
+  ifa->next_triggered = current_time();	/* Available immediately */
+  ifa->want_triggered = 1;		/* All routes in triggered update */
+  tm2_start(ifa->timer, 100 MS);
   ifa->up = 1;
 
   if (!ifa->cf->passive)
@@ -529,7 +529,7 @@ rip_iface_stop(struct rip_iface *ifa)
   WALK_LIST_FIRST(n, ifa->neigh_list)
     rip_remove_neighbor(p, n);
 
-  tm_stop(ifa->timer);
+  tm2_stop(ifa->timer);
   ifa->up = 0;
 }
 
@@ -642,7 +642,7 @@ rip_add_iface(struct rip_proto *p, struct iface *iface, struct rip_iface_config
 
   add_tail(&p->iface_list, NODE ifa);
 
-  ifa->timer = tm_new_set(p->p.pool, rip_iface_timer, ifa, 0, 0);
+  ifa->timer = tm2_new_init(p->p.pool, rip_iface_timer, ifa, 0, 0);
 
   struct object_lock *lock = olock_new(p->p.pool);
   lock->type = OBJLOCK_UDP;
@@ -690,8 +690,8 @@ rip_reconfigure_iface(struct rip_proto *p, struct rip_iface *ifa, struct rip_ifa
 
   rip_iface_update_buffers(ifa);
 
-  if (ifa->next_regular > (now + (bird_clock_t) new->update_time))
-    ifa->next_regular = now + (random() % new->update_time) + 1;
+  if (ifa->next_regular > (current_time() + new->update_time))
+    ifa->next_regular = current_time() + (random() % new->update_time) + 100 MS;
 
   if (new->check_link != old->check_link)
     rip_iface_update_state(ifa);
@@ -816,8 +816,9 @@ rip_timer(timer *t)
   struct rip_iface *ifa;
   struct rip_neighbor *n, *nn;
   struct fib_iterator fit;
-  bird_clock_t next = now + MIN(cf->min_timeout_time, cf->max_garbage_time);
-  bird_clock_t expires = 0;
+  btime now_ = current_time();
+  btime next = now_ + MIN(cf->min_timeout_time, cf->max_garbage_time);
+  btime expires = 0;
 
   TRACE(D_EVENTS, "Main timer fired");
 
@@ -832,7 +833,7 @@ rip_timer(timer *t)
     /* Checking received routes for timeout and for dead neighbors */
     for (rp = &en->routes; rt = *rp; /* rp = &rt->next */)
     {
-      if (!rip_valid_rte(rt) || (rt->expires <= now))
+      if (!rip_valid_rte(rt) || (rt->expires <= now_))
       {
 	rip_remove_rte(p, rp);
 	changed = 1;
@@ -862,7 +863,7 @@ rip_timer(timer *t)
     {
       expires = en->changed + cf->max_garbage_time;
 
-      if (expires <= now)
+      if (expires <= now_)
       {
 	// TRACE(D_EVENTS, "entry is too old: %N", en->n.addr);
 	en->valid = 0;
@@ -890,20 +891,20 @@ rip_timer(timer *t)
       {
 	expires = n->last_seen + n->ifa->cf->timeout_time;
 
-	if (expires <= now)
+	if (expires <= now_)
 	  rip_remove_neighbor(p, n);
 	else
 	  next = MIN(next, expires);
       }
 
-  tm_start(p->timer, MAX(next - now, 1));
+  tm2_start(p->timer, MAX(next - now_, 100 MS));
 }
 
 static inline void
 rip_kick_timer(struct rip_proto *p)
 {
-  if (p->timer->expires TO_S > (now + 1))
-    tm_start(p->timer, 1);	/* Or 100 ms */
+  if (p->timer->expires > (current_time() + 100 MS))
+    tm2_start(p->timer, 100 MS);
 }
 
 /**
@@ -921,7 +922,8 @@ rip_iface_timer(timer *t)
 {
   struct rip_iface *ifa = t->data;
   struct rip_proto *p = ifa->rip;
-  bird_clock_t period = ifa->cf->update_time;
+  btime now_ = current_time();
+  btime period = ifa->cf->update_time;
 
   if (ifa->cf->passive)
     return;
@@ -930,40 +932,40 @@ rip_iface_timer(timer *t)
 
   if (ifa->tx_active)
   {
-    if (now < (ifa->next_regular + period))
-      { tm_start(ifa->timer, 1); return; }
+    if (now_ < (ifa->next_regular + period))
+    { tm2_start(ifa->timer, 100 MS); return; }
 
     /* We are too late, reset is done by rip_send_table() */
     log(L_WARN "%s: Too slow update on %s, resetting", p->p.name, ifa->iface->name);
   }
 
-  if (now >= ifa->next_regular)
+  if (now_ >= ifa->next_regular)
   {
     /* Send regular update, set timer for next period (or following one if necessay) */
     TRACE(D_EVENTS, "Sending regular updates for %s", ifa->iface->name);
     rip_send_table(p, ifa, ifa->addr, 0);
-    ifa->next_regular += period * (1 + ((now - ifa->next_regular) / period));
+    ifa->next_regular += period * (1 + ((now_ - ifa->next_regular) / period));
     ifa->want_triggered = 0;
     p->triggered = 0;
   }
-  else if (ifa->want_triggered && (now >= ifa->next_triggered))
+  else if (ifa->want_triggered && (now_ >= ifa->next_triggered))
   {
     /* Send triggered update, enforce interval between triggered updates */
     TRACE(D_EVENTS, "Sending triggered updates for %s", ifa->iface->name);
     rip_send_table(p, ifa, ifa->addr, ifa->want_triggered);
-    ifa->next_triggered = now + MIN(5, period / 2 + 1);
+    ifa->next_triggered = now_ + MIN(5 S, period / 2);
     ifa->want_triggered = 0;
     p->triggered = 0;
   }
 
-  tm_start(ifa->timer, ifa->want_triggered ? 1 : (ifa->next_regular - now));
+  tm2_start(ifa->timer, ifa->want_triggered ? (1 S) : (ifa->next_regular - now_));
 }
 
 static inline void
 rip_iface_kick_timer(struct rip_iface *ifa)
 {
-  if (ifa->timer->expires TO_S > (now + 1))
-    tm_start(ifa->timer, 1);	/* Or 100 ms */
+  if (ifa->timer->expires > (current_time() + 100 MS))
+    tm2_start(ifa->timer, 100 MS);
 }
 
 static void
@@ -984,7 +986,7 @@ rip_trigger_update(struct rip_proto *p)
       continue;
 
     TRACE(D_EVENTS, "Scheduling triggered updates for %s", ifa->iface->name);
-    ifa->want_triggered = now;
+    ifa->want_triggered = current_time();
     rip_iface_kick_timer(ifa);
   }
 
@@ -1109,7 +1111,7 @@ rip_start(struct proto *P)
   fib_init(&p->rtable, P->pool, cf->rip2 ? NET_IP4 : NET_IP6,
 	   sizeof(struct rip_entry), OFFSETOF(struct rip_entry, n), 0, NULL);
   p->rte_slab = sl_new(P->pool, sizeof(struct rip_rte));
-  p->timer = tm_new_set(P->pool, rip_timer, p, 0, 0);
+  p->timer = tm2_new_init(P->pool, rip_timer, p, 0, 0);
 
   p->rip2 = cf->rip2;
   p->ecmp = cf->ecmp;
@@ -1119,7 +1121,7 @@ rip_start(struct proto *P)
   p->log_pkt_tbf = (struct tbf){ .rate = 1, .burst = 5 };
   p->log_rte_tbf = (struct tbf){ .rate = 4, .burst = 20 };
 
-  tm_start(p->timer, MIN(cf->min_timeout_time, cf->max_garbage_time));
+  tm2_start(p->timer, MIN(cf->min_timeout_time, cf->max_garbage_time));
 
   return PS_UP;
 }
@@ -1194,7 +1196,7 @@ rip_show_interfaces(struct proto *P, char *iff)
   }
 
   cli_msg(-1021, "%s:", p->p.name);
-  cli_msg(-1021, "%-10s %-6s %6s %6s %6s",
+  cli_msg(-1021, "%-10s %-6s %6s %6s %7s",
 	  "Interface", "State", "Metric", "Nbrs", "Timer");
 
   WALK_LIST(ifa, p->iface_list)
@@ -1207,8 +1209,9 @@ rip_show_interfaces(struct proto *P, char *iff)
       if (n->last_seen)
 	nbrs++;
 
-    int timer = MAX(ifa->next_regular - now, 0);
-    cli_msg(-1021, "%-10s %-6s %6u %6u %6u",
+    btime now_ = current_time();
+    btime timer = (ifa->next_regular > now_) ? (ifa->next_regular - now_) : 0;
+    cli_msg(-1021, "%-10s %-6s %6u %6u %7t",
 	    ifa->iface->name, (ifa->up ? "Up" : "Down"), ifa->cf->metric, nbrs, timer);
   }
 
@@ -1230,7 +1233,7 @@ rip_show_neighbors(struct proto *P, char *iff)
   }
 
   cli_msg(-1022, "%s:", p->p.name);
-  cli_msg(-1022, "%-25s %-10s %6s %6s %6s",
+  cli_msg(-1022, "%-25s %-10s %6s %6s %7s",
 	  "IP address", "Interface", "Metric", "Routes", "Seen");
 
   WALK_LIST(ifa, p->iface_list)
@@ -1243,8 +1246,8 @@ rip_show_neighbors(struct proto *P, char *iff)
       if (!n->last_seen)
 	continue;
 
-      int timer = now - n->last_seen;
-      cli_msg(-1022, "%-25I %-10s %6u %6u %6u",
+      btime timer = current_time() - n->last_seen;
+      cli_msg(-1022, "%-25I %-10s %6u %6u %7t",
 	      n->nbr->addr, ifa->iface->name, ifa->cf->metric, n->uc, timer);
     }
   }
@@ -1262,9 +1265,9 @@ rip_dump(struct proto *P)
   i = 0;
   FIB_WALK(&p->rtable, struct rip_entry, en)
   {
-    debug("RIP: entry #%d: %N via %I dev %s valid %d metric %d age %d s\n",
+    debug("RIP: entry #%d: %N via %I dev %s valid %d metric %d age %t\n",
 	  i++, en->n.addr, en->next_hop, en->iface->name,
-	  en->valid, en->metric, now - en->changed);
+	  en->valid, en->metric, current_time() - en->changed);
   }
   FIB_WALK_END;
 
diff --git a/proto/rip/rip.h b/proto/rip/rip.h
index 03dc9138..2371ea31 100644
--- a/proto/rip/rip.h
+++ b/proto/rip/rip.h
@@ -38,9 +38,9 @@
 
 #define RIP_DEFAULT_ECMP_LIMIT	16
 #define RIP_DEFAULT_INFINITY	16
-#define RIP_DEFAULT_UPDATE_TIME	30
-#define RIP_DEFAULT_TIMEOUT_TIME 180
-#define RIP_DEFAULT_GARBAGE_TIME 120
+#define RIP_DEFAULT_UPDATE_TIME	  (30 S_)
+#define RIP_DEFAULT_TIMEOUT_TIME (180 S_)
+#define RIP_DEFAULT_GARBAGE_TIME (120 S_)
 
 
 struct rip_config
@@ -52,8 +52,8 @@ struct rip_config
   u8 ecmp;				/* Maximum number of nexthops in ECMP route, or 0 */
   u8 infinity;				/* Maximum metric value, representing infinity */
 
-  u32 min_timeout_time;			/* Minimum of interface timeout_time */
-  u32 max_garbage_time;			/* Maximum of interface garbage_time */
+  btime min_timeout_time;		/* Minimum of interface timeout_time */
+  btime max_garbage_time;		/* Maximum of interface garbage_time */
 };
 
 struct rip_iface_config
@@ -78,9 +78,9 @@ struct rip_iface_config
   u16 tx_length;			/* TX packet length limit (including headers), 0 for MTU */
   int tx_tos;
   int tx_priority;
-  u32 update_time;			/* Periodic update interval */
-  u32 timeout_time;			/* Route expiration timeout */
-  u32 garbage_time;			/* Unreachable entry GC timeout */
+  btime update_time;			/* Periodic update interval */
+  btime timeout_time;			/* Route expiration timeout */
+  btime garbage_time;			/* Unreachable entry GC timeout */
   list *passwords;			/* Passwords for authentication */
 };
 
@@ -120,14 +120,14 @@ struct rip_iface
   list neigh_list;			/* List of iface neighbors (struct rip_neighbor) */
 
   /* Update scheduling */
-  bird_clock_t next_regular;		/* Next time when regular update should be called */
-  bird_clock_t next_triggered;		/* Next time when triggerd update may be called */
-  bird_clock_t want_triggered;		/* Nonzero if triggered update is scheduled */
+  btime next_regular;			/* Next time when regular update should be called */
+  btime next_triggered;			/* Next time when triggerd update may be called */
+  btime want_triggered;			/* Nonzero if triggered update is scheduled */
 
   /* Active update */
   int tx_active;			/* Update session is active */
   ip_addr tx_addr;			/* Update session destination address */
-  bird_clock_t tx_changed;		/* Minimal changed time for triggered update */
+  btime tx_changed;			/* Minimal changed time for triggered update */
   struct fib_iterator tx_fit;		/* FIB iterator in RIP routing table (p.rtable) */
 };
 
@@ -137,7 +137,7 @@ struct rip_neighbor
   struct rip_iface *ifa;		/* Associated interface, may be NULL if stale */
   struct neighbor *nbr;			/* Associaded core neighbor, may be NULL if stale */
   struct bfd_request *bfd_req;		/* BFD request, if BFD is used */
-  bird_clock_t last_seen;		/* Time of last received and accepted message */
+  btime last_seen;			/* Time of last received and accepted message */
   u32 uc;				/* Use count, number of routes linking the neighbor */
   u32 csn;				/* Last received crypto sequence number */
 };
@@ -153,7 +153,7 @@ struct rip_entry
   struct iface *iface;			/* Outgoing route iface (for next hop) */
   ip_addr next_hop;			/* Outgoing route next hop */
 
-  bird_clock_t changed;			/* Last time when the outgoing route metric changed */
+  btime changed;			/* Last time when the outgoing route metric changed */
 
   struct fib_node n;
 };
@@ -167,7 +167,7 @@ struct rip_rte
   u16 metric;				/* Route metric (after increase) */
   u16 tag;				/* Route tag */
 
-  bird_clock_t expires;			/* Time of route expiration */
+  btime expires;			/* Time of route expiration */
 };
 
 
@@ -211,7 +211,7 @@ void rip_show_neighbors(struct proto *P, char *iff);
 
 /* packets.c */
 void rip_send_request(struct rip_proto *p, struct rip_iface *ifa);
-void rip_send_table(struct rip_proto *p, struct rip_iface *ifa, ip_addr addr, bird_clock_t changed);
+void rip_send_table(struct rip_proto *p, struct rip_iface *ifa, ip_addr addr, btime changed);
 int rip_open_socket(struct rip_iface *ifa);
 
 

From ee528fbd5dc482ceece52832d4a8ea5a08251bfa Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 20 Jun 2017 15:55:39 +0200
Subject: [PATCH 055/198] Timers: Add typecast to unit-converting macros

---
 conf/confbase.Y    | 9 ++++-----
 lib/birdlib.h      | 7 ++++---
 nest/password.c    | 1 +
 proto/radv/radv.c  | 6 +++---
 proto/rip/config.Y | 6 +++---
 proto/rpki/rpki.c  | 8 ++++----
 sysdep/unix/io.c   | 8 ++++----
 7 files changed, 23 insertions(+), 22 deletions(-)

diff --git a/conf/confbase.Y b/conf/confbase.Y
index 390041c4..16330984 100644
--- a/conf/confbase.Y
+++ b/conf/confbase.Y
@@ -80,8 +80,7 @@ CF_DECLS
 %type <iface> ipa_scope
 
 %type <i> expr bool pxlen4
-%type <i32> expr_us
-%type <time> time
+%type <time> expr_us time
 %type <a> ipa
 %type <net> net_ip4_ net_ip6_ net_ip6 net_ip_ net_ip net_or_ipa
 %type <net_ptr> net_ net_any net_vpn4_ net_vpn6_ net_vpn_ net_roa4_ net_roa6_ net_roa_
@@ -137,9 +136,9 @@ expr:
 
 
 expr_us:
-   expr S  { $$ = (u32) $1 * 1000000; }
- | expr MS { $$ = (u32) $1 * 1000; }
- | expr US { $$ = (u32) $1 * 1; }
+   expr S  { $$ = $1 S_; }
+ | expr MS { $$ = $1 MS_; }
+ | expr US { $$ = $1 US_; }
  ;
 
 /* Switches */
diff --git a/lib/birdlib.h b/lib/birdlib.h
index a0826aab..9a77fc7b 100644
--- a/lib/birdlib.h
+++ b/lib/birdlib.h
@@ -71,9 +71,9 @@ static inline int u64_cmp(u64 i1, u64 i2)
 typedef s64 btime;
 typedef s64 bird_clock_t;
 
-#define S_	*1000000
-#define MS_	*1000
-#define US_	*1
+#define S_	* (btime) 1000000
+#define MS_	* (btime) 1000
+#define US_	* (btime) 1
 #define TO_S	/1000000
 #define TO_MS	/1000
 #define TO_US	/1
@@ -82,6 +82,7 @@ typedef s64 bird_clock_t;
 #define S	S_
 #define MS	MS_
 #define US	US_
+#define NS	/1000
 #endif
 
 
diff --git a/nest/password.c b/nest/password.c
index 72f23f5d..6f87af21 100644
--- a/nest/password.c
+++ b/nest/password.c
@@ -10,6 +10,7 @@
 #include "nest/bird.h"
 #include "nest/password.h"
 #include "lib/string.h"
+#include "lib/timer.h"
 #include "lib/mac.h"
 
 struct password_item *last_password_item = NULL;
diff --git a/proto/radv/radv.c b/proto/radv/radv.c
index 978d2091..415cd1d9 100644
--- a/proto/radv/radv.c
+++ b/proto/radv/radv.c
@@ -55,8 +55,8 @@ radv_timer(timer *tm)
 
   /* Update timer */
   ifa->last = current_time();
-  btime t = (btime) ifa->cf->min_ra_int S;
-  btime r = (btime) (ifa->cf->max_ra_int - ifa->cf->min_ra_int) S;
+  btime t = ifa->cf->min_ra_int S;
+  btime r = (ifa->cf->max_ra_int - ifa->cf->min_ra_int) S;
   t += random() % (r + 1);
 
   if (ifa->initial)
@@ -93,7 +93,7 @@ radv_iface_notify(struct radv_iface *ifa, int event)
   }
 
   /* Update timer */
-  btime t = ifa->last + (btime) ifa->cf->min_delay S - current_time();
+  btime t = ifa->last + ifa->cf->min_delay S - current_time();
   tm2_start(ifa->timer, t);
 }
 
diff --git a/proto/rip/config.Y b/proto/rip/config.Y
index ba9ac20b..e3bc4ae3 100644
--- a/proto/rip/config.Y
+++ b/proto/rip/config.Y
@@ -147,9 +147,9 @@ rip_iface_item:
  | SPLIT HORIZON bool	{ RIP_IFACE->split_horizon = $3; }
  | POISON REVERSE bool	{ RIP_IFACE->poison_reverse = $3; }
  | CHECK ZERO bool	{ RIP_IFACE->check_zero = $3; }
- | UPDATE TIME expr	{ RIP_IFACE->update_time = (btime) $3 S_; if ($3<=0) cf_error("Update time must be positive"); }
- | TIMEOUT TIME expr	{ RIP_IFACE->timeout_time = (btime) $3 S_; if ($3<=0) cf_error("Timeout time must be positive"); }
- | GARBAGE TIME expr	{ RIP_IFACE->garbage_time = (btime) $3 S_; if ($3<=0) cf_error("Garbage time must be positive"); }
+ | UPDATE TIME expr	{ RIP_IFACE->update_time = $3 S_; if ($3<=0) cf_error("Update time must be positive"); }
+ | TIMEOUT TIME expr	{ RIP_IFACE->timeout_time = $3 S_; if ($3<=0) cf_error("Timeout time must be positive"); }
+ | GARBAGE TIME expr	{ RIP_IFACE->garbage_time = $3 S_; if ($3<=0) cf_error("Garbage time must be positive"); }
  | ECMP WEIGHT expr	{ RIP_IFACE->ecmp_weight = $3 - 1; if (($3<1) || ($3>256)) cf_error("ECMP weight must be in range 1-256"); }
  | RX BUFFER expr	{ RIP_IFACE->rx_buffer = $3; if (($3<256) || ($3>65535)) cf_error("RX length must be in range 256-65535"); }
  | TX LENGTH expr	{ RIP_IFACE->tx_length = $3; if (($3<256) || ($3>65535)) cf_error("TX length must be in range 256-65535"); }
diff --git a/proto/rpki/rpki.c b/proto/rpki/rpki.c
index cddbf017..f17024fe 100644
--- a/proto/rpki/rpki.c
+++ b/proto/rpki/rpki.c
@@ -318,7 +318,7 @@ rpki_cache_change_state(struct rpki_cache *cache, const enum rpki_cache_state ne
 static void
 rpki_schedule_next_refresh(struct rpki_cache *cache)
 {
-  btime t = (btime) cache->refresh_interval S;
+  btime t = cache->refresh_interval S;
 
   CACHE_DBG(cache, "after %t s", t);
   tm2_start(cache->refresh_timer, t);
@@ -327,7 +327,7 @@ rpki_schedule_next_refresh(struct rpki_cache *cache)
 static void
 rpki_schedule_next_retry(struct rpki_cache *cache)
 {
-  btime t = (btime) cache->retry_interval S;
+  btime t = cache->retry_interval S;
 
   CACHE_DBG(cache, "after %t s", t);
   tm2_start(cache->retry_timer, t);
@@ -337,7 +337,7 @@ static void
 rpki_schedule_next_expire_check(struct rpki_cache *cache)
 {
   /* A minimum time to wait is 1 second */
-  btime t = cache->last_update + (btime) cache->expire_interval S - current_time();
+  btime t = cache->last_update + cache->expire_interval S - current_time();
   t = MAX(t, 1 S);
 
   CACHE_DBG(cache, "after %t s", t);
@@ -483,7 +483,7 @@ rpki_expire_hook(timer *tm)
 
   CACHE_DBG(cache, "%s", rpki_cache_state_to_str(cache->state));
 
-  btime t = cache->last_update + (btime) cache->expire_interval S - current_time();
+  btime t = cache->last_update + cache->expire_interval S - current_time();
   if (t <= 0)
   {
     CACHE_TRACE(D_EVENTS, cache, "All ROAs expired");
diff --git a/sysdep/unix/io.c b/sysdep/unix/io.c
index 99f52aa9..25121d74 100644
--- a/sysdep/unix/io.c
+++ b/sysdep/unix/io.c
@@ -406,7 +406,7 @@ times_init(struct timeloop *loop)
   if ((ts.tv_sec < 0) || (((s64) ts.tv_sec) > ((s64) 1 << 40)))
     log(L_WARN "Monotonic clock is crazy");
 
-  loop->last_time = ((s64) ts.tv_sec S) + (ts.tv_nsec / 1000);
+  loop->last_time = ts.tv_sec S + ts.tv_nsec NS;
   loop->real_time = 0;
 }
 
@@ -420,7 +420,7 @@ times_update(struct timeloop *loop)
   if (rv < 0)
     die("clock_gettime: %m");
 
-  btime new_time = ((s64) ts.tv_sec S) + (ts.tv_nsec / 1000);
+  btime new_time = ts.tv_sec S + ts.tv_nsec NS;
 
   if (new_time < loop->last_time)
     log(L_ERR "Monotonic clock is broken");
@@ -439,7 +439,7 @@ times_update_real_time(struct timeloop *loop)
   if (rv < 0)
     die("clock_gettime: %m");
 
-  loop->real_time = ((s64) ts.tv_sec S) + (ts.tv_nsec / 1000);
+  loop->real_time = ts.tv_sec S + ts.tv_nsec NS;
 }
 
 
@@ -2281,7 +2281,7 @@ io_update_time(void)
   if (rv < 0)
     die("clock_gettime: %m");
 
-  last_time = ((s64) ts.tv_sec S) + (ts.tv_nsec / 1000);
+  last_time = ts.tv_sec S + ts.tv_nsec NS;
 
   if (event_open)
   {

From b32d557a6eca10c1e1dc2f2ab83e201f53d134b4 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 20 Jun 2017 18:03:06 +0200
Subject: [PATCH 056/198] OSPF: Update to new timers

Note that recurrent timers are currently limited to ~1 hour.
---
 proto/ospf/dbdes.c    | 10 +++++-----
 proto/ospf/iface.c    | 28 ++++++++++++++--------------
 proto/ospf/lsalib.h   |  3 +++
 proto/ospf/lsupd.c    | 16 ++++++++--------
 proto/ospf/neighbor.c | 24 ++++++++++++------------
 proto/ospf/ospf.c     |  6 +++---
 proto/ospf/ospf.h     | 12 ++++++------
 proto/ospf/packet.c   |  8 ++++----
 proto/ospf/rt.c       |  6 +++---
 proto/ospf/topology.c | 19 ++++++++++---------
 proto/ospf/topology.h |  2 +-
 11 files changed, 69 insertions(+), 65 deletions(-)

diff --git a/proto/ospf/dbdes.c b/proto/ospf/dbdes.c
index a4452cc8..270259a7 100644
--- a/proto/ospf/dbdes.c
+++ b/proto/ospf/dbdes.c
@@ -279,8 +279,8 @@ ospf_process_dbdes(struct ospf_proto *p, struct ospf_packet *pkt, struct ospf_ne
       req->lsa = lsa;
       req->lsa_body = LSA_BODY_DUMMY;
 
-      if (!tm_active(n->lsrq_timer))
-	tm_start(n->lsrq_timer, 0);
+      if (!tm2_active(n->lsrq_timer))
+	tm2_start(n->lsrq_timer, 0);
     }
   }
 
@@ -366,7 +366,7 @@ ospf_receive_dbdes(struct ospf_packet *pkt, struct ospf_iface *ifa,
       n->options = rcv_options;
       n->myimms &= ~DBDES_MS;
       n->imms = rcv_imms;
-      tm_stop(n->dbdes_timer);
+      tm2_stop(n->dbdes_timer);
       ospf_neigh_sm(n, INM_NEGDONE);
       ospf_send_dbdes(p, n);
       break;
@@ -422,13 +422,13 @@ ospf_receive_dbdes(struct ospf_packet *pkt, struct ospf_iface *ifa,
 
       if (!(n->myimms & DBDES_M) && !(n->imms & DBDES_M))
       {
-	tm_stop(n->dbdes_timer);
+	tm2_stop(n->dbdes_timer);
 	ospf_neigh_sm(n, INM_EXDONE);
 	break;
       }
 
       ospf_send_dbdes(p, n);
-      tm_start(n->dbdes_timer, n->ifa->rxmtint);
+      tm2_start(n->dbdes_timer, n->ifa->rxmtint S);
     }
     else
     {
diff --git a/proto/ospf/iface.c b/proto/ospf/iface.c
index d1f9365e..33ec21fb 100644
--- a/proto/ospf/iface.c
+++ b/proto/ospf/iface.c
@@ -263,13 +263,13 @@ ospf_iface_down(struct ospf_iface *ifa)
     ospf_neigh_sm(n, INM_KILLNBR);
 
   if (ifa->hello_timer)
-    tm_stop(ifa->hello_timer);
+    tm2_stop(ifa->hello_timer);
 
   if (ifa->poll_timer)
-    tm_stop(ifa->poll_timer);
+    tm2_stop(ifa->poll_timer);
 
   if (ifa->wait_timer)
-    tm_stop(ifa->wait_timer);
+    tm2_stop(ifa->wait_timer);
 
   ospf_flush2_lsa(p, &ifa->link_lsa);
   ospf_flush2_lsa(p, &ifa->net_lsa);
@@ -396,15 +396,15 @@ ospf_iface_sm(struct ospf_iface *ifa, int event)
 	{
 	  ospf_iface_chstate(ifa, OSPF_IS_WAITING);
 	  if (ifa->wait_timer)
-	    tm_start(ifa->wait_timer, ifa->waitint);
+	    tm2_start(ifa->wait_timer, ifa->waitint S);
 	}
       }
 
       if (ifa->hello_timer)
-	tm_start(ifa->hello_timer, ifa->helloint);
+	tm2_start(ifa->hello_timer, ifa->helloint S);
 
       if (ifa->poll_timer)
-	tm_start(ifa->poll_timer, ifa->pollint);
+	tm2_start(ifa->poll_timer, ifa->pollint S);
 
       ospf_send_hello(ifa, OHS_HELLO, NULL);
     }
@@ -494,13 +494,13 @@ ospf_iface_add(struct object_lock *lock)
 
   if (! ifa->stub)
   {
-    ifa->hello_timer = tm_new_set(ifa->pool, hello_timer_hook, ifa, 0, ifa->helloint);
+    ifa->hello_timer = tm2_new_init(ifa->pool, hello_timer_hook, ifa, ifa->helloint S, 0);
 
     if (ifa->type == OSPF_IT_NBMA)
-      ifa->poll_timer = tm_new_set(ifa->pool, poll_timer_hook, ifa, 0, ifa->pollint);
+      ifa->poll_timer = tm2_new_init(ifa->pool, poll_timer_hook, ifa, ifa->pollint S, 0);
 
     if ((ifa->type == OSPF_IT_BCAST) || (ifa->type == OSPF_IT_NBMA))
-      ifa->wait_timer = tm_new_set(ifa->pool, wait_timer_hook, ifa, 0, 0);
+      ifa->wait_timer = tm2_new_init(ifa->pool, wait_timer_hook, ifa, 0, 0);
 
     ifa->flood_queue_size = ifa_flood_queue_size(ifa);
     ifa->flood_queue = mb_allocz(ifa->pool, ifa->flood_queue_size * sizeof(void *));
@@ -703,7 +703,7 @@ ospf_iface_new_vlink(struct ospf_proto *p, struct ospf_iface_patt *ip)
 
   add_tail(&p->iface_list, NODE ifa);
 
-  ifa->hello_timer = tm_new_set(ifa->pool, hello_timer_hook, ifa, 0, ifa->helloint);
+  ifa->hello_timer = tm2_new_init(ifa->pool, hello_timer_hook, ifa, ifa->helloint S, 0);
 
   ifa->flood_queue_size = ifa_flood_queue_size(ifa);
   ifa->flood_queue = mb_allocz(ifa->pool, ifa->flood_queue_size * sizeof(void *));
@@ -717,8 +717,8 @@ ospf_iface_change_timer(timer *tm, uint val)
 
   tm->recurrent = val S;
 
-  if (tm_active(tm))
-    tm_start(tm, val);
+  if (tm2_active(tm))
+    tm2_start(tm, val S);
 }
 
 static inline void
@@ -801,8 +801,8 @@ ospf_iface_reconfigure(struct ospf_iface *ifa, struct ospf_iface_patt *new)
 	       ifname, ifa->waitint, new->waitint);
 
     ifa->waitint = new->waitint;
-    if (ifa->wait_timer && ifa->wait_timer->expires)
-      tm_start(ifa->wait_timer, ifa->waitint);
+    if (ifa->wait_timer && tm2_active(ifa->wait_timer))
+      tm2_start(ifa->wait_timer, ifa->waitint S);
   }
 
   /* DEAD TIMER */
diff --git a/proto/ospf/lsalib.h b/proto/ospf/lsalib.h
index 0d477f58..fca7faec 100644
--- a/proto/ospf/lsalib.h
+++ b/proto/ospf/lsalib.h
@@ -60,4 +60,7 @@ void lsa_parse_sum_rt(struct top_hash_entry *en, int ospf2, u32 *drid, u32 *metr
 void lsa_parse_ext(struct top_hash_entry *en, int ospf2, int af, struct ospf_lsa_ext_local *rt);
 int lsa_validate(struct ospf_lsa_header *lsa, u32 lsa_type, int ospf2, void *body);
 
+static inline btime lsa_inst_age(struct top_hash_entry *en)
+{ return current_time() - en->inst_time; }
+
 #endif /* _BIRD_OSPF_LSALIB_H_ */
diff --git a/proto/ospf/lsupd.c b/proto/ospf/lsupd.c
index 157d9628..18811392 100644
--- a/proto/ospf/lsupd.c
+++ b/proto/ospf/lsupd.c
@@ -115,7 +115,7 @@ ospf_lsa_lsrq_down(struct top_hash_entry *req, struct ospf_neighbor *n)
 
   if (EMPTY_SLIST(n->lsrql))
   {
-    tm_stop(n->lsrq_timer);
+    tm2_stop(n->lsrq_timer);
 
     if (n->state == NEIGHBOR_LOADING)
       ospf_neigh_sm(n, INM_LOADDONE);
@@ -136,8 +136,8 @@ ospf_lsa_lsrt_up(struct top_hash_entry *en, struct ospf_neighbor *n)
   ret->lsa = en->lsa;
   ret->lsa_body = LSA_BODY_DUMMY;
 
-  if (!tm_active(n->lsrt_timer))
-    tm_start(n->lsrt_timer, n->ifa->rxmtint);
+  if (!tm2_active(n->lsrt_timer))
+    tm2_start(n->lsrt_timer, n->ifa->rxmtint S);
 }
 
 void
@@ -150,7 +150,7 @@ ospf_lsa_lsrt_down_(struct top_hash_entry *en, struct ospf_neighbor *n, struct t
   ospf_hash_delete(n->lsrth, ret);
 
   if (EMPTY_SLIST(n->lsrtl))
-    tm_stop(n->lsrt_timer);
+    tm2_stop(n->lsrt_timer);
 }
 
 static inline int
@@ -175,8 +175,8 @@ ospf_add_flushed_to_lsrt(struct ospf_proto *p, struct ospf_neighbor *n)
       ospf_lsa_lsrt_up(en, n);
 
   /* If we found any flushed LSA, we send them ASAP */
-  if (tm_active(n->lsrt_timer))
-    tm_start(n->lsrt_timer, 0);
+  if (tm2_active(n->lsrt_timer))
+    tm2_start(n->lsrt_timer, 0);
 }
 
 static int ospf_flood_lsupd(struct ospf_proto *p, struct top_hash_entry **lsa_list, uint lsa_count, uint lsa_min_count, struct ospf_iface *ifa);
@@ -572,7 +572,7 @@ ospf_receive_lsupd(struct ospf_packet *pkt, struct ospf_iface *ifa,
     {
       /* 13. (5a) - enforce minimum time between updates for received LSAs */
       /* We also use this to ratelimit reactions to received self-originated LSAs */
-      if (en && ((now - en->inst_time) < MINLSARRIVAL))
+      if (en && (lsa_inst_age(en) < MINLSARRIVAL))
       {
 	OSPF_TRACE(D_EVENTS, "Skipping LSA received in less that MinLSArrival");
 	continue;
@@ -700,7 +700,7 @@ ospf_receive_lsupd(struct ospf_packet *pkt, struct ospf_iface *ifa,
   if (!EMPTY_SLIST(n->lsrql) && (n->lsrqi == SHEAD(n->lsrql)))
   {
     ospf_send_lsreq(p, n);
-    tm_start(n->lsrq_timer, n->ifa->rxmtint);
+    tm2_start(n->lsrq_timer, n->ifa->rxmtint S);
   }
 
   return;
diff --git a/proto/ospf/neighbor.c b/proto/ospf/neighbor.c
index 74ea63d0..12eb9e84 100644
--- a/proto/ospf/neighbor.c
+++ b/proto/ospf/neighbor.c
@@ -66,10 +66,10 @@ reset_lists(struct ospf_proto *p, struct ospf_neighbor *n)
   ospf_top_free(n->lsrth);
   ospf_reset_lsack_queue(n);
 
-  tm_stop(n->dbdes_timer);
-  tm_stop(n->lsrq_timer);
-  tm_stop(n->lsrt_timer);
-  tm_stop(n->ackd_timer);
+  tm2_stop(n->dbdes_timer);
+  tm2_stop(n->lsrq_timer);
+  tm2_stop(n->lsrt_timer);
+  tm2_stop(n->ackd_timer);
 
   init_lists(p, n);
 }
@@ -94,11 +94,11 @@ ospf_neighbor_new(struct ospf_iface *ifa)
   init_list(&n->ackl[ACKL_DIRECT]);
   init_list(&n->ackl[ACKL_DELAY]);
 
-  n->inactim = tm_new_set(pool, inactivity_timer_hook, n, 0, 0);
-  n->dbdes_timer = tm_new_set(pool, dbdes_timer_hook, n, 0, ifa->rxmtint);
-  n->lsrq_timer = tm_new_set(pool, lsrq_timer_hook, n, 0, ifa->rxmtint);
-  n->lsrt_timer = tm_new_set(pool, lsrt_timer_hook, n, 0, ifa->rxmtint);
-  n->ackd_timer = tm_new_set(pool, ackd_timer_hook, n, 0, ifa->rxmtint / 2);
+  n->inactim = tm2_new_init(pool, inactivity_timer_hook, n, 0, 0);
+  n->dbdes_timer = tm2_new_init(pool, dbdes_timer_hook, n, ifa->rxmtint S, 0);
+  n->lsrq_timer = tm2_new_init(pool, lsrq_timer_hook, n, ifa->rxmtint S, 0);
+  n->lsrt_timer = tm2_new_init(pool, lsrt_timer_hook, n, ifa->rxmtint S, 0);
+  n->ackd_timer = tm2_new_init(pool, ackd_timer_hook, n, ifa->rxmtint S / 2, 0);
 
   return (n);
 }
@@ -185,8 +185,8 @@ ospf_neigh_chstate(struct ospf_neighbor *n, u8 state)
     n->dds++;
     n->myimms = DBDES_IMMS;
 
-    tm_start(n->dbdes_timer, 0);
-    tm_start(n->ackd_timer, ifa->rxmtint / 2);
+    tm2_start(n->dbdes_timer, 0);
+    tm2_start(n->ackd_timer, ifa->rxmtint S / 2);
   }
 
   if (state > NEIGHBOR_EXSTART)
@@ -231,7 +231,7 @@ ospf_neigh_sm(struct ospf_neighbor *n, int event)
       ospf_neigh_chstate(n, NEIGHBOR_INIT);
 
     /* Restart inactivity timer */
-    tm_start(n->inactim, n->ifa->deadint);
+    tm2_start(n->inactim, n->ifa->deadint S);
     break;
 
   case INM_2WAYREC:
diff --git a/proto/ospf/ospf.c b/proto/ospf/ospf.c
index a9081ed0..13f3845b 100644
--- a/proto/ospf/ospf.c
+++ b/proto/ospf/ospf.c
@@ -241,8 +241,8 @@ ospf_start(struct proto *P)
   p->asbr = c->asbr;
   p->ecmp = c->ecmp;
   p->tick = c->tick;
-  p->disp_timer = tm_new_set(P->pool, ospf_disp, p, 0, p->tick);
-  tm_start(p->disp_timer, 1);
+  p->disp_timer = tm2_new_init(P->pool, ospf_disp, p, p->tick S, 0);
+  tm2_start(p->disp_timer, 100 MS);
   p->lsab_size = 256;
   p->lsab_used = 0;
   p->lsab = mb_alloc(P->pool, p->lsab_size);
@@ -677,7 +677,7 @@ ospf_reconfigure(struct proto *P, struct proto_config *CF)
   p->ecmp = new->ecmp;
   p->tick = new->tick;
   p->disp_timer->recurrent = p->tick S;
-  tm_start(p->disp_timer, 1);
+  tm2_start(p->disp_timer, 100 MS);
 
   /* Mark all areas and ifaces */
   WALK_LIST(oa, p->area_list)
diff --git a/proto/ospf/ospf.h b/proto/ospf/ospf.h
index 1530b26a..5aa27a16 100644
--- a/proto/ospf/ospf.h
+++ b/proto/ospf/ospf.h
@@ -64,10 +64,10 @@
 
 #define OSPF_PROTO 89
 
-#define LSREFRESHTIME 1800	/* 30 minutes */
-#define MINLSINTERVAL 5
-#define MINLSARRIVAL 1
-#define LSINFINITY 0xffffff
+#define LSREFRESHTIME		1800	/* 30 minutes */
+#define MINLSINTERVAL		(5 S_)
+#define MINLSARRIVAL		(1 S_)
+#define LSINFINITY		0xffffff
 
 #define OSPF_DEFAULT_TICK 1
 #define OSPF_DEFAULT_STUB_COST 1000
@@ -283,8 +283,8 @@ struct ospf_iface
 				   interface.  LSAs contained in the update */
   u16 helloint;			/* number of seconds between hello sending */
   list *passwords;
-  u32 csn;                      /* Last used crypt seq number */
-  bird_clock_t csn_use;         /* Last time when packet with that CSN was sent */
+  u32 csn;			/* Last used crypt seq number */
+  btime csn_use;		/* Last time when packet with that CSN was sent */
   ip_addr all_routers;		/* Multicast (or broadcast) address for all routers */
   ip_addr des_routers;		/* Multicast (or NULL) address for designated routers */
   ip_addr drip;			/* Designated router IP */
diff --git a/proto/ospf/packet.c b/proto/ospf/packet.c
index 7f9f3b39..38d7a75f 100644
--- a/proto/ospf/packet.c
+++ b/proto/ospf/packet.c
@@ -77,16 +77,16 @@ ospf_pkt_finalize(struct ospf_iface *ifa, struct ospf_packet *pkt, uint *plen)
        reboot when system does not have independent RTC? */
     if (!ifa->csn)
     {
-      ifa->csn = (u32) now;
-      ifa->csn_use = now;
+      ifa->csn = (u32) (current_real_time() TO_S);
+      ifa->csn_use = current_time();
     }
 
     /* We must have sufficient delay between sending a packet and increasing
        CSN to prevent reordering of packets (in a network) with different CSNs */
-    if ((now - ifa->csn_use) > 1)
+    if ((current_time() - ifa->csn_use) > 1 S)
       ifa->csn++;
 
-    ifa->csn_use = now;
+    ifa->csn_use = current_time();
 
     uint auth_len = mac_type_length(pass->alg);
     byte *auth_tail = ((byte *) pkt + *plen);
diff --git a/proto/ospf/rt.c b/proto/ospf/rt.c
index b289d767..36bf0387 100644
--- a/proto/ospf/rt.c
+++ b/proto/ospf/rt.c
@@ -1321,7 +1321,7 @@ ospf_rt_abr2(struct ospf_proto *p)
       if (translate && (oa->translate != TRANS_ON))
       {
 	if (oa->translate == TRANS_WAIT)
-	  tm_stop(oa->translator_timer);
+	  tm2_stop(oa->translator_timer);
 
 	oa->translate = TRANS_ON;
       }
@@ -1329,10 +1329,10 @@ ospf_rt_abr2(struct ospf_proto *p)
       if (!translate && (oa->translate == TRANS_ON))
       {
 	if (oa->translator_timer == NULL)
-	  oa->translator_timer = tm_new_set(p->p.pool, translator_timer_hook, oa, 0, 0);
+	  oa->translator_timer = tm2_new_init(p->p.pool, translator_timer_hook, oa, 0, 0);
 
 	/* Schedule the end of translation */
-	tm_start(oa->translator_timer, oa->ac->transint);
+	tm2_start(oa->translator_timer, oa->ac->transint S);
 	oa->translate = TRANS_WAIT;
       }
     }
diff --git a/proto/ospf/topology.c b/proto/ospf/topology.c
index 8dd91a40..717c8280 100644
--- a/proto/ospf/topology.c
+++ b/proto/ospf/topology.c
@@ -70,7 +70,7 @@ ospf_install_lsa(struct ospf_proto *p, struct ospf_lsa_header *lsa, u32 type, u3
   en->lsa_body = body;
   en->lsa = *lsa;
   en->init_age = en->lsa.age;
-  en->inst_time = now;
+  en->inst_time = current_time();
 
   /*
    * We do not set en->mode. It is either default LSA_M_BASIC, or in a special
@@ -128,7 +128,7 @@ ospf_advance_lsa(struct ospf_proto *p, struct top_hash_entry *en, struct ospf_ls
       en->lsa.sn = lsa->sn + 1;
       en->lsa.age = 0;
       en->init_age = 0;
-      en->inst_time = now;
+      en->inst_time = current_time();
       lsa_generate_checksum(&en->lsa, en->lsa_body);
 
       OSPF_TRACE(D_EVENTS, "Advancing LSA: Type: %04x, Id: %R, Rt: %R, Seq: %08x",
@@ -160,7 +160,7 @@ ospf_advance_lsa(struct ospf_proto *p, struct top_hash_entry *en, struct ospf_ls
       en->lsa = *lsa;
       en->lsa.age = LSA_MAXAGE;
       en->init_age = lsa->age;
-      en->inst_time = now;
+      en->inst_time = current_time();
 
       OSPF_TRACE(D_EVENTS, "Resetting LSA:  Type: %04x, Id: %R, Rt: %R, Seq: %08x",
 		 en->lsa_type, en->lsa.id, en->lsa.rt, en->lsa.sn);
@@ -196,7 +196,7 @@ static int
 ospf_do_originate_lsa(struct ospf_proto *p, struct top_hash_entry *en, void *lsa_body, u16 lsa_blen, u16 lsa_opts)
 {
   /* Enforce MinLSInterval */
-  if ((en->init_age == 0) && en->inst_time && ((en->inst_time + MINLSINTERVAL) > now))
+  if (!en->init_age && en->inst_time && (lsa_inst_age(en) < MINLSINTERVAL))
     return 0;
 
   /* Handle wrapping sequence number */
@@ -237,7 +237,7 @@ ospf_do_originate_lsa(struct ospf_proto *p, struct top_hash_entry *en, void *lsa
   en->lsa.sn++;
   en->lsa.age = 0;
   en->init_age = 0;
-  en->inst_time = now;
+  en->inst_time = current_time();
   lsa_generate_checksum(&en->lsa, en->lsa_body);
 
   OSPF_TRACE(D_EVENTS, "Originating LSA: Type: %04x, Id: %R, Rt: %R, Seq: %08x",
@@ -381,7 +381,7 @@ ospf_refresh_lsa(struct ospf_proto *p, struct top_hash_entry *en)
   en->lsa.sn++;
   en->lsa.age = 0;
   en->init_age = 0;
-  en->inst_time = now;
+  en->inst_time = current_time();
   lsa_generate_checksum(&en->lsa, en->lsa_body);
   ospf_flood_lsa(p, en, NULL);
 }
@@ -476,14 +476,15 @@ void
 ospf_update_lsadb(struct ospf_proto *p)
 {
   struct top_hash_entry *en, *nxt;
-  bird_clock_t real_age;
+  btime now_ = current_time();
+  int real_age;
 
   WALK_SLIST_DELSAFE(en, nxt, p->lsal)
   {
     if (en->next_lsa_body)
       ospf_originate_next_lsa(p, en);
 
-    real_age = en->init_age + (now - en->inst_time);
+    real_age = en->init_age + (now_ - en->inst_time) TO_S;
 
     if (en->lsa.age == LSA_MAXAGE)
     {
@@ -1636,7 +1637,7 @@ ospf_originate_prefix_net_lsa(struct ospf_proto *p, struct ospf_iface *ifa)
 }
 
 static inline int breaks_minlsinterval(struct top_hash_entry *en)
-{ return en && (en->lsa.age < LSA_MAXAGE) && ((en->inst_time + MINLSINTERVAL) > now); }
+{ return en && (en->lsa.age < LSA_MAXAGE) && (lsa_inst_age(en) < MINLSINTERVAL); }
 
 void
 ospf_update_topology(struct ospf_proto *p)
diff --git a/proto/ospf/topology.h b/proto/ospf/topology.h
index d1682c54..ac87334b 100644
--- a/proto/ospf/topology.h
+++ b/proto/ospf/topology.h
@@ -26,7 +26,7 @@ struct top_hash_entry
   void *next_lsa_body;		/* For postponed LSA origination */
   u16 next_lsa_blen;		/* For postponed LSA origination */
   u16 next_lsa_opts;		/* For postponed LSA origination */
-  bird_clock_t inst_time;	/* Time of installation into DB */
+  btime inst_time;		/* Time of installation into DB */
   struct ort *nf;		/* Reference fibnode for sum and ext LSAs, NULL for otherwise */
   struct nexthop *nhs;		/* Computed nexthops - valid only in ospf_rt_spf() */
   ip_addr lb;			/* In OSPFv2, link back address. In OSPFv3, any global address in the area useful for vlinks */

From cc881bd15561224f507ed7162016114ed2b15a1e Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 21 Jun 2017 14:14:51 +0200
Subject: [PATCH 057/198] BGP: Update to new timers

---
 proto/bgp/bgp.c   | 44 ++++++++++++++++++++------------------------
 proto/bgp/bgp.h   |  7 ++++---
 proto/ospf/ospf.h |  8 ++++----
 3 files changed, 28 insertions(+), 31 deletions(-)

diff --git a/proto/bgp/bgp.c b/proto/bgp/bgp.c
index 9823be6b..df4c240a 100644
--- a/proto/bgp/bgp.c
+++ b/proto/bgp/bgp.c
@@ -314,23 +314,24 @@ err1:
 /**
  * bgp_start_timer - start a BGP timer
  * @t: timer
- * @value: time to fire (0 to disable the timer)
+ * @value: time (in seconds) to fire (0 to disable the timer)
  *
  * This functions calls tm_start() on @t with time @value and the amount of
  * randomization suggested by the BGP standard. Please use it for all BGP
  * timers.
  */
 void
-bgp_start_timer(timer *t, int value)
+bgp_start_timer(timer *t, uint value)
 {
   if (value)
   {
-    /* The randomization procedure is specified in RFC 1771: 9.2.3.3 */
-    int randomize = random() % ((value / 4) + 1);
-    tm_start(t, value - randomize);
+    /* The randomization procedure is specified in RFC 4271 section 10 */
+    btime time = value S;
+    btime randomize = random() % ((time / 4) + 1);
+    tm2_start(t, time - randomize);
   }
   else
-    tm_stop(t);
+    tm2_stop(t);
 }
 
 /**
@@ -383,10 +384,10 @@ bgp_update_startup_delay(struct bgp_proto *p)
 
   DBG("BGP: Updating startup delay\n");
 
-  if (p->last_proto_error && ((now - p->last_proto_error) >= (int) cf->error_amnesia_time))
+  if (p->last_proto_error && ((current_time() - p->last_proto_error) >= cf->error_amnesia_time S))
     p->startup_delay = 0;
 
-  p->last_proto_error = now;
+  p->last_proto_error = current_time();
 
   if (cf->disable_after_error)
   {
@@ -516,7 +517,7 @@ bgp_conn_enter_established_state(struct bgp_conn *conn)
   int peer_gr_ready = peer->gr_aware && !(peer->gr_flags & BGP_GRF_RESTART);
 
   if (p->gr_active_num)
-    tm_stop(p->gr_timer);
+    tm2_stop(p->gr_timer);
 
   /* Number of active channels */
   int num = 0;
@@ -615,7 +616,7 @@ bgp_conn_enter_close_state(struct bgp_conn *conn)
   int os = conn->state;
 
   bgp_conn_set_state(conn, BS_CLOSE);
-  tm_stop(conn->keepalive_timer);
+  tm2_stop(conn->keepalive_timer);
   conn->sk->rx_hook = NULL;
 
   /* Timeout for CLOSE state, if we cannot send notification soon then we just hangup */
@@ -778,7 +779,7 @@ bgp_send_open(struct bgp_conn *conn)
   DBG("BGP: Sending open\n");
   conn->sk->rx_hook = bgp_rx;
   conn->sk->tx_hook = bgp_tx;
-  tm_stop(conn->connect_timer);
+  tm2_stop(conn->connect_timer);
   bgp_schedule_packet(conn, NULL, PKT_OPEN);
   bgp_conn_set_state(conn, BS_OPENSENT);
   bgp_start_timer(conn->hold_timer, conn->bgp->cf->initial_hold_time);
@@ -887,9 +888,9 @@ bgp_setup_conn(struct bgp_proto *p, struct bgp_conn *conn)
   conn->last_channel = 0;
   conn->last_channel_count = 0;
 
-  conn->connect_timer	= tm_new_set(p->p.pool, bgp_connect_timeout,	conn, 0, 0);
-  conn->hold_timer 	= tm_new_set(p->p.pool, bgp_hold_timeout,	conn, 0, 0);
-  conn->keepalive_timer	= tm_new_set(p->p.pool, bgp_keepalive_timeout,	conn, 0, 0);
+  conn->connect_timer	= tm2_new_init(p->p.pool, bgp_connect_timeout,	 conn, 0, 0);
+  conn->hold_timer 	= tm2_new_init(p->p.pool, bgp_hold_timeout,	 conn, 0, 0);
+  conn->keepalive_timer	= tm2_new_init(p->p.pool, bgp_keepalive_timeout, conn, 0, 0);
 
   conn->tx_ev = ev_new(p->p.pool);
   conn->tx_ev->hook = bgp_kick_tx;
@@ -1302,13 +1303,8 @@ bgp_start(struct proto *P)
   p->event->hook = bgp_decision;
   p->event->data = p;
 
-  p->startup_timer = tm_new(p->p.pool);
-  p->startup_timer->hook = bgp_startup_timeout;
-  p->startup_timer->data = p;
-
-  p->gr_timer = tm_new(p->p.pool);
-  p->gr_timer->hook = bgp_graceful_restart_timeout;
-  p->gr_timer->data = p;
+  p->startup_timer = tm2_new_init(p->p.pool, bgp_startup_timeout, p, 0, 0);
+  p->gr_timer = tm2_new_init(p->p.pool, bgp_graceful_restart_timeout, p, 0, 0);
 
   p->local_id = proto_get_router_id(P->cf);
   if (p->rr_client)
@@ -2008,16 +2004,16 @@ bgp_show_proto_info(struct proto *P)
     struct bgp_conn *oc = &p->outgoing_conn;
 
     if ((p->start_state < BSS_CONNECT) &&
-	(tm_active(p->startup_timer)))
+	(tm2_active(p->startup_timer)))
       cli_msg(-1006, "    Error wait:       %t/%u",
 	      tm2_remains(p->startup_timer), p->startup_delay);
 
     if ((oc->state == BS_ACTIVE) &&
-	(tm_active(oc->connect_timer)))
+	(tm2_active(oc->connect_timer)))
       cli_msg(-1006, "    Connect delay:    %t/%u",
 	      tm2_remains(oc->connect_timer), p->cf->connect_delay_time);
 
-    if (p->gr_active_num && tm_active(p->gr_timer))
+    if (p->gr_active_num && tm2_active(p->gr_timer))
       cli_msg(-1006, "    Restart timer:    %t/-",
 	      tm2_remains(p->gr_timer));
   }
diff --git a/proto/bgp/bgp.h b/proto/bgp/bgp.h
index 3b38c05f..3d940c22 100644
--- a/proto/bgp/bgp.h
+++ b/proto/bgp/bgp.h
@@ -108,6 +108,7 @@ struct bgp_config {
   int allow_local_pref;			/* Allow LOCAL_PREF in EBGP sessions */
   int gr_mode;				/* Graceful restart mode (BGP_GR_*) */
   int setkey;				/* Set MD5 password to system SA/SP database */
+  /* Times below are in seconds */
   unsigned gr_time;			/* Graceful restart timeout */
   unsigned connect_delay_time;		/* Minimum delay between connect attempts */
   unsigned connect_retry_time;		/* Timeout for connect attempts */
@@ -257,8 +258,8 @@ struct bgp_proto {
   event *event;				/* Event for respawning and shutting process */
   timer *startup_timer;			/* Timer used to delay protocol startup due to previous errors (startup_delay) */
   timer *gr_timer;			/* Timer waiting for reestablishment after graceful restart */
-  unsigned startup_delay;		/* Time to delay protocol startup by due to errors */
-  bird_clock_t last_proto_error;	/* Time of last error that leads to protocol stop */
+  uint startup_delay;			/* Delay (in seconds) of protocol startup due to previous errors */
+  btime last_proto_error;		/* Time of last error that leads to protocol stop */
   u8 last_error_class; 			/* Error class of last error */
   u32 last_error_code;			/* Error code of last error. BGP protocol errors
 					   are encoded as (bgp_err_code << 16 | bgp_err_subcode) */
@@ -422,7 +423,7 @@ extern struct linpool *bgp_linpool;
 extern struct linpool *bgp_linpool2;
 
 
-void bgp_start_timer(timer *t, int value);
+void bgp_start_timer(timer *t, uint value);
 void bgp_check_config(struct bgp_config *c);
 void bgp_error(struct bgp_conn *c, unsigned code, unsigned subcode, byte *data, int len);
 void bgp_close_conn(struct bgp_conn *c);
diff --git a/proto/ospf/ospf.h b/proto/ospf/ospf.h
index 5aa27a16..dbc231b6 100644
--- a/proto/ospf/ospf.h
+++ b/proto/ospf/ospf.h
@@ -269,10 +269,10 @@ struct ospf_iface
   sock *sk;			/* IP socket */
   list neigh_list;		/* List of neighbors (struct ospf_neighbor) */
   u32 cost;			/* Cost of iface */
-  u32 waitint;			/* number of sec before changing state from wait */
-  u32 rxmtint;			/* number of seconds between LSA retransmissions */
-  u32 pollint;			/* Poll interval */
-  u32 deadint;			/* after "deadint" missing hellos is router dead */
+  u32 waitint;			/* Number of seconds before changing state from wait */
+  u32 rxmtint;			/* Number of seconds between LSA retransmissions */
+  u32 pollint;			/* Poll interval in seconds */
+  u32 deadint;			/* After deadint seconds without hellos is router dead */
   u32 iface_id;			/* Interface ID (iface->index or new value for vlinks) */
   u32 vid;			/* ID of peer of virtual link */
   ip_addr vip;			/* IP of peer of virtual link */

From 3e405fb188bee0bb8fcf91f574126771c8661afb Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 21 Jun 2017 14:43:49 +0200
Subject: [PATCH 058/198] Nest: Update to new timers

---
 conf/conf.c        | 10 +++++-----
 conf/conf.h        |  4 ++--
 nest/proto.c       |  9 ++++-----
 sysdep/unix/main.c |  2 +-
 sysdep/unix/unix.h |  2 +-
 5 files changed, 13 insertions(+), 14 deletions(-)

diff --git a/conf/conf.c b/conf/conf.c
index a570fad5..2b5e9c71 100644
--- a/conf/conf.c
+++ b/conf/conf.c
@@ -307,7 +307,7 @@ config_done(void *unused UNUSED)
  * config_commit - commit a configuration
  * @c: new configuration
  * @type: type of reconfiguration (RECONFIG_SOFT or RECONFIG_HARD)
- * @timeout: timeout for undo (or 0 for no timeout)
+ * @timeout: timeout for undo (in seconds; or 0 for no timeout)
  *
  * When a configuration is parsed and prepared for use, the
  * config_commit() function starts the process of reconfiguration.
@@ -331,7 +331,7 @@ config_done(void *unused UNUSED)
  * are accepted.
  */
 int
-config_commit(struct config *c, int type, int timeout)
+config_commit(struct config *c, int type, uint timeout)
 {
   if (shutting_down)
     {
@@ -340,10 +340,10 @@ config_commit(struct config *c, int type, int timeout)
     }
 
   undo_available = 1;
-  if (timeout > 0)
-    tm_start(config_timer, timeout);
+  if (timeout)
+    tm2_start(config_timer, timeout S);
   else
-    tm_stop(config_timer);
+    tm2_stop(config_timer);
 
   if (configuring)
     {
diff --git a/conf/conf.h b/conf/conf.h
index 552d0120..12f51c9d 100644
--- a/conf/conf.h
+++ b/conf/conf.h
@@ -41,7 +41,7 @@ struct config {
   struct timeformat tf_proto;		/* Time format for 'show protocol' */
   struct timeformat tf_log;		/* Time format for the logfile */
   struct timeformat tf_base;		/* Time format for other purposes */
-  u32 gr_wait;				/* Graceful restart wait timeout */
+  u32 gr_wait;				/* Graceful restart wait timeout (sec) */
 
   int cli_debug;			/* Tracing of CLI connections and commands */
   int latency_debug;			/* I/O loop tracks duration of each event */
@@ -68,7 +68,7 @@ struct config *config_alloc(const char *name);
 int config_parse(struct config *);
 int cli_parse(struct config *);
 void config_free(struct config *);
-int config_commit(struct config *, int type, int timeout);
+int config_commit(struct config *, int type, uint timeout);
 int config_confirm(void);
 int config_undo(void);
 void config_init(void);
diff --git a/nest/proto.c b/nest/proto.c
index 8d5edda3..27c8fded 100644
--- a/nest/proto.c
+++ b/nest/proto.c
@@ -1083,9 +1083,8 @@ graceful_restart_init(void)
   }
 
   graceful_restart_state = GRS_ACTIVE;
-  gr_wait_timer = tm_new(proto_pool);
-  gr_wait_timer->hook = graceful_restart_done;
-  tm_start(gr_wait_timer, config->gr_wait);
+  gr_wait_timer = tm2_new_init(proto_pool, graceful_restart_done, NULL, 0, 0);
+  tm2_start(gr_wait_timer, config->gr_wait S);
 }
 
 /**
@@ -1182,7 +1181,7 @@ channel_graceful_restart_unlock(struct channel *c)
   graceful_restart_locks--;
 
   if ((graceful_restart_state == GRS_ACTIVE) && !graceful_restart_locks)
-    tm_start(gr_wait_timer, 0);
+    tm2_start(gr_wait_timer, 0);
 }
 
 
@@ -1329,7 +1328,7 @@ proto_schedule_down(struct proto *p, byte restart, byte code)
 
   p->down_sched = restart ? PDS_RESTART : PDS_DISABLE;
   p->down_code = code;
-  tm_start_max(proto_shutdown_timer, restart ? 2 : 0);
+  tm2_start_max(proto_shutdown_timer, restart ? 250 MS : 0);
 }
 
 
diff --git a/sysdep/unix/main.c b/sysdep/unix/main.c
index 6cab21ac..e3252b43 100644
--- a/sysdep/unix/main.c
+++ b/sysdep/unix/main.c
@@ -303,7 +303,7 @@ cmd_reconfig_undo_notify(void)
 }
 
 void
-cmd_reconfig(char *name, int type, int timeout)
+cmd_reconfig(char *name, int type, uint timeout)
 {
   if (cli_access_restricted())
     return;
diff --git a/sysdep/unix/unix.h b/sysdep/unix/unix.h
index 4b0fb005..cb12fad8 100644
--- a/sysdep/unix/unix.h
+++ b/sysdep/unix/unix.h
@@ -22,7 +22,7 @@ void async_config(void);
 void async_dump(void);
 void async_shutdown(void);
 void cmd_check_config(char *name);
-void cmd_reconfig(char *name, int type, int timeout);
+void cmd_reconfig(char *name, int type, uint timeout);
 void cmd_reconfig_confirm(void);
 void cmd_reconfig_undo(void);
 void cmd_shutdown(void);

From 6b5cd7c05f82e87d67c1b41296021757a3d952a0 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 21 Jun 2017 15:38:11 +0200
Subject: [PATCH 059/198] Sysdep: Remove old timer code

---
 lib/timer.c      |  18 +++
 sysdep/unix/io.c | 283 -----------------------------------------------
 2 files changed, 18 insertions(+), 283 deletions(-)

diff --git a/lib/timer.c b/lib/timer.c
index 3136a56b..338b0a1a 100644
--- a/lib/timer.c
+++ b/lib/timer.c
@@ -7,6 +7,24 @@
  *	Can be freely distributed and used under the terms of the GNU GPL.
  */
 
+/**
+ * DOC: Timers
+ *
+ * Timers are resources which represent a wish of a module to call a function at
+ * the specified time. The timer code does not guarantee exact timing, only that
+ * a timer function will not be called before the requested time.
+ *
+ * In BIRD, time is represented by values of the &btime type which is signed
+ * 64-bit integer interpreted as a relative number of microseconds since some
+ * fixed time point in past. The current time can be obtained by current_time()
+ * function with reasonable accuracy and is monotonic. There is also a current
+ * 'wall-clock' real time obtainable by current_real_time() reported by OS.
+ *
+ * Each timer is described by a &timer structure containing a pointer to the
+ * handler function (@hook), data private to this function (@data), time the
+ * function should be called at (@expires, 0 for inactive timers), for the other
+ * fields see |timer.h|.
+ */
 
 #include <stdio.h>
 #include <stdlib.h>
diff --git a/sysdep/unix/io.c b/sysdep/unix/io.c
index 25121d74..3e56a32d 100644
--- a/sysdep/unix/io.c
+++ b/sysdep/unix/io.c
@@ -103,289 +103,6 @@ tracked_fopen(pool *p, char *name, char *mode)
   return f;
 }
 
-/**
- * DOC: Timers
- *
- * Timers are resources which represent a wish of a module to call
- * a function at the specified time. The platform dependent code
- * doesn't guarantee exact timing, only that a timer function
- * won't be called before the requested time.
- *
- * In BIRD, time is represented by values of the &bird_clock_t type
- * which are integral numbers interpreted as a relative number of seconds since
- * some fixed time point in past. The current time can be read
- * from variable @now with reasonable accuracy and is monotonic. There is also
- * a current 'absolute' time in variable @now_real reported by OS.
- *
- * Each timer is described by a &timer structure containing a pointer
- * to the handler function (@hook), data private to this function (@data),
- * time the function should be called at (@expires, 0 for inactive timers),
- * for the other fields see |timer.h|.
- */
-
-#if 0
-#define NEAR_TIMER_LIMIT 4
-
-static list near_timers, far_timers;
-static bird_clock_t first_far_timer = TIME_INFINITY;
-
-
-/* now must be different from 0, because 0 is a special value in timer->expires */
-bird_clock_t now = 1, now_real, boot_time;
-
-static void
-update_times_plain(void)
-{
-  bird_clock_t new_time = time(NULL);
-  int delta = new_time - now_real;
-
-  if ((delta >= 0) && (delta < 60))
-    now += delta;
-  else if (now_real != 0)
-   log(L_WARN "Time jump, delta %d s", delta);
-
-  now_real = new_time;
-}
-
-static void
-update_times_gettime(void)
-{
-  struct timespec ts;
-  int rv;
-
-  rv = clock_gettime(CLOCK_MONOTONIC, &ts);
-  if (rv != 0)
-    die("clock_gettime: %m");
-
-  if (ts.tv_sec != now) {
-    if (ts.tv_sec < now)
-      log(L_ERR "Monotonic timer is broken");
-
-    now = ts.tv_sec;
-    now_real = time(NULL);
-  }
-}
-
-static int clock_monotonic_available;
-
-static inline void
-update_times(void)
-{
-  if (clock_monotonic_available)
-    update_times_gettime();
-  else
-    update_times_plain();
-}
-
-static inline void
-init_times(void)
-{
- struct timespec ts;
- clock_monotonic_available = (clock_gettime(CLOCK_MONOTONIC, &ts) == 0);
- if (!clock_monotonic_available)
-   log(L_WARN "Monotonic timer is missing");
-}
-
-static void
-tm_free(resource *r)
-{
-  timer *t = (timer *) r;
-
-  tm_stop(t);
-}
-
-static void
-tm_dump(resource *r)
-{
-  timer *t = (timer *) r;
-
-  debug("(code %p, data %p, ", t->hook, t->data);
-  if (t->randomize)
-    debug("rand %d, ", t->randomize);
-  if (t->recurrent)
-    debug("recur %d, ", t->recurrent);
-  if (t->expires)
-    debug("expires in %d sec)\n", t->expires - now);
-  else
-    debug("inactive)\n");
-}
-
-static struct resclass tm_class = {
-  "Timer",
-  sizeof(timer),
-  tm_free,
-  tm_dump,
-  NULL,
-  NULL
-};
-
-/**
- * tm_new - create a timer
- * @p: pool
- *
- * This function creates a new timer resource and returns
- * a pointer to it. To use the timer, you need to fill in
- * the structure fields and call tm_start() to start timing.
- */
-timer *
-tm_new(pool *p)
-{
-  timer *t = ralloc(p, &tm_class);
-  return t;
-}
-
-static inline void
-tm_insert_near(timer *t)
-{
-  node *n = HEAD(near_timers);
-
-  while (n->next && (SKIP_BACK(timer, n, n)->expires < t->expires))
-    n = n->next;
-  insert_node(&t->n, n->prev);
-}
-
-/**
- * tm_start - start a timer
- * @t: timer
- * @after: number of seconds the timer should be run after
- *
- * This function schedules the hook function of the timer to
- * be called after @after seconds. If the timer has been already
- * started, it's @expire time is replaced by the new value.
- *
- * You can have set the @randomize field of @t, the timeout
- * will be increased by a random number of seconds chosen
- * uniformly from range 0 .. @randomize.
- *
- * You can call tm_start() from the handler function of the timer
- * to request another run of the timer. Also, you can set the @recurrent
- * field to have the timer re-added automatically with the same timeout.
- */
-void
-tm_start(timer *t, unsigned after)
-{
-  bird_clock_t when;
-
-  if (t->randomize)
-    after += random() % (t->randomize + 1);
-  when = now + after;
-  if (t->expires == when)
-    return;
-  if (t->expires)
-    rem_node(&t->n);
-  t->expires = when;
-  if (after <= NEAR_TIMER_LIMIT)
-    tm_insert_near(t);
-  else
-    {
-      if (!first_far_timer || first_far_timer > when)
-	first_far_timer = when;
-      add_tail(&far_timers, &t->n);
-    }
-}
-
-/**
- * tm_stop - stop a timer
- * @t: timer
- *
- * This function stops a timer. If the timer is already stopped,
- * nothing happens.
- */
-void
-tm_stop(timer *t)
-{
-  if (t->expires)
-    {
-      rem_node(&t->n);
-      t->expires = 0;
-    }
-}
-
-static void
-tm_dump_them(char *name, list *l)
-{
-  node *n;
-  timer *t;
-
-  debug("%s timers:\n", name);
-  WALK_LIST(n, *l)
-    {
-      t = SKIP_BACK(timer, n, n);
-      debug("%p ", t);
-      tm_dump(&t->r);
-    }
-  debug("\n");
-}
-
-void
-tm_dump_all(void)
-{
-  tm_dump_them("Near", &near_timers);
-  tm_dump_them("Far", &far_timers);
-}
-
-static inline time_t
-tm_first_shot(void)
-{
-  time_t x = first_far_timer;
-
-  if (!EMPTY_LIST(near_timers))
-    {
-      timer *t = SKIP_BACK(timer, n, HEAD(near_timers));
-      if (t->expires < x)
-	x = t->expires;
-    }
-  return x;
-}
-
-void io_log_event(void *hook, void *data);
-
-static void
-tm_shot(void)
-{
-  timer *t;
-  node *n, *m;
-
-  if (first_far_timer <= now)
-    {
-      bird_clock_t limit = now + NEAR_TIMER_LIMIT;
-      first_far_timer = TIME_INFINITY;
-      n = HEAD(far_timers);
-      while (m = n->next)
-	{
-	  t = SKIP_BACK(timer, n, n);
-	  if (t->expires <= limit)
-	    {
-	      rem_node(n);
-	      tm_insert_near(t);
-	    }
-	  else if (t->expires < first_far_timer)
-	    first_far_timer = t->expires;
-	  n = m;
-	}
-    }
-  while ((n = HEAD(near_timers)) -> next)
-    {
-      int delay;
-      t = SKIP_BACK(timer, n, n);
-      if (t->expires > now)
-	break;
-      rem_node(n);
-      delay = t->expires - now;
-      t->expires = 0;
-      if (t->recurrent)
-	{
-	  int i = t->recurrent - delay;
-	  if (i < 0)
-	    i = 0;
-	  tm_start(t, i);
-	}
-      io_log_event(t->hook, t->data);
-      t->hook(t);
-    }
-}
-#endif
-
 
 /*
  *	Time clock

From 8b58f565e4fcd076e2d9fe008c7f2b19e264b319 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Fri, 13 Oct 2017 12:34:08 +0200
Subject: [PATCH 060/198] Babel: Update to new timers

---
 proto/babel/babel.c   | 184 ++++++++++++++++++++++--------------------
 proto/babel/babel.h   |  63 ++++++++-------
 proto/babel/config.Y  |   4 +-
 proto/babel/packets.c |  10 +--
 4 files changed, 136 insertions(+), 125 deletions(-)

diff --git a/proto/babel/babel.c b/proto/babel/babel.c
index bab2e8b3..8a39f022 100644
--- a/proto/babel/babel.c
+++ b/proto/babel/babel.c
@@ -70,7 +70,7 @@ babel_init_entry(void *E)
 {
   struct babel_entry *e = E;
 
-  e->updated = now;
+  e->updated = current_time();
   init_list(&e->sources);
   init_list(&e->routes);
 }
@@ -114,7 +114,7 @@ babel_get_source(struct babel_entry *e, u64 router_id)
 
   s = sl_alloc(p->source_slab);
   s->router_id = router_id;
-  s->expires = now + BABEL_GARBAGE_INTERVAL;
+  s->expires = current_time() + BABEL_GARBAGE_INTERVAL;
   s->seqno = 0;
   s->metric = BABEL_INFINITY;
   add_tail(&e->sources, NODE s);
@@ -127,10 +127,11 @@ babel_expire_sources(struct babel_entry *e)
 {
   struct babel_proto *p = e->proto;
   struct babel_source *n, *nx;
+  btime now_ = current_time();
 
   WALK_LIST_DELSAFE(n, nx, e->sources)
   {
-    if (n->expires && n->expires <= now)
+    if (n->expires && n->expires <= now_)
     {
       rem_node(NODE n);
       sl_free(p->source_slab, n);
@@ -167,7 +168,7 @@ babel_get_route(struct babel_entry *e, struct babel_neighbor *nbr)
   if (nbr)
   {
     r->neigh = nbr;
-    r->expires = now + BABEL_GARBAGE_INTERVAL;
+    r->expires = current_time() + BABEL_GARBAGE_INTERVAL;
     add_tail(&nbr->routes, NODE &r->neigh_route);
   }
 
@@ -208,7 +209,7 @@ babel_expire_route(struct babel_route *r)
   if (r->metric < BABEL_INFINITY)
   {
     r->metric = BABEL_INFINITY;
-    r->expires = now + r->expiry_interval;
+    r->expires = current_time() + r->expiry_interval;
   }
   else
   {
@@ -230,6 +231,7 @@ babel_expire_routes_(struct babel_proto *p UNUSED, struct fib *rtable)
 {
   struct babel_route *r, *rx;
   struct fib_iterator fit;
+  btime now_ = current_time();
 
   FIB_ITERATE_INIT(&fit, rtable);
 
@@ -240,10 +242,10 @@ loop:
 
     WALK_LIST_DELSAFE(r, rx, e->routes)
     {
-      if (r->refresh_time && r->refresh_time <= now)
+      if (r->refresh_time && r->refresh_time <= now_)
 	babel_refresh_route(r);
 
-      if (r->expires && r->expires <= now)
+      if (r->expires && r->expires <= now_)
       {
 	babel_expire_route(r);
 	changed = 1;
@@ -360,15 +362,16 @@ babel_expire_neighbors(struct babel_proto *p)
 {
   struct babel_iface *ifa;
   struct babel_neighbor *nbr, *nbx;
+  btime now_ = current_time();
 
   WALK_LIST(ifa, p->interfaces)
   {
     WALK_LIST_DELSAFE(nbr, nbx, ifa->neigh_list)
     {
-      if (nbr->ihu_expiry && nbr->ihu_expiry <= now)
+      if (nbr->ihu_expiry && nbr->ihu_expiry <= now_)
         babel_expire_ihu(nbr);
 
-      if (nbr->hello_expiry && nbr->hello_expiry <= now)
+      if (nbr->hello_expiry && nbr->hello_expiry <= now_)
         babel_expire_hello(nbr);
     }
   }
@@ -546,7 +549,7 @@ babel_select_route(struct babel_entry *e)
 	  e->n.addr, cur->router_id, cur->metric);
 
     e->selected_in = cur;
-    e->updated = now;
+    e->updated = current_time();
     babel_announce_rte(p, e);
   }
   else if (!cur || cur->metric == BABEL_INFINITY)
@@ -562,7 +565,7 @@ babel_select_route(struct babel_entry *e)
 	    e->n.addr);
 
       e->selected_in->metric = BABEL_INFINITY;
-      e->updated = now;
+      e->updated = current_time();
 
       babel_send_seqno_request(e);
       babel_announce_rte(p, e);
@@ -580,7 +583,7 @@ babel_select_route(struct babel_entry *e)
       TRACE(D_EVENTS, "Flushing route for prefix %N", e->n.addr);
 
       e->selected_in = NULL;
-      e->updated = now;
+      e->updated = current_time();
       babel_announce_rte(p, e);
     }
   }
@@ -614,8 +617,8 @@ babel_build_ihu(union babel_msg *msg, struct babel_iface *ifa, struct babel_neig
   msg->ihu.rxcost = babel_compute_rxcost(n);
   msg->ihu.interval = ifa->cf->ihu_interval;
 
-  TRACE(D_PACKETS, "Sending IHU for %I with rxcost %d interval %d",
-        msg->ihu.addr, msg->ihu.rxcost, msg->ihu.interval);
+  TRACE(D_PACKETS, "Sending IHU for %I with rxcost %d interval %t",
+        msg->ihu.addr, msg->ihu.rxcost, (btime) msg->ihu.interval);
 }
 
 static void
@@ -648,8 +651,8 @@ babel_send_hello(struct babel_iface *ifa, u8 send_ihu)
   msg.hello.seqno = ifa->hello_seqno++;
   msg.hello.interval = ifa->cf->hello_interval;
 
-  TRACE(D_PACKETS, "Sending hello on %s with seqno %d interval %d",
-	ifa->ifname, msg.hello.seqno, msg.hello.interval);
+  TRACE(D_PACKETS, "Sending hello on %s with seqno %d interval %t",
+	ifa->ifname, msg.hello.seqno, (btime) msg.hello.interval);
 
   babel_enqueue(&msg, ifa);
 
@@ -750,7 +753,7 @@ babel_unicast_seqno_request(struct babel_route *r)
  * transmitted entry is updated.
  */
 static void
-babel_send_update_(struct babel_iface *ifa, bird_clock_t changed, struct fib *rtable)
+babel_send_update_(struct babel_iface *ifa, btime changed, struct fib *rtable)
 {
   struct babel_proto *p = ifa->proto;
 
@@ -766,7 +769,7 @@ babel_send_update_(struct babel_iface *ifa, bird_clock_t changed, struct fib *rt
     if ((r->router_id == p->router_id) && (r->seqno < p->update_seqno))
     {
       r->seqno = p->update_seqno;
-      e->updated = now;
+      e->updated = current_time();
     }
 
     /* Skip routes that weren't updated since 'changed' time */
@@ -793,7 +796,7 @@ babel_send_update_(struct babel_iface *ifa, bird_clock_t changed, struct fib *rt
     if (!OUR_ROUTE(r))
     {
       struct babel_source *s = babel_get_source(e, r->router_id);
-      s->expires = now + BABEL_GARBAGE_INTERVAL;
+      s->expires = current_time() + BABEL_GARBAGE_INTERVAL;
 
       if ((msg.update.seqno > s->seqno) ||
 	  ((msg.update.seqno == s->seqno) && (msg.update.metric < s->metric)))
@@ -807,7 +810,7 @@ babel_send_update_(struct babel_iface *ifa, bird_clock_t changed, struct fib *rt
 }
 
 static void
-babel_send_update(struct babel_iface *ifa, bird_clock_t changed)
+babel_send_update(struct babel_iface *ifa, btime changed)
 {
   struct babel_proto *p = ifa->proto;
 
@@ -827,7 +830,7 @@ babel_trigger_iface_update(struct babel_iface *ifa)
   TRACE(D_EVENTS, "Scheduling triggered updates for %s seqno %d",
 	ifa->iface->name, p->update_seqno);
 
-  ifa->want_triggered = now;
+  ifa->want_triggered = current_time();
   babel_iface_kick_timer(ifa);
 }
 
@@ -887,7 +890,7 @@ babel_send_wildcard_retraction(struct babel_iface *ifa)
 
 /* Update hello history according to Appendix A1 of the RFC */
 static void
-babel_update_hello_history(struct babel_neighbor *n, u16 seqno, u16 interval)
+babel_update_hello_history(struct babel_neighbor *n, u16 seqno, uint interval)
 {
   /*
    * Compute the difference between expected and received seqno (modulo 2^16).
@@ -925,16 +928,18 @@ babel_update_hello_history(struct babel_neighbor *n, u16 seqno, u16 interval)
   n->hello_map = (n->hello_map << 1) | 1;
   n->next_hello_seqno = seqno+1;
   if (n->hello_cnt < 16) n->hello_cnt++;
-  n->hello_expiry = now + BABEL_HELLO_EXPIRY_FACTOR(interval);
+  n->hello_expiry = current_time() + BABEL_HELLO_EXPIRY_FACTOR(interval);
 }
 
 static void
 babel_expire_seqno_requests(struct babel_proto *p)
 {
+  btime now_ = current_time();
+
   struct babel_seqno_request *n, *nx;
   WALK_LIST_DELSAFE(n, nx, p->seqno_cache)
   {
-    if ((n->updated + BABEL_SEQNO_REQUEST_EXPIRY) <= now)
+    if ((n->updated + BABEL_SEQNO_REQUEST_EXPIRY) <= now_)
     {
       rem_node(NODE n);
       sl_free(p->seqno_slab, n);
@@ -963,7 +968,7 @@ babel_cache_seqno_request(struct babel_proto *p, net_addr *n,
   net_copy(&r->net, n);
   r->router_id = router_id;
   r->seqno = seqno;
-  r->updated = now;
+  r->updated = current_time();
   add_tail(&p->seqno_cache, NODE r);
 
   return 1;
@@ -1013,8 +1018,8 @@ babel_handle_ack_req(union babel_msg *m, struct babel_iface *ifa)
   struct babel_proto *p = ifa->proto;
   struct babel_msg_ack_req *msg = &m->ack_req;
 
-  TRACE(D_PACKETS, "Handling ACK request nonce %d interval %d",
-	msg->nonce, msg->interval);
+  TRACE(D_PACKETS, "Handling ACK request nonce %d interval %t",
+	msg->nonce, (btime) msg->interval);
 
   babel_send_ack(ifa, msg->sender, msg->nonce);
 }
@@ -1025,8 +1030,8 @@ babel_handle_hello(union babel_msg *m, struct babel_iface *ifa)
   struct babel_proto *p = ifa->proto;
   struct babel_msg_hello *msg = &m->hello;
 
-  TRACE(D_PACKETS, "Handling hello seqno %d interval %d",
-	msg->seqno, msg->interval);
+  TRACE(D_PACKETS, "Handling hello seqno %d interval %t",
+	msg->seqno, (btime) msg->interval);
 
   struct babel_neighbor *n = babel_get_neighbor(ifa, msg->sender);
   babel_update_hello_history(n, msg->seqno, msg->interval);
@@ -1044,12 +1049,12 @@ babel_handle_ihu(union babel_msg *m, struct babel_iface *ifa)
   if ((msg->ae != BABEL_AE_WILDCARD) && !ipa_equal(msg->addr, ifa->addr))
     return;
 
-  TRACE(D_PACKETS, "Handling IHU rxcost %d interval %d",
-	msg->rxcost, msg->interval);
+  TRACE(D_PACKETS, "Handling IHU rxcost %d interval %t",
+	msg->rxcost, (btime) msg->interval);
 
   struct babel_neighbor *n = babel_get_neighbor(ifa, msg->sender);
   n->txcost = msg->rxcost;
-  n->ihu_expiry = now + BABEL_IHU_EXPIRY_FACTOR(msg->interval);
+  n->ihu_expiry = current_time() + BABEL_IHU_EXPIRY_FACTOR(msg->interval);
 }
 
 /**
@@ -1214,9 +1219,9 @@ babel_handle_update(union babel_msg *m, struct babel_iface *ifa)
     r->seqno = msg->seqno;
 
     r->expiry_interval = BABEL_ROUTE_EXPIRY_FACTOR(msg->interval);
-    r->expires = now + r->expiry_interval;
+    r->expires = current_time() + r->expiry_interval;
     if (r->expiry_interval > BABEL_ROUTE_REFRESH_INTERVAL)
-      r->refresh_time = now + r->expiry_interval - BABEL_ROUTE_REFRESH_INTERVAL;
+      r->refresh_time = current_time() + r->expiry_interval - BABEL_ROUTE_REFRESH_INTERVAL;
 
     /* If the route is not feasible at this point, it means it is from another
        neighbour than the one currently selected; so send a unicast seqno
@@ -1256,7 +1261,7 @@ babel_handle_route_request(union babel_msg *m, struct babel_iface *ifa)
   else
   {
     babel_trigger_iface_update(ifa);
-    e->updated = now;
+    e->updated = current_time();
   }
 }
 
@@ -1283,7 +1288,7 @@ babel_handle_seqno_request(union babel_msg *m, struct babel_iface *ifa)
   if ((r->router_id != msg->router_id) || ge_mod64k(r->seqno, msg->seqno))
   {
     babel_trigger_iface_update(ifa);
-    e->updated = now;
+    e->updated = current_time();
     return;
   }
 
@@ -1333,42 +1338,44 @@ babel_iface_timer(timer *t)
 {
   struct babel_iface *ifa = t->data;
   struct babel_proto *p = ifa->proto;
-  bird_clock_t hello_period = ifa->cf->hello_interval;
-  bird_clock_t update_period = ifa->cf->update_interval;
+  btime hello_period = ifa->cf->hello_interval;
+  btime update_period = ifa->cf->update_interval;
+  btime now_ = current_time();
 
-  if (now >= ifa->next_hello)
+  if (now_ >= ifa->next_hello)
   {
     babel_send_hello(ifa, (ifa->cf->type == BABEL_IFACE_TYPE_WIRELESS ||
                            ifa->hello_seqno % BABEL_IHU_INTERVAL_FACTOR == 0));
-    ifa->next_hello +=  hello_period * (1 + (now - ifa->next_hello) / hello_period);
+    ifa->next_hello += hello_period * (1 + (now_ - ifa->next_hello) / hello_period);
   }
 
-  if (now >= ifa->next_regular)
+  if (now_ >= ifa->next_regular)
   {
     TRACE(D_EVENTS, "Sending regular updates on %s", ifa->ifname);
     babel_send_update(ifa, 0);
-    ifa->next_regular += update_period * (1 + (now - ifa->next_regular) / update_period);
+    ifa->next_regular += update_period * (1 + (now_ - ifa->next_regular) / update_period);
     ifa->want_triggered = 0;
     p->triggered = 0;
   }
-  else if (ifa->want_triggered && (now >= ifa->next_triggered))
+  else if (ifa->want_triggered && (now_ >= ifa->next_triggered))
   {
     TRACE(D_EVENTS, "Sending triggered updates on %s", ifa->ifname);
     babel_send_update(ifa, ifa->want_triggered);
-    ifa->next_triggered = now + MIN(5, update_period / 2 + 1);
+    ifa->next_triggered = now_ + MIN(5 S, update_period / 2);
     ifa->want_triggered = 0;
     p->triggered = 0;
   }
 
-  bird_clock_t next_event = MIN(ifa->next_hello, ifa->next_regular);
-  tm_start(ifa->timer, ifa->want_triggered ? 1 : (next_event - now));
+  btime next_event = MIN(ifa->next_hello, ifa->next_regular);
+  if (ifa->want_triggered) next_event = MIN(next_event, ifa->next_triggered);
+  tm2_set(ifa->timer, next_event);
 }
 
 static inline void
 babel_iface_kick_timer(struct babel_iface *ifa)
 {
-  if (ifa->timer->expires TO_S > (now + 1))
-    tm_start(ifa->timer, 1);
+  if (ifa->timer->expires > (current_time() + 100 MS))
+    tm2_start(ifa->timer, 100 MS);
 }
 
 static void
@@ -1378,11 +1385,11 @@ babel_iface_start(struct babel_iface *ifa)
 
   TRACE(D_EVENTS, "Starting interface %s", ifa->ifname);
 
-  ifa->next_hello = now + (random() % ifa->cf->hello_interval) + 1;
-  ifa->next_regular = now + (random() % ifa->cf->update_interval) + 1;
-  ifa->next_triggered = now + MIN(5, ifa->cf->update_interval / 2 + 1);
+  ifa->next_hello = current_time() + (random() % ifa->cf->hello_interval);
+  ifa->next_regular = current_time() + (random() % ifa->cf->update_interval);
+  ifa->next_triggered = current_time() + MIN(5 S, ifa->cf->update_interval / 2);
   ifa->want_triggered = 0;	/* We send an immediate update (below) */
-  tm_start(ifa->timer, 1);
+  tm2_start(ifa->timer, 100 MS);
   ifa->up = 1;
 
   babel_send_hello(ifa, 0);
@@ -1398,6 +1405,7 @@ babel_iface_stop(struct babel_iface *ifa)
   struct babel_neighbor *nbr;
   struct babel_route *r;
   node *n;
+  btime now_ = current_time();
 
   TRACE(D_EVENTS, "Stopping interface %s", ifa->ifname);
 
@@ -1412,12 +1420,12 @@ babel_iface_stop(struct babel_iface *ifa)
     {
       r = SKIP_BACK(struct babel_route, neigh_route, n);
       r->metric = BABEL_INFINITY;
-      r->expires = now + r->expiry_interval;
+      r->expires = now_ + r->expiry_interval;
       babel_select_route(r->e);
     }
   }
 
-  tm_stop(ifa->timer);
+  tm2_stop(ifa->timer);
   ifa->up = 0;
 }
 
@@ -1515,7 +1523,7 @@ babel_add_iface(struct babel_proto *p, struct iface *new, struct babel_iface_con
   init_list(&ifa->neigh_list);
   ifa->hello_seqno = 1;
 
-  ifa->timer = tm_new_set(ifa->pool, babel_iface_timer, ifa, 0, 0);
+  ifa->timer = tm2_new_init(ifa->pool, babel_iface_timer, ifa, 0, 0);
 
   init_list(&ifa->msg_queue);
   ifa->send_event = ev_new(ifa->pool);
@@ -1614,11 +1622,11 @@ babel_reconfigure_iface(struct babel_proto *p, struct babel_iface *ifa, struct b
   if (ipa_zero(ifa->next_hop_ip4) && p->ip4_channel)
     log(L_WARN "%s: Cannot find IPv4 next hop addr on %s", p->p.name, ifa->ifname);
 
-  if (ifa->next_hello > (now + new->hello_interval))
-    ifa->next_hello = now + (random() % new->hello_interval) + 1;
+  if (ifa->next_hello > (current_time() + new->hello_interval))
+    ifa->next_hello = current_time() + (random() % new->hello_interval);
 
-  if (ifa->next_regular > (now + new->update_interval))
-    ifa->next_regular = now + (random() % new->update_interval) + 1;
+  if (ifa->next_regular > (current_time() + new->update_interval))
+    ifa->next_regular = current_time() + (random() % new->update_interval);
 
   if ((new->tx_length != old->tx_length) || (new->rx_buffer != old->rx_buffer))
     babel_iface_update_buffers(ifa);
@@ -1680,18 +1688,19 @@ babel_reconfigure_ifaces(struct babel_proto *p, struct babel_config *cf)
 static void
 babel_dump_source(struct babel_source *s)
 {
-  debug("Source router_id %lR seqno %d metric %d expires %d\n",
-	s->router_id, s->seqno, s->metric, s->expires ? s->expires-now : 0);
+  debug("Source router_id %lR seqno %d metric %d expires %t\n",
+	s->router_id, s->seqno, s->metric,
+	s->expires ? s->expires - current_time() : 0);
 }
 
 static void
 babel_dump_route(struct babel_route *r)
 {
-  debug("Route neigh %I if %s seqno %d metric %d/%d router_id %lR expires %d\n",
+  debug("Route neigh %I if %s seqno %d metric %d/%d router_id %lR expires %t\n",
 	r->neigh ? r->neigh->addr : IPA_NONE,
         r->neigh ? r->neigh->ifa->ifname : "(none)",
-        r->seqno, r->advert_metric, r->metric,
-	r->router_id, r->expires ? r->expires-now : 0);
+        r->seqno, r->advert_metric, r->metric, r->router_id,
+	r->expires ? r->expires - current_time() : 0);
 }
 
 static void
@@ -1717,10 +1726,10 @@ babel_dump_entry(struct babel_entry *e)
 static void
 babel_dump_neighbor(struct babel_neighbor *n)
 {
-  debug("Neighbor %I txcost %d hello_map %x next seqno %d expires %d/%d\n",
+  debug("Neighbor %I txcost %d hello_map %x next seqno %d expires %t/%t\n",
 	n->addr, n->txcost, n->hello_map, n->next_hello_seqno,
-        n->hello_expiry ? n->hello_expiry - now : 0,
-        n->ihu_expiry ? n->ihu_expiry - now : 0);
+	n->hello_expiry ? n->hello_expiry - current_time() : 0,
+        n->ihu_expiry ? n->ihu_expiry - current_time() : 0);
 }
 
 static void
@@ -1728,7 +1737,7 @@ babel_dump_iface(struct babel_iface *ifa)
 {
   struct babel_neighbor *n;
 
-  debug("Babel: Interface %s addr %I rxcost %d type %d hello seqno %d intervals %d %d",
+  debug("Babel: Interface %s addr %I rxcost %d type %d hello seqno %d intervals %t %t",
 	ifa->ifname, ifa->addr, ifa->cf->rxcost, ifa->cf->type, ifa->hello_seqno,
 	ifa->cf->hello_interval, ifa->cf->update_interval);
   debug(" next hop v4 %I next hop v6 %I\n", ifa->next_hop_ip4, ifa->next_hop_ip6);
@@ -1803,7 +1812,7 @@ babel_show_interfaces(struct proto *P, char *iff)
   }
 
   cli_msg(-1023, "%s:", p->p.name);
-  cli_msg(-1023, "%-10s %-6s %7s %6s %6s %-15s %s",
+  cli_msg(-1023, "%-10s %-6s %7s %6s %7s %-15s %s",
 	  "Interface", "State", "RX cost", "Nbrs", "Timer",
 	  "Next hop (v4)", "Next hop (v6)");
 
@@ -1816,8 +1825,8 @@ babel_show_interfaces(struct proto *P, char *iff)
     WALK_LIST(nbr, ifa->neigh_list)
 	nbrs++;
 
-    int timer = MIN(ifa->next_regular, ifa->next_hello) - now;
-    cli_msg(-1023, "%-10s %-6s %7u %6u %6u %-15I %I",
+    btime timer = MIN(ifa->next_regular, ifa->next_hello) - current_time();
+    cli_msg(-1023, "%-10s %-6s %7u %6u %7t %-15I %I",
 	    ifa->iface->name, (ifa->up ? "Up" : "Down"),
 	    ifa->cf->rxcost, nbrs, MAX(timer, 0),
 	    ifa->next_hop_ip4, ifa->next_hop_ip6);
@@ -1842,8 +1851,8 @@ babel_show_neighbors(struct proto *P, char *iff)
   }
 
   cli_msg(-1024, "%s:", p->p.name);
-  cli_msg(-1024, "%-25s %-10s %6s %6s %10s",
-	  "IP address", "Interface", "Metric", "Routes", "Next hello");
+  cli_msg(-1024, "%-25s %-10s %6s %6s %6s %7s",
+	  "IP address", "Interface", "Metric", "Routes", "Hellos", "Expires");
 
   WALK_LIST(ifa, p->interfaces)
   {
@@ -1856,9 +1865,10 @@ babel_show_neighbors(struct proto *P, char *iff)
       WALK_LIST(r, n->routes)
         rts++;
 
-      int timer = n->hello_expiry - now;
-      cli_msg(-1024, "%-25I %-10s %6u %6u %10u",
-	      n->addr, ifa->iface->name, n->txcost, rts, MAX(timer, 0));
+      uint hellos = u32_popcount(n->hello_map);
+      btime timer = n->hello_expiry - current_time();
+      cli_msg(-1024, "%-25I %-10s %6u %6u %6u %7t",
+	      n->addr, ifa->iface->name, n->txcost, rts, hellos, MAX(timer, 0));
     }
   }
 
@@ -1888,8 +1898,8 @@ babel_show_entries_(struct babel_proto *p, struct fib *rtable)
       else
         bsprintf(ridbuf, "%lR", r->router_id);
 
-      int time = r->expires ? r->expires - now : 0;
-      cli_msg(-1025, "%-29N %-23s %6u %5u %7u %7u",
+      btime time = r->expires ? r->expires - current_time() : 0;
+      cli_msg(-1025, "%-29N %-23s %6u %5u %7t %7u",
 	      e->n.addr, ridbuf, r->metric, r->seqno, MAX(time, 0), srcs);
     }
     else
@@ -1948,8 +1958,8 @@ babel_timer(timer *t)
 static inline void
 babel_kick_timer(struct babel_proto *p)
 {
-  if (p->timer->expires TO_S > (now + 1))
-    tm_start(p->timer, 1);
+  if (p->timer->expires > (current_time() + 100 MS))
+    tm2_start(p->timer, 100 MS);
 }
 
 
@@ -2032,9 +2042,9 @@ babel_rt_notify(struct proto *P, struct channel *c UNUSED, struct network *net,
 
     if (r != e->selected_out)
     {
-      e->selected_out = r;
-      e->updated = now;
       babel_trigger_update(p);
+      e->updated = current_time();
+      e->selected_out = r;
     }
   }
   else
@@ -2051,10 +2061,10 @@ babel_rt_notify(struct proto *P, struct channel *c UNUSED, struct network *net,
        * expiry time. This causes a retraction to be sent, and later the route
        * to be flushed once the hold time has passed.
        */
-      e->selected_out->metric = BABEL_INFINITY;
-      e->selected_out->expires = now + BABEL_HOLD_TIME;
-      e->updated = now;
       babel_trigger_update(p);
+      e->updated = current_time();
+      e->selected_out->metric = BABEL_INFINITY;
+      e->selected_out->expires = current_time() + BABEL_HOLD_TIME;
     }
     else
     {
@@ -2121,8 +2131,8 @@ babel_start(struct proto *P)
 	   OFFSETOF(struct babel_entry, n), 0, babel_init_entry);
 
   init_list(&p->interfaces);
-  p->timer = tm_new_set(P->pool, babel_timer, p, 0, 1);
-  tm_start(p->timer, 2);
+  p->timer = tm2_new_init(P->pool, babel_timer, p, 1 S, 0);
+  tm2_start(p->timer, 1 S);
   p->update_seqno = 1;
   p->router_id = proto_get_router_id(&cf->c);
 
diff --git a/proto/babel/babel.h b/proto/babel/babel.h
index fccb60c9..95ea3036 100644
--- a/proto/babel/babel.h
+++ b/proto/babel/babel.h
@@ -32,25 +32,26 @@
 #define BABEL_INFINITY		0xFFFF
 
 
-#define BABEL_HELLO_INTERVAL_WIRED	4	/* Default hello intervals in seconds */
-#define BABEL_HELLO_INTERVAL_WIRELESS	4
+#define BABEL_HELLO_INTERVAL_WIRED	(4 S_)	/* Default hello intervals in seconds */
+#define BABEL_HELLO_INTERVAL_WIRELESS	(4 S_)
 #define BABEL_UPDATE_INTERVAL_FACTOR	4
 #define BABEL_IHU_INTERVAL_FACTOR	3
-#define BABEL_IHU_EXPIRY_FACTOR(X)	((X)*3/2)	/* 1.5 */
-#define BABEL_HELLO_EXPIRY_FACTOR(X)	((X)*3/2)	/* 1.5 */
-#define BABEL_ROUTE_EXPIRY_FACTOR(X)	((X)*7/2)	/* 3.5 */
-#define BABEL_ROUTE_REFRESH_INTERVAL	2	/* Seconds before route expiry to send route request */
-#define BABEL_HOLD_TIME			10	/* Expiry time for our own routes */
+#define BABEL_IHU_EXPIRY_FACTOR(X)	((btime)(X)*3/2)	/* 1.5 */
+#define BABEL_HELLO_EXPIRY_FACTOR(X)	((btime)(X)*3/2)	/* 1.5 */
+#define BABEL_ROUTE_EXPIRY_FACTOR(X)	((btime)(X)*7/2)	/* 3.5 */
+#define BABEL_ROUTE_REFRESH_INTERVAL	(2 S_)	/* Time before route expiry to send route request */
+#define BABEL_HOLD_TIME			(10 S_)	/* Expiry time for our own routes */
+#define BABEL_SEQNO_REQUEST_EXPIRY	(60 S_)
+#define BABEL_GARBAGE_INTERVAL		(300 S_)
 #define BABEL_RXCOST_WIRED		96
 #define BABEL_RXCOST_WIRELESS		256
 #define BABEL_INITIAL_HOP_COUNT		255
-#define BABEL_MAX_SEND_INTERVAL		5
-#define BABEL_TIME_UNITS		100	/* On-wire times are counted in centiseconds */
-#define BABEL_SEQNO_REQUEST_EXPIRY	60
-#define BABEL_GARBAGE_INTERVAL		300
+#define BABEL_MAX_SEND_INTERVAL		5	/* Unused ? */
 
 /* Max interval that will not overflow when carried as 16-bit centiseconds */
-#define BABEL_MAX_INTERVAL		(0xFFFF/BABEL_TIME_UNITS)
+#define BABEL_TIME_UNITS		10000	/* On-wire times are counted in centiseconds */
+#define BABEL_MIN_INTERVAL		(0x0001 * BABEL_TIME_UNITS)
+#define BABEL_MAX_INTERVAL		(0xFFFF * BABEL_TIME_UNITS)
 
 #define BABEL_OVERHEAD		(IP6_HEADER_LENGTH+UDP_HEADER_LENGTH)
 #define BABEL_MIN_MTU		(512 + BABEL_OVERHEAD)
@@ -113,9 +114,9 @@ struct babel_iface_config {
   u8 type;
   u8 check_link;
   uint port;
-  u16 hello_interval;
-  u16 ihu_interval;
-  u16 update_interval;
+  uint hello_interval;			/* Hello interval, in us */
+  uint ihu_interval;			/* IHU interval, in us */
+  uint update_interval;			/* Update interval, in us */
 
   u16 rx_buffer;			/* RX buffer size, 0 for MTU */
   u16 tx_length;			/* TX packet length limit (including headers), 0 for MTU */
@@ -172,10 +173,10 @@ struct babel_iface {
 
   u16 hello_seqno;			/* To be increased on each hello */
 
-  bird_clock_t next_hello;
-  bird_clock_t next_regular;
-  bird_clock_t next_triggered;
-  bird_clock_t want_triggered;
+  btime next_hello;
+  btime next_regular;
+  btime next_triggered;
+  btime want_triggered;
 
   timer *timer;
   event *send_event;
@@ -191,8 +192,8 @@ struct babel_neighbor {
   u16 hello_map;
   u16 next_hello_seqno;
   /* expiry timers */
-  bird_clock_t hello_expiry;
-  bird_clock_t ihu_expiry;
+  btime hello_expiry;
+  btime ihu_expiry;
 
   list routes;				/* Routes this neighbour has sent us (struct babel_route) */
 };
@@ -203,7 +204,7 @@ struct babel_source {
   u64 router_id;
   u16 seqno;
   u16 metric;
-  bird_clock_t expires;
+  btime expires;
 };
 
 struct babel_route {
@@ -217,9 +218,9 @@ struct babel_route {
   u16 metric;
   u64 router_id;
   ip_addr next_hop;
-  bird_clock_t refresh_time;
-  bird_clock_t expires;
-  u16 expiry_interval;
+  btime refresh_time;
+  btime expires;
+  btime expiry_interval;
 };
 
 struct babel_entry {
@@ -227,7 +228,7 @@ struct babel_entry {
   struct babel_route *selected_in;
   struct babel_route *selected_out;
 
-  bird_clock_t updated;
+  btime updated;
 
   list sources;				/* Source entries for this prefix (struct babel_source). */
   list routes;				/* Routes for this prefix (struct babel_route) */
@@ -241,7 +242,7 @@ struct babel_seqno_request {
   net_addr net;
   u64 router_id;
   u16 seqno;
-  bird_clock_t updated;
+  btime updated;
 };
 
 
@@ -252,7 +253,7 @@ struct babel_seqno_request {
 struct babel_msg_ack_req {
   u8 type;
   u16 nonce;
-  u16 interval;
+  uint interval;
   ip_addr sender;
 };
 
@@ -264,7 +265,7 @@ struct babel_msg_ack {
 struct babel_msg_hello {
   u8 type;
   u16 seqno;
-  u16 interval;
+  uint interval;
   ip_addr sender;
 };
 
@@ -272,7 +273,7 @@ struct babel_msg_ihu {
   u8 type;
   u8 ae;
   u16 rxcost;
-  u16 interval;
+  uint interval;
   ip_addr addr;
   ip_addr sender;
 };
@@ -280,7 +281,7 @@ struct babel_msg_ihu {
 struct babel_msg_update {
   u8 type;
   u8 wildcard;
-  u16 interval;
+  uint interval;
   u16 seqno;
   u16 metric;
   u64 router_id;
diff --git a/proto/babel/config.Y b/proto/babel/config.Y
index 93c92790..27c50676 100644
--- a/proto/babel/config.Y
+++ b/proto/babel/config.Y
@@ -89,10 +89,10 @@ babel_iface_finish:
 babel_iface_item:
  | PORT expr { BABEL_IFACE->port = $2; if (($2<1) || ($2>65535)) cf_error("Invalid port number"); }
  | RXCOST expr { BABEL_IFACE->rxcost = $2; if (($2<1) || ($2>65535)) cf_error("Invalid rxcost"); }
- | HELLO INTERVAL expr { BABEL_IFACE->hello_interval = $3; if (($3<1) || ($3>BABEL_MAX_INTERVAL)) cf_error("Invalid hello interval"); }
- | UPDATE INTERVAL expr { BABEL_IFACE->update_interval = $3; if (($3<1) || ($3>BABEL_MAX_INTERVAL)) cf_error("Invalid update interval"); }
  | TYPE WIRED { BABEL_IFACE->type = BABEL_IFACE_TYPE_WIRED; }
  | TYPE WIRELESS { BABEL_IFACE->type = BABEL_IFACE_TYPE_WIRELESS; }
+ | HELLO INTERVAL expr_us { BABEL_IFACE->hello_interval = $3; if (($3<BABEL_MIN_INTERVAL) || ($3>BABEL_MAX_INTERVAL)) cf_error("Hello interval must be in range 10 ms - 655 s"); }
+ | UPDATE INTERVAL expr_us { BABEL_IFACE->update_interval = $3; if (($3<BABEL_MIN_INTERVAL) || ($3>BABEL_MAX_INTERVAL)) cf_error("Update interval must be in range 10 ms - 655 s"); }
  | RX BUFFER expr { BABEL_IFACE->rx_buffer = $3; if (($3<256) || ($3>65535)) cf_error("RX buffer must be in range 256-65535"); }
  | TX LENGTH expr { BABEL_IFACE->tx_length = $3; if (($3<256) || ($3>65535)) cf_error("TX length must be in range 256-65535"); }
  | TX tos { BABEL_IFACE->tx_tos = $2; }
diff --git a/proto/babel/packets.c b/proto/babel/packets.c
index d136efe8..6e8d2eb5 100644
--- a/proto/babel/packets.c
+++ b/proto/babel/packets.c
@@ -162,17 +162,17 @@ bytes_equal(u8 *b1, u8 *b2, uint maxlen)
   return i;
 }
 
-static inline u16
+static inline uint
 get_time16(const void *p)
 {
-  u16 v = get_u16(p) / BABEL_TIME_UNITS;
-  return MAX(1, v);
+  uint v = get_u16(p) * BABEL_TIME_UNITS;
+  return MAX(BABEL_MIN_INTERVAL, v);
 }
 
 static inline void
-put_time16(void *p, u16 v)
+put_time16(void *p, uint v)
 {
-  put_u16(p, v * BABEL_TIME_UNITS);
+  put_u16(p, v / BABEL_TIME_UNITS);
 }
 
 static inline void

From 5ee69d11f2e859b77ff04bb4068f43082fd1794f Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Fri, 13 Oct 2017 19:33:42 +0200
Subject: [PATCH 061/198] Babel: Fix Hello and IHU expiration

---
 proto/babel/babel.c | 6 +++++-
 proto/babel/babel.h | 1 +
 2 files changed, 6 insertions(+), 1 deletion(-)

diff --git a/proto/babel/babel.c b/proto/babel/babel.c
index 8a39f022..10974474 100644
--- a/proto/babel/babel.c
+++ b/proto/babel/babel.c
@@ -343,6 +343,7 @@ static void
 babel_expire_ihu(struct babel_neighbor *nbr)
 {
   nbr->txcost = BABEL_INFINITY;
+  nbr->ihu_expiry = 0;
 }
 
 static void
@@ -353,7 +354,9 @@ babel_expire_hello(struct babel_neighbor *nbr)
   if (nbr->hello_cnt < 16)
     nbr->hello_cnt++;
 
-  if (!nbr->hello_map)
+  if (nbr->hello_map)
+    nbr->hello_expiry += nbr->last_hello_int;
+  else
     babel_flush_neighbor(nbr);
 }
 
@@ -929,6 +932,7 @@ babel_update_hello_history(struct babel_neighbor *n, u16 seqno, uint interval)
   n->next_hello_seqno = seqno+1;
   if (n->hello_cnt < 16) n->hello_cnt++;
   n->hello_expiry = current_time() + BABEL_HELLO_EXPIRY_FACTOR(interval);
+  n->last_hello_int = interval;
 }
 
 static void
diff --git a/proto/babel/babel.h b/proto/babel/babel.h
index 95ea3036..1b857f09 100644
--- a/proto/babel/babel.h
+++ b/proto/babel/babel.h
@@ -191,6 +191,7 @@ struct babel_neighbor {
   u8 hello_cnt;
   u16 hello_map;
   u16 next_hello_seqno;
+  uint last_hello_int;
   /* expiry timers */
   btime hello_expiry;
   btime ihu_expiry;

From 38f4721092f5a835248f9f521d34990ec9883dbd Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Fri, 13 Oct 2017 19:34:34 +0200
Subject: [PATCH 062/198] Babel: Avoid batch seqno updates

---
 proto/babel/babel.c | 11 +++++++++--
 proto/babel/babel.h |  1 +
 2 files changed, 10 insertions(+), 2 deletions(-)

diff --git a/proto/babel/babel.c b/proto/babel/babel.c
index 10974474..54b31c22 100644
--- a/proto/babel/babel.c
+++ b/proto/babel/babel.c
@@ -760,6 +760,13 @@ babel_send_update_(struct babel_iface *ifa, btime changed, struct fib *rtable)
 {
   struct babel_proto *p = ifa->proto;
 
+  /* Update increase was requested */
+  if (p->update_seqno_inc)
+  {
+    p->update_seqno++;
+    p->update_seqno_inc = 0;
+  }
+
   FIB_WALK(rtable, struct babel_entry, e)
   {
     struct babel_route *r = e->selected_out;
@@ -1299,8 +1306,8 @@ babel_handle_seqno_request(union babel_msg *m, struct babel_iface *ifa)
   /* Seqno is larger; check if we own the router id */
   if (msg->router_id == p->router_id)
   {
-    /* Ours; update seqno and trigger global update */
-    p->update_seqno++;
+    /* Ours; seqno increase and trigger global update */
+    p->update_seqno_inc = 1;
     babel_trigger_update(p);
   }
   else
diff --git a/proto/babel/babel.h b/proto/babel/babel.h
index 1b857f09..d0c159f9 100644
--- a/proto/babel/babel.h
+++ b/proto/babel/babel.h
@@ -139,6 +139,7 @@ struct babel_proto {
   list interfaces;			/* Interfaces we really know about (struct babel_iface) */
   u64 router_id;
   u16 update_seqno;			/* To be increased on request */
+  u8 update_seqno_inc;			/* Request for update_seqno increase */
   u8 triggered;				/* For triggering global updates */
 
   slab *route_slab;

From f00221fadbb2c85c835cc5e4e69a0d3ce13d31b3 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Fri, 13 Oct 2017 23:46:41 +0200
Subject: [PATCH 063/198] Babel: Fix unicast seqno requests

---
 proto/babel/babel.c | 44 ++++++++++++++++++--------------------------
 1 file changed, 18 insertions(+), 26 deletions(-)

diff --git a/proto/babel/babel.c b/proto/babel/babel.c
index 54b31c22..0b7ab91b 100644
--- a/proto/babel/babel.c
+++ b/proto/babel/babel.c
@@ -721,28 +721,24 @@ babel_send_seqno_request(struct babel_entry *e)
 }
 
 static void
-babel_unicast_seqno_request(struct babel_route *r)
+babel_unicast_seqno_request(struct babel_entry *e, struct babel_source *s, struct babel_neighbor *nbr)
 {
-  struct babel_entry *e = r->e;
   struct babel_proto *p = e->proto;
-  struct babel_iface *ifa = r->neigh->ifa;
-  struct babel_source *s = NULL;
   union babel_msg msg = {};
 
-  s = babel_find_source(e, r->router_id);
-  if (!s || !babel_cache_seqno_request(p, e->n.addr, r->router_id, s->seqno + 1))
+  if (!s || !babel_cache_seqno_request(p, e->n.addr, s->router_id, s->seqno + 1))
     return;
 
   TRACE(D_PACKETS, "Sending seqno request for %N router-id %lR seqno %d",
-	e->n.addr, r->router_id, s->seqno + 1);
+	e->n.addr, s->router_id, s->seqno + 1);
 
   msg.type = BABEL_TLV_SEQNO_REQUEST;
   msg.seqno_request.hop_count = BABEL_INITIAL_HOP_COUNT;
   msg.seqno_request.seqno = s->seqno + 1;
-  msg.seqno_request.router_id = r->router_id;
+  msg.seqno_request.router_id = s->router_id;
   net_copy(&msg.seqno_request.net, e->n.addr);
 
-  babel_send_unicast(&msg, ifa, r->neigh->addr);
+  babel_send_unicast(&msg, nbr->ifa, nbr->addr);
 }
 
 /**
@@ -1088,9 +1084,9 @@ babel_handle_update(union babel_msg *m, struct babel_iface *ifa)
   struct babel_neighbor *nbr;
   struct babel_entry *e;
   struct babel_source *s;
-  struct babel_route *r;
+  struct babel_route *r, *best;
   node *n;
-  int feasible;
+  int feasible, metric;
 
   if (msg->wildcard)
     TRACE(D_PACKETS, "Handling wildcard retraction", msg->seqno);
@@ -1192,6 +1188,13 @@ babel_handle_update(union babel_msg *m, struct babel_iface *ifa)
   r = babel_find_route(e, nbr); /* the route entry indexed by neighbour */
   s = babel_find_source(e, msg->router_id); /* for feasibility */
   feasible = babel_is_feasible(s, msg->seqno, msg->metric);
+  metric = babel_compute_metric(nbr, msg->metric);
+  best = e->selected_in;
+
+  /* RFC section 3.8.2.2 - Dealing with unfeasible updates */
+  if (!feasible && (metric != BABEL_INFINITY) &&
+      (!best || (r == best) || (metric < best->metric)))
+    babel_unicast_seqno_request(e, s, nbr);
 
   if (!r)
   {
@@ -1201,18 +1204,13 @@ babel_handle_update(union babel_msg *m, struct babel_iface *ifa)
     r = babel_get_route(e, nbr);
     r->advert_metric = msg->metric;
     r->router_id = msg->router_id;
-    r->metric = babel_compute_metric(nbr, msg->metric);
+    r->metric = metric;
     r->next_hop = msg->next_hop;
     r->seqno = msg->seqno;
   }
-  else if (r == r->e->selected_in && !feasible)
+  else if (r == best && !feasible)
   {
-    /*
-     * Route is installed and update is infeasible - we may lose the route,
-     * so send a unicast seqno request (section 3.8.2.2 second paragraph).
-     */
-    babel_unicast_seqno_request(r);
-
+    /* Penultimate paragraph above - ignore or retract */
     if (msg->router_id == r->router_id)
       return;
 
@@ -1223,7 +1221,7 @@ babel_handle_update(union babel_msg *m, struct babel_iface *ifa)
   {
     /* Last paragraph above - update the entry */
     r->advert_metric = msg->metric;
-    r->metric = babel_compute_metric(nbr, msg->metric);
+    r->metric = metric;
     r->next_hop = msg->next_hop;
 
     r->router_id = msg->router_id;
@@ -1233,12 +1231,6 @@ babel_handle_update(union babel_msg *m, struct babel_iface *ifa)
     r->expires = current_time() + r->expiry_interval;
     if (r->expiry_interval > BABEL_ROUTE_REFRESH_INTERVAL)
       r->refresh_time = current_time() + r->expiry_interval - BABEL_ROUTE_REFRESH_INTERVAL;
-
-    /* If the route is not feasible at this point, it means it is from another
-       neighbour than the one currently selected; so send a unicast seqno
-       request to try to get a better route (section 3.8.2.2 last paragraph). */
-    if (!feasible)
-      babel_unicast_seqno_request(r);
   }
 
   babel_select_route(e);

From b47eaefe12d0673af2c7c7ec1a8adff982a958ca Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 25 Oct 2017 17:14:08 +0200
Subject: [PATCH 064/198] Babel: Revamp cost computation and run route
 selection when cost change

Also fix several minor bugs and add 'limit' option for k-out-of-j
link sensing strategy. Change default from 8-of-16 to 12-of-16.
Change IHU expiry factor from 1.5 to 3.5 (as in RFC 6126).
---
 doc/bird.sgml        |  38 +++++---
 proto/babel/babel.c  | 228 +++++++++++++++++++++++++------------------
 proto/babel/babel.h  |  10 +-
 proto/babel/config.Y |   2 +
 4 files changed, 170 insertions(+), 108 deletions(-)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index 65efebb8..0ee86c18 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -1616,6 +1616,7 @@ protocol babel [<name>] {
 	interface <interface pattern> {
 		type <wired|wireless>;
 		rxcost <number>;
+		limit <number>;
 		hello interval <number>;
 		update interval <number>;
 		port <number>;
@@ -1632,23 +1633,34 @@ protocol babel [<name>] {
 
 <descrip>
       <tag><label id="babel-type">type wired|wireless </tag>
-      This option specifies the interface type: Wired or wireless. Wired
-      interfaces are considered more reliable, and so the default hello
-      interval is higher, and a neighbour is considered unreachable after only
-      a small number of "hello" packets are lost. On wireless interfaces,
-      hello packets are sent more often, and the ETX link quality estimation
-      technique is used to compute the metrics of routes discovered over this
-      interface. This technique will gradually degrade the metric of routes
-      when packets are lost rather than the more binary up/down mechanism of
-      wired type links. Default: <cf/wired/.
+      This option specifies the interface type: Wired or wireless. On wired
+      interfaces a neighbor is considered unreachable after a small number of
+      Hello packets are lost, as described by <cf/limit/ option. On wireless
+      interfaces the ETX link quality estimation technique is used to compute
+      the metrics of routes discovered over this interface. This technique will
+      gradually degrade the metric of routes when packets are lost rather than
+      the more binary up/down mechanism of wired type links. Default:
+      <cf/wired/.
 
       <tag><label id="babel-rxcost">rxcost <m/num/</tag>
-      This specifies the RX cost of the interface. The route metrics will be
-      computed from this value with a mechanism determined by the interface
-      <cf/type/. Default: 96 for wired interfaces, 256 for wireless.
+      This option specifies the nominal RX cost of the interface. The effective
+      neighbor costs for route metrics will be computed from this value with a
+      mechanism determined by the interface <cf/type/. Note that in contrast to
+      other routing protocols like RIP or OSPF, the <cf/rxcost/ specifies the
+      cost of RX instead of TX, so it affects primarily neighbors' route
+      selection and not local route selection. Default: 96 for wired interfaces,
+      256 for wireless.
+
+      <tag><label id="babel-limit">limit <m/num/</tag>
+      BIRD keeps track of received Hello messages from each neighbor to
+      establish neighbor reachability. For wired type interfaces, this option
+      specifies how many of last 16 hellos have to be correctly received in
+      order to neighbor is assumed to be up. The option is ignored on wireless
+      type interfaces, where gradual cost degradation is used instead of sharp
+      limit. Default: 12.
 
       <tag><label id="babel-hello">hello interval <m/num/</tag>
-      Interval at which periodic "hello" messages are sent on this interface,
+      Interval at which periodic Hello messages are sent on this interface,
       in seconds. Default: 4 seconds.
 
       <tag><label id="babel-update">update interval <m/num/</tag>
diff --git a/proto/babel/babel.c b/proto/babel/babel.c
index 0b7ab91b..07c82d91 100644
--- a/proto/babel/babel.c
+++ b/proto/babel/babel.c
@@ -57,6 +57,7 @@ static int  babel_cache_seqno_request(struct babel_proto *p, net_addr *n, u64 ro
 static void babel_trigger_iface_update(struct babel_iface *ifa);
 static void babel_trigger_update(struct babel_proto *p);
 static void babel_send_seqno_request(struct babel_entry *e);
+static void babel_update_cost(struct babel_neighbor *n);
 static inline void babel_kick_timer(struct babel_proto *p);
 static inline void babel_iface_kick_timer(struct babel_iface *ifa);
 
@@ -208,7 +209,7 @@ babel_expire_route(struct babel_route *r)
 
   if (r->metric < BABEL_INFINITY)
   {
-    r->metric = BABEL_INFINITY;
+    r->metric = r->advert_metric = BABEL_INFINITY;
     r->expires = current_time() + r->expiry_interval;
   }
   else
@@ -300,15 +301,20 @@ babel_find_neighbor(struct babel_iface *ifa, ip_addr addr)
 static struct babel_neighbor *
 babel_get_neighbor(struct babel_iface *ifa, ip_addr addr)
 {
+  struct babel_proto *p = ifa->proto;
   struct babel_neighbor *nbr = babel_find_neighbor(ifa, addr);
 
   if (nbr)
     return nbr;
 
+  TRACE(D_EVENTS, "New neighbor %I on %s", addr, ifa->iface->name);
+
   nbr = mb_allocz(ifa->pool, sizeof(struct babel_neighbor));
   nbr->ifa = ifa;
   nbr->addr = addr;
+  nbr->rxcost = BABEL_INFINITY;
   nbr->txcost = BABEL_INFINITY;
+  nbr->cost = BABEL_INFINITY;
   init_list(&nbr->routes);
   add_tail(&ifa->neigh_list, NODE nbr);
 
@@ -316,12 +322,11 @@ babel_get_neighbor(struct babel_iface *ifa, ip_addr addr)
 }
 
 static void
-babel_flush_neighbor(struct babel_neighbor *nbr)
+babel_flush_neighbor(struct babel_proto *p, struct babel_neighbor *nbr)
 {
-  struct babel_proto *p = nbr->ifa->proto;
   node *n;
 
-  TRACE(D_EVENTS, "Flushing neighbor %I", nbr->addr);
+  TRACE(D_EVENTS, "Removing neighbor %I on %s", nbr->addr, nbr->ifa->iface->name);
 
   WALK_LIST_FIRST(n, nbr->routes)
   {
@@ -340,24 +345,33 @@ babel_flush_neighbor(struct babel_neighbor *nbr)
 }
 
 static void
-babel_expire_ihu(struct babel_neighbor *nbr)
+babel_expire_ihu(struct babel_proto *p, struct babel_neighbor *nbr)
 {
+  TRACE(D_EVENTS, "IHU from nbr %I on %s expired", nbr->addr, nbr->ifa->iface->name);
+
   nbr->txcost = BABEL_INFINITY;
   nbr->ihu_expiry = 0;
+  babel_update_cost(nbr);
 }
 
 static void
-babel_expire_hello(struct babel_neighbor *nbr)
+babel_expire_hello(struct babel_proto *p, struct babel_neighbor *nbr)
 {
   nbr->hello_map <<= 1;
 
   if (nbr->hello_cnt < 16)
     nbr->hello_cnt++;
 
+  TRACE(D_EVENTS, "Hello from nbr %I on %s expired, %d left",
+	nbr->addr, nbr->ifa->iface->name, u32_popcount(nbr->hello_map));
+
   if (nbr->hello_map)
+  {
     nbr->hello_expiry += nbr->last_hello_int;
+    babel_update_cost(nbr);
+  }
   else
-    babel_flush_neighbor(nbr);
+    babel_flush_neighbor(p, nbr);
 }
 
 static void
@@ -372,10 +386,10 @@ babel_expire_neighbors(struct babel_proto *p)
     WALK_LIST_DELSAFE(nbr, nbx, ifa->neigh_list)
     {
       if (nbr->ihu_expiry && nbr->ihu_expiry <= now_)
-        babel_expire_ihu(nbr);
+        babel_expire_ihu(p, nbr);
 
       if (nbr->hello_expiry && nbr->hello_expiry <= now_)
-        babel_expire_hello(nbr);
+        babel_expire_hello(p, nbr);
     }
   }
 }
@@ -409,66 +423,81 @@ babel_is_feasible(struct babel_source *s, u16 seqno, u16 metric)
     ((seqno == s->seqno) && (metric < s->metric));
 }
 
-static u16
-babel_compute_rxcost(struct babel_neighbor *n)
-{
-  struct babel_iface *ifa = n->ifa;
-  u8 cnt, missed;
-  u16 map=n->hello_map;
-
-  if (!map) return BABEL_INFINITY;
-  cnt = u32_popcount(map); // number of bits set
-  missed = n->hello_cnt-cnt;
-
-  if (ifa->cf->type == BABEL_IFACE_TYPE_WIRELESS)
-  {
-    /* ETX - Appendix 2.2 in the RFC.
-
-       beta = prob. of successful transmission.
-       rxcost = BABEL_RXCOST_WIRELESS/beta
-
-       Since: beta = 1-missed/n->hello_cnt = cnt/n->hello_cnt
-       Then: rxcost = BABEL_RXCOST_WIRELESS * n->hello_cnt / cnt
-   */
-    if (!cnt) return BABEL_INFINITY;
-    return BABEL_RXCOST_WIRELESS * n->hello_cnt / cnt;
-  }
-  else
-  {
-    /* k-out-of-j selection - Appendix 2.1 in the RFC. */
-    DBG("Babel: Missed %d hellos from %I\n", missed, n->addr);
-    /* Link is bad if more than half the expected hellos were lost */
-    return (missed > n->hello_cnt/2) ? BABEL_INFINITY : ifa->cf->rxcost;
-  }
-}
-
-
-static u16
-babel_compute_cost(struct babel_neighbor *n)
-{
-  struct babel_iface *ifa = n->ifa;
-  u16 rxcost = babel_compute_rxcost(n);
-  if (rxcost == BABEL_INFINITY) return rxcost;
-  else if (ifa->cf->type == BABEL_IFACE_TYPE_WIRELESS)
-  {
-    /* ETX - Appendix 2.2 in the RFC */
-    return (MAX(n->txcost, BABEL_RXCOST_WIRELESS) * rxcost)/BABEL_RXCOST_WIRELESS;
-  }
-  else
-  {
-    /* k-out-of-j selection - Appendix 2.1 in the RFC. */
-    return n->txcost;
-  }
-}
-
 /* Simple additive metric - Appendix 3.1 in the RFC */
-static u16
+static inline u16
 babel_compute_metric(struct babel_neighbor *n, uint metric)
 {
-  metric += babel_compute_cost(n);
-  return MIN(metric, BABEL_INFINITY);
+  return MIN(metric + n->cost, BABEL_INFINITY);
 }
 
+static void
+babel_update_cost(struct babel_neighbor *nbr)
+{
+  struct babel_proto *p = nbr->ifa->proto;
+  struct babel_iface_config *cf = nbr->ifa->cf;
+  uint rcv = u32_popcount(nbr->hello_map); // number of bits set
+  uint max = nbr->hello_cnt;
+  uint rxcost = BABEL_INFINITY;	/* Cost to announce in IHU */
+  uint txcost = BABEL_INFINITY;	/* Effective cost for route selection */
+
+  if (!rcv || !nbr->ifa->up)
+    goto done;
+
+  switch (cf->type)
+  {
+  case BABEL_IFACE_TYPE_WIRED:
+    /* k-out-of-j selection - Appendix 2.1 in the RFC. */
+
+    /* Link is bad if less than cf->limit/16 of expected hellos were received */
+    if (rcv * 16 < cf->limit * max)
+      break;
+
+    rxcost =  cf->rxcost;
+    txcost = nbr->txcost;
+    break;
+
+  case BABEL_IFACE_TYPE_WIRELESS:
+    /*
+     * ETX - Appendix 2.2 in the RFC.
+     *
+     * alpha  = prob. of successful transmission estimated by the neighbor
+     * beta   = prob. of successful transmission estimated by the router
+     * rxcost = nominal rxcost of the router / beta
+     * txcost = nominal rxcost of the neighbor / (alpha * beta)
+     *        = received txcost / beta
+     *
+     * Note that received txcost is just neighbor's rxcost. Beta is rcv/max,
+     * we use inverse values of beta (i.e. max/rcv) to stay in integers.
+     */
+    rxcost = MIN( cf->rxcost * max / rcv, BABEL_INFINITY);
+    txcost = MIN(nbr->txcost * max / rcv, BABEL_INFINITY);
+    break;
+  }
+
+done:
+  /* If RX cost changed, send IHU with next Hello */
+  if (rxcost != nbr->rxcost)
+  {
+    nbr->rxcost = rxcost;
+    nbr->ihu_cnt = 0;
+  }
+
+  /* If link cost changed, run route selection */
+  if (txcost != nbr->cost)
+  {
+    TRACE(D_EVENTS, "Cost of nbr %I on %s changed from %u to %u",
+	  nbr->addr, nbr->ifa->iface->name, nbr->cost, txcost);
+
+    nbr->cost = txcost;
+
+    struct babel_route *r; node *n;
+    WALK_LIST2(r, n, nbr->routes, neigh_route)
+    {
+      r->metric = babel_compute_metric(nbr, r->advert_metric);
+      babel_select_route(r->e);
+    }
+  }
+}
 
 /**
  * babel_announce_rte - announce selected route to the core
@@ -506,6 +535,7 @@ babel_announce_rte(struct babel_proto *p, struct babel_entry *e)
     rte->u.babel.router_id = r->router_id;
     rte->pflags = 0;
 
+    r->old_metric = r->metric;
     rte_update2(c, e->n.addr, rte, p->p.main_source);
   }
   else
@@ -535,7 +565,7 @@ static void
 babel_select_route(struct babel_entry *e)
 {
   struct babel_proto *p = e->proto;
-  struct babel_route *r, *cur = e->selected_in;
+  struct babel_route *r, *cur = e->selected_in, *best = e->selected_in;
 
   /* try to find the best feasible route */
   WALK_LIST(r, e->routes)
@@ -544,12 +574,12 @@ babel_select_route(struct babel_entry *e)
         babel_is_feasible(babel_find_source(e, r->router_id), r->seqno, r->advert_metric))
       cur = r;
 
-  if (cur && !OUR_ROUTE(cur) &&
-      ((!e->selected_in && cur->metric < BABEL_INFINITY) ||
-       (e->selected_in && cur->metric < e->selected_in->metric)))
+  if (cur && !OUR_ROUTE(cur) && (cur->metric < BABEL_INFINITY) &&
+      (!best || (cur->metric < best->metric) || ((cur == best) && (cur->metric != cur->old_metric))))
   {
-    TRACE(D_EVENTS, "Picked new route for prefix %N: router id %lR metric %d",
-	  e->n.addr, cur->router_id, cur->metric);
+    if (cur != best)
+      TRACE(D_EVENTS, "Picked new route for prefix %N: router id %lR metric %d",
+	    e->n.addr, cur->router_id, cur->metric);
 
     e->selected_in = cur;
     e->updated = current_time();
@@ -564,10 +594,9 @@ babel_select_route(struct babel_entry *e)
        babel_build_rte() will set the unreachable flag if the metric is BABEL_INFINITY.*/
     if (e->selected_in)
     {
-      TRACE(D_EVENTS, "Lost feasible route for prefix %N",
-	    e->n.addr);
+      TRACE(D_EVENTS, "Lost feasible route for prefix %N", e->n.addr);
 
-      e->selected_in->metric = BABEL_INFINITY;
+      e->selected_in->metric = e->selected_in->advert_metric = BABEL_INFINITY;
       e->updated = current_time();
 
       babel_send_seqno_request(e);
@@ -583,7 +612,7 @@ babel_select_route(struct babel_entry *e)
       /* No route currently selected, and no new one selected; this means we
 	 don't have a route to this destination anymore (and were probably
 	 called from an expiry timer). Remove the route from the nest. */
-      TRACE(D_EVENTS, "Flushing route for prefix %N", e->n.addr);
+      // TRACE(D_EVENTS, "Flushing route for prefix %N", e->n.addr);
 
       e->selected_in = NULL;
       e->updated = current_time();
@@ -617,7 +646,7 @@ babel_build_ihu(union babel_msg *msg, struct babel_iface *ifa, struct babel_neig
 
   msg->type = BABEL_TLV_IHU;
   msg->ihu.addr = n->addr;
-  msg->ihu.rxcost = babel_compute_rxcost(n);
+  msg->ihu.rxcost = n->rxcost;
   msg->ihu.interval = ifa->cf->ihu_interval;
 
   TRACE(D_PACKETS, "Sending IHU for %I with rxcost %d interval %t",
@@ -630,6 +659,7 @@ babel_send_ihu(struct babel_iface *ifa, struct babel_neighbor *n)
   union babel_msg msg = {};
   babel_build_ihu(&msg, ifa, n);
   babel_send_unicast(&msg, ifa, n->addr);
+  n->ihu_cnt = BABEL_IHU_INTERVAL_FACTOR;
 }
 
 static void
@@ -638,14 +668,18 @@ babel_send_ihus(struct babel_iface *ifa)
   struct babel_neighbor *n;
   WALK_LIST(n, ifa->neigh_list)
   {
-    union babel_msg msg = {};
-    babel_build_ihu(&msg, ifa, n);
-    babel_enqueue(&msg, ifa);
+    if (n->hello_cnt && (--n->ihu_cnt <= 0))
+    {
+      union babel_msg msg = {};
+      babel_build_ihu(&msg, ifa, n);
+      babel_enqueue(&msg, ifa);
+      n->ihu_cnt = BABEL_IHU_INTERVAL_FACTOR;
+    }
   }
 }
 
 static void
-babel_send_hello(struct babel_iface *ifa, u8 send_ihu)
+babel_send_hello(struct babel_iface *ifa)
 {
   struct babel_proto *p = ifa->proto;
   union babel_msg msg = {};
@@ -659,8 +693,7 @@ babel_send_hello(struct babel_iface *ifa, u8 send_ihu)
 
   babel_enqueue(&msg, ifa);
 
-  if (send_ihu)
-    babel_send_ihus(ifa);
+  babel_send_ihus(ifa);
 }
 
 static void
@@ -907,7 +940,7 @@ babel_update_hello_history(struct babel_neighbor *n, u16 seqno, uint interval)
 
   u16 delta = ((uint) seqno - (uint) n->next_hello_seqno);
 
-  if (delta == 0)
+  if ((delta == 0) || (n->hello_cnt == 0))
   {
     /* Do nothing */
   }
@@ -934,6 +967,8 @@ babel_update_hello_history(struct babel_neighbor *n, u16 seqno, uint interval)
   n->hello_map = (n->hello_map << 1) | 1;
   n->next_hello_seqno = seqno+1;
   if (n->hello_cnt < 16) n->hello_cnt++;
+
+  /* Update expiration */
   n->hello_expiry = current_time() + BABEL_HELLO_EXPIRY_FACTOR(interval);
   n->last_hello_int = interval;
 }
@@ -1041,8 +1076,13 @@ babel_handle_hello(union babel_msg *m, struct babel_iface *ifa)
 	msg->seqno, (btime) msg->interval);
 
   struct babel_neighbor *n = babel_get_neighbor(ifa, msg->sender);
+  int first_hello = !n->hello_cnt;
+
   babel_update_hello_history(n, msg->seqno, msg->interval);
-  if (ifa->cf->type == BABEL_IFACE_TYPE_WIRELESS)
+  babel_update_cost(n);
+
+  /* Speed up session establishment by sending IHU immediately */
+  if (first_hello)
     babel_send_ihu(ifa, n);
 }
 
@@ -1062,6 +1102,7 @@ babel_handle_ihu(union babel_msg *m, struct babel_iface *ifa)
   struct babel_neighbor *n = babel_get_neighbor(ifa, msg->sender);
   n->txcost = msg->rxcost;
   n->ihu_expiry = current_time() + BABEL_IHU_EXPIRY_FACTOR(msg->interval);
+  babel_update_cost(n);
 }
 
 /**
@@ -1159,7 +1200,7 @@ babel_handle_update(union babel_msg *m, struct babel_iface *ifa)
       WALK_LIST(n, nbr->routes)
       {
 	r = SKIP_BACK(struct babel_route, neigh_route, n);
-	r->metric = BABEL_INFINITY;
+	r->metric = r->advert_metric = BABEL_INFINITY;
 	babel_select_route(r->e);
       }
     }
@@ -1176,7 +1217,7 @@ babel_handle_update(union babel_msg *m, struct babel_iface *ifa)
       if (!r)
 	return;
 
-      r->metric = BABEL_INFINITY;
+      r->metric = r->advert_metric = BABEL_INFINITY;
       babel_select_route(e);
     }
 
@@ -1215,7 +1256,7 @@ babel_handle_update(union babel_msg *m, struct babel_iface *ifa)
       return;
 
     /* Treat as retraction */
-    r->metric = BABEL_INFINITY;
+    r->metric = r->advert_metric = BABEL_INFINITY;
   }
   else
   {
@@ -1347,8 +1388,7 @@ babel_iface_timer(timer *t)
 
   if (now_ >= ifa->next_hello)
   {
-    babel_send_hello(ifa, (ifa->cf->type == BABEL_IFACE_TYPE_WIRELESS ||
-                           ifa->hello_seqno % BABEL_IHU_INTERVAL_FACTOR == 0));
+    babel_send_hello(ifa);
     ifa->next_hello += hello_period * (1 + (now_ - ifa->next_hello) / hello_period);
   }
 
@@ -1395,7 +1435,7 @@ babel_iface_start(struct babel_iface *ifa)
   tm2_start(ifa->timer, 100 MS);
   ifa->up = 1;
 
-  babel_send_hello(ifa, 0);
+  babel_send_hello(ifa);
   babel_send_wildcard_retraction(ifa);
   babel_send_wildcard_request(ifa);
   babel_send_update(ifa, 0);	/* Full update */
@@ -1422,9 +1462,11 @@ babel_iface_stop(struct babel_iface *ifa)
     WALK_LIST(n, nbr->routes)
     {
       r = SKIP_BACK(struct babel_route, neigh_route, n);
-      r->metric = BABEL_INFINITY;
+      r->metric = r->advert_metric = BABEL_INFINITY;
       r->expires = now_ + r->expiry_interval;
-      babel_select_route(r->e);
+
+      if (r == r->e->selected_in)
+	babel_select_route(r->e);
     }
   }
 
@@ -1551,7 +1593,7 @@ babel_remove_iface(struct babel_proto *p, struct babel_iface *ifa)
 
   struct babel_neighbor *n;
   WALK_LIST_FIRST(n, ifa->neigh_list)
-    babel_flush_neighbor(n);
+    babel_flush_neighbor(p, n);
 
   rem_node(NODE ifa);
 
@@ -1871,7 +1913,7 @@ babel_show_neighbors(struct proto *P, char *iff)
       uint hellos = u32_popcount(n->hello_map);
       btime timer = n->hello_expiry - current_time();
       cli_msg(-1024, "%-25I %-10s %6u %6u %6u %7t",
-	      n->addr, ifa->iface->name, n->txcost, rts, hellos, MAX(timer, 0));
+	      n->addr, ifa->iface->name, n->cost, rts, hellos, MAX(timer, 0));
     }
   }
 
diff --git a/proto/babel/babel.h b/proto/babel/babel.h
index d0c159f9..78841254 100644
--- a/proto/babel/babel.h
+++ b/proto/babel/babel.h
@@ -34,9 +34,10 @@
 
 #define BABEL_HELLO_INTERVAL_WIRED	(4 S_)	/* Default hello intervals in seconds */
 #define BABEL_HELLO_INTERVAL_WIRELESS	(4 S_)
+#define BABEL_HELLO_LIMIT		12
 #define BABEL_UPDATE_INTERVAL_FACTOR	4
 #define BABEL_IHU_INTERVAL_FACTOR	3
-#define BABEL_IHU_EXPIRY_FACTOR(X)	((btime)(X)*3/2)	/* 1.5 */
+#define BABEL_IHU_EXPIRY_FACTOR(X)	((btime)(X)*7/2)	/* 3.5 */
 #define BABEL_HELLO_EXPIRY_FACTOR(X)	((btime)(X)*3/2)	/* 1.5 */
 #define BABEL_ROUTE_EXPIRY_FACTOR(X)	((btime)(X)*7/2)	/* 3.5 */
 #define BABEL_ROUTE_REFRESH_INTERVAL	(2 S_)	/* Time before route expiry to send route request */
@@ -112,6 +113,7 @@ struct babel_iface_config {
 
   u16 rxcost;
   u8 type;
+  u8 limit;				/* Minimum number of Hellos to keep link up */
   u8 check_link;
   uint port;
   uint hello_interval;			/* Hello interval, in us */
@@ -188,7 +190,10 @@ struct babel_neighbor {
   struct babel_iface *ifa;
 
   ip_addr addr;
-  u16 txcost;
+  u16 rxcost;				/* Sent in last IHU */
+  u16 txcost;				/* Received in last IHU */
+  u16 cost;				/* Computed neighbor cost */
+  s8 ihu_cnt;				/* IHU countdown, 0 to send it */
   u8 hello_cnt;
   u16 hello_map;
   u16 next_hello_seqno;
@@ -218,6 +223,7 @@ struct babel_route {
   u16 seqno;
   u16 advert_metric;
   u16 metric;
+  u16 old_metric;
   u64 router_id;
   ip_addr next_hop;
   btime refresh_time;
diff --git a/proto/babel/config.Y b/proto/babel/config.Y
index 27c50676..c72c00a3 100644
--- a/proto/babel/config.Y
+++ b/proto/babel/config.Y
@@ -56,6 +56,7 @@ babel_iface_start:
   init_list(&this_ipatt->ipn_list);
   BABEL_IFACE->port = BABEL_PORT;
   BABEL_IFACE->type = BABEL_IFACE_TYPE_WIRED;
+  BABEL_IFACE->limit = BABEL_HELLO_LIMIT;
   BABEL_IFACE->tx_tos = IP_PREC_INTERNET_CONTROL;
   BABEL_IFACE->tx_priority = sk_priority_control;
   BABEL_IFACE->check_link = 1;
@@ -89,6 +90,7 @@ babel_iface_finish:
 babel_iface_item:
  | PORT expr { BABEL_IFACE->port = $2; if (($2<1) || ($2>65535)) cf_error("Invalid port number"); }
  | RXCOST expr { BABEL_IFACE->rxcost = $2; if (($2<1) || ($2>65535)) cf_error("Invalid rxcost"); }
+ | LIMIT expr { BABEL_IFACE->limit = $2; if (($2<1) || ($2>16)) cf_error("Limit must be in range 1-16"); }
  | TYPE WIRED { BABEL_IFACE->type = BABEL_IFACE_TYPE_WIRED; }
  | TYPE WIRELESS { BABEL_IFACE->type = BABEL_IFACE_TYPE_WIRELESS; }
  | HELLO INTERVAL expr_us { BABEL_IFACE->hello_interval = $3; if (($3<BABEL_MIN_INTERVAL) || ($3>BABEL_MAX_INTERVAL)) cf_error("Hello interval must be in range 10 ms - 655 s"); }

From 738a57b69bdff6244cf5093ae3997290e6c11324 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 25 Oct 2017 17:59:57 +0200
Subject: [PATCH 065/198] Babel: Fix hello timeout for short hello intervals

---
 proto/babel/babel.c | 14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)

diff --git a/proto/babel/babel.c b/proto/babel/babel.c
index 07c82d91..2009c0d0 100644
--- a/proto/babel/babel.c
+++ b/proto/babel/babel.c
@@ -355,21 +355,25 @@ babel_expire_ihu(struct babel_proto *p, struct babel_neighbor *nbr)
 }
 
 static void
-babel_expire_hello(struct babel_proto *p, struct babel_neighbor *nbr)
+babel_expire_hello(struct babel_proto *p, struct babel_neighbor *nbr, btime now_)
 {
+again:
   nbr->hello_map <<= 1;
 
   if (nbr->hello_cnt < 16)
     nbr->hello_cnt++;
 
+  nbr->hello_expiry += nbr->last_hello_int;
+
+  /* We may expire multiple hellos if last_hello_int is too short */
+  if (nbr->hello_map && nbr->hello_expiry <= now_)
+    goto again;
+
   TRACE(D_EVENTS, "Hello from nbr %I on %s expired, %d left",
 	nbr->addr, nbr->ifa->iface->name, u32_popcount(nbr->hello_map));
 
   if (nbr->hello_map)
-  {
-    nbr->hello_expiry += nbr->last_hello_int;
     babel_update_cost(nbr);
-  }
   else
     babel_flush_neighbor(p, nbr);
 }
@@ -389,7 +393,7 @@ babel_expire_neighbors(struct babel_proto *p)
         babel_expire_ihu(p, nbr);
 
       if (nbr->hello_expiry && nbr->hello_expiry <= now_)
-        babel_expire_hello(p, nbr);
+        babel_expire_hello(p, nbr, now_);
     }
   }
 }

From 268dc7c8b3c45412fc8f4b0bae451e51c4c48b31 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Fri, 27 Oct 2017 12:20:58 +0200
Subject: [PATCH 066/198] Babel: Remove babel_proto ptr from babel_entry

---
 proto/babel/babel.c | 80 +++++++++++++++++++--------------------------
 proto/babel/babel.h |  1 -
 2 files changed, 34 insertions(+), 47 deletions(-)

diff --git a/proto/babel/babel.c b/proto/babel/babel.c
index 2009c0d0..f5aef492 100644
--- a/proto/babel/babel.c
+++ b/proto/babel/babel.c
@@ -50,13 +50,13 @@ static inline int ge_mod64k(uint a, uint b)
 
 static void babel_dump_entry(struct babel_entry *e);
 static void babel_dump_route(struct babel_route *r);
-static void babel_select_route(struct babel_entry *e);
-static void babel_send_route_request(struct babel_entry *e, struct babel_neighbor *n);
+static void babel_select_route(struct babel_proto *p, struct babel_entry *e);
+static void babel_send_route_request(struct babel_proto *p, struct babel_entry *e, struct babel_neighbor *n);
 static void babel_send_wildcard_request(struct babel_iface *ifa);
 static int  babel_cache_seqno_request(struct babel_proto *p, net_addr *n, u64 router_id, u16 seqno);
 static void babel_trigger_iface_update(struct babel_iface *ifa);
 static void babel_trigger_update(struct babel_proto *p);
-static void babel_send_seqno_request(struct babel_entry *e);
+static void babel_send_seqno_request(struct babel_proto *p, struct babel_entry *e);
 static void babel_update_cost(struct babel_neighbor *n);
 static inline void babel_kick_timer(struct babel_proto *p);
 static inline void babel_iface_kick_timer(struct babel_iface *ifa);
@@ -88,7 +88,6 @@ babel_get_entry(struct babel_proto *p, const net_addr *n)
 {
   struct fib *rtable = (n->type == NET_IP4) ? &p->ip4_rtable : &p->ip6_rtable;
   struct babel_entry *e = fib_get(rtable, n);
-  e->proto = p;
   return e;
 }
 
@@ -105,9 +104,8 @@ babel_find_source(struct babel_entry *e, u64 router_id)
 }
 
 static struct babel_source *
-babel_get_source(struct babel_entry *e, u64 router_id)
+babel_get_source(struct babel_proto *p, struct babel_entry *e, u64 router_id)
 {
-  struct babel_proto *p = e->proto;
   struct babel_source *s = babel_find_source(e, router_id);
 
   if (s)
@@ -124,9 +122,8 @@ babel_get_source(struct babel_entry *e, u64 router_id)
 }
 
 static void
-babel_expire_sources(struct babel_entry *e)
+babel_expire_sources(struct babel_proto *p, struct babel_entry *e)
 {
-  struct babel_proto *p = e->proto;
   struct babel_source *n, *nx;
   btime now_ = current_time();
 
@@ -153,9 +150,8 @@ babel_find_route(struct babel_entry *e, struct babel_neighbor *n)
 }
 
 static struct babel_route *
-babel_get_route(struct babel_entry *e, struct babel_neighbor *nbr)
+babel_get_route(struct babel_proto *p, struct babel_entry *e, struct babel_neighbor *nbr)
 {
-  struct babel_proto *p = e->proto;
   struct babel_route *r = babel_find_route(e, nbr);
 
   if (r)
@@ -177,10 +173,8 @@ babel_get_route(struct babel_entry *e, struct babel_neighbor *nbr)
 }
 
 static void
-babel_flush_route(struct babel_route *r)
+babel_flush_route(struct babel_proto *p, struct babel_route *r)
 {
-  struct babel_proto *p = r->e->proto;
-
   DBG("Babel: Flush route %N router_id %lR neigh %I\n",
       r->e->n.addr, r->router_id, r->neigh ? r->neigh->addr : IPA_NONE);
 
@@ -199,9 +193,8 @@ babel_flush_route(struct babel_route *r)
 }
 
 static void
-babel_expire_route(struct babel_route *r)
+babel_expire_route(struct babel_proto *p, struct babel_route *r)
 {
-  struct babel_proto *p = r->e->proto;
   struct babel_entry *e = r->e;
 
   TRACE(D_EVENTS, "Route expiry timer for %N router-id %lR fired",
@@ -214,21 +207,21 @@ babel_expire_route(struct babel_route *r)
   }
   else
   {
-    babel_flush_route(r);
+    babel_flush_route(p, r);
   }
 }
 
 static void
-babel_refresh_route(struct babel_route *r)
+babel_refresh_route(struct babel_proto *p, struct babel_route *r)
 {
   if (!OUR_ROUTE(r) && (r == r->e->selected_in))
-    babel_send_route_request(r->e, r->neigh);
+    babel_send_route_request(p, r->e, r->neigh);
 
   r->refresh_time = 0;
 }
 
 static void
-babel_expire_routes_(struct babel_proto *p UNUSED, struct fib *rtable)
+babel_expire_routes_(struct babel_proto *p, struct fib *rtable)
 {
   struct babel_route *r, *rx;
   struct fib_iterator fit;
@@ -244,11 +237,11 @@ loop:
     WALK_LIST_DELSAFE(r, rx, e->routes)
     {
       if (r->refresh_time && r->refresh_time <= now_)
-	babel_refresh_route(r);
+	babel_refresh_route(p, r);
 
       if (r->expires && r->expires <= now_)
       {
-	babel_expire_route(r);
+	babel_expire_route(p, r);
 	changed = 1;
       }
     }
@@ -262,11 +255,11 @@ loop:
        */
 
       FIB_ITERATE_PUT(&fit);
-      babel_select_route(e);
+      babel_select_route(p, e);
       goto loop;
     }
 
-    babel_expire_sources(e);
+    babel_expire_sources(p, e);
 
     /* Remove empty entries */
     if (EMPTY_LIST(e->sources) && EMPTY_LIST(e->routes))
@@ -334,10 +327,10 @@ babel_flush_neighbor(struct babel_proto *p, struct babel_neighbor *nbr)
     struct babel_entry *e = r->e;
     int selected = (r == e->selected_in);
 
-    babel_flush_route(r);
+    babel_flush_route(p, r);
 
     if (selected)
-      babel_select_route(e);
+      babel_select_route(p, e);
   }
 
   rem_node(NODE nbr);
@@ -498,7 +491,7 @@ done:
     WALK_LIST2(r, n, nbr->routes, neigh_route)
     {
       r->metric = babel_compute_metric(nbr, r->advert_metric);
-      babel_select_route(r->e);
+      babel_select_route(p, r->e);
     }
   }
 }
@@ -566,9 +559,8 @@ babel_announce_rte(struct babel_proto *p, struct babel_entry *e)
  * route is removed from the nest entirely.
  */
 static void
-babel_select_route(struct babel_entry *e)
+babel_select_route(struct babel_proto *p, struct babel_entry *e)
 {
-  struct babel_proto *p = e->proto;
   struct babel_route *r, *cur = e->selected_in, *best = e->selected_in;
 
   /* try to find the best feasible route */
@@ -603,7 +595,7 @@ babel_select_route(struct babel_entry *e)
       e->selected_in->metric = e->selected_in->advert_metric = BABEL_INFINITY;
       e->updated = current_time();
 
-      babel_send_seqno_request(e);
+      babel_send_seqno_request(p, e);
       babel_announce_rte(p, e);
 
       /* Section 3.6 of the RFC forbids an infeasible from being selected. This
@@ -701,9 +693,8 @@ babel_send_hello(struct babel_iface *ifa)
 }
 
 static void
-babel_send_route_request(struct babel_entry *e, struct babel_neighbor *n)
+babel_send_route_request(struct babel_proto *p, struct babel_entry *e, struct babel_neighbor *n)
 {
-  struct babel_proto *p = e->proto;
   struct babel_iface *ifa = n->ifa;
   union babel_msg msg = {};
 
@@ -732,9 +723,8 @@ babel_send_wildcard_request(struct babel_iface *ifa)
 }
 
 static void
-babel_send_seqno_request(struct babel_entry *e)
+babel_send_seqno_request(struct babel_proto *p, struct babel_entry *e)
 {
-  struct babel_proto *p = e->proto;
   struct babel_route *r = e->selected_in;
   struct babel_iface *ifa = NULL;
   struct babel_source *s = NULL;
@@ -758,9 +748,8 @@ babel_send_seqno_request(struct babel_entry *e)
 }
 
 static void
-babel_unicast_seqno_request(struct babel_entry *e, struct babel_source *s, struct babel_neighbor *nbr)
+babel_unicast_seqno_request(struct babel_proto *p, struct babel_entry *e, struct babel_source *s, struct babel_neighbor *nbr)
 {
-  struct babel_proto *p = e->proto;
   union babel_msg msg = {};
 
   if (!s || !babel_cache_seqno_request(p, e->n.addr, s->router_id, s->seqno + 1))
@@ -838,7 +827,7 @@ babel_send_update_(struct babel_iface *ifa, btime changed, struct fib *rtable)
     /* Update feasibility distance for redistributed routes */
     if (!OUR_ROUTE(r))
     {
-      struct babel_source *s = babel_get_source(e, r->router_id);
+      struct babel_source *s = babel_get_source(p, e, r->router_id);
       s->expires = current_time() + BABEL_GARBAGE_INTERVAL;
 
       if ((msg.update.seqno > s->seqno) ||
@@ -1021,11 +1010,10 @@ babel_cache_seqno_request(struct babel_proto *p, net_addr *n,
 }
 
 static void
-babel_forward_seqno_request(struct babel_entry *e,
+babel_forward_seqno_request(struct babel_proto *p, struct babel_entry *e,
                             struct babel_msg_seqno_request *in,
                             ip_addr sender)
 {
-  struct babel_proto *p = e->proto;
   struct babel_route *r;
 
   TRACE(D_PACKETS, "Forwarding seqno request for %N router-id %lR seqno %d",
@@ -1205,7 +1193,7 @@ babel_handle_update(union babel_msg *m, struct babel_iface *ifa)
       {
 	r = SKIP_BACK(struct babel_route, neigh_route, n);
 	r->metric = r->advert_metric = BABEL_INFINITY;
-	babel_select_route(r->e);
+	babel_select_route(p, r->e);
       }
     }
     else
@@ -1222,7 +1210,7 @@ babel_handle_update(union babel_msg *m, struct babel_iface *ifa)
 	return;
 
       r->metric = r->advert_metric = BABEL_INFINITY;
-      babel_select_route(e);
+      babel_select_route(p, e);
     }
 
     /* Done with retractions */
@@ -1239,14 +1227,14 @@ babel_handle_update(union babel_msg *m, struct babel_iface *ifa)
   /* RFC section 3.8.2.2 - Dealing with unfeasible updates */
   if (!feasible && (metric != BABEL_INFINITY) &&
       (!best || (r == best) || (metric < best->metric)))
-    babel_unicast_seqno_request(e, s, nbr);
+    babel_unicast_seqno_request(p, e, s, nbr);
 
   if (!r)
   {
     if (!feasible)
       return;
 
-    r = babel_get_route(e, nbr);
+    r = babel_get_route(p, e, nbr);
     r->advert_metric = msg->metric;
     r->router_id = msg->router_id;
     r->metric = metric;
@@ -1278,7 +1266,7 @@ babel_handle_update(union babel_msg *m, struct babel_iface *ifa)
       r->refresh_time = current_time() + r->expiry_interval - BABEL_ROUTE_REFRESH_INTERVAL;
   }
 
-  babel_select_route(e);
+  babel_select_route(p, e);
 }
 
 void
@@ -1351,7 +1339,7 @@ babel_handle_seqno_request(union babel_msg *m, struct babel_iface *ifa)
   {
     /* Not ours; forward if TTL allows it */
     if (msg->hop_count > 1)
-      babel_forward_seqno_request(e, msg, msg->sender);
+      babel_forward_seqno_request(p, e, msg, msg->sender);
   }
 }
 
@@ -1470,7 +1458,7 @@ babel_iface_stop(struct babel_iface *ifa)
       r->expires = now_ + r->expiry_interval;
 
       if (r == r->e->selected_in)
-	babel_select_route(r->e);
+	babel_select_route(p, r->e);
     }
   }
 
@@ -2081,7 +2069,7 @@ babel_rt_notify(struct proto *P, struct channel *c UNUSED, struct network *net,
 
     if (new->attrs->src->proto != P)
     {
-      r = babel_get_route(e, NULL);
+      r = babel_get_route(p, e, NULL);
       r->seqno = p->update_seqno;
       r->router_id = p->router_id;
       r->metric = 0;	/* FIXME: should be selectable */
diff --git a/proto/babel/babel.h b/proto/babel/babel.h
index 78841254..ccd3479e 100644
--- a/proto/babel/babel.h
+++ b/proto/babel/babel.h
@@ -232,7 +232,6 @@ struct babel_route {
 };
 
 struct babel_entry {
-  struct babel_proto *proto;
   struct babel_route *selected_in;
   struct babel_route *selected_out;
 

From 672fb78e1272bb49cc2521176effe75e3cb22460 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 8 Nov 2017 14:15:11 +0100
Subject: [PATCH 067/198] Babel: Fix handling of IPv4 retractions

Babel TLV parsing code rejected IPv4 retractions without next-hop,
although next-hop is needed just for regular updates.
---
 proto/babel/packets.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/proto/babel/packets.c b/proto/babel/packets.c
index 6e8d2eb5..8e972db2 100644
--- a/proto/babel/packets.c
+++ b/proto/babel/packets.c
@@ -593,8 +593,8 @@ babel_read_update(struct babel_tlv *hdr, union babel_msg *m,
     if (tlv->omitted && !state->def_ip4_prefix_seen)
       return PARSE_ERROR;
 
-    /* Need next hop for v4 routes */
-    if (ipa_zero(state->next_hop_ip4))
+    /* Update must have next hop, unless it is retraction */
+    if (ipa_zero(state->next_hop_ip4) && (msg->metric != BABEL_INFINITY))
       return PARSE_ERROR;
 
     /* Merge saved prefix and received prefix parts */

From dbf1ed263c1c15f79fb200b4dfe3bffea231f4e4 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 8 Nov 2017 14:35:52 +0100
Subject: [PATCH 068/198] Babel: Fix handling of seqno requests

Old behavior has several deficiencies compared to standard behavior
(no triggered updates for replies, no retransmissions, ...).
---
 proto/babel/babel.c | 309 +++++++++++++++++++++++++-------------------
 proto/babel/babel.h |  24 ++--
 2 files changed, 189 insertions(+), 144 deletions(-)

diff --git a/proto/babel/babel.c b/proto/babel/babel.c
index f5aef492..68b32e8f 100644
--- a/proto/babel/babel.c
+++ b/proto/babel/babel.c
@@ -48,19 +48,20 @@
 static inline int ge_mod64k(uint a, uint b)
 { return (u16)(a - b) < 0x8000; }
 
-static void babel_dump_entry(struct babel_entry *e);
-static void babel_dump_route(struct babel_route *r);
+static void babel_expire_requests(struct babel_proto *p, struct babel_entry *e);
 static void babel_select_route(struct babel_proto *p, struct babel_entry *e);
 static void babel_send_route_request(struct babel_proto *p, struct babel_entry *e, struct babel_neighbor *n);
-static void babel_send_wildcard_request(struct babel_iface *ifa);
-static int  babel_cache_seqno_request(struct babel_proto *p, net_addr *n, u64 router_id, u16 seqno);
-static void babel_trigger_iface_update(struct babel_iface *ifa);
-static void babel_trigger_update(struct babel_proto *p);
-static void babel_send_seqno_request(struct babel_proto *p, struct babel_entry *e);
+static void babel_send_seqno_request(struct babel_proto *p, struct babel_entry *e, struct babel_seqno_request *sr);
 static void babel_update_cost(struct babel_neighbor *n);
 static inline void babel_kick_timer(struct babel_proto *p);
 static inline void babel_iface_kick_timer(struct babel_iface *ifa);
 
+static inline void babel_lock_neighbor(struct babel_neighbor *nbr)
+{ if (nbr) nbr->uc++; }
+
+static inline void babel_unlock_neighbor(struct babel_neighbor *nbr)
+{ if (nbr && !--nbr->uc) mb_free(nbr); }
+
 
 /*
  *	Functions to maintain data structures
@@ -72,6 +73,7 @@ babel_init_entry(void *E)
   struct babel_entry *e = E;
 
   e->updated = current_time();
+  init_list(&e->requests);
   init_list(&e->sources);
   init_list(&e->routes);
 }
@@ -260,9 +262,10 @@ loop:
     }
 
     babel_expire_sources(p, e);
+    babel_expire_requests(p, e);
 
     /* Remove empty entries */
-    if (EMPTY_LIST(e->sources) && EMPTY_LIST(e->routes))
+    if (EMPTY_LIST(e->sources) && EMPTY_LIST(e->routes) && EMPTY_LIST(e->requests))
     {
       FIB_ITERATE_PUT(&fit);
       fib_delete(rtable, e);
@@ -279,6 +282,110 @@ babel_expire_routes(struct babel_proto *p)
   babel_expire_routes_(p, &p->ip6_rtable);
 }
 
+static inline int seqno_request_valid(struct babel_seqno_request *sr)
+{ return !sr->nbr || sr->nbr->ifa; }
+
+/*
+ * Add seqno request to the table of pending requests (RFC 6216 3.2.6) and send
+ * it to network. Do nothing if it is already in the table.
+ */
+
+static void
+babel_add_seqno_request(struct babel_proto *p, struct babel_entry *e,
+			u64 router_id, u16 seqno, u8 hop_count,
+			struct babel_neighbor *nbr)
+{
+  struct babel_seqno_request *sr;
+
+  WALK_LIST(sr, e->requests)
+    if (sr->router_id == router_id)
+    {
+      /* Found matching or newer */
+      if (ge_mod64k(sr->seqno, seqno) && seqno_request_valid(sr))
+	return;
+
+      /* Found older */
+      babel_unlock_neighbor(sr->nbr);
+      rem_node(NODE sr);
+      goto found;
+    }
+
+  /* No entries found */
+  sr = sl_alloc(p->seqno_slab);
+
+found:
+  sr->router_id = router_id;
+  sr->seqno = seqno;
+  sr->hop_count = hop_count;
+  sr->count = 0;
+  sr->expires = current_time() + BABEL_SEQNO_REQUEST_EXPIRY;
+  babel_lock_neighbor(sr->nbr = nbr);
+  add_tail(&e->requests, NODE sr);
+
+  babel_send_seqno_request(p, e, sr);
+}
+
+static void
+babel_remove_seqno_request(struct babel_proto *p, struct babel_seqno_request *sr)
+{
+  babel_unlock_neighbor(sr->nbr);
+  rem_node(NODE sr);
+  sl_free(p->seqno_slab, sr);
+}
+
+static int
+babel_satisfy_seqno_request(struct babel_proto *p, struct babel_entry *e,
+			   u64 router_id, u16 seqno)
+{
+  struct babel_seqno_request *sr;
+
+  WALK_LIST(sr, e->requests)
+    if ((sr->router_id == router_id) && ge_mod64k(seqno, sr->seqno))
+    {
+      /* Found the request, remove it */
+      babel_remove_seqno_request(p, sr);
+      return 1;
+    }
+
+  return 0;
+}
+
+static void
+babel_expire_requests(struct babel_proto *p, struct babel_entry *e)
+{
+  struct babel_seqno_request *sr, *srx;
+  btime now_ = current_time();
+
+  WALK_LIST_DELSAFE(sr, srx, e->requests)
+  {
+    /* Remove seqno requests sent to dead neighbors */
+    if (!seqno_request_valid(sr))
+    {
+      babel_remove_seqno_request(p, sr);
+      continue;
+    }
+
+    /* Handle expired requests - resend or remove */
+    if (sr->expires && sr->expires <= now_)
+    {
+      if (sr->count < BABEL_SEQNO_REQUEST_RETRY)
+      {
+	sr->count++;
+	sr->expires += (BABEL_SEQNO_REQUEST_EXPIRY << sr->count);
+	babel_send_seqno_request(p, e, sr);
+      }
+      else
+      {
+	TRACE(D_EVENTS, "Seqno request for %N router-id %lR expired",
+	      e->n.addr, sr->router_id);
+
+	babel_remove_seqno_request(p, sr);
+	continue;
+      }
+    }
+  }
+}
+
 static struct babel_neighbor *
 babel_find_neighbor(struct babel_iface *ifa, ip_addr addr)
 {
@@ -309,6 +416,7 @@ babel_get_neighbor(struct babel_iface *ifa, ip_addr addr)
   nbr->txcost = BABEL_INFINITY;
   nbr->cost = BABEL_INFINITY;
   init_list(&nbr->routes);
+  babel_lock_neighbor(nbr);
   add_tail(&ifa->neigh_list, NODE nbr);
 
   return nbr;
@@ -333,8 +441,9 @@ babel_flush_neighbor(struct babel_proto *p, struct babel_neighbor *nbr)
       babel_select_route(p, e);
   }
 
+  nbr->ifa = NULL;
   rem_node(NODE nbr);
-  mb_free(nbr);
+  babel_unlock_neighbor(nbr);
 }
 
 static void
@@ -588,14 +697,14 @@ babel_select_route(struct babel_proto *p, struct babel_entry *e)
        (as unreachable), then send a seqno request.
 
        babel_build_rte() will set the unreachable flag if the metric is BABEL_INFINITY.*/
-    if (e->selected_in)
+    if (best)
     {
       TRACE(D_EVENTS, "Lost feasible route for prefix %N", e->n.addr);
 
-      e->selected_in->metric = e->selected_in->advert_metric = BABEL_INFINITY;
+      best->metric = best->advert_metric = BABEL_INFINITY;
       e->updated = current_time();
 
-      babel_send_seqno_request(p, e);
+      babel_add_seqno_request(p, e, best->router_id, best->seqno + 1, 0, NULL);
       babel_announce_rte(p, e);
 
       /* Section 3.6 of the RFC forbids an infeasible from being selected. This
@@ -695,7 +804,6 @@ babel_send_hello(struct babel_iface *ifa)
 static void
 babel_send_route_request(struct babel_proto *p, struct babel_entry *e, struct babel_neighbor *n)
 {
-  struct babel_iface *ifa = n->ifa;
   union babel_msg msg = {};
 
   TRACE(D_PACKETS, "Sending route request for %N to %I",
@@ -704,7 +812,7 @@ babel_send_route_request(struct babel_proto *p, struct babel_entry *e, struct ba
   msg.type = BABEL_TLV_ROUTE_REQUEST;
   net_copy(&msg.route_request.net, e->n.addr);
 
-  babel_send_unicast(&msg, ifa, n->addr);
+  babel_send_unicast(&msg, n->ifa, n->addr);
 }
 
 static void
@@ -723,48 +831,32 @@ babel_send_wildcard_request(struct babel_iface *ifa)
 }
 
 static void
-babel_send_seqno_request(struct babel_proto *p, struct babel_entry *e)
-{
-  struct babel_route *r = e->selected_in;
-  struct babel_iface *ifa = NULL;
-  struct babel_source *s = NULL;
-  union babel_msg msg = {};
-
-  s = babel_find_source(e, r->router_id);
-  if (!s || !babel_cache_seqno_request(p, e->n.addr, r->router_id, s->seqno + 1))
-    return;
-
-  TRACE(D_PACKETS, "Sending seqno request for %N router-id %lR seqno %d",
-	e->n.addr, r->router_id, s->seqno + 1);
-
-  msg.type = BABEL_TLV_SEQNO_REQUEST;
-  msg.seqno_request.hop_count = BABEL_INITIAL_HOP_COUNT;
-  msg.seqno_request.seqno = s->seqno + 1;
-  msg.seqno_request.router_id = r->router_id;
-  net_copy(&msg.seqno_request.net, e->n.addr);
-
-  WALK_LIST(ifa, p->interfaces)
-    babel_enqueue(&msg, ifa);
-}
-
-static void
-babel_unicast_seqno_request(struct babel_proto *p, struct babel_entry *e, struct babel_source *s, struct babel_neighbor *nbr)
+babel_send_seqno_request(struct babel_proto *p, struct babel_entry *e, struct babel_seqno_request *sr)
 {
   union babel_msg msg = {};
 
-  if (!s || !babel_cache_seqno_request(p, e->n.addr, s->router_id, s->seqno + 1))
-    return;
-
-  TRACE(D_PACKETS, "Sending seqno request for %N router-id %lR seqno %d",
-	e->n.addr, s->router_id, s->seqno + 1);
-
   msg.type = BABEL_TLV_SEQNO_REQUEST;
-  msg.seqno_request.hop_count = BABEL_INITIAL_HOP_COUNT;
-  msg.seqno_request.seqno = s->seqno + 1;
-  msg.seqno_request.router_id = s->router_id;
+  msg.seqno_request.hop_count = sr->hop_count ?: BABEL_INITIAL_HOP_COUNT;
+  msg.seqno_request.seqno = sr->seqno;
+  msg.seqno_request.router_id = sr->router_id;
   net_copy(&msg.seqno_request.net, e->n.addr);
 
-  babel_send_unicast(&msg, nbr->ifa, nbr->addr);
+  if (sr->nbr)
+  {
+    TRACE(D_PACKETS, "Sending seqno request for %N router-id %lR seqno %d to %I on %s",
+	  e->n.addr, sr->router_id, sr->seqno, sr->nbr->addr, sr->nbr->ifa->ifname);
+
+    babel_send_unicast(&msg, sr->nbr->ifa, sr->nbr->addr);
+  }
+  else
+  {
+    TRACE(D_PACKETS, "Sending broadcast seqno request for %N router-id %lR seqno %d",
+	  e->n.addr, sr->router_id, sr->seqno);
+
+    struct babel_iface *ifa;
+    WALK_LIST(ifa, p->interfaces)
+      babel_enqueue(&msg, ifa);
+  }
 }
 
 /**
@@ -966,81 +1058,6 @@ babel_update_hello_history(struct babel_neighbor *n, u16 seqno, uint interval)
   n->last_hello_int = interval;
 }
 
-static void
-babel_expire_seqno_requests(struct babel_proto *p)
-{
-  btime now_ = current_time();
-
-  struct babel_seqno_request *n, *nx;
-  WALK_LIST_DELSAFE(n, nx, p->seqno_cache)
-  {
-    if ((n->updated + BABEL_SEQNO_REQUEST_EXPIRY) <= now_)
-    {
-      rem_node(NODE n);
-      sl_free(p->seqno_slab, n);
-    }
-  }
-}
-
-/*
- * Checks the seqno request cache for a matching request and returns failure if
- * found. Otherwise, a new entry is stored in the cache.
- */
-static int
-babel_cache_seqno_request(struct babel_proto *p, net_addr *n,
-                          u64 router_id, u16 seqno)
-{
-  struct babel_seqno_request *r;
-
-  WALK_LIST(r, p->seqno_cache)
-  {
-    if (net_equal(&r->net, n) && (r->router_id == router_id) && (r->seqno == seqno))
-      return 0;
-  }
-
-  /* no entries found */
-  r = sl_alloc(p->seqno_slab);
-  net_copy(&r->net, n);
-  r->router_id = router_id;
-  r->seqno = seqno;
-  r->updated = current_time();
-  add_tail(&p->seqno_cache, NODE r);
-
-  return 1;
-}
-
-static void
-babel_forward_seqno_request(struct babel_proto *p, struct babel_entry *e,
-                            struct babel_msg_seqno_request *in,
-                            ip_addr sender)
-{
-  struct babel_route *r;
-
-  TRACE(D_PACKETS, "Forwarding seqno request for %N router-id %lR seqno %d",
-	e->n.addr, in->router_id, in->seqno);
-
-  WALK_LIST(r, e->routes)
-  {
-    if ((r->router_id == in->router_id) &&
-	!OUR_ROUTE(r) &&
-	!ipa_equal(r->neigh->addr, sender))
-    {
-      if (!babel_cache_seqno_request(p, e->n.addr, in->router_id, in->seqno))
-	return;
-
-      union babel_msg msg = {};
-      msg.type = BABEL_TLV_SEQNO_REQUEST;
-      msg.seqno_request.hop_count = in->hop_count-1;
-      msg.seqno_request.seqno = in->seqno;
-      msg.seqno_request.router_id = in->router_id;
-      net_copy(&msg.seqno_request.net, e->n.addr);
-
-      babel_send_unicast(&msg, r->neigh->ifa, r->neigh->addr);
-      return;
-    }
-  }
-}
-
 
 /*
  *	TLV handlers
@@ -1227,7 +1244,7 @@ babel_handle_update(union babel_msg *m, struct babel_iface *ifa)
   /* RFC section 3.8.2.2 - Dealing with unfeasible updates */
   if (!feasible && (metric != BABEL_INFINITY) &&
       (!best || (r == best) || (metric < best->metric)))
-    babel_unicast_seqno_request(p, e, s, nbr);
+    babel_add_seqno_request(p, e, s->router_id, s->seqno + 1, 0, nbr);
 
   if (!r)
   {
@@ -1264,6 +1281,13 @@ babel_handle_update(union babel_msg *m, struct babel_iface *ifa)
     r->expires = current_time() + r->expiry_interval;
     if (r->expiry_interval > BABEL_ROUTE_REFRESH_INTERVAL)
       r->refresh_time = current_time() + r->expiry_interval - BABEL_ROUTE_REFRESH_INTERVAL;
+
+    /* If received update satisfies seqno request, we send triggered updates */
+    if (babel_satisfy_seqno_request(p, e, msg->router_id, msg->seqno))
+    {
+      babel_trigger_update(p);
+      e->updated = current_time();
+    }
   }
 
   babel_select_route(p, e);
@@ -1301,7 +1325,6 @@ babel_handle_route_request(union babel_msg *m, struct babel_iface *ifa)
   }
 }
 
-
 void
 babel_handle_seqno_request(union babel_msg *m, struct babel_iface *ifa)
 {
@@ -1335,11 +1358,31 @@ babel_handle_seqno_request(union babel_msg *m, struct babel_iface *ifa)
     p->update_seqno_inc = 1;
     babel_trigger_update(p);
   }
-  else
+  else if (msg->hop_count > 1)
   {
     /* Not ours; forward if TTL allows it */
-    if (msg->hop_count > 1)
-      babel_forward_seqno_request(p, e, msg, msg->sender);
+
+    /* Find best admissible route */
+    struct babel_route *r, *best1 = NULL, *best2 = NULL;
+    WALK_LIST(r, e->routes)
+      if ((r->router_id == msg->router_id) && r->neigh && !ipa_equal(r->neigh->addr, msg->sender))
+      {
+	/* Find best feasible route */
+	if (babel_is_feasible(babel_find_source(e, r->router_id), r->seqno, r->advert_metric) &&
+	    (!best1 || r->metric < best1->metric))
+	  best1 = r;
+
+	/* Find best not necessary feasible route */
+	if (!best2 || r->metric < best2->metric)
+	  best2 = r;
+      }
+
+    /* If no route is found, do nothing */
+    r = best1 ?: best2;
+    if (!r)
+      return;
+
+    babel_add_seqno_request(p, e, msg->router_id, msg->seqno, msg->hop_count-1, r->neigh);
   }
 }
 
@@ -1396,7 +1439,7 @@ babel_iface_timer(timer *t)
   {
     TRACE(D_EVENTS, "Sending triggered updates on %s", ifa->ifname);
     babel_send_update(ifa, ifa->want_triggered);
-    ifa->next_triggered = now_ + MIN(5 S, update_period / 2);
+    ifa->next_triggered = now_ + MIN(1 S, update_period / 2);
     ifa->want_triggered = 0;
     p->triggered = 0;
   }
@@ -1422,7 +1465,7 @@ babel_iface_start(struct babel_iface *ifa)
 
   ifa->next_hello = current_time() + (random() % ifa->cf->hello_interval);
   ifa->next_regular = current_time() + (random() % ifa->cf->update_interval);
-  ifa->next_triggered = current_time() + MIN(5 S, ifa->cf->update_interval / 2);
+  ifa->next_triggered = current_time() + MIN(1 S, ifa->cf->update_interval / 2);
   ifa->want_triggered = 0;	/* We send an immediate update (below) */
   tm2_start(ifa->timer, 100 MS);
   ifa->up = 1;
@@ -1988,7 +2031,6 @@ babel_timer(timer *t)
   struct babel_proto *p = t->data;
 
   babel_expire_routes(p);
-  babel_expire_seqno_requests(p);
   babel_expire_neighbors(p);
 }
 
@@ -2177,7 +2219,6 @@ babel_start(struct proto *P)
   p->source_slab = sl_new(P->pool, sizeof(struct babel_source));
   p->msg_slab = sl_new(P->pool, sizeof(struct babel_msg_node));
   p->seqno_slab = sl_new(P->pool, sizeof(struct babel_seqno_request));
-  init_list(&p->seqno_cache);
 
   p->log_pkt_tbf = (struct tbf){ .rate = 1, .burst = 5 };
 
diff --git a/proto/babel/babel.h b/proto/babel/babel.h
index ccd3479e..3ea21269 100644
--- a/proto/babel/babel.h
+++ b/proto/babel/babel.h
@@ -42,7 +42,8 @@
 #define BABEL_ROUTE_EXPIRY_FACTOR(X)	((btime)(X)*7/2)	/* 3.5 */
 #define BABEL_ROUTE_REFRESH_INTERVAL	(2 S_)	/* Time before route expiry to send route request */
 #define BABEL_HOLD_TIME			(10 S_)	/* Expiry time for our own routes */
-#define BABEL_SEQNO_REQUEST_EXPIRY	(60 S_)
+#define BABEL_SEQNO_REQUEST_RETRY	4
+#define BABEL_SEQNO_REQUEST_EXPIRY	(2 S_)
 #define BABEL_GARBAGE_INTERVAL		(300 S_)
 #define BABEL_RXCOST_WIRED		96
 #define BABEL_RXCOST_WIRELESS		256
@@ -147,9 +148,7 @@ struct babel_proto {
   slab *route_slab;
   slab *source_slab;
   slab *msg_slab;
-
   slab *seqno_slab;
-  list seqno_cache;			/* Seqno requests in the cache (struct babel_seqno_request) */
 
   struct tbf log_pkt_tbf;		/* TBF for packet messages */
 };
@@ -190,6 +189,7 @@ struct babel_neighbor {
   struct babel_iface *ifa;
 
   ip_addr addr;
+  uint uc;				/* Reference counter for seqno requests */
   u16 rxcost;				/* Sent in last IHU */
   u16 txcost;				/* Received in last IHU */
   u16 cost;				/* Computed neighbor cost */
@@ -231,12 +231,23 @@ struct babel_route {
   btime expiry_interval;
 };
 
+struct babel_seqno_request {
+  node n;
+  u64 router_id;
+  u16 seqno;
+  u8 hop_count;
+  u8 count;
+  btime expires;
+  struct babel_neighbor *nbr;
+};
+
 struct babel_entry {
   struct babel_route *selected_in;
   struct babel_route *selected_out;
 
   btime updated;
 
+  list requests;
   list sources;				/* Source entries for this prefix (struct babel_source). */
   list routes;				/* Routes for this prefix (struct babel_route) */
 
@@ -244,13 +255,6 @@ struct babel_entry {
 };
 
 /* Stores forwarded seqno requests for duplicate suppression. */
-struct babel_seqno_request {
-  node n;
-  net_addr net;
-  u64 router_id;
-  u16 seqno;
-  btime updated;
-};
 
 
 /*

From 3b3b0910ffb1b212b1c9ea420db6c575a3ecb71a Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 28 Nov 2017 15:11:41 +0100
Subject: [PATCH 069/198] Babel: More changes and bugfixes

Several changes and bugfixes in Babel, namely:

- Exported route parameters stored directly in route table entry
- Exported non-babel routes no longer stored in per-entry route list
- Route update, selection and retraction simplified and fixed
- Route feasibility is evalualated per update and stored with route
- Unreachable route handling fixed, based on hold interval
- Added 'show babel routes' command

Overall, it fixes some issues with proper propagation of triggered
updates, making Babel convergence after topology change almost
instant.
---
 nest/route.h          |   1 +
 proto/babel/babel.c   | 584 +++++++++++++++++++++---------------------
 proto/babel/babel.h   |  36 ++-
 proto/babel/config.Y  |   8 +
 proto/babel/packets.c |   2 +
 5 files changed, 327 insertions(+), 304 deletions(-)

diff --git a/nest/route.h b/nest/route.h
index 2aa2bcd8..a838bd38 100644
--- a/nest/route.h
+++ b/nest/route.h
@@ -236,6 +236,7 @@ typedef struct rte {
 #endif
 #ifdef CONFIG_BABEL
     struct {
+      u16 seqno;			/* Babel seqno */
       u16 metric;			/* Babel metric */
       u64 router_id;			/* Babel router id */
     } babel;
diff --git a/proto/babel/babel.c b/proto/babel/babel.c
index 68b32e8f..768bbda8 100644
--- a/proto/babel/babel.c
+++ b/proto/babel/babel.c
@@ -2,6 +2,8 @@
  *	BIRD -- The Babel protocol
  *
  *	Copyright (c) 2015--2016 Toke Hoiland-Jorgensen
+ * 	(c) 2016--2017 Ondrej Zajicek <santiago@crfreenet.org>
+ *	(c) 2016--2017 CZ.NIC z.s.p.o.
  *
  *	Can be freely distributed and used under the terms of the GNU GPL.
  *
@@ -29,17 +31,14 @@
  *
  * The main route selection is done in babel_select_route(). This is called when
  * an entry is updated by receiving updates from the network or when modified by
- * internal timers. It performs feasibility checks on the available routes for
- * the prefix and selects the one with the lowest metric to be announced to the
- * core.
+ * internal timers. The function selects from feasible and reachable routes the
+ * one with the lowest metric to be announced to the core.
  */
 
 #include <stdlib.h>
 #include "babel.h"
 
 
-#define OUR_ROUTE(r) (r->neigh == NULL)
-
 /*
  * Is one number greater or equal than another mod 2^16? This is based on the
  * definition of serial number space in RFC 1982. Note that arguments are of
@@ -49,7 +48,8 @@ static inline int ge_mod64k(uint a, uint b)
 { return (u16)(a - b) < 0x8000; }
 
 static void babel_expire_requests(struct babel_proto *p, struct babel_entry *e);
-static void babel_select_route(struct babel_proto *p, struct babel_entry *e);
+static void babel_select_route(struct babel_proto *p, struct babel_entry *e, struct babel_route *mod);
+static inline void babel_announce_retraction(struct babel_proto *p, struct babel_entry *e);
 static void babel_send_route_request(struct babel_proto *p, struct babel_entry *e, struct babel_neighbor *n);
 static void babel_send_seqno_request(struct babel_proto *p, struct babel_entry *e, struct babel_seqno_request *sr);
 static void babel_update_cost(struct babel_neighbor *n);
@@ -161,35 +161,35 @@ babel_get_route(struct babel_proto *p, struct babel_entry *e, struct babel_neigh
 
   r = sl_alloc(p->route_slab);
   memset(r, 0, sizeof(*r));
-  r->e = e;
-  add_tail(&e->routes, NODE r);
 
-  if (nbr)
-  {
-    r->neigh = nbr;
-    r->expires = current_time() + BABEL_GARBAGE_INTERVAL;
-    add_tail(&nbr->routes, NODE &r->neigh_route);
-  }
+  r->e = e;
+  r->neigh = nbr;
+  add_tail(&e->routes, NODE r);
+  add_tail(&nbr->routes, NODE &r->neigh_route);
 
   return r;
 }
 
+static inline void
+babel_retract_route(struct babel_proto *p, struct babel_route *r)
+{
+  r->metric = r->advert_metric = BABEL_INFINITY;
+
+  if (r == r->e->selected)
+    babel_select_route(p, r->e, r);
+}
+
 static void
 babel_flush_route(struct babel_proto *p, struct babel_route *r)
 {
   DBG("Babel: Flush route %N router_id %lR neigh %I\n",
-      r->e->n.addr, r->router_id, r->neigh ? r->neigh->addr : IPA_NONE);
+      r->e->n.addr, r->router_id, r->neigh->addr);
 
   rem_node(NODE r);
+  rem_node(&r->neigh_route);
 
-  if (r->neigh)
-    rem_node(&r->neigh_route);
-
-  if (r->e->selected_in == r)
-    r->e->selected_in = NULL;
-
-  if (r->e->selected_out == r)
-    r->e->selected_out = NULL;
+  if (r->e->selected == r)
+    r->e->selected = NULL;
 
   sl_free(p->route_slab, r);
 }
@@ -197,15 +197,15 @@ babel_flush_route(struct babel_proto *p, struct babel_route *r)
 static void
 babel_expire_route(struct babel_proto *p, struct babel_route *r)
 {
-  struct babel_entry *e = r->e;
+  struct babel_config *cf = (void *) p->p.cf;
 
   TRACE(D_EVENTS, "Route expiry timer for %N router-id %lR fired",
-	e->n.addr, r->router_id);
+	r->e->n.addr, r->router_id);
 
   if (r->metric < BABEL_INFINITY)
   {
     r->metric = r->advert_metric = BABEL_INFINITY;
-    r->expires = current_time() + r->expiry_interval;
+    r->expires = current_time() + cf->hold_time;
   }
   else
   {
@@ -216,7 +216,7 @@ babel_expire_route(struct babel_proto *p, struct babel_route *r)
 static void
 babel_refresh_route(struct babel_proto *p, struct babel_route *r)
 {
-  if (!OUR_ROUTE(r) && (r == r->e->selected_in))
+  if (r == r->e->selected)
     babel_send_route_request(p, r->e, r->neigh);
 
   r->refresh_time = 0;
@@ -225,6 +225,7 @@ babel_refresh_route(struct babel_proto *p, struct babel_route *r)
 static void
 babel_expire_routes_(struct babel_proto *p, struct fib *rtable)
 {
+  struct babel_config *cf = (void *) p->p.cf;
   struct babel_route *r, *rx;
   struct fib_iterator fit;
   btime now_ = current_time();
@@ -243,8 +244,8 @@ loop:
 
       if (r->expires && r->expires <= now_)
       {
+	changed = changed || (r == e->selected);
 	babel_expire_route(p, r);
-	changed = 1;
       }
     }
 
@@ -255,9 +256,20 @@ loop:
        * synchronous events babel_select_route() -> nest table change ->
        * babel_rt_notify() -> rtable change, invalidating hidden variables.
        */
-
       FIB_ITERATE_PUT(&fit);
-      babel_select_route(p, e);
+      babel_select_route(p, e, NULL);
+      goto loop;
+    }
+
+    /* Clean up stale entries */
+    if ((e->valid == BABEL_ENTRY_STALE) && ((e->updated + cf->hold_time) <= now_))
+      e->valid = BABEL_ENTRY_DUMMY;
+
+    /* Clean up unreachable route */
+    if (e->unreachable && (!e->valid || (e->router_id == p->router_id)))
+    {
+      FIB_ITERATE_PUT(&fit);
+      babel_announce_retraction(p, e);
       goto loop;
     }
 
@@ -265,7 +277,7 @@ loop:
     babel_expire_requests(p, e);
 
     /* Remove empty entries */
-    if (EMPTY_LIST(e->sources) && EMPTY_LIST(e->routes) && EMPTY_LIST(e->requests))
+    if (!e->valid && EMPTY_LIST(e->routes) && EMPTY_LIST(e->sources) && EMPTY_LIST(e->requests))
     {
       FIB_ITERATE_PUT(&fit);
       fib_delete(rtable, e);
@@ -425,20 +437,16 @@ babel_get_neighbor(struct babel_iface *ifa, ip_addr addr)
 static void
 babel_flush_neighbor(struct babel_proto *p, struct babel_neighbor *nbr)
 {
+  struct babel_route *r;
   node *n;
 
   TRACE(D_EVENTS, "Removing neighbor %I on %s", nbr->addr, nbr->ifa->iface->name);
 
   WALK_LIST_FIRST(n, nbr->routes)
   {
-    struct babel_route *r = SKIP_BACK(struct babel_route, neigh_route, n);
-    struct babel_entry *e = r->e;
-    int selected = (r == e->selected_in);
-
+    r = SKIP_BACK(struct babel_route, neigh_route, n);
+    babel_retract_route(p, r);
     babel_flush_route(p, r);
-
-    if (selected)
-      babel_select_route(p, e);
   }
 
   nbr->ifa = NULL;
@@ -600,7 +608,7 @@ done:
     WALK_LIST2(r, n, nbr->routes, neigh_route)
     {
       r->metric = babel_compute_metric(nbr, r->advert_metric);
-      babel_select_route(p, r->e);
+      babel_select_route(p, r->e, r);
     }
   }
 }
@@ -611,119 +619,151 @@ done:
  * @e: Babel route entry to announce
  *
  * This function announces a Babel entry to the core if it has a selected
- * incoming path, and retracts it otherwise. If the selected entry has infinite
- * metric, the route is announced as unreachable.
+ * incoming path, and retracts it otherwise. If there is no selected route but
+ * the entry is valid and ours, the unreachable route is announced instead.
  */
 static void
 babel_announce_rte(struct babel_proto *p, struct babel_entry *e)
 {
-  struct babel_route *r = e->selected_in;
+  struct babel_route *r = e->selected;
   struct channel *c = (e->n.addr->type == NET_IP4) ? p->ip4_channel : p->ip6_channel;
 
   if (r)
   {
-    rta *ap0 = allocz(RTA_MAX_SIZE);
-    *ap0 = (rta) {
+    rta a0 = {
       .src = p->p.main_source,
       .source = RTS_BABEL,
       .scope = SCOPE_UNIVERSE,
-      .dest = r->metric == BABEL_INFINITY ? RTD_UNREACHABLE : RTD_UNICAST,
+      .dest = RTD_UNICAST,
       .from = r->neigh->addr,
+      .nh.gw = r->next_hop,
       .nh.iface = r->neigh->ifa->iface,
     };
 
-    if (r->metric < BABEL_INFINITY)
-      ap0->nh.gw = r->next_hop;
-
-    rta *a = rta_lookup(ap0);
+    rta *a = rta_lookup(&a0);
     rte *rte = rte_get_temp(a);
+    rte->u.babel.seqno = r->seqno;
     rte->u.babel.metric = r->metric;
     rte->u.babel.router_id = r->router_id;
     rte->pflags = 0;
 
-    r->old_metric = r->metric;
+    e->unreachable = 0;
+    rte_update2(c, e->n.addr, rte, p->p.main_source);
+  }
+  else if (e->valid && (e->router_id != p->router_id))
+  {
+    /* Unreachable */
+    rta a0 = {
+      .src = p->p.main_source,
+      .source = RTS_BABEL,
+      .scope = SCOPE_UNIVERSE,
+      .dest = RTD_UNREACHABLE,
+    };
+
+    rta *a = rta_lookup(&a0);
+    rte *rte = rte_get_temp(a);
+    memset(&rte->u.babel, 0, sizeof(rte->u.babel));
+    rte->pflags = 0;
+    rte->pref = 1;
+
+    e->unreachable = 1;
     rte_update2(c, e->n.addr, rte, p->p.main_source);
   }
   else
   {
     /* Retraction */
+    e->unreachable = 0;
     rte_update2(c, e->n.addr, NULL, p->p.main_source);
   }
 }
 
+/* Special case of babel_announce_rte() just for retraction */
+static inline void
+babel_announce_retraction(struct babel_proto *p, struct babel_entry *e)
+{
+  struct channel *c = (e->n.addr->type == NET_IP4) ? p->ip4_channel : p->ip6_channel;
+  e->unreachable = 0;
+  rte_update2(c, e->n.addr, NULL, p->p.main_source);
+}
+
+
 /**
  * babel_select_route - select best route for given route entry
+ * @p: Babel protocol instance
  * @e: Babel entry to select the best route for
+ * @mod: Babel route that was modified or NULL if unspecified
  *
- * Select the best feasible route for a given prefix among the routes received
- * from peers, and propagate it to the nest. This just selects the feasible
- * route with the lowest metric.
+ * Select the best reachable and feasible route for a given prefix among the
+ * routes received from peers, and propagate it to the nest. This just selects
+ * the reachable and feasible route with the lowest metric, but keeps selected
+ * the old one in case of tie.
  *
  * If no feasible route is available for a prefix that previously had a route
- * selected, a seqno request is sent to try to get a valid route. In the
- * meantime, the route is marked as infeasible in the nest (to blackhole packets
- * going to it, as per the RFC).
+ * selected, a seqno request is sent to try to get a valid route. If the entry
+ * is valid and not owned by us, the unreachable route is announced to the nest
+ * (to blackhole packets going to it, as per section 2.8). It is later removed
+ * by babel_expire_routes(). Otherwise, the route is just removed from the nest.
  *
- * If no feasible route is available, and no previous route is selected, the
- * route is removed from the nest entirely.
+ * Argument @mod is used to optimize best route calculation. When specified, the
+ * function can assume that only the @mod route was modified to avoid full best
+ * route selection and announcement when non-best route was modified in minor
+ * way. The caller is advised to not call babel_select_route() when no change is
+ * done (e.g. periodic route updates) to avoid unnecessary announcements of the
+ * same best route. The caller is not required to call the function in case of a
+ * retraction of a non-best route.
+ *
+ * Note that the function does not active triggered updates. That is done by
+ * babel_rt_notify() when the change is propagated back to Babel.
  */
 static void
-babel_select_route(struct babel_proto *p, struct babel_entry *e)
+babel_select_route(struct babel_proto *p, struct babel_entry *e, struct babel_route *mod)
 {
-  struct babel_route *r, *cur = e->selected_in, *best = e->selected_in;
+  struct babel_route *r, *best = e->selected;
 
-  /* try to find the best feasible route */
-  WALK_LIST(r, e->routes)
-    if (!OUR_ROUTE(r) && /* prevent propagating our own routes back to core */
-	(!cur || r->metric < cur->metric) &&
-        babel_is_feasible(babel_find_source(e, r->router_id), r->seqno, r->advert_metric))
-      cur = r;
-
-  if (cur && !OUR_ROUTE(cur) && (cur->metric < BABEL_INFINITY) &&
-      (!best || (cur->metric < best->metric) || ((cur == best) && (cur->metric != cur->old_metric))))
+  /* Shortcut if only non-best was modified */
+  if (mod && (mod != best))
   {
-    if (cur != best)
-      TRACE(D_EVENTS, "Picked new route for prefix %N: router id %lR metric %d",
-	    e->n.addr, cur->router_id, cur->metric);
-
-    e->selected_in = cur;
-    e->updated = current_time();
-    babel_announce_rte(p, e);
-  }
-  else if (!cur || cur->metric == BABEL_INFINITY)
-  {
-    /* Couldn't find a feasible route. If we have a selected route, that means
-       it just became infeasible; so set it's metric to infinite and install it
-       (as unreachable), then send a seqno request.
-
-       babel_build_rte() will set the unreachable flag if the metric is BABEL_INFINITY.*/
-    if (best)
-    {
-      TRACE(D_EVENTS, "Lost feasible route for prefix %N", e->n.addr);
-
-      best->metric = best->advert_metric = BABEL_INFINITY;
-      e->updated = current_time();
-
-      babel_add_seqno_request(p, e, best->router_id, best->seqno + 1, 0, NULL);
-      babel_announce_rte(p, e);
-
-      /* Section 3.6 of the RFC forbids an infeasible from being selected. This
-	 is cleared after announcing the route to the core to make sure an
-	 unreachable route is propagated first. */
-      e->selected_in = NULL;
-    }
+    /* Either select modified route, or keep old best route */
+    if ((mod->metric < (best ? best->metric : BABEL_INFINITY)) && mod->feasible)
+      best = mod;
     else
-    {
-      /* No route currently selected, and no new one selected; this means we
-	 don't have a route to this destination anymore (and were probably
-	 called from an expiry timer). Remove the route from the nest. */
-      // TRACE(D_EVENTS, "Flushing route for prefix %N", e->n.addr);
-
-      e->selected_in = NULL;
-      e->updated = current_time();
-      babel_announce_rte(p, e);
-    }
+      return;
   }
+  else
+  {
+    /* Selected route may be modified and no longer admissible */
+    if (!best || (best->metric == BABEL_INFINITY) || !best->feasible)
+      best = NULL;
+
+    /* Find the best feasible route from all routes */
+    WALK_LIST(r, e->routes)
+      if ((r->metric < (best ? best->metric : BABEL_INFINITY)) && r->feasible)
+	best = r;
+  }
+
+  if (best)
+  {
+    if (best != e->selected)
+      TRACE(D_EVENTS, "Picked new route for prefix %N: router-id %lR metric %d",
+	    e->n.addr, best->router_id, best->metric);
+  }
+  else if (e->selected)
+  {
+    /*
+     * We have lost all feasible routes. We have to broadcast seqno request
+     * (Section 3.8.2.1) and keep unreachable route for a while (section 2.8).
+     * The later is done automatically by babel_announce_rte().
+     */
+
+    TRACE(D_EVENTS, "Lost feasible route for prefix %N", e->n.addr);
+    if (e->valid && (e->selected->router_id == e->router_id))
+      babel_add_seqno_request(p, e, e->selected->router_id, e->selected->seqno + 1, 0, NULL);
+  }
+  else
+    return;
+
+  e->selected = best;
+  babel_announce_rte(p, e);
 }
 
 /*
@@ -883,16 +923,14 @@ babel_send_update_(struct babel_iface *ifa, btime changed, struct fib *rtable)
 
   FIB_WALK(rtable, struct babel_entry, e)
   {
-    struct babel_route *r = e->selected_out;
-
-    if (!r)
+    if (!e->valid)
       continue;
 
     /* Our own seqno might have changed, in which case we update the routes we
        originate. */
-    if ((r->router_id == p->router_id) && (r->seqno < p->update_seqno))
+    if ((e->router_id == p->router_id) && (e->seqno < p->update_seqno))
     {
-      r->seqno = p->update_seqno;
+      e->seqno = p->update_seqno;
       e->updated = current_time();
     }
 
@@ -901,14 +939,14 @@ babel_send_update_(struct babel_iface *ifa, btime changed, struct fib *rtable)
       continue;
 
     TRACE(D_PACKETS, "Sending update for %N router-id %lR seqno %d metric %d",
-	  e->n.addr, r->router_id, r->seqno, r->metric);
+	  e->n.addr, e->router_id, e->seqno, e->metric);
 
     union babel_msg msg = {};
     msg.type = BABEL_TLV_UPDATE;
     msg.update.interval = ifa->cf->update_interval;
-    msg.update.seqno = r->seqno;
-    msg.update.metric = r->metric;
-    msg.update.router_id = r->router_id;
+    msg.update.seqno = e->seqno;
+    msg.update.metric = e->metric;
+    msg.update.router_id = e->router_id;
     net_copy(&msg.update.net, e->n.addr);
 
     msg.update.next_hop = ((e->n.addr->type == NET_IP4) ?
@@ -917,9 +955,9 @@ babel_send_update_(struct babel_iface *ifa, btime changed, struct fib *rtable)
     babel_enqueue(&msg, ifa);
 
     /* Update feasibility distance for redistributed routes */
-    if (!OUR_ROUTE(r))
+    if (e->router_id != p->router_id)
     {
-      struct babel_source *s = babel_get_source(p, e, r->router_id);
+      struct babel_source *s = babel_get_source(p, e, e->router_id);
       s->expires = current_time() + BABEL_GARBAGE_INTERVAL;
 
       if ((msg.update.seqno > s->seqno) ||
@@ -1164,39 +1202,6 @@ babel_handle_update(union babel_msg *m, struct babel_iface *ifa)
     return;
   }
 
-  /*
-   * RFC section 3.5.4:
-   *
-   * When a Babel node receives an update (id, prefix, seqno, metric) from a
-   * neighbour neigh with a link cost value equal to cost, it checks whether it
-   * already has a routing table entry indexed by (neigh, id, prefix).
-   *
-   * If no such entry exists:
-   *
-   * o if the update is unfeasible, it is ignored;
-   *
-   * o if the metric is infinite (the update is a retraction), the update is
-   *   ignored;
-   *
-   * o otherwise, a new route table entry is created, indexed by (neigh, id,
-   *   prefix), with seqno equal to seqno and an advertised metric equal to the
-   *   metric carried by the update.
-   *
-   * If such an entry exists:
-   *
-   * o if the entry is currently installed and the update is unfeasible, then
-   *   the behaviour depends on whether the router-ids of the two entries match.
-   *   If the router-ids are different, the update is treated as though it were
-   *   a retraction (i.e., as though the metric were FFFF hexadecimal). If the
-   *   router-ids are equal, the update is ignored;
-   *
-   * o otherwise (i.e., if either the update is feasible or the entry is not
-   *   currently installed), then the entry's sequence number, advertised
-   *   metric, metric, and router-id are updated and, unless the advertised
-   *   metric is infinite, the route's expiry timer is reset to a small multiple
-   *   of the Interval value included in the update.
-   */
-
   /* Retraction */
   if (msg->metric == BABEL_INFINITY)
   {
@@ -1209,8 +1214,7 @@ babel_handle_update(union babel_msg *m, struct babel_iface *ifa)
       WALK_LIST(n, nbr->routes)
       {
 	r = SKIP_BACK(struct babel_route, neigh_route, n);
-	r->metric = r->advert_metric = BABEL_INFINITY;
-	babel_select_route(p, r->e);
+	babel_retract_route(p, r);
       }
     }
     else
@@ -1226,71 +1230,56 @@ babel_handle_update(union babel_msg *m, struct babel_iface *ifa)
       if (!r)
 	return;
 
-      r->metric = r->advert_metric = BABEL_INFINITY;
-      babel_select_route(p, e);
+      /* Router-id, next-hop and seqno are ignored for retractions */
+      babel_retract_route(p, r);
     }
 
     /* Done with retractions */
     return;
   }
 
+  /* Regular update */
   e = babel_get_entry(p, &msg->net);
-  r = babel_find_route(e, nbr); /* the route entry indexed by neighbour */
+  r = babel_get_route(p, e, nbr); /* the route entry indexed by neighbour */
   s = babel_find_source(e, msg->router_id); /* for feasibility */
   feasible = babel_is_feasible(s, msg->seqno, msg->metric);
   metric = babel_compute_metric(nbr, msg->metric);
-  best = e->selected_in;
+  best = e->selected;
 
   /* RFC section 3.8.2.2 - Dealing with unfeasible updates */
   if (!feasible && (metric != BABEL_INFINITY) &&
       (!best || (r == best) || (metric < best->metric)))
     babel_add_seqno_request(p, e, s->router_id, s->seqno + 1, 0, nbr);
 
-  if (!r)
+  /* Special case - ignore unfeasible update to best route */
+  if (r == best && !feasible && (msg->router_id == r->router_id))
+    return;
+
+  r->expires = current_time() + BABEL_ROUTE_EXPIRY_FACTOR(msg->interval);
+  r->refresh_time = current_time() + BABEL_ROUTE_REFRESH_FACTOR(msg->interval);
+
+  /* No further processing if there is no change */
+  if ((r->feasible == feasible) && (r->seqno == msg->seqno) &&
+      (r->metric == metric) && (r->advert_metric == msg->metric) &&
+      (r->router_id == msg->router_id) && ipa_equal(r->next_hop, msg->next_hop))
+    return;
+
+  /* Last paragraph above - update the entry */
+  r->feasible = feasible;
+  r->seqno = msg->seqno;
+  r->metric = metric;
+  r->advert_metric = msg->metric;
+  r->router_id = msg->router_id;
+  r->next_hop = msg->next_hop;
+
+  /* If received update satisfies seqno request, we send triggered updates */
+  if (babel_satisfy_seqno_request(p, e, msg->router_id, msg->seqno))
   {
-    if (!feasible)
-      return;
-
-    r = babel_get_route(p, e, nbr);
-    r->advert_metric = msg->metric;
-    r->router_id = msg->router_id;
-    r->metric = metric;
-    r->next_hop = msg->next_hop;
-    r->seqno = msg->seqno;
-  }
-  else if (r == best && !feasible)
-  {
-    /* Penultimate paragraph above - ignore or retract */
-    if (msg->router_id == r->router_id)
-      return;
-
-    /* Treat as retraction */
-    r->metric = r->advert_metric = BABEL_INFINITY;
-  }
-  else
-  {
-    /* Last paragraph above - update the entry */
-    r->advert_metric = msg->metric;
-    r->metric = metric;
-    r->next_hop = msg->next_hop;
-
-    r->router_id = msg->router_id;
-    r->seqno = msg->seqno;
-
-    r->expiry_interval = BABEL_ROUTE_EXPIRY_FACTOR(msg->interval);
-    r->expires = current_time() + r->expiry_interval;
-    if (r->expiry_interval > BABEL_ROUTE_REFRESH_INTERVAL)
-      r->refresh_time = current_time() + r->expiry_interval - BABEL_ROUTE_REFRESH_INTERVAL;
-
-    /* If received update satisfies seqno request, we send triggered updates */
-    if (babel_satisfy_seqno_request(p, e, msg->router_id, msg->seqno))
-    {
-      babel_trigger_update(p);
-      e->updated = current_time();
-    }
+    babel_trigger_update(p);
+    e->updated = current_time();
   }
 
-  babel_select_route(p, e);
+  babel_select_route(p, e, r);
 }
 
 void
@@ -1338,13 +1327,12 @@ babel_handle_seqno_request(union babel_msg *m, struct babel_iface *ifa)
 
   /* Ignore if we have no such entry or entry has infinite metric */
   struct babel_entry *e = babel_find_entry(p, &msg->net);
-  if (!e || !e->selected_out || (e->selected_out->metric == BABEL_INFINITY))
+  if (!e || !e->valid || (e->metric == BABEL_INFINITY))
     return;
 
   /* Trigger update on incoming interface if we have a selected route with
      different router id or seqno no smaller than requested */
-  struct babel_route *r = e->selected_out;
-  if ((r->router_id != msg->router_id) || ge_mod64k(r->seqno, msg->seqno))
+  if ((e->router_id != msg->router_id) || ge_mod64k(e->seqno, msg->seqno))
   {
     babel_trigger_iface_update(ifa);
     e->updated = current_time();
@@ -1365,11 +1353,10 @@ babel_handle_seqno_request(union babel_msg *m, struct babel_iface *ifa)
     /* Find best admissible route */
     struct babel_route *r, *best1 = NULL, *best2 = NULL;
     WALK_LIST(r, e->routes)
-      if ((r->router_id == msg->router_id) && r->neigh && !ipa_equal(r->neigh->addr, msg->sender))
+      if ((r->router_id == msg->router_id) && !ipa_equal(r->neigh->addr, msg->sender))
       {
 	/* Find best feasible route */
-	if (babel_is_feasible(babel_find_source(e, r->router_id), r->seqno, r->advert_metric) &&
-	    (!best1 || r->metric < best1->metric))
+	if ((!best1 || r->metric < best1->metric) && r->feasible)
 	  best1 = r;
 
 	/* Find best not necessary feasible route */
@@ -1483,7 +1470,6 @@ babel_iface_stop(struct babel_iface *ifa)
   struct babel_neighbor *nbr;
   struct babel_route *r;
   node *n;
-  btime now_ = current_time();
 
   TRACE(D_EVENTS, "Stopping interface %s", ifa->ifname);
 
@@ -1497,11 +1483,7 @@ babel_iface_stop(struct babel_iface *ifa)
     WALK_LIST(n, nbr->routes)
     {
       r = SKIP_BACK(struct babel_route, neigh_route, n);
-      r->metric = r->advert_metric = BABEL_INFINITY;
-      r->expires = now_ + r->expiry_interval;
-
-      if (r == r->e->selected_in)
-	babel_select_route(p, r->e);
+      babel_retract_route(p, r);
     }
   }
 
@@ -1777,10 +1759,8 @@ static void
 babel_dump_route(struct babel_route *r)
 {
   debug("Route neigh %I if %s seqno %d metric %d/%d router_id %lR expires %t\n",
-	r->neigh ? r->neigh->addr : IPA_NONE,
-        r->neigh ? r->neigh->ifa->ifname : "(none)",
-        r->seqno, r->advert_metric, r->metric, r->router_id,
-	r->expires ? r->expires - current_time() : 0);
+	r->neigh->addr, r->neigh->ifa->ifname, r->seqno, r->advert_metric, r->metric,
+	r->router_id, r->expires ? r->expires - current_time() : 0);
 }
 
 static void
@@ -1797,8 +1777,7 @@ babel_dump_entry(struct babel_entry *e)
   WALK_LIST(r,e->routes)
   {
     debug(" ");
-    if (r == e->selected_out) debug("*");
-    if (r == e->selected_in) debug("+");
+    if (r == e->selected) debug("*");
     babel_dump_route(r);
   }
 }
@@ -1956,36 +1935,29 @@ babel_show_neighbors(struct proto *P, char *iff)
 }
 
 static void
-babel_show_entries_(struct babel_proto *p, struct fib *rtable)
+babel_show_entries_(struct babel_proto *p UNUSED, struct fib *rtable)
 {
-  struct babel_source *s = NULL;
-  struct babel_route *r = NULL;
-
-  char ridbuf[ROUTER_ID_64_LENGTH+1];
-
   FIB_WALK(rtable, struct babel_entry, e)
   {
-    r = e->selected_in ? e->selected_in : e->selected_out;
+    struct babel_route *r = NULL;
+    uint rts = 0, srcs = 0;
+    node *n;
 
-    int srcs = 0;
-    WALK_LIST(s, e->sources)
+    WALK_LIST(n, e->routes)
+      rts++;
+
+    WALK_LIST(n, e->sources)
       srcs++;
 
-    if (r)
-    {
-      if (r->router_id == p->router_id)
-        bsprintf(ridbuf, "%s", "<self>");
-      else
-        bsprintf(ridbuf, "%lR", r->router_id);
-
-      btime time = r->expires ? r->expires - current_time() : 0;
-      cli_msg(-1025, "%-29N %-23s %6u %5u %7t %7u",
-	      e->n.addr, ridbuf, r->metric, r->seqno, MAX(time, 0), srcs);
-    }
+    if (e->valid)
+      cli_msg(-1025, "%-24N %-23lR %6u %5u %7u %7u",
+	      e->n.addr, e->router_id, e->metric, e->seqno, rts, srcs);
+    else if (r = e->selected)
+      cli_msg(-1025, "%-24N %-23lR %6u %5u %7u %7u",
+	      e->n.addr, r->router_id, r->metric, r->seqno, rts, srcs);
     else
-    {
-      cli_msg(-1025, "%-29N %-44s %7u", e->n.addr, "<pending>", srcs);
-    }
+      cli_msg(-1025, "%-24N %-23s %6s %5s %7u %7u",
+	      e->n.addr, "<none>", "-", "-", rts, srcs);
   }
   FIB_WALK_END;
 }
@@ -2003,8 +1975,8 @@ babel_show_entries(struct proto *P)
   }
 
   cli_msg(-1025, "%s:", p->p.name);
-  cli_msg(-1025, "%-29s %-23s %6s %5s %7s %7s",
-	  "Prefix", "Router ID", "Metric", "Seqno", "Expires", "Sources");
+  cli_msg(-1025, "%-24s %-23s %6s %5s %7s %7s",
+	  "Prefix", "Router ID", "Metric", "Seqno", "Routes", "Sources");
 
   babel_show_entries_(p, &p->ip4_rtable);
   babel_show_entries_(p, &p->ip6_rtable);
@@ -2012,6 +1984,46 @@ babel_show_entries(struct proto *P)
   cli_msg(0, "");
 }
 
+static void
+babel_show_routes_(struct babel_proto *p UNUSED, struct fib *rtable)
+{
+  FIB_WALK(rtable, struct babel_entry, e)
+  {
+    struct babel_route *r;
+    WALK_LIST(r, e->routes)
+    {
+      char c = (r == e->selected) ? '*' : (r->feasible ? '+' : ' ');
+      btime time = r->expires ? r->expires - current_time() : 0;
+      cli_msg(-1025, "%-24N %-25I %-10s %5u %c %5u %7t",
+	      e->n.addr, r->next_hop, r->neigh->ifa->ifname,
+	      r->metric, c, r->seqno, MAX(time, 0));
+    }
+  }
+  FIB_WALK_END;
+}
+
+void
+babel_show_routes(struct proto *P)
+{
+  struct babel_proto *p = (void *) P;
+
+  if (p->p.proto_state != PS_UP)
+  {
+    cli_msg(-1025, "%s: is not up", p->p.name);
+    cli_msg(0, "");
+    return;
+  }
+
+  cli_msg(-1025, "%s:", p->p.name);
+  cli_msg(-1025, "%-24s %-25s %-9s %6s F %5s %7s",
+	  "Prefix", "Nexthop", "Interface", "Metric", "Seqno", "Expires");
+
+  babel_show_routes_(p, &p->ip4_rtable);
+  babel_show_routes_(p, &p->ip6_rtable);
+
+  cli_msg(0, "");
+}
+
 
 /*
  *	Babel protocol glue
@@ -2069,12 +2081,18 @@ babel_prepare_attrs(struct linpool *pool, ea_list *next, uint metric, u64 router
 
 
 static int
-babel_import_control(struct proto *P, struct rte **rt, struct ea_list **attrs, struct linpool *pool)
+babel_import_control(struct proto *P, struct rte **new, struct ea_list **attrs, struct linpool *pool)
 {
   struct babel_proto *p = (void *) P;
+  rte *rt = *new;
+
+  /* Reject our own unreachable routes */
+  if ((rt->attrs->dest == RTD_UNREACHABLE) && (rt->attrs->src->proto == P))
+    return -1;
+
 
   /* Prepare attributes with initial values */
-  if ((*rt)->attrs->source != RTS_BABEL)
+  if (rt->attrs->source != RTS_BABEL)
     *attrs = babel_prepare_attrs(pool, NULL, 0, p->router_id);
 
   return 0;
@@ -2102,65 +2120,50 @@ babel_rt_notify(struct proto *P, struct channel *c UNUSED, struct network *net,
 {
   struct babel_proto *p = (void *) P;
   struct babel_entry *e;
-  struct babel_route *r;
 
   if (new)
   {
     /* Update */
+    uint internal = (new->attrs->src->proto == P);
+    uint rt_seqno = internal ? new->u.babel.seqno : p->update_seqno;
+    uint rt_metric = ea_get_int(attrs, EA_BABEL_METRIC, 0);
+    uint rt_router_id = internal ? new->u.babel.router_id : p->router_id;
+
+    if (rt_metric > BABEL_INFINITY)
+    {
+      log(L_WARN "%s: Invalid babel_metric value %u for route %N",
+	  p->p.name, rt_metric, net->n.addr);
+      rt_metric = BABEL_INFINITY;
+    }
+
     e = babel_get_entry(p, net->n.addr);
 
-    if (new->attrs->src->proto != P)
-    {
-      r = babel_get_route(p, e, NULL);
-      r->seqno = p->update_seqno;
-      r->router_id = p->router_id;
-      r->metric = 0;	/* FIXME: should be selectable */
-    }
-    else
-      r = e->selected_in;
-
-    if (r != e->selected_out)
+    /* Activate triggered updates */
+    if ((e->valid |= BABEL_ENTRY_VALID) ||
+	(e->router_id != rt_router_id))
     {
       babel_trigger_update(p);
       e->updated = current_time();
-      e->selected_out = r;
     }
+
+    e->valid = BABEL_ENTRY_VALID;
+    e->seqno = rt_seqno;
+    e->metric = rt_metric;
+    e->router_id = rt_router_id;
   }
   else
   {
     /* Withdraw */
     e = babel_find_entry(p, net->n.addr);
-    if (!e || !e->selected_out)
+
+    if (!e || e->valid != BABEL_ENTRY_VALID)
       return;
 
-    if (OUR_ROUTE(e->selected_out))
-    {
-      /*
-       * We originate this route, so set its metric to infinity and set an
-       * expiry time. This causes a retraction to be sent, and later the route
-       * to be flushed once the hold time has passed.
-       */
-      babel_trigger_update(p);
-      e->updated = current_time();
-      e->selected_out->metric = BABEL_INFINITY;
-      e->selected_out->expires = current_time() + BABEL_HOLD_TIME;
-    }
-    else
-    {
-      /*
-       * This is a route originating from someone else that was lost; presumably
-       * because an export filter was updated to filter it. This means we can't
-       * set the metric to infinity (it would be overridden on subsequent
-       * updates from the peer originating the route), so just clear the
-       * exported route.
-       *
-       * This causes peers to expire the route after a while (like if we just
-       * shut down), but it's the best we can do in these circumstances; and
-       * since export filters presumably aren't updated that often this is
-       * acceptable.
-       */
-      e->selected_out = NULL;
-    }
+    e->valid = BABEL_ENTRY_STALE;
+    e->metric = BABEL_INFINITY;
+
+    babel_trigger_update(p);
+    e->updated = current_time();
   }
 }
 
@@ -2173,8 +2176,9 @@ babel_rte_better(struct rte *new, struct rte *old)
 static int
 babel_rte_same(struct rte *new, struct rte *old)
 {
-  return ((new->u.babel.router_id == old->u.babel.router_id) &&
-          (new->u.babel.metric == old->u.babel.metric));
+  return ((new->u.babel.seqno == old->u.babel.seqno) &&
+	  (new->u.babel.metric == old->u.babel.metric) &&
+	  (new->u.babel.router_id == old->u.babel.router_id));
 }
 
 
diff --git a/proto/babel/babel.h b/proto/babel/babel.h
index 3ea21269..da3293ef 100644
--- a/proto/babel/babel.h
+++ b/proto/babel/babel.h
@@ -2,6 +2,8 @@
  *	BIRD -- The Babel protocol
  *
  *	Copyright (c) 2015--2016 Toke Hoiland-Jorgensen
+ * 	(c) 2016--2017 Ondrej Zajicek <santiago@crfreenet.org>
+ *	(c) 2016--2017 CZ.NIC z.s.p.o.
  *
  *	Can be freely distributed and used under the terms of the GNU GPL.
  *
@@ -37,11 +39,11 @@
 #define BABEL_HELLO_LIMIT		12
 #define BABEL_UPDATE_INTERVAL_FACTOR	4
 #define BABEL_IHU_INTERVAL_FACTOR	3
+#define BABEL_HOLD_TIME_FACTOR		4	/* How long we keep unreachable route relative to update interval */
 #define BABEL_IHU_EXPIRY_FACTOR(X)	((btime)(X)*7/2)	/* 3.5 */
 #define BABEL_HELLO_EXPIRY_FACTOR(X)	((btime)(X)*3/2)	/* 1.5 */
 #define BABEL_ROUTE_EXPIRY_FACTOR(X)	((btime)(X)*7/2)	/* 3.5 */
-#define BABEL_ROUTE_REFRESH_INTERVAL	(2 S_)	/* Time before route expiry to send route request */
-#define BABEL_HOLD_TIME			(10 S_)	/* Expiry time for our own routes */
+#define BABEL_ROUTE_REFRESH_FACTOR(X)	((btime)(X)*5/2)	/* 2.5 */
 #define BABEL_SEQNO_REQUEST_RETRY	4
 #define BABEL_SEQNO_REQUEST_EXPIRY	(2 S_)
 #define BABEL_GARBAGE_INTERVAL		(300 S_)
@@ -105,8 +107,8 @@ enum babel_ae_type {
 
 struct babel_config {
   struct proto_config c;
-
-  list iface_list;              /* Patterns configured -- keep it first; see babel_reconfigure why */
+  list iface_list;			/* List of iface configs (struct babel_iface_config) */
+  uint hold_time;			/* Time to hold stale entries and unreachable routes */
 };
 
 struct babel_iface_config {
@@ -220,15 +222,14 @@ struct babel_route {
   struct babel_entry    *e;
   struct babel_neighbor *neigh;
 
+  u8 feasible;
   u16 seqno;
-  u16 advert_metric;
   u16 metric;
-  u16 old_metric;
+  u16 advert_metric;
   u64 router_id;
   ip_addr next_hop;
   btime refresh_time;
   btime expires;
-  btime expiry_interval;
 };
 
 struct babel_seqno_request {
@@ -242,19 +243,25 @@ struct babel_seqno_request {
 };
 
 struct babel_entry {
-  struct babel_route *selected_in;
-  struct babel_route *selected_out;
+  struct babel_route *selected;
 
-  btime updated;
-
-  list requests;
-  list sources;				/* Source entries for this prefix (struct babel_source). */
   list routes;				/* Routes for this prefix (struct babel_route) */
+  list sources;				/* Source entries for this prefix (struct babel_source). */
+  list requests;
+
+  u8 valid;				/* Entry validity state (BABEL_ENTRY_*) */
+  u8 unreachable;			/* Unreachable route is announced */
+  u16 seqno;				/* Outgoing seqno */
+  u16 metric;				/* Outgoing metric */
+  u64 router_id;			/* Outgoing router ID */
+  btime updated;			/* Last change of outgoing rte, for triggered updates */
 
   struct fib_node n;
 };
 
-/* Stores forwarded seqno requests for duplicate suppression. */
+#define BABEL_ENTRY_DUMMY	0	/* No outgoing route */
+#define BABEL_ENTRY_VALID	1	/* Valid outgoing route */
+#define BABEL_ENTRY_STALE	2	/* Stale outgoing route, waiting for GC */
 
 
 /*
@@ -346,6 +353,7 @@ void babel_handle_seqno_request(union babel_msg *msg, struct babel_iface *ifa);
 void babel_show_interfaces(struct proto *P, char *iff);
 void babel_show_neighbors(struct proto *P, char *iff);
 void babel_show_entries(struct proto *P);
+void babel_show_routes(struct proto *P);
 
 /* packets.c */
 void babel_enqueue(union babel_msg *msg, struct babel_iface *ifa);
diff --git a/proto/babel/config.Y b/proto/babel/config.Y
index c72c00a3..25ce5ba0 100644
--- a/proto/babel/config.Y
+++ b/proto/babel/config.Y
@@ -2,6 +2,8 @@
  *	BIRD -- Babel Configuration
  *
  *	Copyright (c) 2015-2016 Toke Hoiland-Jorgensen
+ * 	(c) 2016--2017 Ondrej Zajicek <santiago@crfreenet.org>
+ *	(c) 2016--2017 CZ.NIC z.s.p.o.
  *
  *	Can be freely distributed and used under the terms of the GNU GPL.
  */
@@ -32,6 +34,7 @@ babel_proto_start: proto_start BABEL
 {
   this_proto = proto_config_new(&proto_babel, $1);
   init_list(&BABEL_CFG->iface_list);
+  BABEL_CFG->hold_time = 1 S_;
 };
 
 babel_proto_item:
@@ -84,6 +87,8 @@ babel_iface_finish:
   if (!BABEL_IFACE->update_interval)
     BABEL_IFACE->update_interval = MIN_(BABEL_IFACE->hello_interval*BABEL_UPDATE_INTERVAL_FACTOR, BABEL_MAX_INTERVAL);
   BABEL_IFACE->ihu_interval = MIN_(BABEL_IFACE->hello_interval*BABEL_IHU_INTERVAL_FACTOR, BABEL_MAX_INTERVAL);
+
+  BABEL_CFG->hold_time = MAX_(BABEL_CFG->hold_time, BABEL_IFACE->update_interval*BABEL_HOLD_TIME_FACTOR);
 };
 
 
@@ -131,6 +136,9 @@ CF_CLI(SHOW BABEL NEIGHBORS, optsym opttext, [<name>] [\"<interface>\"], [[Show
 CF_CLI(SHOW BABEL ENTRIES, optsym opttext, [<name>], [[Show information about Babel prefix entries]])
 { babel_show_entries(proto_get_named($4, &proto_babel)); };
 
+CF_CLI(SHOW BABEL ROUTES, optsym opttext, [<name>], [[Show information about Babel route entries]])
+{ babel_show_routes(proto_get_named($4, &proto_babel)); };
+
 CF_CODE
 
 CF_END
diff --git a/proto/babel/packets.c b/proto/babel/packets.c
index 8e972db2..4abcf7e4 100644
--- a/proto/babel/packets.c
+++ b/proto/babel/packets.c
@@ -2,6 +2,8 @@
  *	BIRD -- The Babel protocol
  *
  *	Copyright (c) 2015--2016 Toke Hoiland-Jorgensen
+ * 	(c) 2016--2017 Ondrej Zajicek <santiago@crfreenet.org>
+ *	(c) 2016--2017 CZ.NIC z.s.p.o.
  *
  *	Can be freely distributed and used under the terms of the GNU GPL.
  *

From 574b2324275d3292e98a8e329f791eb5c799f7f2 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 28 Nov 2017 17:06:10 +0100
Subject: [PATCH 070/198] Timers: Fix TBF and some last remains

---
 conf/conf.c            |  6 +++---
 lib/birdlib.h          | 30 ++++++++----------------------
 lib/tbf.c              | 31 +++++++++++++++++++------------
 lib/timer.h            |  4 ++--
 nest/proto.c           |  2 +-
 proto/ospf/ospf.h      |  2 +-
 proto/rip/packets.c    |  5 ++++-
 sysdep/linux/netlink.c |  2 +-
 sysdep/unix/io.c       |  2 +-
 sysdep/unix/log.c      |  7 +++----
 sysdep/unix/timer.h    |  4 ++--
 11 files changed, 45 insertions(+), 50 deletions(-)

diff --git a/conf/conf.c b/conf/conf.c
index 2b5e9c71..c68ded7f 100644
--- a/conf/conf.c
+++ b/conf/conf.c
@@ -384,7 +384,7 @@ config_confirm(void)
   if (config_timer->expires == 0)
     return CONF_NOTHING;
 
-  tm_stop(config_timer);
+  tm2_stop(config_timer);
 
   return CONF_CONFIRM;
 }
@@ -420,7 +420,7 @@ config_undo(void)
     return CONF_NOTHING;
 
   undo_available = 0;
-  tm_stop(config_timer);
+  tm2_stop(config_timer);
 
   if (configuring)
     {
@@ -468,7 +468,7 @@ config_init(void)
   config_event = ev_new(&root_pool);
   config_event->hook = config_done;
 
-  config_timer = tm_new(&root_pool);
+  config_timer = tm2_new(&root_pool);
   config_timer->hook = config_timeout;
 }
 
diff --git a/lib/birdlib.h b/lib/birdlib.h
index 9a77fc7b..f41cceb5 100644
--- a/lib/birdlib.h
+++ b/lib/birdlib.h
@@ -69,7 +69,7 @@ static inline int u64_cmp(u64 i1, u64 i2)
 /* Microsecond time */
 
 typedef s64 btime;
-typedef s64 bird_clock_t;
+//typedef s64 bird_clock_t;
 
 #define S_	* (btime) 1000000
 #define MS_	* (btime) 1000
@@ -85,37 +85,23 @@ typedef s64 bird_clock_t;
 #define NS	/1000
 #endif
 
+#define TIME_INFINITY ((s64) 0x7fffffffffffffff)
+
 
 /* Rate limiting */
 
 struct tbf {
-  bird_clock_t timestamp;		/* Last update */
-  u16 count;				/* Available tokens */
+  btime timestamp;			/* Last update */
+  u64 count;				/* Available micro-tokens */
   u16 burst;				/* Max number of tokens */
-  u16 rate;				/* Rate of replenishment */
-  u16 mark;				/* Whether last op was limited */
+  u16 rate;				/* Rate of replenishment (tokens / sec) */
+  u32 drop;				/* Number of failed request since last successful */
 };
 
 /* 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;
-}
+int tbf_limit(struct tbf *f);
 
 
 /* Logging and dying */
diff --git a/lib/tbf.c b/lib/tbf.c
index c1dafee8..e6e84b4f 100644
--- a/lib/tbf.c
+++ b/lib/tbf.c
@@ -10,21 +10,28 @@
 #include "nest/bird.h"
 #include "lib/timer.h"
 
-void
-tbf_update(struct tbf *f)
+int
+tbf_limit(struct tbf *f)
 {
-  bird_clock_t delta = now - f->timestamp;
+  btime delta = current_time() - f->timestamp;
 
-  if (delta == 0)
-    return;
-
-  f->timestamp = now;
-
-  if ((0 < delta) && (delta < f->burst))
+  if (delta > 0)
   {
-    u32 next = f->count + delta * f->rate;
-    f->count = MIN(next, f->burst);
+    u64 next = f->count + delta * f->rate;
+    u64 burst = (u64) f->burst << 20;
+    f->count = MIN(next, burst);
+    f->timestamp += delta;
+  }
+
+  if (f->count < 1000000)
+  {
+    f->drop++;
+    return 1;
   }
   else
-    f->count = f->burst;
+  {
+    f->count -= 1000000;
+    f->drop = 0;
+    return 0;
+  }
 }
diff --git a/lib/timer.h b/lib/timer.h
index 61a2aa94..250bb3cd 100644
--- a/lib/timer.h
+++ b/lib/timer.h
@@ -46,8 +46,8 @@ extern struct timeloop main_timeloop;
 btime current_time(void);
 btime current_real_time(void);
 
-#define now (current_time() TO_S)
-#define now_real (current_real_time() TO_S)
+//#define now (current_time() TO_S)
+//#define now_real (current_real_time() TO_S)
 extern btime boot_time;
 
 timer2 *tm2_new(pool *p);
diff --git a/nest/proto.c b/nest/proto.c
index 27c8fded..72f1f94d 100644
--- a/nest/proto.c
+++ b/nest/proto.c
@@ -1288,7 +1288,7 @@ protos_build(void)
 #endif
 
   proto_pool = rp_new(&root_pool, "Protocols");
-  proto_shutdown_timer = tm_new(proto_pool);
+  proto_shutdown_timer = tm2_new(proto_pool);
   proto_shutdown_timer->hook = proto_shutdown_loop;
 }
 
diff --git a/proto/ospf/ospf.h b/proto/ospf/ospf.h
index dbc231b6..d4571bf6 100644
--- a/proto/ospf/ospf.h
+++ b/proto/ospf/ospf.h
@@ -58,7 +58,7 @@
   log_rl(&p->log_lsa_tbf, L_REMOTE "%s: " msg, p->p.name, args)
 
 #define LOG_LSA2(msg, args...) \
-  do { if (! p->log_lsa_tbf.mark) \
+  do { if (! p->log_lsa_tbf.drop) \
     log(L_REMOTE "%s: " msg, p->p.name, args); } while(0)
 
 
diff --git a/proto/rip/packets.c b/proto/rip/packets.c
index 1518dd3f..4925ca36 100644
--- a/proto/rip/packets.c
+++ b/proto/rip/packets.c
@@ -189,7 +189,10 @@ rip_update_csn(struct rip_proto *p UNUSED, struct rip_iface *ifa)
    * have the same CSN. We are using real time, but enforcing monotonicity.
    */
   if (ifa->cf->auth_type == RIP_AUTH_CRYPTO)
-    ifa->csn = (ifa->csn < (u32) now_real) ? (u32) now_real : ifa->csn + 1;
+  {
+    u32 now_real = (u32) (current_real_time() TO_S);
+    ifa->csn = (ifa->csn < now_real) ? now_real : ifa->csn + 1;
+  }
 }
 
 static void
diff --git a/sysdep/linux/netlink.c b/sysdep/linux/netlink.c
index c9d5cdec..6477d18c 100644
--- a/sysdep/linux/netlink.c
+++ b/sysdep/linux/netlink.c
@@ -151,7 +151,7 @@ nl_open_sock(struct nl_sock *nl)
       nl->fd = socket(PF_NETLINK, SOCK_RAW, NETLINK_ROUTE);
       if (nl->fd < 0)
 	die("Unable to open rtnetlink socket: %m");
-      nl->seq = now;
+      nl->seq = (u32) (current_time() TO_S); /* Or perhaps random_u32() ? */
       nl->rx_buffer = xmalloc(NL_RX_SIZE);
       nl->last_hdr = NULL;
       nl->last_size = 0;
diff --git a/sysdep/unix/io.c b/sysdep/unix/io.c
index 3e56a32d..a196bbe2 100644
--- a/sysdep/unix/io.c
+++ b/sysdep/unix/io.c
@@ -2132,7 +2132,7 @@ io_init(void)
   // XXX init_times();
   // XXX update_times();
   boot_time = current_time();
-  srandom((int) now_real);
+  srandom((uint) (current_real_time() TO_S));
 }
 
 static int short_loops = 0;
diff --git a/sysdep/unix/log.c b/sysdep/unix/log.c
index e564f8f2..f9dccc39 100644
--- a/sysdep/unix/log.c
+++ b/sysdep/unix/log.c
@@ -180,19 +180,18 @@ log_msg(const char *msg, ...)
 void
 log_rl(struct tbf *f, const char *msg, ...)
 {
-  int last_hit = f->mark;
   int class = 1;
   va_list args;
 
   /* Rate limiting is a bit tricky here as it also logs '...' during the first hit */
-  if (tbf_limit(f) && last_hit)
+  if (tbf_limit(f) && (f->drop > 1))
     return;
 
   if (*msg >= 1 && *msg <= 8)
     class = *msg++;
 
   va_start(args, msg);
-  vlog(class, (f->mark ? "..." : msg), args);
+  vlog(class, (f->drop ? "..." : msg), args);
   va_end(args);
 }
 
@@ -332,7 +331,7 @@ void
 mrt_dump_message(struct proto *p, u16 type, u16 subtype, byte *buf, u32 len)
 {
   /* Prepare header */
-  put_u32(buf+0, now_real);
+  put_u32(buf+0, current_real_time() TO_S);
   put_u16(buf+4, type);
   put_u16(buf+6, subtype);
   put_u32(buf+8, len - MRTDUMP_HDR_LENGTH);
diff --git a/sysdep/unix/timer.h b/sysdep/unix/timer.h
index 495d10c7..989574bf 100644
--- a/sysdep/unix/timer.h
+++ b/sysdep/unix/timer.h
@@ -16,7 +16,7 @@
 
 
 typedef struct timer2 timer;
-
+#if 0
 static inline timer *tm_new(pool *p)
 { return (void *) tm2_new(p); }
 
@@ -44,8 +44,8 @@ static inline void tm_start_max(timer *t, bird_clock_t after)
 static inline timer * tm_new_set(pool *p, void (*hook)(timer *), void *data, uint rand, uint rec)
 { return tm2_new_init(p, hook, data, rec S_, rand S_); }
 
+#endif
 
-#define TIME_INFINITY ((s64) 0x7fffffffffffffff)
 
 
 #endif

From a6f79ca57f0b4b296f67c2d063fd85a627b611b8 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 28 Nov 2017 17:43:20 +0100
Subject: [PATCH 071/198] Timers: Revert temporary names and remove old timer.h

---
 conf/conf.c            | 12 +++++-----
 conf/conf.h            |  2 +-
 conf/flowspec.Y        | 10 ---------
 lib/timer.c            | 46 ++++++++++++++++++-------------------
 lib/timer.h            | 42 +++++++++++++++++-----------------
 nest/proto.c           | 12 +++++-----
 nest/protocol.h        |  1 -
 nest/route.h           |  1 -
 proto/babel/babel.c    | 16 ++++++-------
 proto/babel/babel.h    |  2 +-
 proto/bfd/bfd.c        | 37 +++++++++++++++---------------
 proto/bfd/bfd.h        |  4 ++--
 proto/bfd/io.c         |  4 ++--
 proto/bgp/bgp.c        | 36 ++++++++++++++---------------
 proto/ospf/dbdes.c     | 10 ++++-----
 proto/ospf/iface.c     | 28 +++++++++++------------
 proto/ospf/lsupd.c     | 14 ++++++------
 proto/ospf/neighbor.c  | 26 ++++++++++-----------
 proto/ospf/ospf.c      |  6 ++---
 proto/ospf/ospf.h      |  2 +-
 proto/ospf/rt.c        |  6 ++---
 proto/radv/radv.c      |  6 ++---
 proto/radv/radv.h      |  2 +-
 proto/rip/rip.c        | 20 ++++++++---------
 proto/rip/rip.h        |  2 +-
 proto/rpki/rpki.c      | 22 +++++++++---------
 sysdep/linux/netlink.c |  1 -
 sysdep/unix/io.c       |  5 ++---
 sysdep/unix/krt.c      | 30 ++++++++++++-------------
 sysdep/unix/timer.h    | 51 ------------------------------------------
 30 files changed, 195 insertions(+), 261 deletions(-)
 delete mode 100644 sysdep/unix/timer.h

diff --git a/conf/conf.c b/conf/conf.c
index c68ded7f..62477331 100644
--- a/conf/conf.c
+++ b/conf/conf.c
@@ -52,7 +52,7 @@
 #include "lib/resource.h"
 #include "lib/string.h"
 #include "lib/event.h"
-#include "sysdep/unix/timer.h"
+#include "lib/timer.h"
 #include "conf/conf.h"
 #include "filter/filter.h"
 
@@ -341,9 +341,9 @@ config_commit(struct config *c, int type, uint timeout)
 
   undo_available = 1;
   if (timeout)
-    tm2_start(config_timer, timeout S);
+    tm_start(config_timer, timeout S);
   else
-    tm2_stop(config_timer);
+    tm_stop(config_timer);
 
   if (configuring)
     {
@@ -384,7 +384,7 @@ config_confirm(void)
   if (config_timer->expires == 0)
     return CONF_NOTHING;
 
-  tm2_stop(config_timer);
+  tm_stop(config_timer);
 
   return CONF_CONFIRM;
 }
@@ -420,7 +420,7 @@ config_undo(void)
     return CONF_NOTHING;
 
   undo_available = 0;
-  tm2_stop(config_timer);
+  tm_stop(config_timer);
 
   if (configuring)
     {
@@ -468,7 +468,7 @@ config_init(void)
   config_event = ev_new(&root_pool);
   config_event->hook = config_done;
 
-  config_timer = tm2_new(&root_pool);
+  config_timer = tm_new(&root_pool);
   config_timer->hook = config_timeout;
 }
 
diff --git a/conf/conf.h b/conf/conf.h
index 12f51c9d..d6f59eac 100644
--- a/conf/conf.h
+++ b/conf/conf.h
@@ -13,7 +13,7 @@
 #include "lib/ip.h"
 #include "lib/hash.h"
 #include "lib/resource.h"
-#include "sysdep/unix/timer.h"
+#include "lib/timer.h"
 
 
 /* Configuration structure */
diff --git a/conf/flowspec.Y b/conf/flowspec.Y
index 8c72854c..4d259763 100644
--- a/conf/flowspec.Y
+++ b/conf/flowspec.Y
@@ -11,16 +11,6 @@ CF_HDR
 #define PARSER 1
 
 #include "nest/bird.h"
-#include "conf/conf.h"
-#include "lib/resource.h"
-#include "lib/socket.h"
-#include "sysdep/unix/timer.h"
-#include "lib/string.h"
-#include "nest/protocol.h"
-#include "nest/iface.h"
-#include "nest/route.h"
-#include "nest/cli.h"
-#include "filter/filter.h"
 #include "lib/flowspec.h"
 
 
diff --git a/lib/timer.c b/lib/timer.c
index 338b0a1a..05e488c1 100644
--- a/lib/timer.c
+++ b/lib/timer.c
@@ -93,17 +93,17 @@ current_real_time(void)
 
 
 static void
-tm2_free(resource *r)
+tm_free(resource *r)
 {
-  timer2 *t = (timer2 *) r;
+  timer *t = (void *) r;
 
-  tm2_stop(t);
+  tm_stop(t);
 }
 
 static void
-tm2_dump(resource *r)
+tm_dump(resource *r)
 {
-  timer2 *t = (timer2 *) r;
+  timer *t = (void *) r;
 
   debug("(code %p, data %p, ", t->hook, t->data);
   if (t->randomize)
@@ -117,25 +117,25 @@ tm2_dump(resource *r)
 }
 
 
-static struct resclass tm2_class = {
+static struct resclass tm_class = {
   "Timer",
-  sizeof(timer2),
-  tm2_free,
-  tm2_dump,
+  sizeof(timer),
+  tm_free,
+  tm_dump,
   NULL,
   NULL
 };
 
-timer2 *
-tm2_new(pool *p)
+timer *
+tm_new(pool *p)
 {
-  timer2 *t = ralloc(p, &tm2_class);
+  timer *t = ralloc(p, &tm_class);
   t->index = -1;
   return t;
 }
 
 void
-tm2_set(timer2 *t, btime when)
+tm_set(timer *t, btime when)
 {
   struct timeloop *loop = timeloop_current();
   uint tc = timers_count(loop);
@@ -145,17 +145,17 @@ tm2_set(timer2 *t, btime when)
     t->index = ++tc;
     t->expires = when;
     BUFFER_PUSH(loop->timers) = t;
-    HEAP_INSERT(loop->timers.data, tc, timer2 *, TIMER_LESS, TIMER_SWAP);
+    HEAP_INSERT(loop->timers.data, tc, timer *, TIMER_LESS, TIMER_SWAP);
   }
   else if (t->expires < when)
   {
     t->expires = when;
-    HEAP_INCREASE(loop->timers.data, tc, timer2 *, TIMER_LESS, TIMER_SWAP, t->index);
+    HEAP_INCREASE(loop->timers.data, tc, timer *, TIMER_LESS, TIMER_SWAP, t->index);
   }
   else if (t->expires > when)
   {
     t->expires = when;
-    HEAP_DECREASE(loop->timers.data, tc, timer2 *, TIMER_LESS, TIMER_SWAP, t->index);
+    HEAP_DECREASE(loop->timers.data, tc, timer *, TIMER_LESS, TIMER_SWAP, t->index);
   }
 
 #ifdef CONFIG_BFD
@@ -166,13 +166,13 @@ tm2_set(timer2 *t, btime when)
 }
 
 void
-tm2_start(timer2 *t, btime after)
+tm_start(timer *t, btime after)
 {
-  tm2_set(t, current_time() + MAX(after, 0));
+  tm_set(t, current_time() + MAX(after, 0));
 }
 
 void
-tm2_stop(timer2 *t)
+tm_stop(timer *t)
 {
   if (!t->expires)
     return;
@@ -180,7 +180,7 @@ tm2_stop(timer2 *t)
   struct timeloop *loop = timeloop_current();
   uint tc = timers_count(loop);
 
-  HEAP_DELETE(loop->timers.data, tc, timer2 *, TIMER_LESS, TIMER_SWAP, t->index);
+  HEAP_DELETE(loop->timers.data, tc, timer *, TIMER_LESS, TIMER_SWAP, t->index);
   BUFFER_POP(loop->timers);
 
   t->index = -1;
@@ -202,7 +202,7 @@ void
 timers_fire(struct timeloop *loop)
 {
   btime base_time;
-  timer2 *t;
+  timer *t;
 
   times_update(loop);
   base_time = loop->last_time;
@@ -222,10 +222,10 @@ timers_fire(struct timeloop *loop)
       if (t->randomize)
 	when += random() % (t->randomize + 1);
 
-      tm2_set(t, when);
+      tm_set(t, when);
     }
     else
-      tm2_stop(t);
+      tm_stop(t);
 
     /* This is ugly hack, we want to log just timers executed from the main I/O loop */
     if (loop == &main_timeloop)
diff --git a/lib/timer.h b/lib/timer.h
index 250bb3cd..eeb7dcb7 100644
--- a/lib/timer.h
+++ b/lib/timer.h
@@ -7,18 +7,18 @@
  *	Can be freely distributed and used under the terms of the GNU GPL.
  */
 
-#ifndef _BIRD_TIMER2_H_
-#define _BIRD_TIMER2_H_
+#ifndef _BIRD_TIMER_H_
+#define _BIRD_TIMER_H_
 
 #include "nest/bird.h"
 #include "lib/buffer.h"
 #include "lib/resource.h"
 
 
-typedef struct timer2
+typedef struct timer
 {
   resource r;
-  void (*hook)(struct timer2 *);
+  void (*hook)(struct timer *);
   void *data;
 
   btime expires;			/* 0=inactive */
@@ -26,11 +26,11 @@ typedef struct timer2
   uint recurrent;			/* Timer recurrence */
 
   int index;
-} timer2;
+} timer;
 
 struct timeloop
 {
-  BUFFER(timer2 *) timers;
+  BUFFER(timer *) timers;
   btime last_time;
   btime real_time;
 };
@@ -38,7 +38,7 @@ struct timeloop
 static inline uint timers_count(struct timeloop *loop)
 { return loop->timers.used - 1; }
 
-static inline timer2 *timers_first(struct timeloop *loop)
+static inline timer *timers_first(struct timeloop *loop)
 { return (loop->timers.used > 1) ? loop->timers.data[1] : NULL; }
 
 extern struct timeloop main_timeloop;
@@ -50,28 +50,28 @@ btime current_real_time(void);
 //#define now_real (current_real_time() TO_S)
 extern btime boot_time;
 
-timer2 *tm2_new(pool *p);
-void tm2_set(timer2 *t, btime when);
-void tm2_start(timer2 *t, btime after);
-void tm2_stop(timer2 *t);
+timer *tm_new(pool *p);
+void tm_set(timer *t, btime when);
+void tm_start(timer *t, btime after);
+void tm_stop(timer *t);
 
 static inline int
-tm2_active(timer2 *t)
+tm_active(timer *t)
 {
   return t->expires != 0;
 }
 
 static inline btime
-tm2_remains(timer2 *t)
+tm_remains(timer *t)
 {
   btime now_ = current_time();
   return (t->expires > now_) ? (t->expires - now_) : 0;
 }
 
-static inline timer2 *
-tm2_new_init(pool *p, void (*hook)(struct timer2 *), void *data, uint rec, uint rand)
+static inline timer *
+tm_new_init(pool *p, void (*hook)(struct timer *), void *data, uint rec, uint rand)
 {
-  timer2 *t = tm2_new(p);
+  timer *t = tm_new(p);
   t->hook = hook;
   t->data = data;
   t->recurrent = rec;
@@ -80,17 +80,17 @@ tm2_new_init(pool *p, void (*hook)(struct timer2 *), void *data, uint rec, uint
 }
 
 static inline void
-tm2_set_max(timer2 *t, btime when)
+tm_set_max(timer *t, btime when)
 {
   if (when > t->expires)
-    tm2_set(t, when);
+    tm_set(t, when);
 }
 
 static inline void
-tm2_start_max(timer2 *t, btime after)
+tm_start_max(timer *t, btime after)
 {
-  btime rem = tm2_remains(t);
-  tm2_start(t, MAX_(rem, after));
+  btime rem = tm_remains(t);
+  tm_start(t, MAX_(rem, after));
 }
 
 /* In sysdep code */
diff --git a/nest/proto.c b/nest/proto.c
index 72f1f94d..b28ac569 100644
--- a/nest/proto.c
+++ b/nest/proto.c
@@ -1083,8 +1083,8 @@ graceful_restart_init(void)
   }
 
   graceful_restart_state = GRS_ACTIVE;
-  gr_wait_timer = tm2_new_init(proto_pool, graceful_restart_done, NULL, 0, 0);
-  tm2_start(gr_wait_timer, config->gr_wait S);
+  gr_wait_timer = tm_new_init(proto_pool, graceful_restart_done, NULL, 0, 0);
+  tm_start(gr_wait_timer, config->gr_wait S);
 }
 
 /**
@@ -1135,7 +1135,7 @@ graceful_restart_show_status(void)
 
   cli_msg(-24, "Graceful restart recovery in progress");
   cli_msg(-24, "  Waiting for %d channels to recover", graceful_restart_locks);
-  cli_msg(-24, "  Wait timer is %t/%u", tm2_remains(gr_wait_timer), config->gr_wait);
+  cli_msg(-24, "  Wait timer is %t/%u", tm_remains(gr_wait_timer), config->gr_wait);
 }
 
 /**
@@ -1181,7 +1181,7 @@ channel_graceful_restart_unlock(struct channel *c)
   graceful_restart_locks--;
 
   if ((graceful_restart_state == GRS_ACTIVE) && !graceful_restart_locks)
-    tm2_start(gr_wait_timer, 0);
+    tm_start(gr_wait_timer, 0);
 }
 
 
@@ -1288,7 +1288,7 @@ protos_build(void)
 #endif
 
   proto_pool = rp_new(&root_pool, "Protocols");
-  proto_shutdown_timer = tm2_new(proto_pool);
+  proto_shutdown_timer = tm_new(proto_pool);
   proto_shutdown_timer->hook = proto_shutdown_loop;
 }
 
@@ -1328,7 +1328,7 @@ proto_schedule_down(struct proto *p, byte restart, byte code)
 
   p->down_sched = restart ? PDS_RESTART : PDS_DISABLE;
   p->down_code = code;
-  tm2_start_max(proto_shutdown_timer, restart ? 250 MS : 0);
+  tm_start_max(proto_shutdown_timer, restart ? 250 MS : 0);
 }
 
 
diff --git a/nest/protocol.h b/nest/protocol.h
index abd11aff..e843b0b4 100644
--- a/nest/protocol.h
+++ b/nest/protocol.h
@@ -12,7 +12,6 @@
 #include "lib/lists.h"
 #include "lib/resource.h"
 #include "lib/event.h"
-#include "sysdep/unix/timer.h"
 #include "nest/route.h"
 #include "conf/conf.h"
 
diff --git a/nest/route.h b/nest/route.h
index a838bd38..c9e2b3bf 100644
--- a/nest/route.h
+++ b/nest/route.h
@@ -11,7 +11,6 @@
 
 #include "lib/lists.h"
 #include "lib/resource.h"
-#include "sysdep/unix/timer.h"
 #include "lib/net.h"
 
 struct ea_list;
diff --git a/proto/babel/babel.c b/proto/babel/babel.c
index 768bbda8..aa7e8b68 100644
--- a/proto/babel/babel.c
+++ b/proto/babel/babel.c
@@ -1433,14 +1433,14 @@ babel_iface_timer(timer *t)
 
   btime next_event = MIN(ifa->next_hello, ifa->next_regular);
   if (ifa->want_triggered) next_event = MIN(next_event, ifa->next_triggered);
-  tm2_set(ifa->timer, next_event);
+  tm_set(ifa->timer, next_event);
 }
 
 static inline void
 babel_iface_kick_timer(struct babel_iface *ifa)
 {
   if (ifa->timer->expires > (current_time() + 100 MS))
-    tm2_start(ifa->timer, 100 MS);
+    tm_start(ifa->timer, 100 MS);
 }
 
 static void
@@ -1454,7 +1454,7 @@ babel_iface_start(struct babel_iface *ifa)
   ifa->next_regular = current_time() + (random() % ifa->cf->update_interval);
   ifa->next_triggered = current_time() + MIN(1 S, ifa->cf->update_interval / 2);
   ifa->want_triggered = 0;	/* We send an immediate update (below) */
-  tm2_start(ifa->timer, 100 MS);
+  tm_start(ifa->timer, 100 MS);
   ifa->up = 1;
 
   babel_send_hello(ifa);
@@ -1487,7 +1487,7 @@ babel_iface_stop(struct babel_iface *ifa)
     }
   }
 
-  tm2_stop(ifa->timer);
+  tm_stop(ifa->timer);
   ifa->up = 0;
 }
 
@@ -1585,7 +1585,7 @@ babel_add_iface(struct babel_proto *p, struct iface *new, struct babel_iface_con
   init_list(&ifa->neigh_list);
   ifa->hello_seqno = 1;
 
-  ifa->timer = tm2_new_init(ifa->pool, babel_iface_timer, ifa, 0, 0);
+  ifa->timer = tm_new_init(ifa->pool, babel_iface_timer, ifa, 0, 0);
 
   init_list(&ifa->msg_queue);
   ifa->send_event = ev_new(ifa->pool);
@@ -2050,7 +2050,7 @@ static inline void
 babel_kick_timer(struct babel_proto *p)
 {
   if (p->timer->expires > (current_time() + 100 MS))
-    tm2_start(p->timer, 100 MS);
+    tm_start(p->timer, 100 MS);
 }
 
 
@@ -2214,8 +2214,8 @@ babel_start(struct proto *P)
 	   OFFSETOF(struct babel_entry, n), 0, babel_init_entry);
 
   init_list(&p->interfaces);
-  p->timer = tm2_new_init(P->pool, babel_timer, p, 1 S, 0);
-  tm2_start(p->timer, 1 S);
+  p->timer = tm_new_init(P->pool, babel_timer, p, 1 S, 0);
+  tm_start(p->timer, 1 S);
   p->update_seqno = 1;
   p->router_id = proto_get_router_id(&cf->c);
 
diff --git a/proto/babel/babel.h b/proto/babel/babel.h
index da3293ef..1128d261 100644
--- a/proto/babel/babel.h
+++ b/proto/babel/babel.h
@@ -23,7 +23,7 @@
 #include "lib/lists.h"
 #include "lib/socket.h"
 #include "lib/string.h"
-#include "sysdep/unix/timer.h"
+#include "lib/timer.h"
 
 #define EA_BABEL_METRIC		EA_CODE(EAP_BABEL, 0)
 #define EA_BABEL_ROUTER_ID	EA_CODE(EAP_BABEL, 1)
diff --git a/proto/bfd/bfd.c b/proto/bfd/bfd.c
index 8915140d..67ec2270 100644
--- a/proto/bfd/bfd.c
+++ b/proto/bfd/bfd.c
@@ -64,16 +64,15 @@
  * ready, the protocol just creates a BFD request like any other protocol.
  *
  * The protocol uses a new generic event loop (structure &birdloop) from |io.c|,
- * which supports sockets, timers and events like the main loop. Timers
- * (structure &timer2) are new microsecond based timers, while sockets and
- * events are the same. A birdloop is associated with a thread (field @thread)
- * in which event hooks are executed. Most functions for setting event sources
- * (like sk_start() or tm2_start()) must be called from the context of that
- * thread. Birdloop allows to temporarily acquire the context of that thread for
- * the main thread by calling birdloop_enter() and then birdloop_leave(), which
- * also ensures mutual exclusion with all event hooks. Note that resources
- * associated with a birdloop (like timers) should be attached to the
- * independent resource pool, detached from the main resource tree.
+ * which supports sockets, timers and events like the main loop. A birdloop is
+ * associated with a thread (field @thread) in which event hooks are executed.
+ * Most functions for setting event sources (like sk_start() or tm_start()) must
+ * be called from the context of that thread. Birdloop allows to temporarily
+ * acquire the context of that thread for the main thread by calling
+ * birdloop_enter() and then birdloop_leave(), which also ensures mutual
+ * exclusion with all event hooks. Note that resources associated with a
+ * birdloop (like timers) should be attached to the independent resource pool,
+ * detached from the main resource tree.
  *
  * There are two kinds of interaction between the BFD core (running in the BFD
  * thread) and the rest of BFD (running in the main thread). The first kind are
@@ -177,7 +176,7 @@ bfd_session_update_tx_interval(struct bfd_session *s)
     return;
 
   /* Set timer relative to last tx_timer event */
-  tm2_set(s->tx_timer, s->last_tx + tx_int_l);
+  tm_set(s->tx_timer, s->last_tx + tx_int_l);
 }
 
 static void
@@ -191,7 +190,7 @@ bfd_session_update_detection_time(struct bfd_session *s, int kick)
   if (!s->last_rx)
     return;
 
-  tm2_set(s->hold_timer, s->last_rx + timeout);
+  tm_set(s->hold_timer, s->last_rx + timeout);
 }
 
 static void
@@ -212,16 +211,16 @@ bfd_session_control_tx_timer(struct bfd_session *s, int reset)
     goto stop;
 
   /* So TX timer should run */
-  if (reset || !tm2_active(s->tx_timer))
+  if (reset || !tm_active(s->tx_timer))
   {
     s->last_tx = 0;
-    tm2_start(s->tx_timer, 0);
+    tm_start(s->tx_timer, 0);
   }
 
   return;
 
  stop:
-  tm2_stop(s->tx_timer);
+  tm_stop(s->tx_timer);
   s->last_tx = 0;
 }
 
@@ -380,7 +379,7 @@ bfd_find_session_by_addr(struct bfd_proto *p, ip_addr addr)
 }
 
 static void
-bfd_tx_timer_hook(timer2 *t)
+bfd_tx_timer_hook(timer *t)
 {
   struct bfd_session *s = t->data;
 
@@ -389,7 +388,7 @@ bfd_tx_timer_hook(timer2 *t)
 }
 
 static void
-bfd_hold_timer_hook(timer2 *t)
+bfd_hold_timer_hook(timer *t)
 {
   bfd_session_timeout(t->data);
 }
@@ -433,8 +432,8 @@ bfd_add_session(struct bfd_proto *p, ip_addr addr, ip_addr local, struct iface *
   s->passive = ifa->cf->passive;
   s->tx_csn = random_u32();
 
-  s->tx_timer = tm2_new_init(p->tpool, bfd_tx_timer_hook, s, 0, 0);
-  s->hold_timer = tm2_new_init(p->tpool, bfd_hold_timer_hook, s, 0, 0);
+  s->tx_timer = tm_new_init(p->tpool, bfd_tx_timer_hook, s, 0, 0);
+  s->hold_timer = tm_new_init(p->tpool, bfd_hold_timer_hook, s, 0, 0);
   bfd_session_update_tx_interval(s);
   bfd_session_control_tx_timer(s, 1);
 
diff --git a/proto/bfd/bfd.h b/proto/bfd/bfd.h
index 203da437..bc4fe969 100644
--- a/proto/bfd/bfd.h
+++ b/proto/bfd/bfd.h
@@ -140,8 +140,8 @@ struct bfd_session
   btime last_tx;			/* Time of last sent periodic control packet */
   btime last_rx;			/* Time of last received valid control packet */
 
-  timer2 *tx_timer;			/* Periodic control packet timer */
-  timer2 *hold_timer;			/* Timer for session down detection time */
+  timer *tx_timer;			/* Periodic control packet timer */
+  timer *hold_timer;			/* Timer for session down detection time */
 
   list request_list;			/* List of client requests (struct bfd_request) */
   btime last_state_change;		/* Time of last state change */
diff --git a/proto/bfd/io.c b/proto/bfd/io.c
index ab846113..b01cbfce 100644
--- a/proto/bfd/io.c
+++ b/proto/bfd/io.c
@@ -477,7 +477,7 @@ static void *
 birdloop_main(void *arg)
 {
   struct birdloop *loop = arg;
-  timer2 *t;
+  timer *t;
   int rv, timeout;
 
   birdloop_set_current(loop);
@@ -492,7 +492,7 @@ birdloop_main(void *arg)
     if (events_waiting(loop))
       timeout = 0;
     else if (t = timers_first(&loop->time))
-      timeout = (tm2_remains(t) TO_MS) + 1;
+      timeout = (tm_remains(t) TO_MS) + 1;
     else
       timeout = -1;
 
diff --git a/proto/bgp/bgp.c b/proto/bgp/bgp.c
index df4c240a..f4791215 100644
--- a/proto/bgp/bgp.c
+++ b/proto/bgp/bgp.c
@@ -328,10 +328,10 @@ bgp_start_timer(timer *t, uint value)
     /* The randomization procedure is specified in RFC 4271 section 10 */
     btime time = value S;
     btime randomize = random() % ((time / 4) + 1);
-    tm2_start(t, time - randomize);
+    tm_start(t, time - randomize);
   }
   else
-    tm2_stop(t);
+    tm_stop(t);
 }
 
 /**
@@ -517,7 +517,7 @@ bgp_conn_enter_established_state(struct bgp_conn *conn)
   int peer_gr_ready = peer->gr_aware && !(peer->gr_flags & BGP_GRF_RESTART);
 
   if (p->gr_active_num)
-    tm2_stop(p->gr_timer);
+    tm_stop(p->gr_timer);
 
   /* Number of active channels */
   int num = 0;
@@ -616,7 +616,7 @@ bgp_conn_enter_close_state(struct bgp_conn *conn)
   int os = conn->state;
 
   bgp_conn_set_state(conn, BS_CLOSE);
-  tm2_stop(conn->keepalive_timer);
+  tm_stop(conn->keepalive_timer);
   conn->sk->rx_hook = NULL;
 
   /* Timeout for CLOSE state, if we cannot send notification soon then we just hangup */
@@ -779,7 +779,7 @@ bgp_send_open(struct bgp_conn *conn)
   DBG("BGP: Sending open\n");
   conn->sk->rx_hook = bgp_rx;
   conn->sk->tx_hook = bgp_tx;
-  tm2_stop(conn->connect_timer);
+  tm_stop(conn->connect_timer);
   bgp_schedule_packet(conn, NULL, PKT_OPEN);
   bgp_conn_set_state(conn, BS_OPENSENT);
   bgp_start_timer(conn->hold_timer, conn->bgp->cf->initial_hold_time);
@@ -888,9 +888,9 @@ bgp_setup_conn(struct bgp_proto *p, struct bgp_conn *conn)
   conn->last_channel = 0;
   conn->last_channel_count = 0;
 
-  conn->connect_timer	= tm2_new_init(p->p.pool, bgp_connect_timeout,	 conn, 0, 0);
-  conn->hold_timer 	= tm2_new_init(p->p.pool, bgp_hold_timeout,	 conn, 0, 0);
-  conn->keepalive_timer	= tm2_new_init(p->p.pool, bgp_keepalive_timeout, conn, 0, 0);
+  conn->connect_timer	= tm_new_init(p->p.pool, bgp_connect_timeout,	 conn, 0, 0);
+  conn->hold_timer 	= tm_new_init(p->p.pool, bgp_hold_timeout,	 conn, 0, 0);
+  conn->keepalive_timer	= tm_new_init(p->p.pool, bgp_keepalive_timeout, conn, 0, 0);
 
   conn->tx_ev = ev_new(p->p.pool);
   conn->tx_ev->hook = bgp_kick_tx;
@@ -1303,8 +1303,8 @@ bgp_start(struct proto *P)
   p->event->hook = bgp_decision;
   p->event->data = p;
 
-  p->startup_timer = tm2_new_init(p->p.pool, bgp_startup_timeout, p, 0, 0);
-  p->gr_timer = tm2_new_init(p->p.pool, bgp_graceful_restart_timeout, p, 0, 0);
+  p->startup_timer = tm_new_init(p->p.pool, bgp_startup_timeout, p, 0, 0);
+  p->gr_timer = tm_new_init(p->p.pool, bgp_graceful_restart_timeout, p, 0, 0);
 
   p->local_id = proto_get_router_id(P->cf);
   if (p->rr_client)
@@ -2004,18 +2004,18 @@ bgp_show_proto_info(struct proto *P)
     struct bgp_conn *oc = &p->outgoing_conn;
 
     if ((p->start_state < BSS_CONNECT) &&
-	(tm2_active(p->startup_timer)))
+	(tm_active(p->startup_timer)))
       cli_msg(-1006, "    Error wait:       %t/%u",
-	      tm2_remains(p->startup_timer), p->startup_delay);
+	      tm_remains(p->startup_timer), p->startup_delay);
 
     if ((oc->state == BS_ACTIVE) &&
-	(tm2_active(oc->connect_timer)))
+	(tm_active(oc->connect_timer)))
       cli_msg(-1006, "    Connect delay:    %t/%u",
-	      tm2_remains(oc->connect_timer), p->cf->connect_delay_time);
+	      tm_remains(oc->connect_timer), p->cf->connect_delay_time);
 
-    if (p->gr_active_num && tm2_active(p->gr_timer))
+    if (p->gr_active_num && tm_active(p->gr_timer))
       cli_msg(-1006, "    Restart timer:    %t/-",
-	      tm2_remains(p->gr_timer));
+	      tm_remains(p->gr_timer));
   }
   else if (P->proto_state == PS_UP)
   {
@@ -2037,9 +2037,9 @@ bgp_show_proto_info(struct proto *P)
 */
     cli_msg(-1006, "    Source address:   %I", p->source_addr);
     cli_msg(-1006, "    Hold timer:       %t/%u",
-	    tm2_remains(p->conn->hold_timer), p->conn->hold_time);
+	    tm_remains(p->conn->hold_timer), p->conn->hold_time);
     cli_msg(-1006, "    Keepalive timer:  %t/%u",
-	    tm2_remains(p->conn->keepalive_timer), p->conn->keepalive_time);
+	    tm_remains(p->conn->keepalive_timer), p->conn->keepalive_time);
   }
 
   if ((p->last_error_class != BE_NONE) &&
diff --git a/proto/ospf/dbdes.c b/proto/ospf/dbdes.c
index 270259a7..f211935f 100644
--- a/proto/ospf/dbdes.c
+++ b/proto/ospf/dbdes.c
@@ -279,8 +279,8 @@ ospf_process_dbdes(struct ospf_proto *p, struct ospf_packet *pkt, struct ospf_ne
       req->lsa = lsa;
       req->lsa_body = LSA_BODY_DUMMY;
 
-      if (!tm2_active(n->lsrq_timer))
-	tm2_start(n->lsrq_timer, 0);
+      if (!tm_active(n->lsrq_timer))
+	tm_start(n->lsrq_timer, 0);
     }
   }
 
@@ -366,7 +366,7 @@ ospf_receive_dbdes(struct ospf_packet *pkt, struct ospf_iface *ifa,
       n->options = rcv_options;
       n->myimms &= ~DBDES_MS;
       n->imms = rcv_imms;
-      tm2_stop(n->dbdes_timer);
+      tm_stop(n->dbdes_timer);
       ospf_neigh_sm(n, INM_NEGDONE);
       ospf_send_dbdes(p, n);
       break;
@@ -422,13 +422,13 @@ ospf_receive_dbdes(struct ospf_packet *pkt, struct ospf_iface *ifa,
 
       if (!(n->myimms & DBDES_M) && !(n->imms & DBDES_M))
       {
-	tm2_stop(n->dbdes_timer);
+	tm_stop(n->dbdes_timer);
 	ospf_neigh_sm(n, INM_EXDONE);
 	break;
       }
 
       ospf_send_dbdes(p, n);
-      tm2_start(n->dbdes_timer, n->ifa->rxmtint S);
+      tm_start(n->dbdes_timer, n->ifa->rxmtint S);
     }
     else
     {
diff --git a/proto/ospf/iface.c b/proto/ospf/iface.c
index 33ec21fb..29d21a07 100644
--- a/proto/ospf/iface.c
+++ b/proto/ospf/iface.c
@@ -263,13 +263,13 @@ ospf_iface_down(struct ospf_iface *ifa)
     ospf_neigh_sm(n, INM_KILLNBR);
 
   if (ifa->hello_timer)
-    tm2_stop(ifa->hello_timer);
+    tm_stop(ifa->hello_timer);
 
   if (ifa->poll_timer)
-    tm2_stop(ifa->poll_timer);
+    tm_stop(ifa->poll_timer);
 
   if (ifa->wait_timer)
-    tm2_stop(ifa->wait_timer);
+    tm_stop(ifa->wait_timer);
 
   ospf_flush2_lsa(p, &ifa->link_lsa);
   ospf_flush2_lsa(p, &ifa->net_lsa);
@@ -396,15 +396,15 @@ ospf_iface_sm(struct ospf_iface *ifa, int event)
 	{
 	  ospf_iface_chstate(ifa, OSPF_IS_WAITING);
 	  if (ifa->wait_timer)
-	    tm2_start(ifa->wait_timer, ifa->waitint S);
+	    tm_start(ifa->wait_timer, ifa->waitint S);
 	}
       }
 
       if (ifa->hello_timer)
-	tm2_start(ifa->hello_timer, ifa->helloint S);
+	tm_start(ifa->hello_timer, ifa->helloint S);
 
       if (ifa->poll_timer)
-	tm2_start(ifa->poll_timer, ifa->pollint S);
+	tm_start(ifa->poll_timer, ifa->pollint S);
 
       ospf_send_hello(ifa, OHS_HELLO, NULL);
     }
@@ -494,13 +494,13 @@ ospf_iface_add(struct object_lock *lock)
 
   if (! ifa->stub)
   {
-    ifa->hello_timer = tm2_new_init(ifa->pool, hello_timer_hook, ifa, ifa->helloint S, 0);
+    ifa->hello_timer = tm_new_init(ifa->pool, hello_timer_hook, ifa, ifa->helloint S, 0);
 
     if (ifa->type == OSPF_IT_NBMA)
-      ifa->poll_timer = tm2_new_init(ifa->pool, poll_timer_hook, ifa, ifa->pollint S, 0);
+      ifa->poll_timer = tm_new_init(ifa->pool, poll_timer_hook, ifa, ifa->pollint S, 0);
 
     if ((ifa->type == OSPF_IT_BCAST) || (ifa->type == OSPF_IT_NBMA))
-      ifa->wait_timer = tm2_new_init(ifa->pool, wait_timer_hook, ifa, 0, 0);
+      ifa->wait_timer = tm_new_init(ifa->pool, wait_timer_hook, ifa, 0, 0);
 
     ifa->flood_queue_size = ifa_flood_queue_size(ifa);
     ifa->flood_queue = mb_allocz(ifa->pool, ifa->flood_queue_size * sizeof(void *));
@@ -703,7 +703,7 @@ ospf_iface_new_vlink(struct ospf_proto *p, struct ospf_iface_patt *ip)
 
   add_tail(&p->iface_list, NODE ifa);
 
-  ifa->hello_timer = tm2_new_init(ifa->pool, hello_timer_hook, ifa, ifa->helloint S, 0);
+  ifa->hello_timer = tm_new_init(ifa->pool, hello_timer_hook, ifa, ifa->helloint S, 0);
 
   ifa->flood_queue_size = ifa_flood_queue_size(ifa);
   ifa->flood_queue = mb_allocz(ifa->pool, ifa->flood_queue_size * sizeof(void *));
@@ -717,8 +717,8 @@ ospf_iface_change_timer(timer *tm, uint val)
 
   tm->recurrent = val S;
 
-  if (tm2_active(tm))
-    tm2_start(tm, val S);
+  if (tm_active(tm))
+    tm_start(tm, val S);
 }
 
 static inline void
@@ -801,8 +801,8 @@ ospf_iface_reconfigure(struct ospf_iface *ifa, struct ospf_iface_patt *new)
 	       ifname, ifa->waitint, new->waitint);
 
     ifa->waitint = new->waitint;
-    if (ifa->wait_timer && tm2_active(ifa->wait_timer))
-      tm2_start(ifa->wait_timer, ifa->waitint S);
+    if (ifa->wait_timer && tm_active(ifa->wait_timer))
+      tm_start(ifa->wait_timer, ifa->waitint S);
   }
 
   /* DEAD TIMER */
diff --git a/proto/ospf/lsupd.c b/proto/ospf/lsupd.c
index 18811392..a98c9098 100644
--- a/proto/ospf/lsupd.c
+++ b/proto/ospf/lsupd.c
@@ -115,7 +115,7 @@ ospf_lsa_lsrq_down(struct top_hash_entry *req, struct ospf_neighbor *n)
 
   if (EMPTY_SLIST(n->lsrql))
   {
-    tm2_stop(n->lsrq_timer);
+    tm_stop(n->lsrq_timer);
 
     if (n->state == NEIGHBOR_LOADING)
       ospf_neigh_sm(n, INM_LOADDONE);
@@ -136,8 +136,8 @@ ospf_lsa_lsrt_up(struct top_hash_entry *en, struct ospf_neighbor *n)
   ret->lsa = en->lsa;
   ret->lsa_body = LSA_BODY_DUMMY;
 
-  if (!tm2_active(n->lsrt_timer))
-    tm2_start(n->lsrt_timer, n->ifa->rxmtint S);
+  if (!tm_active(n->lsrt_timer))
+    tm_start(n->lsrt_timer, n->ifa->rxmtint S);
 }
 
 void
@@ -150,7 +150,7 @@ ospf_lsa_lsrt_down_(struct top_hash_entry *en, struct ospf_neighbor *n, struct t
   ospf_hash_delete(n->lsrth, ret);
 
   if (EMPTY_SLIST(n->lsrtl))
-    tm2_stop(n->lsrt_timer);
+    tm_stop(n->lsrt_timer);
 }
 
 static inline int
@@ -175,8 +175,8 @@ ospf_add_flushed_to_lsrt(struct ospf_proto *p, struct ospf_neighbor *n)
       ospf_lsa_lsrt_up(en, n);
 
   /* If we found any flushed LSA, we send them ASAP */
-  if (tm2_active(n->lsrt_timer))
-    tm2_start(n->lsrt_timer, 0);
+  if (tm_active(n->lsrt_timer))
+    tm_start(n->lsrt_timer, 0);
 }
 
 static int ospf_flood_lsupd(struct ospf_proto *p, struct top_hash_entry **lsa_list, uint lsa_count, uint lsa_min_count, struct ospf_iface *ifa);
@@ -700,7 +700,7 @@ ospf_receive_lsupd(struct ospf_packet *pkt, struct ospf_iface *ifa,
   if (!EMPTY_SLIST(n->lsrql) && (n->lsrqi == SHEAD(n->lsrql)))
   {
     ospf_send_lsreq(p, n);
-    tm2_start(n->lsrq_timer, n->ifa->rxmtint S);
+    tm_start(n->lsrq_timer, n->ifa->rxmtint S);
   }
 
   return;
diff --git a/proto/ospf/neighbor.c b/proto/ospf/neighbor.c
index 12eb9e84..f2d3505e 100644
--- a/proto/ospf/neighbor.c
+++ b/proto/ospf/neighbor.c
@@ -66,10 +66,10 @@ reset_lists(struct ospf_proto *p, struct ospf_neighbor *n)
   ospf_top_free(n->lsrth);
   ospf_reset_lsack_queue(n);
 
-  tm2_stop(n->dbdes_timer);
-  tm2_stop(n->lsrq_timer);
-  tm2_stop(n->lsrt_timer);
-  tm2_stop(n->ackd_timer);
+  tm_stop(n->dbdes_timer);
+  tm_stop(n->lsrq_timer);
+  tm_stop(n->lsrt_timer);
+  tm_stop(n->ackd_timer);
 
   init_lists(p, n);
 }
@@ -94,11 +94,11 @@ ospf_neighbor_new(struct ospf_iface *ifa)
   init_list(&n->ackl[ACKL_DIRECT]);
   init_list(&n->ackl[ACKL_DELAY]);
 
-  n->inactim = tm2_new_init(pool, inactivity_timer_hook, n, 0, 0);
-  n->dbdes_timer = tm2_new_init(pool, dbdes_timer_hook, n, ifa->rxmtint S, 0);
-  n->lsrq_timer = tm2_new_init(pool, lsrq_timer_hook, n, ifa->rxmtint S, 0);
-  n->lsrt_timer = tm2_new_init(pool, lsrt_timer_hook, n, ifa->rxmtint S, 0);
-  n->ackd_timer = tm2_new_init(pool, ackd_timer_hook, n, ifa->rxmtint S / 2, 0);
+  n->inactim = tm_new_init(pool, inactivity_timer_hook, n, 0, 0);
+  n->dbdes_timer = tm_new_init(pool, dbdes_timer_hook, n, ifa->rxmtint S, 0);
+  n->lsrq_timer = tm_new_init(pool, lsrq_timer_hook, n, ifa->rxmtint S, 0);
+  n->lsrt_timer = tm_new_init(pool, lsrt_timer_hook, n, ifa->rxmtint S, 0);
+  n->ackd_timer = tm_new_init(pool, ackd_timer_hook, n, ifa->rxmtint S / 2, 0);
 
   return (n);
 }
@@ -185,8 +185,8 @@ ospf_neigh_chstate(struct ospf_neighbor *n, u8 state)
     n->dds++;
     n->myimms = DBDES_IMMS;
 
-    tm2_start(n->dbdes_timer, 0);
-    tm2_start(n->ackd_timer, ifa->rxmtint S / 2);
+    tm_start(n->dbdes_timer, 0);
+    tm_start(n->ackd_timer, ifa->rxmtint S / 2);
   }
 
   if (state > NEIGHBOR_EXSTART)
@@ -231,7 +231,7 @@ ospf_neigh_sm(struct ospf_neighbor *n, int event)
       ospf_neigh_chstate(n, NEIGHBOR_INIT);
 
     /* Restart inactivity timer */
-    tm2_start(n->inactim, n->ifa->deadint S);
+    tm_start(n->inactim, n->ifa->deadint S);
     break;
 
   case INM_2WAYREC:
@@ -664,5 +664,5 @@ ospf_sh_neigh_info(struct ospf_neighbor *n)
 
   cli_msg(-1013, "%-1R\t%3u\t%s/%s\t%7t\t%-10s %-1I",
 	  n->rid, n->priority, ospf_ns_names[n->state], pos,
-	  tm2_remains(n->inactim), ifa->ifname, n->ip);
+	  tm_remains(n->inactim), ifa->ifname, n->ip);
 }
diff --git a/proto/ospf/ospf.c b/proto/ospf/ospf.c
index 13f3845b..3ebebdaa 100644
--- a/proto/ospf/ospf.c
+++ b/proto/ospf/ospf.c
@@ -241,8 +241,8 @@ ospf_start(struct proto *P)
   p->asbr = c->asbr;
   p->ecmp = c->ecmp;
   p->tick = c->tick;
-  p->disp_timer = tm2_new_init(P->pool, ospf_disp, p, p->tick S, 0);
-  tm2_start(p->disp_timer, 100 MS);
+  p->disp_timer = tm_new_init(P->pool, ospf_disp, p, p->tick S, 0);
+  tm_start(p->disp_timer, 100 MS);
   p->lsab_size = 256;
   p->lsab_used = 0;
   p->lsab = mb_alloc(P->pool, p->lsab_size);
@@ -677,7 +677,7 @@ ospf_reconfigure(struct proto *P, struct proto_config *CF)
   p->ecmp = new->ecmp;
   p->tick = new->tick;
   p->disp_timer->recurrent = p->tick S;
-  tm2_start(p->disp_timer, 100 MS);
+  tm_start(p->disp_timer, 100 MS);
 
   /* Mark all areas and ifaces */
   WALK_LIST(oa, p->area_list)
diff --git a/proto/ospf/ospf.h b/proto/ospf/ospf.h
index d4571bf6..54eeb74c 100644
--- a/proto/ospf/ospf.h
+++ b/proto/ospf/ospf.h
@@ -18,7 +18,7 @@
 #include "lib/lists.h"
 #include "lib/slists.h"
 #include "lib/socket.h"
-#include "sysdep/unix/timer.h"
+#include "lib/timer.h"
 #include "lib/resource.h"
 #include "nest/protocol.h"
 #include "nest/iface.h"
diff --git a/proto/ospf/rt.c b/proto/ospf/rt.c
index 36bf0387..c0fe218a 100644
--- a/proto/ospf/rt.c
+++ b/proto/ospf/rt.c
@@ -1321,7 +1321,7 @@ ospf_rt_abr2(struct ospf_proto *p)
       if (translate && (oa->translate != TRANS_ON))
       {
 	if (oa->translate == TRANS_WAIT)
-	  tm2_stop(oa->translator_timer);
+	  tm_stop(oa->translator_timer);
 
 	oa->translate = TRANS_ON;
       }
@@ -1329,10 +1329,10 @@ ospf_rt_abr2(struct ospf_proto *p)
       if (!translate && (oa->translate == TRANS_ON))
       {
 	if (oa->translator_timer == NULL)
-	  oa->translator_timer = tm2_new_init(p->p.pool, translator_timer_hook, oa, 0, 0);
+	  oa->translator_timer = tm_new_init(p->p.pool, translator_timer_hook, oa, 0, 0);
 
 	/* Schedule the end of translation */
-	tm2_start(oa->translator_timer, oa->ac->transint S);
+	tm_start(oa->translator_timer, oa->ac->transint S);
 	oa->translate = TRANS_WAIT;
       }
     }
diff --git a/proto/radv/radv.c b/proto/radv/radv.c
index 415cd1d9..fe371ab4 100644
--- a/proto/radv/radv.c
+++ b/proto/radv/radv.c
@@ -65,7 +65,7 @@ radv_timer(timer *tm)
     ifa->initial--;
   }
 
-  tm2_start(ifa->timer, t);
+  tm_start(ifa->timer, t);
 }
 
 static char* ev_name[] = { NULL, "Init", "Change", "RS" };
@@ -94,7 +94,7 @@ radv_iface_notify(struct radv_iface *ifa, int event)
 
   /* Update timer */
   btime t = ifa->last + ifa->cf->min_delay S - current_time();
-  tm2_start(ifa->timer, t);
+  tm_start(ifa->timer, t);
 }
 
 static void
@@ -150,7 +150,7 @@ radv_iface_new(struct radv_proto *p, struct iface *iface, struct radv_iface_conf
 
   add_tail(&p->iface_list, NODE ifa);
 
-  ifa->timer = tm2_new_init(pool, radv_timer, ifa, 0, 0);
+  ifa->timer = tm_new_init(pool, radv_timer, ifa, 0, 0);
 
   struct object_lock *lock = olock_new(pool);
   lock->addr = IPA_NONE;
diff --git a/proto/radv/radv.h b/proto/radv/radv.h
index a4429c60..8324bb67 100644
--- a/proto/radv/radv.h
+++ b/proto/radv/radv.h
@@ -13,7 +13,7 @@
 #include "lib/ip.h"
 #include "lib/lists.h"
 #include "lib/socket.h"
-#include "sysdep/unix/timer.h"
+#include "lib/timer.h"
 #include "lib/resource.h"
 #include "nest/protocol.h"
 #include "nest/iface.h"
diff --git a/proto/rip/rip.c b/proto/rip/rip.c
index 782e6f5c..a3eeaf17 100644
--- a/proto/rip/rip.c
+++ b/proto/rip/rip.c
@@ -509,7 +509,7 @@ rip_iface_start(struct rip_iface *ifa)
   ifa->next_regular = current_time() + (random() % ifa->cf->update_time) + 100 MS;
   ifa->next_triggered = current_time();	/* Available immediately */
   ifa->want_triggered = 1;		/* All routes in triggered update */
-  tm2_start(ifa->timer, 100 MS);
+  tm_start(ifa->timer, 100 MS);
   ifa->up = 1;
 
   if (!ifa->cf->passive)
@@ -529,7 +529,7 @@ rip_iface_stop(struct rip_iface *ifa)
   WALK_LIST_FIRST(n, ifa->neigh_list)
     rip_remove_neighbor(p, n);
 
-  tm2_stop(ifa->timer);
+  tm_stop(ifa->timer);
   ifa->up = 0;
 }
 
@@ -642,7 +642,7 @@ rip_add_iface(struct rip_proto *p, struct iface *iface, struct rip_iface_config
 
   add_tail(&p->iface_list, NODE ifa);
 
-  ifa->timer = tm2_new_init(p->p.pool, rip_iface_timer, ifa, 0, 0);
+  ifa->timer = tm_new_init(p->p.pool, rip_iface_timer, ifa, 0, 0);
 
   struct object_lock *lock = olock_new(p->p.pool);
   lock->type = OBJLOCK_UDP;
@@ -897,14 +897,14 @@ rip_timer(timer *t)
 	  next = MIN(next, expires);
       }
 
-  tm2_start(p->timer, MAX(next - now_, 100 MS));
+  tm_start(p->timer, MAX(next - now_, 100 MS));
 }
 
 static inline void
 rip_kick_timer(struct rip_proto *p)
 {
   if (p->timer->expires > (current_time() + 100 MS))
-    tm2_start(p->timer, 100 MS);
+    tm_start(p->timer, 100 MS);
 }
 
 /**
@@ -933,7 +933,7 @@ rip_iface_timer(timer *t)
   if (ifa->tx_active)
   {
     if (now_ < (ifa->next_regular + period))
-    { tm2_start(ifa->timer, 100 MS); return; }
+    { tm_start(ifa->timer, 100 MS); return; }
 
     /* We are too late, reset is done by rip_send_table() */
     log(L_WARN "%s: Too slow update on %s, resetting", p->p.name, ifa->iface->name);
@@ -958,14 +958,14 @@ rip_iface_timer(timer *t)
     p->triggered = 0;
   }
 
-  tm2_start(ifa->timer, ifa->want_triggered ? (1 S) : (ifa->next_regular - now_));
+  tm_start(ifa->timer, ifa->want_triggered ? (1 S) : (ifa->next_regular - now_));
 }
 
 static inline void
 rip_iface_kick_timer(struct rip_iface *ifa)
 {
   if (ifa->timer->expires > (current_time() + 100 MS))
-    tm2_start(ifa->timer, 100 MS);
+    tm_start(ifa->timer, 100 MS);
 }
 
 static void
@@ -1111,7 +1111,7 @@ rip_start(struct proto *P)
   fib_init(&p->rtable, P->pool, cf->rip2 ? NET_IP4 : NET_IP6,
 	   sizeof(struct rip_entry), OFFSETOF(struct rip_entry, n), 0, NULL);
   p->rte_slab = sl_new(P->pool, sizeof(struct rip_rte));
-  p->timer = tm2_new_init(P->pool, rip_timer, p, 0, 0);
+  p->timer = tm_new_init(P->pool, rip_timer, p, 0, 0);
 
   p->rip2 = cf->rip2;
   p->ecmp = cf->ecmp;
@@ -1121,7 +1121,7 @@ rip_start(struct proto *P)
   p->log_pkt_tbf = (struct tbf){ .rate = 1, .burst = 5 };
   p->log_rte_tbf = (struct tbf){ .rate = 4, .burst = 20 };
 
-  tm2_start(p->timer, MIN(cf->min_timeout_time, cf->max_garbage_time));
+  tm_start(p->timer, MIN(cf->min_timeout_time, cf->max_garbage_time));
 
   return PS_UP;
 }
diff --git a/proto/rip/rip.h b/proto/rip/rip.h
index 2371ea31..55696333 100644
--- a/proto/rip/rip.h
+++ b/proto/rip/rip.h
@@ -24,7 +24,7 @@
 #include "lib/resource.h"
 #include "lib/socket.h"
 #include "lib/string.h"
-#include "sysdep/unix/timer.h"
+#include "lib/timer.h"
 
 
 #define RIP_V1			1
diff --git a/proto/rpki/rpki.c b/proto/rpki/rpki.c
index f17024fe..3145399b 100644
--- a/proto/rpki/rpki.c
+++ b/proto/rpki/rpki.c
@@ -321,7 +321,7 @@ rpki_schedule_next_refresh(struct rpki_cache *cache)
   btime t = cache->refresh_interval S;
 
   CACHE_DBG(cache, "after %t s", t);
-  tm2_start(cache->refresh_timer, t);
+  tm_start(cache->refresh_timer, t);
 }
 
 static void
@@ -330,7 +330,7 @@ rpki_schedule_next_retry(struct rpki_cache *cache)
   btime t = cache->retry_interval S;
 
   CACHE_DBG(cache, "after %t s", t);
-  tm2_start(cache->retry_timer, t);
+  tm_start(cache->retry_timer, t);
 }
 
 static void
@@ -341,28 +341,28 @@ rpki_schedule_next_expire_check(struct rpki_cache *cache)
   t = MAX(t, 1 S);
 
   CACHE_DBG(cache, "after %t s", t);
-  tm2_start(cache->expire_timer, t);
+  tm_start(cache->expire_timer, t);
 }
 
 static void
 rpki_stop_refresh_timer_event(struct rpki_cache *cache)
 {
   CACHE_DBG(cache, "Stop");
-  tm2_stop(cache->refresh_timer);
+  tm_stop(cache->refresh_timer);
 }
 
 static void
 rpki_stop_retry_timer_event(struct rpki_cache *cache)
 {
   CACHE_DBG(cache, "Stop");
-  tm2_stop(cache->retry_timer);
+  tm_stop(cache->retry_timer);
 }
 
 static void UNUSED
 rpki_stop_expire_timer_event(struct rpki_cache *cache)
 {
   CACHE_DBG(cache, "Stop");
-  tm2_stop(cache->expire_timer);
+  tm_stop(cache->expire_timer);
 }
 
 static int
@@ -569,9 +569,9 @@ rpki_init_cache(struct rpki_proto *p, struct rpki_config *cf)
   cache->refresh_interval = cf->refresh_interval;
   cache->retry_interval = cf->retry_interval;
   cache->expire_interval = cf->expire_interval;
-  cache->refresh_timer = tm2_new_init(pool, &rpki_refresh_hook, cache, 0, 0);
-  cache->retry_timer = tm2_new_init(pool, &rpki_retry_hook, cache, 0, 0);
-  cache->expire_timer = tm2_new_init(pool, &rpki_expire_hook, cache, 0, 0);
+  cache->refresh_timer = tm_new_init(pool, &rpki_refresh_hook, cache, 0, 0);
+  cache->retry_timer = tm_new_init(pool, &rpki_retry_hook, cache, 0, 0);
+  cache->expire_timer = tm_new_init(pool, &rpki_expire_hook, cache, 0, 0);
 
   cache->tr_sock = mb_allocz(pool, sizeof(struct rpki_tr_sock));
   cache->tr_sock->cache = cache;
@@ -791,8 +791,8 @@ rpki_get_status(struct proto *P, byte *buf)
 static void
 rpki_show_proto_info_timer(const char *name, uint num, timer *t)
 {
-  if (tm2_active(t))
-    cli_msg(-1006, "  %-16s: %t/%u", name, tm2_remains(t), num);
+  if (tm_active(t))
+    cli_msg(-1006, "  %-16s: %t/%u", name, tm_remains(t), num);
   else
     cli_msg(-1006, "  %-16s: ---", name);
 }
diff --git a/sysdep/linux/netlink.c b/sysdep/linux/netlink.c
index 6477d18c..6cea2fa9 100644
--- a/sysdep/linux/netlink.c
+++ b/sysdep/linux/netlink.c
@@ -21,7 +21,6 @@
 #include "nest/protocol.h"
 #include "nest/iface.h"
 #include "lib/alloca.h"
-#include "sysdep/unix/timer.h"
 #include "sysdep/unix/unix.h"
 #include "sysdep/unix/krt.h"
 #include "lib/socket.h"
diff --git a/sysdep/unix/io.c b/sysdep/unix/io.c
index a196bbe2..7492e031 100644
--- a/sysdep/unix/io.c
+++ b/sysdep/unix/io.c
@@ -34,7 +34,6 @@
 #include "nest/bird.h"
 #include "lib/lists.h"
 #include "lib/resource.h"
-#include "sysdep/unix/timer.h"
 #include "lib/socket.h"
 #include "lib/event.h"
 #include "lib/timer.h"
@@ -2143,7 +2142,7 @@ io_loop(void)
 {
   int poll_tout, timeout;
   int nfds, events, pout;
-  timer2 *t;
+  timer *t;
   sock *s;
   node *n;
   int fdmax = 256;
@@ -2162,7 +2161,7 @@ io_loop(void)
       if (t = timers_first(&main_timeloop))
       {
 	times_update(&main_timeloop);
-	timeout = (tm2_remains(t) TO_MS) + 1;
+	timeout = (tm_remains(t) TO_MS) + 1;
 	poll_tout = MIN(poll_tout, timeout);
       }
 
diff --git a/sysdep/unix/krt.c b/sysdep/unix/krt.c
index 5b1c40f0..052f210a 100644
--- a/sysdep/unix/krt.c
+++ b/sysdep/unix/krt.c
@@ -56,9 +56,9 @@
 #include "nest/route.h"
 #include "nest/protocol.h"
 #include "filter/filter.h"
-#include "sysdep/unix/timer.h"
 #include "conf/conf.h"
 #include "lib/string.h"
+#include "lib/timer.h"
 
 #include "unix.h"
 #include "krt.h"
@@ -115,7 +115,7 @@ kif_force_scan(void)
   if (kif_proto && ((kif_last_shot + 2 S) < current_time()))
     {
       kif_scan(kif_scan_timer);
-      tm2_start(kif_scan_timer, ((struct kif_config *) kif_proto->p.cf)->scan_time);
+      tm_start(kif_scan_timer, ((struct kif_config *) kif_proto->p.cf)->scan_time);
     }
 }
 
@@ -123,7 +123,7 @@ void
 kif_request_scan(void)
 {
   if (kif_proto && (kif_scan_timer->expires > (current_time() + 1 S)))
-    tm2_start(kif_scan_timer, 1 S);
+    tm_start(kif_scan_timer, 1 S);
 }
 
 static struct proto *
@@ -144,9 +144,9 @@ kif_start(struct proto *P)
   kif_sys_start(p);
 
   /* Start periodic interface scanning */
-  kif_scan_timer = tm2_new_init(P->pool, kif_scan, p, KIF_CF->scan_time, 0);
+  kif_scan_timer = tm_new_init(P->pool, kif_scan, p, KIF_CF->scan_time, 0);
   kif_scan(kif_scan_timer);
-  tm2_start(kif_scan_timer, KIF_CF->scan_time);
+  tm_start(kif_scan_timer, KIF_CF->scan_time);
 
   return PS_UP;
 }
@@ -156,7 +156,7 @@ kif_shutdown(struct proto *P)
 {
   struct kif_proto *p = (struct kif_proto *) P;
 
-  tm2_stop(kif_scan_timer);
+  tm_stop(kif_scan_timer);
   kif_sys_shutdown(p);
   kif_proto = NULL;
 
@@ -174,10 +174,10 @@ kif_reconfigure(struct proto *p, struct proto_config *new)
 
   if (o->scan_time != n->scan_time)
     {
-      tm2_stop(kif_scan_timer);
+      tm_stop(kif_scan_timer);
       kif_scan_timer->recurrent = n->scan_time;
       kif_scan(kif_scan_timer);
-      tm2_start(kif_scan_timer, n->scan_time);
+      tm_start(kif_scan_timer, n->scan_time);
     }
 
   if (!EMPTY_LIST(o->iface_list) || !EMPTY_LIST(n->iface_list))
@@ -840,11 +840,11 @@ static void
 krt_scan_timer_start(struct krt_proto *p)
 {
   if (!krt_scan_count)
-    krt_scan_timer = tm2_new_init(krt_pool, krt_scan, NULL, KRT_CF->scan_time, 0);
+    krt_scan_timer = tm_new_init(krt_pool, krt_scan, NULL, KRT_CF->scan_time, 0);
 
   krt_scan_count++;
 
-  tm2_start(krt_scan_timer, 1 S);
+  tm_start(krt_scan_timer, 1 S);
 }
 
 static void
@@ -862,7 +862,7 @@ krt_scan_timer_stop(struct krt_proto *p UNUSED)
 static void
 krt_scan_timer_kick(struct krt_proto *p UNUSED)
 {
-  tm2_start(krt_scan_timer, 0);
+  tm_start(krt_scan_timer, 0);
 }
 
 #else
@@ -882,20 +882,20 @@ krt_scan(timer *t)
 static void
 krt_scan_timer_start(struct krt_proto *p)
 {
-  p->scan_timer = tm2_new_init(p->p.pool, krt_scan, p, KRT_CF->scan_time, 0);
-  tm2_start(p->scan_timer, 1 S);
+  p->scan_timer = tm_new_init(p->p.pool, krt_scan, p, KRT_CF->scan_time, 0);
+  tm_start(p->scan_timer, 1 S);
 }
 
 static void
 krt_scan_timer_stop(struct krt_proto *p)
 {
-  tm2_stop(p->scan_timer);
+  tm_stop(p->scan_timer);
 }
 
 static void
 krt_scan_timer_kick(struct krt_proto *p)
 {
-  tm2_start(p->scan_timer, 0);
+  tm_start(p->scan_timer, 0);
 }
 
 #endif
diff --git a/sysdep/unix/timer.h b/sysdep/unix/timer.h
deleted file mode 100644
index 989574bf..00000000
--- a/sysdep/unix/timer.h
+++ /dev/null
@@ -1,51 +0,0 @@
-/*
- *	BIRD -- Unix Timers
- *
- *	(c) 1998 Martin Mares <mj@ucw.cz>
- *
- *	Can be freely distributed and used under the terms of the GNU GPL.
- */
-
-#ifndef _BIRD_TIMER_H_
-#define _BIRD_TIMER_H_
-
-#include <time.h>
-
-#include "lib/birdlib.h"
-#include "lib/timer.h"
-
-
-typedef struct timer2 timer;
-#if 0
-static inline timer *tm_new(pool *p)
-{ return (void *) tm2_new(p); }
-
-static inline void tm_start(timer *t, bird_clock_t after)
-{ tm2_start(t, after S_); }
-
-static inline void tm_stop(timer *t)
-{ tm2_stop(t); }
-
-// void tm_dump_all(void);
-
-//extern bird_clock_t now; 		/* Relative, monotonic time in seconds */
-//extern bird_clock_t now_real;		/* Time in seconds since fixed known epoch */
-//extern bird_clock_t boot_time;
-
-static inline int tm_active(timer *t)
-{ return tm2_active(t); }
-
-static inline bird_clock_t tm_remains(timer *t)
-{ return tm2_remains(t) TO_S; }
-
-static inline void tm_start_max(timer *t, bird_clock_t after)
-{ tm2_start_max(t, after S_); }
-
-static inline timer * tm_new_set(pool *p, void (*hook)(timer *), void *data, uint rand, uint rec)
-{ return tm2_new_init(p, hook, data, rec S_, rand S_); }
-
-#endif
-
-
-
-#endif

From cd80c9b0700ef9533a19ab19d66c38a1722c4e08 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Thu, 7 Dec 2017 14:21:38 +0100
Subject: [PATCH 072/198] BSD: Fix in the last commit

---
 sysdep/bsd/krt-sock.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/sysdep/bsd/krt-sock.c b/sysdep/bsd/krt-sock.c
index efbfdd5e..08cfea39 100644
--- a/sysdep/bsd/krt-sock.c
+++ b/sysdep/bsd/krt-sock.c
@@ -28,7 +28,6 @@
 #include "nest/route.h"
 #include "nest/protocol.h"
 #include "nest/iface.h"
-#include "sysdep/unix/timer.h"
 #include "sysdep/unix/unix.h"
 #include "sysdep/unix/krt.h"
 #include "lib/string.h"

From ccee67ca3b607130bf441b6060b88525b5e50ad9 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Fri, 8 Dec 2017 02:26:17 +0100
Subject: [PATCH 073/198] BGP: Autoconfigure BGP next hops from preferred
 addresses

---
 proto/bgp/bgp.c | 19 +++++++++++++++++++
 1 file changed, 19 insertions(+)

diff --git a/proto/bgp/bgp.c b/proto/bgp/bgp.c
index 30fe75ba..a26c0cb9 100644
--- a/proto/bgp/bgp.c
+++ b/proto/bgp/bgp.c
@@ -1504,6 +1504,20 @@ bgp_channel_start(struct channel *C)
       c->next_hop_addr = src;
   }
 
+  /* Use preferred addresses associated with interface / source address */
+  if (ipa_zero(c->next_hop_addr))
+  {
+    /* We know the iface for single-hop, we make lookup for multihop */
+    struct neighbor *nbr = p->neigh ?: neigh_find2(&p->p, &src, NULL, 0);
+    struct iface *iface = nbr ? nbr->iface : NULL;
+
+    if (bgp_channel_is_ipv4(c) && iface && iface->addr4)
+      c->next_hop_addr = iface->addr4->ip;
+
+    if (bgp_channel_is_ipv6(c) && iface && iface->addr6)
+      c->next_hop_addr = iface->addr6->ip;
+  }
+
   /* Exit if no feasible next hop address is found */
   if (ipa_zero(c->next_hop_addr))
   {
@@ -2079,6 +2093,11 @@ bgp_show_proto_info(struct proto *P)
     {
       channel_show_info(&c->c);
 
+      if (ipa_zero(c->link_addr))
+	cli_msg(-1006, "    BGP Next hop:   %I", c->next_hop_addr);
+      else
+	cli_msg(-1006, "    BGP Next hop:   %I %I", c->next_hop_addr, c->link_addr);
+
       if (c->igp_table_ip4)
 	cli_msg(-1006, "    IGP IPv4 table: %s", c->igp_table_ip4->name);
 

From 49c7ef3b21e51ae7d1969baa52b4d8fd29b22eeb Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Fri, 8 Dec 2017 14:00:36 +0100
Subject: [PATCH 074/198] BGP: Fix IPv6 MPLS/VPN multicast SAFI

---
 proto/bgp/packets.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/proto/bgp/packets.c b/proto/bgp/packets.c
index 038e89f9..aa69bdd6 100644
--- a/proto/bgp/packets.c
+++ b/proto/bgp/packets.c
@@ -1483,7 +1483,8 @@ bgp_encode_nlri_vpn6(struct bgp_write_state *s, struct bgp_bucket *buck, byte *b
     ADVANCE(pos, size, 1);
 
     /* Encode MPLS labels */
-    bgp_encode_mpls_labels(s, s->mpls_labels, &pos, &size, pos - 1);
+    if (s->mpls)
+      bgp_encode_mpls_labels(s, s->mpls_labels, &pos, &size, pos - 1);
 
     /* Encode route distinguisher */
     put_u64(pos, net->rd);

From eb95b5ec1a5c3bd4e57f3a134828f8de2875cb43 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Fri, 8 Dec 2017 15:16:47 +0100
Subject: [PATCH 075/198] Nest: Minor formatting changes

---
 nest/iface.c   |  2 +-
 nest/proto.c   |  5 +++--
 nest/rt-show.c | 14 +++++++++-----
 3 files changed, 13 insertions(+), 8 deletions(-)

diff --git a/nest/iface.c b/nest/iface.c
index 5e0a2450..a876e9ff 100644
--- a/nest/iface.c
+++ b/nest/iface.c
@@ -880,7 +880,7 @@ if_show_summary(void)
 {
   struct iface *i;
 
-  cli_msg(-2005,  "%-10s %-6s %-18s %s", "Interface", "State", "IPv4 address", "IPv6 address");
+  cli_msg(-2005, "%-10s %-6s %-18s %s", "Interface", "State", "IPv4 address", "IPv6 address");
   WALK_LIST(i, iface_list)
     {
       byte a4[IPA_MAX_TEXT_LENGTH + 17];
diff --git a/nest/proto.c b/nest/proto.c
index a2a2bc7e..d7650517 100644
--- a/nest/proto.c
+++ b/nest/proto.c
@@ -1662,13 +1662,14 @@ proto_cmd_show(struct proto *p, uintptr_t verbose, int cnt)
 
   /* First protocol - show header */
   if (!cnt)
-    cli_msg(-2002, "name     proto    table    state  since       info");
+    cli_msg(-2002, "%-10s %-10s %-10s %-6s %-12s  %s",
+	    "Name", "Proto", "Table", "State", "Since", "Info");
 
   buf[0] = 0;
   if (p->proto->get_status)
     p->proto->get_status(p, buf);
   tm_format_time(tbuf, &config->tf_proto, p->last_state_change);
-  cli_msg(-1002, "%-8s %-8s %-8s %-5s  %-10s  %s",
+  cli_msg(-1002, "%-10s %-10s %-10s %-6s %-12s  %s",
 	  p->name,
 	  p->proto->name,
 	  p->main_channel ? p->main_channel->table->name : "---",
diff --git a/nest/rt-show.c b/nest/rt-show.c
index 9989afa4..41a141a2 100644
--- a/nest/rt-show.c
+++ b/nest/rt-show.c
@@ -63,7 +63,7 @@ rt_show_rte(struct cli *c, byte *ia, rte *e, struct rt_show_data *d, ea_list *tm
   if (d->last_table != d->tab)
     rt_show_table(c, d);
 
-  cli_printf(c, -1007, "%-18s %s [%s %s%s]%s%s", ia, rta_dest_name(a->dest),
+  cli_printf(c, -1007, "%-20s %s [%s %s%s]%s%s", ia, rta_dest_name(a->dest),
 	     a->src->proto->name, tm, from, primary ? (sync_error ? " !" : " *") : "", info);
 
   if (a->dest == RTD_UNICAST)
@@ -71,6 +71,7 @@ rt_show_rte(struct cli *c, byte *ia, rte *e, struct rt_show_data *d, ea_list *tm
     {
       char mpls[MPLS_MAX_LABEL_STACK*12 + 5], *lsp = mpls;
       char *onlink = (nh->flags & RNF_ONLINK) ? " onlink" : "";
+      char weight[16] = "";
 
       if (nh->labels)
         {
@@ -81,11 +82,14 @@ rt_show_rte(struct cli *c, byte *ia, rte *e, struct rt_show_data *d, ea_list *tm
       *lsp = '\0';
 
       if (a->nh.next)
-	cli_printf(c, -1007, "\tvia %I%s on %s%s weight %d",
-		   nh->gw, mpls, nh->iface->name, onlink, nh->weight + 1);
+	bsprintf(weight, " weight %d", nh->weight + 1);
+
+      if (ipa_nonzero(nh->gw))
+	cli_printf(c, -1007, "\tvia %I on %s%s%s%s",
+		   nh->gw, nh->iface->name, mpls, onlink, weight);
       else
-	cli_printf(c, -1007, "\tvia %I%s on %s%s",
-		   nh->gw, mpls, nh->iface->name, onlink);
+	cli_printf(c, -1007, "\tdev %s%s%s",
+		   nh->iface->name, mpls,  onlink, weight);
     }
 
   if (d->verbose)

From 517d05dff17e881b880ee4fd28a72e827c10e8c3 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Fri, 8 Dec 2017 15:59:44 +0100
Subject: [PATCH 076/198] Enable ECMP and Link detection by default

ECMP is not enabled on BSD, where it is not supported by BIRD.
---
 nest/route.h           | 2 ++
 proto/bgp/config.Y     | 1 +
 proto/ospf/config.Y    | 2 ++
 proto/rip/config.Y     | 2 ++
 sysdep/bsd/krt-sock.c  | 4 ++--
 sysdep/linux/netlink.c | 2 ++
 6 files changed, 11 insertions(+), 2 deletions(-)

diff --git a/nest/route.h b/nest/route.h
index bb0cb4a4..792f79dd 100644
--- a/nest/route.h
+++ b/nest/route.h
@@ -309,6 +309,8 @@ int rt_feed_channel(struct channel *c);
 void rt_feed_channel_abort(struct channel *c);
 struct rtable_config *rt_new_table(struct symbol *s, uint addr_type);
 
+/* Default limit for ECMP next hops, defined in sysdep code */
+extern const int rt_default_ecmp;
 
 struct rt_show_data_rtable {
   node n;
diff --git a/proto/bgp/config.Y b/proto/bgp/config.Y
index 941ae5b6..0bd4c554 100644
--- a/proto/bgp/config.Y
+++ b/proto/bgp/config.Y
@@ -58,6 +58,7 @@ bgp_proto_start: proto_start BGP {
      BGP_CFG->gr_mode = BGP_GR_AWARE;
      BGP_CFG->gr_time = 120;
      BGP_CFG->setkey = 1;
+     BGP_CFG->check_link = 1;
    }
  ;
 
diff --git a/proto/ospf/config.Y b/proto/ospf/config.Y
index 98ddf5d0..04d5e2e9 100644
--- a/proto/ospf/config.Y
+++ b/proto/ospf/config.Y
@@ -219,6 +219,7 @@ ospf_proto_start: proto_start ospf_variant
 
   init_list(&OSPF_CFG->area_list);
   init_list(&OSPF_CFG->vlink_list);
+  OSPF_CFG->ecmp = rt_default_ecmp;
   OSPF_CFG->tick = OSPF_DEFAULT_TICK;
   OSPF_CFG->ospf2 = $2;
   OSPF_CFG->af_ext = !$2;
@@ -461,6 +462,7 @@ ospf_iface_start:
   OSPF_PATT->deadc = DEADC_D;
   OSPF_PATT->type = OSPF_IT_UNDEF;
   init_list(&OSPF_PATT->nbma_list);
+  OSPF_PATT->check_link = 1;
   OSPF_PATT->ptp_netmask = 2; /* not specified */
   OSPF_PATT->tx_tos = IP_PREC_INTERNET_CONTROL;
   OSPF_PATT->tx_priority = sk_priority_control;
diff --git a/proto/rip/config.Y b/proto/rip/config.Y
index e3bc4ae3..aff63f03 100644
--- a/proto/rip/config.Y
+++ b/proto/rip/config.Y
@@ -56,6 +56,7 @@ rip_proto_start: proto_start rip_variant
 
   init_list(&RIP_CFG->patt_list);
   RIP_CFG->rip2 = $2;
+  RIP_CFG->ecmp = rt_default_ecmp;
   RIP_CFG->infinity = RIP_DEFAULT_INFINITY;
   RIP_CFG->min_timeout_time = 60 S_;
   RIP_CFG->max_garbage_time = 60 S_;
@@ -92,6 +93,7 @@ rip_iface_start:
   RIP_IFACE->split_horizon = 1;
   RIP_IFACE->poison_reverse = 1;
   RIP_IFACE->check_zero = 1;
+  RIP_IFACE->check_link = 1;
   RIP_IFACE->ttl_security = rip_cfg_is_v2() ? 0 : 1;
   RIP_IFACE->rx_buffer = rip_cfg_is_v2() ? RIP_MAX_PKT_LENGTH : 0;
   RIP_IFACE->tx_length = rip_cfg_is_v2() ? RIP_MAX_PKT_LENGTH : 0;
diff --git a/sysdep/bsd/krt-sock.c b/sysdep/bsd/krt-sock.c
index c36ae558..604cd510 100644
--- a/sysdep/bsd/krt-sock.c
+++ b/sysdep/bsd/krt-sock.c
@@ -33,6 +33,7 @@
 #include "lib/string.h"
 #include "lib/socket.h"
 
+const int rt_default_ecmp = 0;
 
 /*
  * There are significant differences in multiple tables support between BSD variants.
@@ -145,9 +146,8 @@ krt_capable(rte *e)
 {
   rta *a = e->attrs;
 
-  /* XXXX device routes are broken */
   return
-    ((a->dest == RTD_UNICAST && ipa_nonzero(a->nh.gw) && !a->nh.next) /* No multipath support */
+    ((a->dest == RTD_UNICAST && !a->nh.next) /* No multipath support */
 #ifdef RTF_REJECT
      || a->dest == RTD_UNREACHABLE
 #endif
diff --git a/sysdep/linux/netlink.c b/sysdep/linux/netlink.c
index 05c1fa8c..ca5c79a4 100644
--- a/sysdep/linux/netlink.c
+++ b/sysdep/linux/netlink.c
@@ -80,6 +80,8 @@ struct rtvia {
 
 #define krt_ecmp6(p) ((p)->af == AF_INET6)
 
+const int rt_default_ecmp = 16;
+
 /*
  * Structure nl_parse_state keeps state of received route processing. Ideally,
  * we could just independently parse received Netlink messages and immediately

From 3e52d112d714545df3cd97119824ee94c27a0606 Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Wed, 27 Sep 2017 16:55:09 +0200
Subject: [PATCH 077/198] Docs: Update to v2.0

---
 conf/conf.c     |    5 -
 conf/conf.h     |    3 -
 doc/bird.sgml   | 1049 ++++++++++++++++++++++++++++-------------------
 filter/config.Y |   23 +-
 nest/config.Y   |   20 +-
 5 files changed, 636 insertions(+), 464 deletions(-)

diff --git a/conf/conf.c b/conf/conf.c
index 62477331..885e2e7e 100644
--- a/conf/conf.c
+++ b/conf/conf.c
@@ -219,11 +219,6 @@ global_commit(struct config *new, struct config *old)
   if (!old)
     return 0;
 
-  if (!ipa_equal(old->listen_bgp_addr, new->listen_bgp_addr) ||
-      (old->listen_bgp_port != new->listen_bgp_port) ||
-      (old->listen_bgp_flags != new->listen_bgp_flags))
-    log(L_WARN "Reconfiguration of BGP listening socket not implemented, please restart BIRD.");
-
   if (!new->router_id)
     {
       new->router_id = old->router_id;
diff --git a/conf/conf.h b/conf/conf.h
index d6f59eac..f174d352 100644
--- a/conf/conf.h
+++ b/conf/conf.h
@@ -32,9 +32,6 @@ struct config {
   struct iface_patt *router_id_from;	/* Configured list of router ID iface patterns */
 
   u32 router_id;			/* Our Router ID */
-  ip_addr listen_bgp_addr;		/* Listening BGP socket should use this address */
-  unsigned listen_bgp_port;		/* Listening BGP socket should use this port (0 is default) */
-  u32 listen_bgp_flags;			/* Listening BGP socket should use these flags */
   unsigned proto_default_debug;		/* Default protocol debug mask */
   unsigned proto_default_mrtdump;	/* Default protocol mrtdump mask */
   struct timeformat tf_route;		/* Time format for 'show route' */
diff --git a/doc/bird.sgml b/doc/bird.sgml
index ab515191..c50d0cde 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -1,7 +1,7 @@
 <!doctype birddoc system>
 
 <!--
-	BIRD documentation
+	BIRD 2.0 documentation
 
 This documentation can have 4 forms: sgml (this is master copy), html, ASCII
 text and dvi/postscript (generated from sgml using sgmltools). You should always
@@ -20,11 +20,12 @@ configuration - something in config which is not keyword.
 
 <book>
 
-<title>BIRD User's Guide
+<title>BIRD 2.0 User's Guide
 <author>
 Ondrej Filip <it/&lt;feela@network.cz&gt;/,
 Pavel Machek <it/&lt;pavel@ucw.cz&gt;/,
 Martin Mares <it/&lt;mj@ucw.cz&gt;/,
+Jan Matejka <it/&lt;mq@jmq.cz&gt;/,
 Ondrej Zajicek <it/&lt;santiago@crfreenet.org&gt;/
 </author>
 
@@ -88,7 +89,7 @@ supports:
 	<item>both IPv4 and IPv6 protocols
 	<item>multiple routing tables
 	<item>the Border Gateway Protocol (BGPv4)
-	<item>the Routing Information Protocol (RIPv2)
+	<item>the Routing Information Protocol (RIPv2, RIPng)
 	<item>the Open Shortest Path First protocol (OSPFv2, OSPFv3)
 	<item>the Router Advertisements for IPv6 hosts
 	<item>a virtual protocol for exchange of routes between different
@@ -112,11 +113,10 @@ developed and tested under Linux 2.0 to 2.6, and then ported to FreeBSD, NetBSD
 and OpenBSD, porting to other systems (even non-UNIX ones) should be relatively
 easy due to its highly modular architecture.
 
-<p>BIRD supports either IPv4 or IPv6 protocol, but have to be compiled separately
-for each one. Therefore, a dualstack router would run two instances of BIRD (one
-for IPv4 and one for IPv6), with completely separate setups (configuration
-files, tools ...).
-
+<p>BIRD 1.x supported either IPv4 or IPv6 protocol, but had to be compiled separately
+for each one. BIRD~2 supports both of them with a possibility of further extension.
+BIRD~2 supports Linux at least 3.16, FreeBSD 10, NetBSD 7.0, and OpenBSD 5.8.
+Anyway, it will probably work well also on older systems.
 
 <sect>Installing BIRD
 <label id="install">
@@ -133,8 +133,7 @@ and Perl, installing BIRD should be as easy as:
 </code>
 
 <p>You can use <tt>./configure --help</tt> to get a list of configure
-options. The most important ones are: <tt/--enable-ipv6/ which enables building
-of an IPv6 version of BIRD, <tt/--with-protocols=/ to produce a slightly smaller
+options. The most important ones are: <tt/--with-protocols=/ to produce a slightly smaller
 BIRD executable by configuring out routing protocols you don't use, and
 <tt/--prefix=/ to install BIRD to a place different from <file>/usr/local</file>.
 
@@ -217,45 +216,40 @@ is generally easy -- BIRD needs just the standard library, privileges to read
 the config file and create the control socket and the CAP_NET_* capabilities.
 
 
-<chapt>About routing tables
+<chapt>Architecture
+<label id="architecture">
+
+<sect>Routing tables
 <label id="routing-tables">
 
-<p>BIRD has one or more routing tables which may or may not be synchronized with
-OS kernel and which may or may not be synchronized with each other (see the Pipe
-protocol). Each routing table contains a list of known routes. Each route
-consists of:
+<p>The heart of BIRD is a routing table. BIRD has several independent routing tables;
+each of them contains routes of exactly one <m/nettype/ (see below). There are two
+default tables -- <cf/master4/ for IPv4 routes and <cf/master6/ for IPv6 routes.
+Other tables must be explicitly configured.
+
+<p>
+These routing tables are not kernel forwarding tables. No forwarding is done in BIRD.
+If you want to forward packets using the routes in BIRD's tables, you may
+use the Kernel protocol (see below) to synchronize them with kernel's FIBs.
+
+<p>
+Every nettype defines a (kind of) primary key on routes. Every route source
+can supply one route for every possible primary key; new route announcement
+replaces the old route from the same source, keeping other routes intact.
+BIRD always chooses the best route for every primary key among the known
+routes and keeps the others as suboptimal. When the best route is retracted,
+BIRD re-runs the best route selection algorithm to find the current best route.
+
+<p>
+The global best route selection algorithm is (roughly) as follows:
 
 <itemize>
-	<item>network prefix this route is for (network address and prefix
-		length -- the number of bits forming the network part of the
-		address; also known as a netmask)
-	<item>preference of this route
-	<item>IP address of router which told us about this route
-	<item>IP address of router we should forward the packets to using this
-		route
-	<item>other attributes common to all routes
-	<item>dynamic attributes defined by protocols which may or may not be
-		present (typically protocol metrics)
+	<item>Preferences of the routes are compared.
+	<item>Source protocol instance preferences are compared.
+	<item>If source protocols are different (e.g. BGP vs. OSPF), result of the algorithm is undefined.
+	<item>If source protocols are the same (e.g. BGP vs. BGP), the protocol's route selection algorithm is invoked.
 </itemize>
 
-Routing table maintains multiple entries for a network, but at most one entry
-for one network and one protocol. The entry with the highest preference is used
-for routing (we will call such an entry the <it/selected route/). If there are
-more entries with the same preference and they are from the same protocol, the
-protocol decides (typically according to metrics). If they aren't, an internal
-ordering is used to break the tie. You can get the list of route attributes in
-the Route attributes section.
-
-<p>Each protocol is connected to a routing table through two filters which can
-accept, reject and modify the routes. An <it/export/ filter checks routes passed
-from the routing table to the protocol, an <it/import/ filter checks routes in
-the opposite direction. When the routing table gets a route from a protocol, it
-recalculates the selected route and broadcasts it to all protocols connected to
-the table. The protocols typically send the update to other routers in the
-network. Note that although most protocols are interested in receiving just
-selected routes, some protocols (e.g. the <cf/Pipe/ protocol) receive and
-process all entries in routing tables (accepted by filters).
-
 <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
 activated, these lists of entries are kept completely sorted (according to
@@ -268,6 +262,114 @@ 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">
+
+<p>BIRD works with several types of routes. Some of them are typical IP routes,
+others are better described as forwarding rules. We call them all routes,
+regardless on this difference.
+
+<p>Every route consists of several attributes (read more about them in the <ref id="route-attributes" name="Route attributes"> section); the common for all routes are:
+
+<itemize>
+	<item>IP address of router which told us about this route
+	<item>Source protocol instance
+	<item>Route preference
+	<item>Optional attributes defined by protocols
+</itemize>
+
+<p>Other attributes depend on nettypes. Some of them are part of the primary key, these are marked (PK).
+
+<sect1>IPv4 and IPv6 routes
+<label id="ipv4-ipv6-routes">
+
+<p>The traditional routes. Configuration keywords are <cf/ipv4/ and <cf/ipv6/.
+
+<itemize>
+	<item>(PK) Route destination (IP prefix together with its length)
+	<item>Route next hops (see below)
+</itemize>
+
+<sect1>VPN IPv4 and IPv6 routes
+<label id="vpn4-vpn6-routes">
+
+<p>Routes for IPv4 and IPv6 with VPN Route Distinguisher (<rfc id="4364">).
+Configuration keywords are <cf/vpn4/ and <cf/vpn6/.
+
+<itemize>
+	<item>(PK) Route destination (IP prefix together with its length)
+	<item>(PK) Route distinguisher (according to <rfc id="4364">)
+	<item>Route next hops
+</itemize>
+
+<sect1>Route Origin Authorization for IPv4 and IPv6
+<label id="roa4-roa6-routes">
+
+<p>These entries can be used to validate route origination of BGP routes.
+An ROA entry specifies prefixes which could be originated by an AS number.
+Their keywords are <cf/roa4/ and <cf/roa6/.
+
+<itemize>
+	<item>(PK) IP prefix together with its length
+	<item>(PK) Matching prefix maximal length
+	<item>(PK) AS number
+</itemize>
+
+<sect1>Flowspec for IPv4 and IPv6
+<label id="flow4-flow6-routes">
+
+<p>Flowspec rules are a form of firewall and traffic flow control rules
+distributed mostly via BGP. These rules may help the operators stop various
+network attacks in the beginning before eating up the whole bandwidth.
+Configuration keywords are <cf/flow4/ and <cf/flow6/.
+
+<itemize>
+	<item>(PK) IP prefix together with its length
+	<item>(PK) Flow definition data
+	<item>Flow action (encoded internally as BGP communities according to <rfc id="5575">)
+</itemize>
+
+<sect1>MPLS switching rules
+<label id="mpls-routes">
+
+<p>This nettype is currently a stub before implementing more support of <rfc id="3031">.
+BIRD currently does not support any label distribution protocol nor any label assignment method.
+Only the Kernel, Pipe and Static protocols can use MPLS tables.
+Configuration keyword is <cf/mpls/.
+
+<itemize>
+	<item>(PK) MPLS label
+	<item>Route next hops
+</itemize>
+
+<sect1>Route next hops
+<label id="route-next-hop">
+
+<p>This is not a nettype. The route next hop is a complex attribute common for many
+nettypes as you can see before. Every next hop has its assigned device
+(either assumed from its IP address or set explicitly). It may have also
+an IP address and an MPLS stack (one or both independently).
+Maximal MPLS stack depth is set (in compile time) to 8 labels.
+
+<p>Every route (when eligible to have a next hop) can have more than one next hop.
+In that case, every next hop has also its weight.
+
+<sect>Protocols and channels
+<label id="protocols-concept">
+
+<p>BIRD's protocol is an abstract class of producers and consumers of the routes.
+Each protocol may run in multiple instances and bind on one side to route
+tables via channels, on the other side to specified listen sockets (BGP),
+interfaces (Babel, OSPF, RIP), APIs (Kernel, Direct), or nothing (Static, Pipe).
+
+<p>There are also two protocols which do not have any channels -- BFD and Device.
+Both of them are kind of service for other protocols.
+
+<p>Each protocol is connected to a routing table through a channel. Some protocols
+support only one channel (OSPF, RIP), some protocols support more channels (BGP, Direct).
+Each channel has two filters which can accept, reject and modify the routes.
+An <it/export/ filter is applied to routes passed from the routing table to the protocol,
+an <it/import/ filter is applied to routes in the opposite direction.
 
 <sect>Graceful restart
 <label id="graceful-restart">
@@ -304,10 +406,10 @@ variable number of options, they are grouped using the <cf/{ }/ brackets. Each
 option is terminated by a <cf/;/. Configuration is case sensitive. There are two
 ways how to name symbols (like protocol names, filter names, constants etc.). You
 can either use a simple string starting with a letter followed by any
-combination of letters and numbers (e.g. "R123", "myfilter", "bgp5") or you can
+combination of letters and numbers (e.g. <cf/R123/, <cf/myfilter/, <cf/bgp5/) or you can
 enclose the name into apostrophes (<cf/'/) and than you can use any combination
-of numbers, letters. hyphens, dots and colons (e.g. "'1:strange-name'",
-"'-NAME-'", "'cool::name'").
+of numbers, letters. hyphens, dots and colons (e.g. <cf/'1:strange-name'/,
+<cf/'-NAME-'/, <cf/'cool::name'/).
 
 <p>Here is an example of a simple config file. It enables synchronization of
 routing tables with OS kernel, scans for new network interfaces every 10 seconds
@@ -336,11 +438,17 @@ protocol rip {
 <label id="global-opts">
 
 <p><descrip>
-	<tag><label id="opt-include">include "<m/filename/"</tag>
-	This statement causes inclusion of a new file. <m/Filename/ could also
+	<tag><label id="opt-include">include "<m/filename/";</tag>
+	This statement causes inclusion of a new file. The <m/filename/ could also
 	be a wildcard, in that case matching files are included in alphabetic
-	order. The maximal depth is 8. Note that this statement could be used
-	anywhere in the config file, not just as a top-level option.
+	order. The maximal depth is 8. Note that this statement can be used
+	anywhere in the config file, even inside other options, but always on the beginning of line.
+	In the following example, the first semicolon belongs to the <cf/include/, the second to <cf/ipv6 table/.
+	If the <file/tablename.conf/ contains exactly one token (the name of the table), this construction is correct:
+<code>
+ipv6 table
+include "tablename.conf";;
+</code>
 
 	<tag><label id="opt-log">log "<m/filename/"|syslog [name <m/name/]|stderr all|{ <m/list of classes/ }</tag>
 	Set logging of messages having the given class (either <cf/all/ or
@@ -397,7 +505,7 @@ protocol rip {
 	<tag><label id="opt-function">function <m/name/ (<m/parameters/) <m/local variables/ { <m/commands/ }</tag>
 	Define a function. You can learn more about functions in the following chapter.
 
-	<tag><label id="opt-protocol">protocol rip|ospf|bgp|<m/.../ [<m/name/ [from <m/name2/]] { <m>protocol options</m> }</tag>
+	<tag><label id="opt-protocol">protocol rip [|ng]|ospf [v2|v3]|bgp|<m/.../ [<m/name/ [from <m/name2/]] { <m>protocol options</m> }</tag>
 	Define a protocol instance called <cf><m/name/</cf> (or with a name like
 	"rip5" generated automatically if you don't specify any
 	<cf><m/name/</cf>). You can learn more about configuring protocols in
@@ -406,7 +514,7 @@ protocol rip {
 	<cf><m/name2/</cf> You can run more than one instance of most protocols
 	(like RIP or BGP). By default, no instances are configured.
 
-	<tag><label id="opt-template">template rip|bgp|<m/.../ [<m/name/ [from <m/name2/]] { <m>protocol options</m> }</tag>
+	<tag><label id="opt-template">template rip [|ng]|ospf [v2|v3]|bgp|<m/.../ [<m/name/ [from <m/name2/]] { <m>protocol options</m> }</tag>
 	Define a protocol template instance called <m/name/ (or with a name like
 	"bgp1" generated automatically if you don't specify any	<m/name/).
 	Protocol templates can be used to group common options when many
@@ -423,25 +531,15 @@ protocol rip {
 
 	<tag><label id="opt-router-id">router id <m/IPv4 address/</tag>
 	Set BIRD's router ID. It's a world-wide unique identification of your
-	router, usually one of router's IPv4 addresses. Default: in IPv4
-	version, the lowest IP address of a non-loopback interface. In IPv6
-	version, this option is mandatory.
+	router, usually one of router's IPv4 addresses. Default: the lowest
+	IPv4 address of a non-loopback interface.
 
 	<tag><label id="opt-router-id-from">router id from [-] [ "<m/mask/" ] [ <m/prefix/ ] [, <m/.../]</tag>
-	Set BIRD's router ID based on an IP address of an interface specified by
-	an interface pattern. The option is applicable for IPv4 version only.
+	Set BIRD's router ID based on an IPv4 address of an interface specified by
+	an interface pattern.
 	See <ref id="proto-iface" name="interface"> section for detailed
 	description of interface patterns with extended clauses.
 
-	<tag><label id="opt-listen-bgp">listen bgp [address <m/address/] [port <m/port/] [dual]</tag>
-	This option allows to specify address and port where BGP protocol should
-	listen. It is global option as listening socket is common to all BGP
-	instances. Default is to listen on all addresses (0.0.0.0) and port 179.
-	In IPv6 mode, option <cf/dual/ can be used to specify that BGP socket
-	should accept both IPv4 and IPv6 connections (but even in that case,
-	BIRD would accept IPv6 routes only). Such behavior was default in older
-	versions of BIRD.
-
 	<tag><label id="opt-graceful-restart">graceful restart wait <m/number/</tag>
 	During graceful restart recovery, BIRD waits for convergence of routing
 	protocols. This option allows to specify a timeout for the recovery to
@@ -476,12 +574,13 @@ protocol rip {
 	<cf/protocol/ times, and the <cf/iso long ms/ format for <cf/base/ and
 	<cf/log/ times.
 
-	<tag><label id="opt-table">table <m/name/ [sorted]</tag>
-	Create a new routing table. The default routing table is created
+	<tag><label id="opt-table"><m/nettype/ table <m/name/ [sorted]</tag>
+	Create a new routing table. The default routing tables <cf/master4/ and <cf/master6/ are created
 	implicitly, other routing tables have to be added by this command.
 	Option <cf/sorted/ can be used to enable sorting of routes, see
 	<ref id="dsc-table-sorted" name="sorted table"> description for details.
 
+<!--
 	<tag><label id="opt-roa-table">roa table <m/name/ [ { <m/roa table options .../ } ]</tag>
 	Create a new ROA (Route Origin Authorization) table. ROA tables can be
 	used to validate route origination of BGP routes. A ROA table contains
@@ -495,12 +594,12 @@ protocol rip {
 	<m/num/</cf>, which can be used to populate the ROA table with static
 	ROA entries. The option may be used multiple times. Other entries can be
 	added dynamically by <cf/add roa/ command.
+-->
 
 	<tag><label id="opt-eval">eval <m/expr/</tag>
-	Evaluates given filter expression. It is used by us for	testing of filters.
+	Evaluates given filter expression. It is used by the developers for testing of filters.
 </descrip>
 
-
 <sect>Protocol options
 <label id="protocol-opts">
 
@@ -549,58 +648,10 @@ agreement").
 	This option can be used to override global router id for a given
 	protocol. Default: uses global router id.
 
-	<tag><label id="proto-import">import all | none | filter <m/name/ | filter { <m/filter commands/ } | where <m/filter expression/</tag>
-	Specify a filter to be used for filtering routes coming from the
-	protocol to the routing table. <cf/all/ is shorthand for <cf/where true/
-	and <cf/none/ is shorthand for <cf/where false/. Default: <cf/all/.
-
-	<tag><label id="proto-export">export <m/filter/</tag>
-	This is similar to the <cf>import</cf> keyword, except that it works in
-	the direction from the routing table to the protocol. Default: <cf/none/.
-
-	<tag><label id="proto-import-keep-filtered">import keep filtered <m/switch/</tag>
-	Usually, if an import filter rejects a route, the route is forgotten.
-	When this option is active, these routes are kept in the routing table,
-	but they are hidden and not propagated to other protocols. But it is
-	possible to show them using <cf/show route filtered/. Note that this
-	option does not work for the pipe protocol. Default: off.
-
-	<tag><label id="proto-import-limit">import limit [<m/number/ | off ] [action warn | block | restart | disable]</tag>
-	Specify an import route limit (a maximum number of routes imported from
-	the protocol) and optionally the action to be taken when the limit is
-	hit. Warn action just prints warning log message. Block action discards
-	new routes coming from the protocol. Restart and disable actions shut
-	the protocol down like appropriate commands. Disable is the default
-	action if an action is not explicitly specified. Note that limits are
-	reset during protocol reconfigure, reload or restart. Default: <cf/off/.
-
-	<tag><label id="proto-receive-limit">receive limit [<m/number/ | off ] [action warn | block | restart | disable]</tag>
-	Specify an receive route limit (a maximum number of routes received from
-	the protocol and remembered). It works almost identically to <cf>import
-	limit</cf> option, the only difference is that if <cf/import keep
-	filtered/ option is active, filtered routes are counted towards the
-	limit and blocked routes are forgotten, as the main purpose of the
-	receive limit is to protect routing tables from overflow. Import limit,
-	on the contrary, counts accepted routes only and routes blocked by the
-	limit are handled like filtered routes. Default: <cf/off/.
-
-	<tag><label id="proto-export-limit">export limit [ <m/number/ | off ] [action warn | block | restart | disable]</tag>
-	Specify an export route limit, works similarly to the <cf>import
-	limit</cf> option, but for the routes exported to the protocol. This
-	option is experimental, there are some problems in details of its
-	behavior -- the number of exported routes can temporarily exceed the
-	limit without triggering it during protocol reload, exported routes
-	counter ignores route blocking and block action also blocks route
-	updates of already accepted routes -- and these details will probably
-	change in the future. Default: <cf/off/.
-
 	<tag><label id="proto-description">description "<m/text/"</tag>
 	This is an optional description of the protocol. It is displayed as a
 	part of the output of 'show route all' command.
 
-	<tag><label id="proto-table">table <m/name/</tag>
-	Connect this protocol to a non-default routing table.
-
 	<tag><label id="proto-vrf">vrf "<m/text/"</tag>
 	Associate the protocol with specific VRF. The protocol will be
 	restricted to interfaces assigned to the VRF and will use sockets bound
@@ -609,6 +660,11 @@ agreement").
 	<cf/table/ option. Note that the VRF support in BIRD and Linux kernel
 	(4.11) is still in development and is currently problematic outside of
 	multihop BGP.
+
+	<tag><label id="proto-channel"><m/channel name/ [{<m/channel config/}]</tag>
+	Every channel must be explicitly stated. See the protocol-specific configuration
+	for the list of supported channel names.
+	See the <ref id="channel-opts" name="channel configuration section"> for channel definition.
 </descrip>
 
 <p>There are several options that give sense only with certain protocols:
@@ -638,7 +694,7 @@ agreement").
 	option is used.
 
 	This option is allowed in Babel, BFD, Direct, OSPF, RAdv and RIP
-	protocols, but in OSPF protocol it is used in the <cf/area/ subsection.
+	protocols. In OSPF protocol it is used in the <cf/area/ subsection.
 
 	Default: none.
 
@@ -726,135 +782,81 @@ agreement").
 
 </descrip>
 
+<sect>Channel options
+<label id="channel-opts">
 
-<sect>Flowspec network type
-<label id="flowspec-network-type">
+<p>Every channel belongs to a protocol and is configured inside its block. The
+minimal channel config is empty, then it uses the default values.
+The channel's name implies its nettype.
 
-<p>The flow specification are rules for routers and firewalls for filtering
-purpose. It is described by <rfc id="5575">. There are 3 types of arguments:
-<m/inet4/ or <m/inet6/ prefixes, bitmasks matching expressions and numbers
-matching expressions.
+<descrip>
+	<tag><label id="proto-table">table <m/name/</tag>
+	Connect this channel to a non-default routing table. Default 
 
-Bitmasks matching is written using <m/value/<cf>/</cf><m/mask/ or
-<cf/!/<m/value/<cf>/</cf><m/mask/ pairs. It means that <cf/(/<m/data/ <cf/&/
-<m/mask/<cf/)/ is or is not equal to <m/value/.
+	<tag><label id="proto-import">import all | none | filter <m/name/ | filter { <m/filter commands/ } | where <m/boolean filter expression/</tag>
+	Specify a filter to be used for filtering routes coming from the
+	protocol to the routing table. <cf/all/ is for keeping all routes,
+	<cf/none/ is for dropping all routes. Default: <cf/all/.
 
-Numbers matching is a matching sequence of numbers and ranges separeted by a
-commas (<cf/,/) (e.g. <cf/10,20,30/). Ranges can be written using double dots
-<cf/../ notation (e.g. <cf/80..90,120..124/). An alternative notation are
-sequence of one or more pairs of relational operators and values separated by
-logical operators <cf/&&/ or <cf/||/. Allowed relational operators are <cf/=/,
-<cf/!=/, <cf/</, <cf/<=/, <cf/>/, <cf/>=/, <cf/true/ and <cf/false/.
+	<tag><label id="proto-export">export <m/filter/</tag>
+	This is similar to the <cf>import</cf> keyword, except that it works in
+	the direction from the routing table to the protocol. Default: <cf/none/.
 
-<sect1>IPv4 Flowspec
+	<tag><label id="proto-import-keep-filtered">import keep filtered <m/switch/</tag>
+	Usually, if an import filter rejects a route, the route is forgotten.
+	When this option is active, these routes are kept in the routing table,
+	but they are hidden and not propagated to other protocols. But it is
+	possible to show them using <cf/show route filtered/. Note that this
+	option does not work for the pipe protocol. Default: off.
 
-<p><descrip>
-	<tag><label id="flow-dst">dst <m/inet4/</tag>
-	Set a matching destination prefix (e.g. <cf>dst 192.168.0.0/16</cf>).
-	Only this option is mandatory in IPv4 Flowspec.
+	<tag><label id="proto-import-limit">import limit [<m/number/ | off ] [action warn | block | restart | disable]</tag>
+	Specify an import route limit (a maximum number of routes imported from
+	the protocol) and optionally the action to be taken when the limit is
+	hit. Warn action just prints warning log message. Block action discards
+	new routes coming from the protocol. Restart and disable actions shut
+	the protocol down like appropriate commands. Disable is the default
+	action if an action is not explicitly specified. Note that limits are
+	reset during protocol reconfigure, reload or restart. Default: <cf/off/.
 
-	<tag><label id="flow-src">src <m/inet4/</tag>
-	Set a matching source prefix (e.g. <cf>src 10.0.0.0/8</cf>).
+	<tag><label id="proto-receive-limit">receive limit [<m/number/ | off ] [action warn | block | restart | disable]</tag>
+	Specify an receive route limit (a maximum number of routes received from
+	the protocol and remembered). It works almost identically to <cf>import
+	limit</cf> option, the only difference is that if <cf/import keep
+	filtered/ option is active, filtered routes are counted towards the
+	limit and blocked routes are forgotten, as the main purpose of the
+	receive limit is to protect routing tables from overflow. Import limit,
+	on the contrary, counts accepted routes only and routes blocked by the
+	limit are handled like filtered routes. Default: <cf/off/.
 
-	<tag><label id="flow-proto">proto <m/numbers-match/</tag>
-	Set a matching IP protocol numbers (e.g.  <cf/proto 6/).
-
-	<tag><label id="flow-port">port <m/numbers-match/</tag>
-	Set a matching source or destination TCP/UDP port numbers (e.g.
-	<cf>port 1..1023,1194,3306</cf>).
-
-	<tag><label id="flow-dport">dport <m/numbers-match/</tag>
-	Set a mating destination port numbers (e.g. <cf>dport 49151</cf>).
-
-	<tag><label id="flow-sport">sport <m/numbers-match/</tag>
-	Set a matching source port numbers (e.g. <cf>sport = 0</cf>).
-
-	<tag><label id="flow-icmp-type">icmp type <m/numbers-match/</tag>
-	Set a matching type field number of an ICMP packet (e.g. <cf>icmp type
-	3</cf>)
-
-	<tag><label id="flow-icmp-code">icmp code <m/numbers-match/</tag>
-	Set a matching code field number of an ICMP packet (e.g. <cf>icmp code
-	1</cf>)
-
-	<tag><label id="flow-tcp-flags">tcp flags <m/bitmask-match/</tag>
-	Set a matching bitmask for TCP header flags (aka control bits) (e.g.
-	<cf>tcp flags 0x03/0x0f;</cf>). The maximum length of mask is 12 bits
-	(0xfff).
-
-	<tag><label id="flow-length">length <m/numbers-match/</tag>
-	Set a matching packet length (e.g. <cf>length > 1500;</cf>)
-
-	<tag><label id="flow-dscp">dscp <m/numbers-match/</tag>
-	Set a matching DiffServ Code Point number (e.g. <cf>length > 1500;</cf>).
-
-	<tag><label id="flow-fragment">fragment <m/fragmentation-type/</tag>
-	Set a matching type of packet fragmentation. Allowed fragmentation
-	types are <cf/dont_fragment/, <cf/is_fragment/, <cf/first_fragment/,
-	<cf/last_fragment/ (e.g. <cf>fragment is_fragment &&
-	!dont_fragment</cf>).
+	<tag><label id="proto-export-limit">export limit [ <m/number/ | off ] [action warn | block | restart | disable]</tag>
+	Specify an export route limit, works similarly to the <cf>import
+	limit</cf> option, but for the routes exported to the protocol. This
+	option is experimental, there are some problems in details of its
+	behavior -- the number of exported routes can temporarily exceed the
+	limit without triggering it during protocol reload, exported routes
+	counter ignores route blocking and block action also blocks route
+	updates of already accepted routes -- and these details will probably
+	change in the future. Default: <cf/off/.
 </descrip>
 
-<p><code>
-protocol static {
-	flow4;
-
-	route flow4 {
-		dst 10.0.0.0/8;
-		port > 24 && < 30 || 40..50,60..70,80 && >= 90;
-		tcp flags 0x03/0x0f;
-		length > 1024;
-		dscp = 63;
-		fragment dont_fragment, is_fragment || !first_fragment;
-	} drop;
+<p>This is a trivial example of RIP configured for IPv6 on all interfaces:
+<code>
+protocol rip ng {
+	ipv6;
+	interface "*";
 }
 </code>
 
-<sect1>Differences for IPv6 Flowspec
-
-<p>Flowspec IPv6 are same as Flowspec IPv4 with a few exceptions.
-<itemize>
-	<item>Prefixes <m/inet6/ can be specified not only with prefix length,
-	but with prefix <cf/offset/ <m/num/ too (e.g.
-	<cf>::1234:5678:9800:0000/101 offset 64</cf>). Offset means to don't
-	care of <m/num/ first bits.
-	<item>IPv6 Flowspec hasn't mandatory any flowspec component.
-	<item>In IPv6 packets, there is a matching the last next header value
-	for a matching IP protocol number (e.g. <cf>next header 6</cf>).
-	<item>It is not possible to set <cf>dont_fragment</cf> as a type of
-	packet fragmentation.
-</itemize>
-
-<p><descrip>
-	<tag><label id="flow6-dst">dst <m/inet6/ [offset <m/num/]</tag>
-	Set a matching destination IPv6 prefix (e.g. <cf>dst
-	::1c77:3769:27ad:a11a/128 offset 64</cf>).
-
-	<tag><label id="flow6-src">src <m/inet6/ [offset <m/num/]</tag>
-	Set a matching source IPv6 prefix (e.g. <cf>src fe80::/64</cf>).
-
-	<tag><label id="flow6-next-header">next header <m/numbers-match/</tag>
-	Set a matching IP protocol numbers (e.g. <cf>next header != 6</cf>).
-
-	<tag><label id="flow6-label">label <m/bitmask-match/</tag>
-	Set a 20-bit bitmask for matching Flow Label field in IPv6 packets
-	(e.g. <cf>label 0x8e5/0x8e5</cf>).
-</descrip>
-
-<p><code>
-protocol static {
-	flow6;
-
-	route flow6 {
-		dst fec0:1122:3344:5566:7788:99aa:bbcc:ddee/128;
-		src 0000:0000:0000:0001:1234:5678:9800:0000/101 offset 63;
-		next header = 23;
-		sport > 24 && < 30 || = 40 || 50,60,70..80;
-		dport = 50;
-		tcp flags 0x03/0x0f, !0/0xff || 0x33/0x33;
-		fragment !is_fragment || !first_fragment;
-		label 0xaaaa/0xaaaa && 0x33/0x33;
-	} drop;
+<p>And this is a non-trivial example.
+<code>
+protocol rip ng {
+	ipv6 {
+		table mytable6;
+		import filter { ... };
+		export filter { ... };
+		import limit 50;
+	};
+	interface "*";
 }
 </code>
 
@@ -898,6 +900,7 @@ This argument can be omitted if there exists only a single instance.
 	and protocol status, possibly giving verbose information, if <cf/all/ is
 	specified.
 
+	<!-- TODO: Move these protocol-specific remote control commands to the protocol sections -->
 	<tag><label id="cli-show-ospf-iface">show ospf interface [<m/name/] ["<m/interface/"]</tag>
 	Show detailed information about OSPF interfaces.
 
@@ -1054,7 +1057,6 @@ This argument can be omitted if there exists only a single instance.
 	Evaluate given expression.
 </descrip>
 
-
 <chapt>Filters
 <label id="filters">
 
@@ -1185,9 +1187,10 @@ foot).
 	a shell pattern (represented also as a string).
 
 	<tag><label id="type-ip">ip</tag>
-	This type can hold a single IP address. Depending on the compile-time
-	configuration of BIRD you are using, it is either an IPv4 or IPv6
-	address; this may be checked by <cf>.is_ip4</cf> which returns <cf/bool/.
+	This type can hold a single IP address. The IPv4 addresses are stored
+	as IPv4-Mapped IPv6 addresses so one data type for both of them is used.
+	Whether the address is IPv4 or not may be checked by 
+	<cf>.is_ip4</cf> which returns <cf/bool/.
         IP addresses are written in the standard notation
 	(<cf/10.20.30.40/ or <cf/fec0:3:4::1/). You can apply special operator
 	<cf>.mask(<M>num</M>)</cf> on values of type ip. It masks out all but
@@ -1204,7 +1207,7 @@ foot).
 	<cf/NET_IP4/ and <cf/NET_IP6/ prefixes hold an IP prefix. The literals
 	are written as <cf><m/ipaddress//<m/pxlen/</cf>,
 	or <cf><m>ipaddress</m>/<m>netmask</m></cf>. There are two special
-	operators on IP prefixes: <cf/.ip/ which extracts the IP address from
+	operators on these: <cf/.ip/ which extracts the IP address from
 	the pair, and <cf/.len/, which separates prefix length from the pair.
 	So <cf>1.2.0.0/16.len = 16</cf> is true.
 
@@ -1222,6 +1225,13 @@ foot).
 	<cf/NET_FLOW4/ and <cf/NET_FLOW6/ hold an IP prefix together with a
 	flowspec rule. Filters currently don't support flowspec parsing.
 
+	<cf/NET_MPLS/ holds a single MPLS label and its handling is currently not implemented.
+
+	<tag><label id="type-vpnrd">vpnrd</tag>
+	This is a route distinguisher according to <rfc id="4364">. There are
+	three kinds of RD's: <cf><m/asn/:<m/32bit int/</cf>, <cf><m/asn4/:<m/16bit int/</cf>
+	and <cf><m/IPv4 address/:<m/32bit int/</cf>
+
 	<tag><label id="type-ec">ec</tag>
 	This is a specialized type used to represent BGP extended community
 	values. It is essentially a 64bit value, literals of this type are
@@ -1318,6 +1328,10 @@ foot).
 	<cf>192.168.0.0/16{16,24}</cf> and <cf>192.168.0.0/16 ge 24</cf> as
 	<cf>192.168.0.0/16{24,32}</cf>.
 
+	It's possible to mix IPv4 and IPv6 prefixes/addresses in a prefix/ip set
+	but its behavior may change between versions without any warning; don't do
+	it unless you are more than sure what you are doing. (Really, don't do it.)
+
 	<tag><label id="type-enum">enum</tag>
 	Enumeration types are fixed sets of possibilities. You can't define your
 	own variables of such type, but some route attributes are of enumeration
@@ -1340,7 +1354,7 @@ foot).
 
 	<cf><m/P/.len</cf> returns the length of path <m/P/.
 
-	<cf><m/P/.empty</cf> resets path <m/P/ to empty path.
+	<cf><m/P/.empty()</cf> makes the path <m/P/ empty.
 
 	<cf>prepend(<m/P/,<m/A/)</cf> prepends ASN <m/A/ to path <m/P/ and
 	returns the result.
@@ -1369,8 +1383,6 @@ foot).
 	expressions can also contain integer expressions enclosed in parenthesis
 	and integer variables, for example <tt>[= * 4 (1+2) a =]</tt>. You can
         also use ranges, for example <tt>[= * 3..5 2 100..200 * =]</tt>.
-        There is also old (deprecated) syntax that uses / .. / instead of [= .. =]
-        and ? instead of *.
 
 	<tag><label id="type-clist">clist</tag>
 	Clist is similar to a set, except that unlike other sets, it can be
@@ -1380,7 +1392,7 @@ foot).
 
 	<cf><m/C/.len</cf> returns the length of clist <m/C/.
 
-	<cf><m/C/.empty</cf> resets clist <m/C/ to empty clist.
+	<cf><m/C/.empty</cf> makes the list <m/C/ empty.
 
 	<cf>add(<m/C/,<m/P/)</cf> adds pair (or quad) <m/P/ to clist <m/C/ and
 	returns the result. If item <m/P/ is already in clist <m/C/, it does
@@ -1418,13 +1430,12 @@ foot).
 	lclists, with LCs instead of pairs as arguments.
 </descrip>
 
-
 <sect>Operators
 <label id="operators">
 
 <p>The filter language supports common integer operators <cf>(+,-,*,/)</cf>,
 parentheses <cf/(a*(b+c))/, comparison <cf/(a=b, a!=b, a&lt;b, a&gt;=b)/.
-Logical operations include unary not (<cf/!/), and (<cf/&amp;&amp;/) and or
+Logical operations include unary not (<cf/!/), and (<cf/&amp;&amp;/), and or
 (<cf/&verbar;&verbar;/). Special operators include (<cf/&tilde;/,
 <cf/!&tilde;/) for "is (not) element of a set" operation - it can be used on
 element and set of elements of the same type (returning true if element is
@@ -1455,11 +1466,11 @@ prefix and an ASN as arguments.
 
 <p>Filters support two control structures: conditions and case switches.
 
-<p>Syntax of a condition is: <cf>if <M>boolean expression</M> then <m/command1/;
-else <m/command2/;</cf> and you can use <cf>{ <m/command_1/; <m/command_2/;
+<p>Syntax of a condition is: <cf>if <M>boolean expression</M> then <m/commandT/;
+else <m/commandF/;</cf> and you can use <cf>{ <m/command1/; <m/command2/;
 <M>...</M> }</cf> instead of either command. The <cf>else</cf> clause may be
-omitted. If the <cf><m>boolean expression</m></cf> is true, <m/command1/ is
-executed, otherwise <m/command2/ is executed.
+omitted. If the <cf><m>boolean expression</m></cf> is true, <m/commandT/ is
+executed, otherwise <m/commandF/ is executed.
 
 <p>The <cf>case</cf> is similar to case from Pascal. Syntax is <cf>case
 <m/expr/ { else: | <m/num_or_prefix [ .. num_or_prefix]/: <m/statement/ ; [
@@ -1498,8 +1509,8 @@ clist for most purposes.
 
 <descrip>
 	<tag><label id="rta-net"><m/prefix/ net</tag>
-	Network the route is talking about. Read-only. (See the chapter about
-	routing tables.)
+	The network prefix or anything else the route is talking about.
+	The primary key of the table. Read-only. (See the <ref id="routes" name="chapter about routes">.)
 
 	<tag><label id="rta-scope"><m/enum/ scope</tag>
 	The scope of the route. Possible values: <cf/SCOPE_HOST/ for routes
@@ -1530,11 +1541,6 @@ clist for most purposes.
 	<cf/RTS_OSPF_IA/, <cf/RTS_OSPF_EXT1/, <cf/RTS_OSPF_EXT2/, <cf/RTS_BGP/,
 	<cf/RTS_PIPE/, <cf/RTS_BABEL/.
 
-	<tag><label id="rta-cast"><m/enum/ cast</tag>
-	Route type (Currently <cf/RTC_UNICAST/ for normal routes,
-	<cf/RTC_BROADCAST/, <cf/RTC_MULTICAST/, <cf/RTC_ANYCAST/ will be used in
-	the future for broadcast, multicast and anycast routes). Read-only.
-
 	<tag><label id="rta-dest"><m/enum/ dest</tag>
 	Type of destination the packets should be sent to
 	(<cf/RTD_ROUTER/ for forwarding to a neighboring router,
@@ -1566,7 +1572,7 @@ clist for most purposes.
 	1 metric.
 </descrip>
 
-<p>There also exist some protocol-specific attributes which are described in the
+<p>There also exist protocol-specific attributes which are described in the
 corresponding protocol sections.
 
 
@@ -1626,6 +1632,8 @@ other protocols, but supports the following per-interface configuration options:
 
 <code>
 protocol babel [<name>] {
+	ipv4 { <channel config> };
+	ipv6 { <channel config> };
 	interface <interface pattern> {
 		type <wired|wireless>;
 		rxcost <number>;
@@ -1645,6 +1653,9 @@ protocol babel [<name>] {
 </code>
 
 <descrip>
+      <tag><label id="babel-channel">ipv4|ipv6 <m/channel config/</tag>
+      The supported channels are IPv4 and IPv6. 
+
       <tag><label id="babel-type">type wired|wireless </tag>
       This option specifies the interface type: Wired or wireless. On wired
       interfaces a neighbor is considered unreachable after a small number of
@@ -1757,6 +1768,11 @@ protocol babel {
 }
 </code>
 
+<sect1>Known issues
+<label id="babel-issues">
+
+<p>When retracting a route, Babel sends instead an unreachable route for a little while (according to RFC). The
+interaction of this behavior with other protocols is not well tested and strange things may happen.
 
 <sect>BFD
 <label id="bfd">
@@ -1785,10 +1801,7 @@ the BFD session went down).
 advanced features like the echo mode or authentication are not implemented), IP
 transport for BFD as defined in <rfc id="5881"> and <rfc id="5883"> and
 interaction with client protocols as defined in <rfc id="5882">.
-
-<p>Note that BFD implementation in BIRD is currently a new feature in
-development, expect some rough edges and possible UI and configuration changes
-in the future. Also note that we currently support at most one protocol instance.
+We currently support at most one protocol instance.
 
 <p>BFD packets are sent with a dynamic source port number. Linux systems use by
 default a bit different dynamic port range than the IANA approved one
@@ -2064,8 +2077,8 @@ determine immediate next hops for routes and to know their internal distances to
 boundary routers for the purpose of BGP route selection. In BIRD, there is
 usually one routing table used for both IGP routes and BGP routes.
 
-<sect1>Configuration
-<label id="bgp-config">
+<sect1>Protocol configuration
+<label id="bgp-proto-config">
 
 <p>Each instance of the BGP corresponds to one neighboring router. This allows
 to set routing policy and all the other parameters differently for each neighbor
@@ -2130,52 +2143,6 @@ using the following configuration parameters:
 	same address family and using the same local port) should have set
 	<cf/strict bind/, or none of them. Default: disabled.
 
-	<tag><label id="bgp-next-hop-self">next hop self</tag>
-	Avoid calculation of the Next Hop attribute and always advertise our own
-	source address as a next hop. This needs to be used only occasionally to
-	circumvent misconfigurations of other routers. Default: disabled.
-
-	<tag><label id="bgp-next-hop-keep">next hop keep</tag>
-	Forward the received Next Hop attribute even in situations where the
-	local address should be used instead, like when the route is sent to an
-	interface with a different subnet. Default: disabled.
-
-	<tag><label id="bgp-missing-lladdr">missing lladdr self|drop|ignore</tag>
-	Next Hop attribute in BGP-IPv6 sometimes contains just the global IPv6
-	address, but sometimes it has to contain both global and link-local IPv6
-	addresses. This option specifies what to do if BIRD have to send both
-	addresses but does not know link-local address. This situation might
-	happen when routes from other protocols are exported to BGP, or when
-	improper updates are received from BGP peers. <cf/self/ means that BIRD
-	advertises its own local address instead. <cf/drop/ means that BIRD
-	skips that prefixes and logs error. <cf/ignore/ means that BIRD ignores
-	the problem and sends just the global address (and therefore forms
-	improper BGP update). Default: <cf/self/, unless BIRD is configured as a
-	route server (option <cf/rs client/), in that case default is <cf/ignore/,
-	because route servers usually do not forward packets themselves.
-
-	<tag><label id="bgp-gateway">gateway direct|recursive</tag>
-	For received routes, their <cf/gw/ (immediate next hop) attribute is
-	computed from received <cf/bgp_next_hop/ attribute. This option
-	specifies how it is computed. Direct mode means that the IP address from
-	<cf/bgp_next_hop/ is used if it is directly reachable, otherwise the
-	neighbor IP address is used. Recursive mode means that the gateway is
-	computed by an IGP routing table lookup for the IP address from
-	<cf/bgp_next_hop/. Note that there is just one level of indirection in
-	recursive mode - the route obtained by the lookup must not be recursive
-	itself, to prevent mutually recursive routes.
-
-	Recursive mode is the behavior specified by the BGP
-	standard. Direct mode is simpler, does not require any routes in a
-	routing table, and was used in older versions of BIRD, but does not
-	handle well nontrivial iBGP setups and multihop. Recursive mode is
-	incompatible with <ref id="dsc-table-sorted" name="sorted tables">. Default:
-	<cf/direct/ for direct sessions, <cf/recursive/ for multihop sessions.
-
-	<tag><label id="bgp-igp-table">igp table <m/name/</tag>
-	Specifies a table that is used as an IGP routing table. Default: the
-	same as the table BGP is connected to.
-
 	<tag><label id="bgp-check-link">check link <M>switch</M></tag>
 	BGP could use hardware link state into consideration.  If enabled,
 	BIRD tracks the link state of the associated interface and when link
@@ -2260,24 +2227,6 @@ using the following configuration parameters:
 	example does not prepend its AS number to AS PATH attribute and
 	keeps MED attribute). Default: disabled.
 
-	<tag><label id="bgp-secondary">secondary <m/switch/</tag>
-	Usually, if an export filter rejects a selected route, no other route is
-	propagated for that network. This option allows to try the next route in
-	order until one that is accepted is found or all routes for that network
-	are rejected. This can be used for route servers that need to propagate
-	different tables to each client but do not want to have these tables
-	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-add-paths">add paths <m/switch/|rx|tx</tag>
-	Standard BGP can propagate only one path (route) per destination network
-	(usually the selected one). This option controls the add-path protocol
-	extension, which allows to advertise any number of paths to a
-	destination. Note that to be active, add-path has to be enabled on both
-	sides of the BGP session, but it could be enabled separately for RX and
-	TX direction. When active, all available routes accepted by the export
-	filter are advertised to the neighbor. Default: off.
-
 	<tag><label id="bgp-allow-local-pref">allow bgp_local_pref <m/switch/</tag>
 	A standard BGP implementation do not send the Local Preference attribute
 	to eBGP neighbors and ignore this attribute if received from eBGP
@@ -2309,20 +2258,6 @@ using the following configuration parameters:
 	related procedures. Note that even when disabled, BIRD can send route
 	refresh requests.  Default: on.
 
-	<tag><label id="bgp-graceful-restart">graceful restart <m/switch/|aware</tag>
-	When a BGP speaker restarts or crashes, neighbors will discard all
-	received paths from the speaker, which disrupts packet forwarding even
-	when the forwarding plane of the speaker remains intact. <rfc
-	id="4724"> specifies an optional graceful restart mechanism to
-	alleviate this issue. This option controls the mechanism. It has three
-	states: Disabled, when no support is provided. Aware, when the graceful
-	restart support is announced and the support for restarting neighbors
-	is provided, but no local graceful restart is allowed (i.e.
-	receiving-only role). Enabled, when the full graceful restart
-	support is provided (i.e. both restarting and receiving role). Note
-	that proper support for local graceful restart requires also
-	configuration of other protocols.  Default: aware.
-
 	<tag><label id="bgp-graceful-restart-time">graceful restart time <m/number/</tag>
 	The restart time is announced in the BGP graceful restart capability
 	and specifies how long the neighbor would wait for the BGP session to
@@ -2455,6 +2390,118 @@ using the following configuration parameters:
 	versions of BIRD).
 </descrip>
 
+<sect1>Channel configuration
+<label id="bgp-channel-config">
+
+<p>BGP supports several AFIs and SAFIs over one connection. Every AFI/SAFI
+announced to the peer corresponds to one channel. The table of supported AFI/SAFIs
+together with their appropriate channels follows.
+
+<tabular ca="l|l|l|r|r">
+  <bf/Channel name/   | <bf/Table nettype/ | <bf/IGP table allowed/  | <bf/AFI/ | <bf/SAFI/
+@ <hline>
+  <cf/ipv4/	      | <cf/ipv4/          | <cf/ipv4/ and <cf/ipv6/ | 1        | 1
+@ <cf/ipv6/           | <cf/ipv6/          | <cf/ipv4/ and <cf/ipv6/ | 2        | 1
+@ <cf/ipv4 multicast/ | <cf/ipv4/          | <cf/ipv4/ and <cf/ipv6/ | 1        | 2
+@ <cf/ipv6 multicast/ | <cf/ipv6/          | <cf/ipv4/ and <cf/ipv6/ | 2        | 2
+@ <cf/ipv4 mpls/      | <cf/ipv4/          | <cf/ipv4/ and <cf/ipv6/ | 1        | 4
+@ <cf/ipv6 mpls/      | <cf/ipv6/          | <cf/ipv4/ and <cf/ipv6/ | 2        | 4
+@ <cf/vpn4 mpls/      | <cf/vpn4/          | <cf/ipv4/ and <cf/ipv6/ | 1        | 128
+@ <cf/vpn6 mpls/      | <cf/vpn6/          | <cf/ipv4/ and <cf/ipv6/ | 2        | 128
+@ <cf/vpn4 multicast/ | <cf/vpn4/          | <cf/ipv4/ and <cf/ipv6/ | 1        | 129
+@ <cf/vpn6 multicast/ | <cf/vpn6/          | <cf/ipv4/ and <cf/ipv6/ | 2        | 129
+@ <cf/flow4/	      | <cf/flow4/         | ---                     | 1        | 133
+@ <cf/flow6/          | <cf/flow6/         | ---                     | 2        | 133
+</tabular>
+
+<p>BGP's channels have additional config options (together with the common ones):
+
+<descrip>
+	<tag><label id="bgp-next-hop-keep">next hop keep</tag>
+	Forward the received Next Hop attribute even in situations where the
+	local address should be used instead, like when the route is sent to an
+	interface with a different subnet. Default: disabled.
+
+	<tag><label id="bgp-next-hop-self">next hop self</tag>
+	Avoid calculation of the Next Hop attribute and always advertise our own
+	source address as a next hop. This needs to be used only occasionally to
+	circumvent misconfigurations of other routers. Default: disabled.
+
+	<tag><label id="bgp-next-hop-address">next hop address <m/ip/</tag>
+	Avoid calculation of the Next Hop attribute and always advertise this address
+	as a next hop.
+
+	<tag><label id="bgp-missing-lladdr">missing lladdr self|drop|ignore</tag>
+	Next Hop attribute in BGP-IPv6 sometimes contains just the global IPv6
+	address, but sometimes it has to contain both global and link-local IPv6
+	addresses. This option specifies what to do if BIRD have to send both
+	addresses but does not know link-local address. This situation might
+	happen when routes from other protocols are exported to BGP, or when
+	improper updates are received from BGP peers. <cf/self/ means that BIRD
+	advertises its own local address instead. <cf/drop/ means that BIRD
+	skips that prefixes and logs error. <cf/ignore/ means that BIRD ignores
+	the problem and sends just the global address (and therefore forms
+	improper BGP update). Default: <cf/self/, unless BIRD is configured as a
+	route server (option <cf/rs client/), in that case default is <cf/ignore/,
+	because route servers usually do not forward packets themselves.
+
+	<tag><label id="bgp-gateway">gateway direct|recursive</tag>
+	For received routes, their <cf/gw/ (immediate next hop) attribute is
+	computed from received <cf/bgp_next_hop/ attribute. This option
+	specifies how it is computed. Direct mode means that the IP address from
+	<cf/bgp_next_hop/ is used if it is directly reachable, otherwise the
+	neighbor IP address is used. Recursive mode means that the gateway is
+	computed by an IGP routing table lookup for the IP address from
+	<cf/bgp_next_hop/. Note that there is just one level of indirection in
+	recursive mode - the route obtained by the lookup must not be recursive
+	itself, to prevent mutually recursive routes.
+
+	Recursive mode is the behavior specified by the BGP
+	standard. Direct mode is simpler, does not require any routes in a
+	routing table, and was used in older versions of BIRD, but does not
+	handle well nontrivial iBGP setups and multihop. Recursive mode is
+	incompatible with <ref id="dsc-table-sorted" name="sorted tables">. Default:
+	<cf/direct/ for direct sessions, <cf/recursive/ for multihop sessions.
+
+	<tag><label id="bgp-igp-table">igp table <m/name/</tag>
+	Specifies a table that is used as an IGP routing table. The type of this
+	table must be as allowed in the table above. This option is allowed once
+	for every allowed table type. Default: the same as the main table
+	the channel is connected to (if eligible).
+
+	<tag><label id="bgp-secondary">secondary <m/switch/</tag>
+	Usually, if an export filter rejects a selected route, no other route is
+	propagated for that network. This option allows to try the next route in
+	order until one that is accepted is found or all routes for that network
+	are rejected. This can be used for route servers that need to propagate
+	different tables to each client but do not want to have these tables
+	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-add-paths">add paths <m/switch/|rx|tx</tag>
+	Standard BGP can propagate only one path (route) per destination network
+	(usually the selected one). This option controls the add-path protocol
+	extension, which allows to advertise any number of paths to a
+	destination. Note that to be active, add-path has to be enabled on both
+	sides of the BGP session, but it could be enabled separately for RX and
+	TX direction. When active, all available routes accepted by the export
+	filter are advertised to the neighbor. Default: off.
+
+	<tag><label id="bgp-graceful-restart">graceful restart <m/switch/|aware</tag>
+	When a BGP speaker restarts or crashes, neighbors will discard all
+	received paths from the speaker, which disrupts packet forwarding even
+	when the forwarding plane of the speaker remains intact. <rfc
+	id="4724"> specifies an optional graceful restart mechanism to
+	alleviate this issue. This option controls the mechanism. It has three
+	states: Disabled, when no support is provided. Aware, when the graceful
+	restart support is announced and the support for restarting neighbors
+	is provided, but no local graceful restart is allowed (i.e.
+	receiving-only role). Enabled, when the full graceful restart
+	support is provided (i.e. both restarting and receiving role). Note
+	that proper support for local graceful restart requires also
+	configuration of other protocols.  Default: aware.
+</descrip>
+
 <sect1>Attributes
 <label id="bgp-attr">
 
@@ -2549,20 +2596,38 @@ protocol bgp {
 	local as 65000;			     # Use a private AS number
 	neighbor 198.51.100.130 as 64496;    # Our neighbor ...
 	multihop;			     # ... which is connected indirectly
-	export filter {			     # We use non-trivial export rules
-		if source = RTS_STATIC then { # Export only static routes
-			# Assign our community
-			bgp_community.add((65000,64501));
-			# Artificially increase path length
-			# by advertising local AS number twice
-			if bgp_path ~ [= 65000 =] then
-				bgp_path.prepend(65000);
-			accept;
-		}
-		reject;
-	};
-	import all;
 	source address 198.51.100.14;	# Use a non-standard source address
+	ipv4 {
+		export filter {			     # We use non-trivial export rules
+			if source = RTS_STATIC then { # Export only static routes
+				# Assign our community
+				bgp_community.add((65000,64501));
+				# Artificially increase path length
+				# by advertising local AS number twice
+				if bgp_path ~ [= 65000 =] then
+					bgp_path.prepend(65000);
+				accept;
+			}
+			reject;
+		};
+		import all;
+		next hop self; # advertise this router as next hop
+		igp table myigptable4; # IGP table for routes with IPv4 nexthops
+		igp table myigptable6; # IGP table for routes with IPv6 nexthops
+	};
+	ipv6 {
+		export filter mylargefilter; # We use a named filter
+		import all;
+		missing lladdr self;
+		igp table myigptable4; # IGP table for routes with IPv4 nexthops
+		igp table myigptable6; # IGP table for routes with IPv6 nexthops
+	};
+	ipv4 multicast {
+		import all;
+		export filter someotherfilter;
+		table mymulticasttable4; # Another IPv4 table, dedicated for multicast
+		igp table myigptable4;
+	};
 }
 </code>
 
@@ -2572,7 +2637,7 @@ protocol bgp {
 
 <p>The Device protocol is not a real routing protocol. It doesn't generate any
 routes and it only serves as a module for getting information about network
-interfaces from the kernel.
+interfaces from the kernel. This protocol supports no channel.
 
 <p>Except for very unusual circumstances, you probably should include this
 protocol in the configuration since almost all other protocols require network
@@ -2616,13 +2681,13 @@ protocol device {
 }
 </code>
 
-
 <sect>Direct
 <label id="direct">
 
 <p>The Direct protocol is a simple generator of device routes for all the
 directly connected networks according to the list of interfaces provided by the
-kernel via the Device protocol.
+kernel via the Device protocol. The Direct protocol supports IPv4 and IPv6 channels
+in any count.
 
 <p>The question is whether it is a good idea to have such device routes in BIRD
 routing table. OS kernel usually handles device routes for directly connected
@@ -2667,10 +2732,12 @@ on Linux systems BIRD cannot change non-BIRD route in the kernel routing table.
 <p><code>
 protocol direct {
 	interface "-arc*", "*";		# Exclude the ARCnets
+	ipv4;
+	ipv4 { table myothertable; };
+	ipv6;
 }
 </code>
 
-
 <sect>Kernel
 <label id="krt">
 
@@ -2703,6 +2770,8 @@ routing table, there are two limitations - it is not possible to connect more
 kernel protocols to the same routing table and changing route destination
 (gateway) in an export filter of a kernel protocol does not work. Both
 limitations can be overcome using another routing table and the pipe protocol.
+The Kernel protocol obviously supports only two channels -- IPv4 and IPv6; only
+one of them can be configured in each protocol instance.
 
 <sect1>Configuration
 <label id="krt-config">
@@ -2832,14 +2901,18 @@ protocol kernel {		# Primary routing table
 	learn;			# Learn alien routes from the kernel
 	persist;		# Don't remove routes on bird shutdown
 	scan time 10;		# Scan kernel routing table every 10 seconds
-	import all;
-	export all;
+	ipv4 {
+		import all;
+		export all;
+	};
 }
 
 protocol kernel {		# Secondary routing table
-	table auxtable;
 	kernel table 100;
-	export all;
+	ipv4 {
+		table auxtable;
+		export all;
+	};
 }
 </code>
 
@@ -2890,6 +2963,8 @@ different id. These definitions includes many other switches and multiple
 definitions of interfaces. Definition of interface may contain many switches and
 constant definitions and list of neighbors on nonbroadcast networks.
 
+<p>OSPF v2 needs one IPv4 channel, OSPF v3 needs one IPv6 channel.
+
 <code>
 protocol ospf [v2|v3] &lt;name&gt; {
 	rfc1583compat &lt;switch&gt;;
@@ -3360,12 +3435,14 @@ network. This attribute is read-only. Default is <cf/ospf_metric2 = 10000/ and
 protocol ospf MyOSPF {
 	rfc1583compat yes;
 	tick 2;
-	export filter {
-		if source = RTS_BGP then {
-			ospf_metric1 = 100;
-			accept;
-		}
-		reject;
+	ipv4 {
+		export filter {
+			if source = RTS_BGP then {
+				ospf_metric1 = 100;
+				accept;
+			}
+			reject;
+		};
 	};
 	area 0.0.0.0 {
 		interface "eth*" {
@@ -3433,7 +3510,8 @@ routes to be passed from a table declared as primary (i.e., the one the pipe is
 connected to using the <cf/table/ configuration keyword) to the secondary one
 (declared using <cf/peer table/) and vice versa, depending on what's allowed by
 the filters. Export filters control export of routes from the primary table to
-the secondary one, import filters control the opposite direction.
+the secondary one, import filters control the opposite direction. Both tables
+must be of the same nettype.
 
 <p>The Pipe protocol may work in the transparent mode mode or in the opaque
 mode. In the transparent mode, the Pipe protocol retransmits all routes from
@@ -3463,6 +3541,10 @@ exporting a selected subset of one table to another one.
 <sect1>Configuration
 <label id="pipe-config">
 
+<p>Technically, the Pipe protocol is just a channel connected to a table on both sides.
+Therefore, the configuration block for <cf/protocol pipe/ shall directly include
+standard channel config options; see the example below.
+
 <p><descrip>
 	<tag><label id="pipe-peer-table">peer table <m/table/</tag>
 	Defines secondary routing table to connect to. The primary one is
@@ -3500,29 +3582,33 @@ table as1;				# Define the tables
 table as2;
 
 protocol kernel kern1 {			# Synchronize them with the kernel
-	table as1;
+	ipv4 { table as1; };
 	kernel table 1;
 }
 
 protocol kernel kern2 {
-	table as2;
+	ipv4 { table as2; };
 	kernel table 2;
 }
 
 protocol bgp bgp1 {			# The outside connections
-	table as1;
+	ipv4 {
+		table as1;
+		export all;
+		import all;
+	};
 	local as 1;
 	neighbor 192.168.0.1 as 1001;
-	export all;
-	import all;
 }
 
 protocol bgp bgp2 {
-	table as2;
+	ipv4 {
+		table as2;
+		export all;
+		import all;
+	};
 	local as 2;
 	neighbor 10.0.0.1 as 1002;
-	export all;
-	import all;
 }
 
 protocol pipe {				# The Pipe
@@ -3563,6 +3649,8 @@ addresses and choose a default route. BIRD implements router behavior as defined
 in <rfc id="4861">, router preferences and specific routes (<rfc id="4191">),
 and DNS extensions (<rfc id="6106">).
 
+<p>The RAdv protocols supports only IPv6 channels.
+
 <sect1>Configuration
 <label id="radv-config">
 
@@ -3874,7 +3962,6 @@ protocol radv {
 }
 </code>
 
-
 <sect>RIP
 <label id="rip">
 
@@ -4199,11 +4286,10 @@ protocol rpki [&lt;name&gt;] {
 }
 </code>
 
-<p>Alse note that you have to specify ROA table into which will be imported
-routes from a cache server. If you want to import only IPv4 prefixes you have
-to specify only roa4 table. Similarly with IPv6 prefixes only. If you want to
-fetch both IPv4 and even IPv6 ROAs you have to specify both types of ROA
-tables.
+<p>Alse note that you have to specify the ROA channel.
+If you want to import only IPv4 prefixes you have
+to specify only roa4 channel. Similarly with IPv6 prefixes only. If you want to
+fetch both IPv4 and even IPv6 ROAs you have to specify both channels.
 
 <sect2>RPKI protocol options
 
@@ -4336,21 +4422,10 @@ return packets as undeliverable if they are in your IP block, you don't have any
 specific destination for them and you don't want to send them out through the
 default route to prevent routing loops).
 
-<p>There are four types of static routes: `classical' routes telling to forward
-packets to a neighboring router (single path or multipath, possibly weighted),
-device routes specifying forwarding to hosts on a directly connected network,
-recursive routes computing their nexthops by doing route table lookups for a
-given IP, and special routes (sink, blackhole etc.)  which specify a special
-action to be done instead of forwarding the packet.
-
-<p>When the particular destination is not available (the interface is down or
-the next hop of the route is not a neighbor at the moment), Static just
-uninstalls the route from the table it is connected to and adds it again as soon
-as the destination becomes adjacent again.
-
 <p>There are three classes of definitions in Static protocol configuration --
 global options, static route definitions, and per-route options. Usually, the
 definition of the protocol contains mainly a list of static routes.
+Static routes have no specific attributes.
 
 <p>Global options:
 
@@ -4369,32 +4444,166 @@ definition of the protocol contains mainly a list of static routes.
 
 <p>Route definitions (each may also contain a block of per-route options):
 
+<sect1>Regular routes; MPLS switching rules</sect1>
+
+There exist several types of routes; keep in mind that <m/prefix/ syntax is
+<ref id="type-prefix" name="dependent on network type">.
+
 <descrip>
-	<tag><label id="static-route-via-ip">route <m/prefix/ via <m/ip/ [mpls <m/num/[/<m/num/[/<m/num/[...]]]]</tag>
-	Static single path route through a neighboring router. For link-local next hops,
-	interface can be specified as a part of the address (e.g.,
-	<cf/via fe80::1234%eth0/). MPLS labels should be specified in outer-first order.
-
-	<tag><label id="static-route-via-mpath">route <m/prefix/ via <m/ip/ [mpls <m/num/[/<m/num/[/<m/num/[...]]]] [weight <m/num/] [bfd <m/switch/] [via ...]</tag>
-	Static multipath route. Contains several nexthops (gateways), possibly
-	with their weights and MPLS labels.
-
-	<tag><label id="static-route-via-iface">route <m/prefix/ via <m/"interface"/</tag>
-	Static device route through an interface to hosts on a directly
-	connected network.
-
-	<tag><label id="static-route-recursive">route <m/prefix/ recursive <m/ip/</tag>
-	Static recursive route, its nexthop depends on a route table lookup for
-	given IP address.
-
-	<tag><label id="static-route-drop">route <m/prefix/ blackhole|unreachable|prohibit</tag>
+	<tag>route <m/prefix/ via <m/ip/|<m/"interface"/ [mpls <m/num/[/<m/num/[/<m/num/[...]]]]</tag>
+	Next hop routes may bear one or more <ref id="route-next-hop" name="next hops">.
+	Every next hop is preceded by <cf/via/ and configured as shown.
+	<tag>route <m/prefix/ recursive <m/ip/ [mpls <m/num/[/<m/num/[/<m/num/[...]]]]</tag>
+	Recursive nexthop resolves the given IP in the configured IGP table and
+	uses that route's next hop. The MPLS stacks are concatenated; on top is
+	the IGP's nexthop stack and on bottom is this route's stack.
+	<tag>route <m/prefix/ blackhole|unreachable|prohibit</tag>
 	Special routes specifying to silently drop the packet, return it as
 	unreachable or return it as administratively prohibited. First two
 	targets are also known as <cf/drop/ and <cf/reject/.
 </descrip>
 
-<p>Per-route options:
+<p>When the particular destination is not available (the interface is down or
+the next hop of the route is not a neighbor at the moment), Static just
+uninstalls the route from the table it is connected to and adds it again as soon
+as the destination becomes adjacent again.
 
+<sect1>Route Origin Authorization
+
+<p>The ROA config is just <cf>route <m/prefix/ max <m/int/ as <m/int/</cf> with no nexthop.
+
+<sect1>Flowspec
+<label id="flowspec-network-type">
+
+<p>The flow specification are rules for routers and firewalls for filtering
+purpose. It is described by <rfc id="5575">. There are 3 types of arguments:
+<m/inet4/ or <m/inet6/ prefixes, bitmasks matching expressions and numbers
+matching expressions.
+
+Bitmasks matching is written using <m/value/<cf>/</cf><m/mask/ or
+<cf/!/<m/value/<cf>/</cf><m/mask/ pairs. It means that <cf/(/<m/data/ <cf/&/
+<m/mask/<cf/)/ is or is not equal to <m/value/.
+
+Numbers matching is a matching sequence of numbers and ranges separeted by a
+commas (<cf/,/) (e.g. <cf/10,20,30/). Ranges can be written using double dots
+<cf/../ notation (e.g. <cf/80..90,120..124/). An alternative notation are
+sequence of one or more pairs of relational operators and values separated by
+logical operators <cf/&&/ or <cf/||/. Allowed relational operators are <cf/=/,
+<cf/!=/, <cf/</, <cf/<=/, <cf/>/, <cf/>=/, <cf/true/ and <cf/false/.
+
+<sect2>IPv4 Flowspec
+
+<p><descrip>
+	<tag><label id="flow-dst">dst <m/inet4/</tag>
+	Set a matching destination prefix (e.g. <cf>dst 192.168.0.0/16</cf>).
+	Only this option is mandatory in IPv4 Flowspec.
+
+	<tag><label id="flow-src">src <m/inet4/</tag>
+	Set a matching source prefix (e.g. <cf>src 10.0.0.0/8</cf>).
+
+	<tag><label id="flow-proto">proto <m/numbers-match/</tag>
+	Set a matching IP protocol numbers (e.g.  <cf/proto 6/).
+
+	<tag><label id="flow-port">port <m/numbers-match/</tag>
+	Set a matching source or destination TCP/UDP port numbers (e.g.
+	<cf>port 1..1023,1194,3306</cf>).
+
+	<tag><label id="flow-dport">dport <m/numbers-match/</tag>
+	Set a mating destination port numbers (e.g. <cf>dport 49151</cf>).
+
+	<tag><label id="flow-sport">sport <m/numbers-match/</tag>
+	Set a matching source port numbers (e.g. <cf>sport = 0</cf>).
+
+	<tag><label id="flow-icmp-type">icmp type <m/numbers-match/</tag>
+	Set a matching type field number of an ICMP packet (e.g. <cf>icmp type
+	3</cf>)
+
+	<tag><label id="flow-icmp-code">icmp code <m/numbers-match/</tag>
+	Set a matching code field number of an ICMP packet (e.g. <cf>icmp code
+	1</cf>)
+
+	<tag><label id="flow-tcp-flags">tcp flags <m/bitmask-match/</tag>
+	Set a matching bitmask for TCP header flags (aka control bits) (e.g.
+	<cf>tcp flags 0x03/0x0f;</cf>). The maximum length of mask is 12 bits
+	(0xfff).
+
+	<tag><label id="flow-length">length <m/numbers-match/</tag>
+	Set a matching packet length (e.g. <cf>length > 1500;</cf>)
+
+	<tag><label id="flow-dscp">dscp <m/numbers-match/</tag>
+	Set a matching DiffServ Code Point number (e.g. <cf>length > 1500;</cf>).
+
+	<tag><label id="flow-fragment">fragment <m/fragmentation-type/</tag>
+	Set a matching type of packet fragmentation. Allowed fragmentation
+	types are <cf/dont_fragment/, <cf/is_fragment/, <cf/first_fragment/,
+	<cf/last_fragment/ (e.g. <cf>fragment is_fragment &&
+	!dont_fragment</cf>).
+</descrip>
+
+<p><code>
+protocol static {
+	flow4;
+
+	route flow4 {
+		dst 10.0.0.0/8;
+		port > 24 && < 30 || 40..50,60..70,80 && >= 90;
+		tcp flags 0x03/0x0f;
+		length > 1024;
+		dscp = 63;
+		fragment dont_fragment, is_fragment || !first_fragment;
+	} drop;
+}
+</code>
+
+<sect2>Differences for IPv6 Flowspec
+
+<p>Flowspec IPv6 are same as Flowspec IPv4 with a few exceptions.
+<itemize>
+	<item>Prefixes <m/inet6/ can be specified not only with prefix length,
+	but with prefix <cf/offset/ <m/num/ too (e.g.
+	<cf>::1234:5678:9800:0000/101 offset 64</cf>). Offset means to don't
+	care of <m/num/ first bits.
+	<item>IPv6 Flowspec hasn't mandatory any flowspec component.
+	<item>In IPv6 packets, there is a matching the last next header value
+	for a matching IP protocol number (e.g. <cf>next header 6</cf>).
+	<item>It is not possible to set <cf>dont_fragment</cf> as a type of
+	packet fragmentation.
+</itemize>
+
+<p><descrip>
+	<tag><label id="flow6-dst">dst <m/inet6/ [offset <m/num/]</tag>
+	Set a matching destination IPv6 prefix (e.g. <cf>dst
+	::1c77:3769:27ad:a11a/128 offset 64</cf>).
+
+	<tag><label id="flow6-src">src <m/inet6/ [offset <m/num/]</tag>
+	Set a matching source IPv6 prefix (e.g. <cf>src fe80::/64</cf>).
+
+	<tag><label id="flow6-next-header">next header <m/numbers-match/</tag>
+	Set a matching IP protocol numbers (e.g. <cf>next header != 6</cf>).
+
+	<tag><label id="flow6-label">label <m/bitmask-match/</tag>
+	Set a 20-bit bitmask for matching Flow Label field in IPv6 packets
+	(e.g. <cf>label 0x8e5/0x8e5</cf>).
+</descrip>
+
+<p><code>
+protocol static {
+	flow6 { table myflow6; };
+
+	route flow6 {
+		dst fec0:1122:3344:5566:7788:99aa:bbcc:ddee/128;
+		src 0000:0000:0000:0001:1234:5678:9800:0000/101 offset 63;
+		next header = 23;
+		sport > 24 && < 30 || = 40 || 50,60,70..80;
+		dport = 50;
+		tcp flags 0x03/0x0f, !0/0xff || 0x33/0x33;
+		fragment !is_fragment || !first_fragment;
+		label 0xaaaa/0xaaaa && 0x33/0x33;
+	} drop;
+}
+</code>
+
+<sect1>Per-route options
 <descrip>
 	<tag><label id="static-route-bfd">bfd <m/switch/</tag>
 	The Static protocol could use BFD protocol for next hop liveness
@@ -4420,16 +4629,14 @@ definition of the protocol contains mainly a list of static routes.
 	exported to the OSPF protocol.
 </descrip>
 
-<p>Static routes have no specific attributes.
-
-<p>Example static config might look like this:
+<sect1>Example static config
 
 <p><code>
 protocol static {
-	table testable;			# Connect to a non-default routing table
+	ipv4 { table testable; };	# Connect to a non-default routing table
 	check link;			# Advertise routes only if link is up
 	route 0.0.0.0/0 via 198.51.100.130; # Default route
-	route 10.0.0.0/8 multipath	# Multipath route
+	route 10.0.0.0/8		# Multipath route
 		via 198.51.100.10 weight 2
 		via 198.51.100.20 bfd	# BFD-controlled next hop
 		via 192.0.2.1;
@@ -4446,7 +4653,6 @@ protocol static {
 }
 </code>
 
-
 <chapt>Conclusions
 <label id="conclusion">
 
@@ -4460,7 +4666,6 @@ versions of BIRD:
 <itemize>
 <item>Opaque LSA's
 <item>Route aggregation and flap dampening
-<item>Multipath routes
 <item>Multicast routing protocols
 <item>Ports to other systems
 </itemize>
diff --git a/filter/config.Y b/filter/config.Y
index 12e593d0..cd5a5b33 100644
--- a/filter/config.Y
+++ b/filter/config.Y
@@ -421,7 +421,7 @@ CF_KEYWORDS(FUNCTION, PRINT, PRINTN, UNSET, RETURN,
 %type <v> set_atom switch_atom fipa
 %type <px> fprefix
 %type <s> decls declsn one_decl function_params
-%type <h> bgp_path bgp_path_tail1 bgp_path_tail2
+%type <h> bgp_path bgp_path_tail
 %type <t> get_cf_position
 
 CF_GRAMMAR
@@ -763,25 +763,18 @@ bgp_path_expr:
  ;
 
 bgp_path:
-   PO  bgp_path_tail1 PC  { $$ = $2; }
- | '/' bgp_path_tail2 '/' { $$ = $2; }
+   PO  bgp_path_tail PC  { $$ = $2; }
  ;
 
-bgp_path_tail1:
-   NUM bgp_path_tail1		{ $$ = cfg_allocz(sizeof(struct f_path_mask)); $$->next = $2; $$->kind = PM_ASN; $$->val = $1; }
- | NUM DDOT NUM bgp_path_tail1	{ $$ = cfg_allocz(sizeof(struct f_path_mask)); $$->next = $4; $$->kind = PM_ASN_RANGE; $$->val = $1; $$->val2 = $3; }
- | '*' bgp_path_tail1		{ $$ = cfg_allocz(sizeof(struct f_path_mask)); $$->next = $2; $$->kind = PM_ASTERISK; }
- | '?' bgp_path_tail1		{ $$ = cfg_allocz(sizeof(struct f_path_mask)); $$->next = $2; $$->kind = PM_QUESTION; }
- | bgp_path_expr bgp_path_tail1	{ $$ = cfg_allocz(sizeof(struct f_path_mask)); $$->next = $2; $$->kind = PM_ASN_EXPR; $$->val = (uintptr_t) $1; }
+bgp_path_tail:
+   NUM bgp_path_tail		{ $$ = cfg_allocz(sizeof(struct f_path_mask)); $$->next = $2; $$->kind = PM_ASN; $$->val = $1; }
+ | NUM DDOT NUM bgp_path_tail	{ $$ = cfg_allocz(sizeof(struct f_path_mask)); $$->next = $4; $$->kind = PM_ASN_RANGE; $$->val = $1; $$->val2 = $3; }
+ | '*' bgp_path_tail		{ $$ = cfg_allocz(sizeof(struct f_path_mask)); $$->next = $2; $$->kind = PM_ASTERISK; }
+ | '?' bgp_path_tail		{ $$ = cfg_allocz(sizeof(struct f_path_mask)); $$->next = $2; $$->kind = PM_QUESTION; }
+ | bgp_path_expr bgp_path_tail	{ $$ = cfg_allocz(sizeof(struct f_path_mask)); $$->next = $2; $$->kind = PM_ASN_EXPR; $$->val = (uintptr_t) $1; }
  | 				{ $$ = NULL; }
  ;
 
-bgp_path_tail2:
-   NUM bgp_path_tail2 { $$ = cfg_allocz(sizeof(struct f_path_mask)); $$->next = $2; $$->kind = PM_ASN; $$->val = $1; }
- | '?' bgp_path_tail2 { $$ = cfg_allocz(sizeof(struct f_path_mask)); $$->next = $2; $$->kind = PM_ASTERISK; }
- | 		      { $$ = NULL; }
- ;
-
 constant:
    NUM    { $$ = f_new_inst(); $$->code = 'c'; $$->aux = T_INT;  $$->a2.i = $1; }
  | TRUE   { $$ = f_new_inst(); $$->code = 'c'; $$->aux = T_BOOL; $$->a2.i = 1;  }
diff --git a/nest/config.Y b/nest/config.Y
index 555c9e05..432dd9c7 100644
--- a/nest/config.Y
+++ b/nest/config.Y
@@ -71,7 +71,7 @@ CF_KEYWORDS(RECEIVE, LIMIT, ACTION, WARN, BLOCK, RESTART, DISABLE, KEEP, FILTERE
 CF_KEYWORDS(PASSWORD, FROM, PASSIVE, TO, ID, EVENTS, PACKETS, PROTOCOLS, INTERFACES)
 CF_KEYWORDS(ALGORITHM, KEYED, HMAC, MD5, SHA1, SHA256, SHA384, SHA512)
 CF_KEYWORDS(PRIMARY, STATS, COUNT, BY, FOR, COMMANDS, PREEXPORT, NOEXPORT, GENERATE)
-CF_KEYWORDS(LISTEN, BGP, V6ONLY, DUAL, ADDRESS, PORT, PASSWORDS, DESCRIPTION, SORTED)
+CF_KEYWORDS(BGP, PASSWORDS, DESCRIPTION, SORTED)
 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)
@@ -124,24 +124,6 @@ idval:
    }
  ;
 
-
-CF_ADDTO(conf, listen)
-
-listen: LISTEN BGP listen_opts ';' ;
-
-listen_opts:
-   /* Nothing */
- | listen_opts listen_opt
- ;
-
-listen_opt:
-   ADDRESS ipa { new_config->listen_bgp_addr = $2; }
- | PORT expr { new_config->listen_bgp_port = $2; }
- | V6ONLY { new_config->listen_bgp_flags = 0; }
- | DUAL { new_config->listen_bgp_flags = 1; }
- ;
-
-
 CF_ADDTO(conf, gr_opts)
 
 gr_opts: GRACEFUL RESTART WAIT expr ';' { new_config->gr_wait = $4; } ;

From dea9886454c1c0953b5977dd8a96718be465b962 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Fri, 8 Dec 2017 17:00:47 +0100
Subject: [PATCH 078/198] BGP: Link check just for single-hop

---
 proto/bgp/bgp.c    | 4 ++++
 proto/bgp/config.Y | 2 +-
 2 files changed, 5 insertions(+), 1 deletion(-)

diff --git a/proto/bgp/bgp.c b/proto/bgp/bgp.c
index a26c0cb9..f6d26639 100644
--- a/proto/bgp/bgp.c
+++ b/proto/bgp/bgp.c
@@ -1621,6 +1621,10 @@ bgp_postconfig(struct proto_config *CF)
   if (cf->multihop < 0)
     cf->multihop = internal ? 64 : 0;
 
+  /* Link check for single-hop BGP by default */
+  if (cf->check_link < 0)
+    cf->check_link = !cf->multihop;
+
 
   if (!cf->local_as)
     cf_error("Local AS number must be set");
diff --git a/proto/bgp/config.Y b/proto/bgp/config.Y
index 0bd4c554..6ce0f1aa 100644
--- a/proto/bgp/config.Y
+++ b/proto/bgp/config.Y
@@ -58,7 +58,7 @@ bgp_proto_start: proto_start BGP {
      BGP_CFG->gr_mode = BGP_GR_AWARE;
      BGP_CFG->gr_time = 120;
      BGP_CFG->setkey = 1;
-     BGP_CFG->check_link = 1;
+     BGP_CFG->check_link = -1;
    }
  ;
 

From ed1d853e5147376086e25f5edae9804cf242d6e0 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Fri, 8 Dec 2017 17:31:33 +0100
Subject: [PATCH 079/198] Filter: Remove old BGP path mask syntax from tests

---
 filter/test.conf | 10 ++--------
 1 file changed, 2 insertions(+), 8 deletions(-)

diff --git a/filter/test.conf b/filter/test.conf
index e25105d3..989dab14 100644
--- a/filter/test.conf
+++ b/filter/test.conf
@@ -589,14 +589,11 @@ function mkpath(int a; int b)
 
 function t_path()
 bgpmask pm1;
-bgpmask pm2;
 bgppath p2;
 {
-	pm1 =  / 4 3 2 1 /;
-	pm2 = [= 4 3 2 1 =];
+	pm1 = [= 4 3 2 1 =];
 
-	bt_assert(pm1 = pm2);
-	bt_assert(format(pm2) = "[= 4 3 2 1 =]");
+	bt_assert(format(pm1) = "[= 4 3 2 1 =]");
 
 	bt_assert(+empty+ = +empty+);
 	bt_assert(10 !~ +empty+);
@@ -609,17 +606,14 @@ bgppath p2;
 	bt_assert(format(p2) = "(path 4 3 2 1)");
 	bt_assert(p2.len = 4);
 	bt_assert(p2 ~ pm1);
-	bt_assert(p2 ~ pm2);
 	bt_assert(3 ~ p2);
 	bt_assert(p2 ~ [2, 10..20]);
 	bt_assert(p2 ~ [4, 10..20]);
 
 	p2 = prepend(p2, 5);
 	bt_assert(p2 !~ pm1);
-	bt_assert(p2 !~ pm2);
 	bt_assert(10 !~ p2);
 	bt_assert(p2 !~ [8, ten..(2*ten)]);
-	bt_assert(p2 ~  / ? 4 3 2 1 /);
 	bt_assert(p2 ~ [= * 4 3 * 1 =]);
 	bt_assert(p2 ~ [= (3+2) (2*2) 3 2 1 =]);
 	bt_assert(p2 ~ mkpath(5, 4));

From 7fc55925beb06059759294e0e9b7bae45465395f Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Sun, 10 Dec 2017 00:55:34 +0100
Subject: [PATCH 080/198] Several minor fixes

---
 doc/bird.sgml       | 34 ++++++++++++++++++------------
 lib/net.h           |  7 +++++--
 nest/iface.c        |  4 ++--
 nest/rt-table.c     |  4 +++-
 proto/bgp/attrs.c   | 18 ++++++++++++++--
 proto/bgp/bgp.c     | 50 ++++++++++++++++++++++++---------------------
 proto/bgp/bgp.h     |  2 ++
 proto/bgp/packets.c |  8 ++++----
 proto/ospf/config.Y |  2 +-
 9 files changed, 81 insertions(+), 48 deletions(-)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index c50d0cde..3ec6612a 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -2044,6 +2044,7 @@ avoid routing loops.
 <item> <rfc id="7911"> - Advertisement of Multiple Paths in BGP
 <item> <rfc id="7947"> - Internet Exchange BGP Route Server
 <item> <rfc id="8092"> - BGP Large Communities Attribute
+<item> <rfc id="8203"> - BGP Administrative Shutdown Communication
 </itemize>
 
 <sect1>Route selection rules
@@ -2258,6 +2259,20 @@ using the following configuration parameters:
 	related procedures. Note that even when disabled, BIRD can send route
 	refresh requests.  Default: on.
 
+	<tag><label id="bgp-graceful-restart">graceful restart <m/switch/|aware</tag>
+	When a BGP speaker restarts or crashes, neighbors will discard all
+	received paths from the speaker, which disrupts packet forwarding even
+	when the forwarding plane of the speaker remains intact. <rfc id="4724">
+	specifies an optional graceful restart mechanism to alleviate this
+	issue. This option controls the mechanism. It has three states:
+	Disabled, when no support is provided. Aware, when the graceful restart
+	support is announced and the support for restarting neighbors is
+	provided, but no local graceful restart is allowed (i.e. receiving-only
+	role). Enabled, when the full graceful restart support is provided
+	(i.e. both restarting and receiving role). Restarting role could be also
+	configured per-channel. Note that proper support for local graceful
+	restart requires also configuration of other protocols. Default: aware.
+
 	<tag><label id="bgp-graceful-restart-time">graceful restart time <m/number/</tag>
 	The restart time is announced in the BGP graceful restart capability
 	and specifies how long the neighbor would wait for the BGP session to
@@ -2487,19 +2502,12 @@ together with their appropriate channels follows.
 	TX direction. When active, all available routes accepted by the export
 	filter are advertised to the neighbor. Default: off.
 
-	<tag><label id="bgp-graceful-restart">graceful restart <m/switch/|aware</tag>
-	When a BGP speaker restarts or crashes, neighbors will discard all
-	received paths from the speaker, which disrupts packet forwarding even
-	when the forwarding plane of the speaker remains intact. <rfc
-	id="4724"> specifies an optional graceful restart mechanism to
-	alleviate this issue. This option controls the mechanism. It has three
-	states: Disabled, when no support is provided. Aware, when the graceful
-	restart support is announced and the support for restarting neighbors
-	is provided, but no local graceful restart is allowed (i.e.
-	receiving-only role). Enabled, when the full graceful restart
-	support is provided (i.e. both restarting and receiving role). Note
-	that proper support for local graceful restart requires also
-	configuration of other protocols.  Default: aware.
+	<tag><label id="bgp-graceful-restart-c">graceful restart <m/switch/</tag>
+	Although BGP graceful restart is configured mainly by protocol-wide
+	<ref id="bgp-graceful-restart" name="options">, it is possible to
+	configure restarting role per AFI/SAFI pair by this channel option.
+	The option is ignored if graceful restart is disabled by protocol-wide
+	option. Default: off in aware mode, on in full mode.
 </descrip>
 
 <sect1>Attributes
diff --git a/lib/net.h b/lib/net.h
index cae3a749..2138f275 100644
--- a/lib/net.h
+++ b/lib/net.h
@@ -230,11 +230,14 @@ static inline int net_type_match(const net_addr *a, u32 mask)
 static inline int net_is_ip(const net_addr *a)
 { return (a->type == NET_IP4) || (a->type == NET_IP6); }
 
+static inline int net_is_vpn(const net_addr *a)
+{ return (a->type == NET_VPN4) || (a->type == NET_VPN6); }
+
 static inline int net_is_roa(const net_addr *a)
 { return (a->type == NET_ROA4) || (a->type == NET_ROA6); }
 
-static inline int net_is_vpn(const net_addr *a)
-{ return (a->type == NET_VPN4) || (a->type == NET_VPN6); }
+static inline int net_is_flow(const net_addr *a)
+{ return (a->type == NET_FLOW4) || (a->type == NET_FLOW6); }
 
 
 static inline ip4_addr net4_prefix(const net_addr *a)
diff --git a/nest/iface.c b/nest/iface.c
index a876e9ff..303a52b2 100644
--- a/nest/iface.c
+++ b/nest/iface.c
@@ -849,7 +849,7 @@ if_show(void)
       else if (i->master_index)
 	bsprintf(mbuf, " master=#%u", i->master_index);
 
-      cli_msg(-1001, "%s %s (index=%d%s)", i->name, (i->flags & IF_UP) ? "Up" : "Down", i->index, mbuf);
+      cli_msg(-1001, "%s %s (index=%d%s)", i->name, (i->flags & IF_UP) ? "up" : "down", i->index, mbuf);
       if (!(i->flags & IF_MULTIACCESS))
 	type = "PtP";
       else
@@ -897,7 +897,7 @@ if_show_summary(void)
 	a6[0] = 0;
 
       cli_msg(-1005, "%-10s %-6s %-18s %s",
-	      i->name, (i->flags & IF_UP) ? "Up" : "Down", a4, a6);
+	      i->name, (i->flags & IF_UP) ? "up" : "down", a4, a6);
     }
   cli_msg(0, "");
 }
diff --git a/nest/rt-table.c b/nest/rt-table.c
index c42d3a97..b0dd6d3f 100644
--- a/nest/rt-table.c
+++ b/nest/rt-table.c
@@ -897,7 +897,9 @@ rte_validate(rte *e)
     return 0;
   }
 
-  c = net_classify(n->n.addr);
+  /* FIXME: better handling different nettypes */
+  c = !net_is_flow(n->n.addr) ?
+    net_classify(n->n.addr): (IADDR_HOST | SCOPE_UNIVERSE);
   if ((c < 0) || !(c & IADDR_HOST) || ((c & IADDR_SCOPE_MASK) <= SCOPE_LINK))
   {
     log(L_WARN "Ignoring bogus route %N received via %s",
diff --git a/proto/bgp/attrs.c b/proto/bgp/attrs.c
index 882ba44e..dea3c4a6 100644
--- a/proto/bgp/attrs.c
+++ b/proto/bgp/attrs.c
@@ -64,8 +64,6 @@
  * format - Optional hook that converts eattr to textual representation.
  */
 
-// XXXX review pool usage : c->c.proto->pool
-
 
 struct bgp_attr_desc {
   const char *name;
@@ -1175,6 +1173,22 @@ bgp_init_bucket_table(struct bgp_channel *c)
   c->withdraw_bucket = NULL;
 }
 
+void
+bgp_free_bucket_table(struct bgp_channel *c)
+{
+  HASH_FREE(c->bucket_hash);
+
+  struct bgp_bucket *b;
+  WALK_LIST_FIRST(b, c->bucket_queue)
+  {
+    rem_node(&b->send_node);
+    mb_free(b);
+  }
+
+  mb_free(c->withdraw_bucket);
+  c->withdraw_bucket = NULL;
+}
+
 static struct bgp_bucket *
 bgp_get_bucket(struct bgp_channel *c, ea_list *new)
 {
diff --git a/proto/bgp/bgp.c b/proto/bgp/bgp.c
index f6d26639..705c5e91 100644
--- a/proto/bgp/bgp.c
+++ b/proto/bgp/bgp.c
@@ -98,6 +98,7 @@
  * <item> <rfc id="7911"> - Advertisement of Multiple Paths in BGP
  * <item> <rfc id="7947"> - Internet Exchange BGP Route Server
  * <item> <rfc id="8092"> - BGP Large Communities Attribute
+ * <item> <rfc id="8203"> - BGP Administrative Shutdown Communication
  * </itemize>
 */
 
@@ -601,10 +602,6 @@ bgp_conn_leave_established_state(struct bgp_proto *p)
   BGP_TRACE(D_EVENTS, "BGP session closed");
   p->conn = NULL;
 
-  // XXXX free these tables to avoid memory leak during graceful restart
-  // bgp_free_prefix_table(p);
-  // bgp_free_bucket_table(p);
-
   if (p->p.proto_state == PS_UP)
     bgp_stop(p, 0, NULL, 0);
 }
@@ -664,6 +661,10 @@ bgp_handle_graceful_restart(struct bgp_proto *p)
   struct bgp_channel *c;
   WALK_LIST(c, p->p.channels)
   {
+    /* FIXME: perhaps check for channel state instead of disabled flag? */
+    if (c->c.disabled)
+      continue;
+
     if (c->gr_ready)
     {
       if (c->gr_active)
@@ -679,6 +680,13 @@ bgp_handle_graceful_restart(struct bgp_proto *p)
       rt_refresh_begin(c->c.table, &c->c);
       rt_refresh_end(c->c.table, &c->c);
     }
+
+    /* Reset bucket and prefix tables */
+    bgp_free_bucket_table(c);
+    bgp_free_prefix_table(c);
+    bgp_init_bucket_table(c);
+    bgp_init_prefix_table(c);
+    c->packets_to_send = 0;
   }
 
   proto_notify_state(&p->p, PS_START);
@@ -1315,7 +1323,7 @@ bgp_start(struct proto *P)
   p->source_addr = p->cf->local_ip;
   p->link_addr = IPA_NONE;
 
-  /* XXXX */
+  /* Lock all channels when in GR recovery mode */
   if (p->p.gr_recovery && p->cf->gr_mode)
   {
     struct bgp_channel *c;
@@ -1546,10 +1554,9 @@ bgp_channel_shutdown(struct channel *C)
 {
   struct bgp_channel *c = (void *) C;
 
-  /* XXXX: cleanup bucket and prefix tables */
-
   c->next_hop_addr = IPA_NONE;
   c->link_addr = IPA_NONE;
+  c->packets_to_send = 0;
 }
 
 static void
@@ -2064,17 +2071,12 @@ bgp_show_proto_info(struct proto *P)
     bgp_show_capabilities(p, p->conn->local_caps);
     cli_msg(-1006, "    Neighbor capabilities");
     bgp_show_capabilities(p, p->conn->remote_caps);
-/* XXXX
-      cli_msg(-1006, "    Session:          %s%s%s%s%s%s%s%s",
-	      p->is_internal ? "internal" : "external",
-	      p->cf->multihop ? " multihop" : "",
-	      p->rr_client ? " route-reflector" : "",
-	      p->rs_client ? " route-server" : "",
-	      p->as4_session ? " AS4" : "",
-	      p->add_path_rx ? " add-path-rx" : "",
-	      p->add_path_tx ? " add-path-tx" : "",
-	      p->ext_messages ? " ext-messages" : "");
-*/
+    cli_msg(-1006, "    Session:          %s%s%s%s%s",
+	    p->is_internal ? "internal" : "external",
+	    p->cf->multihop ? " multihop" : "",
+	    p->rr_client ? " route-reflector" : "",
+	    p->rs_client ? " route-server" : "",
+	    p->as4_session ? " AS4" : "");
     cli_msg(-1006, "    Source address:   %I", p->source_addr);
     cli_msg(-1006, "    Hold timer:       %t/%u",
 	    tm_remains(p->conn->hold_timer), p->conn->hold_time);
@@ -2091,16 +2093,18 @@ bgp_show_proto_info(struct proto *P)
   }
 
   {
-    /* XXXX ?? */
     struct bgp_channel *c;
     WALK_LIST(c, p->p.channels)
     {
       channel_show_info(&c->c);
 
-      if (ipa_zero(c->link_addr))
-	cli_msg(-1006, "    BGP Next hop:   %I", c->next_hop_addr);
-      else
-	cli_msg(-1006, "    BGP Next hop:   %I %I", c->next_hop_addr, c->link_addr);
+      if (c->c.channel_state == CS_UP)
+      {
+	if (ipa_zero(c->link_addr))
+	  cli_msg(-1006, "    BGP Next hop:   %I", c->next_hop_addr);
+	else
+	  cli_msg(-1006, "    BGP Next hop:   %I %I", c->next_hop_addr, c->link_addr);
+      }
 
       if (c->igp_table_ip4)
 	cli_msg(-1006, "    IGP IPv4 table: %s", c->igp_table_ip4->name);
diff --git a/proto/bgp/bgp.h b/proto/bgp/bgp.h
index 40c4b3f0..1310582b 100644
--- a/proto/bgp/bgp.h
+++ b/proto/bgp/bgp.h
@@ -492,11 +492,13 @@ int bgp_encode_attrs(struct bgp_write_state *s, ea_list *attrs, byte *buf, byte
 ea_list * bgp_decode_attrs(struct bgp_parse_state *s, byte *data, uint len);
 
 void bgp_init_bucket_table(struct bgp_channel *c);
+void bgp_free_bucket_table(struct bgp_channel *c);
 void bgp_free_bucket(struct bgp_channel *c, struct bgp_bucket *b);
 void bgp_defer_bucket(struct bgp_channel *c, struct bgp_bucket *b);
 void bgp_withdraw_bucket(struct bgp_channel *c, struct bgp_bucket *b);
 
 void bgp_init_prefix_table(struct bgp_channel *c);
+void bgp_free_prefix_table(struct bgp_channel *c);
 void bgp_free_prefix(struct bgp_channel *c, struct bgp_prefix *bp);
 
 int bgp_rte_better(struct rte *, struct rte *);
diff --git a/proto/bgp/packets.c b/proto/bgp/packets.c
index aa69bdd6..ca942880 100644
--- a/proto/bgp/packets.c
+++ b/proto/bgp/packets.c
@@ -1637,8 +1637,8 @@ bgp_decode_nlri_flow4(struct bgp_parse_state *s, byte *pos, uint len, rta *a)
     uint pxlen = data[1];
 
     // FIXME: Use some generic function
-    memcpy(&px, data, BYTES(pxlen));
-    px = ip4_and(px, ip4_mkmask(pxlen));
+    memcpy(&px, data+2, BYTES(pxlen));
+    px = ip4_and(ip4_ntoh(px), ip4_mkmask(pxlen));
 
     /* Prepare the flow */
     net_addr *n = alloca(sizeof(struct net_addr_flow4) + flen);
@@ -1729,8 +1729,8 @@ bgp_decode_nlri_flow6(struct bgp_parse_state *s, byte *pos, uint len, rta *a)
     uint pxlen = data[1];
 
     // FIXME: Use some generic function
-    memcpy(&px, data, BYTES(pxlen));
-    px = ip6_and(px, ip6_mkmask(pxlen));
+    memcpy(&px, data+2, BYTES(pxlen));
+    px = ip6_and(ip6_ntoh(px), ip6_mkmask(pxlen));
 
     /* Prepare the flow */
     net_addr *n = alloca(sizeof(struct net_addr_flow6) + flen);
diff --git a/proto/ospf/config.Y b/proto/ospf/config.Y
index 04d5e2e9..ce409a3a 100644
--- a/proto/ospf/config.Y
+++ b/proto/ospf/config.Y
@@ -499,7 +499,7 @@ CF_ADDTO(dynamic_attr, OSPF_TAG { $$ = f_new_dynamic_attr(EAF_TYPE_INT | EAF_TEM
 CF_ADDTO(dynamic_attr, OSPF_ROUTER_ID { $$ = f_new_dynamic_attr(EAF_TYPE_ROUTER_ID | EAF_TEMP, T_QUAD, EA_OSPF_ROUTER_ID); })
 
 CF_CLI_HELP(SHOW OSPF, ..., [[Show information about OSPF protocol]]);
-CF_CLI(SHOW OSPF, optsym, [<name>], [[Show information about OSPF protocol XXX]])
+CF_CLI(SHOW OSPF, optsym, [<name>], [[Show information about OSPF protocol]])
 { ospf_sh(proto_get_named($3, &proto_ospf)); };
 
 CF_CLI(SHOW OSPF NEIGHBORS, optsym opttext, [<name>] [\"<interface>\"], [[Show information about OSPF neighbors]])

From a32a7b58cebc2838de2ffd94e679d3e1ec9d493a Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Sun, 10 Dec 2017 13:16:31 +0100
Subject: [PATCH 081/198] Lib: Fix macro/keyword collisions

Old code breaks with some versions of bison
---
 lib/buffer.h          | 6 +++++-
 lib/flowspec.h        | 2 +-
 lib/timer.h           | 2 +-
 proto/static/static.h | 2 +-
 4 files changed, 8 insertions(+), 4 deletions(-)

diff --git a/lib/buffer.h b/lib/buffer.h
index 6fc18852..cd9bab86 100644
--- a/lib/buffer.h
+++ b/lib/buffer.h
@@ -13,10 +13,14 @@
 #include "lib/resource.h"
 #include "sysdep/config.h"
 
-#define BUFFER(type)		struct { type *data; uint used, size; }
+#define BUFFER_(type)		struct { type *data; uint used, size; }
 #define BUFFER_TYPE(v)		typeof(* (v).data)
 #define BUFFER_SIZE(v)		((v).size * sizeof(* (v).data))
 
+#ifndef PARSER
+#define BUFFER(type) BUFFER_(type)
+#endif
+
 #define BUFFER_INIT(v,pool,isize)					\
   ({									\
     (v).used = 0;							\
diff --git a/lib/flowspec.h b/lib/flowspec.h
index 4fe23da1..fa90c70d 100644
--- a/lib/flowspec.h
+++ b/lib/flowspec.h
@@ -91,7 +91,7 @@ const byte *flow6_next_part(const byte *pos, const byte *end);
 
 /* A data structure for keep a state of flow builder */
 struct flow_builder {
-  BUFFER(byte) data;
+  BUFFER_(byte) data;
   enum flow_type this_type;
   enum flow_type last_type;
   u16 last_op_offset;			/* Position of last operator in data.data */
diff --git a/lib/timer.h b/lib/timer.h
index eeb7dcb7..ed8f0d02 100644
--- a/lib/timer.h
+++ b/lib/timer.h
@@ -30,7 +30,7 @@ typedef struct timer
 
 struct timeloop
 {
-  BUFFER(timer *) timers;
+  BUFFER_(timer *) timers;
   btime last_time;
   btime real_time;
 };
diff --git a/proto/static/static.h b/proto/static/static.h
index b202c0b1..a3c30b87 100644
--- a/proto/static/static.h
+++ b/proto/static/static.h
@@ -25,7 +25,7 @@ struct static_proto {
   struct proto p;
 
   struct event *event;			/* Event for announcing updated routes */
-  BUFFER(struct static_route *) marked;	/* Routes marked for reannouncement */
+  BUFFER_(struct static_route *) marked; /* Routes marked for reannouncement */
   rtable *igp_table_ip4;		/* Table for recursive IPv4 next hop lookups */
   rtable *igp_table_ip6;		/* Table for recursive IPv6 next hop lookups */
 };

From 67a2eb9177fe0b8a6854775a4fad1f7768f6a94a Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Sun, 10 Dec 2017 13:18:36 +0100
Subject: [PATCH 082/198] Lib: Check size of nets

---
 lib/net.c          | 18 ++++++++++++++++++
 lib/net.h          |  6 ++++--
 sysdep/unix/main.c |  1 +
 3 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/lib/net.c b/lib/net.c
index a00ff272..9335b78f 100644
--- a/lib/net.c
+++ b/lib/net.c
@@ -288,3 +288,21 @@ net_in_netX(const net_addr *a, const net_addr *n)
 
   return (net_pxlen(n) <= net_pxlen(a)) && ipa_in_netX(net_prefix(a), n);
 }
+
+#define CHECK_NET(T,S) \
+  ({ if (sizeof(T) != S) die("sizeof %s is %d/%d", #T, (int) sizeof(T), S); })
+
+void
+net_init(void)
+{
+  CHECK_NET(net_addr,		24);
+  CHECK_NET(net_addr_ip4,	 8);
+  CHECK_NET(net_addr_ip6,	20);
+  CHECK_NET(net_addr_vpn4,	16);
+  CHECK_NET(net_addr_vpn6,	32);
+  CHECK_NET(net_addr_roa4,	16);
+  CHECK_NET(net_addr_roa6,	28);
+  CHECK_NET(net_addr_flow4,	 8);
+  CHECK_NET(net_addr_flow6,	20);
+  CHECK_NET(net_addr_mpls,	 8);
+}
diff --git a/lib/net.h b/lib/net.h
index 2138f275..4b2077ae 100644
--- a/lib/net.h
+++ b/lib/net.h
@@ -45,7 +45,7 @@ typedef struct net_addr {
   u8 type;
   u8 pxlen;
   u16 length;
-  u8 data[16];
+  u8 data[20];
   u64 align[0];
 } net_addr;
 
@@ -76,6 +76,7 @@ typedef struct net_addr_vpn6 {
   u8 pxlen;
   u16 length;
   ip6_addr prefix;
+  u32 padding;
   u64 rd;
 } net_addr_vpn6;
 
@@ -152,7 +153,7 @@ extern const u16 net_max_text_length[];
   ((net_addr_vpn4) { NET_VPN4, pxlen, sizeof(net_addr_vpn4), prefix, rd })
 
 #define NET_ADDR_VPN6(prefix,pxlen,rd) \
-  ((net_addr_vpn6) { NET_VPN6, pxlen, sizeof(net_addr_vpn6), prefix, rd })
+  ((net_addr_vpn6) { NET_VPN6, pxlen, sizeof(net_addr_vpn6), prefix, 0, rd })
 
 #define NET_ADDR_ROA4(prefix,pxlen,max_pxlen,asn) \
   ((net_addr_roa4) { NET_ROA4, pxlen, sizeof(net_addr_roa4), prefix, max_pxlen, asn })
@@ -544,5 +545,6 @@ static inline int net_in_net_ip6(const net_addr_ip6 *a, const net_addr_ip6 *b)
 int ipa_in_netX(const ip_addr A, const net_addr *N);
 int net_in_netX(const net_addr *A, const net_addr *N);
 
+void net_init(void);
 
 #endif
diff --git a/sysdep/unix/main.c b/sysdep/unix/main.c
index e3252b43..2251d3fb 100644
--- a/sysdep/unix/main.c
+++ b/sysdep/unix/main.c
@@ -820,6 +820,7 @@ main(int argc, char **argv)
     log_init_debug("");
   log_switch(debug_flag, NULL, NULL);
 
+  net_init();
   resource_init();
   timer_init();
   olock_init();

From 6b0f5f68a8ae35b05abb144cf998bf537dfa283b Mon Sep 17 00:00:00 2001
From: Maria Jan Matejka <mq@jmq.cz>
Date: Mon, 25 Sep 2017 13:00:05 +0200
Subject: [PATCH 083/198] Switchoff for MPLS in kernel.

---
 aclocal.m4              | 35 ++++++++++++++++++++
 configure.ac            | 33 ++++++++++++-------
 sysdep/linux/lwtunnel.h | 45 --------------------------
 sysdep/linux/netlink.c  | 71 ++++++++++++++++++++++++++---------------
 sysdep/unix/krt.c       |  4 +++
 5 files changed, 106 insertions(+), 82 deletions(-)
 delete mode 100644 sysdep/linux/lwtunnel.h

diff --git a/aclocal.m4 b/aclocal.m4
index 365bfa81..8160e539 100644
--- a/aclocal.m4
+++ b/aclocal.m4
@@ -31,6 +31,41 @@ AC_DEFUN([BIRD_CHECK_PTHREADS],
   CFLAGS="$bird_tmp_cflags"
 ])
 
+AC_DEFUN([BIRD_CHECK_MPLS_KERNEL],
+[
+  AC_CACHE_CHECK(
+    [for Linux MPLS headers],
+    [bird_cv_mpls_kernel],
+    [
+      AC_COMPILE_IFELSE(
+	[
+	  AC_LANG_PROGRAM(
+	    [
+	      #include <linux/lwtunnel.h>
+	      #include <linux/netlink.h>
+	      #include <linux/rtnetlink.h>
+	      #include <sys/socket.h>
+	      void t(int arg);
+	    ],
+	    [
+	      t(AF_MPLS);
+	      t(RTA_VIA);
+	      t(RTA_NEWDST);
+	      t(RTA_ENCAP_TYPE);
+	      t(RTA_ENCAP);
+	      struct rtvia rtvia;
+	      t(LWTUNNEL_ENCAP_MPLS);
+	    ]
+	  )
+	],
+	[bird_cv_mpls_kernel=yes],
+	[bird_cv_mpls_kernel=no]
+      )
+    ]
+  )
+])
+
+
 AC_DEFUN([BIRD_CHECK_GCC_OPTION],
 [
   bird_tmp_cflags="$CFLAGS"
diff --git a/configure.ac b/configure.ac
index 5514fb56..68e68282 100644
--- a/configure.ac
+++ b/configure.ac
@@ -36,6 +36,12 @@ AC_ARG_ENABLE([libssh],
   [enable_libssh=try]
 )
 
+AC_ARG_ENABLE([mpls-kernel],
+  [AS_HELP_STRING([--enable-mpls-kernel], [enable MPLS support in kernel protocol @<:@try@:>@])],
+  [],
+  [enable_mpls_kernel=try]
+)
+
 AC_ARG_WITH([protocols],
   [AS_HELP_STRING([--with-protocols=LIST], [include specified routing protocols @<:@all@:>@])],
   [],
@@ -239,6 +245,20 @@ if test "$enable_libssh" != no ; then
   fi
 fi
 
+if test "$enable_mpls_kernel" != no ; then
+  BIRD_CHECK_MPLS_KERNEL
+
+  if test "$bird_cv_mpls_kernel" = yes ; then
+    AC_DEFINE([HAVE_MPLS_KERNEL], [1], [Define to 1 if kernel is MPLS capable])
+  elif test "$enable_mpls_kernel" = yes ; then
+    AC_MSG_ERROR([Kernel MPLS support not found.])
+  fi
+
+  if test "$enable_mpls_kernel" = try ; then
+    enable_mpls_kernel="$bird_cv_mpls_kernel"
+  fi
+fi
+
 all_protocols="$proto_bfd babel bgp ospf pipe radv rip $proto_rpki static"
 
 all_protocols=`echo $all_protocols | sed 's/ /,/g'`
@@ -287,18 +307,6 @@ esac
 AC_CHECK_HEADERS_ONCE([alloca.h syslog.h])
 AC_CHECK_MEMBERS([struct sockaddr.sa_len], [], [], [#include <sys/socket.h>])
 
-AC_CHECK_HEADERS([linux/lwtunnel.h],
-  [AC_DEFINE([HAVE_LWTUNNEL], [1], [Define to 1 if you have the <linux/lwtunnel.h> header file.])],
-  [],
-  [] dnl Force new AC_CHECK_HEADER semantics
-)
-
-AC_CHECK_MEMBERS([struct rtvia.rtvia_family],
-  [AC_DEFINE([HAVE_STRUCT_RTVIA], [1], [Define to 1 if you have rtvia structure.])],
-  [],
-  [#include <linux/rtnetlink.h>]
-)
-
 AC_C_BIGENDIAN(
   [AC_DEFINE([CPU_BIG_ENDIAN], [1], [Define to 1 if cpu is big endian])],
   [AC_DEFINE([CPU_LITTLE_ENDIAN], [1], [Define to 1 if cpu is little endian])],
@@ -396,6 +404,7 @@ AC_MSG_RESULT([        System configuration:	$sysdesc])
 AC_MSG_RESULT([        Debugging:		$enable_debug])
 AC_MSG_RESULT([        POSIX threads:		$enable_pthreads])
 AC_MSG_RESULT([        Routing protocols:	$protocols])
+AC_MSG_RESULT([        Kernel MPLS support:	$enable_mpls_kernel])
 AC_MSG_RESULT([        Client:			$enable_client])
 
 rm -f $objdir/.*-stamp
diff --git a/sysdep/linux/lwtunnel.h b/sysdep/linux/lwtunnel.h
deleted file mode 100644
index 7cea3cbf..00000000
--- a/sysdep/linux/lwtunnel.h
+++ /dev/null
@@ -1,45 +0,0 @@
-#ifndef _LWTUNNEL_H_
-#define _LWTUNNEL_H_
-
-#include <linux/types.h>
-
-enum lwtunnel_encap_types {
-	LWTUNNEL_ENCAP_NONE,
-	LWTUNNEL_ENCAP_MPLS,
-	LWTUNNEL_ENCAP_IP,
-	LWTUNNEL_ENCAP_ILA,
-	LWTUNNEL_ENCAP_IP6,
-	__LWTUNNEL_ENCAP_MAX,
-};
-
-#define LWTUNNEL_ENCAP_MAX (__LWTUNNEL_ENCAP_MAX - 1)
-
-enum lwtunnel_ip_t {
-	LWTUNNEL_IP_UNSPEC,
-	LWTUNNEL_IP_ID,
-	LWTUNNEL_IP_DST,
-	LWTUNNEL_IP_SRC,
-	LWTUNNEL_IP_TTL,
-	LWTUNNEL_IP_TOS,
-	LWTUNNEL_IP_FLAGS,
-	LWTUNNEL_IP_PAD,
-	__LWTUNNEL_IP_MAX,
-};
-
-#define LWTUNNEL_IP_MAX (__LWTUNNEL_IP_MAX - 1)
-
-enum lwtunnel_ip6_t {
-	LWTUNNEL_IP6_UNSPEC,
-	LWTUNNEL_IP6_ID,
-	LWTUNNEL_IP6_DST,
-	LWTUNNEL_IP6_SRC,
-	LWTUNNEL_IP6_HOPLIMIT,
-	LWTUNNEL_IP6_TC,
-	LWTUNNEL_IP6_FLAGS,
-	LWTUNNEL_IP6_PAD,
-	__LWTUNNEL_IP6_MAX,
-};
-
-#define LWTUNNEL_IP6_MAX (__LWTUNNEL_IP6_MAX - 1)
-
-#endif /* _LWTUNNEL_H_ */
diff --git a/sysdep/linux/netlink.c b/sysdep/linux/netlink.c
index ca5c79a4..afb9dc8b 100644
--- a/sysdep/linux/netlink.c
+++ b/sysdep/linux/netlink.c
@@ -30,14 +30,12 @@
 
 #include <asm/types.h>
 #include <linux/if.h>
-#ifdef HAVE_LWTUNNEL
-#include <linux/lwtunnel.h>
-#else
-#include "sysdep/linux/lwtunnel.h"
-#endif
 #include <linux/netlink.h>
 #include <linux/rtnetlink.h>
 
+#ifdef HAVE_MPLS_KERNEL
+#include <linux/lwtunnel.h>
+#endif
 
 #ifndef MSG_TRUNC			/* Hack: Several versions of glibc miss this one :( */
 #define MSG_TRUNC 0x20
@@ -59,13 +57,6 @@
 #define RTA_VIA	 18
 #endif
 
-#ifndef HAVE_STRUCT_RTVIA
-struct rtvia {
-	unsigned short	rtvia_family;
-	u8		rtvia_addr[0];
-};
-#endif
-
 #ifndef RTA_NEWDST
 #define RTA_NEWDST  19
 #endif
@@ -361,9 +352,11 @@ static struct nl_want_attrs nexthop_attr_want6[BIRD_RTA_MAX] = {
   [RTA_ENCAP]	  = { 1, 0, 0 },
 };
 
+#ifdef HAVE_MPLS_KERNEL
 static struct nl_want_attrs encap_mpls_want[BIRD_RTA_MAX] = {
   [RTA_DST]       = { 1, 0, 0 },
 };
+#endif
 
 static struct nl_want_attrs rtm_attr_want4[BIRD_RTA_MAX] = {
   [RTA_DST]	  = { 1, 1, sizeof(ip4_addr) },
@@ -394,6 +387,7 @@ static struct nl_want_attrs rtm_attr_want6[BIRD_RTA_MAX] = {
   [RTA_ENCAP]	  = { 1, 0, 0 },
 };
 
+#ifdef HAVE_MPLS_KERNEL
 static struct nl_want_attrs rtm_attr_want_mpls[BIRD_RTA_MAX] = {
   [RTA_DST]	  = { 1, 1, sizeof(u32) },
   [RTA_IIF]	  = { 1, 1, sizeof(u32) },
@@ -405,6 +399,7 @@ static struct nl_want_attrs rtm_attr_want_mpls[BIRD_RTA_MAX] = {
   [RTA_VIA]	  = { 1, 0, 0 },
   [RTA_NEWDST]	  = { 1, 0, 0 },
 };
+#endif
 
 
 static int
@@ -456,6 +451,7 @@ static inline ip_addr rta_get_ipa(struct rtattr *a)
     return ipa_from_ip6(rta_get_ip6(a));
 }
 
+#ifdef HAVE_MPLS_KERNEL
 static inline ip_addr rta_get_via(struct rtattr *a)
 {
   struct rtvia *v = RTA_DATA(a);
@@ -474,6 +470,7 @@ static inline int rta_get_mpls(struct rtattr *a, u32 *stack)
 
   return mpls_get(RTA_DATA(a), RTA_PAYLOAD(a) & ~0x3, stack);
 }
+#endif
 
 struct rtattr *
 nl_add_attr(struct nlmsghdr *h, uint bufsize, uint code, const void *data, uint dlen)
@@ -542,6 +539,7 @@ nl_add_attr_ipa(struct nlmsghdr *h, uint bufsize, int code, ip_addr ipa)
     nl_add_attr_ip6(h, bufsize, code, ipa_to_ip6(ipa));
 }
 
+#ifdef HAVE_MPLS_KERNEL
 static inline void
 nl_add_attr_mpls(struct nlmsghdr *h, uint bufsize, int code, int len, u32 *stack)
 {
@@ -583,6 +581,7 @@ nl_add_attr_via(struct nlmsghdr *h, uint bufsize, ip_addr ipa)
 
   nl_close_attr(h, nest);
 }
+#endif
 
 static inline struct rtnexthop *
 nl_open_nexthop(struct nlmsghdr *h, uint bufsize)
@@ -605,8 +604,9 @@ nl_close_nexthop(struct nlmsghdr *h, struct rtnexthop *nh)
 }
 
 static inline void
-nl_add_nexthop(struct nlmsghdr *h, uint bufsize, struct nexthop *nh, int af)
+nl_add_nexthop(struct nlmsghdr *h, uint bufsize, struct nexthop *nh, int af UNUSED)
 {
+#ifdef HAVE_MPLS_KERNEL
   if (nh->labels > 0)
     if (af == AF_MPLS)
       nl_add_attr_mpls(h, bufsize, RTA_NEWDST, nh->labels, nh->label);
@@ -618,6 +618,11 @@ nl_add_nexthop(struct nlmsghdr *h, uint bufsize, struct nexthop *nh, int af)
       nl_add_attr_via(h, bufsize, nh->gw);
     else
       nl_add_attr_ipa(h, bufsize, RTA_GATEWAY, nh->gw);
+#else
+
+  if (ipa_nonzero(nh->gw))
+    nl_add_attr_ipa(h, bufsize, RTA_GATEWAY, nh->gw);
+#endif
 }
 
 static void
@@ -716,20 +721,22 @@ nl_parse_multipath(struct krt_proto *p, struct rtattr *ra, int af)
       else
 	rv->gw = IPA_NONE;
 
+#ifdef HAVE_MPLS_KERNEL
       if (a[RTA_ENCAP_TYPE])
-	{
-	  if (rta_get_u16(a[RTA_ENCAP_TYPE]) != LWTUNNEL_ENCAP_MPLS) {
-	    log(L_WARN "KRT: Unknown encapsulation method %d in multipath", rta_get_u16(a[RTA_ENCAP_TYPE]));
-	    return NULL;
-	  }
-
-	  struct rtattr *enca[BIRD_RTA_MAX];
-	  nl_attr_len = RTA_PAYLOAD(a[RTA_ENCAP]);
-	  nl_parse_attrs(RTA_DATA(a[RTA_ENCAP]), encap_mpls_want, enca, sizeof(enca));
-	  rv->labels = rta_get_mpls(enca[RTA_DST], rv->label);
-	  break;
+      {
+	if (rta_get_u16(a[RTA_ENCAP_TYPE]) != LWTUNNEL_ENCAP_MPLS) {
+	  log(L_WARN "KRT: Unknown encapsulation method %d in multipath", rta_get_u16(a[RTA_ENCAP_TYPE]));
+	  return NULL;
 	}
 
+	struct rtattr *enca[BIRD_RTA_MAX];
+	nl_attr_len = RTA_PAYLOAD(a[RTA_ENCAP]);
+	nl_parse_attrs(RTA_DATA(a[RTA_ENCAP]), encap_mpls_want, enca, sizeof(enca));
+	rv->labels = rta_get_mpls(enca[RTA_DST], rv->label);
+	break;
+      }
+#endif
+
 
       len -= NLMSG_ALIGN(nh->rtnh_len);
       nh = RTNH_NEXT(nh);
@@ -1216,12 +1223,14 @@ nl_send_route(struct krt_proto *p, rte *e, struct ea_list *eattrs, int op, int d
   r->r.rtm_dst_len = net_pxlen(net->n.addr);
   r->r.rtm_protocol = RTPROT_BIRD;
   r->r.rtm_scope = RT_SCOPE_NOWHERE;
+#ifdef HAVE_MPLS_KERNEL
   if (p->af == AF_MPLS)
   {
     u32 label = net_mpls(net->n.addr);
     nl_add_attr_mpls(&r->h, rsize, RTA_DST, 1, &label);
   }
   else
+#endif
     nl_add_attr_ipa(&r->h, rsize, RTA_DST, net_prefix(net->n.addr));
 
   /*
@@ -1490,6 +1499,7 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h)
 	net_fill_ip6(&dst, IP6_NONE, 0);
       break;
 
+#ifdef HAVE_MPLS_KERNEL
     case AF_MPLS:
       if (!nl_parse_attrs(RTM_RTA(i), rtm_attr_want_mpls, a, sizeof(a)))
 	return;
@@ -1502,6 +1512,7 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h)
 
       net_fill_mpls(&dst, rta_mpls_stack[0]);
       break;
+#endif
 
     default:
       return;
@@ -1595,11 +1606,17 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h)
 	  return;
 	}
 
-      if ((i->rtm_family != AF_MPLS) && a[RTA_GATEWAY] || (i->rtm_family == AF_MPLS) && a[RTA_VIA])
+      if ((i->rtm_family != AF_MPLS) && a[RTA_GATEWAY]
+#ifdef HAVE_MPLS_KERNEL
+	  || (i->rtm_family == AF_MPLS) && a[RTA_VIA]
+#endif
+	  )
 	{
+#ifdef HAVE_MPLS_KERNEL
 	  if (i->rtm_family == AF_MPLS)
 	    ra->nh.gw = rta_get_via(a[RTA_VIA]);
 	  else
+#endif
 	    ra->nh.gw = rta_get_ipa(a[RTA_GATEWAY]);
 
 	  /* Silently skip strange 6to4 routes */
@@ -1637,6 +1654,7 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h)
       return;
     }
 
+#ifdef HAVE_MPLS_KERNEL
   int labels = 0;
   if ((i->rtm_family == AF_MPLS) && a[RTA_NEWDST] && !ra->nh.next)
     labels = rta_get_mpls(a[RTA_NEWDST], ra->nh.label);
@@ -1666,6 +1684,7 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h)
   }
   else
     ra->nh.labels = labels;
+#endif
 
   rte *e = rte_get_temp(ra);
   e->net = net;
@@ -1816,6 +1835,7 @@ krt_do_scan(struct krt_proto *p UNUSED)	/* CONFIG_ALL_TABLES_AT_ONCE => p is NUL
       log(L_DEBUG "nl_scan_fire: Unknown packet received (type=%d)", h->nlmsg_type);
   nl_parse_end(&s);
 
+#ifdef HAVE_MPLS_KERNEL
   nl_parse_begin(&s, 1, 1);
   nl_request_dump(AF_MPLS, RTM_GETROUTE);
   while (h = nl_get_scan())
@@ -1824,6 +1844,7 @@ krt_do_scan(struct krt_proto *p UNUSED)	/* CONFIG_ALL_TABLES_AT_ONCE => p is NUL
     else
       log(L_DEBUG "nl_scan_fire: Unknown packet received (type=%d)", h->nlmsg_type);
   nl_parse_end(&s);
+#endif
 }
 
 /*
diff --git a/sysdep/unix/krt.c b/sysdep/unix/krt.c
index 052f210a..29d2d01e 100644
--- a/sysdep/unix/krt.c
+++ b/sysdep/unix/krt.c
@@ -1211,7 +1211,11 @@ struct protocol proto_unix_kernel = {
   .template =		"kernel%d",
   .attr_class =		EAP_KRT,
   .preference =		DEF_PREF_INHERITED,
+#ifdef HAVE_MPLS_KERNEL
   .channel_mask =	NB_IP | NB_MPLS,
+#else
+  .channel_mask =	NB_IP,
+#endif
   .proto_size =		sizeof(struct krt_proto),
   .config_size =	sizeof(struct krt_config),
   .preconfig =		krt_preconfig,

From cf3e3845c205dacf2720cdb76a9a004a816d2e6f Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Sun, 10 Dec 2017 22:47:38 +0100
Subject: [PATCH 084/198] Doc: Documentation update

---
 doc/bird.sgml     | 339 +++++++++++++++++++++++-----------------------
 sysdep/unix/krt.h |   2 +-
 2 files changed, 167 insertions(+), 174 deletions(-)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index 3ec6612a..8309e051 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -91,6 +91,7 @@ supports:
 	<item>the Border Gateway Protocol (BGPv4)
 	<item>the Routing Information Protocol (RIPv2, RIPng)
 	<item>the Open Shortest Path First protocol (OSPFv2, OSPFv3)
+	<item>the Babel Routing Protocol
 	<item>the Router Advertisements for IPv6 hosts
 	<item>a virtual protocol for exchange of routes between different
 		routing tables on a single host
@@ -228,17 +229,17 @@ default tables -- <cf/master4/ for IPv4 routes and <cf/master6/ for IPv6 routes.
 Other tables must be explicitly configured.
 
 <p>
-These routing tables are not kernel forwarding tables. No forwarding is done in BIRD.
-If you want to forward packets using the routes in BIRD's tables, you may
-use the Kernel protocol (see below) to synchronize them with kernel's FIBs.
+These routing tables are not kernel forwarding tables. No forwarding is done by
+BIRD. If you want to forward packets using the routes in BIRD tables, you may
+use the Kernel protocol (see below) to synchronize them with kernel FIBs.
 
 <p>
-Every nettype defines a (kind of) primary key on routes. Every route source
-can supply one route for every possible primary key; new route announcement
-replaces the old route from the same source, keeping other routes intact.
-BIRD always chooses the best route for every primary key among the known
-routes and keeps the others as suboptimal. When the best route is retracted,
-BIRD re-runs the best route selection algorithm to find the current best route.
+Every nettype defines a (kind of) primary key on routes. Every route source can
+supply one route for every possible primary key; new route announcement replaces
+the old route from the same source, keeping other routes intact. BIRD always
+chooses the best route for each primary key among the known routes and keeps the
+others as suboptimal. When the best route is retracted, BIRD re-runs the best
+route selection algorithm to find the current best route.
 
 <p>
 The global best route selection algorithm is (roughly) as follows:
@@ -246,12 +247,12 @@ The global best route selection algorithm is (roughly) as follows:
 <itemize>
 	<item>Preferences of the routes are compared.
 	<item>Source protocol instance preferences are compared.
-	<item>If source protocols are different (e.g. BGP vs. OSPF), result of the algorithm is undefined.
 	<item>If source protocols are the same (e.g. BGP vs. BGP), the protocol's route selection algorithm is invoked.
+	<item>If source protocols are different (e.g. BGP vs. OSPF), result of the algorithm is undefined.
 </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
+<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
 activated, these lists of entries are kept completely sorted (according to
 preference or some protocol-dependent metric). This is needed for some features
 of some protocols (e.g. <cf/secondary/ option of BGP protocol, which allows to
@@ -267,9 +268,11 @@ is that it is slightly more computationally expensive.
 
 <p>BIRD works with several types of routes. Some of them are typical IP routes,
 others are better described as forwarding rules. We call them all routes,
-regardless on this difference.
+regardless of this difference.
 
-<p>Every route consists of several attributes (read more about them in the <ref id="route-attributes" name="Route attributes"> section); the common for all routes are:
+<p>Every route consists of several attributes (read more about them in the
+<ref id="route-attributes" name="Route attributes"> section); the common for all
+routes are:
 
 <itemize>
 	<item>IP address of router which told us about this route
@@ -281,7 +284,7 @@ regardless on this difference.
 <p>Other attributes depend on nettypes. Some of them are part of the primary key, these are marked (PK).
 
 <sect1>IPv4 and IPv6 routes
-<label id="ipv4-ipv6-routes">
+<label id="ip-routes">
 
 <p>The traditional routes. Configuration keywords are <cf/ipv4/ and <cf/ipv6/.
 
@@ -291,7 +294,7 @@ regardless on this difference.
 </itemize>
 
 <sect1>VPN IPv4 and IPv6 routes
-<label id="vpn4-vpn6-routes">
+<label id="vpn-routes">
 
 <p>Routes for IPv4 and IPv6 with VPN Route Distinguisher (<rfc id="4364">).
 Configuration keywords are <cf/vpn4/ and <cf/vpn6/.
@@ -303,10 +306,10 @@ Configuration keywords are <cf/vpn4/ and <cf/vpn6/.
 </itemize>
 
 <sect1>Route Origin Authorization for IPv4 and IPv6
-<label id="roa4-roa6-routes">
+<label id="roa-routes">
 
 <p>These entries can be used to validate route origination of BGP routes.
-An ROA entry specifies prefixes which could be originated by an AS number.
+A ROA entry specifies prefixes which could be originated by an AS number.
 Their keywords are <cf/roa4/ and <cf/roa6/.
 
 <itemize>
@@ -316,7 +319,7 @@ Their keywords are <cf/roa4/ and <cf/roa6/.
 </itemize>
 
 <sect1>Flowspec for IPv4 and IPv6
-<label id="flow4-flow6-routes">
+<label id="flow-routes">
 
 <p>Flowspec rules are a form of firewall and traffic flow control rules
 distributed mostly via BGP. These rules may help the operators stop various
@@ -357,12 +360,12 @@ In that case, every next hop has also its weight.
 <sect>Protocols and channels
 <label id="protocols-concept">
 
-<p>BIRD's protocol is an abstract class of producers and consumers of the routes.
+<p>BIRD protocol is an abstract class of producers and consumers of the routes.
 Each protocol may run in multiple instances and bind on one side to route
 tables via channels, on the other side to specified listen sockets (BGP),
 interfaces (Babel, OSPF, RIP), APIs (Kernel, Direct), or nothing (Static, Pipe).
 
-<p>There are also two protocols which do not have any channels -- BFD and Device.
+<p>There are also two protocols that do not have any channels -- BFD and Device.
 Both of them are kind of service for other protocols.
 
 <p>Each protocol is connected to a routing table through a channel. Some protocols
@@ -439,12 +442,14 @@ protocol rip {
 
 <p><descrip>
 	<tag><label id="opt-include">include "<m/filename/";</tag>
-	This statement causes inclusion of a new file. The <m/filename/ could also
-	be a wildcard, in that case matching files are included in alphabetic
-	order. The maximal depth is 8. Note that this statement can be used
-	anywhere in the config file, even inside other options, but always on the beginning of line.
-	In the following example, the first semicolon belongs to the <cf/include/, the second to <cf/ipv6 table/.
-	If the <file/tablename.conf/ contains exactly one token (the name of the table), this construction is correct:
+	This statement causes inclusion of a new file. The <m/filename/ could
+	also be a wildcard, in that case matching files are included in
+	alphabetic order. The maximal depth is 8. Note that this statement can
+	be used anywhere in the config file, even inside other options, but
+	always on the beginning of line. In the following example, the first
+	semicolon belongs to the <cf/include/, the second to <cf/ipv6 table/.
+	If the <file/tablename.conf/ contains exactly one token (the name of the
+	table), this construction is correct:
 <code>
 ipv6 table
 include "tablename.conf";;
@@ -505,7 +510,7 @@ include "tablename.conf";;
 	<tag><label id="opt-function">function <m/name/ (<m/parameters/) <m/local variables/ { <m/commands/ }</tag>
 	Define a function. You can learn more about functions in the following chapter.
 
-	<tag><label id="opt-protocol">protocol rip [|ng]|ospf [v2|v3]|bgp|<m/.../ [<m/name/ [from <m/name2/]] { <m>protocol options</m> }</tag>
+	<tag><label id="opt-protocol">protocol rip|ospf|bgp|<m/.../ [<m/name/ [from <m/name2/]] { <m>protocol options</m> }</tag>
 	Define a protocol instance called <cf><m/name/</cf> (or with a name like
 	"rip5" generated automatically if you don't specify any
 	<cf><m/name/</cf>). You can learn more about configuring protocols in
@@ -514,7 +519,7 @@ include "tablename.conf";;
 	<cf><m/name2/</cf> You can run more than one instance of most protocols
 	(like RIP or BGP). By default, no instances are configured.
 
-	<tag><label id="opt-template">template rip [|ng]|ospf [v2|v3]|bgp|<m/.../ [<m/name/ [from <m/name2/]] { <m>protocol options</m> }</tag>
+	<tag><label id="opt-template">template rip|ospf|bgp|<m/.../ [<m/name/ [from <m/name2/]] { <m>protocol options</m> }</tag>
 	Define a protocol template instance called <m/name/ (or with a name like
 	"bgp1" generated automatically if you don't specify any	<m/name/).
 	Protocol templates can be used to group common options when many
@@ -575,31 +580,17 @@ include "tablename.conf";;
 	<cf/log/ times.
 
 	<tag><label id="opt-table"><m/nettype/ table <m/name/ [sorted]</tag>
-	Create a new routing table. The default routing tables <cf/master4/ and <cf/master6/ are created
-	implicitly, other routing tables have to be added by this command.
-	Option <cf/sorted/ can be used to enable sorting of routes, see
-	<ref id="dsc-table-sorted" name="sorted table"> description for details.
-
-<!--
-	<tag><label id="opt-roa-table">roa table <m/name/ [ { <m/roa table options .../ } ]</tag>
-	Create a new ROA (Route Origin Authorization) table. ROA tables can be
-	used to validate route origination of BGP routes. A ROA table contains
-	ROA entries, each consist of a network prefix, a max prefix length and
-	an AS number. A ROA entry specifies prefixes which could be originated
-	by that AS number. ROA tables could be filled with data from RPKI (<rfc
-	id="6480">) or from public databases like Whois. ROA tables are
-	examined by <cf/roa_check()/ operator in filters.
-
-	Currently, there is just one option, <cf>roa <m/prefix/ max <m/num/ as
-	<m/num/</cf>, which can be used to populate the ROA table with static
-	ROA entries. The option may be used multiple times. Other entries can be
-	added dynamically by <cf/add roa/ command.
--->
+	Create a new routing table. The default routing tables <cf/master4/ and
+	<cf/master6/ are created implicitly, other routing tables have to be
+	added by this command.  Option <cf/sorted/ can be used to enable sorting
+	of routes, see <ref id="dsc-table-sorted" name="sorted table">
+	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>Protocol options
 <label id="protocol-opts">
 
@@ -614,10 +605,6 @@ disable it. An empty <m/switch/ is equivalent to <cf/on/ ("silence means
 agreement").
 
 <descrip>
-	<tag><label id="proto-preference">preference <m/expr/</tag>
-	Sets the preference of routes generated by this protocol. Default:
-	protocol dependent.
-
 	<tag><label id="proto-disabled">disabled <m/switch/</tag>
 	Disables the protocol. You can change the disable/enable status from the
 	command line interface without needing to touch the configuration.
@@ -662,9 +649,10 @@ agreement").
 	multihop BGP.
 
 	<tag><label id="proto-channel"><m/channel name/ [{<m/channel config/}]</tag>
-	Every channel must be explicitly stated. See the protocol-specific configuration
-	for the list of supported channel names.
-	See the <ref id="channel-opts" name="channel configuration section"> for channel definition.
+	Every channel must be explicitly stated. See the protocol-specific
+	configuration for the list of supported channel names. See the
+	<ref id="channel-opts" name="channel configuration section"> for channel
+	definition.
 </descrip>
 
 <p>There are several options that give sense only with certain protocols:
@@ -693,7 +681,7 @@ agreement").
 	options, in that case for given interface the first matching interface
 	option is used.
 
-	This option is allowed in Babel, BFD, Direct, OSPF, RAdv and RIP
+	This option is allowed in Babel, BFD, Device, Direct, OSPF, RAdv and RIP
 	protocols. In OSPF protocol it is used in the <cf/area/ subsection.
 
 	Default: none.
@@ -782,16 +770,22 @@ agreement").
 
 </descrip>
 
+
 <sect>Channel options
 <label id="channel-opts">
 
 <p>Every channel belongs to a protocol and is configured inside its block. The
-minimal channel config is empty, then it uses the default values.
-The channel's name implies its nettype.
+minimal channel config is empty, then it uses the default values. The name of
+the channel implies its nettype.
 
 <descrip>
 	<tag><label id="proto-table">table <m/name/</tag>
-	Connect this channel to a non-default routing table. Default 
+	Specify a table to which the channel is connected. Default: the first
+	table of given nettype.
+
+	<tag><label id="proto-preference">preference <m/expr/</tag>
+	Sets the preference of routes generated by the protocol and imported
+	through this channel. Default: protocol dependent.
 
 	<tag><label id="proto-import">import all | none | filter <m/name/ | filter { <m/filter commands/ } | where <m/boolean filter expression/</tag>
 	Specify a filter to be used for filtering routes coming from the
@@ -1187,14 +1181,13 @@ foot).
 	a shell pattern (represented also as a string).
 
 	<tag><label id="type-ip">ip</tag>
-	This type can hold a single IP address. The IPv4 addresses are stored
-	as IPv4-Mapped IPv6 addresses so one data type for both of them is used.
-	Whether the address is IPv4 or not may be checked by 
-	<cf>.is_ip4</cf> which returns <cf/bool/.
-        IP addresses are written in the standard notation
-	(<cf/10.20.30.40/ or <cf/fec0:3:4::1/). You can apply special operator
-	<cf>.mask(<M>num</M>)</cf> on values of type ip. It masks out all but
-	first <cf><M>num</M></cf> bits from the IP address. So
+	This type can hold a single IP address. The IPv4 addresses are stored as
+	IPv4-Mapped IPv6 addresses so one data type for both of them is used.
+	Whether the address is IPv4 or not may be checked by <cf>.is_ip4</cf>
+	which returns <cf/bool/. IP addresses are written in the standard
+	notation (<cf/10.20.30.40/ or <cf/fec0:3:4::1/). You can apply special
+	operator <cf>.mask(<M>num</M>)</cf> on values of type ip. It masks out
+	all but first <cf><M>num</M></cf> bits from the IP address. So
 	<cf/1.2.3.4.mask(8) = 1.0.0.0/ is true.
 
 	<tag><label id="type-prefix">prefix</tag>
@@ -1207,8 +1200,8 @@ foot).
 	<cf/NET_IP4/ and <cf/NET_IP6/ prefixes hold an IP prefix. The literals
 	are written as <cf><m/ipaddress//<m/pxlen/</cf>,
 	or <cf><m>ipaddress</m>/<m>netmask</m></cf>. There are two special
-	operators on these: <cf/.ip/ which extracts the IP address from
-	the pair, and <cf/.len/, which separates prefix length from the pair.
+	operators on these: <cf/.ip/ which extracts the IP address from the
+	pair, and <cf/.len/, which separates prefix length from the pair.
 	So <cf>1.2.0.0/16.len = 16</cf> is true.
 
 	<cf/NET_VPN4/ and <cf/NET_VPN6/ prefixes hold an IP prefix with VPN
@@ -1225,7 +1218,8 @@ foot).
 	<cf/NET_FLOW4/ and <cf/NET_FLOW6/ hold an IP prefix together with a
 	flowspec rule. Filters currently don't support flowspec parsing.
 
-	<cf/NET_MPLS/ holds a single MPLS label and its handling is currently not implemented.
+	<cf/NET_MPLS/ holds a single MPLS label and its handling is currently
+	not implemented.
 
 	<tag><label id="type-vpnrd">vpnrd</tag>
 	This is a route distinguisher according to <rfc id="4364">. There are
@@ -1328,7 +1322,7 @@ foot).
 	<cf>192.168.0.0/16{16,24}</cf> and <cf>192.168.0.0/16 ge 24</cf> as
 	<cf>192.168.0.0/16{24,32}</cf>.
 
-	It's possible to mix IPv4 and IPv6 prefixes/addresses in a prefix/ip set
+	It is possible to mix IPv4 and IPv6 prefixes/addresses in a prefix/ip set
 	but its behavior may change between versions without any warning; don't do
 	it unless you are more than sure what you are doing. (Really, don't do it.)
 
@@ -1509,8 +1503,9 @@ clist for most purposes.
 
 <descrip>
 	<tag><label id="rta-net"><m/prefix/ net</tag>
-	The network prefix or anything else the route is talking about.
-	The primary key of the table. Read-only. (See the <ref id="routes" name="chapter about routes">.)
+	The network prefix or anything else the route is talking about. The
+	primary key of the routing table. Read-only. (See the <ref id="routes"
+	name="chapter about routes">.)
 
 	<tag><label id="rta-scope"><m/enum/ scope</tag>
 	The scope of the route. Possible values: <cf/SCOPE_HOST/ for routes
@@ -1638,8 +1633,8 @@ protocol babel [<name>] {
 		type <wired|wireless>;
 		rxcost <number>;
 		limit <number>;
-		hello interval <number>;
-		update interval <number>;
+		hello interval <time>;
+		update interval <time>;
 		port <number>;
 		tx class|dscp <number>;
 		tx priority <number>;
@@ -1654,7 +1649,7 @@ protocol babel [<name>] {
 
 <descrip>
       <tag><label id="babel-channel">ipv4|ipv6 <m/channel config/</tag>
-      The supported channels are IPv4 and IPv6. 
+      The supported channels are IPv4 and IPv6.
 
       <tag><label id="babel-type">type wired|wireless </tag>
       This option specifies the interface type: Wired or wireless. On wired
@@ -1683,13 +1678,13 @@ protocol babel [<name>] {
       type interfaces, where gradual cost degradation is used instead of sharp
       limit. Default: 12.
 
-      <tag><label id="babel-hello">hello interval <m/num/</tag>
+      <tag><label id="babel-hello">hello interval <m/time/ s|ms</tag>
       Interval at which periodic Hello messages are sent on this interface,
-      in seconds. Default: 4 seconds.
+      with time units. Default: 4 seconds.
 
-      <tag><label id="babel-update">update interval <m/num/</tag>
-      Interval at which periodic (full) updates are sent. Default: 4 times the
-      hello interval.
+      <tag><label id="babel-update">update interval <m/time/ s|ms</tag>
+      Interval at which periodic (full) updates are sent, with time
+      units. Default: 4 times the hello interval.
 
       <tag><label id="babel-port">port <m/number/</tag>
       This option selects an UDP port to operate on. The default is to operate
@@ -1722,9 +1717,7 @@ protocol babel [<name>] {
 
       <tag><label id="babel-next-hop-ipv4">next hop ipv4 <m/address/</tag>
       Set the next hop address advertised for IPv4 routes advertised on this
-      interface. If not set, the first IPv4 address found on the interface will
-      be used, so it should only be necessary to set this option if this
-      auto-detection fails or finds the wrong address.
+      interface. Default: the preferred IPv4 address of the interface.
 
       <tag><label id="babel-next-hop-ipv6">next hop ipv6 <m/address/</tag>
       Set the next hop address advertised for IPv6 routes advertised on this
@@ -1771,8 +1764,10 @@ protocol babel {
 <sect1>Known issues
 <label id="babel-issues">
 
-<p>When retracting a route, Babel sends instead an unreachable route for a little while (according to RFC). The
-interaction of this behavior with other protocols is not well tested and strange things may happen.
+<p>When retracting a route, Babel generates an unreachable route for a little
+while (according to RFC). The interaction of this behavior with other protocols
+is not well tested and strange things may happen.
+
 
 <sect>BFD
 <label id="bfd">
@@ -2149,7 +2144,7 @@ using the following configuration parameters:
 	BIRD tracks the link state of the associated interface and when link
 	disappears (e.g. an ethernet cable is unplugged), the BGP session is
 	immediately shut down. Note that this option cannot be used with
-	multihop BGP. Default: disabled.
+	multihop BGP. Default: enabled for direct BGP, disabled otherwise.
 
 	<tag><label id="bgp-bfd">bfd <M>switch</M></tag>
 	BGP could use BFD protocol as an advisory mechanism for neighbor
@@ -2386,7 +2381,7 @@ using the following configuration parameters:
 
 	<tag><label id="bgp-igp-metric">igp metric <m/switch/</tag>
 	Enable comparison of internal distances to boundary routers during best
- 	route selection. Default: on.
+	route selection. Default: on.
 
 	<tag><label id="bgp-prefer-older">prefer older <m/switch/</tag>
 	Standard route selection algorithm breaks ties by comparing router IDs.
@@ -2601,10 +2596,9 @@ some of them (marked with `<tt/O/') are optional.
 
 <p><code>
 protocol bgp {
-	local as 65000;			     # Use a private AS number
+	local 198.51.100.14 as 65000;	     # Use a private AS number
 	neighbor 198.51.100.130 as 64496;    # Our neighbor ...
 	multihop;			     # ... which is connected indirectly
-	source address 198.51.100.14;	# Use a non-standard source address
 	ipv4 {
 		export filter {			     # We use non-trivial export rules
 			if source = RTS_STATIC then { # Export only static routes
@@ -2655,7 +2649,6 @@ interfaces to be defined for them to work with.
 <label id="device-config">
 
 <p><descrip>
-
 	<tag><label id="device-scan-time">scan time <m/number/</tag>
 	Time in seconds between two scans of the network interface list. On
 	systems where we are notified about interface status changes
@@ -2663,19 +2656,26 @@ interfaces to be defined for them to work with.
 	list only in order to avoid confusion by lost notification messages,
 	so the default time is set to a large value.
 
-	<tag><label id="device-primary">primary [ "<m/mask/" ] <m/prefix/</tag>
-	If a network interface has more than one network address, BIRD has to
-	choose one of them as a primary one. By default, BIRD chooses the
-	lexicographically smallest address as the primary one.
+	<tag><label id="device-iface">interface <m/pattern/ [, <m/.../]</tag>
 
-	This option allows to specify which network address should be chosen as
-	a primary one. Network addresses that match <m/prefix/ are preferred to
-	non-matching addresses. If more <cf/primary/ options are used, the first
-	one has the highest preference. If "<m/mask/" is specified, then such
-	<cf/primary/ option is relevant only to matching network interfaces.
+	By default, the Device protocol handles all interfaces without any
+	configuration. Interface definitions allow to specify optional
+	parameters for specific interfaces. See <ref id="proto-iface"
+	name="interface"> common option for detailed description. Currently only
+	one interface option is available:
 
-	In all cases, an address marked by operating system as secondary cannot
-	be chosen as the primary one.
+	<tag><label id="device-preferred">preferred <m/ip/</tag>
+	If a network interface has more than one IP address, BIRD chooses one of
+	them as a preferred one. Preferred IP address is used as source address
+	for packets or announced next hop by routing protocols. Precisely, BIRD
+	chooses one preferred IPv4 address, one preferred IPv6 address and one
+	preferred link-local IPv6 address. By default, BIRD chooses the first
+	found IP address as the preferred one.
+
+	This option allows to specify which IP address should be preferred. May
+	be used multiple times for different address classes (IPv4, IPv6, IPv6
+	link-local). In all cases, an address marked by operating system as
+	secondary cannot be chosen as the primary one.
 </descrip>
 
 <p>As the Device protocol doesn't generate any routes, it cannot have
@@ -2684,18 +2684,21 @@ any attributes. Example configuration looks like this:
 <p><code>
 protocol device {
 	scan time 10;		# Scan the interfaces often
-	primary "eth0" 192.168.1.1;
-	primary 192.168.0.0/16;
+	interface "eth0" {
+		preferred 192.168.1.1;
+		preferred 2001:db8:1:10::1;
+	};
 }
 </code>
 
+
 <sect>Direct
 <label id="direct">
 
 <p>The Direct protocol is a simple generator of device routes for all the
 directly connected networks according to the list of interfaces provided by the
-kernel via the Device protocol. The Direct protocol supports IPv4 and IPv6 channels
-in any count.
+kernel via the Device protocol. The Direct protocol supports both IPv4 and IPv6
+channels.
 
 <p>The question is whether it is a good idea to have such device routes in BIRD
 routing table. OS kernel usually handles device routes for directly connected
@@ -2739,13 +2742,13 @@ on Linux systems BIRD cannot change non-BIRD route in the kernel routing table.
 
 <p><code>
 protocol direct {
-	interface "-arc*", "*";		# Exclude the ARCnets
 	ipv4;
-	ipv4 { table myothertable; };
 	ipv6;
+	interface "-arc*", "*";		# Exclude the ARCnets
 }
 </code>
 
+
 <sect>Kernel
 <label id="krt">
 
@@ -2778,8 +2781,9 @@ routing table, there are two limitations - it is not possible to connect more
 kernel protocols to the same routing table and changing route destination
 (gateway) in an export filter of a kernel protocol does not work. Both
 limitations can be overcome using another routing table and the pipe protocol.
-The Kernel protocol obviously supports only two channels -- IPv4 and IPv6; only
-one of them can be configured in each protocol instance.
+
+<p>The Kernel protocol supports both IPv4 and IPv6 channels; only one of them
+can be configured in each protocol instance.
 
 <sect1>Configuration
 <label id="krt-config">
@@ -2798,13 +2802,6 @@ one of them can be configured in each protocol instance.
 	routing daemons or by the system administrator. This is possible only on
 	systems which support identification of route authorship.
 
-	<tag><label id="krt-device-routes">device routes <m/switch/</tag>
-	Enable export of device routes to the kernel routing table. By default,
-	such routes are rejected (with the exception of explicitly configured
-	device routes from the static protocol) regardless of the export filter
-	to protect device routes in kernel routing table (managed by OS itself)
-	from accidental overwriting or erasing.
-
 	<tag><label id="krt-kernel-table">kernel table <m/number/</tag>
 	Select which kernel table should this particular instance of the Kernel
 	protocol work with. Available only on systems supporting multiple
@@ -2971,7 +2968,11 @@ different id. These definitions includes many other switches and multiple
 definitions of interfaces. Definition of interface may contain many switches and
 constant definitions and list of neighbors on nonbroadcast networks.
 
-<p>OSPF v2 needs one IPv4 channel, OSPF v3 needs one IPv6 channel.
+<p>OSPFv2 needs one IPv4 channel. OSPFv3 needs either one IPv6 channel, or one
+IPv4 channel (<rfc id="5838">). Therefore, it is possible to use OSPFv3 for both
+IPv4 and Pv6 routing, but it is necessary to have two protocol instances anyway.
+If no channel is configured, appropriate channel is defined with default
+parameters.
 
 <code>
 protocol ospf [v2|v3] &lt;name&gt; {
@@ -3114,8 +3115,8 @@ protocol ospf [v2|v3] &lt;name&gt; {
 	(equal-cost multipath) routes. Such routes are used when there are
 	several directions to the destination, each with the same (computed)
 	cost. This option also allows to specify a limit on maximum number of
-	nexthops in one route. By default, ECMP is disabled. If enabled,
-	default	value of the limit is 16.
+	nexthops in one route. By default, ECMP is enabled if supported by
+	Kernel. Default value of the limit is 16.
 
 	<tag><label id="ospf-merge-external">merge external <M>switch</M></tag>
 	This option specifies whether OSPF should merge external routes from
@@ -3354,7 +3355,7 @@ protocol ospf [v2|v3] &lt;name&gt; {
 	are immediately considered unreachable and only the address of the iface
 	(instead of whole network prefix) is propagated. It is possible that
 	some hardware drivers or platforms do not implement this feature.
-	Default value is no.
+	Default value is yes.
 
 	<tag><label id="ospf-bfd">bfd <M>switch</M></tag>
 	OSPF could use BFD protocol as an advisory mechanism for neighbor
@@ -3441,8 +3442,6 @@ network. This attribute is read-only. Default is <cf/ospf_metric2 = 10000/ and
 
 <p><code>
 protocol ospf MyOSPF {
-	rfc1583compat yes;
-	tick 2;
 	ipv4 {
 		export filter {
 			if source = RTS_BGP then {
@@ -3549,17 +3548,14 @@ exporting a selected subset of one table to another one.
 <sect1>Configuration
 <label id="pipe-config">
 
-<p>Technically, the Pipe protocol is just a channel connected to a table on both sides.
-Therefore, the configuration block for <cf/protocol pipe/ shall directly include
-standard channel config options; see the example below.
+<p>Essentially, the Pipe protocol is just a channel connected to a table on both
+sides. Therefore, the configuration block for <cf/protocol pipe/ shall directly
+include standard channel config options; see the example below.
 
 <p><descrip>
 	<tag><label id="pipe-peer-table">peer table <m/table/</tag>
 	Defines secondary routing table to connect to. The primary one is
 	selected by the <cf/table/ keyword.
-
-	<tag><label id="pipe-mode">mode opaque|transparent</tag>
-	Specifies the mode for the pipe to work in. Default is transparent.
 </descrip>
 
 <sect1>Attributes
@@ -3586,35 +3582,27 @@ Pipe protocol while decreasing their preferences and correcting their BGP paths
 to reflect the AS boundary crossing.
 
 <code>
-table as1;				# Define the tables
-table as2;
+ipv4 table as1;				# Define the tables
+ipv4 table as2;
 
 protocol kernel kern1 {			# Synchronize them with the kernel
-	ipv4 { table as1; };
+	ipv4 { table as1; export all; };
 	kernel table 1;
 }
 
 protocol kernel kern2 {
-	ipv4 { table as2; };
+	ipv4 { table as2; export all; };
 	kernel table 2;
 }
 
 protocol bgp bgp1 {			# The outside connections
-	ipv4 {
-		table as1;
-		export all;
-		import all;
-	};
+	ipv4 { table as1; export all; };
 	local as 1;
 	neighbor 192.168.0.1 as 1001;
 }
 
 protocol bgp bgp2 {
-	ipv4 {
-		table as2;
-		export all;
-		import all;
-	};
+	ipv4 { table as2; export all; };
 	local as 2;
 	neighbor 10.0.0.1 as 1002;
 }
@@ -3657,7 +3645,7 @@ addresses and choose a default route. BIRD implements router behavior as defined
 in <rfc id="4861">, router preferences and specific routes (<rfc id="4191">),
 and DNS extensions (<rfc id="6106">).
 
-<p>The RAdv protocols supports only IPv6 channels.
+<p>The RAdv protocols supports just IPv6 channel.
 
 <sect1>Configuration
 <label id="radv-config">
@@ -3827,7 +3815,6 @@ definitions, prefix definitions and DNS definitions:
 	option above. Default: no.
 </descrip>
 
-
 <p>Prefix specific options
 
 <descrip>
@@ -3879,7 +3866,6 @@ definitions, prefix definitions and DNS definitions:
 	valid DNS servers. Default: 3 * <cf/max ra interval/.
 </descrip>
 
-
 <p>DNSSL specific options:
 
 <descrip>
@@ -3917,10 +3903,10 @@ definitions, prefix definitions and DNS definitions:
 <label id="radv-exam">
 
 <p><code>
-table radv_routes;			# Manually configured routes go here
+ipv6 table radv_routes;			# Manually configured routes go here
 
 protocol static {
-	table radv_routes;
+	ipv6 { table radv_routes; };
 
 	route 2001:0DB8:4000::/48 unreachable;
 	route 2001:0DB8:4010::/48 unreachable;
@@ -3933,8 +3919,7 @@ protocol static {
 
 protocol radv {
 	propagate routes yes;		# Propagate the routes from the radv_routes table
-	table radv_routes;
-	export all;
+	ipv6 { table radv_routes; export all; };
 
 	interface "eth2" {
 		max ra interval 5;	# Fast failover with more routers
@@ -3970,6 +3955,7 @@ protocol radv {
 }
 </code>
 
+
 <sect>RIP
 <label id="rip">
 
@@ -4003,6 +3989,9 @@ definitions, most RIP options are interface specific. RIPng (RIP for IPv6)
 protocol instance can be configured by using <cf/rip ng/ instead of just
 <cf/rip/ as a protocol type.
 
+<p>RIP needs one IPv4 channel. RIPng needs one IPv6 channel. If no channel is
+configured, appropriate channel is defined with default parameters.
+
 <code>
 protocol rip [ng] [&lt;name&gt;] {
 	infinity &lt;number&gt;;
@@ -4053,8 +4042,8 @@ protocol rip [ng] [&lt;name&gt;] {
 	(equal-cost multipath) routes. Such routes are used when there are
 	several directions to the destination, each with the same (computed)
 	cost. This option also allows to specify a limit on maximum number of
-	nexthops in one route. By default, ECMP is disabled. If enabled,
-	default	value of the limit is 16.
+	nexthops in one route. By default, ECMP is enabled if supported by
+	Kernel. Default value of the limit is 16.
 
 	<tag><label id="rip-iface">interface <m/pattern/ [, <m/.../] { <m/options/ }</tag>
 	Interface definitions specify a set of interfaces on which the
@@ -4200,7 +4189,7 @@ protocol rip [ng] [&lt;name&gt;] {
 	unplugged), neighbors are immediately considered unreachable and all
 	routes received from them are withdrawn. It is possible that some
 	hardware drivers or platforms do not implement this feature.
-	Default: no.
+	Default: yes.
 </descrip>
 
 <sect1>Attributes
@@ -4227,8 +4216,10 @@ protocol rip [ng] [&lt;name&gt;] {
 
 <p><code>
 protocol rip {
-	import all;
-	export all;
+	ipv4 {
+		import all;
+		export all;
+	};
 	interface "eth*" {
 		metric 2;
 		port 1520;
@@ -4241,6 +4232,7 @@ protocol rip {
 }
 </code>
 
+
 <sect>RPKI
 
 <sect1>Introduction
@@ -4248,7 +4240,7 @@ protocol rip {
 <p>The Resource Public Key Infrastructure (RPKI) is mechanism for origin
 validation of BGP routes (RFC 6480). BIRD supports only so-called RPKI-based
 origin validation. There is implemented RPKI to Router (RPKI-RTR) protocol (RFC
-6810).  It uses some of the RPKI data to allow a router to verify that the
+6810). It uses some of the RPKI data to allow a router to verify that the
 autonomous system announcing an IP address prefix is in fact authorized to do
 so. This is not crypto checked so can be violated. But it should prevent the
 vast majority of accidental hijackings on the Internet today, e.g. the famous
@@ -4294,10 +4286,10 @@ protocol rpki [&lt;name&gt;] {
 }
 </code>
 
-<p>Alse note that you have to specify the ROA channel.
-If you want to import only IPv4 prefixes you have
-to specify only roa4 channel. Similarly with IPv6 prefixes only. If you want to
-fetch both IPv4 and even IPv6 ROAs you have to specify both channels.
+<p>Alse note that you have to specify the ROA channel. If you want to import
+only IPv4 prefixes you have to specify only roa4 channel. Similarly with IPv6
+prefixes only. If you want to fetch both IPv4 and even IPv6 ROAs you have to
+specify both channels.
 
 <sect2>RPKI protocol options
 
@@ -4365,21 +4357,20 @@ roa6 table r6;
 
 protocol rpki {
 	debug all;
-	
+
 	roa4 { table r4; };
 	roa6 { table r6; };
 
 	# Please, do not use rpki-validator.realmv6.org in production
 	remote "rpki-validator.realmv6.org" port 8282;
-	
+
 	retry keep 5;
 	refresh keep 30;
 	expire 600;
 }
 
-filter peer_in {
-	if (roa_check(r4, net, bgp_path.last) = ROA_INVALID ||
-	    roa_check(r6, net, bgp_path.last) = ROA_INVALID) then
+filter peer_in_v4 {
+	if (roa_check(r4, net, bgp_path.last) = ROA_INVALID) then
 	{
 		print "Ignore invalid ROA ", net, " for ASN ", bgp_path.last;
 		reject;
@@ -4391,7 +4382,7 @@ protocol bgp {
 	debug all;
 	local as 65000;
 	neighbor 192.168.2.1 as 65001;
-	import filter peer_in;
+	ipv4 { import filter peer_in_v4; };
 }
 </code>
 
@@ -4402,23 +4393,22 @@ roa6 table r6;
 
 protocol rpki {
 	debug all;
-	
+
 	roa4 { table r4; };
 	roa6 { table r6; };
-	
+
 	remote 127.0.0.1 port 2345;
 	transport ssh {
 		bird private key "/home/birdgeek/.ssh/id_rsa";
 		remote public key "/home/birdgeek/.ssh/known_hosts";
 		user "birdgeek";
 	};
-	
+
 	# Default interval values
 }
 </code>
 
 
-
 <sect>Static
 <label id="static">
 
@@ -4461,10 +4451,12 @@ There exist several types of routes; keep in mind that <m/prefix/ syntax is
 	<tag>route <m/prefix/ via <m/ip/|<m/"interface"/ [mpls <m/num/[/<m/num/[/<m/num/[...]]]]</tag>
 	Next hop routes may bear one or more <ref id="route-next-hop" name="next hops">.
 	Every next hop is preceded by <cf/via/ and configured as shown.
+
 	<tag>route <m/prefix/ recursive <m/ip/ [mpls <m/num/[/<m/num/[/<m/num/[...]]]]</tag>
 	Recursive nexthop resolves the given IP in the configured IGP table and
 	uses that route's next hop. The MPLS stacks are concatenated; on top is
 	the IGP's nexthop stack and on bottom is this route's stack.
+
 	<tag>route <m/prefix/ blackhole|unreachable|prohibit</tag>
 	Special routes specifying to silently drop the packet, return it as
 	unreachable or return it as administratively prohibited. First two
@@ -4559,7 +4551,7 @@ protocol static {
 		length > 1024;
 		dscp = 63;
 		fragment dont_fragment, is_fragment || !first_fragment;
-	} drop;
+	};
 }
 </code>
 
@@ -4607,7 +4599,7 @@ protocol static {
 		tcp flags 0x03/0x0f, !0/0xff || 0x33/0x33;
 		fragment !is_fragment || !first_fragment;
 		label 0xaaaa/0xaaaa && 0x33/0x33;
-	} drop;
+	};
 }
 </code>
 
@@ -4661,6 +4653,7 @@ protocol static {
 }
 </code>
 
+
 <chapt>Conclusions
 <label id="conclusion">
 
diff --git a/sysdep/unix/krt.h b/sysdep/unix/krt.h
index eda83b91..3bfccfc2 100644
--- a/sysdep/unix/krt.h
+++ b/sysdep/unix/krt.h
@@ -49,7 +49,7 @@ struct krt_config {
   btime scan_time;		/* How often we re-scan routes */
   int persist;			/* Keep routes when we exit */
   int learn;			/* Learn routes from other sources */
-  int devroutes;		/* Allow export of device routes */
+  int devroutes;		/* XXX: remove */
   int graceful_restart;		/* Regard graceful restart recovery */
 };
 

From c99050cce228b214c2856c6eebefbefffccade7f Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Mon, 11 Dec 2017 01:04:15 +0100
Subject: [PATCH 085/198] KRT: Fix bug in multipath handling

---
 sysdep/linux/netlink.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/sysdep/linux/netlink.c b/sysdep/linux/netlink.c
index afb9dc8b..76ffe78b 100644
--- a/sysdep/linux/netlink.c
+++ b/sysdep/linux/netlink.c
@@ -677,8 +677,10 @@ nl_parse_multipath(struct krt_proto *p, struct rtattr *ra, int af)
 	nh_buf_size = nh_buf_size ? (nh_buf_size * 2) : 4;
 	nh_buffer = xrealloc(nh_buffer, nh_buf_size * NEXTHOP_MAX_SIZE);
       }
-      *last = rv = nh_buffer + nh_buf_used++;
-      rv->next = NULL;
+      /* FIXME: This is really ugly */
+      *last = rv = (void *) (((byte *) nh_buffer) + (nh_buf_used++ * NEXTHOP_MAX_SIZE));
+      memset(rv, 0, NEXTHOP_MAX_SIZE);
+      // rv->next = NULL;
       last = &(rv->next);
 
       rv->flags = 0;

From d4eada9e0f3dc1f1e22b83fc8a8361ae256592ff Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Mon, 11 Dec 2017 02:05:35 +0100
Subject: [PATCH 086/198] NEWS and version update

---
 NEWS            | 57 +++++++++++++++++++++++++------------------------
 sysdep/config.h |  2 +-
 2 files changed, 30 insertions(+), 29 deletions(-)

diff --git a/NEWS b/NEWS
index 9d8e734a..55df8146 100644
--- a/NEWS
+++ b/NEWS
@@ -1,45 +1,46 @@
-Version 2.0.0-pre1 (2017-04-29)
+Version 2.0.0 (2017-12-11)
+  o Integrated IPv4 + IPv6 design
   o Support for MPLS next hops
-  o VPNv4 and VPNv6 network types
+  o Support for VPNv4 and VPNv6 networks
+  o Microsecond timers infrastructure
+  o Basic VRF support
+  o Babel: Support for dual-stack IPv4/IPv6
+  o Babel: Many improvements and bugfixes
+  o Major BGP protocol redesign
+  o Full support for Multiprotocol BGP
+  o BGP multicast support (SAFI 2)
+  o BGP flowspec support (RFC 5575)
   o BGP with MPLS labels (RFC 3107)
   o BGP MPLS/VPN support (RFC 4364)
   o BGP 6PE - IPv6 NLRI over IPv4 MPLS (RFC 4798)
   o BGP IPv4 NLRI with an IPv6 Next Hop (RFC 5549)
   o BGP Confederations (RFC 5065)
-  o BGP: Simplify igp table options
+  o BGP Shutdown communication (RFC 8203)
   o BGP: Allow exchanging LOCAL_PREF with eBGP peers
   o BGP: Allow to specify interface for regular sessions
-  o Babel support restored
-  o Static: Minor overhaul
-  o Netlink: Default kernel metric changed to 32
-  o Flowspec: Limit tcp mask length to 12 bits
-  o Update of show route command
-
-  Notes:
-
-  Definitions of OSPFv2, OSPFv3 and RIP NG protocols now use keywords
-  'ospf v2', 'ospf v3' and 'rip ng' instead of 'ospf2', 'ospf3' and 'ripng'.
-
-  Flows and ROAs no longer use phony next hops, so there is no need to use
-  'drop' or 'unreachable' in their static route definitions.
-
-  See doc/bird.conf.example2 for configuration examples.
-
-
-Version 2.0.0-pre0 (2016-12-07)
-  o Integrated IPv4 + IPv6 design
-  o Major BGP protocol redesign
-  o BGP multicast support (SAFI 2)
-  o BGP flowspec support (RFC 5575)
+  o OSPF: Support of address families in OSPFv3
+  o OSPF: Enable ECMP and Link detection by default
+  o RAdv: Support for more specific routes (RFC 4191)
+  o RAdv: Proper handling of prefix retraction
+  o RIP: Enable ECMP and Link detection by default
+  o Redesign of RPKI handling
   o New RPKI-Router protocol
+  o Static: Minor overhaul
+  o Static: Support for all new route types
+  o Kenrel: Default Linux kernel metric changed to 32
+  o Kernel: Fix IPv6 ECMP handling with Linux 4.11+
+  o Update of show route command
+  o BIRD client persistent history
   o New build system
   o Unit tests
+  o ...
 
   Notes:
 
-  Protocols and tables are now connected using explicit channels, most related
-  protocol options (table, import, export, ...) are now channel options. See
-  doc/bird.conf.example2 for configuration examples.
+  Tables are now defined with appropriate net type keyword. Protocols and tables
+  are now connected by explicit channels, most related protocol options (table,
+  import, export, ...) are now channel options. See doc/bird.conf.example2 for
+  configuration examples. Some options were removed/replaced.
 
 
 Version 1.6.3 (2016-12-21)
diff --git a/sysdep/config.h b/sysdep/config.h
index fb710d86..460cac8b 100644
--- a/sysdep/config.h
+++ b/sysdep/config.h
@@ -7,7 +7,7 @@
 #define _BIRD_CONFIG_H_
 
 /* BIRD version */
-#define BIRD_VERSION "2.0.0-pre1"
+#define BIRD_VERSION "2.0.0"
 
 /* Include parameters determined by configure script */
 #include "sysdep/autoconf.h"

From fa5c09a2e708ed505ca140531b98e695fdaf989c Mon Sep 17 00:00:00 2001
From: Ondrej Filip <feela@network.cz>
Date: Mon, 11 Dec 2017 09:36:21 +0100
Subject: [PATCH 087/198] Changes to be able to build 2.0.0

---
 tools/gendist | 21 +++++++++++----------
 1 file changed, 11 insertions(+), 10 deletions(-)

diff --git a/tools/gendist b/tools/gendist
index 09d5d055..0275db1a 100755
--- a/tools/gendist
+++ b/tools/gendist
@@ -4,26 +4,27 @@
 #  (c) 2000--2004 Martin Mares <mj@ucw.cz>
 #
 
+VERSION=`sed <sysdep/config.h '/BIRD_VERSION/!d;s/^.*"\(.*\)"$/\1/'`
+REL=bird-$VERSION
+DREL=bird-doc-$VERSION
+T=/tmp/bird
 set -e
 AC=`if [ -x /usr/bin/autoconf2.50 ] ; then echo autoconf2.50 ; else echo autoconf ; fi`
 $AC
 ./configure
 make docs
-make distclean
-$AC
-rm -rf autom4te*cache
-VERSION=`sed <sysdep/config.h '/BIRD_VERSION/!d;s/^.*"\(.*\)"$/\1/'`
-REL=bird-$VERSION
-DREL=bird-doc-$VERSION
-T=/tmp/bird
-echo Building $REL
 rm -rf $T/$REL $T/$DREL
 mkdir -p $T/$REL $T/$DREL $T/$DREL/doc
+mv obj/doc/*.pdf $T/$DREL/doc
+make distclean
+find . -name "*~" -exec rm -f '{}' '+'
+
+$AC
+rm -rf autom4te*cache
+echo Building $REL
 cp -a . $T/$REL
 echo Generating ChangeLog
 git log  >$T/$REL/ChangeLog
-mv $T/$REL/doc/*.ps $T/$DREL/doc
-mv $T/$REL/doc/*.pdf $T/$DREL/doc
 rm -f $T/$REL/bird.conf*
 rm -rf $T/$REL/.git/
 rm -rf `find $T/$REL -name CVS -o -name tmp` $T/$REL/{misc,rfc,doc/slides,doc/slt2001,doc/old,doc/*.out}

From 66acbc8d7f80ce3c197549922b60fca093129047 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 12 Dec 2017 00:05:49 +0100
Subject: [PATCH 088/198] Revive FIB and kernel MPLS code

---
 conf/confbase.Y        | 11 +++++++++--
 nest/config.Y          |  1 +
 nest/rt-fib.c          |  3 +++
 sysdep/linux/netlink.c | 27 +++++++++++++++++----------
 4 files changed, 30 insertions(+), 12 deletions(-)

diff --git a/conf/confbase.Y b/conf/confbase.Y
index 16330984..7e0537c5 100644
--- a/conf/confbase.Y
+++ b/conf/confbase.Y
@@ -83,7 +83,7 @@ CF_DECLS
 %type <time> expr_us time
 %type <a> ipa
 %type <net> net_ip4_ net_ip6_ net_ip6 net_ip_ net_ip net_or_ipa
-%type <net_ptr> net_ net_any net_vpn4_ net_vpn6_ net_vpn_ net_roa4_ net_roa6_ net_roa_
+%type <net_ptr> net_ net_any net_vpn4_ net_vpn6_ net_vpn_ net_roa4_ net_roa6_ net_roa_ net_mpls_
 %type <mls> label_stack_start label_stack
 
 %type <t> text opttext
@@ -96,7 +96,7 @@ CF_DECLS
 %left '!'
 %nonassoc '.'
 
-CF_KEYWORDS(DEFINE, ON, OFF, YES, NO, S, MS, US, PORT, VPN)
+CF_KEYWORDS(DEFINE, ON, OFF, YES, NO, S, MS, US, PORT, VPN, MPLS)
 
 CF_GRAMMAR
 
@@ -234,6 +234,12 @@ net_roa6_: net_ip6_ MAX NUM AS NUM
     cf_error("Invalid max prefix length %u", $3);
 };
 
+net_mpls_: MPLS NUM
+{
+  $$ = cfg_alloc(sizeof(net_addr_roa6));
+  net_fill_mpls($$, $2);
+}
+
 net_ip_: net_ip4_ | net_ip6_ ;
 net_vpn_: net_vpn4_ | net_vpn6_ ;
 net_roa_: net_roa4_ | net_roa6_ ;
@@ -243,6 +249,7 @@ net_:
  | net_vpn_
  | net_roa_
  | net_flow_
+ | net_mpls_
  ;
 
 
diff --git a/nest/config.Y b/nest/config.Y
index 432dd9c7..5c4f2393 100644
--- a/nest/config.Y
+++ b/nest/config.Y
@@ -140,6 +140,7 @@ net_type:
  | ROA6 { $$ = NET_ROA6; }
  | FLOW4{ $$ = NET_FLOW4; }
  | FLOW6{ $$ = NET_FLOW6; }
+ | MPLS { $$ = NET_MPLS; }
  ;
 
 CF_ENUM(T_ENUM_NETTYPE, NET_, IP4, IP6, VPN4, VPN6, ROA4, ROA6, FLOW4, FLOW6)
diff --git a/nest/rt-fib.c b/nest/rt-fib.c
index 45f47261..60c3da55 100644
--- a/nest/rt-fib.c
+++ b/nest/rt-fib.c
@@ -213,6 +213,7 @@ fib_hash(struct fib *f, const net_addr *a)
   case NET_ROA6: return FIB_HASH(f, a, roa6);
   case NET_FLOW4: return FIB_HASH(f, a, flow4);
   case NET_FLOW6: return FIB_HASH(f, a, flow6);
+  case NET_MPLS: return FIB_HASH(f, a, mpls);
   default: bug("invalid type");
   }
 }
@@ -249,6 +250,7 @@ fib_find(struct fib *f, const net_addr *a)
   case NET_ROA6: return FIB_FIND(f, a, roa6);
   case NET_FLOW4: return FIB_FIND(f, a, flow4);
   case NET_FLOW6: return FIB_FIND(f, a, flow6);
+  case NET_MPLS: return FIB_FIND(f, a, mpls);
   default: bug("invalid type");
   }
 }
@@ -268,6 +270,7 @@ fib_insert(struct fib *f, const net_addr *a, struct fib_node *e)
   case NET_ROA6: FIB_INSERT(f, a, e, roa6); return;
   case NET_FLOW4: FIB_INSERT(f, a, e, flow4); return;
   case NET_FLOW6: FIB_INSERT(f, a, e, flow6); return;
+  case NET_MPLS: FIB_INSERT(f, a, e, mpls); return;
   default: bug("invalid type");
   }
 }
diff --git a/sysdep/linux/netlink.c b/sysdep/linux/netlink.c
index 76ffe78b..bd3dbbd9 100644
--- a/sysdep/linux/netlink.c
+++ b/sysdep/linux/netlink.c
@@ -561,25 +561,20 @@ nl_add_attr_mpls_encap(struct nlmsghdr *h, uint bufsize, int len, u32 *stack)
 static inline void
 nl_add_attr_via(struct nlmsghdr *h, uint bufsize, ip_addr ipa)
 {
-  struct rtattr *nest = nl_open_attr(h, bufsize, RTA_VIA);
-  struct rtvia *via = RTA_DATA(nest);
-
-  h->nlmsg_len += sizeof(*via);
+  struct rtvia *via = alloca(sizeof(struct rtvia) + 16);
 
   if (ipa_is_ip4(ipa))
   {
     via->rtvia_family = AF_INET;
     put_ip4(via->rtvia_addr, ipa_to_ip4(ipa));
-    h->nlmsg_len += sizeof(ip4_addr);
+    nl_add_attr(h, bufsize, RTA_VIA, via, sizeof(struct rtvia) + 4);
   }
   else
   {
     via->rtvia_family = AF_INET6;
     put_ip6(via->rtvia_addr, ipa_to_ip6(ipa));
-    h->nlmsg_len += sizeof(ip6_addr);
+    nl_add_attr(h, bufsize, RTA_VIA, via, sizeof(struct rtvia) + 16);
   }
-
-  nl_close_attr(h, nest);
 }
 #endif
 
@@ -1228,8 +1223,16 @@ nl_send_route(struct krt_proto *p, rte *e, struct ea_list *eattrs, int op, int d
 #ifdef HAVE_MPLS_KERNEL
   if (p->af == AF_MPLS)
   {
+    /*
+     * Kernel MPLS code is a bit picky. We must:
+     * 1) Always set RT_SCOPE_UNIVERSE and RTN_UNICAST (even for RTM_DELROUTE)
+     * 2) Never use RTA_PRIORITY
+     */
+
     u32 label = net_mpls(net->n.addr);
     nl_add_attr_mpls(&r->h, rsize, RTA_DST, 1, &label);
+    r->r.rtm_scope = RT_SCOPE_UNIVERSE;
+    r->r.rtm_type = RTN_UNICAST;
   }
   else
 #endif
@@ -1247,7 +1250,9 @@ nl_send_route(struct krt_proto *p, rte *e, struct ea_list *eattrs, int op, int d
   else
     nl_add_attr_u32(&r->h, rsize, RTA_TABLE, krt_table_id(p));
 
-  if (a->source == RTS_DUMMY)
+  if (p->af == AF_MPLS)
+    priority = 0;
+  else if (a->source == RTS_DUMMY)
     priority = e->u.krt.metric;
   else if (KRT_CF->sys.metric)
     priority = KRT_CF->sys.metric;
@@ -1262,7 +1267,9 @@ nl_send_route(struct krt_proto *p, rte *e, struct ea_list *eattrs, int op, int d
     goto dest;
 
   /* Default scope is LINK for device routes, UNIVERSE otherwise */
-  if (ea = ea_find(eattrs, EA_KRT_SCOPE))
+  if (p->af == AF_MPLS)
+    r->r.rtm_scope = RT_SCOPE_UNIVERSE;
+  else if (ea = ea_find(eattrs, EA_KRT_SCOPE))
     r->r.rtm_scope = ea->u.data;
   else
     r->r.rtm_scope = (dest == RTD_UNICAST && ipa_zero(nh->gw)) ? RT_SCOPE_LINK : RT_SCOPE_UNIVERSE;

From b5257bea853850809be7f03eb0e1dbb398c56c34 Mon Sep 17 00:00:00 2001
From: Ondrej Filip <feela@network.cz>
Date: Tue, 12 Dec 2017 10:43:56 +0100
Subject: [PATCH 089/198] Removed '--enable-ipv6' reference.

---
 README | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/README b/README
index 5a6d934d..7b8fbf2e 100644
--- a/README
+++ b/README
@@ -6,7 +6,7 @@
                 (c) 1998--2008  Martin Mares <mj@ucw.cz>
                 (c) 1998--2000  Pavel Machek <pavel@ucw.cz>
                 (c) 1998--2008  Ondrej Filip <feela@network.cz>
-                (c) 2009--2016  CZ.NIC z.s.p.o.
+                (c) 2009--2017  CZ.NIC z.s.p.o.
 
 ================================================================================
 
@@ -19,7 +19,7 @@ Public License.
 What do we support
 ==================
 
-	o  Both IPv4 and IPv6 (use --enable-ipv6 when configuring)
+	o  Both IPv4 and IPv6
 	o  Multiple routing tables
 	o  Border Gateway Protocol (BGPv4)
 	o  Routing Information Protocol (RIPv2, RIPng)

From cb5df823acdc54f94b0b85094cb59ac68c83c33a Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 12 Dec 2017 15:22:01 +0100
Subject: [PATCH 090/198] Minor CI tweak

---
 .gitlab-ci.yml | 38 ++------------------------------------
 1 file changed, 2 insertions(+), 36 deletions(-)

diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
index 05989484..482c5004 100644
--- a/.gitlab-ci.yml
+++ b/.gitlab-ci.yml
@@ -243,107 +243,73 @@ docker_ubuntu-16_04-amd64:
   - MAKE=make
   - which gmake 2>/dev/null >/dev/null && MAKE=gmake
   - $MAKE
-  # Run tests if they are available (eg. don't fail if "check" isn't a valid make target)
-  - $MAKE check || [ "$?" = 2 ]
+  # Run tests if they are available
+  - $MAKE check
 
 build-debian-7-amd64:
-  variables:
-    IPV6: "no"
   <<: *debian-7-amd64_env
   <<: *build_job
 
 build-debian-8-amd64:
-  variables:
-    IPV6: "no"
   <<: *debian-8-amd64_env
   <<: *build_job
 
 build-debian-9-amd64:
-  variables:
-    IPV6: "no"
   <<: *debian-9-amd64_env
   <<: *build_job
 
 build-debian-testing-amd64:
-  variables:
-    IPV6: "no"
   <<: *debian-testing-amd64_env
   <<: *build_job
 
 build-fedora-25-amd64:
-  variables:
-    IPV6: "no"
   <<: *fedora-25-amd64_env
   <<: *build_job
 
 build-fedora-26-amd64:
-  variables:
-    IPV6: "no"
   <<: *fedora-26-amd64_env
   <<: *build_job
 
 build-centos-6-amd64:
-  variables:
-    IPV6: "no"
   <<: *centos-6-amd64_env
   <<: *build_job
 
 build-centos-7-amd64:
-  variables:
-    IPV6: "no"
   <<: *centos-7-amd64_env
   <<: *build_job
 
 build-opensuse-42_3-amd64:
-  variables:
-    IPV6: "no"
   <<: *opensuse-42_3-amd64_env
   <<: *build_job
 
 build-ubuntu-14_04-amd64:
-  variables:
-    IPV6: "no"
   <<: *ubuntu-14_04-amd64_env
   <<: *build_job
 
 build-ubuntu-16_04-amd64:
-  variables:
-    IPV6: "no"
   <<: *ubuntu-16_04-amd64_env
   <<: *build_job
 
 build-debian-7-i386:
-  variables:
-    IPV6: "no"
   <<: *debian-7-i386_env
   <<: *build_job
 
 build-debian-8-i386:
-  variables:
-    IPV6: "no"
   <<: *debian-8-i386_env
   <<: *build_job
 
 build-debian-9-i386:
-  variables:
-    IPV6: "no"
   <<: *debian-9-i386_env
   <<: *build_job
 
 build-debian-testing-i386:
-  variables:
-    IPV6: "no"
   <<: *debian-testing-i386_env
   <<: *build_job
 
 build-freebsd-11-amd64:
-  variables:
-    IPV6: "no"
   <<: *freebsd-11-amd64_env
   <<: *build_job
 
 build-freebsd-11-i386:
-  variables:
-    IPV6: "no"
   <<: *freebsd-11-i386_env
   <<: *build_job

From ac48e72bf6f9f491824e2de59a035c93aab8f81b Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 12 Dec 2017 15:56:31 +0100
Subject: [PATCH 091/198] Fix some minor issues

---
 lib/flowspec_test.c   |  4 ++--
 lib/net.h             |  4 ++--
 sysdep/bsd/krt-sock.Y |  4 ++--
 sysdep/bsd/krt-sock.c | 10 +++++++---
 sysdep/bsd/krt-sys.h  |  2 +-
 5 files changed, 14 insertions(+), 10 deletions(-)

diff --git a/lib/flowspec_test.c b/lib/flowspec_test.c
index 69bc279d..dd71dc7b 100644
--- a/lib/flowspec_test.c
+++ b/lib/flowspec_test.c
@@ -70,8 +70,8 @@ t_first_part(void)
   net_addr_flow4 *f;
   NET_ADDR_FLOW4_(f, ip4_build(10,0,0,1), 24, ((byte[]) { 0x00, 0x00, 0xab }));
 
-  const byte const *under240 = &f->data[1];
-  const byte const *above240 = &f->data[2];
+  const byte *under240 = &f->data[1];
+  const byte *above240 = &f->data[2];
 
   /* Case 0x00 0x00 */
   bt_assert(flow4_first_part(f) == NULL);
diff --git a/lib/net.h b/lib/net.h
index 4b2077ae..69f00641 100644
--- a/lib/net.h
+++ b/lib/net.h
@@ -358,10 +358,10 @@ static inline int net_zero_roa6(const net_addr_roa6 *a)
 { return !a->pxlen && ip6_zero(a->prefix) && !a->max_pxlen && !a->asn; }
 
 static inline int net_zero_flow4(const net_addr_flow4 *a)
-{ return !a->pxlen && ip4_zero(a->prefix) && !a->data; }
+{ return !a->pxlen && ip4_zero(a->prefix) && (a->length == sizeof(net_addr_flow4)); }
 
 static inline int net_zero_flow6(const net_addr_flow6 *a)
-{ return !a->pxlen && ip6_zero(a->prefix) && !a->data; }
+{ return !a->pxlen && ip6_zero(a->prefix) && (a->length == sizeof(net_addr_flow6)); }
 
 static inline int net_zero_mpls(const net_addr_mpls *a)
 { return !a->label; }
diff --git a/sysdep/bsd/krt-sock.Y b/sysdep/bsd/krt-sock.Y
index 0218f188..81422c79 100644
--- a/sysdep/bsd/krt-sock.Y
+++ b/sysdep/bsd/krt-sock.Y
@@ -20,8 +20,8 @@ kern_sys_item:
    KERNEL TABLE expr {
 	if ($3 && (krt_max_tables == 1))
 	  cf_error("Multiple kernel routing tables not supported");
- 	if ($3 < 0 || $3 >= krt_max_tables)
-	  cf_error("Kernel table id must be in range 0-%d", krt_max_tables - 1);
+	if ($3 >= krt_max_tables)
+	  cf_error("Kernel table id must be in range 0-%u", krt_max_tables - 1);
 
 	THIS_KRT->sys.table_id = $3;
    }
diff --git a/sysdep/bsd/krt-sock.c b/sysdep/bsd/krt-sock.c
index 604cd510..0a52cfbd 100644
--- a/sysdep/bsd/krt-sock.c
+++ b/sysdep/bsd/krt-sock.c
@@ -74,11 +74,11 @@ const int rt_default_ecmp = 0;
 
 /* Dynamic max number of tables */
 
-int krt_max_tables;
+uint krt_max_tables;
 
 #ifdef KRT_USE_SYSCTL_NET_FIBS
 
-static int
+static uint
 krt_get_max_tables(void)
 {
   int fibs;
@@ -90,7 +90,11 @@ krt_get_max_tables(void)
     return 1;
   }
 
-  return MIN(fibs, KRT_MAX_TABLES);
+  /* Should not happen */
+  if (fibs < 1)
+    return 1;
+
+  return (uint) MIN(fibs, KRT_MAX_TABLES);
 }
 
 #else
diff --git a/sysdep/bsd/krt-sys.h b/sysdep/bsd/krt-sys.h
index ed667e80..aa6cc72e 100644
--- a/sysdep/bsd/krt-sys.h
+++ b/sysdep/bsd/krt-sys.h
@@ -31,7 +31,7 @@ static inline void kif_sys_copy_config(struct kif_config *d UNUSED, struct kif_c
 
 /* Kernel routes */
 
-extern int krt_max_tables;
+extern uint krt_max_tables;
 
 struct krt_params {
   int table_id;				/* Kernel table ID we sync with */

From 1e11918c8c56e3505193f4e6426c1a34aaae3941 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 12 Dec 2017 19:51:36 +0100
Subject: [PATCH 092/198] Lib: Save/restore state for linpools

Also change linpool.current ptr to really point to thr current chunk.
---
 lib/mempool.c  | 69 ++++++++++++++++++++++++++++++++++++++++++--------
 lib/resource.h |  7 +++++
 2 files changed, 65 insertions(+), 11 deletions(-)

diff --git a/lib/mempool.c b/lib/mempool.c
index 3cf9c2d3..5a8f2a69 100644
--- a/lib/mempool.c
+++ b/lib/mempool.c
@@ -11,7 +11,7 @@
  *
  * Linear memory pools are collections of memory blocks which
  * support very fast allocation of new blocks, but are able to free only
- * the whole collection at once.
+ * the whole collection at once (or in stack order).
  *
  * Example: Each configuration is described by a complex system of structures,
  * linked lists and function trees which are all allocated from a single linear
@@ -37,7 +37,7 @@ const int lp_chunk_size = sizeof(struct lp_chunk);
 struct linpool {
   resource r;
   byte *ptr, *end;
-  struct lp_chunk *first, *current, **plast;	/* Normal (reusable) chunks */
+  struct lp_chunk *first, *current;		/* Normal (reusable) chunks */
   struct lp_chunk *first_large;			/* Large chunks */
   uint chunk_size, threshold, total, total_large;
 };
@@ -69,7 +69,6 @@ linpool
 *lp_new(pool *p, uint blk)
 {
   linpool *m = ralloc(p, &lp_class);
-  m->plast = &m->first;
   m->chunk_size = blk;
   m->threshold = 3*blk/4;
   return m;
@@ -114,22 +113,25 @@ lp_alloc(linpool *m, uint size)
 	}
       else
 	{
-	  if (m->current)
+	  if (m->current && m->current->next)
 	    {
 	      /* Still have free chunks from previous incarnation (before lp_flush()) */
-	      c = m->current;
-	      m->current = c->next;
+	      c = m->current->next;
 	    }
 	  else
 	    {
 	      /* Need to allocate a new chunk */
 	      c = xmalloc(sizeof(struct lp_chunk) + m->chunk_size);
 	      m->total += m->chunk_size;
-	      *m->plast = c;
-	      m->plast = &c->next;
 	      c->next = NULL;
 	      c->size = m->chunk_size;
+
+	      if (m->current)
+		m->current->next = c;
+	      else
+		m->first = c;
 	    }
+	  m->current = c;
 	  m->ptr = c->data + size;
 	  m->end = c->data + m->chunk_size;
 	}
@@ -190,9 +192,11 @@ lp_flush(linpool *m)
 {
   struct lp_chunk *c;
 
-  /* Relink all normal chunks to free list and free all large chunks */
-  m->ptr = m->end = NULL;
-  m->current = m->first;
+  /* Move ptr to the first chunk and free all large chunks */
+  m->current = c = m->first;
+  m->ptr = c ? c->data : NULL;
+  m->end = c ? c->data + m->chunk_size : NULL;
+
   while (c = m->first_large)
     {
       m->first_large = c->next;
@@ -201,6 +205,49 @@ lp_flush(linpool *m)
   m->total_large = 0;
 }
 
+/**
+ * lp_save - save the state of a linear memory pool
+ * @m: linear memory pool
+ * @p: state buffer
+ *
+ * This function saves the state of a linear memory pool. Saved state can be
+ * used later to restore the pool (to free memory allocated since).
+ */
+void
+lp_save(linpool *m, lp_state *p)
+{
+  p->current = m->current;
+  p->large = m->first_large;
+  p->ptr = m->ptr;
+}
+
+/**
+ * lp_restore - restore the state of a linear memory pool
+ * @m: linear memory pool
+ * @p: saved state
+ *
+ * This function restores the state of a linear memory pool, freeing all memory
+ * allocated since the state was saved. Note that the function cannot un-free
+ * the memory, therefore the function also invalidates other states that were
+ * saved between (on the same pool).
+ */
+void
+lp_restore(linpool *m, lp_state *p)
+{
+  struct lp_chunk *c;
+
+  /* Move ptr to the saved pos and free all newer large chunks */
+  m->current = c = p->current;
+  m->ptr = p->ptr;
+  m->end = c ? c->data + m->chunk_size : NULL;
+
+  while ((c = m->first_large) && (c != p->large))
+    {
+      m->first_large = c->next;
+      xfree(c);
+    }
+}
+
 static void
 lp_free(resource *r)
 {
diff --git a/lib/resource.h b/lib/resource.h
index 761c6adc..d9d4bb8f 100644
--- a/lib/resource.h
+++ b/lib/resource.h
@@ -59,11 +59,18 @@ void mb_free(void *);
 
 typedef struct linpool linpool;
 
+typedef struct lp_state {
+  void *current, *large;
+  byte *ptr;
+} lp_state;
+
 linpool *lp_new(pool *, unsigned blk);
 void *lp_alloc(linpool *, unsigned size);	/* Aligned */
 void *lp_allocu(linpool *, unsigned size);	/* Unaligned */
 void *lp_allocz(linpool *, unsigned size);	/* With clear */
 void lp_flush(linpool *);			/* Free everything, but leave linpool */
+void lp_save(linpool *m, lp_state *p);		/* Save state */
+void lp_restore(linpool *m, lp_state *p);	/* Restore state */
 
 extern const int lp_chunk_size;
 #define LP_GAS		    1024

From 71c51aa4ab0daa3490f9a488f505eb25102c4705 Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Wed, 13 Dec 2017 10:28:50 +0100
Subject: [PATCH 093/198] Doc: Fixed misc sgml bugs, no content change

---
 doc/bird.sgml | 17 ++++++++++++-----
 1 file changed, 12 insertions(+), 5 deletions(-)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index 8309e051..ebafcf92 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -2007,9 +2007,10 @@ table it wishes to export along with complete path information (a list of AS'es
 the packet will travel through if it uses the particular route) in order to
 avoid routing loops.
 
-<sect1>Supported standards:
+<sect1>Supported standards
 <label id="bgp-standards">
 
+<p>
 <itemize>
 <item> <rfc id="4271"> - Border Gateway Protocol 4 (BGP)
 <item> <rfc id="1997"> - BGP Communities Attribute
@@ -2407,9 +2408,10 @@ using the following configuration parameters:
 announced to the peer corresponds to one channel. The table of supported AFI/SAFIs
 together with their appropriate channels follows.
 
+<table loc="h">
 <tabular ca="l|l|l|r|r">
   <bf/Channel name/   | <bf/Table nettype/ | <bf/IGP table allowed/  | <bf/AFI/ | <bf/SAFI/
-@ <hline>
+@<hline>
   <cf/ipv4/	      | <cf/ipv4/          | <cf/ipv4/ and <cf/ipv6/ | 1        | 1
 @ <cf/ipv6/           | <cf/ipv6/          | <cf/ipv4/ and <cf/ipv6/ | 2        | 1
 @ <cf/ipv4 multicast/ | <cf/ipv4/          | <cf/ipv4/ and <cf/ipv6/ | 1        | 2
@@ -2423,6 +2425,7 @@ together with their appropriate channels follows.
 @ <cf/flow4/	      | <cf/flow4/         | ---                     | 1        | 133
 @ <cf/flow6/          | <cf/flow6/         | ---                     | 2        | 133
 </tabular>
+</table>
 
 <p>BGP's channels have additional config options (together with the common ones):
 
@@ -4255,6 +4258,7 @@ RFC 6811, but we don't support it yet! You can use a BIRD's client command
 routes.
 
 <sect1>Supported transports
+<p>
 <itemize>
         <item>Unprotected transport over TCP uses a port 323. The cache server
         and BIRD router should be on the same trusted and controlled network
@@ -4292,7 +4296,7 @@ prefixes only. If you want to fetch both IPv4 and even IPv6 ROAs you have to
 specify both channels.
 
 <sect2>RPKI protocol options
-
+<p>
 <descrip>
         <tag>remote <m/ip/ | "<m/hostname/" [port <m/num/]</tag> Specifies
         a destination address of the cache server.  Can be specified by an IP
@@ -4335,6 +4339,7 @@ specify both channels.
 </descrip>
 
 <sect3>SSH transport options
+<p>
 <descrip>
 	<tag>bird private key "<m>/path/to/id_rsa</m>"</tag>
 	A path to the BIRD's private SSH key for authentication.
@@ -4387,6 +4392,7 @@ protocol bgp {
 </code>
 
 <sect2>SSHv2 transport encryption
+<p>
 <code>
 roa4 table r4;
 roa6 table r6;
@@ -4442,9 +4448,9 @@ Static routes have no specific attributes.
 
 <p>Route definitions (each may also contain a block of per-route options):
 
-<sect1>Regular routes; MPLS switching rules</sect1>
+<sect1>Regular routes; MPLS switching rules
 
-There exist several types of routes; keep in mind that <m/prefix/ syntax is
+<p>There exist several types of routes; keep in mind that <m/prefix/ syntax is
 <ref id="type-prefix" name="dependent on network type">.
 
 <descrip>
@@ -4604,6 +4610,7 @@ protocol static {
 </code>
 
 <sect1>Per-route options
+<p>
 <descrip>
 	<tag><label id="static-route-bfd">bfd <m/switch/</tag>
 	The Static protocol could use BFD protocol for next hop liveness

From 772beb7308a53558e153223bc42742f719ff4a40 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 13 Dec 2017 14:49:55 +0100
Subject: [PATCH 094/198] Lib: Minor fix

---
 lib/mempool.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/lib/mempool.c b/lib/mempool.c
index 5a8f2a69..758882ce 100644
--- a/lib/mempool.c
+++ b/lib/mempool.c
@@ -244,6 +244,7 @@ lp_restore(linpool *m, lp_state *p)
   while ((c = m->first_large) && (c != p->large))
     {
       m->first_large = c->next;
+      m->total_large -= c->size;
       xfree(c);
     }
 }

From 3e7923507b40b42c80dc621aff9d896106bae8c7 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 13 Dec 2017 15:26:29 +0100
Subject: [PATCH 095/198] Netlink: Use linpool instead of static buffer

---
 sysdep/linux/netlink.c | 39 +++++----------------------------------
 1 file changed, 5 insertions(+), 34 deletions(-)

diff --git a/sysdep/linux/netlink.c b/sysdep/linux/netlink.c
index bd3dbbd9..72ff13fa 100644
--- a/sysdep/linux/netlink.c
+++ b/sysdep/linux/netlink.c
@@ -645,13 +645,8 @@ nl_add_multipath(struct nlmsghdr *h, uint bufsize, struct nexthop *nh, int af)
 }
 
 static struct nexthop *
-nl_parse_multipath(struct krt_proto *p, struct rtattr *ra, int af)
+nl_parse_multipath(struct nl_parse_state *s, struct krt_proto *p, struct rtattr *ra, int af)
 {
-  /* Temporary buffer for multicast nexthops */
-  static struct nexthop *nh_buffer;
-  static int nh_buf_size;	/* in number of structures */
-  static int nh_buf_used;
-
   struct rtattr *a[BIRD_RTA_MAX];
   struct rtnexthop *nh = RTA_DATA(ra);
   struct nexthop *rv, *first, **last;
@@ -659,7 +654,6 @@ nl_parse_multipath(struct krt_proto *p, struct rtattr *ra, int af)
 
   first = NULL;
   last = &first;
-  nh_buf_used = 0;
 
   while (len)
     {
@@ -667,18 +661,9 @@ nl_parse_multipath(struct krt_proto *p, struct rtattr *ra, int af)
       if ((len < sizeof(*nh)) || (len < nh->rtnh_len))
 	return NULL;
 
-      if (nh_buf_used == nh_buf_size)
-      {
-	nh_buf_size = nh_buf_size ? (nh_buf_size * 2) : 4;
-	nh_buffer = xrealloc(nh_buffer, nh_buf_size * NEXTHOP_MAX_SIZE);
-      }
-      /* FIXME: This is really ugly */
-      *last = rv = (void *) (((byte *) nh_buffer) + (nh_buf_used++ * NEXTHOP_MAX_SIZE));
-      memset(rv, 0, NEXTHOP_MAX_SIZE);
-      // rv->next = NULL;
+      *last = rv = lp_allocz(s->pool, NEXTHOP_MAX_SIZE);
       last = &(rv->next);
 
-      rv->flags = 0;
       rv->weight = nh->rtnh_hops;
       rv->iface = if_find_by_index(nh->rtnh_ifindex);
       if (!rv->iface)
@@ -1395,20 +1380,6 @@ krt_replace_rte(struct krt_proto *p, net *n, rte *new, rte *old, struct ea_list
     n->n.flags &= ~KRF_SYNC_ERROR;
 }
 
-
-static inline struct nexthop *
-nl_alloc_nexthop(struct nl_parse_state *s, ip_addr gw, struct iface *iface, byte weight)
-{
-  struct nexthop *nh = lp_alloc(s->pool, sizeof(struct nexthop));
-
-  nh->gw = gw;
-  nh->iface = iface;
-  nh->next = NULL;
-  nh->weight = weight;
-
-  return nh;
-}
-
 static int
 nl_mergable_route(struct nl_parse_state *s, net *net, struct krt_proto *p, uint priority, uint krt_type)
 {
@@ -1597,7 +1568,7 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h)
 
       if (a[RTA_MULTIPATH])
         {
-	  struct nexthop *nh = nl_parse_multipath(p, a[RTA_MULTIPATH], i->rtm_family);
+	  struct nexthop *nh = nl_parse_multipath(s, p, a[RTA_MULTIPATH], i->rtm_family);
 	  if (!nh)
 	    {
 	      log(L_ERR "KRT: Received strange multipath route %N", net->n.addr);
@@ -1784,8 +1755,8 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h)
    * Ideally, now we would send the received route to the rest of kernel code.
    * But IPv6 ECMP routes before 4.11 are sent as a sequence of routes, so we
    * postpone it and merge next hops until the end of the sequence. Note that
-   * proper multipath updates are rejected by nl_mergable_route(), so it is
-   * always the first case for them.
+   * when doing merging of next hops, we expect the new route to be unipath.
+   * Otherwise, we ignore additional next hops in nexthop_insert().
    */
 
   if (!s->net)

From d807ea087f8d60e25eaef8c10168a40ca6545c57 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 13 Dec 2017 15:57:44 +0100
Subject: [PATCH 096/198] BGP: Fix non-transitive ext communities

---
 nest/a-set.c      | 7 +++++++
 nest/attrs.h      | 2 ++
 proto/bgp/attrs.c | 4 +++-
 3 files changed, 12 insertions(+), 1 deletion(-)

diff --git a/nest/a-set.c b/nest/a-set.c
index a2fb6953..048e522d 100644
--- a/nest/a-set.c
+++ b/nest/a-set.c
@@ -536,6 +536,13 @@ ec_set_sort(struct linpool *pool, struct adata *src)
   return dst;
 }
 
+void
+ec_set_sort_x(struct adata *set)
+{
+  /* Sort in place */
+  qsort(set->data, set->length / 8, 8, ec_set_cmp);
+}
+
 
 static int
 lc_set_cmp(const void *X, const void *Y)
diff --git a/nest/attrs.h b/nest/attrs.h
index f66d4f04..102f378a 100644
--- a/nest/attrs.h
+++ b/nest/attrs.h
@@ -197,4 +197,6 @@ struct adata *int_set_sort(struct linpool *pool, struct adata *src);
 struct adata *ec_set_sort(struct linpool *pool, struct adata *src);
 struct adata *lc_set_sort(struct linpool *pool, struct adata *src);
 
+void ec_set_sort_x(struct adata *set); /* Sort in place */
+
 #endif
diff --git a/proto/bgp/attrs.c b/proto/bgp/attrs.c
index dea3c4a6..fdc981ca 100644
--- a/proto/bgp/attrs.c
+++ b/proto/bgp/attrs.c
@@ -550,10 +550,12 @@ bgp_decode_mp_unreach_nlri(struct bgp_parse_state *s, uint code UNUSED, uint fla
 static void
 bgp_export_ext_community(struct bgp_export_state *s, eattr *a)
 {
+  a->u.ptr = ec_set_del_nontrans(s->pool, a->u.ptr);
+
   if (a->u.ptr->length == 0)
     UNSET(a);
 
-  a->u.ptr = ec_set_sort(s->pool, a->u.ptr);
+  ec_set_sort_x(a->u.ptr);
 }
 
 static void

From c36a298c21f1aa5b4e61d2d4740811529175df09 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 13 Dec 2017 19:18:30 +0100
Subject: [PATCH 097/198] Use git describe for BIRD version

Based on patch from Pavel Tvrdik
---
 Makefile.in     | 5 +++++
 sysdep/config.h | 7 +++++++
 2 files changed, 12 insertions(+)

diff --git a/Makefile.in b/Makefile.in
index e1f8d9c2..fdd5e6c7 100644
--- a/Makefile.in
+++ b/Makefile.in
@@ -21,6 +21,11 @@ INSTALL=@INSTALL@
 INSTALL_PROGRAM=@INSTALL_PROGRAM@
 INSTALL_DATA=@INSTALL_DATA@
 
+git-label:=$(strip $(shell git describe --always --dirty=-x 2>/dev/null))
+ifneq ($(git-label),)
+        CFLAGS += -DGIT_LABEL="$(git-label)"
+endif
+
 client=$(addprefix $(exedir)/,@CLIENT@)
 daemon=$(exedir)/bird
 protocols=@protocols@
diff --git a/sysdep/config.h b/sysdep/config.h
index 460cac8b..96e371f8 100644
--- a/sysdep/config.h
+++ b/sysdep/config.h
@@ -6,8 +6,15 @@
 #ifndef _BIRD_CONFIG_H_
 #define _BIRD_CONFIG_H_
 
+#define XSTR2(X) #X
+#define XSTR1(X) XSTR2(X)
+
 /* BIRD version */
+#ifdef GIT_LABEL
+#define BIRD_VERSION XSTR1(GIT_LABEL)
+#else
 #define BIRD_VERSION "2.0.0"
+#endif
 
 /* Include parameters determined by configure script */
 #include "sysdep/autoconf.h"

From abd4367f483bc29a03cce8bcfc70a83ac5d13cdc Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Thu, 14 Dec 2017 21:52:07 +0100
Subject: [PATCH 098/198] Minor cleanup

---
 lib/timer.c | 1 +
 nest/bfd.h  | 2 +-
 2 files changed, 2 insertions(+), 1 deletion(-)

diff --git a/lib/timer.c b/lib/timer.c
index 05e488c1..ed731d26 100644
--- a/lib/timer.c
+++ b/lib/timer.c
@@ -28,6 +28,7 @@
 
 #include <stdio.h>
 #include <stdlib.h>
+#include <time.h>
 
 #include "nest/bird.h"
 
diff --git a/nest/bfd.h b/nest/bfd.h
index b0ebe31d..04d6c001 100644
--- a/nest/bfd.h
+++ b/nest/bfd.h
@@ -46,7 +46,7 @@ static inline void cf_check_bfd(int use UNUSED) { }
 
 #else
 
-static inline struct bfd_request * bfd_request_session(pool *p, ip_addr addr, ip_addr local, struct iface *iface, void (*hook)(struct bfd_request *), void *data) { return NULL; }
+static inline struct bfd_request * bfd_request_session(pool *p UNUSED, ip_addr addr UNUSED, ip_addr local UNUSED, struct iface *iface UNUSED, void (*hook)(struct bfd_request *) UNUSED, void *data UNUSED) { return NULL; }
 
 static inline void cf_check_bfd(int use) { if (use) cf_error("BFD not available"); }
 

From 8396094156c2ddae279ec9c265100e28acb65a8d Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Thu, 14 Dec 2017 22:15:01 +0100
Subject: [PATCH 099/198] Minor cleanups

---
 nest/config.Y     | 2 +-
 proto/radv/radv.c | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/nest/config.Y b/nest/config.Y
index f51be6ea..358c7745 100644
--- a/nest/config.Y
+++ b/nest/config.Y
@@ -227,7 +227,7 @@ proto_item:
  | IMPORT LIMIT limit_spec { this_proto->in_limit = $3; }
  | EXPORT LIMIT limit_spec { this_proto->out_limit = $3; }
  | IMPORT KEEP FILTERED bool { this_proto->in_keep_filtered = $4; }
- | VRF TEXT { this_proto->vrf = if_get_by_name($2); }
+ | VRF text { this_proto->vrf = if_get_by_name($2); }
  | TABLE rtable { this_proto->table = $2; }
  | ROUTER ID idval { this_proto->router_id = $3; }
  | DESCRIPTION text { this_proto->dsc = $2; }
diff --git a/proto/radv/radv.c b/proto/radv/radv.c
index d53b65e3..7e8950c5 100644
--- a/proto/radv/radv.c
+++ b/proto/radv/radv.c
@@ -169,7 +169,7 @@ radv_prepare_prefixes(struct radv_iface *ifa)
   {
     if (pfx->valid && !pfx->mark)
     {
-      RADV_TRACE(D_EVENTS, "Invalidating prefix %I/$d on %s",
+      RADV_TRACE(D_EVENTS, "Invalidating prefix %I/%d on %s",
 		 pfx->prefix, pfx->len, ifa->iface->name);
 
       pfx->valid = 0;

From 3013fc57bdc97b0031c4f4250b9bdd172f4c0f4e Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Sat, 16 Dec 2017 00:42:56 +0100
Subject: [PATCH 100/198] Netlink: Fix memory leak

---
 sysdep/linux/netlink.c | 11 -----------
 1 file changed, 11 deletions(-)

diff --git a/sysdep/linux/netlink.c b/sysdep/linux/netlink.c
index 72ff13fa..4cb51519 100644
--- a/sysdep/linux/netlink.c
+++ b/sysdep/linux/netlink.c
@@ -1666,14 +1666,6 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h)
     ra->nh.labels = labels;
 #endif
 
-  rte *e = rte_get_temp(ra);
-  e->net = net;
-  e->u.krt.src = src;
-  e->u.krt.proto = i->rtm_protocol;
-  e->u.krt.seen = 0;
-  e->u.krt.best = 0;
-  e->u.krt.metric = 0;
-
   if (i->rtm_scope != def_scope)
     {
       ea_list *ea = lp_alloc(s->pool, sizeof(ea_list) + sizeof(eattr));
@@ -1687,9 +1679,6 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h)
       ea->attrs[0].u.data = i->rtm_scope;
     }
 
-  if (a[RTA_PRIORITY])
-    e->u.krt.metric = rta_get_u32(a[RTA_PRIORITY]);
-
   if (a[RTA_PREFSRC])
     {
       ip_addr ps = rta_get_ipa(a[RTA_PREFSRC]);

From e87a95d97d18eb12d005312f27b082f0821d3923 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Sat, 16 Dec 2017 16:31:43 +0100
Subject: [PATCH 101/198] Minor fixes for debug mode

---
 nest/iface.c      |  3 ++-
 nest/rt-fib.c     | 12 +++---------
 proto/bfd/io.c    |  2 +-
 proto/ospf/ospf.c |  2 +-
 4 files changed, 7 insertions(+), 12 deletions(-)

diff --git a/nest/iface.c b/nest/iface.c
index 303a52b2..54c16c58 100644
--- a/nest/iface.c
+++ b/nest/iface.c
@@ -93,7 +93,8 @@ if_dump(struct iface *i)
   WALK_LIST(a, i->addrs)
     {
       ifa_dump(a);
-      ASSERT((a != i->addr) == !(a->flags & IA_PRIMARY));
+      ASSERT(!!(a->flags & IA_PRIMARY) ==
+	     ((a == i->addr4) || (a == i->addr6) || (a == i->llv6)));
     }
 }
 
diff --git a/nest/rt-fib.c b/nest/rt-fib.c
index 60c3da55..24a7facc 100644
--- a/nest/rt-fib.c
+++ b/nest/rt-fib.c
@@ -568,22 +568,17 @@ found:
 void
 fib_check(struct fib *f)
 {
-#if 0
-  uint i, ec, lo, nulls;
+  uint i, ec, nulls;
 
   ec = 0;
   for(i=0; i<f->hash_size; i++)
     {
       struct fib_node *n;
-      lo = 0;
       for(n=f->hash_table[i]; n; n=n->next)
 	{
 	  struct fib_iterator *j, *j0;
-	  uint h0 = ipa_hash(n->prefix);
-	  if (h0 < lo)
-	    bug("fib_check: discord in hash chains");
-	  lo = h0;
-	  if ((h0 >> f->hash_shift) != i)
+	  uint h0 = fib_hash(f, n->addr);
+	  if (h0 != i)
 	    bug("fib_check: mishashed %x->%x (order %d)", h0, i, f->hash_order);
 	  j0 = (struct fib_iterator *) n;
 	  nulls = 0;
@@ -604,7 +599,6 @@ fib_check(struct fib *f)
     }
   if (ec != f->entries)
     bug("fib_check: invalid entry count (%d != %d)", ec, f->entries);
-#endif
   return;
 }
 
diff --git a/proto/bfd/io.c b/proto/bfd/io.c
index b01cbfce..1cd9365a 100644
--- a/proto/bfd/io.c
+++ b/proto/bfd/io.c
@@ -298,7 +298,7 @@ sockets_prepare(struct birdloop *loop)
 
   struct pollfd *pfd = loop->poll_fd.data;
   sock **psk = loop->poll_sk.data;
-  int i = 0;
+  uint i = 0;
   node *n;
 
   WALK_LIST(n, loop->sock_list)
diff --git a/proto/ospf/ospf.c b/proto/ospf/ospf.c
index 3ebebdaa..df6c452e 100644
--- a/proto/ospf/ospf.c
+++ b/proto/ospf/ospf.c
@@ -1179,7 +1179,7 @@ ospf_sh_state(struct proto *P, int verbose, int reachable)
   /* We store interesting area-scoped LSAs in array hea and
      global-scoped (LSA_T_EXT) LSAs in array hex */
 
-  int num = p->gr->hash_entries;
+  uint num = p->gr->hash_entries;
   struct top_hash_entry *hea[num];
   struct top_hash_entry *hex[verbose ? num : 0];
   struct top_hash_entry *he;

From a63e78c31a518f4441b2f87e1947c12e1e49aba1 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Mon, 18 Dec 2017 23:15:07 +0100
Subject: [PATCH 102/198] Fix build without limited protocol set

---
 sysdep/unix/krt.Y | 1 +
 1 file changed, 1 insertion(+)

diff --git a/sysdep/unix/krt.Y b/sysdep/unix/krt.Y
index 6d90db64..1cb28389 100644
--- a/sysdep/unix/krt.Y
+++ b/sysdep/unix/krt.Y
@@ -40,6 +40,7 @@ kif_set_preferred(ip_addr ip)
 CF_DECLS
 
 CF_KEYWORDS(KERNEL, PERSIST, SCAN, TIME, LEARN, DEVICE, ROUTES, GRACEFUL, RESTART, KRT_SOURCE, KRT_METRIC, MERGE, PATHS)
+CF_KEYWORDS(INTERFACE, PREFERRED)
 
 %type <i> kern_mp_limit
 

From 4842eeaad32aeeeb1774ec9840c6da03952425ff Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Thu, 21 Dec 2017 00:16:52 +0100
Subject: [PATCH 103/198] Minor fix in documentation

---
 doc/bird.sgml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index ebafcf92..c725081a 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -1348,7 +1348,7 @@ foot).
 
 	<cf><m/P/.len</cf> returns the length of path <m/P/.
 
-	<cf><m/P/.empty()</cf> makes the path <m/P/ empty.
+	<cf><m/P/.empty</cf> makes the path <m/P/ empty.
 
 	<cf>prepend(<m/P/,<m/A/)</cf> prepends ASN <m/A/ to path <m/P/ and
 	returns the result.

From cce6ba4daa2e00d6e1881a831d05cdc51249b4a7 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 2 Jan 2018 14:11:59 +0100
Subject: [PATCH 104/198] Remove libhistory check

According to GNU Readline developers, if we link with libreadline then
there is no need to link with libhistory at all.
---
 configure.ac | 8 +-------
 1 file changed, 1 insertion(+), 7 deletions(-)

diff --git a/configure.ac b/configure.ac
index 91f89edf..d135dc04 100644
--- a/configure.ac
+++ b/configure.ac
@@ -361,12 +361,6 @@ if test "$enable_client" = yes ; then
     [$TINFO_LIBS]
   )
 
-  AC_SEARCH_LIBS([add_history], [history readline],
-    [HISTORY_LIBS="$LIBS"; LIBS=""],
-    [AC_MSG_ERROR([The client requires GNU Readline library. Either install the library or use --disable-client to compile without the client.])],
-    [$TINFO_LIBS]
-  )
-
   AC_CHECK_LIB([readline], [rl_crlf],
     [AC_DEFINE([HAVE_RL_CRLF], [1], [Define to 1 if you have rl_crlf()])],
     [],
@@ -380,7 +374,7 @@ if test "$enable_client" = yes ; then
   )
 
   LIBS="$BASE_LIBS"
-  CLIENT_LIBS="$HISTORY_LIBS $READLINE_LIBS $TINFO_LIBS"
+  CLIENT_LIBS="$READLINE_LIBS $TINFO_LIBS"
 fi
 AC_SUBST([CLIENT])
 AC_SUBST([CLIENT_LIBS])

From e62cd033079c4bc988a467f4122c7c276c77fdde Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 2 Jan 2018 14:30:08 +0100
Subject: [PATCH 105/198] BGP: Fix graceful restart timer

Should use remote value, not local value.
---
 proto/bgp/bgp.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/proto/bgp/bgp.c b/proto/bgp/bgp.c
index 705c5e91..8da5b0aa 100644
--- a/proto/bgp/bgp.c
+++ b/proto/bgp/bgp.c
@@ -689,8 +689,11 @@ bgp_handle_graceful_restart(struct bgp_proto *p)
     c->packets_to_send = 0;
   }
 
+  /* p->gr_ready -> at least one active channel is c->gr_ready */
+  ASSERT(p->gr_active_num > 0);
+
   proto_notify_state(&p->p, PS_START);
-  bgp_start_timer(p->gr_timer, p->conn->local_caps->gr_time);
+  bgp_start_timer(p->gr_timer, p->conn->remote_caps->gr_time);
 }
 
 /**

From d493d0f180e3df1f33d344d8b28cc1743201210b Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 2 Jan 2018 16:57:45 +0100
Subject: [PATCH 106/198] BGP: Fix unknown attribute handling

---
 nest/route.h        | 40 ++++++++++++++++++++++++++++++++++++++++
 proto/bgp/attrs.c   |  3 ++-
 proto/bgp/packets.c |  2 ++
 3 files changed, 44 insertions(+), 1 deletion(-)

diff --git a/nest/route.h b/nest/route.h
index 792f79dd..43d7d696 100644
--- a/nest/route.h
+++ b/nest/route.h
@@ -555,6 +555,46 @@ uint ea_hash(ea_list *e);	/* Calculate 16-bit hash value */
 ea_list *ea_append(ea_list *to, ea_list *what);
 void ea_format_bitfield(struct eattr *a, byte *buf, int bufsize, const char **names, int min, int max);
 
+static inline eattr *
+ea_set_attr(ea_list **to, struct linpool *pool, uint id, uint flags, uint type, uintptr_t val)
+{
+  ea_list *a = lp_alloc(pool, sizeof(ea_list) + sizeof(eattr));
+  eattr *e = &a->attrs[0];
+
+  a->flags = EALF_SORTED;
+  a->count = 1;
+  a->next = *to;
+  *to = a;
+
+  e->id = id;
+  e->type = type;
+  e->flags = flags;
+
+  if (type & EAF_EMBEDDED)
+    e->u.data = (u32) val;
+  else
+    e->u.ptr = (struct adata *) val;
+
+  return e;
+}
+
+static inline void
+ea_set_attr_u32(ea_list **to, struct linpool *pool, uint id, uint flags, uint type, u32 val)
+{ ea_set_attr(to, pool, id, flags, type, (uintptr_t) val); }
+
+static inline void
+ea_set_attr_ptr(ea_list **to, struct linpool *pool, uint id, uint flags, uint type, struct adata *val)
+{ ea_set_attr(to, pool, id, flags, type, (uintptr_t) val); }
+
+static inline void
+ea_set_attr_data(ea_list **to, struct linpool *pool, uint id, uint flags, uint type, void *data, uint len)
+{
+  struct adata *a = lp_alloc_adata(pool, len);
+  memcpy(a->data, data, len);
+  ea_set_attr(to, pool, id, flags, type, (uintptr_t) a);
+}
+
+
 #define NEXTHOP_MAX_SIZE (sizeof(struct nexthop) + sizeof(u32)*MPLS_MAX_LABEL_STACK)
 
 static inline size_t nexthop_size(const struct nexthop *nh)
diff --git a/proto/bgp/attrs.c b/proto/bgp/attrs.c
index fdc981ca..0f41f818 100644
--- a/proto/bgp/attrs.c
+++ b/proto/bgp/attrs.c
@@ -701,7 +701,8 @@ bgp_format_mpls_label_stack(eattr *a, byte *buf, uint size)
 static inline void
 bgp_decode_unknown(struct bgp_parse_state *s, uint code, uint flags, byte *data, uint len, ea_list **to)
 {
-  bgp_set_attr_data(to, s->pool, code, flags, data, len);
+  /* Cannot use bgp_set_attr_data() as it works on known attributes only */
+  ea_set_attr_data(to, s->pool, EA_CODE(EAP_BGP, code), flags, EAF_TYPE_OPAQUE, data, len);
 }
 
 
diff --git a/proto/bgp/packets.c b/proto/bgp/packets.c
index ca942880..95a974eb 100644
--- a/proto/bgp/packets.c
+++ b/proto/bgp/packets.c
@@ -2284,6 +2284,8 @@ bgp_rx_update(struct bgp_conn *conn, byte *pkt, uint len)
 
   if (s.attr_len)
     ea = bgp_decode_attrs(&s, s.attrs, s.attr_len);
+  else
+    ea = NULL;
 
   /* Check for End-of-RIB marker */
   if (!s.attr_len && !s.ip_unreach_len && !s.ip_reach_len)

From 0ff86d054efa8005c5df943acf6d2122781d3175 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 3 Jan 2018 14:12:00 +0100
Subject: [PATCH 107/198] ROA: Fix reconfiguration

---
 nest/rt-roa.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/nest/rt-roa.c b/nest/rt-roa.c
index 0fd89667..bf457e30 100644
--- a/nest/rt-roa.c
+++ b/nest/rt-roa.c
@@ -306,10 +306,10 @@ void
 roa_commit(struct config *new, struct config *old)
 {
   struct roa_table_config *cf;
-  struct roa_table *t;
+  struct roa_table *t, *tx;
 
   if (old)
-    WALK_LIST(t, roa_table_list)
+    WALK_LIST_DELSAFE(t, tx, roa_table_list)
       {
 	struct symbol *sym = cf_find_symbol(new, t->name);
 	if (sym && sym->class == SYM_ROA)

From 8f8671bcde738890d312ab91336f101465b3eeac Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 3 Jan 2018 15:44:05 +0100
Subject: [PATCH 108/198] Filter: Handle undefined BGP paths as empty

The same is already done for clists. Also fixes defined() to work
properly for paths and clists.
---
 doc/bird.sgml   |  4 ++--
 filter/filter.c | 28 +++++++++++++++++++++++-----
 2 files changed, 25 insertions(+), 7 deletions(-)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index c725081a..3a2607b4 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -1498,8 +1498,8 @@ if 1234 = i then printn "."; else {
 like it accesses variables. Attempts to access undefined attribute result in a
 runtime error; you can check if an attribute is defined by using the
 <cf>defined( <m>attribute</m> )</cf> operator. One notable exception to this
-rule are attributes of clist type, where undefined value is regarded as empty
-clist for most purposes.
+rule are attributes of bgppath and *clist types, where undefined value is
+regarded as empty bgppath/*clist for most purposes.
 
 <descrip>
 	<tag><label id="rta-net"><m/prefix/ net</tag>
diff --git a/filter/filter.c b/filter/filter.c
index aef4f44a..4e17f974 100644
--- a/filter/filter.c
+++ b/filter/filter.c
@@ -54,6 +54,17 @@
 
 void (*bt_assert_hook)(int result, struct f_inst *assert);
 
+static struct adata undef_adata;	/* adata of length 0 used for undefined */
+
+/* Special undef value for paths and clists */
+static inline int
+undef_value(struct f_val v)
+{
+  return ((v.type == T_PATH) || (v.type == T_CLIST) ||
+	  (v.type == T_ECLIST) || (v.type == T_LCLIST)) &&
+    (v.val.ad == &undef_adata);
+}
+
 static struct adata *
 adata_empty(struct linpool *pool, int l)
 {
@@ -815,7 +826,7 @@ interpret(struct f_inst *what)
   case P('d','e'):
     ONEARG;
     res.type = T_BOOL;
-    res.val.i = (v1.type != T_VOID);
+    res.val.i = (v1.type != T_VOID) && !undef_value(v1);
     break;
   case 'T':
     ONEARG;
@@ -1001,24 +1012,31 @@ interpret(struct f_inst *what)
 	e = ea_find((*f_rte)->attrs->eattrs, code);
 
       if (!e) {
-	/* A special case: undefined int_set looks like empty int_set */
+	/* A special case: undefined as_path looks like empty as_path */
+	if ((what->aux & EAF_TYPE_MASK) == EAF_TYPE_AS_PATH) {
+	  res.type = T_PATH;
+	  res.val.ad = &undef_adata;
+	  break;
+	}
+
+	/* The same special case for int_set */
 	if ((what->aux & EAF_TYPE_MASK) == EAF_TYPE_INT_SET) {
 	  res.type = T_CLIST;
-	  res.val.ad = adata_empty(f_pool, 0);
+	  res.val.ad = &undef_adata;
 	  break;
 	}
 
 	/* The same special case for ec_set */
 	if ((what->aux & EAF_TYPE_MASK) == EAF_TYPE_EC_SET) {
 	  res.type = T_ECLIST;
-	  res.val.ad = adata_empty(f_pool, 0);
+	  res.val.ad = &undef_adata;
 	  break;
 	}
 
 	/* The same special case for lc_set */
 	if ((what->aux & EAF_TYPE_MASK) == EAF_TYPE_LC_SET) {
 	  res.type = T_LCLIST;
-	  res.val.ad = adata_empty(f_pool, 0);
+	  res.val.ad = &undef_adata;
 	  break;
 	}
 

From 94f9be80c3686284942ba73670780d27730da997 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 9 Jan 2018 14:36:11 +0100
Subject: [PATCH 109/198] Nest: Fix filter reconfiguration

Function filter_same() must be called with arguments in proper order,
otherwise it breaks the new filter, causing crash during route
processing.
---
 nest/proto.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/nest/proto.c b/nest/proto.c
index d7650517..e103fec6 100644
--- a/nest/proto.c
+++ b/nest/proto.c
@@ -513,8 +513,9 @@ channel_reconfigure(struct channel *c, struct channel_config *cf)
   if ((c->table != cf->table->table) || (cf->ra_mode && (c->ra_mode != cf->ra_mode)))
     return 0;
 
-  int import_changed = !filter_same(c->in_filter, cf->in_filter);
-  int export_changed = !filter_same(c->out_filter, cf->out_filter);
+  /* Note that filter_same() requires arguments in (new, old) order */
+  int import_changed = !filter_same(cf->in_filter, c->in_filter);
+  int export_changed = !filter_same(cf->out_filter, c->out_filter);
 
   if (c->preference != cf->preference)
     import_changed = 1;

From 09c1e370b3084f7acb7c3777427670a69945368a Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Tue, 9 Jan 2018 16:46:00 +0100
Subject: [PATCH 110/198] Moved freebsd cflags and ldflags to configure

---
 .gitlab-ci.yml | 6 ------
 configure.ac   | 2 ++
 2 files changed, 2 insertions(+), 6 deletions(-)

diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
index 482c5004..ff11dda0 100644
--- a/.gitlab-ci.yml
+++ b/.gitlab-ci.yml
@@ -211,9 +211,6 @@ docker_ubuntu-16_04-amd64:
 
 # TODO We want to copy these BSDs to our own virtual machines, to make sure someone doesn't update them by accident.
 .freebsd-11-i386: &freebsd-11-i386_env
-  variables:
-    CPPFLAGS: "-I/usr/local/include"
-    LDFLAGS: "-L/usr/local/lib"
   tags:
   - freebsd
   - i386
@@ -223,9 +220,6 @@ docker_ubuntu-16_04-amd64:
   #- tags
 
 .freebsd-11-amd64: &freebsd-11-amd64_env
-  variables:
-    CPPFLAGS: "-I/usr/local/include"
-    LDFLAGS: "-L/usr/local/lib"
   tags:
   - freebsd
   - amd64
diff --git a/configure.ac b/configure.ac
index f73e6264..f4de8f93 100644
--- a/configure.ac
+++ b/configure.ac
@@ -181,6 +181,8 @@ else
       ;;
     freebsd*)
       sysdesc=bsd
+      CPPFLAGS="$CPPFLAGS -I/usr/local/include"
+      LDFLAGS="$LDFLAGS -L/usr/local/lib"
       ;;
     kfreebsd*)
       sysdesc=bsd

From 72163bd5f3ccefc1edda585f6f605c37e774a0b8 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 9 Jan 2018 18:42:22 +0100
Subject: [PATCH 111/198] Nest: Allow modification of channels inherited from
 templates

Multiple definitions of same channels are forbidden, but inherited
channel can be redefined. In such case channel options are merged.
---
 doc/bird.sgml       | 66 +++++++++++++++++++++++++++++++++------------
 nest/config.Y       |  2 +-
 nest/proto.c        | 25 ++++++++++++++---
 nest/protocol.h     |  4 ++-
 proto/bgp/config.Y  | 16 ++++++-----
 proto/ospf/config.Y |  7 ++---
 proto/pipe/config.Y |  2 +-
 proto/radv/radv.c   |  2 +-
 proto/rip/rip.c     |  2 +-
 proto/rpki/rpki.c   |  2 +-
 10 files changed, 93 insertions(+), 35 deletions(-)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index 3a2607b4..bde825c3 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -407,31 +407,33 @@ extensive way.
 a comment, whitespace characters are treated as a single space. If there's a
 variable number of options, they are grouped using the <cf/{ }/ brackets. Each
 option is terminated by a <cf/;/. Configuration is case sensitive. There are two
-ways how to name symbols (like protocol names, filter names, constants etc.). You
-can either use a simple string starting with a letter followed by any
-combination of letters and numbers (e.g. <cf/R123/, <cf/myfilter/, <cf/bgp5/) or you can
-enclose the name into apostrophes (<cf/'/) and than you can use any combination
-of numbers, letters. hyphens, dots and colons (e.g. <cf/'1:strange-name'/,
-<cf/'-NAME-'/, <cf/'cool::name'/).
+ways how to name symbols (like protocol names, filter names, constants etc.).
+You can either use a simple string starting with a letter followed by any
+combination of letters and numbers (e.g. <cf/R123/, <cf/myfilter/, <cf/bgp5/) or
+you can enclose the name into apostrophes (<cf/'/) and than you can use any
+combination of numbers, letters. hyphens, dots and colons (e.g.
+<cf/'1:strange-name'/, <cf/'-NAME-'/, <cf/'cool::name'/).
 
 <p>Here is an example of a simple config file. It enables synchronization of
-routing tables with OS kernel, scans for new network interfaces every 10 seconds
-and runs RIP on all network interfaces found.
+routing tables with OS kernel, learns network interfaces and runs RIP on all
+network interfaces found.
 
 <code>
 protocol kernel {
+	ipv4 {
+		export all;	# Default is export none
+	};
 	persist;		# Don't remove routes on BIRD shutdown
-	scan time 20;		# Scan kernel routing table every 20 seconds
-	export all;		# Default is export none
 }
 
 protocol device {
-	scan time 10;		# Scan interfaces every 10 seconds
 }
 
 protocol rip {
-	export all;
-	import all;
+	ipv4 {
+		import all;
+		export all;
+	};
 	interface "*";
 }
 </code>
@@ -775,8 +777,10 @@ agreement").
 <label id="channel-opts">
 
 <p>Every channel belongs to a protocol and is configured inside its block. The
-minimal channel config is empty, then it uses the default values. The name of
-the channel implies its nettype.
+minimal channel config is empty, then it uses default values. The name of the
+channel implies its nettype. Channel definitions can be inherited from protocol
+templates. Multiple definitions of the same channel are forbidden, but channels
+inherited from templates can be updated by new definitions.
 
 <descrip>
 	<tag><label id="proto-table">table <m/name/</tag>
@@ -841,7 +845,7 @@ protocol rip ng {
 }
 </code>
 
-<p>And this is a non-trivial example.
+<p>This is a non-trivial example.
 <code>
 protocol rip ng {
 	ipv6 {
@@ -854,6 +858,33 @@ protocol rip ng {
 }
 </code>
 
+<p>And this is even more complicated example using templates.
+<code>
+template bgp {
+	local 198.51.100.14 as 65000;
+
+	ipv4 {
+		table mytable4;
+		import filter { ... };
+	};
+	ipv6 {
+		table mytable6;
+		import filter { ... };
+	};
+}
+
+protocol bgp from  {
+	neighbor 198.51.100.130 as 64496;
+
+	# IPv4 channel is inherited as-is, while IPv6
+	# channel is adjusted by export filter option
+	ipv6 {
+		export filter { ... };
+	};
+}
+</code>
+
+
 <chapt>Remote control
 <label id="remote-control">
 
@@ -1051,6 +1082,7 @@ This argument can be omitted if there exists only a single instance.
 	Evaluate given expression.
 </descrip>
 
+
 <chapt>Filters
 <label id="filters">
 
@@ -2427,7 +2459,7 @@ together with their appropriate channels follows.
 </tabular>
 </table>
 
-<p>BGP's channels have additional config options (together with the common ones):
+<p>BGP channels have additional config options (together with the common ones):
 
 <descrip>
 	<tag><label id="bgp-next-hop-keep">next hop keep</tag>
diff --git a/nest/config.Y b/nest/config.Y
index 5c4f2393..044aba2b 100644
--- a/nest/config.Y
+++ b/nest/config.Y
@@ -214,7 +214,7 @@ proto_item:
 
 channel_start: net_type
 {
-  $$ = this_channel = channel_config_new(NULL, $1, this_proto);
+  $$ = this_channel = channel_config_get(NULL, net_label[$1], $1, this_proto);
 };
 
 channel_item:
diff --git a/nest/proto.c b/nest/proto.c
index e103fec6..d584cb93 100644
--- a/nest/proto.c
+++ b/nest/proto.c
@@ -455,11 +455,10 @@ const struct channel_class channel_basic = {
 };
 
 void *
-channel_config_new(const struct channel_class *cc, uint net_type, struct proto_config *proto)
+channel_config_new(const struct channel_class *cc, const char *name, uint net_type, struct proto_config *proto)
 {
   struct channel_config *cf = NULL;
   struct rtable_config *tab = NULL;
-  const char *name = NULL;
 
   if (net_type)
   {
@@ -470,7 +469,6 @@ channel_config_new(const struct channel_class *cc, uint net_type, struct proto_c
       cf_error("Different channel type");
 
     tab = new_config->def_tables[net_type];
-    name = net_label[net_type];
   }
 
   if (!cc)
@@ -479,6 +477,7 @@ channel_config_new(const struct channel_class *cc, uint net_type, struct proto_c
   cf = cfg_allocz(cc->config_size);
   cf->name = name;
   cf->channel = cc;
+  cf->parent = proto;
   cf->table = tab;
   cf->out_filter = FILTER_REJECT;
 
@@ -491,6 +490,26 @@ channel_config_new(const struct channel_class *cc, uint net_type, struct proto_c
   return cf;
 }
 
+void *
+channel_config_get(const struct channel_class *cc, const char *name, uint net_type, struct proto_config *proto)
+{
+  struct channel_config *cf;
+
+  /* We are using name as token, so no strcmp() */
+  WALK_LIST(cf, proto->channels)
+    if (cf->name == name)
+    {
+      /* Allow to redefine channel only if inherited from template */
+      if (cf->parent == proto)
+	cf_error("Multiple %s channels", name);
+
+      cf->parent = proto;
+      return cf;
+    }
+
+  return channel_config_new(cc, name, net_type, proto);
+}
+
 struct channel_config *
 channel_copy_config(struct channel_config *src, struct proto_config *proto)
 {
diff --git a/nest/protocol.h b/nest/protocol.h
index c8f37367..9afd3a0a 100644
--- a/nest/protocol.h
+++ b/nest/protocol.h
@@ -461,6 +461,7 @@ struct channel_config {
   const char *name;
   const struct channel_class *channel;
 
+  struct proto_config *parent;		/* Where channel is defined (proto or template) */
   struct rtable_config *table;		/* Table we're attached to */
   struct filter *in_filter, *out_filter; /* Attached filters */
   struct channel_limit rx_limit;	/* Limit for receiving routes from protocol
@@ -585,7 +586,8 @@ static inline void channel_open(struct channel *c) { channel_set_state(c, CS_UP)
 static inline void channel_close(struct channel *c) { channel_set_state(c, CS_FLUSHING); }
 
 void channel_request_feeding(struct channel *c);
-void *channel_config_new(const struct channel_class *cc, uint net_type, struct proto_config *proto);
+void *channel_config_new(const struct channel_class *cc, const char *name, uint net_type, struct proto_config *proto);
+void *channel_config_get(const struct channel_class *cc, const char *name, uint net_type, struct proto_config *proto);
 int channel_reconfigure(struct channel *c, struct channel_config *cf);
 
 
diff --git a/proto/bgp/config.Y b/proto/bgp/config.Y
index 6ce0f1aa..4e819eb7 100644
--- a/proto/bgp/config.Y
+++ b/proto/bgp/config.Y
@@ -158,12 +158,16 @@ bgp_channel_start: bgp_afi
   if (!desc)
     cf_error("Unknown AFI/SAFI");
 
-  this_channel = channel_config_new(&channel_bgp, desc->net, this_proto);
-  BGP_CC->c.name = desc->name;
-  BGP_CC->c.ra_mode = RA_UNDEF;
-  BGP_CC->afi = $1;
-  BGP_CC->desc = desc;
-  BGP_CC->gr_able = 0xff;	/* undefined */
+  this_channel = channel_config_get(&channel_bgp, desc->name, desc->net, this_proto);
+
+  /* New channel */
+  if (!BGP_CC->desc)
+  {
+    BGP_CC->c.ra_mode = RA_UNDEF;
+    BGP_CC->afi = $1;
+    BGP_CC->desc = desc;
+    BGP_CC->gr_able = 0xff;	/* undefined */
+  }
 };
 
 bgp_channel_item:
diff --git a/proto/ospf/config.Y b/proto/ospf/config.Y
index ce409a3a..005f4381 100644
--- a/proto/ospf/config.Y
+++ b/proto/ospf/config.Y
@@ -84,8 +84,8 @@ ospf_proto_finish(void)
   /* Define default channel */
   if (EMPTY_LIST(this_proto->channels))
   {
-    this_proto->net_type = ospf_cfg_is_v2() ? NET_IP4 : NET_IP6;
-    channel_config_new(NULL, this_proto->net_type, this_proto);
+    uint net_type = this_proto->net_type = ospf_cfg_is_v2() ? NET_IP4 : NET_IP6;
+    channel_config_new(NULL, net_label[net_type], net_type, this_proto);
   }
 
   /* Propagate global instance ID to interfaces */
@@ -238,7 +238,8 @@ ospf_af_mc:
 /* We redefine proto_channel to add multicast flag */
 ospf_channel_start: net_type ospf_af_mc
 {
-  $$ = this_channel = channel_config_new(NULL, $1, this_proto);
+  /* TODO: change name for multicast channels */
+  $$ = this_channel = channel_config_get(NULL, net_label[$1], $1, this_proto);
 
   /* Save the multicast flag */
   if (this_channel == proto_cf_main_channel(this_proto))
diff --git a/proto/pipe/config.Y b/proto/pipe/config.Y
index f51ee575..4f96fdcb 100644
--- a/proto/pipe/config.Y
+++ b/proto/pipe/config.Y
@@ -25,7 +25,7 @@ CF_ADDTO(proto, pipe_proto '}' { this_channel = NULL; } )
 pipe_proto_start: proto_start PIPE
 {
   this_proto = proto_config_new(&proto_pipe, $1);
-  this_channel = channel_config_new(NULL, 0, this_proto);
+  this_channel = channel_config_new(NULL, NULL, 0, this_proto);
   this_channel->in_filter = FILTER_ACCEPT;
   this_channel->out_filter = FILTER_ACCEPT;
 };
diff --git a/proto/radv/radv.c b/proto/radv/radv.c
index 0a2a3e78..8a79dfaf 100644
--- a/proto/radv/radv.c
+++ b/proto/radv/radv.c
@@ -569,7 +569,7 @@ radv_postconfig(struct proto_config *CF)
 
   /* Define default channel */
   if (EMPTY_LIST(CF->channels))
-    channel_config_new(NULL, NET_IP6, CF);
+    channel_config_new(NULL, net_label[NET_IP6], NET_IP6, CF);
 }
 
 static struct proto *
diff --git a/proto/rip/rip.c b/proto/rip/rip.c
index a3eeaf17..85e37cea 100644
--- a/proto/rip/rip.c
+++ b/proto/rip/rip.c
@@ -1078,7 +1078,7 @@ rip_postconfig(struct proto_config *CF)
 
   /* Define default channel */
   if (EMPTY_LIST(CF->channels))
-    channel_config_new(NULL, CF->net_type, CF);
+    channel_config_new(NULL, net_label[CF->net_type], CF->net_type, CF);
 }
 
 static struct proto *
diff --git a/proto/rpki/rpki.c b/proto/rpki/rpki.c
index 3145399b..74860071 100644
--- a/proto/rpki/rpki.c
+++ b/proto/rpki/rpki.c
@@ -901,7 +901,7 @@ rpki_postconfig(struct proto_config *CF)
 {
   /* Define default channel */
   if (EMPTY_LIST(CF->channels))
-    channel_config_new(NULL, CF->net_type, CF);
+    channel_config_new(NULL, net_label[CF->net_type], CF->net_type, CF);
 }
 
 static void

From 2e507a745733b0cee2ea7d302cc9d626cb2e0c80 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 10 Jan 2018 16:17:37 +0100
Subject: [PATCH 112/198] Use non-fatal asserts even for regular build

---
 lib/birdlib.h              | 4 ++--
 nest/proto.c               | 2 +-
 proto/rpki/ssh_transport.c | 2 --
 proto/rpki/tcp_transport.c | 2 --
 4 files changed, 3 insertions(+), 7 deletions(-)

diff --git a/lib/birdlib.h b/lib/birdlib.h
index f41cceb5..428b3209 100644
--- a/lib/birdlib.h
+++ b/lib/birdlib.h
@@ -152,9 +152,9 @@ void debug(const char *msg, ...);	/* Printf to debug output */
 #endif
 
 #ifdef DEBUGGING
-#define ASSERT(x) do { if (!(x)) bug("Assertion `%s' failed at %s:%d", #x, __FILE__, __LINE__); } while(0)
+#define ASSERT(x) do { if (!(x)) bug("Assertion '%s' failed at %s:%d", #x, __FILE__, __LINE__); } while(0)
 #else
-#define ASSERT(x) do { } while(0)
+#define ASSERT(x) do { if (!(x)) log(L_BUG "Assertion '%s' failed at %s:%d", #x, __FILE__, __LINE__); } while(0)
 #endif
 
 /* Pseudorandom numbers */
diff --git a/nest/proto.c b/nest/proto.c
index d584cb93..15d6f4de 100644
--- a/nest/proto.c
+++ b/nest/proto.c
@@ -437,7 +437,7 @@ static void
 channel_request_reload(struct channel *c)
 {
   ASSERT(c->channel_state == CS_UP);
-  // ASSERT(channel_reloadable(c));
+  ASSERT(channel_reloadable(c));
 
   c->proto->reload_routes(c);
 
diff --git a/proto/rpki/ssh_transport.c b/proto/rpki/ssh_transport.c
index cd49ab90..13e061fc 100644
--- a/proto/rpki/ssh_transport.c
+++ b/proto/rpki/ssh_transport.c
@@ -42,8 +42,6 @@ rpki_tr_ssh_open(struct rpki_tr_sock *tr)
 static const char *
 rpki_tr_ssh_ident(struct rpki_tr_sock *tr)
 {
-  ASSERT(tr != NULL);
-
   struct rpki_cache *cache = tr->cache;
   struct rpki_config *cf = (void *) cache->p->p.cf;
   struct rpki_tr_ssh_config *ssh_cf = (void *) cf->tr_config.spec;
diff --git a/proto/rpki/tcp_transport.c b/proto/rpki/tcp_transport.c
index 6c05964a..0d3c7e99 100644
--- a/proto/rpki/tcp_transport.c
+++ b/proto/rpki/tcp_transport.c
@@ -37,8 +37,6 @@ rpki_tr_tcp_open(struct rpki_tr_sock *tr)
 static const char *
 rpki_tr_tcp_ident(struct rpki_tr_sock *tr)
 {
-  ASSERT(tr != NULL);
-
   struct rpki_cache *cache = tr->cache;
   struct rpki_config *cf = (void *) cache->p->p.cf;
 

From 4d3679613179637df5ef999073aba2d187540a57 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Sun, 14 Jan 2018 14:30:38 +0100
Subject: [PATCH 113/198] KRT: Fix direct routes for BSD

Old way to set direct routes is to use local IP as gateway, but that does
not work properly on newer FreeBSDs. Now we use sockaddr_dl containing
interface index as gateway.
---
 sysdep/bsd/krt-sock.c | 28 ++++++++++++++++++++++------
 1 file changed, 22 insertions(+), 6 deletions(-)

diff --git a/sysdep/bsd/krt-sock.c b/sysdep/bsd/krt-sock.c
index 0a52cfbd..e56dd616 100644
--- a/sysdep/bsd/krt-sock.c
+++ b/sysdep/bsd/krt-sock.c
@@ -186,6 +186,16 @@ struct ks_msg
     memcpy(p, body, (l > sizeof(*p) ? sizeof(*p) : l));\
     body += l;}
 
+static inline void
+sockaddr_fill_dl(struct sockaddr_dl *sa, struct iface *ifa)
+{
+  uint len = OFFSETOF(struct sockaddr_dl, sdl_data);
+  memset(sa, 0, len);
+  sa->sdl_len = len;
+  sa->sdl_family = AF_LINK;
+  sa->sdl_index = ifa->index;
+}
+
 static int
 krt_send_route(struct krt_proto *p, int cmd, rte *e)
 {
@@ -291,11 +301,8 @@ krt_send_route(struct krt_proto *p, int cmd, rte *e)
   {
     /* Fallback for all other valid cases */
 
-#ifdef RTF_CLONING
-    if (cmd == RTM_ADD && (i->flags & IF_MULTIACCESS) != IF_MULTIACCESS)	/* PTP */
-      msg.rtm.rtm_flags |= RTF_CLONING;
-#endif
-
+#if __OpenBSD__
+    /* Keeping temporarily old code for OpenBSD */
     struct ifa *addr = (net->n.addr->type == NET_IP4) ? i->addr4 : (i->addr6 ?: i->llv6);
 
     if (!addr)
@@ -304,7 +311,16 @@ krt_send_route(struct krt_proto *p, int cmd, rte *e)
       return -1;
     }
 
-    sockaddr_fill(&gate, af, addr->ip, i, 0);
+    /* Embed interface ID to link-local address */
+    ip_addr gw = addr->ip;
+    if (ipa_is_link_local(gw))
+      _I0(gw) = 0xfe800000 | (i->index & 0x0000ffff);
+
+    sockaddr_fill(&gate, af, gw, i, 0);
+#else
+    sockaddr_fill_dl(&gate, i);
+#endif
+
     msg.rtm.rtm_addrs |= RTA_GATEWAY;
     break;
   }

From 4db4ac7243bf54187029abda0b42cc9d29757d13 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Sun, 14 Jan 2018 21:52:58 +0100
Subject: [PATCH 114/198] NEWS and version update

---
 NEWS            | 5 +++++
 misc/bird.spec  | 2 +-
 sysdep/config.h | 2 +-
 3 files changed, 7 insertions(+), 2 deletions(-)

diff --git a/NEWS b/NEWS
index 55df8146..c25b11f6 100644
--- a/NEWS
+++ b/NEWS
@@ -1,3 +1,8 @@
+Version 2.0.1 (2018-01-14)
+  o Linux MPLS kernel support
+  o Better handling of channels inherited from templates
+  o Many bugfixes
+
 Version 2.0.0 (2017-12-11)
   o Integrated IPv4 + IPv6 design
   o Support for MPLS next hops
diff --git a/misc/bird.spec b/misc/bird.spec
index 1da0c813..7b645655 100644
--- a/misc/bird.spec
+++ b/misc/bird.spec
@@ -1,6 +1,6 @@
 Summary: BIRD Internet Routing Daemon
 Name: bird
-Version: 2.0.0
+Version: 2.0.1
 Release: 1
 Copyright: GPL
 Group: Networking/Daemons
diff --git a/sysdep/config.h b/sysdep/config.h
index 96e371f8..a552e6b6 100644
--- a/sysdep/config.h
+++ b/sysdep/config.h
@@ -13,7 +13,7 @@
 #ifdef GIT_LABEL
 #define BIRD_VERSION XSTR1(GIT_LABEL)
 #else
-#define BIRD_VERSION "2.0.0"
+#define BIRD_VERSION "2.0.1"
 #endif
 
 /* Include parameters determined by configure script */

From 3831b619661d08d935fd78656732cd2f339ff811 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 16 Jan 2018 04:14:49 +0100
Subject: [PATCH 115/198] BGP: Require explicit import and export policies for
 EBGP channels

To comply with RFC 8212 requirements.
---
 doc/bird.sgml      | 24 +++++++++++++++++++-----
 filter/filter.h    |  1 +
 proto/bgp/bgp.c    | 17 +++++++++++++++++
 proto/bgp/config.Y |  2 ++
 4 files changed, 39 insertions(+), 5 deletions(-)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index bde825c3..04f9258b 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -794,11 +794,13 @@ inherited from templates can be updated by new definitions.
 	<tag><label id="proto-import">import all | none | filter <m/name/ | filter { <m/filter commands/ } | where <m/boolean filter expression/</tag>
 	Specify a filter to be used for filtering routes coming from the
 	protocol to the routing table. <cf/all/ is for keeping all routes,
-	<cf/none/ is for dropping all routes. Default: <cf/all/.
+	<cf/none/ is for dropping all routes. Default: <cf/all/ (except for
+	EBGP).
 
 	<tag><label id="proto-export">export <m/filter/</tag>
 	This is similar to the <cf>import</cf> keyword, except that it works in
-	the direction from the routing table to the protocol. Default: <cf/none/.
+	the direction from the routing table to the protocol. Default: <cf/none/
+	(except for EBGP).
 
 	<tag><label id="proto-import-keep-filtered">import keep filtered <m/switch/</tag>
 	Usually, if an import filter rejects a route, the route is forgotten.
@@ -866,10 +868,12 @@ template bgp {
 	ipv4 {
 		table mytable4;
 		import filter { ... };
+		export none;
 	};
 	ipv6 {
 		table mytable6;
 		import filter { ... };
+		export none;
 	};
 }
 
@@ -2073,6 +2077,7 @@ avoid routing loops.
 <item> <rfc id="7947"> - Internet Exchange BGP Route Server
 <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
 </itemize>
 
 <sect1>Route selection rules
@@ -2459,6 +2464,12 @@ together with their appropriate channels follows.
 </tabular>
 </table>
 
+<p>Due to <rfc id="8212">, external BGP protocol requires explicit configuration
+of import and export policies (in contrast to other protocols, where default
+policies of <cf/import all/ and <cf/export none/ are used in absence of explicit
+configuration). Note that blanket policies like <cf/all/ or <cf/none/ can still
+be used in explicit configuration.
+
 <p>BGP channels have additional config options (together with the common ones):
 
 <descrip>
@@ -3631,13 +3642,13 @@ protocol kernel kern2 {
 }
 
 protocol bgp bgp1 {			# The outside connections
-	ipv4 { table as1; export all; };
+	ipv4 { table as1; import all; export all; };
 	local as 1;
 	neighbor 192.168.0.1 as 1001;
 }
 
 protocol bgp bgp2 {
-	ipv4 { table as2; export all; };
+	ipv4 { table as2; import all; export all; };
 	local as 2;
 	neighbor 10.0.0.1 as 1002;
 }
@@ -4419,7 +4430,10 @@ protocol bgp {
 	debug all;
 	local as 65000;
 	neighbor 192.168.2.1 as 65001;
-	ipv4 { import filter peer_in_v4; };
+	ipv4 {
+		import filter peer_in_v4;
+		export none;
+	};
 }
 </code>
 
diff --git a/filter/filter.h b/filter/filter.h
index c277b67c..89cd80e6 100644
--- a/filter/filter.h
+++ b/filter/filter.h
@@ -122,6 +122,7 @@ void val_format(struct f_val v, buffer *buf);
 
 #define FILTER_ACCEPT NULL
 #define FILTER_REJECT ((void *) 1)
+#define FILTER_UNDEF  ((void *) 2)	/* Used in BGP */
 
 /* Type numbers must be in 0..0xff range */
 #define T_MASK 0xff
diff --git a/proto/bgp/bgp.c b/proto/bgp/bgp.c
index 8da5b0aa..9db26050 100644
--- a/proto/bgp/bgp.c
+++ b/proto/bgp/bgp.c
@@ -99,6 +99,7 @@
  * <item> <rfc id="7947"> - Internet Exchange BGP Route Server
  * <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
  * </itemize>
 */
 
@@ -113,6 +114,7 @@
 #include "nest/cli.h"
 #include "nest/locks.h"
 #include "conf/conf.h"
+#include "filter/filter.h"
 #include "lib/socket.h"
 #include "lib/resource.h"
 #include "lib/string.h"
@@ -1621,6 +1623,7 @@ bgp_postconfig(struct proto_config *CF)
 {
   struct bgp_config *cf = (void *) CF;
   int internal = (cf->local_as == cf->remote_as);
+  int interior = internal || cf->confederation_member;
 
   /* Do not check templates at all */
   if (cf->c.class == SYM_TEMPLATE)
@@ -1677,6 +1680,20 @@ bgp_postconfig(struct proto_config *CF)
   struct bgp_channel_config *cc;
   WALK_LIST(cc, CF->channels)
   {
+    /* Handle undefined import filter */
+    if (cc->c.in_filter == FILTER_UNDEF)
+      if (interior)
+	cc->c.in_filter = FILTER_ACCEPT;
+      else
+	cf_error("EBGP requires explicit import policy");
+
+    /* Handle undefined export filter */
+    if (cc->c.out_filter == FILTER_UNDEF)
+      if (interior)
+	cc->c.out_filter = FILTER_REJECT;
+      else
+	cf_error("EBGP requires explicit export policy");
+
     /* Disable after error incompatible with restart limit action */
     if ((cc->c.in_limit.action == PLA_RESTART) && cf->disable_after_error)
       cc->c.in_limit.action = PLA_DISABLE;
diff --git a/proto/bgp/config.Y b/proto/bgp/config.Y
index 4e819eb7..04e6d666 100644
--- a/proto/bgp/config.Y
+++ b/proto/bgp/config.Y
@@ -163,6 +163,8 @@ bgp_channel_start: bgp_afi
   /* New channel */
   if (!BGP_CC->desc)
   {
+    BGP_CC->c.in_filter = FILTER_UNDEF;
+    BGP_CC->c.out_filter = FILTER_UNDEF;
     BGP_CC->c.ra_mode = RA_UNDEF;
     BGP_CC->afi = $1;
     BGP_CC->desc = desc;

From 68d0048b3d8d17cfa30846cfe06ea159c3c49a17 Mon Sep 17 00:00:00 2001
From: Ondrej Filip <feela@network.cz>
Date: Tue, 16 Jan 2018 10:45:03 +0100
Subject: [PATCH 116/198] Notice about RFC 8212 added

---
 NEWS | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/NEWS b/NEWS
index c25b11f6..a00e0826 100644
--- a/NEWS
+++ b/NEWS
@@ -1,6 +1,7 @@
-Version 2.0.1 (2018-01-14)
+Version 2.0.1 (2018-01-16)
   o Linux MPLS kernel support
   o Better handling of channels inherited from templates
+  o Default EBGP Route Propagation Behavior without Policies (RFC 8212)
   o Many bugfixes
 
 Version 2.0.0 (2017-12-11)

From c2febfa33294be0f457b49f5453d8066a7d88a5d Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 16 Jan 2018 14:18:57 +0100
Subject: [PATCH 117/198] Add note to NEWS

---
 NEWS | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/NEWS b/NEWS
index a00e0826..5bac8527 100644
--- a/NEWS
+++ b/NEWS
@@ -4,6 +4,12 @@ Version 2.0.1 (2018-01-16)
   o Default EBGP Route Propagation Behavior without Policies (RFC 8212)
   o Many bugfixes
 
+  Notes:
+
+  To satisfy requirements of RFC 8212, external BGP protocols now require
+  explicit configuration of import and export policies.
+
+
 Version 2.0.0 (2017-12-11)
   o Integrated IPv4 + IPv6 design
   o Support for MPLS next hops

From 6f46465af1b3d21ca67e3b4379640c008fc9d1a1 Mon Sep 17 00:00:00 2001
From: Ondrej Filip <feela@network.cz>
Date: Tue, 16 Jan 2018 14:36:46 +0100
Subject: [PATCH 118/198] Error in version guessing

---
 tools/gendist | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/gendist b/tools/gendist
index 0275db1a..eb2355d3 100755
--- a/tools/gendist
+++ b/tools/gendist
@@ -4,7 +4,7 @@
 #  (c) 2000--2004 Martin Mares <mj@ucw.cz>
 #
 
-VERSION=`sed <sysdep/config.h '/BIRD_VERSION/!d;s/^.*"\(.*\)"$/\1/'`
+VERSION=`grep 'BIRD_VERSION \"' sysdep/config.h | sed '/BIRD_VERSION/!d;s/^.*"\(.*\)"$/\1/'`
 REL=bird-$VERSION
 DREL=bird-doc-$VERSION
 T=/tmp/bird

From b94057911554e04df9b709f8354e2e220131096a Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 16 Jan 2018 16:20:01 +0100
Subject: [PATCH 119/198] Filter: Allow silent filter execution

A filter should log messages only if executed explicitly (e.g., during
route export or route import). When a filter is executed for technical
reasons (e.g., to establish whether a route was exported before), it
should run silently.
---
 filter/filter.c   | 9 ++++++---
 filter/filter.h   | 1 +
 nest/rt-table.c   | 9 ++++++---
 sysdep/unix/krt.c | 2 +-
 4 files changed, 14 insertions(+), 7 deletions(-)

diff --git a/filter/filter.c b/filter/filter.c
index 1e8f7d5a..b0c56046 100644
--- a/filter/filter.c
+++ b/filter/filter.c
@@ -590,7 +590,8 @@ f_rta_cow(void)
 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); \
+    if (!(f_flags & FF_SILENT)) \
+      log_rl(&rl_runtime_err, L_ERR "filters, line %d: %s", what->lineno, x); \
     res.type = T_RETURN; \
     res.val.i = F_ERROR; \
     return res; \
@@ -889,7 +890,8 @@ interpret(struct f_inst *what)
     break;
   case P('p',','):
     ONEARG;
-    if (what->a2.i == F_NOP || (what->a2.i != F_NONL && what->a1.p))
+    if ((what->a2.i == F_NOP || (what->a2.i != F_NONL && what->a1.p)) &&
+	!(f_flags & FF_SILENT))
       log_commit(*L_INFO, &f_buf);
 
     switch (what->a2.i) {
@@ -1723,7 +1725,8 @@ f_run(struct filter *filter, struct rte **rte, struct ea_list **tmp_attrs, struc
 
 
   if (res.type != T_RETURN) {
-    log_rl(&rl_runtime_err, L_ERR "Filter %s did not return accept nor reject. Make up your mind", filter->name);
+    if (!(f_flags & FF_SILENT))
+      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 );
diff --git a/filter/filter.h b/filter/filter.h
index 72b37461..efb4b978 100644
--- a/filter/filter.h
+++ b/filter/filter.h
@@ -227,5 +227,6 @@ struct f_trie
 #define NEW_F_VAL struct f_val * val; val = cfg_alloc(sizeof(struct f_val));
 
 #define FF_FORCE_TMPATTR 1		/* Force all attributes to be temporary */
+#define FF_SILENT 2			/* Silent filter execution */
 
 #endif
diff --git a/nest/rt-table.c b/nest/rt-table.c
index bcb48b53..28fe5baa 100644
--- a/nest/rt-table.c
+++ b/nest/rt-table.c
@@ -273,7 +273,8 @@ export_filter_(struct announce_hook *ah, rte *rt0, rte **rt_free, ea_list **tmpa
     }
 
   v = filter && ((filter == FILTER_REJECT) ||
-		 (f_run(filter, &rt, tmpa, pool, FF_FORCE_TMPATTR) > F_ACCEPT));
+		 (f_run(filter, &rt, tmpa, pool,
+			FF_FORCE_TMPATTR | (silent ? FF_SILENT : 0)) > F_ACCEPT));
   if (v)
     {
       if (silent)
@@ -1298,7 +1299,8 @@ rt_examine(rtable *t, ip_addr prefix, int pxlen, struct proto *p, struct filter
   ea_list *tmpa = rte_make_tmp_attrs(rt, rte_update_pool);
   int v = p->import_control ? p->import_control(p, &rt, &tmpa, rte_update_pool) : 0;
   if (v == RIC_PROCESS)
-    v = (f_run(filter, &rt, &tmpa, rte_update_pool, FF_FORCE_TMPATTR) <= F_ACCEPT);
+    v = (f_run(filter, &rt, &tmpa, rte_update_pool,
+	       FF_FORCE_TMPATTR | FF_SILENT) <= F_ACCEPT);
 
    /* Discard temporary rte */
   if (rt != n->routes)
@@ -2493,7 +2495,8 @@ rt_show_net(struct cli *c, net *n, struct rt_show_data *d)
 	       * 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);
+		(f_run(a->out_filter, &e, &tmpa, rte_update_pool,
+		       FF_FORCE_TMPATTR | FF_SILENT) <= F_ACCEPT);
 
 	      if (do_export != (d->export_mode == RSEM_EXPORT))
 		goto skip;
diff --git a/sysdep/unix/krt.c b/sysdep/unix/krt.c
index 07a55c0d..526c0cab 100644
--- a/sysdep/unix/krt.c
+++ b/sysdep/unix/krt.c
@@ -632,7 +632,7 @@ krt_export_net(struct krt_proto *p, net *net, rte **rt_free, ea_list **tmpa)
   if (filter == FILTER_ACCEPT)
     goto accept;
 
-  if (f_run(filter, &rt, tmpa, krt_filter_lp, FF_FORCE_TMPATTR) > F_ACCEPT)
+  if (f_run(filter, &rt, tmpa, krt_filter_lp, FF_FORCE_TMPATTR | FF_SILENT) > F_ACCEPT)
     goto reject;
 
 

From 63472779ad4ecdecbcfedf2d2bb40abc2f8c84b0 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 16 Jan 2018 19:17:04 +0100
Subject: [PATCH 120/198] BGP: Implement 'disable after cease' option

The option allows to specify that some cease subcodes should
disable the protocol when received.
---
 proto/bgp/bgp.h     |  1 +
 proto/bgp/config.Y  | 30 ++++++++++++++++++++++++++++++
 proto/bgp/packets.c | 10 ++++++++++
 3 files changed, 41 insertions(+)

diff --git a/proto/bgp/bgp.h b/proto/bgp/bgp.h
index 22a150ab..b3db8b7e 100644
--- a/proto/bgp/bgp.h
+++ b/proto/bgp/bgp.h
@@ -62,6 +62,7 @@ struct bgp_config {
   unsigned error_delay_time_min;	/* Time to wait after an error is detected */
   unsigned error_delay_time_max;
   unsigned disable_after_error;		/* Disable the protocol when error is detected */
+  u32 disable_after_cease;		/* Disable it when cease is received, bitfield */
 
   char *password;			/* Password used for MD5 authentication */
   struct rtable_config *igp_table;	/* Table used for recursive next hop lookups */
diff --git a/proto/bgp/config.Y b/proto/bgp/config.Y
index 55c602f1..075403a3 100644
--- a/proto/bgp/config.Y
+++ b/proto/bgp/config.Y
@@ -29,6 +29,12 @@ CF_KEYWORDS(BGP, LOCAL, NEIGHBOR, AS, HOLD, TIME, CONNECT, RETRY,
 	SECONDARY, ALLOW, BFD, ADD, PATHS, RX, TX, GRACEFUL, RESTART, AWARE,
 	CHECK, LINK, PORT, EXTENDED, MESSAGES, SETKEY, BGP_LARGE_COMMUNITY)
 
+CF_KEYWORDS(CEASE, PREFIX, LIMIT, HIT, ADMINISTRATIVE, SHUTDOWN, RESET, PEER,
+	CONFIGURATION, CHANGE, DECONFIGURED, CONNECTION, REJECTED, COLLISION,
+	OUT, OF, RESOURCES)
+
+%type<i> bgp_cease_mask bgp_cease_list bgp_cease_flag
+
 CF_GRAMMAR
 
 CF_ADDTO(proto, bgp_proto '}' { bgp_check_config(BGP_CFG); } )
@@ -64,6 +70,29 @@ bgp_nbr_opts:
  | bgp_nbr_opts AS expr { BGP_CFG->remote_as = $3; }
  ;
 
+bgp_cease_mask:
+   /* true -> all except connection collision */
+   bool { $$ = $1 ? ~(1 << 7) : 0; }
+ | '{' bgp_cease_list '}' { $$ = $2; }
+ ;
+
+bgp_cease_list:
+   bgp_cease_flag
+ | bgp_cease_list ',' bgp_cease_flag { $$ = $1 | $3; }
+ ;
+
+bgp_cease_flag:
+   CEASE			{ $$ = 1 << 0; }
+ | PREFIX LIMIT HIT		{ $$ = 1 << 1; }
+ | ADMINISTRATIVE SHUTDOWN	{ $$ = 1 << 2; }
+ | PEER DECONFIGURED		{ $$ = 1 << 3; }
+ | ADMINISTRATIVE RESET		{ $$ = 1 << 4; }
+ | CONNECTION REJECTED		{ $$ = 1 << 5; }
+ | CONFIGURATION CHANGE		{ $$ = 1 << 6; }
+ | CONNECTION COLLISION		{ $$ = 1 << 7; }
+ | OUT OF RESOURCES		{ $$ = 1 << 8; }
+ ;
+
 bgp_proto:
    bgp_proto_start proto_name '{'
  | bgp_proto proto_item ';'
@@ -107,6 +136,7 @@ bgp_proto:
  | bgp_proto ERROR FORGET TIME expr ';' { BGP_CFG->error_amnesia_time = $5; }
  | bgp_proto ERROR WAIT TIME expr ',' expr ';' { BGP_CFG->error_delay_time_min = $5; BGP_CFG->error_delay_time_max = $7; }
  | bgp_proto DISABLE AFTER ERROR bool ';' { BGP_CFG->disable_after_error = $5; }
+ | bgp_proto DISABLE AFTER CEASE bgp_cease_mask ';' { BGP_CFG->disable_after_cease = $5; }
  | bgp_proto ENABLE ROUTE REFRESH bool ';' { BGP_CFG->enable_refresh = $5; }
  | bgp_proto ENABLE AS4 bool ';' { BGP_CFG->enable_as4 = $4; }
  | bgp_proto ENABLE EXTENDED MESSAGES bool ';' { BGP_CFG->enable_extended_messages = $5; }
diff --git a/proto/bgp/packets.c b/proto/bgp/packets.c
index af3b15b5..f0049d3a 100644
--- a/proto/bgp/packets.c
+++ b/proto/bgp/packets.c
@@ -1607,6 +1607,16 @@ bgp_rx_notification(struct bgp_conn *conn, byte *pkt, uint len)
       bgp_update_startup_delay(p);
       bgp_stop(p, 0, NULL, 0);
     }
+  else
+    {
+      uint subcode_bit = 1 << ((subcode <= 8) ? subcode : 0);
+      if (p->cf->disable_after_cease & subcode_bit)
+      {
+	log(L_INFO "%s: Disabled after Cease notification", p->p.name);
+	p->startup_delay = 0;
+	p->p.disabled = 1;
+      }
+    }
 }
 
 static void

From c591810d46ae31c9e46768388878321e78e4ef94 Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Tue, 16 Jan 2018 14:46:06 +0100
Subject: [PATCH 121/198] Pipe: fixed template bug

When pipe inherited from template, every channel config was lost.
---
 proto/pipe/config.Y | 14 ++++++++++----
 1 file changed, 10 insertions(+), 4 deletions(-)

diff --git a/proto/pipe/config.Y b/proto/pipe/config.Y
index 4f96fdcb..b3c332be 100644
--- a/proto/pipe/config.Y
+++ b/proto/pipe/config.Y
@@ -25,13 +25,19 @@ CF_ADDTO(proto, pipe_proto '}' { this_channel = NULL; } )
 pipe_proto_start: proto_start PIPE
 {
   this_proto = proto_config_new(&proto_pipe, $1);
-  this_channel = channel_config_new(NULL, NULL, 0, this_proto);
-  this_channel->in_filter = FILTER_ACCEPT;
-  this_channel->out_filter = FILTER_ACCEPT;
+}
+proto_name
+{
+  this_channel = proto_cf_main_channel(this_proto);
+  if (!this_channel) {
+    this_channel = channel_config_new(NULL, NULL, 0, this_proto);
+    this_channel->in_filter = FILTER_ACCEPT;
+    this_channel->out_filter = FILTER_ACCEPT;
+  }
 };
 
 pipe_proto:
-   pipe_proto_start proto_name '{'
+   pipe_proto_start '{'
  | pipe_proto proto_item ';'
  | pipe_proto channel_item ';'
  | pipe_proto PEER TABLE rtable ';' { PIPE_CFG->peer = $4; }

From 8adaf730c0e2dae2debd7aa09633899e28428b11 Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Tue, 16 Jan 2018 16:10:13 +0100
Subject: [PATCH 122/198] Pipe: show export state

---
 proto/pipe/pipe.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/proto/pipe/pipe.c b/proto/pipe/pipe.c
index 310f3c01..49ff52e2 100644
--- a/proto/pipe/pipe.c
+++ b/proto/pipe/pipe.c
@@ -249,6 +249,8 @@ pipe_show_stats(struct pipe_proto *p)
 	  s2->imp_withdraws_ignored, s2->imp_withdraws_accepted);
 }
 
+static const char *pipe_feed_state[] = { [ES_DOWN] = "down", [ES_FEEDING] = "feed", [ES_READY] = "up" };
+
 static void
 pipe_show_proto_info(struct proto *P)
 {
@@ -257,6 +259,8 @@ pipe_show_proto_info(struct proto *P)
   cli_msg(-1006, "  Channel %s", "main");
   cli_msg(-1006, "    Table:          %s", p->pri->table->name);
   cli_msg(-1006, "    Peer table:     %s", p->sec->table->name);
+  cli_msg(-1006, "    Import state:   %s", pipe_feed_state[p->sec->export_state]);
+  cli_msg(-1006, "    Export state:   %s", pipe_feed_state[p->pri->export_state]);
   cli_msg(-1006, "    Import filter:  %s", filter_name(p->sec->out_filter));
   cli_msg(-1006, "    Export filter:  %s", filter_name(p->pri->out_filter));
 

From def6efa1ef98966424a73923832c6912acefc0ba Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 23 Jan 2018 14:26:18 +0100
Subject: [PATCH 123/198] Doc: Fix example

---
 doc/bird.conf.example2 | 11 +----------
 1 file changed, 1 insertion(+), 10 deletions(-)

diff --git a/doc/bird.conf.example2 b/doc/bird.conf.example2
index d7048b64..815651c7 100644
--- a/doc/bird.conf.example2
+++ b/doc/bird.conf.example2
@@ -28,20 +28,15 @@ flow6 table flowtab6;
 
 
 protocol device {
-	scan time 10;
 }
 
 protocol kernel kernel4 {
-	scan time 20;
-
 	ipv4 {
 		export all;
 	};
 }
 
 protocol kernel kernel6 {
-	scan time 20;
-
 	ipv6 {
 		export all;
 	};
@@ -169,8 +164,6 @@ protocol pipe {
 }
 
 protocol ospf v2 ospf4 {
-#	ecmp;
-
 	ipv4 {
 		import all;
 #		export where source = RTS_STATIC;
@@ -186,8 +179,6 @@ protocol ospf v2 ospf4 {
 
 
 protocol ospf v3 ospf6 {
-#	ecmp;
-
 	ipv6 {
 		import all;
 #		export where source = RTS_STATIC;
@@ -251,7 +242,7 @@ protocol bgp {
 	};
 
 	# IPv6 with MPLS labels (2/4)
-	ipv6 multicast {
+	ipv6 mpls {
 		# explicit IPv6 table
 		table mtab6;
 		import all;

From e5ff7929c4d85b84496cdfc46f006b8cd1b4c0e5 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 23 Jan 2018 14:48:07 +0100
Subject: [PATCH 124/198] KRT: Remove useless option

---
 sysdep/unix/krt.Y | 1 -
 sysdep/unix/krt.c | 2 +-
 sysdep/unix/krt.h | 1 -
 3 files changed, 1 insertion(+), 3 deletions(-)

diff --git a/sysdep/unix/krt.Y b/sysdep/unix/krt.Y
index 1cb28389..31f5f1d4 100644
--- a/sysdep/unix/krt.Y
+++ b/sysdep/unix/krt.Y
@@ -78,7 +78,6 @@ kern_item:
 	cf_error("Learning of kernel routes not supported on this platform");
 #endif
    }
- | DEVICE ROUTES bool { THIS_KRT->devroutes = $3; }
  | GRACEFUL RESTART bool { THIS_KRT->graceful_restart = $3; }
  | MERGE PATHS bool kern_mp_limit {
       krt_set_merge_paths(this_channel, $3, $4);
diff --git a/sysdep/unix/krt.c b/sysdep/unix/krt.c
index 29d2d01e..d4c0db72 100644
--- a/sysdep/unix/krt.c
+++ b/sysdep/unix/krt.c
@@ -1159,7 +1159,7 @@ krt_reconfigure(struct proto *p, struct proto_config *CF)
     return 0;
 
   /* persist, graceful restart need not be the same */
-  return o->scan_time == n->scan_time && o->learn == n->learn && o->devroutes == n->devroutes;
+  return o->scan_time == n->scan_time && o->learn == n->learn;
 }
 
 struct proto_config *
diff --git a/sysdep/unix/krt.h b/sysdep/unix/krt.h
index 3bfccfc2..8d3add0e 100644
--- a/sysdep/unix/krt.h
+++ b/sysdep/unix/krt.h
@@ -49,7 +49,6 @@ struct krt_config {
   btime scan_time;		/* How often we re-scan routes */
   int persist;			/* Keep routes when we exit */
   int learn;			/* Learn routes from other sources */
-  int devroutes;		/* XXX: remove */
   int graceful_restart;		/* Regard graceful restart recovery */
 };
 

From ace3072e09e445b2fd8554492b80bea5cc1f3411 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 23 Jan 2018 15:12:43 +0100
Subject: [PATCH 125/198] KRT: Fix option 'merge paths'

---
 sysdep/unix/krt.Y | 12 +-----------
 sysdep/unix/krt.c |  9 ++++++++-
 sysdep/unix/krt.h |  1 +
 3 files changed, 10 insertions(+), 12 deletions(-)

diff --git a/sysdep/unix/krt.Y b/sysdep/unix/krt.Y
index 31f5f1d4..9aac8668 100644
--- a/sysdep/unix/krt.Y
+++ b/sysdep/unix/krt.Y
@@ -16,16 +16,6 @@ CF_DEFINES
 #define THIS_KIF ((struct kif_config *) this_proto)
 #define KIF_IFACE ((struct kif_iface_config *) this_ipatt)
 
-static void
-krt_set_merge_paths(struct channel_config *cc, uint merge, uint limit)
-{
-  if ((limit <= 0) || (limit > 255))
-    cf_error("Merge paths limit must be in range 1-255");
-
-  cc->ra_mode = merge ? RA_MERGED : RA_OPTIMAL;
-  cc->merge_limit = limit;
-}
-
 static void
 kif_set_preferred(ip_addr ip)
 {
@@ -80,7 +70,7 @@ kern_item:
    }
  | GRACEFUL RESTART bool { THIS_KRT->graceful_restart = $3; }
  | MERGE PATHS bool kern_mp_limit {
-      krt_set_merge_paths(this_channel, $3, $4);
+      THIS_KRT->merge_paths = $3 ? $4 : 0;
 #ifndef KRT_ALLOW_MERGE_PATHS
       if ($3)
 	cf_error("Path merging not supported on this platform");
diff --git a/sysdep/unix/krt.c b/sysdep/unix/krt.c
index d4c0db72..bfdbfa4e 100644
--- a/sysdep/unix/krt.c
+++ b/sysdep/unix/krt.c
@@ -1059,11 +1059,18 @@ krt_postconfig(struct proto_config *CF)
     cf_error("All kernel syncers must use the same table scan interval");
 #endif
 
-  struct rtable_config *tab = proto_cf_main_channel(CF)->table;
+  struct channel_config *cc = proto_cf_main_channel(CF);
+  struct rtable_config *tab = cc->table;
   if (tab->krt_attached)
     cf_error("Kernel syncer (%s) already attached to table %s", tab->krt_attached->name, tab->name);
   tab->krt_attached = CF;
 
+  if (cf->merge_paths)
+  {
+    cc->ra_mode = RA_MERGED;
+    cc->merge_limit = cf->merge_paths;
+  }
+
   krt_sys_postconfig(cf);
 }
 
diff --git a/sysdep/unix/krt.h b/sysdep/unix/krt.h
index 8d3add0e..b627882d 100644
--- a/sysdep/unix/krt.h
+++ b/sysdep/unix/krt.h
@@ -50,6 +50,7 @@ struct krt_config {
   int persist;			/* Keep routes when we exit */
   int learn;			/* Learn routes from other sources */
   int graceful_restart;		/* Regard graceful restart recovery */
+  int merge_paths;		/* Exported routes are merged for ECMP */
 };
 
 struct krt_proto {

From d6cf996151307d083c30e4ecde0f1d7449b19253 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 23 Jan 2018 17:05:45 +0100
Subject: [PATCH 126/198] IO: Fix socket priority

On Linux, setting the ToS will also set the priority and the range of
accepted values is quite limited (masked by 0x1e). Therefore, 0xc0 is
translated to a priority of 0, not something we want, overriding the
"7" priority which was set previously explicitely. To avoid that, just
move setting priority later in the code.

Thanks to Vincent Bernat for the patch.
---
 sysdep/unix/io.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/sysdep/unix/io.c b/sysdep/unix/io.c
index 6d6a0990..53a37a50 100644
--- a/sysdep/unix/io.c
+++ b/sysdep/unix/io.c
@@ -1238,10 +1238,6 @@ sk_setup(sock *s)
 #endif
   }
 
-  if (s->priority >= 0)
-    if (sk_set_priority(s, s->priority) < 0)
-      return -1;
-
   if (sk_is_ipv4(s))
   {
     if (s->flags & SKF_LADDR_RX)
@@ -1292,6 +1288,11 @@ sk_setup(sock *s)
 	return -1;
   }
 
+  /* Must be after sk_set_tos4() as setting ToS on Linux also mangles priority */
+  if (s->priority >= 0)
+    if (sk_set_priority(s, s->priority) < 0)
+      return -1;
+
   return 0;
 }
 

From 345e50d59ff299e466fab6d0e0d37c210d624d96 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 24 Jan 2018 13:55:12 +0100
Subject: [PATCH 127/198] Nest: remove duplicate function

---
 nest/rt-fib.c | 22 ++++------------------
 1 file changed, 4 insertions(+), 18 deletions(-)

diff --git a/nest/rt-fib.c b/nest/rt-fib.c
index 24a7facc..c09f2759 100644
--- a/nest/rt-fib.c
+++ b/nest/rt-fib.c
@@ -92,8 +92,7 @@ fib_ht_free(struct fib_node **h)
 }
 
 
-static u32
-fib_hash(struct fib *f, const net_addr *a);
+static inline u32 fib_hash(struct fib *f, const net_addr *a);
 
 /**
  * fib_init - initialize a new FIB
@@ -198,24 +197,11 @@ fib_rehash(struct fib *f, int step)
   })
 
 
-static u32
+static inline u32
 fib_hash(struct fib *f, const net_addr *a)
 {
-  ASSERT(f->addr_type == a->type);
-
-  switch (f->addr_type)
-  {
-  case NET_IP4: return FIB_HASH(f, a, ip4);
-  case NET_IP6: return FIB_HASH(f, a, ip6);
-  case NET_VPN4: return FIB_HASH(f, a, vpn4);
-  case NET_VPN6: return FIB_HASH(f, a, vpn6);
-  case NET_ROA4: return FIB_HASH(f, a, roa4);
-  case NET_ROA6: return FIB_HASH(f, a, roa6);
-  case NET_FLOW4: return FIB_HASH(f, a, flow4);
-  case NET_FLOW6: return FIB_HASH(f, a, flow6);
-  case NET_MPLS: return FIB_HASH(f, a, mpls);
-  default: bug("invalid type");
-  }
+  /* Same as FIB_HASH() */
+  return net_hash(a) >> f->hash_shift;
 }
 
 void *

From 85ad5855a02e8b185a61bbcb601f005d2e6747db Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Mon, 29 Jan 2018 12:49:37 +0100
Subject: [PATCH 128/198] Nest: Fix corner case in recursive next hop lookup

Thanks to Svenne Krap for the bugreport.
---
 nest/rt-table.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/nest/rt-table.c b/nest/rt-table.c
index 4640d6a9..6ee261e8 100644
--- a/nest/rt-table.c
+++ b/nest/rt-table.c
@@ -2392,12 +2392,13 @@ static int
 rt_update_hostentry(rtable *tab, struct hostentry *he)
 {
   rta *old_src = he->src;
+  int direct = 0;
   int pxlen = 0;
 
   /* Reset the hostentry */
   he->src = NULL;
-  he->nexthop_linkable = 0;
   he->dest = RTD_UNREACHABLE;
+  he->nexthop_linkable = 0;
   he->igp_metric = 0;
 
   net_addr he_addr;
@@ -2417,9 +2418,7 @@ rt_update_hostentry(rtable *tab, struct hostentry *he)
 	  goto done;
 	}
 
-      he->dest = a->dest;
-      he->nexthop_linkable = 1;
-      if (he->dest == RTD_UNICAST)
+      if (a->dest == RTD_UNICAST)
 	{
 	  for (struct nexthop *nh = &(a->nh); nh; nh = nh->next)
 	    if (ipa_zero(nh->gw))
@@ -2432,12 +2431,13 @@ rt_update_hostentry(rtable *tab, struct hostentry *he)
 		    goto done;
 		  }
 
-		he->nexthop_linkable = 0;
-		break;
+		direct++;
 	      }
 	}
 
       he->src = rta_clone(a);
+      he->dest = a->dest;
+      he->nexthop_linkable = !direct;
       he->igp_metric = rt_get_igp_metric(e);
     }
 

From 28b3b551222ab58456a067a9be4790824cdbb60e Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 6 Feb 2018 16:08:45 +0100
Subject: [PATCH 129/198] KRT: Fix IPv6 route learn

Internal table used for route learn was created with non-matching net
type for IPv6 kernel proto.

Thanks to Toke Hoiland-Jorgensen for the bugreport
---
 bird.conf         |  7 ++++---
 nest/route.h      |  2 +-
 nest/rt-table.c   | 19 ++++++++-----------
 sysdep/unix/krt.c |  8 +++++++-
 4 files changed, 20 insertions(+), 16 deletions(-)

diff --git a/bird.conf b/bird.conf
index 410f190c..e383c934 100644
--- a/bird.conf
+++ b/bird.conf
@@ -22,13 +22,14 @@ protocol direct {
 
 # Feed routes to kernel FIB
 protocol kernel {
-	ipv4 { export all; };
-#	learn;			# Learn all routes from the kernel
+	ipv4 { export all; import all; };
+	learn;			# Learn all routes from the kernel
 #	scan time 10;		# Scan kernel tables every 10 seconds
 }
 
 protocol kernel {
-	ipv6;
+	ipv6 { import all; };
+	learn;
 }
 
 # Static route feed
diff --git a/nest/route.h b/nest/route.h
index 43d7d696..1c86110b 100644
--- a/nest/route.h
+++ b/nest/route.h
@@ -282,7 +282,7 @@ void rt_preconfig(struct config *);
 void rt_commit(struct config *new, struct config *old);
 void rt_lock_table(rtable *);
 void rt_unlock_table(rtable *);
-void rt_setup(pool *, rtable *, char *, struct rtable_config *);
+void rt_setup(pool *, rtable *, struct rtable_config *);
 static inline net *net_find(rtable *tab, const net_addr *addr) { return (net *) fib_find(&tab->fib, addr); }
 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; }
diff --git a/nest/rt-table.c b/nest/rt-table.c
index 6ee261e8..784f6cfb 100644
--- a/nest/rt-table.c
+++ b/nest/rt-table.c
@@ -1599,22 +1599,19 @@ rt_event(void *ptr)
 }
 
 void
-rt_setup(pool *p, rtable *t, char *name, struct rtable_config *cf)
+rt_setup(pool *p, rtable *t, struct rtable_config *cf)
 {
   bzero(t, sizeof(*t));
-  t->name = name;
+  t->name = cf->name;
   t->config = cf;
-  t->addr_type = cf ? cf->addr_type : NET_IP4;
+  t->addr_type = cf->addr_type;
   fib_init(&t->fib, p, t->addr_type, sizeof(net), OFFSETOF(net, n), 0, NULL);
   init_list(&t->channels);
 
-  if (cf)
-    {
-      t->rt_event = ev_new(p);
-      t->rt_event->hook = rt_event;
-      t->rt_event->data = t;
-      t->gc_time = current_time();
-    }
+  t->rt_event = ev_new(p);
+  t->rt_event->hook = rt_event;
+  t->rt_event->data = t;
+  t->gc_time = current_time();
 }
 
 /**
@@ -2090,7 +2087,7 @@ rt_commit(struct config *new, struct config *old)
       {
 	rtable *t = mb_alloc(rt_table_pool, sizeof(struct rtable));
 	DBG("\t%s: created\n", r->name);
-	rt_setup(rt_table_pool, t, r->name, r);
+	rt_setup(rt_table_pool, t, r);
 	add_tail(&routing_tables, &t->n);
 	r->table = t;
       }
diff --git a/sysdep/unix/krt.c b/sysdep/unix/krt.c
index e7bd79e3..a3cbb336d 100644
--- a/sysdep/unix/krt.c
+++ b/sysdep/unix/krt.c
@@ -506,7 +506,13 @@ static void
 krt_learn_init(struct krt_proto *p)
 {
   if (KRT_CF->learn)
-    rt_setup(p->p.pool, &p->krt_table, "Inherited", NULL);
+  {
+    struct rtable_config *cf = mb_allocz(p->p.pool, sizeof(struct rtable_config));
+    cf->name = "Inherited";
+    cf->addr_type = p->p.net_type;
+
+    rt_setup(p->p.pool, &p->krt_table, cf);
+  }
 }
 
 static void

From a82f692e5844d5efdc091a796dc0e8ae8ab5a322 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 6 Feb 2018 17:43:55 +0100
Subject: [PATCH 130/198] Nest: Trivial whitespace cleanup

---
 lib/net.c          | 64 +++++++++++++++++++++++-----------------------
 nest/a-path.c      |  2 +-
 nest/a-path_test.c |  6 ++---
 nest/cli.c         |  4 +--
 nest/config.Y      |  8 +++---
 nest/iface.c       |  2 +-
 nest/mrtdump.h     |  1 -
 nest/neighbor.c    |  2 +-
 nest/proto-hooks.c |  2 +-
 nest/protocol.h    |  2 +-
 nest/route.h       |  6 ++---
 nest/rt-attr.c     |  2 +-
 nest/rt-dev.c      | 10 ++++----
 nest/rt-fib.c      |  2 +-
 nest/rt-show.c     |  5 ++--
 nest/rt-table.c    | 10 ++++----
 16 files changed, 63 insertions(+), 65 deletions(-)

diff --git a/lib/net.c b/lib/net.c
index 9335b78f..3e5d62e2 100644
--- a/lib/net.c
+++ b/lib/net.c
@@ -6,50 +6,50 @@
 
 
 const char * const net_label[] = {
-  [NET_IP4] 	= "ipv4",
-  [NET_IP6] 	= "ipv6",
-  [NET_VPN4] 	= "vpn4",
-  [NET_VPN6] 	= "vpn6",
-  [NET_ROA4] 	= "roa4",
-  [NET_ROA6] 	= "roa6",
-  [NET_FLOW4] 	= "flow4",
-  [NET_FLOW6] 	= "flow6",
+  [NET_IP4]	= "ipv4",
+  [NET_IP6]	= "ipv6",
+  [NET_VPN4]	= "vpn4",
+  [NET_VPN6]	= "vpn6",
+  [NET_ROA4]	= "roa4",
+  [NET_ROA6]	= "roa6",
+  [NET_FLOW4]	= "flow4",
+  [NET_FLOW6]	= "flow6",
   [NET_MPLS]	= "mpls",
 };
 
 const u16 net_addr_length[] = {
-  [NET_IP4] 	= sizeof(net_addr_ip4),
-  [NET_IP6] 	= sizeof(net_addr_ip6),
-  [NET_VPN4] 	= sizeof(net_addr_vpn4),
-  [NET_VPN6] 	= sizeof(net_addr_vpn6),
-  [NET_ROA4] 	= sizeof(net_addr_roa4),
-  [NET_ROA6] 	= sizeof(net_addr_roa6),
-  [NET_FLOW4] 	= 0,
-  [NET_FLOW6] 	= 0,
+  [NET_IP4]	= sizeof(net_addr_ip4),
+  [NET_IP6]	= sizeof(net_addr_ip6),
+  [NET_VPN4]	= sizeof(net_addr_vpn4),
+  [NET_VPN6]	= sizeof(net_addr_vpn6),
+  [NET_ROA4]	= sizeof(net_addr_roa4),
+  [NET_ROA6]	= sizeof(net_addr_roa6),
+  [NET_FLOW4]	= 0,
+  [NET_FLOW6]	= 0,
   [NET_MPLS]	= sizeof(net_addr_mpls),
 };
 
 const u8 net_max_prefix_length[] = {
-  [NET_IP4] 	= IP4_MAX_PREFIX_LENGTH,
-  [NET_IP6] 	= IP6_MAX_PREFIX_LENGTH,
-  [NET_VPN4] 	= IP4_MAX_PREFIX_LENGTH,
-  [NET_VPN6] 	= IP6_MAX_PREFIX_LENGTH,
-  [NET_ROA4] 	= IP4_MAX_PREFIX_LENGTH,
-  [NET_ROA6] 	= IP6_MAX_PREFIX_LENGTH,
-  [NET_FLOW4] 	= IP4_MAX_PREFIX_LENGTH,
-  [NET_FLOW6] 	= IP6_MAX_PREFIX_LENGTH,
+  [NET_IP4]	= IP4_MAX_PREFIX_LENGTH,
+  [NET_IP6]	= IP6_MAX_PREFIX_LENGTH,
+  [NET_VPN4]	= IP4_MAX_PREFIX_LENGTH,
+  [NET_VPN6]	= IP6_MAX_PREFIX_LENGTH,
+  [NET_ROA4]	= IP4_MAX_PREFIX_LENGTH,
+  [NET_ROA6]	= IP6_MAX_PREFIX_LENGTH,
+  [NET_FLOW4]	= IP4_MAX_PREFIX_LENGTH,
+  [NET_FLOW6]	= IP6_MAX_PREFIX_LENGTH,
   [NET_MPLS]	= 0,
 };
 
 const u16 net_max_text_length[] = {
-  [NET_IP4] 	= 18,	/* "255.255.255.255/32" */
-  [NET_IP6] 	= 43,	/* "ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff/128" */
-  [NET_VPN4] 	= 40,	/* "4294967296:4294967296 255.255.255.255/32" */
-  [NET_VPN6] 	= 65,	/* "4294967296:4294967296 ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff/128" */
-  [NET_ROA4] 	= 34,	/* "255.255.255.255/32-32 AS4294967295" */
-  [NET_ROA6] 	= 60,	/* "ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff/128-128 AS4294967295" */
-  [NET_FLOW4] 	= 0,	/* "flow4 { ... }" */
-  [NET_FLOW6] 	= 0,	/* "flow6 { ... }" */
+  [NET_IP4]	= 18,	/* "255.255.255.255/32" */
+  [NET_IP6]	= 43,	/* "ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff/128" */
+  [NET_VPN4]	= 40,	/* "4294967296:4294967296 255.255.255.255/32" */
+  [NET_VPN6]	= 65,	/* "4294967296:4294967296 ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff/128" */
+  [NET_ROA4]	= 34,	/* "255.255.255.255/32-32 AS4294967295" */
+  [NET_ROA6]	= 60,	/* "ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff/128-128 AS4294967295" */
+  [NET_FLOW4]	= 0,	/* "flow4 { ... }" */
+  [NET_FLOW6]	= 0,	/* "flow6 { ... }" */
   [NET_MPLS]	= 7,	/* "1048575" */
 };
 
diff --git a/nest/a-path.c b/nest/a-path.c
index 32ffc72c..c0d16c30 100644
--- a/nest/a-path.c
+++ b/nest/a-path.c
@@ -810,7 +810,7 @@ as_path_match(const struct adata *path, struct f_path_mask *mask)
 	case PM_ASN_RANGE:
 	  val = mask->val;
 	  val2 = mask->val2;
-          goto step;
+	  goto step;
 	case PM_QUESTION:
 	step:
 	  nh = nl = -1;
diff --git a/nest/a-path_test.c b/nest/a-path_test.c
index 5e122396..a71b48ba 100644
--- a/nest/a-path_test.c
+++ b/nest/a-path_test.c
@@ -43,14 +43,14 @@ t_as_path_match(void)
       bt_debug("Prepending ASN: %10u \n", val);
 
       if (i == 0)
-        first_prepended = val;
+	first_prepended = val;
       if (i == AS_PATH_LENGTH-1)
-        last_prepended = val;
+	last_prepended = val;
 
       mask[i].kind = PM_ASN;
       mask[i].val  = val;
       if (i)
-        mask[i].next = &mask[i-1];
+	mask[i].next = &mask[i-1];
     }
 
     bt_assert_msg(as_path_match(as_path, &mask[AS_PATH_LENGTH-1]), "Mask should match with AS path");
diff --git a/nest/cli.c b/nest/cli.c
index ad81d384..c421cc7e 100644
--- a/nest/cli.c
+++ b/nest/cli.c
@@ -60,7 +60,7 @@
  * the new one. When the consumer processes everything in the buffer
  * queue, it calls cli_written(), tha frees all buffers (except the
  * first one) and schedules cli.event .
- * 
+ *
  */
 
 #include "nest/bird.h"
@@ -136,7 +136,7 @@ cli_printf(cli *c, int code, char *msg, ...)
     }
   else if (cd == CLI_ASYNC_CODE)
     {
-      size = 1; buf[0] = '+'; 
+      size = 1; buf[0] = '+';
       errcode = cd;
     }
   else
diff --git a/nest/config.Y b/nest/config.Y
index 044aba2b..af5114f5 100644
--- a/nest/config.Y
+++ b/nest/config.Y
@@ -151,7 +151,7 @@ CF_ENUM(T_ENUM_NETTYPE, NET_, IP4, IP6, VPN4, VPN6, ROA4, ROA6, FLOW4, FLOW6)
 CF_ADDTO(conf, table)
 
 table_sorted:
-          { $$ = 0; }
+	  { $$ = 0; }
  | SORTED { $$ = 1; }
  ;
 
@@ -454,9 +454,9 @@ password_item:
 password_item_begin:
    PASSWORD text {
      if (!this_p_list) {
-     	this_p_list = cfg_alloc(sizeof(list));
-     	init_list(this_p_list);
-        password_id = 1;
+	this_p_list = cfg_alloc(sizeof(list));
+	init_list(this_p_list);
+	password_id = 1;
      }
      this_p_item = cfg_alloc(sizeof (struct password_item));
      this_p_item->password = $2;
diff --git a/nest/iface.c b/nest/iface.c
index 54c16c58..a633f748 100644
--- a/nest/iface.c
+++ b/nest/iface.c
@@ -317,7 +317,7 @@ if_update(struct iface *new)
 	    new->sysdep = i->sysdep;
 	    memcpy(&new->addrs, &i->addrs, sizeof(i->addrs));
 	    memcpy(i, new, sizeof(*i));
-	    i->flags &= ~IF_UP; 	/* IF_TMP_DOWN will be added later */
+	    i->flags &= ~IF_UP;		/* IF_TMP_DOWN will be added later */
 	    goto newif;
 	  }
 
diff --git a/nest/mrtdump.h b/nest/mrtdump.h
index 73932553..28b3bdfd 100644
--- a/nest/mrtdump.h
+++ b/nest/mrtdump.h
@@ -28,4 +28,3 @@
 void mrt_dump_message(struct proto *p, u16 type, u16 subtype, byte *buf, u32 len);
 
 #endif
-
diff --git a/nest/neighbor.c b/nest/neighbor.c
index fb05d96c..4f93e29e 100644
--- a/nest/neighbor.c
+++ b/nest/neighbor.c
@@ -156,7 +156,7 @@ neigh_find2(struct proto *p, ip_addr *a, struct iface *ifa, unsigned flags)
     WALK_LIST(i, iface_list)
       if ((!p->vrf || p->vrf == i->master) &&
 	  ((scope = if_connected(a, i, &addr)) >= 0))
-        {
+	{
 	  ifa = i;
 	  break;
 	}
diff --git a/nest/proto-hooks.c b/nest/proto-hooks.c
index 92863f8e..71cddd64 100644
--- a/nest/proto-hooks.c
+++ b/nest/proto-hooks.c
@@ -281,7 +281,7 @@ int import_control(struct proto *p, rte **e, ea_list **attrs, struct linpool *po
 
 /**
  * rte_recalculate - prepare routes for comparison
- * @table: a routing table 
+ * @table: a routing table
  * @net: a network entry
  * @new: new route for the network
  * @old: old route for the network
diff --git a/nest/protocol.h b/nest/protocol.h
index 9afd3a0a..8a22d76b 100644
--- a/nest/protocol.h
+++ b/nest/protocol.h
@@ -337,7 +337,7 @@ void proto_notify_state(struct proto *p, unsigned state);
  *
  *		HUNGRY    ---->   FEEDING
  *		 ^		     |
- *		 | 		     V
+ *		 |		     V
  *		FLUSHING  <----   HAPPY
  *
  *	States:	HUNGRY	Protocol either administratively down (i.e.,
diff --git a/nest/route.h b/nest/route.h
index 1c86110b..79127519 100644
--- a/nest/route.h
+++ b/nest/route.h
@@ -74,7 +74,7 @@ static inline struct fib_node * fib_user_to_node(struct fib *f, void *e)
 void fib_init(struct fib *f, pool *p, uint addr_type, uint node_size, uint node_offset, uint hash_order, fib_init_fn init);
 void *fib_find(struct fib *, const net_addr *);	/* Find or return NULL if doesn't exist */
 void *fib_get_chain(struct fib *f, const net_addr *a); /* Find first node in linked list from hash table */
-void *fib_get(struct fib *, const net_addr *); 	/* Find or create new if nonexistent */
+void *fib_get(struct fib *, const net_addr *);	/* Find or create new if nonexistent */
 void *fib_route(struct fib *, const net_addr *); /* Longest-match routing lookup */
 void fib_delete(struct fib *, void *);	/* Remove fib entry */
 void fib_free(struct fib *);		/* Destroy the fib */
@@ -104,7 +104,7 @@ void fit_put_next(struct fib *f, struct fib_iterator *i, struct fib_node *n, uin
 	type *z;						\
 	for(;;) {						\
 	  if (!fn_)						\
-            {							\
+	    {							\
 	       if (++hpos_ >= count_)				\
 		 break;						\
 	       fn_ = (fib)->hash_table[hpos_];			\
@@ -663,7 +663,7 @@ extern struct protocol *attr_class_to_protocol[EAP_MAX];
  *	Default protocol preferences
  */
 
-#define DEF_PREF_DIRECT	    	240	/* Directly connected */
+#define DEF_PREF_DIRECT		240	/* Directly connected */
 #define DEF_PREF_STATIC		200	/* Static route */
 #define DEF_PREF_OSPF		150	/* OSPF intra-area, inter-area and type 1 external routes */
 #define DEF_PREF_BABEL		130	/* Babel */
diff --git a/nest/rt-attr.c b/nest/rt-attr.c
index 761ba9fe..881687de 100644
--- a/nest/rt-attr.c
+++ b/nest/rt-attr.c
@@ -1204,7 +1204,7 @@ rta_dump(rta *a)
   static char *rts[] = { "RTS_DUMMY", "RTS_STATIC", "RTS_INHERIT", "RTS_DEVICE",
 			 "RTS_STAT_DEV", "RTS_REDIR", "RTS_RIP",
 			 "RTS_OSPF", "RTS_OSPF_IA", "RTS_OSPF_EXT1",
-                         "RTS_OSPF_EXT2", "RTS_BGP", "RTS_PIPE", "RTS_BABEL" };
+			 "RTS_OSPF_EXT2", "RTS_BGP", "RTS_PIPE", "RTS_BABEL" };
   static char *rtd[] = { "", " DEV", " HOLE", " UNREACH", " PROHIBIT" };
 
   debug("p=%s uc=%d %s %s%s h=%04x",
diff --git a/nest/rt-dev.c b/nest/rt-dev.c
index b3d5bf97..718c4578 100644
--- a/nest/rt-dev.c
+++ b/nest/rt-dev.c
@@ -159,13 +159,13 @@ dev_copy_config(struct proto_config *dest, struct proto_config *src)
 }
 
 struct protocol proto_device = {
-  .name = 		"Direct",
-  .template = 		"direct%d",
+  .name =		"Direct",
+  .template =		"direct%d",
   .preference =		DEF_PREF_DIRECT,
   .channel_mask =	NB_IP,
   .proto_size =		sizeof(struct rt_dev_proto),
   .config_size =	sizeof(struct rt_dev_config),
-  .init = 		dev_init,
-  .reconfigure = 	dev_reconfigure,
-  .copy_config = 	dev_copy_config
+  .init =		dev_init,
+  .reconfigure =	dev_reconfigure,
+  .copy_config =	dev_copy_config
 };
diff --git a/nest/rt-fib.c b/nest/rt-fib.c
index c09f2759..169d6a4f 100644
--- a/nest/rt-fib.c
+++ b/nest/rt-fib.c
@@ -603,7 +603,7 @@ fib_histogram(struct fib *f)
       for (e = f->hash_table[i]; e != NULL; e = e->next)
 	j++;
       if (j > 0)
-        log(L_WARN "Histogram line %d: %d", i, j);
+	log(L_WARN "Histogram line %d: %d", i, j);
     }
 
   log(L_WARN "Histogram dump end");
diff --git a/nest/rt-show.c b/nest/rt-show.c
index 70dabc1f..1f1b73d2 100644
--- a/nest/rt-show.c
+++ b/nest/rt-show.c
@@ -74,7 +74,7 @@ rt_show_rte(struct cli *c, byte *ia, rte *e, struct rt_show_data *d, ea_list *tm
       char weight[16] = "";
 
       if (nh->labels)
-        {
+	{
 	  lsp += bsprintf(lsp, " mpls %d", nh->label[0]);
 	  for (int i=1;i<nh->labels; i++)
 	    lsp += bsprintf(lsp, "/%d", nh->label[i]);
@@ -129,7 +129,7 @@ rt_show_net(struct cli *c, net *n, struct rt_show_data *d)
 
       /* Special case for merged export */
       if ((d->export_mode == RSEM_EXPORT) && (ec->ra_mode == RA_MERGED))
-        {
+	{
 	  rte *rt_free;
 	  e = rt_export_merged(ec, n, &rt_free, &tmpa, c->show_pool, 1);
 	  pass = 1;
@@ -419,4 +419,3 @@ rt_show(struct rt_show_data *d)
       cli_msg(8001, "Network not found");
   }
 }
-
diff --git a/nest/rt-table.c b/nest/rt-table.c
index 784f6cfb..0f53c93f 100644
--- a/nest/rt-table.c
+++ b/nest/rt-table.c
@@ -614,9 +614,9 @@ rt_notify_accepted(struct channel *c, net *net, rte *new_changed, rte *old_chang
 	old_meet = 1;
     }
 
-  /* 
+  /*
    * Second, handle the feed case. That means we do not care for
-   * old_best. It is NULL for feed, and the new_best for refeed. 
+   * old_best. It is NULL for feed, and the new_best for refeed.
    * For refeed, there is a hack similar to one in rt_notify_basic()
    * to ensure withdraws in case of changed filters
    */
@@ -825,7 +825,7 @@ rt_notify_merged(struct channel *c, net *net, rte *new_changed, rte *old_changed
  * @new_best: the new best route for the same network
  * @old_best: the previous best route for the same network
  * @before_old: The previous route before @old for the same network.
- * 		If @before_old is NULL @old was the first.
+ *		If @before_old is NULL @old was the first.
  *
  * This function gets a routing table update and announces it
  * to all protocols that acccepts given type of route announcement
@@ -1387,7 +1387,7 @@ rte_update2(struct channel *c, const net_addr *n, rte *new, struct rte_src *src)
 
 /* Independent call to rte_announce(), used from next hop
    recalculation, outside of rte_update(). new must be non-NULL */
-static inline void 
+static inline void
 rte_announce_i(rtable *tab, unsigned type, net *net, rte *new, rte *old,
 	       rte *new_best, rte *old_best)
 {
@@ -2357,7 +2357,7 @@ if_local_addr(ip_addr a, struct iface *i)
   return 0;
 }
 
-static u32 
+static u32
 rt_get_igp_metric(rte *rt)
 {
   eattr *ea = ea_find(rt->attrs->eattrs, EA_GEN_IGP_METRIC);

From be17805c0bbd37e865dc9b17b56e8e8d210c2c6c Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 13 Feb 2018 16:27:57 +0100
Subject: [PATCH 131/198] Add support for source-specific IPv6 routes to BIRD
 core

This patch adds support for source-specific IPv6 routes to BIRD core.
This is based on Dean Luga's original patch, with the review comments
addressed. SADR support is added to network address parsing in confbase.Y
and to the kernel protocol on Linux.

Currently there is no way to mix source-specific and non-source-specific
routes (i.e., SADR tables cannot be connected to non-SADR tables).

Thanks to Toke Hoiland-Jorgensen for the original patch.
Minor changes by Ondrej Santiago Zajicek.
---
 conf/confbase.Y        | 24 ++++++++++-
 lib/net.c              | 23 +++++++++++
 lib/net.h              | 90 +++++++++++++++++++++++++++++++++++++-----
 nest/config.Y          | 13 ++++--
 nest/rt-fib.c          |  2 +
 nest/rt-table.c        | 42 ++++++++++++++++++++
 sysdep/cf/linux.h      |  1 +
 sysdep/linux/netlink.c | 43 ++++++++++++++++----
 sysdep/unix/krt.c      | 25 ++++++++----
 9 files changed, 233 insertions(+), 30 deletions(-)

diff --git a/conf/confbase.Y b/conf/confbase.Y
index 7e0537c5..c2d647eb 100644
--- a/conf/confbase.Y
+++ b/conf/confbase.Y
@@ -83,7 +83,7 @@ CF_DECLS
 %type <time> expr_us time
 %type <a> ipa
 %type <net> net_ip4_ net_ip6_ net_ip6 net_ip_ net_ip net_or_ipa
-%type <net_ptr> net_ net_any net_vpn4_ net_vpn6_ net_vpn_ net_roa4_ net_roa6_ net_roa_ net_mpls_
+%type <net_ptr> net_ net_any net_vpn4_ net_vpn6_ net_vpn_ net_roa4_ net_roa6_ net_roa_ net_ip6_sadr_ net_mpls_
 %type <mls> label_stack_start label_stack
 
 %type <t> text opttext
@@ -96,7 +96,7 @@ CF_DECLS
 %left '!'
 %nonassoc '.'
 
-CF_KEYWORDS(DEFINE, ON, OFF, YES, NO, S, MS, US, PORT, VPN, MPLS)
+CF_KEYWORDS(DEFINE, ON, OFF, YES, NO, S, MS, US, PORT, VPN, MPLS, FROM)
 
 CF_GRAMMAR
 
@@ -206,6 +206,25 @@ net_ip6_: IP6 '/' NUM
 	     n->prefix, n->pxlen, ip6_and(n->prefix, ip6_mkmask(n->pxlen)), n->pxlen);
 };
 
+net_ip6_sadr_: IP6 '/' NUM FROM IP6 '/' NUM
+{
+  if ($3 > IP6_MAX_PREFIX_LENGTH)
+    cf_error("Invalid prefix length %u", $3);
+
+  if ($7 > IP6_MAX_PREFIX_LENGTH)
+    cf_error("Invalid prefix length %u", $7);
+
+  $$ = cfg_alloc(sizeof(net_addr_ip6_sadr));
+  net_fill_ip6_sadr($$, $1, $3, $5, $7);
+
+  net_addr_ip6_sadr *n = (void *) $$;
+  if (!net_validate_ip6_sadr(n))
+    cf_error("Invalid SADR IPv6 prefix %I6/%d from %I6/%d, maybe you wanted %I6/%d from %I6/%d",
+	     n->dst_prefix, n->dst_pxlen, n->src_prefix, n->src_pxlen,
+	     ip6_and(n->dst_prefix, ip6_mkmask(n->dst_pxlen)), n->dst_pxlen,
+	     ip6_and(n->src_prefix, ip6_mkmask(n->src_pxlen)), n->src_pxlen);
+};
+
 net_vpn4_: VPN_RD net_ip4_
 {
   $$ = cfg_alloc(sizeof(net_addr_vpn4));
@@ -249,6 +268,7 @@ net_:
  | net_vpn_
  | net_roa_
  | net_flow_
+ | net_ip6_sadr_
  | net_mpls_
  ;
 
diff --git a/lib/net.c b/lib/net.c
index 3e5d62e2..976ddbcc 100644
--- a/lib/net.c
+++ b/lib/net.c
@@ -14,6 +14,7 @@ const char * const net_label[] = {
   [NET_ROA6]	= "roa6",
   [NET_FLOW4]	= "flow4",
   [NET_FLOW6]	= "flow6",
+  [NET_IP6_SADR]= "ipv6-sadr",
   [NET_MPLS]	= "mpls",
 };
 
@@ -26,6 +27,7 @@ const u16 net_addr_length[] = {
   [NET_ROA6]	= sizeof(net_addr_roa6),
   [NET_FLOW4]	= 0,
   [NET_FLOW6]	= 0,
+  [NET_IP6_SADR]= sizeof(net_addr_ip6_sadr),
   [NET_MPLS]	= sizeof(net_addr_mpls),
 };
 
@@ -38,6 +40,7 @@ const u8 net_max_prefix_length[] = {
   [NET_ROA6]	= IP6_MAX_PREFIX_LENGTH,
   [NET_FLOW4]	= IP4_MAX_PREFIX_LENGTH,
   [NET_FLOW6]	= IP6_MAX_PREFIX_LENGTH,
+  [NET_IP6_SADR]= IP6_MAX_PREFIX_LENGTH,
   [NET_MPLS]	= 0,
 };
 
@@ -50,6 +53,7 @@ const u16 net_max_text_length[] = {
   [NET_ROA6]	= 60,	/* "ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff/128-128 AS4294967295" */
   [NET_FLOW4]	= 0,	/* "flow4 { ... }" */
   [NET_FLOW6]	= 0,	/* "flow6 { ... }" */
+  [NET_IP6_SADR]= 92,	/* "ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff/128 from ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff/128" */
   [NET_MPLS]	= 7,	/* "1048575" */
 };
 
@@ -102,6 +106,8 @@ net_format(const net_addr *N, char *buf, int buflen)
     return flow4_net_format(buf, buflen, &n->flow4);
   case NET_FLOW6:
     return flow6_net_format(buf, buflen, &n->flow6);
+  case NET_IP6_SADR:
+    return bsnprintf(buf, buflen, "%I6/%d from %I6/%d", n->ip6_sadr.dst_prefix, n->ip6_sadr.dst_pxlen, n->ip6_sadr.src_prefix, n->ip6_sadr.src_pxlen);
   case NET_MPLS:
     return bsnprintf(buf, buflen, "%u", n->mpls.label);
   }
@@ -124,6 +130,7 @@ net_pxmask(const net_addr *a)
   case NET_VPN6:
   case NET_ROA6:
   case NET_FLOW6:
+  case NET_IP6_SADR:
     return ipa_from_ip6(ip6_mkmask(net6_pxlen(a)));
 
   case NET_MPLS:
@@ -156,6 +163,8 @@ net_compare(const net_addr *a, const net_addr *b)
     return net_compare_flow4((const net_addr_flow4 *) a, (const net_addr_flow4 *) b);
   case NET_FLOW6:
     return net_compare_flow6((const net_addr_flow6 *) a, (const net_addr_flow6 *) b);
+  case NET_IP6_SADR:
+    return net_compare_ip6_sadr((const net_addr_ip6_sadr *) a, (const net_addr_ip6_sadr *) b);
   case NET_MPLS:
     return net_compare_mpls((const net_addr_mpls *) a, (const net_addr_mpls *) b);
   }
@@ -177,6 +186,7 @@ net_hash(const net_addr *n)
   case NET_ROA6: return NET_HASH(n, roa6);
   case NET_FLOW4: return NET_HASH(n, flow4);
   case NET_FLOW6: return NET_HASH(n, flow6);
+  case NET_IP6_SADR: return NET_HASH(n, ip6_sadr);
   case NET_MPLS: return NET_HASH(n, mpls);
   default: bug("invalid type");
   }
@@ -198,6 +208,7 @@ net_validate(const net_addr *n)
   case NET_ROA6: return NET_VALIDATE(n, roa6);
   case NET_FLOW4: return NET_VALIDATE(n, flow4);
   case NET_FLOW6: return NET_VALIDATE(n, flow6);
+  case NET_IP6_SADR: return NET_VALIDATE(n, ip6_sadr);
   case NET_MPLS: return NET_VALIDATE(n, mpls);
   default: return 0;
   }
@@ -222,6 +233,9 @@ net_normalize(net_addr *N)
   case NET_FLOW6:
     return net_normalize_ip6(&n->ip6);
 
+  case NET_IP6_SADR:
+    return net_normalize_ip6_sadr(&n->ip6_sadr);
+
   case NET_MPLS:
     return;
   }
@@ -246,6 +260,9 @@ net_classify(const net_addr *N)
   case NET_FLOW6:
     return ip6_zero(n->ip6.prefix) ? (IADDR_HOST | SCOPE_UNIVERSE) : ip6_classify(&n->ip6.prefix);
 
+  case NET_IP6_SADR:
+    return ip6_zero(n->ip6_sadr.dst_prefix) ? (IADDR_HOST | SCOPE_UNIVERSE) : ip6_classify(&n->ip6_sadr.dst_prefix);
+
   case NET_MPLS:
     return IADDR_HOST | SCOPE_UNIVERSE;
   }
@@ -274,6 +291,11 @@ ipa_in_netX(const ip_addr a, const net_addr *n)
     return ip6_zero(ip6_and(ip6_xor(ipa_to_ip6(a), net6_prefix(n)),
 			    ip6_mkmask(net6_pxlen(n))));
 
+  case NET_IP6_SADR:
+    if (ipa_is_ip4(a)) return 0;
+    return ip6_zero(ip6_and(ip6_xor(ipa_to_ip6(a), net6_prefix(n)),
+			    ip6_mkmask(net6_pxlen(n))));
+
   case NET_MPLS:
   default:
     return 0;
@@ -304,5 +326,6 @@ net_init(void)
   CHECK_NET(net_addr_roa6,	28);
   CHECK_NET(net_addr_flow4,	 8);
   CHECK_NET(net_addr_flow6,	20);
+  CHECK_NET(net_addr_ip6_sadr,	40);
   CHECK_NET(net_addr_mpls,	 8);
 }
diff --git a/lib/net.h b/lib/net.h
index 69f00641..ad4000fd 100644
--- a/lib/net.h
+++ b/lib/net.h
@@ -21,8 +21,9 @@
 #define NET_ROA6	6
 #define NET_FLOW4	7
 #define NET_FLOW6	8
-#define NET_MPLS	9
-#define NET_MAX		10
+#define NET_IP6_SADR	9
+#define NET_MPLS	10
+#define NET_MAX		11
 
 #define NB_IP4		(1 << NET_IP4)
 #define NB_IP6		(1 << NET_IP6)
@@ -32,12 +33,13 @@
 #define NB_ROA6		(1 << NET_ROA6)
 #define NB_FLOW4	(1 << NET_FLOW4)
 #define NB_FLOW6	(1 << NET_FLOW6)
+#define NB_IP6_SADR	(1 << NET_IP6_SADR)
 #define NB_MPLS		(1 << NET_MPLS)
 
 #define NB_IP		(NB_IP4 | NB_IP6)
 #define NB_VPN		(NB_VPN4 | NB_VPN6)
 #define NB_FLOW		(NB_FLOW4 | NB_FLOW6)
-#define NB_DEST		(NB_IP | NB_VPN | NB_MPLS)
+#define NB_DEST		(NB_IP | NB_IP6_SADR | NB_VPN | NB_MPLS)
 #define NB_ANY		0xffffffff
 
 
@@ -121,6 +123,15 @@ typedef struct net_addr_mpls {
   u32 label;
 } net_addr_mpls;
 
+typedef struct net_addr_ip6_sadr {
+  u8 type;
+  u8 dst_pxlen;
+  u16 length;
+  ip6_addr dst_prefix;
+  s32 src_pxlen; /* s32 to avoid padding */
+  ip6_addr src_prefix;
+} net_addr_ip6_sadr;
+
 typedef union net_addr_union {
   net_addr n;
   net_addr_ip4 ip4;
@@ -131,6 +142,7 @@ typedef union net_addr_union {
   net_addr_roa6 roa6;
   net_addr_flow4 flow4;
   net_addr_flow6 flow6;
+  net_addr_ip6_sadr ip6_sadr;
   net_addr_mpls mpls;
 } net_addr_union;
 
@@ -167,6 +179,9 @@ extern const u16 net_max_text_length[];
 #define NET_ADDR_FLOW6(prefix,pxlen,dlen) \
   ((net_addr_flow6) { NET_FLOW6, pxlen, sizeof(net_addr_ip6) + dlen, prefix })
 
+#define NET_ADDR_IP6_SADR(dst_prefix,dst_pxlen,src_prefix,src_pxlen) \
+  ((net_addr_ip6_sadr) { NET_IP6_SADR, dst_pxlen, sizeof(net_addr_ip6_sadr), dst_prefix, src_pxlen, src_prefix })
+
 #define NET_ADDR_MPLS(label) \
   ((net_addr_mpls) { NET_MPLS, 20, sizeof(net_addr_mpls), label })
 
@@ -189,6 +204,9 @@ static inline void net_fill_roa4(net_addr *a, ip4_addr prefix, uint pxlen, uint
 static inline void net_fill_roa6(net_addr *a, ip6_addr prefix, uint pxlen, uint max_pxlen, u32 asn)
 { *(net_addr_roa6 *)a = NET_ADDR_ROA6(prefix, pxlen, max_pxlen, asn); }
 
+static inline void net_fill_ip6_sadr(net_addr *a, ip6_addr dst_prefix, uint dst_pxlen, ip6_addr src_prefix, uint src_pxlen)
+{ *(net_addr_ip6_sadr *)a = NET_ADDR_IP6_SADR(dst_prefix, dst_pxlen, src_prefix, src_pxlen); }
+
 static inline void net_fill_mpls(net_addr *a, u32 label)
 { *(net_addr_mpls *)a = NET_ADDR_MPLS(label); }
 
@@ -222,6 +240,16 @@ static inline void net_fill_flow6(net_addr *a, ip6_addr prefix, uint pxlen, byte
   memcpy(f->data, data, dlen);
 }
 
+/* Make NET_IP6_SADR from NET_IP6, assuming there is enough space */
+static inline void net_make_ip6_sadr(net_addr *a)
+{
+  net_addr_ip6_sadr *n = (void *) a;
+  n->type = NET_IP6_SADR;
+  n->length = sizeof(net_addr_ip6_sadr);
+  n->src_pxlen = 0;
+  n->src_prefix = IP6_NONE;
+}
+
 static inline int net_val_match(u8 type, u32 mask)
 { return !!((1 << type) & mask); }
 
@@ -261,6 +289,7 @@ static inline ip_addr net_prefix(const net_addr *a)
   case NET_VPN6:
   case NET_ROA6:
   case NET_FLOW6:
+  case NET_IP6_SADR:
     return ipa_from_ip6(net6_prefix(a));
 
   case NET_MPLS:
@@ -328,6 +357,9 @@ static inline int net_equal_flow4(const net_addr_flow4 *a, const net_addr_flow4
 static inline int net_equal_flow6(const net_addr_flow6 *a, const net_addr_flow6 *b)
 { return net_equal((const net_addr *) a, (const net_addr *) b); }
 
+static inline int net_equal_ip6_sadr(const net_addr_ip6_sadr *a, const net_addr_ip6_sadr *b)
+{ return !memcmp(a, b, sizeof(net_addr_ip6_sadr)); }
+
 static inline int net_equal_mpls(const net_addr_mpls *a, const net_addr_mpls *b)
 { return !memcmp(a, b, sizeof(net_addr_mpls)); }
 
@@ -338,6 +370,12 @@ static inline int net_equal_prefix_roa4(const net_addr_roa4 *a, const net_addr_r
 static inline int net_equal_prefix_roa6(const net_addr_roa6 *a, const net_addr_roa6 *b)
 { return ip6_equal(a->prefix, b->prefix) && (a->pxlen == b->pxlen); }
 
+static inline int net_equal_dst_ip6_sadr(const net_addr_ip6_sadr *a, const net_addr_ip6_sadr *b)
+{ return ip6_equal(a->dst_prefix, b->dst_prefix) && (a->dst_pxlen == b->dst_pxlen); }
+
+static inline int net_equal_src_ip6_sadr(const net_addr_ip6_sadr *a, const net_addr_ip6_sadr *b)
+{ return ip6_equal(a->src_prefix, b->src_prefix) && (a->src_pxlen == b->src_pxlen); }
+
 
 static inline int net_zero_ip4(const net_addr_ip4 *a)
 { return !a->pxlen && ip4_zero(a->prefix); }
@@ -391,6 +429,13 @@ static inline int net_compare_flow4(const net_addr_flow4 *a, const net_addr_flow
 static inline int net_compare_flow6(const net_addr_flow6 *a, const net_addr_flow6 *b)
 { return ip6_compare(a->prefix, b->prefix) ?: uint_cmp(a->pxlen, b->pxlen) ?: uint_cmp(a->length, b->length) ?: memcmp(a->data, b->data, a->length - sizeof(net_addr_flow6)); }
 
+static inline int net_compare_ip6_sadr(const net_addr_ip6_sadr *a, const net_addr_ip6_sadr *b)
+{
+  return
+    ip6_compare(a->dst_prefix, b->dst_prefix) ?: uint_cmp(a->dst_pxlen, b->dst_pxlen) ?:
+    ip6_compare(a->src_prefix, b->src_prefix) ?: uint_cmp(a->src_pxlen, b->src_pxlen);
+}
+
 static inline int net_compare_mpls(const net_addr_mpls *a, const net_addr_mpls *b)
 { return uint_cmp(a->label, b->label); }
 
@@ -424,6 +469,9 @@ static inline void net_copy_flow4(net_addr_flow4 *dst, const net_addr_flow4 *src
 static inline void net_copy_flow6(net_addr_flow6 *dst, const net_addr_flow6 *src)
 { memcpy(dst, src, src->length); }
 
+static inline void net_copy_ip6_sadr(net_addr_ip6_sadr *dst, const net_addr_ip6_sadr *src)
+{ memcpy(dst, src, sizeof(net_addr_ip6_sadr)); }
+
 static inline void net_copy_mpls(net_addr_mpls *dst, const net_addr_mpls *src)
 { memcpy(dst, src, sizeof(net_addr_mpls)); }
 
@@ -456,6 +504,9 @@ static inline u32 net_hash_flow4(const net_addr_flow4 *n)
 static inline u32 net_hash_flow6(const net_addr_flow6 *n)
 { return ip6_hash(n->prefix) ^ ((u32) n->pxlen << 26); }
 
+static inline u32 net_hash_ip6_sadr(const net_addr_ip6_sadr *n)
+{ return net_hash_ip6((net_addr_ip6 *) n); }
+
 static inline u32 net_hash_mpls(const net_addr_mpls *n)
 { return n->label; }
 
@@ -508,6 +559,9 @@ static inline int net_validate_flow6(const net_addr_flow6 *n)
 static inline int net_validate_mpls(const net_addr_mpls *n)
 { return n->label < (1 << 20); }
 
+static inline int net_validate_ip6_sadr(const net_addr_ip6_sadr *n)
+{ return net_validate_px6(n->dst_prefix, n->dst_pxlen) && net_validate_px6(n->src_prefix, n->src_pxlen); }
+
 int net_validate(const net_addr *N);
 
 
@@ -523,6 +577,12 @@ static inline void net_normalize_vpn4(net_addr_vpn4 *n)
 static inline void net_normalize_vpn6(net_addr_vpn6 *n)
 { net_normalize_ip6((net_addr_ip6 *) n); }
 
+static inline void net_normalize_ip6_sadr(net_addr_ip6_sadr *n)
+{
+  n->dst_prefix = ip6_and(n->dst_prefix, ip6_mkmask(n->dst_pxlen));
+  n->src_prefix = ip6_and(n->src_prefix, ip6_mkmask(n->src_pxlen));
+}
+
 void net_normalize(net_addr *N);
 
 
@@ -530,17 +590,29 @@ int net_classify(const net_addr *N);
 int net_format(const net_addr *N, char *buf, int buflen);
 int rd_format(const u64 rd, char *buf, int buflen);
 
-static inline int ipa_in_net_ip4(ip4_addr a, const net_addr_ip4 *n)
-{ return ip4_zero(ip4_and(ip4_xor(a, n->prefix), ip4_mkmask(n->pxlen))); }
+static inline int ipa_in_px4(ip4_addr a, ip4_addr prefix, uint pxlen)
+{ return ip4_zero(ip4_and(ip4_xor(a, prefix), ip4_mkmask(pxlen))); }
 
-static inline int net_in_net_ip4(const net_addr_ip4 *a, const net_addr_ip4 *b)
-{ return (a->pxlen >= b->pxlen) && ipa_in_net_ip4(a->prefix, b); }
+static inline int ipa_in_px6(ip6_addr a, ip6_addr prefix, uint pxlen)
+{ return ip6_zero(ip6_and(ip6_xor(a, prefix), ip6_mkmask(pxlen))); }
+
+static inline int ipa_in_net_ip4(ip4_addr a, const net_addr_ip4 *n)
+{ return ipa_in_px4(a, n->prefix, n->pxlen); }
 
 static inline int ipa_in_net_ip6(ip6_addr a, const net_addr_ip6 *n)
-{ return ip6_zero(ip6_and(ip6_xor(a, n->prefix), ip6_mkmask(n->pxlen))); }
+{ return ipa_in_px6(a, n->prefix, n->pxlen); }
+
+static inline int net_in_net_ip4(const net_addr_ip4 *a, const net_addr_ip4 *b)
+{ return (a->pxlen >= b->pxlen) && ipa_in_px4(a->prefix, b->prefix, b->pxlen); }
 
 static inline int net_in_net_ip6(const net_addr_ip6 *a, const net_addr_ip6 *b)
-{ return (a->pxlen >= b->pxlen) && ipa_in_net_ip6(a->prefix, b); }
+{ return (a->pxlen >= b->pxlen) && ipa_in_px6(a->prefix, b->prefix, b->pxlen); }
+
+static inline int net_in_net_dst_ip6_sadr(const net_addr_ip6_sadr *a, const net_addr_ip6_sadr *b)
+{ return (a->dst_pxlen >= b->dst_pxlen) && ipa_in_px6(a->dst_prefix, b->dst_prefix, b->dst_pxlen); }
+
+static inline int net_in_net_src_ip6_sadr(const net_addr_ip6_sadr *a, const net_addr_ip6_sadr *b)
+{ return (a->src_pxlen >= b->src_pxlen) && ipa_in_px6(a->src_prefix, b->src_prefix, b->src_pxlen); }
 
 int ipa_in_netX(const ip_addr A, const net_addr *N);
 int net_in_netX(const net_addr *A, const net_addr *N);
diff --git a/nest/config.Y b/nest/config.Y
index af5114f5..ab09a10c 100644
--- a/nest/config.Y
+++ b/nest/config.Y
@@ -66,7 +66,7 @@ CF_DECLS
 
 CF_KEYWORDS(ROUTER, ID, PROTOCOL, TEMPLATE, PREFERENCE, DISABLED, DEBUG, ALL, OFF, DIRECT)
 CF_KEYWORDS(INTERFACE, IMPORT, EXPORT, FILTER, NONE, VRF, TABLE, STATES, ROUTES, FILTERS)
-CF_KEYWORDS(IPV4, IPV6, VPN4, VPN6, ROA4, ROA6)
+CF_KEYWORDS(IPV4, IPV6, VPN4, VPN6, ROA4, ROA6, FLOW4, FLOW6, SADR, MPLS)
 CF_KEYWORDS(RECEIVE, LIMIT, ACTION, WARN, BLOCK, RESTART, DISABLE, KEEP, FILTERED)
 CF_KEYWORDS(PASSWORD, FROM, PASSIVE, TO, ID, EVENTS, PACKETS, PROTOCOLS, INTERFACES)
 CF_KEYWORDS(ALGORITHM, KEYED, HMAC, MD5, SHA1, SHA256, SHA384, SHA512)
@@ -77,7 +77,7 @@ CF_KEYWORDS(TIMEFORMAT, ISO, SHORT, LONG, ROUTE, PROTOCOL, BASE, LOG, S, MS, US)
 CF_KEYWORDS(GRACEFUL, RESTART, WAIT, MAX, FLUSH, AS)
 
 /* For r_args_channel */
-CF_KEYWORDS(IPV4, IPV4_MC, IPV4_MPLS, IPV6, IPV6_MC, IPV6_MPLS, VPN4, VPN4_MC, VPN4_MPLS, VPN6, VPN6_MC, VPN6_MPLS, ROA4, ROA6, FLOW4, FLOW6, MPLS, PRI, SEC)
+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)
 
 CF_ENUM(T_ENUM_RTS, RTS_, DUMMY, STATIC, INHERIT, DEVICE, STATIC_DEVICE, REDIRECT,
 	RIP, OSPF, OSPF_IA, OSPF_EXT1, OSPF_EXT2, BGP, PIPE, BABEL)
@@ -134,6 +134,7 @@ gr_opts: GRACEFUL RESTART WAIT expr ';' { new_config->gr_wait = $4; } ;
 net_type:
    IPV4 { $$ = NET_IP4; }
  | IPV6 { $$ = NET_IP6; }
+ | IPV6 SADR { $$ = NET_IP6_SADR; }
  | VPN4 { $$ = NET_VPN4; }
  | VPN6 { $$ = NET_VPN6; }
  | ROA4 { $$ = NET_ROA4; }
@@ -143,7 +144,7 @@ net_type:
  | MPLS { $$ = NET_MPLS; }
  ;
 
-CF_ENUM(T_ENUM_NETTYPE, NET_, IP4, IP6, VPN4, VPN6, ROA4, ROA6, FLOW4, FLOW6)
+CF_ENUM(T_ENUM_NETTYPE, NET_, IP4, IP6, VPN4, VPN6, ROA4, ROA6, FLOW4, FLOW6, IP6_SADR)
 
 
 /* Creation of routing tables */
@@ -625,6 +626,7 @@ r_args_for:
   }
  | net_vpn4_
  | net_vpn6_
+ | net_ip6_sadr_
  | VPN_RD IP4 {
     $$ = cfg_alloc(sizeof(net_addr_vpn4));
     net_fill_vpn4($$, $2, IP4_MAX_PREFIX_LENGTH, $1);
@@ -633,6 +635,10 @@ r_args_for:
     $$ = cfg_alloc(sizeof(net_addr_vpn6));
     net_fill_vpn6($$, $2, IP6_MAX_PREFIX_LENGTH, $1);
   }
+ | IP6 FROM IP6 {
+    $$ = cfg_alloc(sizeof(net_addr_ip6_sadr));
+    net_fill_ip6_sadr($$, $1, IP6_MAX_PREFIX_LENGTH, $3, IP6_MAX_PREFIX_LENGTH);
+  }
  | SYM {
      if ($1->class == (SYM_CONSTANT | T_IP))
      {
@@ -666,6 +672,7 @@ r_args_channel:
  | IPV6		{ $$ = "ipv6"; }
  | IPV6_MC	{ $$ = "ipv6-mc"; }
  | IPV6_MPLS	{ $$ = "ipv6-mpls"; }
+ | IPV6_SADR	{ $$ = "ipv6-sadr"; }
  | VPN4		{ $$ = "vpn4"; }
  | VPN4_MC	{ $$ = "vpn4-mc"; }
  | VPN4_MPLS	{ $$ = "vpn4-mpls"; }
diff --git a/nest/rt-fib.c b/nest/rt-fib.c
index 169d6a4f..18ccbfc3 100644
--- a/nest/rt-fib.c
+++ b/nest/rt-fib.c
@@ -236,6 +236,7 @@ fib_find(struct fib *f, const net_addr *a)
   case NET_ROA6: return FIB_FIND(f, a, roa6);
   case NET_FLOW4: return FIB_FIND(f, a, flow4);
   case NET_FLOW6: return FIB_FIND(f, a, flow6);
+  case NET_IP6_SADR: return FIB_FIND(f, a, ip6_sadr);
   case NET_MPLS: return FIB_FIND(f, a, mpls);
   default: bug("invalid type");
   }
@@ -256,6 +257,7 @@ fib_insert(struct fib *f, const net_addr *a, struct fib_node *e)
   case NET_ROA6: FIB_INSERT(f, a, e, roa6); return;
   case NET_FLOW4: FIB_INSERT(f, a, e, flow4); return;
   case NET_FLOW6: FIB_INSERT(f, a, e, flow6); return;
+  case NET_IP6_SADR: FIB_INSERT(f, a, e, ip6_sadr); return;
   case NET_MPLS: FIB_INSERT(f, a, e, mpls); return;
   default: bug("invalid type");
   }
diff --git a/nest/rt-table.c b/nest/rt-table.c
index 0f53c93f..686d0e84 100644
--- a/nest/rt-table.c
+++ b/nest/rt-table.c
@@ -85,6 +85,45 @@ net_route_ip6(rtable *t, net_addr_ip6 *n)
   return r;
 }
 
+static inline void *
+net_route_ip6_sadr(rtable *t, net_addr_ip6_sadr *n)
+{
+  struct fib_node *fn;
+
+  while (1)
+  {
+    net *best = NULL;
+    int best_pxlen = 0;
+
+    /* We need to do dst first matching. Since sadr addresses are hashed on dst
+       prefix only, find the hash table chain and go through it to find the
+       match with the smallest matching src prefix. */
+    for (fn = fib_get_chain(&t->fib, (net_addr *) n); fn; fn = fn->next)
+    {
+      net_addr_ip6_sadr *a = (void *) fn->addr;
+
+      if (net_equal_dst_ip6_sadr(n, a) &&
+	  net_in_net_src_ip6_sadr(n, a) &&
+	  (a->src_pxlen >= best_pxlen))
+      {
+	best = fib_node_to_user(&t->fib, fn);
+	best_pxlen = a->src_pxlen;
+      }
+    }
+
+    if (best)
+      return best;
+
+    if (!n->dst_pxlen)
+      break;
+
+    n->dst_pxlen--;
+    ip6_clrbit(&n->dst_prefix, n->dst_pxlen);
+  }
+
+  return NULL;
+}
+
 void *
 net_route(rtable *tab, const net_addr *n)
 {
@@ -105,6 +144,9 @@ net_route(rtable *tab, const net_addr *n)
   case NET_ROA6:
     return net_route_ip6(tab, (net_addr_ip6 *) n0);
 
+  case NET_IP6_SADR:
+    return net_route_ip6_sadr(tab, (net_addr_ip6_sadr *) n0);
+
   default:
     return NULL;
   }
diff --git a/sysdep/cf/linux.h b/sysdep/cf/linux.h
index 3a3a15da..047d3764 100644
--- a/sysdep/cf/linux.h
+++ b/sysdep/cf/linux.h
@@ -10,6 +10,7 @@
 #define CONFIG_SELF_CONSCIOUS
 #define CONFIG_MULTIPLE_TABLES
 #define CONFIG_ALL_TABLES_AT_ONCE
+#define CONFIG_IP6_SADR_KERNEL
 
 #define CONFIG_MC_PROPER_SRC
 #define CONFIG_UNIX_DONTROUTE
diff --git a/sysdep/linux/netlink.c b/sysdep/linux/netlink.c
index 4cb51519..84591eb2 100644
--- a/sysdep/linux/netlink.c
+++ b/sysdep/linux/netlink.c
@@ -374,6 +374,7 @@ static struct nl_want_attrs rtm_attr_want4[BIRD_RTA_MAX] = {
 
 static struct nl_want_attrs rtm_attr_want6[BIRD_RTA_MAX] = {
   [RTA_DST]	  = { 1, 1, sizeof(ip6_addr) },
+  [RTA_SRC]	  = { 1, 1, sizeof(ip6_addr) },
   [RTA_IIF]	  = { 1, 1, sizeof(u32) },
   [RTA_OIF]	  = { 1, 1, sizeof(u32) },
   [RTA_GATEWAY]	  = { 1, 1, sizeof(ip6_addr) },
@@ -1221,8 +1222,18 @@ nl_send_route(struct krt_proto *p, rte *e, struct ea_list *eattrs, int op, int d
   }
   else
 #endif
+  {
     nl_add_attr_ipa(&r->h, rsize, RTA_DST, net_prefix(net->n.addr));
 
+    /* Add source address for IPv6 SADR routes */
+    if (net->n.addr->type == NET_IP6_SADR)
+    {
+      net_addr_ip6_sadr *a = (void *) &net->n.addr;
+      nl_add_attr_ip6(&r->h, rsize, RTA_SRC, a->src_prefix);
+      r->r.rtm_src_len = a->src_pxlen;
+    }
+  }
+
   /*
    * Strange behavior for RTM_DELROUTE:
    * 1) rtm_family is ignored in IPv6, works for IPv4
@@ -1447,12 +1458,12 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h)
   struct rtattr *a[BIRD_RTA_MAX];
   int new = h->nlmsg_type == RTM_NEWROUTE;
 
-  net_addr dst;
+  net_addr dst, src = {};
   u32 oif = ~0;
   u32 table_id;
   u32 priority = 0;
   u32 def_scope = RT_SCOPE_UNIVERSE;
-  int src;
+  int krt_src;
 
   if (!(i = nl_checkin(h, sizeof(*i))))
     return;
@@ -1477,6 +1488,11 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h)
 	net_fill_ip6(&dst, rta_get_ip6(a[RTA_DST]), i->rtm_dst_len);
       else
 	net_fill_ip6(&dst, IP6_NONE, 0);
+
+      if (a[RTA_SRC])
+	net_fill_ip6(&src, rta_get_ip6(a[RTA_SRC]), i->rtm_src_len);
+      else
+	net_fill_ip6(&src, IP6_NONE, 0);
       break;
 
 #ifdef HAVE_MPLS_KERNEL
@@ -1511,6 +1527,9 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h)
   if (!p)
     SKIP("unknown table %d\n", table);
 
+  if (a[RTA_SRC] && (p->p.net_type != NET_IP6_SADR))
+    SKIP("src prefix for non-SADR channel\n");
+
   if (a[RTA_IIF])
     SKIP("IIF set\n");
 
@@ -1533,25 +1552,33 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h)
       SKIP("proto unspec\n");
 
     case RTPROT_REDIRECT:
-      src = KRT_SRC_REDIRECT;
+      krt_src = KRT_SRC_REDIRECT;
       break;
 
     case RTPROT_KERNEL:
-      src = KRT_SRC_KERNEL;
+      krt_src = KRT_SRC_KERNEL;
       return;
 
     case RTPROT_BIRD:
       if (!s->scan)
 	SKIP("echo\n");
-      src = KRT_SRC_BIRD;
+      krt_src = KRT_SRC_BIRD;
       break;
 
     case RTPROT_BOOT:
     default:
-      src = KRT_SRC_ALIEN;
+      krt_src = KRT_SRC_ALIEN;
     }
 
-  net *net = net_get(p->p.main_channel->table, &dst);
+  net_addr *n = &dst;
+  if (p->p.net_type == NET_IP6_SADR)
+  {
+    n = alloca(sizeof(net_addr_ip6_sadr));
+    net_fill_ip6_sadr(n, net6_prefix(&dst), net6_pxlen(&dst),
+		      net6_prefix(&src), net6_pxlen(&src));
+  }
+
+  net *net = net_get(p->p.main_channel->table, n);
 
   if (s->net && !nl_mergable_route(s, net, p, priority, i->rtm_type))
     nl_announce_route(s);
@@ -1755,7 +1782,7 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h)
     s->attrs = ra;
     s->proto = p;
     s->new = new;
-    s->krt_src = src;
+    s->krt_src = krt_src;
     s->krt_type = i->rtm_type;
     s->krt_proto = i->rtm_protocol;
     s->krt_metric = priority;
diff --git a/sysdep/unix/krt.c b/sysdep/unix/krt.c
index a3cbb336d..b4fb1967 100644
--- a/sysdep/unix/krt.c
+++ b/sysdep/unix/krt.c
@@ -1108,10 +1108,11 @@ krt_start(struct proto *P)
 
   switch (p->p.net_type)
   {
-  case NET_IP4:	p->af = AF_INET; break;
-  case NET_IP6:	p->af = AF_INET6; break;
+  case NET_IP4:		p->af = AF_INET; break;
+  case NET_IP6:		p->af = AF_INET6; break;
+  case NET_IP6_SADR:	p->af = AF_INET6; break;
 #ifdef AF_MPLS
-  case NET_MPLS: p->af = AF_MPLS; break;
+  case NET_MPLS:	p->af = AF_MPLS; break;
 #endif
   default: log(L_ERR "KRT: Tried to start with strange net type: %d", p->p.net_type); return PS_START; break;
   }
@@ -1219,16 +1220,24 @@ krt_get_attr(eattr *a, byte *buf, int buflen)
 }
 
 
+#ifdef CONFIG_IP6_SADR_KERNEL
+#define MAYBE_IP6_SADR	NB_IP6_SADR
+#else
+#define MAYBE_IP6_SADR	0
+#endif
+
+#ifdef HAVE_MPLS_KERNEL
+#define MAYBE_MPLS	NB_MPLS
+#else
+#define MAYBE_MPLS	0
+#endif
+
 struct protocol proto_unix_kernel = {
   .name =		"Kernel",
   .template =		"kernel%d",
   .attr_class =		EAP_KRT,
   .preference =		DEF_PREF_INHERITED,
-#ifdef HAVE_MPLS_KERNEL
-  .channel_mask =	NB_IP | NB_MPLS,
-#else
-  .channel_mask =	NB_IP,
-#endif
+  .channel_mask =	NB_IP | MAYBE_IP6_SADR | MAYBE_MPLS,
   .proto_size =		sizeof(struct krt_proto),
   .config_size =	sizeof(struct krt_config),
   .preconfig =		krt_preconfig,

From 185a0a51f8aed635eecac0cfbd837dd262a8add0 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 13 Feb 2018 16:39:36 +0100
Subject: [PATCH 132/198] Babel: Add source-specific routing support

This patch adds support for source-specific routing to the Babel protocol.
It changes the protocol to support both NET_IP6 and NET_IP6_SADR channels
for IPv6 addresses. If only a NET_IP6 channel is configured,
source-specific updates are ignored. Otherwise, non-source-specific
routes are simply treated as source-specific routes with SADR prefix 0.

Thanks to Toke Hoiland-Jorgensen for the original patch.
Minor changes by Ondrej Santiago Zajicek.
---
 proto/babel/babel.c   |  58 +++++++++++----
 proto/babel/babel.h   |  25 ++++++-
 proto/babel/packets.c | 169 +++++++++++++++++++++++++++++++++++++++---
 3 files changed, 224 insertions(+), 28 deletions(-)

diff --git a/proto/babel/babel.c b/proto/babel/babel.c
index aa7e8b68..a60ffe78 100644
--- a/proto/babel/babel.c
+++ b/proto/babel/babel.c
@@ -1935,8 +1935,10 @@ babel_show_neighbors(struct proto *P, char *iff)
 }
 
 static void
-babel_show_entries_(struct babel_proto *p UNUSED, struct fib *rtable)
+babel_show_entries_(struct babel_proto *p, struct fib *rtable)
 {
+  int width = babel_sadr_enabled(p) ? -54 : -24;
+
   FIB_WALK(rtable, struct babel_entry, e)
   {
     struct babel_route *r = NULL;
@@ -1950,13 +1952,13 @@ babel_show_entries_(struct babel_proto *p UNUSED, struct fib *rtable)
       srcs++;
 
     if (e->valid)
-      cli_msg(-1025, "%-24N %-23lR %6u %5u %7u %7u",
+      cli_msg(-1025, "%-*N %-23lR %6u %5u %7u %7u", width,
 	      e->n.addr, e->router_id, e->metric, e->seqno, rts, srcs);
     else if (r = e->selected)
-      cli_msg(-1025, "%-24N %-23lR %6u %5u %7u %7u",
+      cli_msg(-1025, "%-*N %-23lR %6u %5u %7u %7u", width,
 	      e->n.addr, r->router_id, r->metric, r->seqno, rts, srcs);
     else
-      cli_msg(-1025, "%-24N %-23s %6s %5s %7u %7u",
+      cli_msg(-1025, "%-*N %-23s %6s %5s %7u %7u", width,
 	      e->n.addr, "<none>", "-", "-", rts, srcs);
   }
   FIB_WALK_END;
@@ -1966,6 +1968,7 @@ void
 babel_show_entries(struct proto *P)
 {
   struct babel_proto *p = (void *) P;
+  int width = babel_sadr_enabled(p) ? -54 : -24;
 
   if (p->p.proto_state != PS_UP)
   {
@@ -1975,7 +1978,7 @@ babel_show_entries(struct proto *P)
   }
 
   cli_msg(-1025, "%s:", p->p.name);
-  cli_msg(-1025, "%-24s %-23s %6s %5s %7s %7s",
+  cli_msg(-1025, "%-*s %-23s %6s %5s %7s %7s", width,
 	  "Prefix", "Router ID", "Metric", "Seqno", "Routes", "Sources");
 
   babel_show_entries_(p, &p->ip4_rtable);
@@ -1985,8 +1988,10 @@ babel_show_entries(struct proto *P)
 }
 
 static void
-babel_show_routes_(struct babel_proto *p UNUSED, struct fib *rtable)
+babel_show_routes_(struct babel_proto *p, struct fib *rtable)
 {
+  int width = babel_sadr_enabled(p) ? -54 : -24;
+
   FIB_WALK(rtable, struct babel_entry, e)
   {
     struct babel_route *r;
@@ -1994,7 +1999,7 @@ babel_show_routes_(struct babel_proto *p UNUSED, struct fib *rtable)
     {
       char c = (r == e->selected) ? '*' : (r->feasible ? '+' : ' ');
       btime time = r->expires ? r->expires - current_time() : 0;
-      cli_msg(-1025, "%-24N %-25I %-10s %5u %c %5u %7t",
+      cli_msg(-1025, "%-*N %-25I %-10s %5u %c %5u %7t", width,
 	      e->n.addr, r->next_hop, r->neigh->ifa->ifname,
 	      r->metric, c, r->seqno, MAX(time, 0));
     }
@@ -2006,6 +2011,7 @@ void
 babel_show_routes(struct proto *P)
 {
   struct babel_proto *p = (void *) P;
+  int width = babel_sadr_enabled(p) ? -54 : -24;
 
   if (p->p.proto_state != PS_UP)
   {
@@ -2015,7 +2021,7 @@ babel_show_routes(struct proto *P)
   }
 
   cli_msg(-1025, "%s:", p->p.name);
-  cli_msg(-1025, "%-24s %-25s %-9s %6s F %5s %7s",
+  cli_msg(-1025, "%-*s %-25s %-9s %6s F %5s %7s", width,
 	  "Prefix", "Nexthop", "Interface", "Metric", "Seqno", "Expires");
 
   babel_show_routes_(p, &p->ip4_rtable);
@@ -2182,14 +2188,32 @@ babel_rte_same(struct rte *new, struct rte *old)
 }
 
 
+static void
+babel_postconfig(struct proto_config *CF)
+{
+  struct babel_config *cf = (void *) CF;
+  struct channel_config *ip4, *ip6, *ip6_sadr;
+
+  ip4 = proto_cf_find_channel(CF, NET_IP4);
+  ip6 = proto_cf_find_channel(CF, NET_IP6);
+  ip6_sadr = proto_cf_find_channel(CF, NET_IP6_SADR);
+
+  if (ip6 && ip6_sadr)
+    cf_error("Both ipv6 and ipv6-sadr channels");
+
+  cf->ip4_channel = ip4;
+  cf->ip6_channel = ip6 ?: ip6_sadr;
+}
+
 static struct proto *
 babel_init(struct proto_config *CF)
 {
   struct proto *P = proto_new(CF);
   struct babel_proto *p = (void *) P;
+  struct babel_config *cf = (void *) CF;
 
-  proto_configure_channel(P, &p->ip4_channel, proto_cf_find_channel(CF, NET_IP4));
-  proto_configure_channel(P, &p->ip6_channel, proto_cf_find_channel(CF, NET_IP6));
+  proto_configure_channel(P, &p->ip4_channel, cf->ip4_channel);
+  proto_configure_channel(P, &p->ip6_channel, cf->ip6_channel);
 
   P->if_notify = babel_if_notify;
   P->rt_notify = babel_rt_notify;
@@ -2207,10 +2231,11 @@ babel_start(struct proto *P)
 {
   struct babel_proto *p = (void *) P;
   struct babel_config *cf = (void *) P->cf;
+  u8 ip6_type = cf->ip6_channel ? cf->ip6_channel->net_type : NET_IP6;
 
   fib_init(&p->ip4_rtable, P->pool, NET_IP4, sizeof(struct babel_entry),
 	   OFFSETOF(struct babel_entry, n), 0, babel_init_entry);
-  fib_init(&p->ip6_rtable, P->pool, NET_IP6, sizeof(struct babel_entry),
+  fib_init(&p->ip6_rtable, P->pool, ip6_type, sizeof(struct babel_entry),
 	   OFFSETOF(struct babel_entry, n), 0, babel_init_entry);
 
   init_list(&p->interfaces);
@@ -2258,11 +2283,15 @@ babel_reconfigure(struct proto *P, struct proto_config *CF)
 {
   struct babel_proto *p = (void *) P;
   struct babel_config *new = (void *) CF;
+  u8 ip6_type = new->ip6_channel ? new->ip6_channel->net_type : NET_IP6;
 
   TRACE(D_EVENTS, "Reconfiguring");
 
-  if (!proto_configure_channel(P, &p->ip4_channel, proto_cf_find_channel(CF, NET_IP4)) ||
-      !proto_configure_channel(P, &p->ip6_channel, proto_cf_find_channel(CF, NET_IP6)))
+  if (p->ip6_rtable.addr_type != ip6_type)
+    return 0;
+
+  if (!proto_configure_channel(P, &p->ip4_channel, new->ip4_channel) ||
+      !proto_configure_channel(P, &p->ip6_channel, new->ip6_channel))
     return 0;
 
   p->p.cf = CF;
@@ -2280,9 +2309,10 @@ struct protocol proto_babel = {
   .template =		"babel%d",
   .attr_class =		EAP_BABEL,
   .preference =		DEF_PREF_BABEL,
-  .channel_mask =	NB_IP,
+  .channel_mask =	NB_IP | NB_IP6_SADR,
   .proto_size =		sizeof(struct babel_proto),
   .config_size =	sizeof(struct babel_config),
+  .postconfig =		babel_postconfig,
   .init =		babel_init,
   .dump =		babel_dump,
   .start =		babel_start,
diff --git a/proto/babel/babel.h b/proto/babel/babel.h
index 1128d261..b194ce30 100644
--- a/proto/babel/babel.h
+++ b/proto/babel/babel.h
@@ -85,7 +85,10 @@ enum babel_tlv_type {
 
 enum babel_subtlv_type {
   BABEL_SUBTLV_PAD1		= 0,
-  BABEL_SUBTLV_PADN		= 1
+  BABEL_SUBTLV_PADN		= 1,
+
+  /* Mandatory subtlvs */
+  BABEL_SUBTLV_SOURCE_PREFIX    = 128,
 };
 
 enum babel_iface_type {
@@ -109,6 +112,9 @@ struct babel_config {
   struct proto_config c;
   list iface_list;			/* List of iface configs (struct babel_iface_config) */
   uint hold_time;			/* Time to hold stale entries and unreachable routes */
+
+  struct channel_config *ip4_channel;
+  struct channel_config *ip6_channel;
 };
 
 struct babel_iface_config {
@@ -303,7 +309,10 @@ struct babel_msg_update {
   u16 seqno;
   u16 metric;
   u64 router_id;
-  net_addr net;
+  union {
+    net_addr net;
+    net_addr_ip6_sadr net_sadr;
+  };
   ip_addr next_hop;
   ip_addr sender;
 };
@@ -311,7 +320,10 @@ struct babel_msg_update {
 struct babel_msg_route_request {
   u8 type;
   u8 full;
-  net_addr net;
+  union {
+    net_addr net;
+    net_addr_ip6_sadr net_sadr;
+  };
 };
 
 struct babel_msg_seqno_request {
@@ -319,7 +331,10 @@ struct babel_msg_seqno_request {
   u8 hop_count;
   u16 seqno;
   u64 router_id;
-  net_addr net;
+  union {
+    net_addr net;
+    net_addr_ip6_sadr net_sadr;
+  };
   ip_addr sender;
 };
 
@@ -339,6 +354,8 @@ struct babel_msg_node {
   union babel_msg msg;
 };
 
+static inline int babel_sadr_enabled(struct babel_proto *p)
+{ return p->ip6_rtable.addr_type == NET_IP6_SADR; }
 
 /* babel.c */
 void babel_handle_ack_req(union babel_msg *msg, struct babel_iface *ifa);
diff --git a/proto/babel/packets.c b/proto/babel/packets.c
index dd86222a..59678678 100644
--- a/proto/babel/packets.c
+++ b/proto/babel/packets.c
@@ -105,6 +105,13 @@ struct babel_tlv_seqno_request {
   u8 addr[0];
 } PACKED;
 
+struct babel_subtlv_source_prefix {
+  u8 type;
+  u8 length;
+  u8 plen;
+  u8 addr[0];
+} PACKED;
+
 
 /* Hello flags */
 #define BABEL_HF_UNICAST	0x8000
@@ -127,6 +134,7 @@ struct babel_parse_state {
   u8 def_ip6_prefix_seen;	/* def_ip6_prefix is valid */
   u8 def_ip4_prefix_seen;	/* def_ip4_prefix is valid */
   u8 current_tlv_endpos;	/* End of self-terminating TLVs (offset from start) */
+  u8 sadr_enabled;
 };
 
 enum parse_result {
@@ -237,6 +245,7 @@ static int babel_read_next_hop(struct babel_tlv *hdr, union babel_msg *msg, stru
 static int babel_read_update(struct babel_tlv *hdr, union babel_msg *msg, struct babel_parse_state *state);
 static int babel_read_route_request(struct babel_tlv *hdr, union babel_msg *msg, struct babel_parse_state *state);
 static int babel_read_seqno_request(struct babel_tlv *hdr, union babel_msg *msg, struct babel_parse_state *state);
+static int babel_read_source_prefix(struct babel_tlv *hdr, union babel_msg *msg, struct babel_parse_state *state);
 
 static uint babel_write_ack(struct babel_tlv *hdr, union babel_msg *msg, struct babel_write_state *state, uint max_len);
 static uint babel_write_hello(struct babel_tlv *hdr, union babel_msg *msg, struct babel_write_state *state, uint max_len);
@@ -244,6 +253,7 @@ static uint babel_write_ihu(struct babel_tlv *hdr, union babel_msg *msg, struct
 static uint babel_write_update(struct babel_tlv *hdr, union babel_msg *msg, struct babel_write_state *state, uint max_len);
 static uint babel_write_route_request(struct babel_tlv *hdr, union babel_msg *msg, struct babel_write_state *state, uint max_len);
 static uint babel_write_seqno_request(struct babel_tlv *hdr, union babel_msg *msg, struct babel_write_state *state, uint max_len);
+static int babel_write_source_prefix(struct babel_tlv *hdr, net_addr *net, uint max_len);
 
 struct babel_tlv_data {
   u8 min_length;
@@ -640,6 +650,9 @@ babel_read_update(struct babel_tlv *hdr, union babel_msg *m,
     ip6_addr prefix6 = get_ip6(buf);
     net_fill_ip6(&msg->net, prefix6, tlv->plen);
 
+    if (state->sadr_enabled)
+      net_make_ip6_sadr(&msg->net);
+
     if (tlv->flags & BABEL_UF_DEF_PREFIX)
     {
       put_ip6(state->def_ip6_prefix, prefix6);
@@ -770,12 +783,21 @@ babel_write_update(struct babel_tlv *hdr, union babel_msg *m,
   put_u16(&tlv->seqno, msg->seqno);
   put_u16(&tlv->metric, msg->metric);
 
+  if (msg->net.type == NET_IP6_SADR)
+  {
+    int l = babel_write_source_prefix(hdr, &msg->net, max_len - (len0 + len));
+    if (l < 0)
+      return 0;
+
+    len += l;
+  }
+
   return len0 + len;
 }
 
 static int
 babel_read_route_request(struct babel_tlv *hdr, union babel_msg *m,
-                         struct babel_parse_state *state UNUSED)
+                         struct babel_parse_state *state)
 {
   struct babel_tlv_route_request *tlv = (void *) hdr;
   struct babel_msg_route_request *msg = &m->route_request;
@@ -812,6 +834,10 @@ babel_read_route_request(struct babel_tlv *hdr, union babel_msg *m,
 
     read_ip6_px(&msg->net, tlv->addr, tlv->plen);
     state->current_tlv_endpos += BYTES(tlv->plen);
+
+    if (state->sadr_enabled)
+      net_make_ip6_sadr(&msg->net);
+
     return PARSE_SUCCESS;
 
   case BABEL_AE_IP6_LL:
@@ -856,6 +882,15 @@ babel_write_route_request(struct babel_tlv *hdr, union babel_msg *m,
     put_ip6_px(tlv->addr, &msg->net);
   }
 
+  if (msg->net.type == NET_IP6_SADR)
+  {
+    int l = babel_write_source_prefix(hdr, &msg->net, max_len - len);
+    if (l < 0)
+      return 0;
+
+    len += l;
+  }
+
   return len;
 }
 
@@ -900,6 +935,10 @@ babel_read_seqno_request(struct babel_tlv *hdr, union babel_msg *m,
 
     read_ip6_px(&msg->net, tlv->addr, tlv->plen);
     state->current_tlv_endpos += BYTES(tlv->plen);
+
+    if (state->sadr_enabled)
+      net_make_ip6_sadr(&msg->net);
+
     return PARSE_SUCCESS;
 
   case BABEL_AE_IP6_LL:
@@ -943,38 +982,147 @@ babel_write_seqno_request(struct babel_tlv *hdr, union babel_msg *m,
   tlv->hop_count = msg->hop_count;
   put_u64(&tlv->router_id, msg->router_id);
 
+  if (msg->net.type == NET_IP6_SADR)
+  {
+    int l = babel_write_source_prefix(hdr, &msg->net, max_len - len);
+    if (l < 0)
+      return 0;
+
+    len += l;
+  }
+
   return len;
 }
 
+static int
+babel_read_source_prefix(struct babel_tlv *hdr, union babel_msg *msg,
+			 struct babel_parse_state *state UNUSED)
+{
+  struct babel_subtlv_source_prefix *tlv = (void *) hdr;
+  net_addr_ip6_sadr *net;
+
+  /*
+   * We would like to skip the sub-TLV if SADR is not enabled, but we do not
+   * know AF of the enclosing TLV yet. We will do that later.
+   */
+
+  /* Check internal consistency */
+  if ((tlv->length < 1) ||
+      (tlv->plen > IP6_MAX_PREFIX_LENGTH) ||
+      (tlv->length < (1 + BYTES(tlv->plen))))
+    return PARSE_ERROR;
+
+  /* Plen MUST NOT be 0 */
+  if (tlv->plen == 0)
+    return PARSE_ERROR;
+
+  switch(msg->type)
+  {
+  case BABEL_TLV_UPDATE:
+    /* Wildcard updates with source prefix MUST be silently ignored */
+    if (msg->update.wildcard)
+      return PARSE_IGNORE;
+
+    net = (void *) &msg->update.net;
+    break;
+
+  case BABEL_TLV_ROUTE_REQUEST:
+    /* Wildcard requests with source addresses MUST be silently ignored */
+    if (msg->route_request.full)
+      return PARSE_IGNORE;
+
+    net = (void *) &msg->route_request.net;
+    break;
+
+  case BABEL_TLV_SEQNO_REQUEST:
+    net = (void *) &msg->seqno_request.net;
+    break;
+
+  default:
+    return PARSE_ERROR;
+  }
+
+  /* If SADR is active, the net has appropriate type */
+  if (net->type != NET_IP6_SADR)
+    return PARSE_IGNORE;
+
+  /* Duplicate Source Prefix sub-TLV; SHOULD ignore whole TLV */
+  if (net->src_pxlen > 0)
+    return PARSE_IGNORE;
+
+  net_addr_ip6 src;
+  read_ip6_px((void *) &src, tlv->addr, tlv->plen);
+  net->src_prefix = src.prefix;
+  net->src_pxlen = src.pxlen;
+
+  return PARSE_SUCCESS;
+}
+
+static int
+babel_write_source_prefix(struct babel_tlv *hdr, net_addr *n, uint max_len)
+{
+  struct babel_subtlv_source_prefix *tlv = (void *) NEXT_TLV(hdr);
+  net_addr_ip6_sadr *net = (void *) n;
+
+  /* Do not use this sub-TLV for default prefix */
+  if (net->src_pxlen == 0)
+    return 0;
+
+  uint len = sizeof(*tlv) + BYTES(net->src_pxlen);
+
+  if (len > max_len)
+    return -1;
+
+  TLV_HDR(tlv, BABEL_SUBTLV_SOURCE_PREFIX, len);
+  hdr->length += len;
+
+  net_addr_ip6 src = NET_ADDR_IP6(net->src_prefix, net->src_pxlen);
+  tlv->plen = src.pxlen;
+  put_ip6_px(tlv->addr, (void *) &src);
+
+  return len;
+}
+
+
 static inline int
 babel_read_subtlvs(struct babel_tlv *hdr,
-		   union babel_msg *msg UNUSED,
+		   union babel_msg *msg,
 		   struct babel_parse_state *state)
 {
   struct babel_tlv *tlv;
+  byte *pos, *end = (byte *) hdr + TLV_LENGTH(hdr);
+  int res;
 
   for (tlv = (void *) hdr + state->current_tlv_endpos;
-       (void *) tlv < (void *) hdr + TLV_LENGTH(hdr);
+       (byte *) tlv < end;
        tlv = NEXT_TLV(tlv))
   {
+    /* Ugly special case */
+    if (tlv->type == BABEL_TLV_PAD1)
+      continue;
+
+    /* The end of the common TLV header */
+    pos = (byte *)tlv + sizeof(struct babel_tlv);
+    if ((pos > end) || (pos + tlv->length > end))
+      return PARSE_ERROR;
+
     /*
      * The subtlv type space is non-contiguous (due to the mandatory bit), so
      * use a switch for dispatch instead of the mapping array we use for TLVs
      */
     switch (tlv->type)
     {
-    case BABEL_SUBTLV_PAD1:
-    case BABEL_SUBTLV_PADN:
-      /* FIXME: Framing errors in PADN are silently ignored, see babel_process_packet() */
+    case BABEL_SUBTLV_SOURCE_PREFIX:
+      res = babel_read_source_prefix(tlv, msg, state);
+      if (res != PARSE_SUCCESS)
+	return res;
       break;
 
+    case BABEL_SUBTLV_PADN:
     default:
       /* Unknown mandatory subtlv; PARSE_IGNORE ignores the whole TLV */
-      if (tlv->type > 128)
-      {
-	DBG("Babel: Mandatory subtlv %d found; skipping TLV\n", tlv->type);
+      if (tlv->type >= 128)
 	return PARSE_IGNORE;
-      }
       break;
     }
   }
@@ -1197,6 +1345,7 @@ babel_process_packet(struct babel_pkt_header *pkt, int len,
     .ifa	  = ifa,
     .saddr	  = saddr,
     .next_hop_ip6 = saddr,
+    .sadr_enabled = babel_sadr_enabled(p),
   };
 
   if ((pkt->magic != BABEL_MAGIC) || (pkt->version != BABEL_VERSION))

From 5ce7adfcf96787878d60668cf33acf7349a69c0b Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 13 Feb 2018 16:42:03 +0100
Subject: [PATCH 133/198] Babel: Fix accidental bitwise or assignment

Fix an accidental bitwise or assignment that was supposed to be a
comparison.

Thanks to Toke Hoiland-Jorgensen for the patch.
---
 proto/babel/babel.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/proto/babel/babel.c b/proto/babel/babel.c
index a60ffe78..88c4711e 100644
--- a/proto/babel/babel.c
+++ b/proto/babel/babel.c
@@ -2145,7 +2145,7 @@ babel_rt_notify(struct proto *P, struct channel *c UNUSED, struct network *net,
     e = babel_get_entry(p, net->n.addr);
 
     /* Activate triggered updates */
-    if ((e->valid |= BABEL_ENTRY_VALID) ||
+    if ((e->valid != BABEL_ENTRY_VALID) ||
 	(e->router_id != rt_router_id))
     {
       babel_trigger_update(p);

From d5144ea9bf01f0b19cba40ab06aa05a11d70aa3e Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 13 Feb 2018 17:00:24 +0100
Subject: [PATCH 134/198] Add cscope Makefile target

For those who prefer cscope to etags

Thanks to Toke Hoiland-Jorgensen for the patch.
---
 .gitignore  | 1 +
 Makefile.in | 5 ++++-
 2 files changed, 5 insertions(+), 1 deletion(-)

diff --git a/.gitignore b/.gitignore
index 0dcebfd1..3b734f49 100644
--- a/.gitignore
+++ b/.gitignore
@@ -12,3 +12,4 @@
 /configure
 /sysdep/autoconf.h.in
 /sysdep/autoconf.h.in~
+/cscope.*
diff --git a/Makefile.in b/Makefile.in
index fdd5e6c7..c8168bbe 100644
--- a/Makefile.in
+++ b/Makefile.in
@@ -58,7 +58,7 @@ endif
 docgoals := docs userdocs progdocs
 testgoals := check test tests tests_run
 cleangoals := clean distclean testsclean
-.PHONY: all daemon cli $(docgoals) $(testgoals) $(cleangoals) tags
+.PHONY: all daemon cli $(docgoals) $(testgoals) $(cleangoals) tags cscope
 all: daemon cli
 
 daemon: $(daemon)
@@ -162,6 +162,9 @@ endif
 tags:
 	cd $(srcdir) ; etags -lc `find $(dirs) -name *.[chY]`
 
+cscope:
+	cd $(srcdir) ; find $(dirs) -name *.[chY] > cscope.files ; cscope -b
+
 # Install
 
 install: all

From 1561ee799cfe79d208ce9588e487da4b62a88dad Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 13 Feb 2018 19:52:22 +0100
Subject: [PATCH 135/198] Handle properly enums for extended attributes

---
 filter/f-util.c | 4 ++--
 filter/filter.c | 9 +++++----
 2 files changed, 7 insertions(+), 6 deletions(-)

diff --git a/filter/f-util.c b/filter/f-util.c
index 661941ec..52c13223 100644
--- a/filter/f-util.c
+++ b/filter/f-util.c
@@ -24,11 +24,11 @@ f_new_inst(void)
 }
 
 struct f_inst *
-f_new_dynamic_attr(int type, int f_type UNUSED, int code)
+f_new_dynamic_attr(int type, int f_type, int code)
 {
   /* FIXME: Remove the f_type parameter? */
   struct f_inst *f = f_new_inst();
-  f->aux = type;
+  f->aux = (f_type << 8) | type;
   f->a2.i = code;
   return f;
 }
diff --git a/filter/filter.c b/filter/filter.c
index 42e60a68..8cf90b53 100644
--- a/filter/filter.c
+++ b/filter/filter.c
@@ -1005,6 +1005,7 @@ interpret(struct f_inst *what)
     {
       eattr *e = NULL;
       u16 code = what->a2.i;
+      int f_type = what->aux >> 8;
 
       if (!(f_flags & FF_FORCE_TMPATTR))
 	e = ea_find((*f_rte)->attrs->eattrs, code);
@@ -1049,7 +1050,7 @@ interpret(struct f_inst *what)
 
       switch (what->aux & EAF_TYPE_MASK) {
       case EAF_TYPE_INT:
-	res.type = T_INT;
+	res.type = f_type;
 	res.val.i = e->u.data;
 	break;
       case EAF_TYPE_ROUTER_ID:
@@ -1099,18 +1100,18 @@ interpret(struct f_inst *what)
     {
       struct ea_list *l = lp_alloc(f_pool, sizeof(struct ea_list) + sizeof(eattr));
       u16 code = what->a2.i;
+      int f_type = what->aux >> 8;
 
       l->next = NULL;
       l->flags = EALF_SORTED;
       l->count = 1;
       l->attrs[0].id = code;
       l->attrs[0].flags = 0;
-      l->attrs[0].type = what->aux | EAF_ORIGINATED | EAF_FRESH;
+      l->attrs[0].type = (what->aux & 0xff) | EAF_ORIGINATED | EAF_FRESH;
 
       switch (what->aux & EAF_TYPE_MASK) {
       case EAF_TYPE_INT:
-	// Enums are also ints, so allow them in.
-	if (v1.type != T_INT && (v1.type < T_ENUM_LO || v1.type > T_ENUM_HI))
+	if (v1.type != f_type)
 	  runtime( "Setting int attribute to non-int value" );
 	l->attrs[0].u.data = v1.val.i;
 	break;

From 2d6d4b80539be13aa53c6751fb33689b263e4010 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 7 Mar 2018 17:35:24 +0100
Subject: [PATCH 136/198] Babel: Fix build with restricted protocol set

All keywords used in Babel config have to be declared locally.

Thanks to Leo Vandewoestijne for the bugreport.
---
 conf/confbase.Y      | 7 ++++++-
 proto/babel/config.Y | 5 +++--
 proto/ospf/config.Y  | 6 ------
 3 files changed, 9 insertions(+), 9 deletions(-)

diff --git a/conf/confbase.Y b/conf/confbase.Y
index 96b32028..9c08a41a 100644
--- a/conf/confbase.Y
+++ b/conf/confbase.Y
@@ -77,7 +77,7 @@ CF_DECLS
 %type <time> datetime
 %type <a> ipa
 %type <px> prefix prefix_or_ipa
-%type <t> text
+%type <t> text opttext
 %type <t> text_or_none
 
 %nonassoc PREFIX_DUMMY
@@ -200,6 +200,11 @@ text:
    }
  ;
 
+opttext:
+    TEXT
+ | /* empty */ { $$ = NULL; }
+ ;
+
 text_or_none:
    TEXT { $$ = $1; }
  |      { $$ = NULL; }
diff --git a/proto/babel/config.Y b/proto/babel/config.Y
index b6170852..dcc0847e 100644
--- a/proto/babel/config.Y
+++ b/proto/babel/config.Y
@@ -20,8 +20,9 @@ CF_DEFINES
 
 CF_DECLS
 
-CF_KEYWORDS(BABEL, METRIC, RXCOST, HELLO, UPDATE, INTERVAL, PORT, WIRED,
-WIRELESS, RX, TX, BUFFER, LENGTH, CHECK, LINK, BABEL_METRIC)
+CF_KEYWORDS(BABEL, INTERFACE, METRIC, RXCOST, HELLO, UPDATE, INTERVAL, PORT,
+	TYPE, WIRED, WIRELESS, RX, TX, BUFFER, PRIORITY, LENGTH, CHECK, LINK,
+	BABEL_METRIC, SHOW, INTERFACES, NEIGHBORS, ENTRIES)
 
 CF_GRAMMAR
 
diff --git a/proto/ospf/config.Y b/proto/ospf/config.Y
index 7b35b191..46b8b238 100644
--- a/proto/ospf/config.Y
+++ b/proto/ospf/config.Y
@@ -146,7 +146,6 @@ CF_KEYWORDS(WAIT, DELAY, LSADB, ECMP, LIMIT, WEIGHT, NSSA, TRANSLATOR, STABILITY
 CF_KEYWORDS(GLOBAL, LSID, ROUTER, SELF, INSTANCE, REAL, NETMASK, TX, PRIORITY, LENGTH)
 CF_KEYWORDS(SECONDARY, MERGE, LSA, SUPPRESSION)
 
-%type <t> opttext
 %type <ld> lsadb_args
 %type <i> nbma_eligible
 
@@ -415,11 +414,6 @@ ospf_iface:
   ospf_iface_start ospf_iface_patt_list ospf_iface_opt_list { ospf_iface_finish(); }
  ;
 
-opttext:
-    TEXT
- | /* empty */ { $$ = NULL; }
- ;
-
 CF_ADDTO(dynamic_attr, OSPF_METRIC1 { $$ = f_new_dynamic_attr(EAF_TYPE_INT | EAF_TEMP, T_INT, EA_OSPF_METRIC1); })
 CF_ADDTO(dynamic_attr, OSPF_METRIC2 { $$ = f_new_dynamic_attr(EAF_TYPE_INT | EAF_TEMP, T_INT, EA_OSPF_METRIC2); })
 CF_ADDTO(dynamic_attr, OSPF_TAG { $$ = f_new_dynamic_attr(EAF_TYPE_INT | EAF_TEMP, T_INT, EA_OSPF_TAG); })

From 0575c7db723523701d5582e2a9058cb5c46951c9 Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Tue, 6 Mar 2018 16:03:35 +0100
Subject: [PATCH 137/198] Config: Dropped the ipv4:netmask4 syntax for IPv4
 prefixes.

---
 conf/confbase.Y   | 4 ----
 doc/bird.sgml     | 3 +--
 filter/test.conf2 | 6 +++---
 3 files changed, 4 insertions(+), 9 deletions(-)

diff --git a/conf/confbase.Y b/conf/confbase.Y
index c2d647eb..55615528 100644
--- a/conf/confbase.Y
+++ b/conf/confbase.Y
@@ -177,10 +177,6 @@ pxlen4:
      if ($2 > IP4_MAX_PREFIX_LENGTH) cf_error("Invalid prefix length %u", $2);
      $$ = $2;
    }
- | ':' IP4 {
-     $$ = ip4_masklen($2);
-     if ($$ == 255) cf_error("Invalid netmask %I4", $2);
-   }
  ;
 
 net_ip4_: IP4 pxlen4
diff --git a/doc/bird.sgml b/doc/bird.sgml
index 04f9258b..5ed816e8 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -1234,8 +1234,7 @@ foot).
 	operator <cf/.type/. The type may be:
 
 	<cf/NET_IP4/ and <cf/NET_IP6/ prefixes hold an IP prefix. The literals
-	are written as <cf><m/ipaddress//<m/pxlen/</cf>,
-	or <cf><m>ipaddress</m>/<m>netmask</m></cf>. There are two special
+	are written as <cf><m/ipaddress//<m/pxlen/</cf>. There are two special
 	operators on these: <cf/.ip/ which extracts the IP address from the
 	pair, and <cf/.len/, which separates prefix length from the pair.
 	So <cf>1.2.0.0/16.len = 16</cf> is true.
diff --git a/filter/test.conf2 b/filter/test.conf2
index 2a5d896b..83554952 100644
--- a/filter/test.conf2
+++ b/filter/test.conf2
@@ -76,8 +76,8 @@ protocol static {
 	route 62.168.0.0/25 reject;
 	route 1.2.3.4/32 via 195.113.31.124;
 	route 10.0.0.0/8 reject;
-	route 10.1.1.0:255.255.255.0 via 62.168.0.3;
-	route 10.1.2.0:255.255.255.0 via 62.168.0.3;
-	route 10.1.3.0:255.255.255.0 via 62.168.0.4;
+	route 10.1.1.0/24 via 62.168.0.3;
+	route 10.1.2.0/24 via 62.168.0.3;
+	route 10.1.3.0/24 via 62.168.0.4;
 	route 10.2.0.0/24 via "arc0";
 }

From f2f5a7d9455f938fb14e31315a879c3be2c5d28a Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Tue, 6 Mar 2018 16:04:56 +0100
Subject: [PATCH 138/198] Filter: the test conf checks also a bit of BGP args

Uncommented an old test.
---
 filter/test.conf2 | 29 ++++++++++++++---------------
 1 file changed, 14 insertions(+), 15 deletions(-)

diff --git a/filter/test.conf2 b/filter/test.conf2
index 83554952..48515020 100644
--- a/filter/test.conf2
+++ b/filter/test.conf2
@@ -55,21 +55,20 @@ protocol static {
 			rip_metric = rip_metric + 5;
 			print rip_metric;
 
-#
-# TODO: uncomment this part after finishing BGP integration version
-#
-#			bgp_community = -empty-;
-#			print "hi";
-#			bgp_community = add(bgp_community, (1,2));
-#			print "hello";
-#			bgp_community = add(bgp_community, (2,3));
-#			bgp_community.add((4,5));
-#			print "community = ", bgp_community;
-#			bgp_community.delete((2,3));
-#			print "community = ", bgp_community;
-#			bgp_community.empty;
-#			print "community = ", bgp_community;
-#			print "done";
+			bgp_community = -empty-;
+			print "hi";
+			bgp_community = add(bgp_community, (1,2));
+			print "hello";
+			bgp_community = add(bgp_community, (2,3));
+			bgp_community.add((4,5));
+			print "community = ", bgp_community;
+			bgp_community.delete((2,3));
+			print "community = ", bgp_community;
+			bgp_community.empty;
+			print "community = ", bgp_community;
+			print "done";
+
+			accept;
 		};
 	};
 	route 0.0.0.0/0 via 195.113.31.113;

From 5a14df395053f4094a1e3ebea98e3487cbfc0e63 Mon Sep 17 00:00:00 2001
From: Maria Jan Matejka <mq@jmq.cz>
Date: Thu, 19 Oct 2017 12:39:44 +0200
Subject: [PATCH 139/198] Filter: Instruction codes named as enum

The two-letter instructions were quite messy but they could be easily
read from memory dumps. Now GDB (since 2012) supports pretty printing
enum values and GCC checks the switch construction for missing enum
values so we are converting the nice two-byte values to enums.

Anyway, the enum still keeps the old two-byte values to be able to read
the instruction codes even without GDB from plain memory dump.
---
 conf/confbase.Y |   2 +
 filter/config.Y | 234 +++++++++++++++++++++---------------------------
 filter/f-util.c |  44 +++++----
 filter/filter.c | 194 ++++++++++++++++++++-------------------
 filter/filter.h |  83 ++++++++++++++++-
 5 files changed, 305 insertions(+), 252 deletions(-)

diff --git a/conf/confbase.Y b/conf/confbase.Y
index 9c08a41a..b8deed54 100644
--- a/conf/confbase.Y
+++ b/conf/confbase.Y
@@ -43,6 +43,8 @@ CF_DECLS
   char *t;
   struct rtable_config *r;
   struct f_inst *x;
+  struct f_dynamic_attr fda;
+  struct f_static_attr fsa;
   struct filter *f;
   struct f_tree *e;
   struct f_trie *trie;
diff --git a/filter/config.Y b/filter/config.Y
index c9f9c753..1ef5a3a8 100644
--- a/filter/config.Y
+++ b/filter/config.Y
@@ -12,8 +12,6 @@ CF_HDR
 
 CF_DEFINES
 
-#define P(a,b) ((a << 8) | b)
-
 static inline u32 pair(u32 a, u32 b) { return (a << 16) | b; }
 static inline u32 pair_a(u32 p) { return p >> 16; }
 static inline u32 pair_b(u32 p) { return p & 0xFFFF; }
@@ -157,12 +155,11 @@ f_new_lc_item(u32 f1, u32 t1, u32 f2, u32 t2, u32 f3, u32 t3)
 }
 
 static inline struct f_inst *
-f_generate_empty(struct f_inst *dyn)
+f_generate_empty(struct f_dynamic_attr dyn)
 {
-  struct f_inst *e = f_new_inst();
-  e->code = 'E';
+  struct f_inst *e = f_new_inst(FI_EMPTY);
 
-  switch (dyn->aux & EAF_TYPE_MASK) {
+  switch (dyn.type & EAF_TYPE_MASK) {
     case EAF_TYPE_AS_PATH:
       e->aux = T_PATH;
       break;
@@ -179,9 +176,9 @@ f_generate_empty(struct f_inst *dyn)
       cf_error("Can't empty that attribute");
   }
 
-  dyn->code = P('e','S');
-  dyn->a1.p = e;
-  return dyn;
+  struct f_inst *s = f_new_inst_da(FI_EA_SET, dyn);
+  s->a1.p = e;
+  return s;
 }
 
 
@@ -190,21 +187,19 @@ f_generate_dpair(struct f_inst *t1, struct f_inst *t2)
 {
   struct f_inst *rv;
 
-  if ((t1->code == 'c') && (t2->code == 'c')) {
+  if ((t1->fi_code == FI_CONSTANT) && (t2->fi_code == FI_CONSTANT)) {
     if ((t1->aux != T_INT) || (t2->aux != T_INT))
       cf_error( "Can't operate with value of non-integer type in pair constructor");
 
     check_u16(t1->a2.i);
     check_u16(t2->a2.i);
 
-    rv = f_new_inst();
-    rv->code = 'c';
+    rv = f_new_inst(FI_CONSTANT);
     rv->aux = T_PAIR;
     rv->a2.i = pair(t1->a2.i, t2->a2.i);
   }
   else {
-    rv = f_new_inst();
-    rv->code = P('m', 'p');
+    rv = f_new_inst(FI_PAIR_CONSTRUCT);
     rv->a1.p = t1;
     rv->a2.p = t2;
   }
@@ -219,7 +214,7 @@ f_generate_ec(u16 kind, struct f_inst *tk, struct f_inst *tv)
   int c1 = 0, c2 = 0, ipv4_used = 0;
   u32 key = 0, val2 = 0;
 
-  if (tk->code == 'c') {
+  if (tk->fi_code == FI_CONSTANT) {
     c1 = 1;
 
     if (tk->aux == T_INT) {
@@ -234,7 +229,7 @@ f_generate_ec(u16 kind, struct f_inst *tk, struct f_inst *tv)
 
 #ifndef IPV6
   /* IP->Quad implicit conversion */
-  else if (tk->code == 'C') {
+  else if (tk->fi_code == FI_CONSTANT_INDIRECT) {
     c1 = 1;
     struct f_val *val = tk->a1.p;
 
@@ -252,7 +247,7 @@ f_generate_ec(u16 kind, struct f_inst *tk, struct f_inst *tv)
   }
 #endif
 
-  if (tv->code == 'c') {
+  if (tv->fi_code == FI_CONSTANT) {
     if (tv->aux != T_INT)
       cf_error("Can't operate with value of non-integer type in EC constructor");
     c2 = 1;
@@ -278,15 +273,13 @@ f_generate_ec(u16 kind, struct f_inst *tk, struct f_inst *tv)
     }
 
     NEW_F_VAL;
-    rv = f_new_inst();
-    rv->code = 'C';
+    rv = f_new_inst(FI_CONSTANT_INDIRECT);
     rv->a1.p = val;
     val->type = T_EC;
     val->val.ec = ec;
   }
   else {
-    rv = f_new_inst();
-    rv->code = P('m','c');
+    rv = f_new_inst(FI_EC_CONSTRUCT);
     rv->aux = kind;
     rv->a1.p = tk;
     rv->a2.p = tv;
@@ -300,12 +293,11 @@ f_generate_lc(struct f_inst *t1, struct f_inst *t2, struct f_inst *t3)
 {
   struct f_inst *rv;
 
-  if ((t1->code == 'c') && (t2->code == 'c') && (t3->code == 'c')) {
+  if ((t1->fi_code == FI_CONSTANT) && (t2->fi_code == FI_CONSTANT) && (t3->fi_code == FI_CONSTANT)) {
     if ((t1->aux != T_INT) || (t2->aux != T_INT) || (t3->aux != T_INT))
       cf_error( "LC - Can't operate with value of non-integer type in tuple constructor");
 
-    rv = f_new_inst();
-    rv->code = 'C';
+    rv = f_new_inst(FI_CONSTANT_INDIRECT);
 
     NEW_F_VAL;
     rv->a1.p = val;
@@ -316,7 +308,7 @@ f_generate_lc(struct f_inst *t1, struct f_inst *t2, struct f_inst *t3)
   {
     rv = cfg_allocz(sizeof(struct f_inst3));
     rv->lineno = ifs->lino;
-    rv->code = P('m','l');
+    rv->fi_code = FI_LC_CONSTRUCT;
     rv->a1.p = t1;
     rv->a2.p = t2;
     INST3(rv).p = t3;
@@ -348,7 +340,9 @@ CF_KEYWORDS(FUNCTION, PRINT, PRINTN, UNSET, RETURN,
 %nonassoc THEN
 %nonassoc ELSE
 
-%type <x> term block cmds cmds_int cmd function_body constant constructor print_one print_list var_list var_listn dynamic_attr static_attr function_call symbol bgp_path_expr
+%type <x> term block cmds cmds_int cmd function_body constant constructor print_one print_list var_list var_listn function_call symbol bgp_path_expr
+%type <fda> dynamic_attr
+%type <fsa> static_attr
 %type <f> filter filter_body where_filter
 %type <i> type break_command ec_kind
 %type <i32> cnum
@@ -461,16 +455,13 @@ where_filter:
      /* Construct 'IF term THEN ACCEPT; REJECT;' */
      struct filter *f = cfg_alloc(sizeof(struct filter));
      struct f_inst *i, *acc, *rej;
-     acc = f_new_inst();		/* ACCEPT */
-     acc->code = P('p',',');
+     acc = f_new_inst(FI_PRINT_AND_DIE);	/* ACCEPT */
      acc->a1.p = NULL;
      acc->a2.i = F_ACCEPT;
-     rej = f_new_inst();		/* REJECT */
-     rej->code = P('p',',');
+     rej = f_new_inst(FI_PRINT_AND_DIE);	/* REJECT */
      rej->a1.p = NULL;
      rej->a2.i = F_REJECT;
-     i = f_new_inst();			/* IF */
-     i->code = '?';
+     i = f_new_inst(FI_CONDITION);			/* IF */
      i->a1.p = $2;
      i->a2.p = acc;
      i->next = rej;
@@ -489,8 +480,7 @@ function_body:
    decls '{' cmds '}' {
      if ($1) {
        /* Prepend instruction to clear local variables */
-       $$ = f_new_inst();
-       $$->code = P('c','v');
+       $$ = f_new_inst(FI_CLEAR_LOCAL_VARS);
        $$->a1.p = $1;
        $$->next = $3;
      } else
@@ -680,7 +670,7 @@ switch_body: /* EMPTY */ { $$ = NULL; }
  }
  ;
 
-/* CONST '(' expr ')' { $$ = f_new_inst(); $$->code = 'c'; $$->aux = T_INT; $$->a2.i = $3; } */
+/* CONST '(' expr ')' { $$ = f_new_inst(FI_CONSTANT); $$->aux = T_INT; $$->a2.i = $3; } */
 
 bgp_path_expr:
    symbol       { $$ = $1; }
@@ -708,17 +698,17 @@ bgp_path_tail2:
  ;
 
 constant:
-   NUM    { $$ = f_new_inst(); $$->code = 'c'; $$->aux = T_INT;  $$->a2.i = $1; }
- | TRUE   { $$ = f_new_inst(); $$->code = 'c'; $$->aux = T_BOOL; $$->a2.i = 1;  }
- | FALSE  { $$ = f_new_inst(); $$->code = 'c'; $$->aux = T_BOOL; $$->a2.i = 0;  }
- | TEXT   { $$ = f_new_inst(); $$->code = 'c'; $$->aux = T_STRING; $$->a2.p = $1; }
- | fipa	   { NEW_F_VAL; $$ = f_new_inst(); $$->code = 'C'; $$->a1.p = val; *val = $1; }
- | fprefix_s {NEW_F_VAL; $$ = f_new_inst(); $$->code = 'C'; $$->a1.p = val; *val = $1; }
- | RTRID  { $$ = f_new_inst(); $$->code = 'c'; $$->aux = T_QUAD;  $$->a2.i = $1; }
- | '[' set_items ']' { DBG( "We've got a set here..." ); $$ = f_new_inst(); $$->code = 'c'; $$->aux = T_SET; $$->a2.p = build_tree($2); DBG( "ook\n" ); }
- | '[' fprefix_set ']' { $$ = f_new_inst(); $$->code = 'c'; $$->aux = T_PREFIX_SET;  $$->a2.p = $2; }
- | ENUM	  { $$ = f_new_inst(); $$->code = 'c'; $$->aux = $1 >> 16; $$->a2.i = $1 & 0xffff; }
- | bgp_path { NEW_F_VAL; $$ = f_new_inst(); $$->code = 'C'; val->type = T_PATH_MASK; val->val.path_mask = $1; $$->a1.p = val; }
+   NUM    { $$ = f_new_inst(FI_CONSTANT); $$->aux = T_INT;  $$->a2.i = $1; }
+ | TRUE   { $$ = f_new_inst(FI_CONSTANT); $$->aux = T_BOOL; $$->a2.i = 1;  }
+ | FALSE  { $$ = f_new_inst(FI_CONSTANT); $$->aux = T_BOOL; $$->a2.i = 0;  }
+ | TEXT   { $$ = f_new_inst(FI_CONSTANT); $$->aux = T_STRING; $$->a2.p = $1; }
+ | fipa	   { NEW_F_VAL; $$ = f_new_inst(FI_CONSTANT_INDIRECT); $$->a1.p = val; *val = $1; }
+ | fprefix_s {NEW_F_VAL; $$ = f_new_inst(FI_CONSTANT_INDIRECT); $$->a1.p = val; *val = $1; }
+ | RTRID  { $$ = f_new_inst(FI_CONSTANT); $$->aux = T_QUAD;  $$->a2.i = $1; }
+ | '[' set_items ']' { DBG( "We've got a set here..." ); $$ = f_new_inst(FI_CONSTANT); $$->aux = T_SET; $$->a2.p = build_tree($2); DBG( "ook\n" ); }
+ | '[' fprefix_set ']' { $$ = f_new_inst(FI_CONSTANT); $$->aux = T_PREFIX_SET;  $$->a2.p = $2; }
+ | ENUM	  { $$ = f_new_inst(FI_CONSTANT); $$->aux = $1 >> 16; $$->a2.i = $1 & 0xffff; }
+ | bgp_path { NEW_F_VAL; $$ = f_new_inst(FI_CONSTANT_INDIRECT); val->type = T_PATH_MASK; val->val.path_mask = $1; $$->a1.p = val; }
  ;
 
 constructor:
@@ -733,7 +723,7 @@ constructor:
  *  For such cases, we force the dynamic_attr list to contain
  *  at least an invalid token, so it is syntantically correct.
  */
-CF_ADDTO(dynamic_attr, INVALID_TOKEN { $$ = NULL; })
+CF_ADDTO(dynamic_attr, INVALID_TOKEN { $$ = (struct f_dynamic_attr) {}; })
 
 rtadot: /* EMPTY, we are not permitted RTA. prefix */
  ;
@@ -745,8 +735,7 @@ function_call:
      if ($1->class != SYM_FUNCTION)
        cf_error("You can't call something which is not a function. Really.");
      DBG("You are calling function %s\n", $1->name);
-     $$ = f_new_inst();
-     $$->code = P('c','a');
+     $$ = f_new_inst(FI_CALL);
      $$->a1.p = inst;
      $$->a2.p = $1->def;
      sym = $1->aux2;
@@ -763,11 +752,9 @@ function_call:
 
 symbol:
    SYM {
-     $$ = f_new_inst();
-
      switch ($1->class & 0xff00) {
-       case SYM_CONSTANT: $$->code = 'C'; break;
-       case SYM_VARIABLE: $$->code = 'V'; break;
+       case SYM_CONSTANT: $$ = f_new_inst(FI_CONSTANT_INDIRECT); break;
+       case SYM_VARIABLE: $$ = f_new_inst(FI_VARIABLE); break;
        default: cf_error("%s: variable expected.", $1->name);
      }
 
@@ -776,53 +763,53 @@ symbol:
    }
 
 static_attr:
-   FROM    { $$ = f_new_inst(); $$->aux = T_IP;         $$->a2.i = SA_FROM;	$$->a1.i = 1; }
- | GW      { $$ = f_new_inst(); $$->aux = T_IP;         $$->a2.i = SA_GW;	$$->a1.i = 1; }
- | NET     { $$ = f_new_inst(); $$->aux = T_PREFIX;     $$->a2.i = SA_NET; }
- | PROTO   { $$ = f_new_inst(); $$->aux = T_STRING;     $$->a2.i = SA_PROTO; }
- | SOURCE  { $$ = f_new_inst(); $$->aux = T_ENUM_RTS;   $$->a2.i = SA_SOURCE; }
- | SCOPE   { $$ = f_new_inst(); $$->aux = T_ENUM_SCOPE; $$->a2.i = SA_SCOPE;	$$->a1.i = 1; }
- | CAST    { $$ = f_new_inst(); $$->aux = T_ENUM_RTC;   $$->a2.i = SA_CAST; }
- | DEST    { $$ = f_new_inst(); $$->aux = T_ENUM_RTD;   $$->a2.i = SA_DEST;	$$->a1.i = 1; }
- | IFNAME  { $$ = f_new_inst(); $$->aux = T_STRING;     $$->a2.i = SA_IFNAME; }
- | IFINDEX { $$ = f_new_inst(); $$->aux = T_INT;        $$->a2.i = SA_IFINDEX; }
+   FROM    { $$ = f_new_static_attr(T_IP,         SA_FROM,	1); }
+ | GW      { $$ = f_new_static_attr(T_IP,         SA_GW,	1); }
+ | NET     { $$ = f_new_static_attr(T_PREFIX,     SA_NET,	0); }
+ | PROTO   { $$ = f_new_static_attr(T_STRING,     SA_PROTO,	0); }
+ | SOURCE  { $$ = f_new_static_attr(T_ENUM_RTS,   SA_SOURCE,	0); }
+ | SCOPE   { $$ = f_new_static_attr(T_ENUM_SCOPE, SA_SCOPE,	1); }
+ | CAST    { $$ = f_new_static_attr(T_ENUM_RTC,   SA_CAST,	0); }
+ | DEST    { $$ = f_new_static_attr(T_ENUM_RTD,   SA_DEST,	1); }
+ | IFNAME  { $$ = f_new_static_attr(T_STRING,     SA_IFNAME,	0); }
+ | IFINDEX { $$ = f_new_static_attr(T_INT,        SA_IFINDEX,	0); }
  ;
 
 term:
-   '(' term ')'      { $$ = $2; }
- | term '+' term     { $$ = f_new_inst(); $$->code = '+';        $$->a1.p = $1; $$->a2.p = $3; }
- | term '-' term     { $$ = f_new_inst(); $$->code = '-';        $$->a1.p = $1; $$->a2.p = $3; }
- | term '*' term     { $$ = f_new_inst(); $$->code = '*';        $$->a1.p = $1; $$->a2.p = $3; }
- | term '/' term     { $$ = f_new_inst(); $$->code = '/';        $$->a1.p = $1; $$->a2.p = $3; }
- | term AND term     { $$ = f_new_inst(); $$->code = '&';        $$->a1.p = $1; $$->a2.p = $3; }
- | term OR  term     { $$ = f_new_inst(); $$->code = '|';        $$->a1.p = $1; $$->a2.p = $3; }
- | term '=' term     { $$ = f_new_inst(); $$->code = P('=','='); $$->a1.p = $1; $$->a2.p = $3; }
- | term NEQ term { $$ = f_new_inst(); $$->code = P('!','=');     $$->a1.p = $1; $$->a2.p = $3; }
- | term '<' term     { $$ = f_new_inst(); $$->code = '<';        $$->a1.p = $1; $$->a2.p = $3; }
- | term LEQ term { $$ = f_new_inst(); $$->code = P('<','=');     $$->a1.p = $1; $$->a2.p = $3; }
- | term '>' term     { $$ = f_new_inst(); $$->code = '<';        $$->a1.p = $3; $$->a2.p = $1; }
- | term GEQ term { $$ = f_new_inst(); $$->code = P('<','=');     $$->a1.p = $3; $$->a2.p = $1; }
- | term '~' term     { $$ = f_new_inst(); $$->code = '~';        $$->a1.p = $1; $$->a2.p = $3; }
- | term NMA term { $$ = f_new_inst(); $$->code = P('!','~');     $$->a1.p = $1; $$->a2.p = $3; }
- | '!' term { $$ = f_new_inst(); $$->code = '!'; $$->a1.p = $2; }
- | DEFINED '(' term ')' { $$ = f_new_inst(); $$->code = P('d','e');  $$->a1.p = $3; }
+   '(' term ')'		{ $$ = $2; }
+ | term '+' term	{ $$ = f_new_inst(FI_ADD);	$$->a1.p = $1; $$->a2.p = $3; }
+ | term '-' term	{ $$ = f_new_inst(FI_SUBTRACT);	$$->a1.p = $1; $$->a2.p = $3; }
+ | term '*' term	{ $$ = f_new_inst(FI_MULTIPLY);	$$->a1.p = $1; $$->a2.p = $3; }
+ | term '/' term	{ $$ = f_new_inst(FI_DIVIDE);	$$->a1.p = $1; $$->a2.p = $3; }
+ | term AND term	{ $$ = f_new_inst(FI_AND);	$$->a1.p = $1; $$->a2.p = $3; }
+ | term OR  term	{ $$ = f_new_inst(FI_OR);	$$->a1.p = $1; $$->a2.p = $3; }
+ | term '=' term	{ $$ = f_new_inst(FI_EQ);	$$->a1.p = $1; $$->a2.p = $3; }
+ | term NEQ term	{ $$ = f_new_inst(FI_NEQ);	$$->a1.p = $1; $$->a2.p = $3; }
+ | term '<' term	{ $$ = f_new_inst(FI_LT);	$$->a1.p = $1; $$->a2.p = $3; }
+ | term LEQ term	{ $$ = f_new_inst(FI_LTE);	$$->a1.p = $1; $$->a2.p = $3; }
+ | term '>' term	{ $$ = f_new_inst(FI_LT);	$$->a1.p = $3; $$->a2.p = $1; }
+ | term GEQ term	{ $$ = f_new_inst(FI_LTE);	$$->a1.p = $3; $$->a2.p = $1; }
+ | term '~' term	{ $$ = f_new_inst(FI_MATCH);	$$->a1.p = $1; $$->a2.p = $3; }
+ | term NMA term	{ $$ = f_new_inst(FI_NOT_MATCH);$$->a1.p = $1; $$->a2.p = $3; }
+ | '!' term		{ $$ = f_new_inst(FI_NOT);	$$->a1.p = $2; }
+ | DEFINED '(' term ')' { $$ = f_new_inst(FI_DEFINED);	$$->a1.p = $3; }
 
  | symbol   { $$ = $1; }
  | constant { $$ = $1; }
  | constructor { $$ = $1; }
 
- | PREFERENCE { $$ = f_new_inst(); $$->code = 'P'; }
+ | PREFERENCE { $$ = f_new_inst(FI_PREF_GET); }
 
- | rtadot static_attr { $$ = $2; $$->code = 'a'; }
+ | rtadot static_attr { $$ = f_new_inst_sa(FI_RTA_GET, $2); }
 
- | rtadot dynamic_attr { $$ = $2; $$->code = P('e','a'); }
+ | rtadot dynamic_attr { $$ = f_new_inst_da(FI_EA_GET, $2); }
 
- | term '.' IP { $$ = f_new_inst(); $$->code = P('c','p'); $$->a1.p = $1; $$->aux = T_IP; }
- | term '.' LEN { $$ = f_new_inst(); $$->code = 'L'; $$->a1.p = $1; }
- | term '.' MASK '(' term ')' { $$ = f_new_inst(); $$->code = P('i','M'); $$->a1.p = $1; $$->a2.p = $5; }
- | term '.' FIRST { $$ = f_new_inst(); $$->code = P('a','f'); $$->a1.p = $1; }
- | term '.' LAST  { $$ = f_new_inst(); $$->code = P('a','l'); $$->a1.p = $1; }
- | term '.' LAST_NONAGGREGATED  { $$ = f_new_inst(); $$->code = P('a','L'); $$->a1.p = $1; }
+ | term '.' IP { $$ = f_new_inst(FI_IP); $$->a1.p = $1; $$->aux = T_IP; }
+ | term '.' LEN { $$ = f_new_inst(FI_LENGTH); $$->a1.p = $1; }
+ | term '.' MASK '(' term ')' { $$ = f_new_inst(FI_IP_MASK); $$->a1.p = $1; $$->a2.p = $5; }
+ | term '.' FIRST { $$ = f_new_inst(FI_AS_PATH_FIRST); $$->a1.p = $1; }
+ | term '.' LAST  { $$ = f_new_inst(FI_AS_PATH_LAST); $$->a1.p = $1; }
+ | term '.' LAST_NONAGGREGATED  { $$ = f_new_inst(FI_AS_PATH_LAST_NAG); $$->a1.p = $1; }
 
 /* Communities */
 /* This causes one shift/reduce conflict
@@ -832,20 +819,18 @@ term:
  | rtadot dynamic_attr '.' RESET{ }
 */
 
- | '+' EMPTY '+' { $$ = f_new_inst(); $$->code = 'E'; $$->aux = T_PATH; }
- | '-' EMPTY '-' { $$ = f_new_inst(); $$->code = 'E'; $$->aux = T_CLIST; }
- | '-' '-' EMPTY '-' '-' { $$ = f_new_inst(); $$->code = 'E'; $$->aux = T_ECLIST; }
- | '-' '-' '-' EMPTY '-' '-' '-' { $$ = f_new_inst(); $$->code = 'E'; $$->aux = T_LCLIST; }
- | PREPEND '(' term ',' term ')' { $$ = f_new_inst(); $$->code = P('A','p'); $$->a1.p = $3; $$->a2.p = $5; }
- | ADD '(' term ',' term ')' { $$ = f_new_inst(); $$->code = P('C','a'); $$->a1.p = $3; $$->a2.p = $5; $$->aux = 'a'; }
- | DELETE '(' term ',' term ')' { $$ = f_new_inst(); $$->code = P('C','a'); $$->a1.p = $3; $$->a2.p = $5; $$->aux = 'd'; }
- | FILTER '(' term ',' term ')' { $$ = f_new_inst(); $$->code = P('C','a'); $$->a1.p = $3; $$->a2.p = $5; $$->aux = 'f'; }
+ | '+' EMPTY '+' { $$ = f_new_inst(FI_EMPTY); $$->aux = T_PATH; }
+ | '-' EMPTY '-' { $$ = f_new_inst(FI_EMPTY); $$->aux = T_CLIST; }
+ | '-' '-' EMPTY '-' '-' { $$ = f_new_inst(FI_EMPTY); $$->aux = T_ECLIST; }
+ | '-' '-' '-' EMPTY '-' '-' '-' { $$ = f_new_inst(FI_EMPTY); $$->aux = T_LCLIST; }
+ | PREPEND '(' term ',' term ')' { $$ = f_new_inst(FI_PATH_PREPEND); $$->a1.p = $3; $$->a2.p = $5; }
+ | ADD '(' term ',' term ')' { $$ = f_new_inst(FI_CLIST_ADD_DEL); $$->a1.p = $3; $$->a2.p = $5; $$->aux = 'a'; }
+ | DELETE '(' term ',' term ')' { $$ = f_new_inst(FI_CLIST_ADD_DEL); $$->a1.p = $3; $$->a2.p = $5; $$->aux = 'd'; }
+ | FILTER '(' term ',' term ')' { $$ = f_new_inst(FI_CLIST_ADD_DEL); $$->a1.p = $3; $$->a2.p = $5; $$->aux = 'f'; }
 
  | ROA_CHECK '(' SYM ')' { $$ = f_generate_roa_check($3, NULL, NULL); }
  | ROA_CHECK '(' SYM ',' term ',' term ')' { $$ = f_generate_roa_check($3, $5, $7); }
 
-/* | term '.' LEN { $$->code = P('P','l'); } */
-
 /* function_call is inlined here */
  | SYM '(' var_list ')' {
      struct symbol *sym;
@@ -853,8 +838,7 @@ term:
      if ($1->class != SYM_FUNCTION)
        cf_error("You can't call something which is not a function. Really.");
      DBG("You are calling function %s\n", $1->name);
-     $$ = f_new_inst();
-     $$->code = P('c','a');
+     $$ = f_new_inst(FI_CALL);
      $$->a1.p = inst;
      $$->a2.p = $1->def;
      sym = $1->aux2;
@@ -879,7 +863,7 @@ break_command:
  ;
 
 print_one:
-   term { $$ = f_new_inst(); $$->code = 'p'; $$->a1.p = $1; $$->a2.p = NULL; }
+   term { $$ = f_new_inst(FI_PRINT); $$->a1.p = $1; $$->a2.p = NULL; }
  ;
 
 print_list: /* EMPTY */ { $$ = NULL; }
@@ -893,15 +877,13 @@ print_list: /* EMPTY */ { $$ = NULL; }
  ;
 
 var_listn: term {
-     $$ = f_new_inst();
-     $$->code = 's';
+     $$ = f_new_inst(FI_SET);
      $$->a1.p = NULL;
      $$->a2.p = $1;
      $$->next = NULL;
    }
  | term ',' var_listn {
-     $$ = f_new_inst();
-     $$->code = 's';
+     $$ = f_new_inst(FI_SET);
      $$->a1.p = NULL;
      $$->a2.p = $1;
      $$->next = $3;
@@ -914,74 +896,64 @@ var_list: /* EMPTY */ { $$ = NULL; }
 
 cmd:
    IF term THEN block {
-     $$ = f_new_inst();
-     $$->code = '?';
+     $$ = f_new_inst(FI_CONDITION);
      $$->a1.p = $2;
      $$->a2.p = $4;
    }
  | IF term THEN block ELSE block {
-     struct f_inst *i = f_new_inst();
-     i->code = '?';
+     struct f_inst *i = f_new_inst(FI_CONDITION);
      i->a1.p = $2;
      i->a2.p = $4;
-     $$ = f_new_inst();
-     $$->code = '?';
+     $$ = f_new_inst(FI_CONDITION);
      $$->a1.p = i;
      $$->a2.p = $6;
    }
  | SYM '=' term ';' {
-     $$ = f_new_inst();
      DBG( "Ook, we'll set value\n" );
      if (($1->class & ~T_MASK) != SYM_VARIABLE)
        cf_error( "You may set only variables." );
-     $$->code = 's';
+     $$ = f_new_inst(FI_SET);
      $$->a1.p = $1;
      $$->a2.p = $3;
    }
  | RETURN term ';' {
-     $$ = f_new_inst();
      DBG( "Ook, we'll return the value\n" );
-     $$->code = 'r';
+     $$ = f_new_inst(FI_RETURN);
      $$->a1.p = $2;
    }
  | rtadot dynamic_attr '=' term ';' {
-     $$ = $2;
-     $$->code = P('e','S');
+     $$ = f_new_inst_da(FI_EA_SET, $2);
      $$->a1.p = $4;
    }
  | rtadot static_attr '=' term ';' {
-     $$ = $2;
+     $$ = f_new_inst_sa(FI_RTA_SET, $2);
      if (!$$->a1.i)
        cf_error( "This static attribute is read-only.");
-     $$->code = P('a','S');
      $$->a1.p = $4;
    }
  | PREFERENCE '=' term ';' {
-     $$ = f_new_inst();
-     $$->code = P('P','S');
+     $$ = f_new_inst(FI_PREF_SET);
      $$->a1.p = $3;
    }
  | UNSET '(' rtadot dynamic_attr ')' ';' {
-     $$ = $4;
+     $$ = f_new_inst_da(FI_EA_SET, $4);
      $$->aux = EAF_TYPE_UNDEF | EAF_TEMP;
-     $$->code = P('e','S');
      $$->a1.p = NULL;
    }
- | break_command print_list ';' { $$ = f_new_inst(); $$->code = P('p',','); $$->a1.p = $2; $$->a2.i = $1; }
+ | break_command print_list ';' { $$ = f_new_inst(FI_PRINT_AND_DIE); $$->a1.p = $2; $$->a2.i = $1; }
  | function_call ';' { $$ = $1; }
  | CASE term '{' switch_body '}' {
-      $$ = f_new_inst();
-      $$->code = P('S','W');
+      $$ = f_new_inst(FI_SWITCH);
       $$->a1.p = $2;
       $$->a2.p = build_tree( $4 );
    }
 
 
  | rtadot dynamic_attr '.' EMPTY ';' { $$ = f_generate_empty($2); }
- | rtadot dynamic_attr '.' PREPEND '(' term ')' ';'   { $$ = f_generate_complex( P('A','p'), 'x', $2, $6 ); }
- | rtadot dynamic_attr '.' ADD '(' term ')' ';'       { $$ = f_generate_complex( P('C','a'), 'a', $2, $6 ); }
- | rtadot dynamic_attr '.' DELETE '(' term ')' ';'    { $$ = f_generate_complex( P('C','a'), 'd', $2, $6 ); }
- | rtadot dynamic_attr '.' FILTER '(' term ')' ';'    { $$ = f_generate_complex( P('C','a'), 'f', $2, $6 ); }
+ | rtadot dynamic_attr '.' PREPEND '(' term ')' ';'   { $$ = f_generate_complex( FI_PATH_PREPEND, 'x', $2, $6 ); }
+ | rtadot dynamic_attr '.' ADD '(' term ')' ';'       { $$ = f_generate_complex( FI_CLIST_ADD_DEL, 'a', $2, $6 ); }
+ | rtadot dynamic_attr '.' DELETE '(' term ')' ';'    { $$ = f_generate_complex( FI_CLIST_ADD_DEL, 'd', $2, $6 ); }
+ | rtadot dynamic_attr '.' FILTER '(' term ')' ';'    { $$ = f_generate_complex( FI_CLIST_ADD_DEL, 'f', $2, $6 ); }
  ;
 
 CF_END
diff --git a/filter/f-util.c b/filter/f-util.c
index def2b248..42b08868 100644
--- a/filter/f-util.c
+++ b/filter/f-util.c
@@ -2,6 +2,7 @@
  *	Filters: utility functions
  *
  *	Copyright 1998 Pavel Machek <pavel@ucw.cz>
+ *		  2017 Jan Maria Matejka <mq@ucw.cz>
  *
  *	Can be freely distributed and used under the terms of the GNU GPL.
  */
@@ -13,43 +14,48 @@
 #define P(a,b) ((a<<8) | b)
 
 struct f_inst *
-f_new_inst(void)
+f_new_inst(enum f_instruction_code fi_code)
 {
   struct f_inst * ret;
-  ret = cfg_alloc(sizeof(struct f_inst));
-  ret->code = ret->aux = 0;
-  ret->arg1 = ret->arg2 = ret->next = NULL;
+  ret = cfg_allocz(sizeof(struct f_inst));
+  ret->fi_code = fi_code;
   ret->lineno = ifs->lino;
   return ret;
 }
 
 struct f_inst *
-f_new_dynamic_attr(int type, int f_type UNUSED, int code)
+f_new_inst_da(enum f_instruction_code fi_code, struct f_dynamic_attr da)
 {
-  /* FIXME: Remove the f_type parameter? */
-  struct f_inst *f = f_new_inst();
-  f->aux = type;
-  f->a2.i = code;
-  return f;
+  struct f_inst *ret = f_new_inst(fi_code);
+  ret->aux = da.type;
+  ret->a2.i = da.ea_code;
+  return ret;
+}
+
+struct f_inst *
+f_new_inst_sa(enum f_instruction_code fi_code, struct f_static_attr sa)
+{
+  struct f_inst *ret = f_new_inst(fi_code);
+  ret->aux = sa.f_type;
+  ret->a2.i = sa.sa_code;
+  ret->a1.i = sa.readonly;
+  return ret;
 }
 
 /*
  * Generate set_dynamic( operation( get_dynamic(), argument ) )
  */
 struct f_inst *
-f_generate_complex(int operation, int operation_aux, struct f_inst *dyn, struct f_inst *argument)
+f_generate_complex(int operation, int operation_aux, struct f_dynamic_attr da, struct f_inst *argument)
 {
-  struct f_inst *set_dyn = f_new_inst(),
-                *oper = f_new_inst(),
-                *get_dyn = dyn;
+  struct f_inst *set_dyn = f_new_inst_da(FI_EA_SET, da),
+                *oper = f_new_inst(operation),
+                *get_dyn = f_new_inst_da(FI_EA_GET, da);
 
-  *set_dyn = *get_dyn;
-  get_dyn->code = P('e','a');
-  oper->code = operation;
   oper->aux = operation_aux;
   oper->a1.p = get_dyn;
   oper->a2.p = argument;
-  set_dyn->code = P('e','S');
+
   set_dyn->a1.p = oper;
   return set_dyn;
 }
@@ -59,7 +65,7 @@ struct f_inst *
 f_generate_roa_check(struct symbol *sym, struct f_inst *prefix, struct f_inst *asn)
 {
   struct f_inst_roa_check *ret = cfg_allocz(sizeof(struct f_inst_roa_check));
-  ret->i.code = P('R','C');
+  ret->i.fi_code = FI_ROA_CHECK;
   ret->i.lineno = ifs->lino;
   ret->i.arg1 = prefix;
   ret->i.arg2 = asn;
diff --git a/filter/filter.c b/filter/filter.c
index b0c56046..db90941f 100644
--- a/filter/filter.c
+++ b/filter/filter.c
@@ -46,8 +46,6 @@
 #include "conf/conf.h"
 #include "filter/filter.h"
 
-#define P(a,b) ((a<<8) | b)
-
 #define CMP_ERROR 999
 
 static struct adata *
@@ -644,13 +642,13 @@ interpret(struct f_inst *what)
   if (!what)
     return res;
 
-  switch(what->code) {
-  case ',':
+  switch(what->fi_code) {
+  case FI_COMMA:
     TWOARGS;
     break;
 
 /* Binary operators */
-  case '+':
+  case FI_ADD:
     TWOARGS_C;
     switch (res.type = v1.type) {
     case T_VOID: runtime( "Can't operate with values of type void" );
@@ -658,7 +656,7 @@ interpret(struct f_inst *what)
     default: runtime( "Usage of unknown type" );
     }
     break;
-  case '-':
+  case FI_SUBTRACT:
     TWOARGS_C;
     switch (res.type = v1.type) {
     case T_VOID: runtime( "Can't operate with values of type void" );
@@ -666,7 +664,7 @@ interpret(struct f_inst *what)
     default: runtime( "Usage of unknown type" );
     }
     break;
-  case '*':
+  case FI_MULTIPLY:
     TWOARGS_C;
     switch (res.type = v1.type) {
     case T_VOID: runtime( "Can't operate with values of type void" );
@@ -674,7 +672,7 @@ interpret(struct f_inst *what)
     default: runtime( "Usage of unknown type" );
     }
     break;
-  case '/':
+  case FI_DIVIDE:
     TWOARGS_C;
     switch (res.type = v1.type) {
     case T_VOID: runtime( "Can't operate with values of type void" );
@@ -684,12 +682,12 @@ interpret(struct f_inst *what)
     }
     break;
 
-  case '&':
-  case '|':
+  case FI_AND:
+  case FI_OR:
     ARG(v1, a1.p);
     if (v1.type != T_BOOL)
       runtime( "Can't do boolean operation on non-booleans" );
-    if (v1.val.i == (what->code == '|')) {
+    if (v1.val.i == (what->fi_code == FI_OR)) {
       res.type = T_BOOL;
       res.val.i = v1.val.i;
       break;
@@ -702,7 +700,7 @@ interpret(struct f_inst *what)
     res.val.i = v2.val.i;
     break;
 
-  case P('m','p'):
+  case FI_PAIR_CONSTRUCT:
     TWOARGS;
     if ((v1.type != T_INT) || (v2.type != T_INT))
       runtime( "Can't operate with value of non-integer type in pair constructor" );
@@ -714,7 +712,7 @@ interpret(struct f_inst *what)
     res.type = T_PAIR;
     break;
 
-  case P('m','c'):
+  case FI_EC_CONSTRUCT:
     {
       TWOARGS;
 
@@ -762,7 +760,7 @@ interpret(struct f_inst *what)
       break;
     }
 
-  case P('m','l'):
+  case FI_LC_CONSTRUCT:
     {
       TWOARGS;
 
@@ -798,12 +796,12 @@ interpret(struct f_inst *what)
     res.val.i = (x); \
     break;
 
-  case P('!','='): SAME(!i);
-  case P('=','='): SAME(i);
-  case '<': COMPARE(i==-1);
-  case P('<','='): COMPARE(i!=1);
+  case FI_NEQ: SAME(!i);
+  case FI_EQ: SAME(i);
+  case FI_LT: COMPARE(i==-1);
+  case FI_LTE: COMPARE(i!=1);
 
-  case '!':
+  case FI_NOT:
     ONEARG;
     if (v1.type != T_BOOL)
       runtime( "Not applied to non-boolean" );
@@ -811,7 +809,7 @@ interpret(struct f_inst *what)
     res.val.i = !res.val.i;
     break;
 
-  case '~':
+  case FI_MATCH:
     TWOARGS;
     res.type = T_BOOL;
     res.val.i = val_in_range(v1, v2);
@@ -820,7 +818,7 @@ interpret(struct f_inst *what)
     res.val.i = !!res.val.i;
     break;
 
-  case P('!','~'):
+  case FI_NOT_MATCH:
     TWOARGS;
     res.type = T_BOOL;
     res.val.i = val_in_range(v1, v2);
@@ -829,14 +827,14 @@ interpret(struct f_inst *what)
     res.val.i = !res.val.i;
     break;
 
-  case P('d','e'):
+  case FI_DEFINED:
     ONEARG;
     res.type = T_BOOL;
     res.val.i = (v1.type != T_VOID);
     break;
 
   /* Set to indirect value, a1 = variable, a2 = value */
-  case 's':
+  case FI_SET:
     ARG(v2, a2.p);
     sym = what->a1.p;
     vp = sym->def;
@@ -855,7 +853,7 @@ interpret(struct f_inst *what)
     break;
 
     /* some constants have value in a2, some in *a1.p, strange. */
-  case 'c':	/* integer (or simple type) constant, string, set, or prefix_set */
+  case FI_CONSTANT:	/* integer (or simple type) constant, string, set, or prefix_set */
     res.type = what->aux;
 
     if (res.type == T_PREFIX_SET)
@@ -867,15 +865,15 @@ interpret(struct f_inst *what)
     else
       res.val.i = what->a2.i;
     break;
-  case 'V':
-  case 'C':
+  case FI_VARIABLE:
+  case FI_CONSTANT_INDIRECT:
     res = * ((struct f_val *) what->a1.p);
     break;
-  case 'p':
+  case FI_PRINT:
     ONEARG;
     val_format(v1, &f_buf);
     break;
-  case '?':	/* ? has really strange error value, so we can implement if ... else nicely :-) */
+  case FI_CONDITION:	/* ? has really strange error value, so we can implement if ... else nicely :-) */
     ONEARG;
     if (v1.type != T_BOOL)
       runtime( "If requires boolean expression" );
@@ -885,10 +883,10 @@ interpret(struct f_inst *what)
     } else res.val.i = 1;
     res.type = T_BOOL;
     break;
-  case '0':
+  case FI_NOP:
     debug( "No operation\n" );
     break;
-  case P('p',','):
+  case FI_PRINT_AND_DIE:
     ONEARG;
     if ((what->a2.i == F_NOP || (what->a2.i != F_NONL && what->a1.p)) &&
 	!(f_flags & FF_SILENT))
@@ -911,7 +909,7 @@ interpret(struct f_inst *what)
       bug( "unknown return type: Can't happen");
     }
     break;
-  case 'a':	/* rta access */
+  case FI_RTA_GET:	/* rta access */
     {
       ACCESS_RTE;
       struct rta *rta = (*f_rte)->attrs;
@@ -936,7 +934,7 @@ interpret(struct f_inst *what)
       }
     }
     break;
-  case P('a','S'):
+  case FI_RTA_SET:
     ACCESS_RTE;
     ONEARG;
     if (what->aux != v1.type)
@@ -988,7 +986,7 @@ interpret(struct f_inst *what)
       }
     }
     break;
-  case P('e','a'):	/* Access to extended attributes */
+  case FI_EA_GET:	/* Access to extended attributes */
     ACCESS_RTE;
     {
       eattr *e = NULL;
@@ -1074,7 +1072,7 @@ interpret(struct f_inst *what)
       }
     }
     break;
-  case P('e','S'):
+  case FI_EA_SET:
     ACCESS_RTE;
     ONEARG;
     {
@@ -1180,12 +1178,12 @@ interpret(struct f_inst *what)
       }
     }
     break;
-  case 'P':
+  case FI_PREF_GET:
     ACCESS_RTE;
     res.type = T_INT;
     res.val.i = (*f_rte)->pref;
     break;
-  case P('P','S'):
+  case FI_PREF_SET:
     ACCESS_RTE;
     ONEARG;
     if (v1.type != T_INT)
@@ -1195,7 +1193,7 @@ interpret(struct f_inst *what)
     f_rte_cow();
     (*f_rte)->pref = v1.val.i;
     break;
-  case 'L':	/* Get length of */
+  case FI_LENGTH:	/* Get length of */
     ONEARG;
     res.type = T_INT;
     switch(v1.type) {
@@ -1207,7 +1205,7 @@ interpret(struct f_inst *what)
     default: runtime( "Prefix, path, clist or eclist expected" );
     }
     break;
-  case P('c','p'):	/* Convert prefix to ... */
+  case FI_IP:	/* Convert prefix to ... */
     ONEARG;
     if (v1.type != T_PREFIX)
       runtime( "Prefix expected" );
@@ -1218,7 +1216,7 @@ interpret(struct f_inst *what)
     default: bug( "Unknown prefix to conversion" );
     }
     break;
-  case P('a','f'):	/* Get first ASN from AS PATH */
+  case FI_AS_PATH_FIRST:	/* Get first ASN from AS PATH */
     ONEARG;
     if (v1.type != T_PATH)
       runtime( "AS path expected" );
@@ -1228,7 +1226,7 @@ interpret(struct f_inst *what)
     res.type = T_INT;
     res.val.i = as;
     break;
-  case P('a','l'):	/* Get last ASN from AS PATH */
+  case FI_AS_PATH_LAST:	/* Get last ASN from AS PATH */
     ONEARG;
     if (v1.type != T_PATH)
       runtime( "AS path expected" );
@@ -1238,7 +1236,7 @@ interpret(struct f_inst *what)
     res.type = T_INT;
     res.val.i = as;
     break;
-  case P('a','L'):	/* Get last ASN from non-aggregated part of AS PATH */
+  case FI_AS_PATH_LAST_NAG:	/* Get last ASN from non-aggregated part of AS PATH */
     ONEARG;
     if (v1.type != T_PATH)
       runtime( "AS path expected" );
@@ -1246,23 +1244,23 @@ interpret(struct f_inst *what)
     res.type = T_INT;
     res.val.i = as_path_get_last_nonaggregated(v1.val.ad);
     break;
-  case 'r':
+  case FI_RETURN:
     ONEARG;
     res = v1;
     res.type |= T_RETURN;
     return res;
-  case P('c','a'): /* CALL: this is special: if T_RETURN and returning some value, mask it out  */
+  case FI_CALL: /* CALL: this is special: if T_RETURN and returning some value, mask it out  */
     ONEARG;
     res = interpret(what->a2.p);
     if (res.type == T_RETURN)
       return res;
     res.type &= ~T_RETURN;
     break;
-  case P('c','v'):	/* Clear local variables */
+  case FI_CLEAR_LOCAL_VARS:	/* Clear local variables */
     for (sym = what->a1.p; sym != NULL; sym = sym->aux2)
       ((struct f_val *) sym->def)->type = T_VOID;
     break;
-  case P('S','W'):
+  case FI_SWITCH:
     ONEARG;
     {
       struct f_tree *t = find_tree(what->a2.p, v1);
@@ -1281,7 +1279,7 @@ interpret(struct f_inst *what)
 	return res;
     }
     break;
-  case P('i','M'): /* IP.MASK(val) */
+  case FI_IP_MASK: /* IP.MASK(val) */
     TWOARGS;
     if (v2.type != T_INT)
       runtime( "Integer expected");
@@ -1294,11 +1292,11 @@ interpret(struct f_inst *what)
     }
     break;
 
-  case 'E':	/* Create empty attribute */
+  case FI_EMPTY:	/* Create empty attribute */
     res.type = what->aux;
     res.val.ad = adata_empty(f_pool, 0);
     break;
-  case P('A','p'):	/* Path prepend */
+  case FI_PATH_PREPEND:	/* Path prepend */
     TWOARGS;
     if (v1.type != T_PATH)
       runtime("Can't prepend to non-path");
@@ -1309,7 +1307,7 @@ interpret(struct f_inst *what)
     res.val.ad = as_path_prepend(f_pool, v1.val.ad, v2.val.i);
     break;
 
-  case P('C','a'):	/* (Extended) Community list add or delete */
+  case FI_CLIST_ADD_DEL:	/* (Extended) Community list add or delete */
     TWOARGS;
     if (v1.type == T_PATH)
     {
@@ -1477,7 +1475,7 @@ interpret(struct f_inst *what)
 
     break;
 
-  case P('R','C'):	/* ROA Check */
+  case FI_ROA_CHECK:	/* ROA Check */
     if (what->arg1)
     {
       TWOARGS;
@@ -1511,7 +1509,7 @@ interpret(struct f_inst *what)
     break;
 
   default:
-    bug( "Unknown instruction %d (%c)", what->code, what->code & 0xff);
+    bug( "Unknown instruction %d (%c)", what->fi_code, what->fi_code & 0xff);
   }
   if (what->next)
     return interpret(what->next);
@@ -1541,38 +1539,38 @@ i_same(struct f_inst *f1, struct f_inst *f2)
     return 1;
   if (f1->aux != f2->aux)
     return 0;
-  if (f1->code != f2->code)
+  if (f1->fi_code != f2->fi_code)
     return 0;
   if (f1 == f2)		/* It looks strange, but it is possible with call rewriting trickery */
     return 1;
 
-  switch(f1->code) {
-  case ',': /* fall through */
-  case '+':
-  case '-':
-  case '*':
-  case '/':
-  case '|':
-  case '&':
-  case P('m','p'):
-  case P('m','c'):
-  case P('!','='):
-  case P('=','='):
-  case '<':
-  case P('<','='): TWOARGS; break;
+  switch(f1->fi_code) {
+  case FI_COMMA: /* fall through */
+  case FI_ADD:
+  case FI_SUBTRACT:
+  case FI_MULTIPLY:
+  case FI_DIVIDE:
+  case FI_OR:
+  case FI_AND:
+  case FI_PAIR_CONSTRUCT:
+  case FI_EC_CONSTRUCT:
+  case FI_NEQ:
+  case FI_EQ:
+  case FI_LT:
+  case FI_LTE: TWOARGS; break;
 
-  case '!': ONEARG; break;
-  case P('!', '~'):
-  case '~': TWOARGS; break;
-  case P('d','e'): ONEARG; break;
+  case FI_NOT: ONEARG; break;
+  case FI_NOT_MATCH:
+  case FI_MATCH: TWOARGS; break;
+  case FI_DEFINED: ONEARG; break;
 
-  case P('m','l'):
+  case FI_LC_CONSTRUCT:
     TWOARGS;
     if (!i_same(INST3(f1).p, INST3(f2).p))
       return 0;
     break;
 
-  case 's':
+  case FI_SET:
     ARG(v2, a2.p);
     {
       struct symbol *s1, *s2;
@@ -1585,7 +1583,7 @@ i_same(struct f_inst *f1, struct f_inst *f2)
     }
     break;
 
-  case 'c':
+  case FI_CONSTANT:
     switch (f1->aux) {
 
     case T_PREFIX_SET:
@@ -1608,43 +1606,43 @@ i_same(struct f_inst *f1, struct f_inst *f2)
     }
     break;
 
-  case 'C':
+  case FI_CONSTANT_INDIRECT:
     if (!val_same(* (struct f_val *) f1->a1.p, * (struct f_val *) f2->a1.p))
       return 0;
     break;
 
-  case 'V':
+  case FI_VARIABLE:
     if (strcmp((char *) f1->a2.p, (char *) f2->a2.p))
       return 0;
     break;
-  case 'p': case 'L': ONEARG; break;
-  case '?': TWOARGS; break;
-  case '0': case 'E': break;
-  case P('p',','): ONEARG; A2_SAME; break;
-  case 'P':
-  case 'a': A2_SAME; break;
-  case P('e','a'): A2_SAME; break;
-  case P('P','S'):
-  case P('a','S'):
-  case P('e','S'): ONEARG; A2_SAME; break;
+  case FI_PRINT: case FI_LENGTH: ONEARG; break;
+  case FI_CONDITION: TWOARGS; break;
+  case FI_NOP: case FI_EMPTY: break;
+  case FI_PRINT_AND_DIE: ONEARG; A2_SAME; break;
+  case FI_PREF_GET:
+  case FI_RTA_GET: A2_SAME; break;
+  case FI_EA_GET: A2_SAME; break;
+  case FI_PREF_SET:
+  case FI_RTA_SET:
+  case FI_EA_SET: ONEARG; A2_SAME; break;
 
-  case 'r': ONEARG; break;
-  case P('c','p'): ONEARG; break;
-  case P('c','a'): /* Call rewriting trickery to avoid exponential behaviour */
+  case FI_RETURN: ONEARG; break;
+  case FI_IP: ONEARG; break;
+  case FI_CALL: /* Call rewriting trickery to avoid exponential behaviour */
              ONEARG;
 	     if (!i_same(f1->a2.p, f2->a2.p))
 	       return 0;
 	     f2->a2.p = f1->a2.p;
 	     break;
-  case P('c','v'): break; /* internal instruction */
-  case P('S','W'): ONEARG; if (!same_tree(f1->a2.p, f2->a2.p)) return 0; break;
-  case P('i','M'): TWOARGS; break;
-  case P('A','p'): TWOARGS; break;
-  case P('C','a'): TWOARGS; break;
-  case P('a','f'):
-  case P('a','l'):
-  case P('a','L'): ONEARG; break;
-  case P('R','C'):
+  case FI_CLEAR_LOCAL_VARS: break; /* internal instruction */
+  case FI_SWITCH: ONEARG; if (!same_tree(f1->a2.p, f2->a2.p)) return 0; break;
+  case FI_IP_MASK: TWOARGS; break;
+  case FI_PATH_PREPEND: TWOARGS; break;
+  case FI_CLIST_ADD_DEL: TWOARGS; break;
+  case FI_AS_PATH_FIRST:
+  case FI_AS_PATH_LAST:
+  case FI_AS_PATH_LAST_NAG: ONEARG; break;
+  case FI_ROA_CHECK:
     TWOARGS;
     /* Does not really make sense - ROA check resuls may change anyway */
     if (strcmp(((struct f_inst_roa_check *) f1)->rtc->name,
@@ -1652,7 +1650,7 @@ i_same(struct f_inst *f1, struct f_inst *f2)
       return 0;
     break;
   default:
-    bug( "Unknown instruction %d in same (%c)", f1->code, f1->code & 0xff);
+    bug( "Unknown instruction %d in same (%c)", f1->fi_code, f1->fi_code & 0xff);
   }
   return i_same(f1->next, f2->next);
 }
diff --git a/filter/filter.h b/filter/filter.h
index efb4b978..09f96084 100644
--- a/filter/filter.h
+++ b/filter/filter.h
@@ -14,9 +14,67 @@
 #include "nest/route.h"
 #include "nest/attrs.h"
 
+/* Filter instruction types */
+
+#define FI__TWOCHAR(a,b)	((a<<8) | b)
+#define FI__LIST \
+  F(FI_COMMA,			  0, ',') \
+  F(FI_ADD,			  0, '+') \
+  F(FI_SUBTRACT,		  0, '-') \
+  F(FI_MULTIPLY,		  0, '*') \
+  F(FI_DIVIDE,			  0, '/') \
+  F(FI_AND,			  0, '&') \
+  F(FI_OR,			  0, '|') \
+  F(FI_PAIR_CONSTRUCT,		'm', 'p') \
+  F(FI_EC_CONSTRUCT,		'm', 'c') \
+  F(FI_LC_CONSTRUCT,		'm', 'l') \
+  F(FI_NEQ,			'!', '=') \
+  F(FI_EQ,			'=', '=') \
+  F(FI_LT,			  0, '<') \
+  F(FI_LTE,			'<', '=') \
+  F(FI_NOT,			  0, '!') \
+  F(FI_MATCH,			  0, '~') \
+  F(FI_NOT_MATCH,		'!', '~') \
+  F(FI_DEFINED,			'd', 'e') \
+  F(FI_SET,			  0, 's') \
+  F(FI_CONSTANT,		  0, 'c') \
+  F(FI_VARIABLE,		  0, 'V') \
+  F(FI_CONSTANT_INDIRECT,	  0, 'C') \
+  F(FI_PRINT,			  0, 'p') \
+  F(FI_CONDITION,		  0, '?') \
+  F(FI_NOP,			  0, '0') \
+  F(FI_PRINT_AND_DIE,		'p', ',') \
+  F(FI_RTA_GET,			  0, 'a') \
+  F(FI_RTA_SET,			'a', 'S') \
+  F(FI_EA_GET,			'e', 'a') \
+  F(FI_EA_SET,			'e', 'S') \
+  F(FI_PREF_GET,		  0, 'P') \
+  F(FI_PREF_SET,		'P', 'S') \
+  F(FI_LENGTH,			  0, 'L') \
+  F(FI_IP,			'c', 'p') \
+  F(FI_AS_PATH_FIRST,		'a', 'f') \
+  F(FI_AS_PATH_LAST,		'a', 'l') \
+  F(FI_AS_PATH_LAST_NAG,	'a', 'L') \
+  F(FI_RETURN,			  0, 'r') \
+  F(FI_CALL,			'c', 'a') \
+  F(FI_CLEAR_LOCAL_VARS,	'c', 'V') \
+  F(FI_SWITCH,			'S', 'W') \
+  F(FI_IP_MASK,			'i', 'M') \
+  F(FI_EMPTY,			  0, 'E') \
+  F(FI_PATH_PREPEND,		'A', 'p') \
+  F(FI_CLIST_ADD_DEL,		'C', 'a') \
+  F(FI_ROA_CHECK,		'R', 'C')
+
+enum f_instruction_code {
+#define F(c,a,b) \
+  c = FI__TWOCHAR(a,b),
+FI__LIST
+#undef F
+} PACKED;
+
 struct f_inst {		/* Instruction */
   struct f_inst *next;	/* Structure is 16 bytes, anyway */
-  u16 code;
+  enum f_instruction_code fi_code;
   u16 aux;
   union {
     int i;
@@ -75,15 +133,32 @@ struct f_val {
   } val;
 };
 
+struct f_dynamic_attr {
+  int type;
+  int f_type;
+  int ea_code;
+};
+
+struct f_static_attr {
+  int f_type;
+  int sa_code;
+  int readonly;
+};
+
 struct filter {
   char *name;
   struct f_inst *root;
 };
 
-struct f_inst *f_new_inst(void);
-struct f_inst *f_new_dynamic_attr(int type, int f_type, int code);	/* Type as core knows it, type as filters know it, and code of dynamic attribute */
+struct f_inst *f_new_inst(enum f_instruction_code fi_code);
+struct f_inst *f_new_inst_da(enum f_instruction_code fi_code, struct f_dynamic_attr da);
+struct f_inst *f_new_inst_sa(enum f_instruction_code fi_code, struct f_static_attr sa);
+static inline struct f_dynamic_attr f_new_dynamic_attr(int type, int f_type, int code) /* Type as core knows it, type as filters know it, and code of dynamic attribute */
+{ return (struct f_dynamic_attr) { .type = type, .f_type = f_type, .ea_code = code }; }   /* f_type currently unused; will be handy for static type checking */
+static inline struct f_static_attr f_new_static_attr(int f_type, int code, int readonly)
+{ return (struct f_static_attr) { .f_type = f_type, .sa_code = code, .readonly = readonly }; }
 struct f_tree *f_new_tree(void);
-struct f_inst *f_generate_complex(int operation, int operation_aux, struct f_inst *dyn, struct f_inst *argument);
+struct f_inst *f_generate_complex(int operation, int operation_aux, struct f_dynamic_attr da, struct f_inst *argument);
 struct f_inst *f_generate_roa_check(struct symbol *sym, struct f_inst *prefix, struct f_inst *asn);
 
 

From 7c601e6b7b7696b24ce5f5715fa14dbb91c71d6e Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Wed, 29 Nov 2017 11:38:01 +0100
Subject: [PATCH 140/198] Filter: recursion to loop

It was supposed to do tail-recursion in interpret() but it didn't
compile as such. Converting it to loop makes a significant filter
performance improvement for flat filters.
---
 filter/filter.c | 10 +++-------
 1 file changed, 3 insertions(+), 7 deletions(-)

diff --git a/filter/filter.c b/filter/filter.c
index db90941f..85721dbd 100644
--- a/filter/filter.c
+++ b/filter/filter.c
@@ -633,15 +633,13 @@ static struct f_val
 interpret(struct f_inst *what)
 {
   struct symbol *sym;
-  struct f_val v1, v2, res, *vp;
+  struct f_val v1, v2, res = { .type = T_VOID }, *vp;
   unsigned u1, u2;
   int i;
   u32 as;
 
+  for ( ; what; what = what->next) {
   res.type = T_VOID;
-  if (!what)
-    return res;
-
   switch(what->fi_code) {
   case FI_COMMA:
     TWOARGS;
@@ -1510,9 +1508,7 @@ interpret(struct f_inst *what)
 
   default:
     bug( "Unknown instruction %d (%c)", what->fi_code, what->fi_code & 0xff);
-  }
-  if (what->next)
-    return interpret(what->next);
+  }}
   return res;
 }
 

From 74bfd2f97c0a95b6fb73a67d9334e54a90695c58 Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Tue, 13 Mar 2018 12:08:37 +0100
Subject: [PATCH 141/198] Filters: Removed FI_COMMA, not used for 19 years.

This instruction was removed in the commit linked below
and never used ever again. Rest in peace.

commit 84c7e1943f0dbf896b1dd8d02a21120aa00463f4
Author: Pavel Machek <pavel@ucw.cz>
Date:   Tue Mar 2 19:49:28 1999 +0000
---
 filter/filter.c | 7 +------
 filter/filter.h | 1 -
 2 files changed, 1 insertion(+), 7 deletions(-)

diff --git a/filter/filter.c b/filter/filter.c
index 85721dbd..023f7e2f 100644
--- a/filter/filter.c
+++ b/filter/filter.c
@@ -641,10 +641,6 @@ interpret(struct f_inst *what)
   for ( ; what; what = what->next) {
   res.type = T_VOID;
   switch(what->fi_code) {
-  case FI_COMMA:
-    TWOARGS;
-    break;
-
 /* Binary operators */
   case FI_ADD:
     TWOARGS_C;
@@ -1541,8 +1537,7 @@ i_same(struct f_inst *f1, struct f_inst *f2)
     return 1;
 
   switch(f1->fi_code) {
-  case FI_COMMA: /* fall through */
-  case FI_ADD:
+  case FI_ADD: /* fall through */
   case FI_SUBTRACT:
   case FI_MULTIPLY:
   case FI_DIVIDE:
diff --git a/filter/filter.h b/filter/filter.h
index 09f96084..1d0f389e 100644
--- a/filter/filter.h
+++ b/filter/filter.h
@@ -18,7 +18,6 @@
 
 #define FI__TWOCHAR(a,b)	((a<<8) | b)
 #define FI__LIST \
-  F(FI_COMMA,			  0, ',') \
   F(FI_ADD,			  0, '+') \
   F(FI_SUBTRACT,		  0, '-') \
   F(FI_MULTIPLY,		  0, '*') \

From e8bc64e308586b6502090da2775af84cd760ed0d Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Wed, 28 Feb 2018 16:57:50 +0100
Subject: [PATCH 142/198] Filter: make bgpmask literals real constructors

The bgpmask literals can include expressions. This is OK but they have
to be interpreted as soon as the code is run, not in the time the code
is used as value.

This led to strange behavior like rewriting bgpmasks when they shan't
be rewritten:

	function mask_generator(int as)
	{
		return [= * as * =];
	}

	function another()
	bgpmask m1;
	bgpmask m2;
	{
		m1 = mask_generator(10);
		m2 = mask_generator(20);
		if (m1 == m2) {
			print("strange"); # this would happen
		}
	}

Moreover, sending this to CLI would cause stack overflow and knock down the
whole BIRD, as soon as there is at least one route to execute the given
filter on.

	show route filter bgpmask mmm; bgppath ppp; { ppp = +empty+; mmm = [= (ppp ~ mmm) =]; print(mmm); accept; }

The magic match operator (~) inside the bgpmask literal would try to
resolve mmm, which points to the same bgpmask so it would resolve
itself, call the magic match operator and vice versa.

After this patch, the bgpmask literal will get resolved as soon as it's
assigned to mmm and it also will return a type error as bool is not
convertible to ASN in BIRD.
---
 filter/config.Y | 22 +++++++++++++++++++++-
 filter/filter.c | 39 +++++++++++++++++++++++++++++----------
 filter/filter.h |  2 +-
 nest/a-path.c   |  3 +--
 4 files changed, 52 insertions(+), 14 deletions(-)

diff --git a/filter/config.Y b/filter/config.Y
index 1ef5a3a8..7c077b65 100644
--- a/filter/config.Y
+++ b/filter/config.Y
@@ -317,6 +317,26 @@ f_generate_lc(struct f_inst *t1, struct f_inst *t2, struct f_inst *t3)
   return rv;
 }
 
+static inline struct f_inst *
+f_generate_path_mask(struct f_path_mask *t)
+{
+  for (struct f_path_mask *tt = t; tt; tt = tt->next) {
+    if (tt->kind == PM_ASN_EXPR) {
+      struct f_inst *mrv = f_new_inst(FI_PATHMASK_CONSTRUCT);
+      mrv->a1.p = t;
+      return mrv;
+    }
+  }
+
+  NEW_F_VAL;
+  val->type = T_PATH_MASK;
+  val->val.path_mask = t;
+
+  struct f_inst *rv = f_new_inst(FI_CONSTANT_INDIRECT);
+  rv->a1.p = val;
+
+  return rv;
+}
 
 
 CF_DECLS
@@ -708,13 +728,13 @@ constant:
  | '[' set_items ']' { DBG( "We've got a set here..." ); $$ = f_new_inst(FI_CONSTANT); $$->aux = T_SET; $$->a2.p = build_tree($2); DBG( "ook\n" ); }
  | '[' fprefix_set ']' { $$ = f_new_inst(FI_CONSTANT); $$->aux = T_PREFIX_SET;  $$->a2.p = $2; }
  | ENUM	  { $$ = f_new_inst(FI_CONSTANT); $$->aux = $1 >> 16; $$->a2.i = $1 & 0xffff; }
- | bgp_path { NEW_F_VAL; $$ = f_new_inst(FI_CONSTANT_INDIRECT); val->type = T_PATH_MASK; val->val.path_mask = $1; $$->a1.p = val; }
  ;
 
 constructor:
    '(' term ',' term ')' { $$ = f_generate_dpair($2, $4); }
  | '(' ec_kind ',' term ',' term ')' { $$ = f_generate_ec($2, $4, $6); }
  | '(' term ',' term ',' term ')' { $$ = f_generate_lc($2, $4, $6); }
+ | bgp_path { $$ = f_generate_path_mask($1); }
  ;
 
 
diff --git a/filter/filter.c b/filter/filter.c
index 023f7e2f..8b66b57e 100644
--- a/filter/filter.c
+++ b/filter/filter.c
@@ -82,8 +82,7 @@ pm_format(struct f_path_mask *p, buffer *buf)
       break;
 
     case PM_ASN_EXPR:
-      buffer_print(buf, "%u ", f_eval_asn((struct f_inst *) p->val));
-      break;
+      ASSERT(0);
     }
 
     p = p->next;
@@ -772,6 +771,32 @@ interpret(struct f_inst *what)
       break;
     }
 
+  case FI_PATHMASK_CONSTRUCT:
+    {
+      struct f_path_mask *tt = what->a1.p, *vbegin, **vv = &vbegin;
+
+      while (tt) {
+	*vv = lp_alloc(f_pool, sizeof(struct f_path_mask));
+	if (tt->kind == PM_ASN_EXPR) {
+	  struct f_val res = interpret((struct f_inst *) tt->val);
+	  (*vv)->kind = PM_ASN;
+	  if (res.type != T_INT) {
+	    runtime( "Error resolving path mask template: value not an integer" );
+	    return (struct f_val) { .type = T_VOID };
+	  }
+
+	  (*vv)->val = res.val.i;
+	} else {
+	  **vv = *tt;
+	}
+	tt = tt->next;
+	vv = &((*vv)->next);
+      }
+
+      res = (struct f_val) { .type = T_PATH_MASK, .val.path_mask = vbegin };
+      break;
+    }
+
 /* Relational operators */
 
 #define COMPARE(x) \
@@ -1550,6 +1575,8 @@ i_same(struct f_inst *f1, struct f_inst *f2)
   case FI_LT:
   case FI_LTE: TWOARGS; break;
 
+  case FI_PATHMASK_CONSTRUCT: if (!pm_same(f1->a1.p, f2->a1.p)) return 0; break;
+
   case FI_NOT: ONEARG; break;
   case FI_NOT_MATCH:
   case FI_MATCH: TWOARGS; break;
@@ -1771,14 +1798,6 @@ f_eval_int(struct f_inst *expr)
   return res.val.i;
 }
 
-u32
-f_eval_asn(struct f_inst *expr)
-{
-  /* Called as a part of another interpret call, therefore no log_reset() */
-  struct f_val res = interpret(expr);
-  return (res.type == T_INT) ? res.val.i : 0;
-}
-
 /**
  * filter_same - compare two filters
  * @new: first filter to be compared
diff --git a/filter/filter.h b/filter/filter.h
index 1d0f389e..572459d8 100644
--- a/filter/filter.h
+++ b/filter/filter.h
@@ -27,6 +27,7 @@
   F(FI_PAIR_CONSTRUCT,		'm', 'p') \
   F(FI_EC_CONSTRUCT,		'm', 'c') \
   F(FI_LC_CONSTRUCT,		'm', 'l') \
+  F(FI_PATHMASK_CONSTRUCT,	'm', 'P') \
   F(FI_NEQ,			'!', '=') \
   F(FI_EQ,			'=', '=') \
   F(FI_LT,			  0, '<') \
@@ -196,7 +197,6 @@ int f_run(struct filter *filter, struct rte **rte, struct ea_list **tmp_attrs, s
 struct f_val f_eval_rte(struct f_inst *expr, struct rte **rte, struct linpool *tmp_pool);
 struct f_val f_eval(struct f_inst *expr, struct linpool *tmp_pool);
 uint f_eval_int(struct f_inst *expr);
-u32 f_eval_asn(struct f_inst *expr);
 
 char *filter_name(struct filter *filter);
 int filter_same(struct filter *new, struct filter *old);
diff --git a/nest/a-path.c b/nest/a-path.c
index b453f702..0272c6d7 100644
--- a/nest/a-path.c
+++ b/nest/a-path.c
@@ -535,8 +535,7 @@ as_path_match(struct adata *path, struct f_path_mask *mask)
 	  val2 = val = mask->val;
 	  goto step;
 	case PM_ASN_EXPR:
-	  val2 = val = f_eval_asn((struct f_inst *) mask->val);
-	  goto step;
+	  ASSERT(0);
 	case PM_ASN_RANGE:
 	  val = mask->val;
 	  val2 = mask->val2;

From 7a8ae228f907400a33375c07b31f3e42a89834e7 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Sat, 17 Mar 2018 17:14:02 +0100
Subject: [PATCH 143/198] Doc: Update BGP documentation

Thanks to Joshua McQuistan for the bugreport.
---
 doc/bird.sgml | 16 +++++++++-------
 1 file changed, 9 insertions(+), 7 deletions(-)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index 5ed816e8..c5d74228 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -2118,22 +2118,24 @@ to set routing policy and all the other parameters differently for each neighbor
 using the following configuration parameters:
 
 <descrip>
-	<tag><label id="bgp-local">local [<m/ip/] as <m/number/</tag>
+	<tag><label id="bgp-local">local [<m/ip/] [port <m/number/] [as <m/number/]</tag>
 	Define which AS we are part of. (Note that contrary to other IP routers,
 	BIRD is able to act as a router located in multiple AS'es simultaneously,
 	but in such cases you need to tweak the BGP paths manually in the filters
 	to get consistent behavior.) Optional <cf/ip/ argument specifies a source
-	address, equivalent to the <cf/source address/ option (see below). This
-	parameter is mandatory.
+	address, equivalent to the <cf/source address/ option (see below).
+	Optional </cf/port> argument specifies the local BGP port instead of
+	standard port 179. The parameter may be used multiple times with
+	different sub-options (e.g., both <cf/local 10.0.0.1 as 65000;/ and
+	<cf/local 10.0.0.1; local as 65000;/ are valid). This parameter is
+	mandatory.
 
 	<tag><label id="bgp-neighbor">neighbor [<m/ip/] [port <m/number/] [as <m/number/]</tag>
 	Define neighboring router this instance will be talking to and what AS
 	it is located in. In case the neighbor is in the same AS as we are, we
 	automatically switch to iBGP. Optionally, the remote port may also be
-	specified. The parameter may be used multiple times with different
-	sub-options (e.g., both <cf/neighbor 10.0.0.1 as 65000;/ and
-	<cf/neighbor 10.0.0.1; neighbor as 65000;/ are valid). This parameter is
-	mandatory.
+	specified. Like <cf/local/ parameter, this parameter may also be used
+	multiple times with different sub-options. This parameter is mandatory.
 
 	<tag><label id="bgp-iface">interface <m/string/</tag>
 	Define interface we should use for link-local BGP IPv6 sessions.

From 159d619caf2311b62dc0d876f22c42b5d23e86c0 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Sat, 17 Mar 2018 22:25:06 +0100
Subject: [PATCH 144/198] Doc: SADR documentation

---
 doc/bird.sgml | 61 ++++++++++++++++++++++++++++++++++++++-------------
 1 file changed, 46 insertions(+), 15 deletions(-)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index c5d74228..2ca44e38 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -293,6 +293,20 @@ routes are:
 	<item>Route next hops (see below)
 </itemize>
 
+<sect1>IPv6 source-specific routes
+<label id="ip-sadr-routes">
+
+<p>The IPv6 routes containing both destination and source prefix. They are used
+for source-specific routing (SSR), also called source-address dependent routing
+(SADR), see <rfc id="8043">. Currently limited mostly to the Babel protocol.
+Configuration keyword is <cf/ipv6 sadr/.
+
+<itemize>
+	<item>(PK) Route destination (IP prefix together with its length)
+	<item>(PK) Route source (IP prefix together with its length)
+	<item>Route next hops (see below)
+</itemize>
+
 <sect1>VPN IPv4 and IPv6 routes
 <label id="vpn-routes">
 
@@ -1239,6 +1253,12 @@ foot).
 	pair, and <cf/.len/, which separates prefix length from the pair.
 	So <cf>1.2.0.0/16.len = 16</cf> is true.
 
+	<cf/NET_IP6_SADR/ nettype holds both destination and source IPv6
+	prefix. The literals are written as <cf><m/ipaddress//<m/pxlen/ from
+	<m/ipaddress//<m/pxlen/</cf>, where the first part is the destination
+	prefix and the second art is the source prefix. They support the same
+	operators as IP prefixes, but just for the destination part.
+
 	<cf/NET_VPN4/ and <cf/NET_VPN6/ prefixes hold an IP prefix with VPN
 	Route Distinguisher (<rfc id="4364">). They support the same special
 	operators as IP prefixes, and also <cf/.rd/ which extracts the Route
@@ -1459,6 +1479,7 @@ foot).
 	lclists, with LCs instead of pairs as arguments.
 </descrip>
 
+
 <sect>Operators
 <label id="operators">
 
@@ -1651,19 +1672,25 @@ cases desirable.
 routes over the same IPv6 transport. For sending and receiving Babel packets,
 only a link-local IPv6 address is needed.
 
-<p>BIRD does not implement any Babel extensions, but will coexist with
-implementations using extensions (and will just ignore extension messages).
+<p>BIRD implements an extension for IPv6 source-specific routing (SSR or SADR),
+but must be configured accordingly to use it. SADR-enabled Babel router can
+interoperate with non-SADR Babel router, but the later would ignore routes
+with specific (non-zero) source prefix.
 
 <sect1>Configuration
 <label id="babel-config">
 
-<p>Babel supports no global configuration options apart from those common to all
-other protocols, but supports the following per-interface configuration options:
+<p>The Babel protocol support both IPv4 and IPv6 channels; both can be
+configured simultaneously. It can also be configured with <ref
+id="ip-sadr-routes" name="IPv6 SADR"> channel instead of regular IPv6
+channel, in such case SADR support is enabled. Babel supports no global
+configuration options apart from those common to all other protocols, but
+supports the following per-interface configuration options:
 
 <code>
 protocol babel [<name>] {
 	ipv4 { <channel config> };
-	ipv6 { <channel config> };
+	ipv6 [sadr] { <channel config> };
 	interface <interface pattern> {
 		type <wired|wireless>;
 		rxcost <number>;
@@ -1683,8 +1710,8 @@ protocol babel [<name>] {
 </code>
 
 <descrip>
-      <tag><label id="babel-channel">ipv4|ipv6 <m/channel config/</tag>
-      The supported channels are IPv4 and IPv6.
+      <tag><label id="babel-channel">ipv4 | ipv6 [sadr] <m/channel config/</tag>
+      The supported channels are IPv4, IPv6, and IPv6 SADR.
 
       <tag><label id="babel-type">type wired|wireless </tag>
       This option specifies the interface type: Wired or wireless. On wired
@@ -2124,7 +2151,7 @@ using the following configuration parameters:
 	but in such cases you need to tweak the BGP paths manually in the filters
 	to get consistent behavior.) Optional <cf/ip/ argument specifies a source
 	address, equivalent to the <cf/source address/ option (see below).
-	Optional </cf/port> argument specifies the local BGP port instead of
+	Optional <cf/port/ argument specifies the local BGP port instead of
 	standard port 179. The parameter may be used multiple times with
 	different sub-options (e.g., both <cf/local 10.0.0.1 as 65000;/ and
 	<cf/local 10.0.0.1; local as 65000;/ are valid). This parameter is
@@ -2745,16 +2772,17 @@ protocol device {
 <p>The Direct protocol is a simple generator of device routes for all the
 directly connected networks according to the list of interfaces provided by the
 kernel via the Device protocol. The Direct protocol supports both IPv4 and IPv6
-channels.
+channels; both can be configured simultaneously. It can also be configured with
+<ref id="ip-sadr-routes" name="IPv6 SADR"> channel instead of regular IPv6
+channel in order to be used together with SADR-enabled Babel protocol.
 
 <p>The question is whether it is a good idea to have such device routes in BIRD
 routing table. OS kernel usually handles device routes for directly connected
 networks by itself so we don't need (and don't want) to export these routes to
 the kernel protocol. OSPF protocol creates device routes for its interfaces
-itself and BGP protocol is usually used for exporting aggregate routes. Although
-there are some use cases that use the direct protocol (like abusing eBGP as an
-IGP routing protocol), in most cases it is not needed to have these device
-routes in BIRD routing table and to use the direct protocol.
+itself and BGP protocol is usually used for exporting aggregate routes. But the
+Direct protocol is necessary for distance-vector protocols like RIP or Babel to
+announce local networks.
 
 <p>There is one notable case when you definitely want to use the direct protocol
 -- running BIRD on BSD systems. Having high priority device routes for directly
@@ -2829,8 +2857,10 @@ kernel protocols to the same routing table and changing route destination
 (gateway) in an export filter of a kernel protocol does not work. Both
 limitations can be overcome using another routing table and the pipe protocol.
 
-<p>The Kernel protocol supports both IPv4 and IPv6 channels; only one of them
-can be configured in each protocol instance.
+<p>The Kernel protocol supports both IPv4 and IPv6 channels; only one channel
+can be configured in each protocol instance. On Linux, it also supports <ref
+id="ip-sadr-routes" name="IPv6 SADR"> and <ref id="mpls-routes" name="MPLS">
+channels.
 
 <sect1>Configuration
 <label id="krt-config">
@@ -4281,6 +4311,7 @@ protocol rip {
 
 
 <sect>RPKI
+<label id="rpki">
 
 <sect1>Introduction
 

From 364d5823eac85178361bad188d89949b7e0d321c Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Sun, 18 Mar 2018 02:56:51 +0100
Subject: [PATCH 145/198] Nest: SADR support for Direct

---
 nest/rt-dev.c | 41 +++++++++++++++++++++++++++++++++--------
 nest/rt-dev.h |  3 +++
 2 files changed, 36 insertions(+), 8 deletions(-)

diff --git a/nest/rt-dev.c b/nest/rt-dev.c
index 718c4578..66f458e7 100644
--- a/nest/rt-dev.c
+++ b/nest/rt-dev.c
@@ -31,6 +31,7 @@ dev_ifa_notify(struct proto *P, uint flags, struct ifa *ad)
   struct rt_dev_proto *p = (void *) P;
   struct rt_dev_config *cf = (void *) P->cf;
   struct channel *c;
+  net_addr *net = &ad->prefix;
 
   if (!EMPTY_LIST(cf->iface_list) &&
       !iface_patt_find(&cf->iface_list, ad->iface, ad))
@@ -53,13 +54,20 @@ dev_ifa_notify(struct proto *P, uint flags, struct ifa *ad)
   if (!c)
     return;
 
+  /* For IPv6 SADR, replace regular prefix with SADR prefix */
+  if (c->net_type == NET_IP6_SADR)
+  {
+    net = alloca(sizeof(net_addr_ip6_sadr));
+    net_fill_ip6_sadr(net, net6_prefix(&ad->prefix), net6_pxlen(&ad->prefix), IP6_NONE, 0);
+  }
+
   if (flags & IF_CHANGE_DOWN)
     {
       DBG("dev_if_notify: %s:%I going down\n", ad->iface->name, ad->ip);
 
       /* Use iface ID as local source ID */
       struct rte_src *src = rt_get_source(P, ad->iface->index);
-      rte_update2(c, &ad->prefix, NULL, src);
+      rte_update2(c, net, NULL, src);
     }
   else if (flags & IF_CHANGE_UP)
     {
@@ -85,7 +93,7 @@ dev_ifa_notify(struct proto *P, uint flags, struct ifa *ad)
       a = rta_lookup(&a0);
       e = rte_get_temp(a);
       e->pflags = 0;
-      rte_update2(c, &ad->prefix, e, src);
+      rte_update2(c, net, e, src);
     }
 }
 
@@ -107,16 +115,32 @@ dev_if_notify(struct proto *p, uint c, struct iface *iface)
   }
 }
 
+static void
+dev_postconfig(struct proto_config *CF)
+{
+  struct rt_dev_config *cf = (void *) CF;
+  struct channel_config *ip4, *ip6, *ip6_sadr;
+
+  ip4 = proto_cf_find_channel(CF, NET_IP4);
+  ip6 = proto_cf_find_channel(CF, NET_IP6);
+  ip6_sadr = proto_cf_find_channel(CF, NET_IP6_SADR);
+
+  if (ip6 && ip6_sadr)
+    cf_error("Both ipv6 and ipv6-sadr channels");
+
+  cf->ip4_channel = ip4;
+  cf->ip6_channel = ip6 ?: ip6_sadr;
+}
 
 static struct proto *
 dev_init(struct proto_config *CF)
 {
   struct proto *P = proto_new(CF);
   struct rt_dev_proto *p = (void *) P;
-  // struct rt_dev_config *cf = (void *) CF;
+  struct rt_dev_config *cf = (void *) CF;
 
-  proto_configure_channel(P, &p->ip4_channel, proto_cf_find_channel(CF, NET_IP4));
-  proto_configure_channel(P, &p->ip6_channel, proto_cf_find_channel(CF, NET_IP6));
+  proto_configure_channel(P, &p->ip4_channel, cf->ip4_channel);
+  proto_configure_channel(P, &p->ip6_channel, cf->ip6_channel);
 
   P->if_notify = dev_if_notify;
   P->ifa_notify = dev_ifa_notify;
@@ -136,8 +160,8 @@ dev_reconfigure(struct proto *P, struct proto_config *CF)
     return 0;
 
   return
-    proto_configure_channel(P, &p->ip4_channel, proto_cf_find_channel(CF, NET_IP4)) &&
-    proto_configure_channel(P, &p->ip6_channel, proto_cf_find_channel(CF, NET_IP6));
+    proto_configure_channel(P, &p->ip4_channel, n->ip4_channel) &&
+    proto_configure_channel(P, &p->ip6_channel, n->ip6_channel);
 
   return 1;
 }
@@ -162,9 +186,10 @@ struct protocol proto_device = {
   .name =		"Direct",
   .template =		"direct%d",
   .preference =		DEF_PREF_DIRECT,
-  .channel_mask =	NB_IP,
+  .channel_mask =	NB_IP | NB_IP6_SADR,
   .proto_size =		sizeof(struct rt_dev_proto),
   .config_size =	sizeof(struct rt_dev_config),
+  .postconfig =		dev_postconfig,
   .init =		dev_init,
   .reconfigure =	dev_reconfigure,
   .copy_config =	dev_copy_config
diff --git a/nest/rt-dev.h b/nest/rt-dev.h
index 20b88a64..5d91242a 100644
--- a/nest/rt-dev.h
+++ b/nest/rt-dev.h
@@ -13,6 +13,9 @@ struct rt_dev_config {
   struct proto_config c;
   list iface_list;		/* list of struct iface_patt */
   int check_link;
+
+  struct channel_config *ip4_channel;
+  struct channel_config *ip6_channel;
 };
 
 struct rt_dev_proto {

From bcb4af81fc8ea0acf7c2fa5a6854cd3c23d92d9f Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Sun, 18 Mar 2018 13:48:47 +0100
Subject: [PATCH 146/198] Nest: Fix table reconfiguration when nettype changes

Thanks to Toke Hoiland-Jorgensen for the bugreport.
---
 nest/rt-table.c | 12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)

diff --git a/nest/rt-table.c b/nest/rt-table.c
index 686d0e84..b885c6e3 100644
--- a/nest/rt-table.c
+++ b/nest/rt-table.c
@@ -2076,6 +2076,13 @@ rt_unlock_table(rtable *r)
     }
 }
 
+static struct rtable_config *
+rt_find_table_config(struct config *cf, char *name)
+{
+  struct symbol *sym = cf_find_symbol(cf, name);
+  return (sym && (sym->class == SYM_TABLE)) ? sym->def : NULL;
+}
+
 /**
  * rt_commit - commit new routing table configuration
  * @new: new configuration
@@ -2101,11 +2108,10 @@ rt_commit(struct config *new, struct config *old)
 	  rtable *ot = o->table;
 	  if (!ot->deleted)
 	    {
-	      struct symbol *sym = cf_find_symbol(new, o->name);
-	      if (sym && sym->class == SYM_TABLE && !new->shutdown)
+	      r = rt_find_table_config(new, o->name);
+	      if (r && (r->addr_type == o->addr_type) && !new->shutdown)
 		{
 		  DBG("\t%s: same\n", o->name);
-		  r = sym->def;
 		  r->table = ot;
 		  ot->name = r->name;
 		  ot->config = r;

From 3b522a1e5cd034196cd0f5c0eab4d9e87f0f2a8a Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 20 Mar 2018 19:28:26 +0100
Subject: [PATCH 147/198] Doc: Redesign default config file

The old one does not work with 2.0.x.
---
 doc/bird.conf.example | 358 ++++++++++++++++++++----------------------
 1 file changed, 170 insertions(+), 188 deletions(-)

diff --git a/doc/bird.conf.example b/doc/bird.conf.example
index 62c65ce9..384270e3 100644
--- a/doc/bird.conf.example
+++ b/doc/bird.conf.example
@@ -1,222 +1,204 @@
-/*
- *	This is an example configuration file
- *	(for version 1.x.x, obsolete)
- */
-
-# Yes, even shell-like comments work...
+# This is a basic configuration file, which contains boilerplate options and
+# some basic examples. It allows the BIRD daemon to start but will not cause
+# anything else to happen.
+#
+# Please refer to the BIRD User's Guide documentation, which is also available
+# online at http://bird.network.cz/ in HTML format, for more information on
+# configuring BIRD and adding routing protocols.
 
 # Configure logging
-#log syslog { debug, trace, info, remote, warning, error, auth, fatal, bug };
-#log stderr all;
-#log "tmp" all;
+log syslog all;
+# log "/var/log/bird.log" { debug, trace, info, remote, warning, error, auth, fatal, bug };
 
-# Override router ID
-#router id 198.51.100.1;
+# Set router ID. It is a unique identification of your router, usually one of
+# IPv4 addresses of the router. It is recommended to configure it explicitly.
+# router id 198.51.100.1;
 
-# You can define your own symbols...
-#define xyzzy = (120+10);
-#define '1a-a1' = (30+40);
-
-# Define a route filter...
-#filter test_filter {
-#	if net ~ 10.0.0.0/16 then accept;
-#	else reject;
-#}
-
-#filter sink { reject; }
-#filter okay { accept; }
-
-#include "filters.conf";
-
-# Define another routing table
-#table testable;
-
-# Turn on global debugging of all protocols
-#debug protocols all;
+# Turn on global debugging of all protocols (all messages or just selected classes)
+# debug protocols all;
+# debug protocols { events, states };
 
 # Turn on internal watchdog
-#watchdog warning 5 s;
-#watchdog timeout 30 s;
+# watchdog warning 5 s;
+# watchdog timeout 30 s;
 
-# The direct protocol automatically generates device routes to
-# all network interfaces. Can exist in as many instances as you wish
-# if you want to populate multiple routing tables with device routes.
-#protocol direct {
-#	interface "-eth*", "*";	# Restrict network interfaces it works with
-#}
+# You can define your own constants
+# define my_asn = 65000;
+# define my_addr = 198.51.100.1;
 
-# This pseudo-protocol performs synchronization between BIRD's routing
-# tables and the kernel. If your kernel supports multiple routing tables
-# (as Linux 2.2.x does), you can run multiple instances of the kernel
-# protocol and synchronize different kernel tables with different BIRD tables.
-protocol kernel {
-#	learn;			# Learn all alien routes from the kernel
-	persist;		# Don't remove routes on bird shutdown
-	scan time 20;		# Scan kernel routing table every 20 seconds
-#	import none;		# Default is import all
-	export all;		# Default is export none
-#	kernel table 5;		# Kernel table to synchronize with (default: main)
-}
+# Tables master4 and master6 are defined by default
+# ipv4 table master4;
+# ipv6 table master6;
 
-# This pseudo-protocol watches all interface up/down events.
+# Define more tables, e.g. for policy routing or as MRIB
+# ipv4 table mrib4;
+# ipv6 table mrib6;
+
+# The Device protocol is not a real routing protocol. It does not generate any
+# routes and it only serves as a module for getting information about network
+# interfaces from the kernel. It is necessary in almost any configuration.
 protocol device {
-	scan time 10;		# Scan interfaces every 10 seconds
 }
 
-# Static routes (again, there can be multiple instances, so that you
-# can disable/enable various groups of static routes on the fly).
+# The direct protocol is not a real routing protocol. It automatically generates
+# direct routes to all network interfaces. Can exist in as many instances as you
+# wish if you want to populate multiple routing tables with direct routes.
+protocol direct {
+	disabled;		# Disable by default
+	ipv4;			# Connect to default IPv4 table
+	ipv6;			# ... and to default IPv6 table
+}
+
+# The Kernel protocol is not a real routing protocol. Instead of communicating
+# with other routers in the network, it performs synchronization of BIRD
+# routing tables with the OS kernel. One instance per table.
+protocol kernel {
+	ipv4 {			# Connect protocol to IPv4 table by channel
+#	      table master4;	# Default IPv4 table is master4
+#	      import all;	# Import to table, default is import all
+	      export all;	# Export to protocol. default is export none
+	};
+#	learn;			# Learn alien routes from the kernel
+#	kernel table 10;	# Kernel table to synchronize with (default: main)
+}
+
+# Another instance for IPv6, skipping default options
+protocol kernel {
+	ipv6 { export all; };
+}
+
+# Static routes (Again, there can be multiple instances, for different address
+# families and to disable/enable various groups of static routes on the fly).
 protocol static {
-#	disabled;		# Disable by default
-#	table testable;		# Connect to a non-default table
-#	preference 1000;	# Default preference of routes
-#	debug { states, routes, filters, interfaces, events, packets };
-#	debug all;
-#	route 0.0.0.0/0 via 198.51.100.13;
-#	route 198.51.100.0/25 unreachable;
+	ipv4;			# Again, IPv4 channel with default options
+
+#	route 0.0.0.0/0 via 198.51.100.10;
+#	route 192.0.2.0/24 blackhole;
 #	route 10.0.0.0/8 unreachable;
-#	route 10.1.1.0:255.255.255.0 via 198.51.100.3;
-#	route 10.1.2.0:255.255.255.0 via 198.51.100.3;
-#	route 10.1.3.0:255.255.255.0 via 198.51.100.4;
-#	route 10.2.0.0/24 via "arc0";
+#	route 10.2.0.0/24 via "eth0";
+#	# Static routes can be defined with optional attributes
+#	route 10.1.1.0/24 via 198.51.100.3 { rip_metric = 3; };
+#	route 10.1.2.0/24 via 198.51.100.3 { ospf_metric1 = 100; };
+#	route 10.1.3.0/24 via 198.51.100.4 { ospf_metric2 = 100; };
 }
 
-# Pipe protocol connects two routing tables... Beware of loops.
-#protocol pipe {
-#	peer table testable;
-# Define what routes do we export to this protocol / import from it.
-#	import all;		# default is all
-#	export all;		# default is none
-#	import none;		# If you wish to disable imports
-#	import filter test_filter;		# Use named filter
-#	import where source = RTS_DEVICE;	# Use explicit filter
-#}
+# Pipe protocol connects two routing tables. Beware of loops.
+# protocol pipe {
+#	table master4;		# No ipv4/ipv6 channel definition like in other protocols
+#	peer table mrib4;
+#	import all;		# Direction peer table -> table
+#	export all;		# Direction table -> peer table
+# }
 
-# RIP aka Rest In Pieces...
-#protocol rip MyRIP {	# You can also use an explicit name
-#	preference xyzzy;
-#	debug all;
-#	port 1520;
-#	period 7;
-#	infinity 16;
-#	garbage time 60;
-#	interface "*" { mode broadcast; };
-#	honor neighbor;		# To whom do we agree to send the routing table
-#	honor always;
-#	honor never;
-#	passwords {
-#		password "nazdar";
+# RIP example, both RIP and RIPng are supported
+# protocol rip {
+#	ipv4 {
+#		# Export direct, static routes and ones from RIP itself
+#		import all;
+#		export where source ~ [ RTS_DEVICE, RTS_STATIC, RTS_RIP ];
 #	};
-#	authentication none;
-#	import filter { print "importing"; accept; };
-#	export filter { print "exporting"; accept; };
-#}
+#	interface "eth*" {
+#	  	update time 10;			# Default period is 30
+#		timeout time 60;		# Default timeout is 180
+#		authentication cryptographic;	# No authentication by default
+#		password "hello" { algorithm hmac sha256; }; # Default is MD5
+#	};
+# }
 
-#protocol ospf MyOSPF {
-#       tick 2;
-#	rfc1583compat yes;
-#	area 0.0.0.0 {
-#		stub no;
+# OSPF example, both OSPFv2 and OSPFv3 are supported
+# protocol ospf v3 {
+#  	ipv6 {
+#		import all;
+#		export where source = RTS_STATIC;
+#	};
+#	area 0 {
 #		interface "eth*" {
-#			hello 9;
-#			retransmit 6;
-#			cost 10;
-#			transmit delay 5;
-#			dead count 5;
-#			wait 50;
-#			type broadcast;
-#			authentication simple;
-#			password "pass";
+#			type broadcast;		# Detected by default
+#			cost 10;		# Interface metric
+#			hello 5;		# Default hello perid 10 is too long
 #		};
-#		interface "arc0" {
-#			rx buffer large;
-#			type nonbroadcast;
-#			poll 14;
-#			dead 75;
-#			neighbors {
-#				10.1.1.2 eligible;
-#				10.1.1.4;
-#			};
-#			strict nonbroadcast yes;
+#		interface "tun*" {
+#			type ptp;		# PtP mode, avoids DR selection
+#			cost 100;		# Interface metric
+#			hello 5;		# Default hello perid 10 is too long
 #		};
-#		interface "xxx0" {
-#                       passwords {
-#				password "abc" {
-#					id 1;
-#					generate to "22-04-2003 11:00:06";
-#					accept to "17-01-2004 12:01:05";
-#				};
-#				password "def" {
-#					id 2;
-#					generate from "22-04-2003 11:00:07";
-#					accept from "17-01-2003 12:01:05";
-#				};
-#			};
-#                       authentication cryptographic;
+#		interface "dummy0" {
+#			stub;			# Stub interface, just propagate it
 #		};
 #	};
-#	area 20 {
-#		stub 1;
-#		interface "ppp1" {
-#			hello 8;
-#			authentication none;
-#		};
-#               interface "fr*";
-#               virtual link 192.168.0.1 {
-#                       password "sdsdffsdfg";
-#                       authentication cryptographic;
-#               };
-#	};
 #}
 
+# Define simple filter as an example for BGP import filter
+# See https://gitlab.labs.nic.cz/labs/bird/wikis/BGP_filtering for more examples
+# filter rt_import
+# {
+#	if bgp_path.first != 64496 then accept;
+#	if bgp_path.len > 64 then accept;
+#	if bgp_next_hop != from then accept;
+#	reject;
+# }
 
-#protocol bgp {
-#	disabled;
+# BGP example, explicit name 'uplink1' is used instead of default 'bgp1'
+# protocol bgp uplink1 {
 #	description "My BGP uplink";
-#	local as 65000;
-#	neighbor 198.51.100.130 as 64496;
-#	multihop;
-#	hold time 240;
-#	startup hold time 240;
-#	connect retry time 120;
-#	keepalive time 80;	# defaults to hold time / 3
-#	start delay time 5;	# How long do we wait before initial connect
-#	error wait time 60, 300;# Minimum and maximum time we wait after an error (when consecutive
-#				# errors occur, we increase the delay exponentially ...
-#	error forget time 300;	# ... until this timeout expires)
-#	disable after error;	# Disable the protocol automatically when an error occurs
-#	next hop self;		# Disable next hop processing and always advertise our local address as nexthop
-#	path metric 1;		# Prefer routes with shorter paths (like Cisco does)
-#	default bgp_med 0;	# MED value we use for comparison when none is defined
-#	default bgp_local_pref 0;	# The same for local preference
-#	source address 198.51.100.14;	# What local address we use for the TCP connection
+#	local 198.51.100.1 as 65000;
+#	neighbor 198.51.100.10 as 64496;
+#	hold time 90;		# Default is 240
 #	password "secret";	# Password used for MD5 authentication
-#	rr client;		# I am a route reflector and the neighor is my client
-#	rr cluster id 1.0.0.1;	# Use this value for cluster id instead of my router id
-#	export where source=RTS_STATIC;
-#	export filter {
-#		if source = RTS_STATIC then {
-#			bgp_community = -empty-; bgp_community = add(bgp_community,(65000,5678));
-#			bgp_origin = 0;
-#			bgp_community = -empty-; bgp_community.add((65000,5678));
-#			if (65000,64501) ~ bgp_community then
-#				bgp_community.add((0, 1));
-#			if bgp_path ~ [= 65000 =] then
-#				bgp_path.prepend(65000);
-#			accept;
-#		}
-#		reject;
+#
+#	ipv4 {			# regular IPv4 unicast (1/1)
+#		import filter rt_import;
+#		export where source ~ [ RTS_STATIC, RTS_BGP ];
+#	};
+#
+#	ipv6 {			# regular IPv6 unicast (2/1)
+#		import filter rt_import;
+#		export filter {	# The same as 'where' expression above
+#			if source ~ [ RTS_STATIC, RTS_BGP ]
+#			then accept;
+#			else reject;
+#		};
+#	};
+#
+#	ipv4 multicast {	# IPv4 multicast topology (1/2)
+#		table mrib4;	# explicit IPv4 table
+#		import filter rt_import;
+#		export all;
+#	};
+#
+#	ipv6 multicast {	# IPv6 multicast topology (2/2)
+#		table mrib6;	# explicit IPv6 table
+#		import filter rt_import;
+#		export all;
 #	};
 #}
-#
-# Template usage example
-#template bgp rr_client {
-#	disabled;
-#	local as 65000;
-#	multihop;
+
+# Template example. Using templates to define IBGP route reflector clients.
+# template bgp rr_clients {
+#	local 10.0.0.1 as 65000;
+#	neighbor as 65000;
 #	rr client;
 #	rr cluster id 1.0.0.1;
-#}
 #
-#protocol bgp rr_abcd from rr_client {
-#	neighbor 10.1.4.7 as 65000;
-#}
+#	ipv4 {
+#		import all;
+#		export where source = RTS_BGP;
+#	};
+#
+#	ipv6 {
+#		import all;
+#		export where source = RTS_BGP;
+#	};
+# }
+#
+# protocol bgp client1 from rr_clients {
+#	neighbor 10.0.1.1;
+# }
+#
+# protocol bgp client2 from rr_clients {
+#	neighbor 10.0.2.1;
+# }
+#
+# protocol bgp client3 from rr_clients {
+#	neighbor 10.0.3.1;
+# }

From a177e4dd04b29dc364d2c5505fe2c483981c6498 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 21 Mar 2018 16:32:51 +0100
Subject: [PATCH 148/198] Doc: Minor update

---
 doc/bird.sgml | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index 2ca44e38..903456f0 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -660,9 +660,9 @@ agreement").
 	restricted to interfaces assigned to the VRF and will use sockets bound
 	to the VRF. Appropriate VRF interface must exist on OS level. For kernel
 	protocol, an appropriate table still must be explicitly selected by
-	<cf/table/ option. Note that the VRF support in BIRD and Linux kernel
-	(4.11) is still in development and is currently problematic outside of
-	multihop BGP.
+	<cf/table/ option. Note that for proper VRF support it is necessary to
+	use Linux kernel version at least 4.14, older versions have limited VRF
+	implementation.
 
 	<tag><label id="proto-channel"><m/channel name/ [{<m/channel config/}]</tag>
 	Every channel must be explicitly stated. See the protocol-specific

From 4841804fffd8bf669b2445ec3328b6a49eed31f9 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 21 Mar 2018 15:17:56 +0100
Subject: [PATCH 149/198] NEWS and version update

---
 NEWS            | 12 ++++++++++++
 misc/bird.spec  |  2 +-
 sysdep/config.h |  2 +-
 3 files changed, 14 insertions(+), 2 deletions(-)

diff --git a/NEWS b/NEWS
index 5bac8527..7431a809 100644
--- a/NEWS
+++ b/NEWS
@@ -1,3 +1,15 @@
+Version 2.0.2 (2018-03-xx)
+  o Source-specific routing support for Linux kernel and Babel
+  o BGP: New option 'disable after cease'
+  o Filter: Allow silent filter execution
+  o Filter: Fixed stack overflow in BGP mask expressions.
+  o Several bugfixes
+
+  Notes:
+
+  Syntax prefix:netmask for IPv4 prefixes was dropped. Just use prefix/pxlen.
+
+
 Version 2.0.1 (2018-01-16)
   o Linux MPLS kernel support
   o Better handling of channels inherited from templates
diff --git a/misc/bird.spec b/misc/bird.spec
index 7b645655..bf5c14bb 100644
--- a/misc/bird.spec
+++ b/misc/bird.spec
@@ -1,6 +1,6 @@
 Summary: BIRD Internet Routing Daemon
 Name: bird
-Version: 2.0.1
+Version: 2.0.2
 Release: 1
 Copyright: GPL
 Group: Networking/Daemons
diff --git a/sysdep/config.h b/sysdep/config.h
index a552e6b6..393f0281 100644
--- a/sysdep/config.h
+++ b/sysdep/config.h
@@ -13,7 +13,7 @@
 #ifdef GIT_LABEL
 #define BIRD_VERSION XSTR1(GIT_LABEL)
 #else
-#define BIRD_VERSION "2.0.1"
+#define BIRD_VERSION "2.0.2"
 #endif
 
 /* Include parameters determined by configure script */

From 4d3d34f5997128824c376a097eee60954c3611cf Mon Sep 17 00:00:00 2001
From: Ondrej Filip <feela@network.cz>
Date: Thu, 22 Mar 2018 13:30:10 +0100
Subject: [PATCH 150/198] Date added.

---
 NEWS | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/NEWS b/NEWS
index 7431a809..8719a488 100644
--- a/NEWS
+++ b/NEWS
@@ -1,4 +1,4 @@
-Version 2.0.2 (2018-03-xx)
+Version 2.0.2 (2018-03-22)
   o Source-specific routing support for Linux kernel and Babel
   o BGP: New option 'disable after cease'
   o Filter: Allow silent filter execution

From 6807320a0f732829c475ab230857147bf4f0cc7c Mon Sep 17 00:00:00 2001
From: Ondrej Filip <feela@network.cz>
Date: Fri, 23 Mar 2018 17:24:58 +0100
Subject: [PATCH 151/198] Autoconf replaced by autoreconf

---
 tools/gendist | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/gendist b/tools/gendist
index eb2355d3..2ac59030 100755
--- a/tools/gendist
+++ b/tools/gendist
@@ -9,7 +9,7 @@ REL=bird-$VERSION
 DREL=bird-doc-$VERSION
 T=/tmp/bird
 set -e
-AC=`if [ -x /usr/bin/autoconf2.50 ] ; then echo autoconf2.50 ; else echo autoconf ; fi`
+AC=autoreconf
 $AC
 ./configure
 make docs

From 157f6c2aaddbf8668e33ed6efcd874a6677fa005 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Sat, 24 Mar 2018 01:53:03 +0100
Subject: [PATCH 152/198] Doc: Remove some superfluous slashes

---
 doc/bird.sgml | 40 ++++++++++++++++++++--------------------
 1 file changed, 20 insertions(+), 20 deletions(-)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index 903456f0..78569d28 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -2587,17 +2587,17 @@ be used in explicit configuration.
 some of them (marked with `<tt/O/') are optional.
 
 <descrip>
-	<tag><label id="rta-bgp-path">bgppath bgp_path/</tag>
+	<tag><label id="rta-bgp-path">bgppath bgp_path</tag>
 	Sequence of AS numbers describing the AS path the packet will travel
 	through when forwarded according to the particular route. In case of
 	internal BGP it doesn't contain the number of the local AS.
 
-	<tag><label id="rta-bgp-local-pref">int bgp_local_pref/ [I]</tag>
+	<tag><label id="rta-bgp-local-pref">int bgp_local_pref [I]</tag>
 	Local preference value used for selection among multiple BGP routes (see
 	the selection rules above). It's used as an additional metric which is
 	propagated through the whole local AS.
 
-	<tag><label id="rta-bgp-med">int bgp_med/ [O]</tag>
+	<tag><label id="rta-bgp-med">int bgp_med [O]</tag>
 	The Multiple Exit Discriminator of the route is an optional attribute
 	which is used on external (inter-AS) links to convey to an adjacent AS
 	the optimal entry point into the local AS. The received attribute is
@@ -2608,20 +2608,20 @@ some of them (marked with `<tt/O/') are optional.
 	external BGP instance. See <rfc id="4451"> for further discussion of
 	BGP MED attribute.
 
-	<tag><label id="rta-bgp-origin">enum bgp_origin/</tag>
+	<tag><label id="rta-bgp-origin">enum bgp_origin</tag>
 	Origin of the route: either <cf/ORIGIN_IGP/ if the route has originated
 	in an interior routing protocol or <cf/ORIGIN_EGP/ if it's been imported
 	from the <tt>EGP</tt> protocol (nowadays it seems to be obsolete) or
 	<cf/ORIGIN_INCOMPLETE/ if the origin is unknown.
 
-	<tag><label id="rta-bgp-next-hop">ip bgp_next_hop/</tag>
+	<tag><label id="rta-bgp-next-hop">ip bgp_next_hop</tag>
 	Next hop to be used for forwarding of packets to this destination. On
 	internal BGP connections, it's an address of the originating router if
 	it's inside the local AS or a boundary router the packet will leave the
 	AS through if it's an exterior route, so each BGP speaker within the AS
 	has a chance to use the shortest interior path possible to this point.
 
-	<tag><label id="rta-bgp-atomic-aggr">void bgp_atomic_aggr/ [O]</tag>
+	<tag><label id="rta-bgp-atomic-aggr">void bgp_atomic_aggr [O]</tag>
 	This is an optional attribute which carries no value, but the sole
 	presence of which indicates that the route has been aggregated from
 	multiple routes by some router on the path from the originator.
@@ -2629,7 +2629,7 @@ some of them (marked with `<tt/O/') are optional.
 <!-- we don't handle aggregators right since they are of a very obscure type
 	<tag>bgp_aggregator</tag>
 -->
-	<tag><label id="rta-bgp-community">clist bgp_community/ [O]</tag>
+	<tag><label id="rta-bgp-community">clist bgp_community [O]</tag>
 	List of community values associated with the route. Each such value is a
 	pair (represented as a <cf/pair/ data type inside the filters) of 16-bit
 	integers, the first of them containing the number of the AS which
@@ -2640,14 +2640,14 @@ some of them (marked with `<tt/O/') are optional.
 	freedom about which community attributes it defines and what will their
 	semantics be.
 
-	<tag><label id="rta-bgp-ext-community">eclist bgp_ext_community/ [O]</tag>
+	<tag><label id="rta-bgp-ext-community">eclist bgp_ext_community [O]</tag>
 	List of extended community values associated with the route. Extended
 	communities have similar usage as plain communities, but they have an
 	extended range (to allow 4B ASNs) and a nontrivial structure with a type
 	field. Individual community values are represented using an <cf/ec/ data
 	type inside the filters.
 
-	<tag><label id="rta-bgp-large-community">lclist <cf/bgp_large_community/ [O]</tag>
+	<tag><label id="rta-bgp-large-community">lclist bgp_large_community [O]</tag>
 	List of large community values associated with the route. Large BGP
 	communities is another variant of communities, but contrary to extended
 	communities they behave very much the same way as regular communities,
@@ -2655,12 +2655,12 @@ some of them (marked with `<tt/O/') are optional.
 	Individual community values are represented using an <cf/lc/ data type
 	inside the filters.
 
-	<tag><label id="rta-bgp-originator-id">quad bgp_originator_id/ [I, O]</tag>
+	<tag><label id="rta-bgp-originator-id">quad bgp_originator_id [I, O]</tag>
 	This attribute is created by the route reflector when reflecting the
 	route and contains the router ID of the originator of the route in the
 	local AS.
 
-	<tag><label id="rta-bgp-cluster-list">clist bgp_cluster_list/ [I, O]</tag>
+	<tag><label id="rta-bgp-cluster-list">clist bgp_cluster_list [I, O]</tag>
 	This attribute contains a list of cluster IDs of route reflectors. Each
 	route reflector prepends its cluster ID when reflecting the route.
 </descrip>
@@ -2922,26 +2922,26 @@ translated to appropriate system (and OS-specific) route attributes. We support
 these attributes:
 
 <descrip>
-	<tag><label id="rta-krt-source">int krt_source/</tag>
+	<tag><label id="rta-krt-source">int krt_source</tag>
 	The original source of the imported kernel route. The value is
 	system-dependent. On Linux, it is a value of the protocol field of the
 	route. See /etc/iproute2/rt_protos for common values. On BSD, it is
 	based on STATIC and PROTOx flags. The attribute is read-only.
 
-	<tag><label id="rta-krt-metric">int krt_metric/</tag> (Linux)
+	<tag><label id="rta-krt-metric">int krt_metric</tag> (Linux)
 	The kernel metric of the route. When multiple same routes are in a
 	kernel routing table, the Linux kernel chooses one with lower metric.
 	Note that preferred way to set kernel metric is to use protocol option
 	<cf/metric/, unless per-route metric values are needed.
 
-	<tag><label id="rta-krt-prefsrc">ip krt_prefsrc/</tag> (Linux)
+	<tag><label id="rta-krt-prefsrc">ip krt_prefsrc</tag> (Linux)
 	The preferred source address. Used in source address selection for
 	outgoing packets. Has to be one of the IP addresses of the router.
 
-	<tag><label id="rta-krt-realm">int krt_realm/</tag> (Linux)
+	<tag><label id="rta-krt-realm">int krt_realm</tag> (Linux)
 	The realm of the route. Can be used for traffic classification.
 
-	<tag><label id="rta-krt-scope">int krt_scope/</tag> (Linux IPv4)
+	<tag><label id="rta-krt-scope">int krt_scope</tag> (Linux IPv4)
 	The scope of the route. Valid values are 0-254, although Linux kernel
 	may reject some values depending on route type and nexthop. It is
 	supposed to represent `indirectness' of the route, where nexthops of
@@ -3963,13 +3963,13 @@ definitions, prefix definitions and DNS definitions:
 <p>RAdv defines two route attributes:
 
 <descrip>
-	<tag><label id="rta-ra-preference">enum ra_preference/</tag>
+	<tag><label id="rta-ra-preference">enum ra_preference</tag>
 	The preference of the route. The value can be <it/RA_PREF_LOW/,
 	<it/RA_PREF_MEDIUM/ or <it/RA_PREF_HIGH/. If the attribute is not set,
 	the <ref id="radv-iface-route-preference" name="route preference">
 	option is used.
 
-	<tag><label id="rta-ra-lifetime">int ra_lifetime/</tag>
+	<tag><label id="rta-ra-lifetime">int ra_lifetime</tag>
 	The advertised lifetime of the route, in seconds. The special value of
 	0xffffffff represents infinity. If the attribute is not set, the
 	<ref id="radv-iface-route-lifetime" name="route lifetime">
@@ -4275,13 +4275,13 @@ protocol rip [ng] [&lt;name&gt;] {
 <p>RIP defines two route attributes:
 
 <descrip>
-	<tag>int <cf/rip_metric/</tag>
+	<tag><label id="rta-rip-metric">int rip_metric</tag>
 	RIP metric of the route (ranging from 0 to <cf/infinity/). When routes
 	from different RIP instances are available and all of them have the same
 	preference, BIRD prefers the route with lowest <cf/rip_metric/. When a
 	non-RIP route is exported to RIP, the default metric is 1.
 
-	<tag><label id="rta-rip-tag">int rip_tag/</tag>
+	<tag><label id="rta-rip-tag">int rip_tag</tag>
 	RIP route tag: a 16-bit number which can be used to carry additional
 	information with the route (for example, an originating AS number in
 	case of external routes). When a non-RIP route is exported to RIP, the

From a63d20aa872364873d0fbfa03d1c3c6d9a1889eb Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 3 Apr 2018 16:53:58 +0200
Subject: [PATCH 153/198] Doc: Documentation for BGP extended next hop feature

Thanks to Arvin Gan for the bugreport.
---
 doc/bird.sgml | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index 78569d28..fde47e65 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -2562,6 +2562,15 @@ 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-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
+	IPv4 next hops with IPv6 prefixes and vice versa. For IPv4 / VPNv4
+	channels, the behavior is controlled by the Extended Next Hop Encoding
+	capability, as described in <rfc id="5549">. For IPv6 / VPNv6 channels,
+	just IPv4-mapped IPv6 addresses are used, as described in
+	<rfc id="4798"> and <rfc id="4659">. Default: off.
+
 	<tag><label id="bgp-add-paths">add paths <m/switch/|rx|tx</tag>
 	Standard BGP can propagate only one path (route) per destination network
 	(usually the selected one). This option controls the add-path protocol

From c408d807a374b521dc66c434ca0dc5987820646f Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 3 Apr 2018 17:31:45 +0200
Subject: [PATCH 154/198] Doc: Documentation for BGP disable after cease option

---
 doc/bird.sgml | 13 +++++++++++++
 1 file changed, 13 insertions(+)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index fde47e65..1191fa03 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -2384,6 +2384,19 @@ using the following configuration parameters:
 	disable the instance automatically and wait for an administrator to fix
 	the problem manually. Default: off.
 
+	<tag><label id="bgp-disable-after-cease">disable after cease <m/switch/|<m/set-of-flags/</tag>
+	When a Cease notification is received, disable the instance
+	automatically and wait for an administrator to fix the problem manually.
+	When used with <m/switch/ argument, it means handle every Cease subtype
+	with the exception of <cf/connection collision/. Default: off.
+
+	The <m/set-of-flags/ allows to narrow down relevant Cease subtypes. The
+	syntax is <cf>{<m/flag/ [, <m/.../] }</cf>, where flags are: <cf/cease/,
+	<cf/prefix limit hit/, <cf/administrative shutdown/,
+	<cf/peer deconfigured/, <cf/administrative reset/,
+	<cf/connection rejected/, <cf/configuration change/,
+	<cf/connection collision/, <cf/out of resources/.
+
 	<tag><label id="bgp-hold-time">hold time <m/number/</tag>
 	Time in seconds to wait for a Keepalive message from the other side
 	before considering the connection stale. Default: depends on agreement

From f3a8cf050e6181e158dcde2fe885d7bf220eedc3 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Thu, 12 Apr 2018 16:55:56 +0200
Subject: [PATCH 155/198] BGP: Fix extended next hop handling

For IPv4 with extended next hop, we use MP-BGP format and therefore no
independent NEXT_HOP attribute.

Thanks to Arvin Gan for the bugreport.
---
 proto/bgp/attrs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/proto/bgp/attrs.c b/proto/bgp/attrs.c
index 0f41f818..9003feb2 100644
--- a/proto/bgp/attrs.c
+++ b/proto/bgp/attrs.c
@@ -295,7 +295,7 @@ bgp_encode_next_hop(struct bgp_write_state *s, eattr *a, byte *buf, uint size)
    * store it and encode it later by AFI-specific hooks.
    */
 
-  if (s->channel->afi == BGP_AF_IPV4)
+  if ((s->channel->afi == BGP_AF_IPV4) && !s->channel->ext_next_hop)
   {
     ASSERT(a->u.ptr->length == sizeof(ip_addr));
 

From 4727d1db9d83a8f1025481cbcc06a7e4c8ec9f33 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 25 Apr 2018 15:50:57 +0200
Subject: [PATCH 156/198] OSPF: Support of authentication trailer for OSPFv3

Implement RFC 7166, crypthographic authentication for OSPFv3
analogous to authentication used for OSPFv2.
---
 proto/ospf/config.Y   |  15 +--
 proto/ospf/dbdes.c    |  12 ++-
 proto/ospf/hello.c    |  13 ++-
 proto/ospf/iface.c    |  13 ++-
 proto/ospf/neighbor.c |   2 -
 proto/ospf/ospf.h     |  40 ++++++--
 proto/ospf/packet.c   | 216 +++++++++++++++++++++++++++++++++++++-----
 7 files changed, 265 insertions(+), 46 deletions(-)

diff --git a/proto/ospf/config.Y b/proto/ospf/config.Y
index 005f4381..0b09966d 100644
--- a/proto/ospf/config.Y
+++ b/proto/ospf/config.Y
@@ -37,7 +37,7 @@ ospf_iface_finish(void)
 
   ip->passwords = get_passwords();
 
-  if ((ip->autype == OSPF_AUTH_CRYPT) && (ip->helloint < 5))
+  if (ospf_cfg_is_v2() && (ip->autype == OSPF_AUTH_CRYPT) && (ip->helloint < 5))
     log(L_WARN "Hello or poll interval less that 5 makes cryptographic authenication prone to replay attacks");
 
   if ((ip->autype == OSPF_AUTH_NONE) && (ip->passwords != NULL))
@@ -54,6 +54,9 @@ ospf_iface_finish(void)
       /* Set default OSPF crypto algorithms */
       if (!pass->alg && (ip->autype == OSPF_AUTH_CRYPT))
 	pass->alg = ospf_cfg_is_v2() ? ALG_MD5 : ALG_HMAC_SHA256;
+
+      if (ospf_cfg_is_v3() && ip->autype && (pass->alg < ALG_HMAC))
+	cf_error("Keyed hash algorithms are not allowed, use HMAC algorithms");
     }
   }
 }
@@ -181,7 +184,7 @@ static inline void
 ospf_check_auth(void)
 {
   if (ospf_cfg_is_v3())
-    cf_error("Authentication not supported in OSPFv3");
+    cf_error("Plaintext authentication not supported in OSPFv3");
 }
 
 
@@ -346,8 +349,8 @@ ospf_vlink_item:
  | DEAD COUNT expr { OSPF_PATT->deadc = $3 ; if ($3<=1) cf_error("Dead count must be greater than one"); }
  | AUTHENTICATION NONE { OSPF_PATT->autype = OSPF_AUTH_NONE;  }
  | AUTHENTICATION SIMPLE { OSPF_PATT->autype = OSPF_AUTH_SIMPLE; ospf_check_auth(); }
- | AUTHENTICATION CRYPTOGRAPHIC { OSPF_PATT->autype = OSPF_AUTH_CRYPT; ospf_check_auth(); }
- | password_list { ospf_check_auth(); }
+ | AUTHENTICATION CRYPTOGRAPHIC { OSPF_PATT->autype = OSPF_AUTH_CRYPT; }
+ | password_list
  ;
 
 ospf_vlink_start: VIRTUAL LINK idval
@@ -396,7 +399,7 @@ ospf_iface_item:
  | NEIGHBORS '{' nbma_list '}'
  | AUTHENTICATION NONE { OSPF_PATT->autype = OSPF_AUTH_NONE; }
  | AUTHENTICATION SIMPLE { OSPF_PATT->autype = OSPF_AUTH_SIMPLE; ospf_check_auth(); }
- | AUTHENTICATION CRYPTOGRAPHIC { OSPF_PATT->autype = OSPF_AUTH_CRYPT; ospf_check_auth(); }
+ | AUTHENTICATION CRYPTOGRAPHIC { OSPF_PATT->autype = OSPF_AUTH_CRYPT; }
  | RX BUFFER NORMAL { OSPF_PATT->rx_buffer = 0; }
  | RX BUFFER LARGE { OSPF_PATT->rx_buffer = OSPF_MAX_PKT_SIZE; }
  | RX BUFFER expr { OSPF_PATT->rx_buffer = $3; if (($3 < OSPF_MIN_PKT_SIZE) || ($3 > OSPF_MAX_PKT_SIZE)) cf_error("Buffer size must be in range 256-65535"); }
@@ -406,7 +409,7 @@ ospf_iface_item:
  | TTL SECURITY bool { OSPF_PATT->ttl_security = $3; }
  | TTL SECURITY TX ONLY { OSPF_PATT->ttl_security = 2; }
  | BFD bool { OSPF_PATT->bfd = $2; cf_check_bfd($2); }
- | password_list { ospf_check_auth(); }
+ | password_list
  ;
 
 pref_list:
diff --git a/proto/ospf/dbdes.c b/proto/ospf/dbdes.c
index f211935f..2626a24c 100644
--- a/proto/ospf/dbdes.c
+++ b/proto/ospf/dbdes.c
@@ -14,7 +14,7 @@
 struct ospf_dbdes2_packet
 {
   struct ospf_packet hdr;
-  union ospf_auth auth;
+  union ospf_auth2 auth;
 
   u16 iface_mtu;
   u8 options;
@@ -38,6 +38,13 @@ struct ospf_dbdes3_packet
 };
 
 
+uint
+ospf_dbdes3_options(struct ospf_packet *pkt)
+{
+  struct ospf_dbdes3_packet *ps = (void *) pkt;
+  return ntohl(ps->options);
+}
+
 static inline uint
 ospf_dbdes_hdrlen(struct ospf_proto *p)
 {
@@ -45,7 +52,6 @@ ospf_dbdes_hdrlen(struct ospf_proto *p)
     sizeof(struct ospf_dbdes2_packet) : sizeof(struct ospf_dbdes3_packet);
 }
 
-
 static void
 ospf_dbdes_body(struct ospf_proto *p, struct ospf_packet *pkt,
 		struct ospf_lsa_header **body, uint *count)
@@ -129,7 +135,7 @@ ospf_prepare_dbdes(struct ospf_proto *p, struct ospf_neighbor *n)
   else /* OSPFv3 */
   {
     struct ospf_dbdes3_packet *ps = (void *) pkt;
-    ps->options = htonl(ifa->oa->options);
+    ps->options = htonl(ifa->oa->options | (ifa->autype == OSPF_AUTH_CRYPT ? OPT_AT : 0));
     ps->iface_mtu = htons(iface_mtu);
     ps->padding = 0;
     ps->imms = 0;	/* Will be set later */
diff --git a/proto/ospf/hello.c b/proto/ospf/hello.c
index e706ea0f..523c24ad 100644
--- a/proto/ospf/hello.c
+++ b/proto/ospf/hello.c
@@ -14,7 +14,7 @@
 struct ospf_hello2_packet
 {
   struct ospf_packet hdr;
-  union ospf_auth auth;
+  union ospf_auth2 auth;
 
   u32 netmask;
   u16 helloint;
@@ -42,6 +42,13 @@ struct ospf_hello3_packet
 };
 
 
+uint
+ospf_hello3_options(struct ospf_packet *pkt)
+{
+  struct ospf_hello3_packet *ps = (void *) pkt;
+  return ntohl(ps->options) & 0x00FFFFFF;
+}
+
 void
 ospf_send_hello(struct ospf_iface *ifa, int kind, struct ospf_neighbor *dirn)
 {
@@ -86,9 +93,10 @@ ospf_send_hello(struct ospf_iface *ifa, int kind, struct ospf_neighbor *dirn)
   else
   {
     struct ospf_hello3_packet *ps = (void *) pkt;
+    u32 options = ifa->oa->options | (ifa->autype == OSPF_AUTH_CRYPT ? OPT_AT : 0);
 
     ps->iface_id = htonl(ifa->iface_id);
-    ps->options = ntohl(ifa->oa->options | (ifa->priority << 24));
+    ps->options = ntohl(options | (ifa->priority << 24));
     ps->helloint = ntohs(ifa->helloint);
     ps->deadint = htons(ifa->deadint);
     ps->dr = htonl(ifa->drid);
@@ -334,7 +342,6 @@ ospf_receive_hello(struct ospf_packet *pkt, struct ospf_iface *ifa,
   n->priority = rcv_priority;
   n->iface_id = rcv_iface_id;
 
-
   /* Update inactivity timer */
   ospf_neigh_sm(n, INM_HELLOREC);
 
diff --git a/proto/ospf/iface.c b/proto/ospf/iface.c
index e3d8d61b..19bcfa15 100644
--- a/proto/ospf/iface.c
+++ b/proto/ospf/iface.c
@@ -61,7 +61,10 @@ ifa_tx_hdrlen(struct ospf_iface *ifa)
 
   /* Relevant just for OSPFv2 */
   if (ifa->autype == OSPF_AUTH_CRYPT)
+  {
+    hlen += ospf_is_v2(p) ? 0 : sizeof(struct ospf_auth3);
     hlen += max_mac_length(ifa->passwords);
+  }
 
   return hlen;
 }
@@ -137,7 +140,7 @@ ospf_sk_open(struct ospf_iface *ifa)
     goto err;
 
   /* 12 is an offset of the checksum in an OSPFv3 packet */
-  if (ospf_is_v3(p))
+  if (ospf_is_v3(p) && !ifa->autype)
     if (sk_set_ipv6_checksum(sk, 12) < 0)
       goto err;
 
@@ -828,6 +831,14 @@ ospf_iface_reconfigure(struct ospf_iface *ifa, struct ospf_iface_patt *new)
   {
     OSPF_TRACE(D_EVENTS, "Changing authentication type of %s", ifname);
     ifa->autype = new->autype;
+
+    /* For OSPFv3, we need to update checksum calculation by OS */
+    if (ospf_is_v3(p) && ifa->sk)
+      if (sk_set_ipv6_checksum(ifa->sk, ifa->autype ? -1 : 12) < 0)
+      {
+	sk_log_error(ifa->sk, p->p.name);
+	return 0;
+      }
   }
 
   /* Update passwords */
diff --git a/proto/ospf/neighbor.c b/proto/ospf/neighbor.c
index f2d3505e..7ce682b0 100644
--- a/proto/ospf/neighbor.c
+++ b/proto/ospf/neighbor.c
@@ -84,8 +84,6 @@ ospf_neighbor_new(struct ospf_iface *ifa)
   n->pool = pool;
   n->ifa = ifa;
   add_tail(&ifa->neigh_list, NODE n);
-  n->adj = 0;
-  n->csn = 0;
   n->state = NEIGHBOR_DOWN;
 
   init_lists(p, n);
diff --git a/proto/ospf/ospf.h b/proto/ospf/ospf.h
index 54eeb74c..f26ed99c 100644
--- a/proto/ospf/ospf.h
+++ b/proto/ospf/ospf.h
@@ -69,6 +69,9 @@
 #define MINLSARRIVAL		(1 S_)
 #define LSINFINITY		0xffffff
 
+#define OSPF_PKT_TYPES		5	/* HELLO_P .. LSACK_P */
+#define OSPF3_CRYPTO_ID		1	/* OSPFv3 Cryptographic Protocol ID */
+
 #define OSPF_DEFAULT_TICK 1
 #define OSPF_DEFAULT_STUB_COST 1000
 #define OSPF_DEFAULT_ECMP_LIMIT 16
@@ -222,6 +225,7 @@ struct ospf_proto
   u8 merge_external;		/* Should i merge external routes? */
   u8 asbr;			/* May i originate any ext/NSSA lsa? */
   u8 ecmp;			/* Maximal number of nexthops in ECMP route, or 0 */
+  u64 csn64;			/* Last used cryptographic sequence number */
   struct ospf_area *backbone;	/* If exists */
   event *flood_event;		/* Event for flooding LS updates */
   void *lsab;			/* LSA buffer used when originating router LSAs */
@@ -254,8 +258,6 @@ struct ospf_area
   struct fib rtr;		/* Routing tables for routers */
 };
 
-
-
 struct ospf_iface
 {
   node n;
@@ -387,7 +389,8 @@ struct ospf_neighbor
   struct bfd_request *bfd_req;	/* BFD request, if BFD is used */
   void *ldd_buffer;		/* Last database description packet */
   u32 ldd_bsize;		/* Buffer size for ldd_buffer */
-  u32 csn;                      /* Last received crypt seq number (for MD5) */
+  u32 csn;			/* OSPFv2: Last received crypt seq number */
+  u64 csn64[OSPF_PKT_TYPES];	/* OSPFv3: Last received CSN for each type of packet */
 };
 
 
@@ -422,6 +425,7 @@ struct ospf_neighbor
 #define OSPF_AUTH_SIMPLE 1
 #define OSPF_AUTH_CRYPT	2
 
+#define OSPF3_AUTH_HMAC 1	/* HMAC Cryptographic Authentication */
 
 /* OSPF neighbor states */
 #define NEIGHBOR_DOWN	0
@@ -459,10 +463,12 @@ struct ospf_neighbor
 #define OPT_MC		0x0004	/* Related to MOSPF, not used and obsolete */
 #define OPT_N		0x0008	/* Related to NSSA */
 #define OPT_P		0x0008	/* OSPFv2, flags P and N share position, see NSSA RFC */
-#define OPT_EA		0x0010	/* OSPFv2, external attributes, not used and obsolete */
+#define OPT_L_V2	0x0010	/* OSPFv2, link-local signaling, not used */
 #define OPT_R		0x0010	/* OSPFv3, originator is active router */
 #define OPT_DC		0x0020	/* Related to demand circuits, not used */
 #define OPT_AF		0x0100	/* OSPFv3 Address Families (RFC 5838) */
+#define OPT_L_V3	0x0200	/* OSPFv3, link-local signaling */
+#define OPT_AT          0x0400	/* OSPFv3, authentication trailer */
 
 /* Router-LSA VEB flags are are stored together with links (OSPFv2) or options (OSPFv3) */
 #define OPT_RT_B	(0x01 << 24)
@@ -489,20 +495,38 @@ struct ospf_packet
   u8 autype;			/* Undefined for OSPFv3 */
 };
 
+struct ospf_lls
+{
+  u16 checksum;
+  u16 length;
+  byte data[0];
+};
+
 struct ospf_auth_crypto
 {
   u16 zero;
   u8 keyid;
   u8 len;
-  u32 csn;
+  u32 csn;			/* Cryptographic sequence number (32-bit) */
 };
 
-union ospf_auth
+union ospf_auth2
 {
   u8 password[8];
   struct ospf_auth_crypto c32;
 };
 
+struct ospf_auth3
+{
+  u16 type;			/* Authentication type (OSPF3_AUTH_*) */
+  u16 length;			/* Authentication trailer length (header + data) */
+  u16 reserved;
+  u16 sa_id;			/* Security association identifier (key_id) */
+  u64 csn;			/* Cryptographic sequence number (64-bit) */
+  byte data[0];			/* Authentication data */
+};
+
+
 /* Packet types */
 #define HELLO_P		1	/* Hello */
 #define DBDES_P		2	/* Database description */
@@ -957,7 +981,7 @@ static inline void ospf_send_to_des(struct ospf_iface *ifa)
 #endif
 
 static inline uint ospf_pkt_hdrlen(struct ospf_proto *p)
-{ return ospf_is_v2(p) ? (sizeof(struct ospf_packet) + sizeof(union ospf_auth)) : sizeof(struct ospf_packet); }
+{ return ospf_is_v2(p) ? (sizeof(struct ospf_packet) + sizeof(union ospf_auth2)) : sizeof(struct ospf_packet); }
 
 static inline void * ospf_tx_buffer(struct ospf_iface *ifa)
 { return ifa->sk->tbuf; }
@@ -969,11 +993,13 @@ static inline void * ospf_tx_buffer(struct ospf_iface *ifa)
 
 void ospf_send_hello(struct ospf_iface *ifa, int kind, struct ospf_neighbor *dirn);
 void ospf_receive_hello(struct ospf_packet *pkt, struct ospf_iface *ifa, struct ospf_neighbor *n, ip_addr faddr);
+uint ospf_hello3_options(struct ospf_packet *pkt);
 
 /* dbdes.c */
 void ospf_send_dbdes(struct ospf_proto *p, struct ospf_neighbor *n);
 void ospf_rxmt_dbdes(struct ospf_proto *p, struct ospf_neighbor *n);
 void ospf_receive_dbdes(struct ospf_packet *pkt, struct ospf_iface *ifa, struct ospf_neighbor *n);
+uint ospf_dbdes3_options(struct ospf_packet *pkt);
 
 /* lsreq.c */
 void ospf_send_lsreq(struct ospf_proto *p, struct ospf_neighbor *n);
diff --git a/proto/ospf/packet.c b/proto/ospf/packet.c
index 38d7a75f..db387661 100644
--- a/proto/ospf/packet.c
+++ b/proto/ospf/packet.c
@@ -29,23 +29,24 @@ ospf_pkt_fill_hdr(struct ospf_iface *ifa, void *buf, u8 h_type)
   pkt->areaid = htonl(ifa->oa->areaid);
   pkt->checksum = 0;
   pkt->instance_id = ifa->instance_id;
-  pkt->autype = ifa->autype;
+  pkt->autype = ospf_is_v2(p) ? ifa->autype : 0;
 }
 
 /* We assume OSPFv2 in ospf_pkt_finalize() */
 static void
-ospf_pkt_finalize(struct ospf_iface *ifa, struct ospf_packet *pkt, uint *plen)
+ospf_pkt_finalize2(struct ospf_iface *ifa, struct ospf_packet *pkt, uint *plen)
 {
+  struct ospf_proto *p = ifa->oa->po;
   struct password_item *pass = NULL;
-  union ospf_auth *auth = (void *) (pkt + 1);
-
-  pkt->checksum = 0;
-  pkt->autype = ifa->autype;
-  bzero(auth, sizeof(union ospf_auth));
+  union ospf_auth2 *auth = (void *) (pkt + 1);
+  memset(auth, 0, sizeof(union ospf_auth2));
 
   /* Compatibility note: auth may contain anything if autype is
      none, but nonzero values do not work with Mikrotik OSPF */
 
+  pkt->checksum = 0;
+  pkt->autype = ifa->autype;
+
   switch (ifa->autype)
   {
   case OSPF_AUTH_SIMPLE:
@@ -60,7 +61,7 @@ ospf_pkt_finalize(struct ospf_iface *ifa, struct ospf_packet *pkt, uint *plen)
   case OSPF_AUTH_NONE:
     {
       void *body = (void *) (auth + 1);
-      uint blen = *plen - sizeof(struct ospf_packet) - sizeof(union ospf_auth);
+      uint blen = *plen - sizeof(struct ospf_packet) - sizeof(union ospf_auth2);
       pkt->checksum = ipsum_calculate(pkt, sizeof(struct ospf_packet), body, blen, NULL);
     }
     break;
@@ -69,7 +70,7 @@ ospf_pkt_finalize(struct ospf_iface *ifa, struct ospf_packet *pkt, uint *plen)
     pass = password_find(ifa->passwords, 0);
     if (!pass)
     {
-      log(L_ERR "No suitable password found for authentication");
+      log(L_ERR "%s: No suitable password found for authentication", p->p.name);
       return;
     }
 
@@ -105,8 +106,7 @@ ospf_pkt_finalize(struct ospf_iface *ifa, struct ospf_packet *pkt, uint *plen)
     else
       memset32(auth_tail, HMAC_MAGIC, auth_len / 4);
 
-    mac_fill(pass->alg, pass->password, pass->length,
-	     (byte *) pkt, *plen, auth_tail);
+    mac_fill(pass->alg, pass->password, pass->length, (byte *) pkt, *plen, auth_tail);
     break;
 
   default:
@@ -114,13 +114,63 @@ ospf_pkt_finalize(struct ospf_iface *ifa, struct ospf_packet *pkt, uint *plen)
   }
 }
 
-
-/* We assume OSPFv2 in ospf_pkt_checkauth() */
-static int
-ospf_pkt_checkauth(struct ospf_neighbor *n, struct ospf_iface *ifa, struct ospf_packet *pkt, uint len)
+/*
+ * Return an extra packet size that should be added to a final packet size
+ */
+static void
+ospf_pkt_finalize3(struct ospf_iface *ifa, struct ospf_packet *pkt, uint *plen, ip_addr src)
 {
   struct ospf_proto *p = ifa->oa->po;
-  union ospf_auth *auth = (void *) (pkt + 1);
+  struct ospf_auth3 *auth = (void *) ((byte *) pkt + *plen);
+
+  pkt->checksum = 0;
+  pkt->autype = 0;
+
+  if (ifa->autype != OSPF_AUTH_CRYPT)
+    return;
+
+  struct password_item *pass = password_find(ifa->passwords, 0);
+  if (!pass)
+  {
+    log(L_ERR "%s: No suitable password found for authentication", p->p.name);
+    return;
+  }
+
+  /* FIXME: Ensure persistence */
+  p->csn64++;
+
+  uint mac_len = mac_type_length(pass->alg);
+  uint auth_len = sizeof(struct ospf_auth3) + mac_len;
+  *plen += auth_len;
+
+  ASSERT(*plen < ifa->sk->tbsize);
+
+  memset(auth, 0, sizeof(struct ospf_auth3));
+  auth->type = htons(OSPF3_AUTH_HMAC);
+  auth->length = htons(auth_len);
+  auth->reserved = 0;
+  auth->sa_id = htons(pass->id);
+  put_u64(&auth->csn, p->csn64);
+
+  /* Initialize with src IP address padded with HMAC_MAGIC */
+  put_ip6(auth->data, ipa_to_ip6(src));
+  memset32(auth->data + 16, HMAC_MAGIC, (mac_len - 16) / 4);
+
+  /* Attach OSPFv3 Cryptographic Protocol ID to the key */
+  uint pass_len = pass->length + 2;
+  byte *pass_key = alloca(pass_len);
+  memcpy(pass_key, pass->password, pass->length);
+  put_u16(pass_key + pass->length, OSPF3_CRYPTO_ID);
+
+  mac_fill(pass->alg, pass_key, pass_len, (byte *) pkt, *plen, auth->data);
+}
+
+
+static int
+ospf_pkt_checkauth2(struct ospf_neighbor *n, struct ospf_iface *ifa, struct ospf_packet *pkt, uint len)
+{
+  struct ospf_proto *p = ifa->oa->po;
+  union ospf_auth2 *auth = (void *) (pkt + 1);
   struct password_item *pass = NULL;
   const char *err_dsc = NULL;
   uint err_val = 0;
@@ -200,6 +250,119 @@ drop:
   return 0;
 }
 
+static int
+ospf_pkt_checkauth3(struct ospf_neighbor *n, struct ospf_iface *ifa, struct ospf_packet *pkt, uint len, ip_addr src)
+{
+  struct ospf_proto *p = ifa->oa->po;
+  const char *err_dsc = NULL;
+  uint err_val = 0;
+
+  uint plen = ntohs(pkt->length);
+  uint opts, lls_present, auth_present;
+
+  /*
+   * When autentication is not enabled, ignore the trailer. This is different
+   * from OSPFv2, but it is necessary in order to support migration modes. Note
+   * that regular authenticated packets do not have valid checksum and will be
+   * dropped by OS on non-authenticated ifaces.
+   */
+  if (ifa->autype != OSPF_AUTH_CRYPT)
+    return 1;
+
+  switch(pkt->type)
+  {
+  case HELLO_P:
+    opts = ospf_hello3_options(pkt);
+    lls_present = opts & OPT_L_V3;
+    auth_present = opts & OPT_AT;
+    break;
+
+  case DBDES_P:
+    opts = ospf_dbdes3_options(pkt);
+    lls_present = opts & OPT_L_V3;
+    auth_present = opts & OPT_AT;
+    break;
+
+  default:
+    lls_present = 0;
+    auth_present = n->options & OPT_AT;
+  }
+
+  if (!auth_present)
+    DROP1("missing authentication trailer");
+
+  if (lls_present)
+  {
+    if ((plen + sizeof(struct ospf_lls)) > len)
+      DROP("packet length mismatch", len);
+
+    struct ospf_lls *lls = (void *) ((byte *) pkt + plen);
+    plen += ntohs(lls->length);
+  }
+
+  if ((plen + sizeof(struct ospf_auth3)) > len)
+    DROP("packet length mismatch", len);
+
+  struct ospf_auth3 *auth = (void *) ((byte *) pkt + plen);
+
+  uint rcv_auth_type = ntohs(auth->type);
+  if (rcv_auth_type != OSPF3_AUTH_HMAC)
+    DROP("authentication method mismatch", rcv_auth_type);
+
+  uint rcv_auth_len = ntohs(auth->length);
+  if (plen + rcv_auth_len > len)
+    DROP("packet length mismatch", len);
+
+  uint rcv_key_id = ntohs(auth->sa_id);
+  struct password_item *pass = password_find_by_id(ifa->passwords, rcv_key_id);
+  if (!pass)
+    DROP("no suitable password found", rcv_key_id);
+
+  uint mac_len = mac_type_length(pass->alg);
+  if (rcv_auth_len != (sizeof(struct ospf_auth3) + mac_len))
+    DROP("wrong authentication length", rcv_auth_len);
+
+  uint pt = pkt->type - 1;
+  u64 rcv_csn = get_u64(&auth->csn);
+  if (n && (rcv_csn <= n->csn64[pt]))
+  {
+    /* We want to report both new and old CSN */
+    LOG_PKT_AUTH("Authentication failed for nbr %R on %s - "
+		 "lower sequence number (rcv %u, old %u)",
+		 n->rid, ifa->ifname, (uint) rcv_csn, (uint) n->csn64[pt]);
+    return 0;
+  }
+
+  /* Save the received authentication data */
+  byte *auth_data = alloca(mac_len);
+  memcpy(auth_data, auth->data, mac_len);
+
+  /* Initialize with src IP address padded with HMAC_MAGIC */
+  put_ip6(auth->data, ipa_to_ip6(src));
+  memset32(auth->data + 16, HMAC_MAGIC, (mac_len - 16) / 4);
+
+  /* Attach OSPFv3 Cryptographic Protocol ID to the key */
+  uint pass_len = pass->length + 2;
+  byte *pass_key = alloca(pass_len);
+  memcpy(pass_key, pass->password, pass->length);
+  put_u16(pass_key + pass->length, OSPF3_CRYPTO_ID);
+
+  if (!mac_verify(pass->alg, pass_key, pass_len,
+		  (byte *) pkt, plen + rcv_auth_len, auth_data))
+    DROP("wrong authentication code", pass->id);
+
+  if (n)
+    n->csn64[pt] = rcv_csn;
+
+  return 1;
+
+drop:
+  LOG_PKT_AUTH("Authentication failed for nbr %R on %s - %s (%u)",
+	       (n ? n->rid : ntohl(pkt->routerid)), ifa->ifname, err_dsc, err_val);
+
+  return 0;
+}
+
 /**
  * ospf_rx_hook
  * @sk: socket we received the packet.
@@ -298,12 +461,12 @@ ospf_rx_hook(sock *sk, uint len)
 
   if (ospf_is_v2(p) && (pkt->autype != OSPF_AUTH_CRYPT))
   {
-    uint hlen = sizeof(struct ospf_packet) + sizeof(union ospf_auth);
+    uint hlen = sizeof(struct ospf_packet) + sizeof(union ospf_auth2);
     uint blen = plen - hlen;
     void *body = ((void *) pkt) + hlen;
 
     if (!ipsum_verify(pkt, sizeof(struct ospf_packet), body, blen, NULL))
-      DROP1("invalid checksum");
+      DROP("invalid checksum", ntohs(pkt->checksum));
   }
 
   /* Third, we resolve associated iface and handle vlinks. */
@@ -401,8 +564,14 @@ found:
     return 1;
   }
 
+  /* Check packet type here, ospf_pkt_checkauth3() expects valid values */
+  if (pkt->type < HELLO_P || pkt->type > LSACK_P)
+    DROP("invalid packet type", pkt->type);
+
   /* ospf_pkt_checkauth() has its own error logging */
-  if (ospf_is_v2(p) && !ospf_pkt_checkauth(n, ifa, pkt, len))
+  if ((ospf_is_v2(p) ?
+       !ospf_pkt_checkauth2(n, ifa, pkt, len) :
+       !ospf_pkt_checkauth3(n, ifa, pkt, len, sk->faddr)))
     return 1;
 
   switch (pkt->type)
@@ -426,9 +595,6 @@ found:
   case LSACK_P:
     ospf_receive_lsack(pkt, ifa, n);
     break;
-
-  default:
-    DROP("invalid packet type", pkt->type);
   };
   return 1;
 
@@ -452,7 +618,7 @@ ospf_tx_hook(sock * sk)
 void
 ospf_err_hook(sock * sk, int err)
 {
-  struct ospf_iface *ifa= (struct ospf_iface *) (sk->data);
+  struct ospf_iface *ifa = (struct ospf_iface *) (sk->data);
   struct ospf_proto *p = ifa->oa->po;
   log(L_ERR "%s: Socket error on %s: %M", p->p.name, ifa->ifname, err);
 }
@@ -472,7 +638,9 @@ ospf_send_to(struct ospf_iface *ifa, ip_addr dst)
   uint plen = ntohs(pkt->length);
 
   if (ospf_is_v2(ifa->oa->po))
-    ospf_pkt_finalize(ifa, pkt, &plen);
+    ospf_pkt_finalize2(ifa, pkt, &plen);
+  else
+    ospf_pkt_finalize3(ifa, pkt, &plen, sk->saddr);
 
   int done = sk_send_to(sk, plen, dst, 0);
   if (!done)

From 823ad12191e66e243dd088a81c66e4a518563e40 Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Fri, 27 Apr 2018 14:38:41 +0200
Subject: [PATCH 157/198] Filter: Added missing instruction comparators.

These instructions caused SIGABORTs on reconfiguration.
---
 filter/filter.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/filter/filter.c b/filter/filter.c
index bb3146e7..881ba420 100644
--- a/filter/filter.c
+++ b/filter/filter.c
@@ -1712,7 +1712,10 @@ i_same(struct f_inst *f1, struct f_inst *f2)
   case FI_EA_SET: ONEARG; A2_SAME; break;
 
   case FI_RETURN: ONEARG; break;
+  case FI_ROA_MAXLEN: ONEARG; break;
+  case FI_ROA_ASN: ONEARG; break;
   case FI_IP: ONEARG; break;
+  case FI_IS_V4: ONEARG; break;
   case FI_ROUTE_DISTINGUISHER: ONEARG; break;
   case FI_CALL: /* Call rewriting trickery to avoid exponential behaviour */
              ONEARG;
@@ -1735,6 +1738,8 @@ i_same(struct f_inst *f1, struct f_inst *f2)
 	       ((struct f_inst_roa_check *) f2)->rtc->name))
       return 0;
     break;
+  case FI_FORMAT: ONEARG; break;
+  case FI_ASSERT: ONEARG; break;
   default:
     bug( "Unknown instruction %d in same (%c)", f1->fi_code, f1->fi_code & 0xff);
   }

From 29958745c8f58cffd24e1793702524ce84545e5a Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Thu, 3 May 2018 11:14:49 +0200
Subject: [PATCH 158/198] Makefile: Only set git version if BIRD is build from
 its repository.
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Thanks to Toke Høiland-Jørgensen <toke@toke.dk> for reporting this bug.
---
 Makefile.in | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/Makefile.in b/Makefile.in
index c8168bbe..eb6cc5c9 100644
--- a/Makefile.in
+++ b/Makefile.in
@@ -21,11 +21,6 @@ INSTALL=@INSTALL@
 INSTALL_PROGRAM=@INSTALL_PROGRAM@
 INSTALL_DATA=@INSTALL_DATA@
 
-git-label:=$(strip $(shell git describe --always --dirty=-x 2>/dev/null))
-ifneq ($(git-label),)
-        CFLAGS += -DGIT_LABEL="$(git-label)"
-endif
-
 client=$(addprefix $(exedir)/,@CLIENT@)
 daemon=$(exedir)/bird
 protocols=@protocols@
@@ -42,6 +37,11 @@ srcdir := @srcdir@
 objdir := @objdir@
 exedir := @exedir@
 
+git-label:=$(strip $(shell cd $(srcdir) && [ "$$(git rev-parse --show-toplevel)" = "$$(readlink -f .)" ] && git describe --always --dirty=-x 2>/dev/null))
+ifneq ($(git-label),)
+        CFLAGS += -DGIT_LABEL="$(git-label)"
+endif
+
 ifeq ($(objdir),.)
   objdir := $(realpath .)
 endif

From 23b079043bea5899b49a750c4616aff5b332c50d Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Thu, 3 May 2018 16:02:29 +0200
Subject: [PATCH 159/198] Babel: Fix type of route entry router ID

The router ID being assigned to routes was a uint, which discards the
upper 32 bits. This also has the nice side effect of echoing the wrong
router ID back to other routers.

Thanks to Toke Hoiland-Jorgensen for the patch.
---
 proto/babel/babel.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/proto/babel/babel.c b/proto/babel/babel.c
index 88c4711e..797a83d4 100644
--- a/proto/babel/babel.c
+++ b/proto/babel/babel.c
@@ -2133,7 +2133,7 @@ babel_rt_notify(struct proto *P, struct channel *c UNUSED, struct network *net,
     uint internal = (new->attrs->src->proto == P);
     uint rt_seqno = internal ? new->u.babel.seqno : p->update_seqno;
     uint rt_metric = ea_get_int(attrs, EA_BABEL_METRIC, 0);
-    uint rt_router_id = internal ? new->u.babel.router_id : p->router_id;
+    u64 rt_router_id = internal ? new->u.babel.router_id : p->router_id;
 
     if (rt_metric > BABEL_INFINITY)
     {

From 70fab17837dbb4c5848681e4c6b9b90891891130 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Thu, 3 May 2018 16:55:11 +0200
Subject: [PATCH 160/198] Babel: Add option to randomize router ID

When a Babel node restarts, it loses its sequence number, which can cause
its routes to be rejected by peers until the state is cleared out by other
nodes in the network (which can take on the order of minutes).

There are two ways to fix this: Having stable storage to keep the sequence
number across restarts, or picking a different router ID each time.

This implements the latter, by introducing a new option that will cause
BIRD to randomize a high 32 bits of router ID every time it starts up.
This avoids the problem at the cost of not having stable router IDs in
the network.

Thanks to Toke Hoiland-Jorgensen for the patch.
---
 doc/bird.sgml        | 10 ++++++++++
 proto/babel/babel.c  | 11 +++++++++++
 proto/babel/babel.h  |  1 +
 proto/babel/config.Y |  3 ++-
 4 files changed, 24 insertions(+), 1 deletion(-)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index 1191fa03..ae308d4c 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -1691,6 +1691,7 @@ supports the following per-interface configuration options:
 protocol babel [<name>] {
 	ipv4 { <channel config> };
 	ipv6 [sadr] { <channel config> };
+        randomize router id <switch>;
 	interface <interface pattern> {
 		type <wired|wireless>;
 		rxcost <number>;
@@ -1713,6 +1714,15 @@ protocol babel [<name>] {
       <tag><label id="babel-channel">ipv4 | ipv6 [sadr] <m/channel config/</tag>
       The supported channels are IPv4, IPv6, and IPv6 SADR.
 
+      <tag><label id="babel-random-router-id">randomize router id <m/switch/</tag>
+      If enabled, Bird will randomize the top 32 bits of its router ID whenever
+      the protocol instance starts up. If a Babel node restarts, it loses its
+      sequence number, which can cause its routes to be rejected by peers until
+      the state is cleared out by other nodes in the network (which can take on
+      the order of minutes). Enabling this option causes Bird to pick a random
+      router ID every time it starts up, which avoids this problem at the cost
+      of not having stable router IDs in the network. Default: no.
+
       <tag><label id="babel-type">type wired|wireless </tag>
       This option specifies the interface type: Wired or wireless. On wired
       interfaces a neighbor is considered unreachable after a small number of
diff --git a/proto/babel/babel.c b/proto/babel/babel.c
index 797a83d4..ce05191c 100644
--- a/proto/babel/babel.c
+++ b/proto/babel/babel.c
@@ -2226,6 +2226,14 @@ babel_init(struct proto_config *CF)
   return P;
 }
 
+static inline void
+babel_randomize_router_id(struct babel_proto *p)
+{
+  p->router_id &= (u64) 0xffffffff;
+  p->router_id |= ((u64) random()) << 32;
+  TRACE(D_EVENTS, "Randomized router ID to %lR", p->router_id);
+}
+
 static int
 babel_start(struct proto *P)
 {
@@ -2244,6 +2252,9 @@ babel_start(struct proto *P)
   p->update_seqno = 1;
   p->router_id = proto_get_router_id(&cf->c);
 
+  if (cf->randomize_router_id)
+    babel_randomize_router_id(p);
+
   p->route_slab = sl_new(P->pool, sizeof(struct babel_route));
   p->source_slab = sl_new(P->pool, sizeof(struct babel_source));
   p->msg_slab = sl_new(P->pool, sizeof(struct babel_msg_node));
diff --git a/proto/babel/babel.h b/proto/babel/babel.h
index b194ce30..e5c9cd5b 100644
--- a/proto/babel/babel.h
+++ b/proto/babel/babel.h
@@ -112,6 +112,7 @@ struct babel_config {
   struct proto_config c;
   list iface_list;			/* List of iface configs (struct babel_iface_config) */
   uint hold_time;			/* Time to hold stale entries and unreachable routes */
+  u8 randomize_router_id;
 
   struct channel_config *ip4_channel;
   struct channel_config *ip6_channel;
diff --git a/proto/babel/config.Y b/proto/babel/config.Y
index 7adfb4bb..205b4e4f 100644
--- a/proto/babel/config.Y
+++ b/proto/babel/config.Y
@@ -25,7 +25,7 @@ CF_DECLS
 CF_KEYWORDS(BABEL, INTERFACE, METRIC, RXCOST, HELLO, UPDATE, INTERVAL, PORT,
 	TYPE, WIRED, WIRELESS, RX, TX, BUFFER, PRIORITY, LENGTH, CHECK, LINK,
 	NEXT, HOP, IPV4, IPV6, BABEL_METRIC, SHOW, INTERFACES, NEIGHBORS,
-	ENTRIES)
+	ENTRIES, RANDOMIZE, ROUTER, ID)
 
 CF_GRAMMAR
 
@@ -42,6 +42,7 @@ babel_proto_item:
    proto_item
  | proto_channel
  | INTERFACE babel_iface
+ | RANDOMIZE ROUTER ID bool { BABEL_CFG->randomize_router_id = $4; }
  ;
 
 babel_proto_opts:

From eaf63d314d50cba5b2cfa8f18de64a91d3131b94 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Thu, 3 May 2018 17:07:39 +0200
Subject: [PATCH 161/198] Better initialization of random generator

Use full time precision to initialize random generator. The old
code was prone to initialize it to the same values in specific
circumstances (boot without RTC, multiple VMs starting at once).
---
 sysdep/unix/io.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/sysdep/unix/io.c b/sysdep/unix/io.c
index 012deaf0..11a0d6f1 100644
--- a/sysdep/unix/io.c
+++ b/sysdep/unix/io.c
@@ -2144,7 +2144,9 @@ io_init(void)
   // XXX init_times();
   // XXX update_times();
   boot_time = current_time();
-  srandom((uint) (current_real_time() TO_S));
+
+  u64 now = (u64) current_real_time();
+  srandom((uint) (now ^ (now >> 32)));
 }
 
 static int short_loops = 0;

From b24b781117179f301116837f0a39468343e4805b Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 16 May 2018 11:19:29 +0200
Subject: [PATCH 162/198] Filter: Add support for src filter op to access SADR
 source prefix

The patch allows to use 'net.src' to access SADR source prefix
from filters.

Thanks to Toke Hoiland-Jorgensen for the original patch for srclen.
---
 filter/config.Y |  3 ++-
 filter/filter.c | 15 +++++++++++++++
 filter/filter.h |  1 +
 lib/net.h       |  2 ++
 4 files changed, 20 insertions(+), 1 deletion(-)

diff --git a/filter/config.Y b/filter/config.Y
index f8170a83..e01e02ef 100644
--- a/filter/config.Y
+++ b/filter/config.Y
@@ -411,7 +411,7 @@ CF_KEYWORDS(FUNCTION, PRINT, PRINTN, UNSET, RETURN,
 	TRUE, FALSE, RT, RO, UNKNOWN, GENERIC,
 	FROM, GW, NET, MASK, PROTO, SOURCE, SCOPE, DEST, IFNAME, IFINDEX,
 	PREFERENCE,
-	ROA_CHECK, ASN,
+	ROA_CHECK, ASN, SRC,
 	IS_V4, IS_V6,
 	LEN, MAXLEN,
 	DEFINED,
@@ -898,6 +898,7 @@ term:
  | term '.' LEN { $$ = f_new_inst(FI_LENGTH); $$->a1.p = $1; }
  | term '.' MAXLEN { $$ = f_new_inst(FI_ROA_MAXLEN); $$->a1.p = $1; }
  | term '.' ASN { $$ = f_new_inst(FI_ROA_ASN); $$->a1.p = $1; }
+ | term '.' SRC { $$ = f_new_inst(FI_SADR_SRC); $$->a1.p = $1; }
  | term '.' MASK '(' term ')' { $$ = f_new_inst(FI_IP_MASK); $$->a1.p = $1; $$->a2.p = $5; }
  | term '.' FIRST { $$ = f_new_inst(FI_AS_PATH_FIRST); $$->a1.p = $1; }
  | term '.' LAST  { $$ = f_new_inst(FI_AS_PATH_LAST); $$->a1.p = $1; }
diff --git a/filter/filter.c b/filter/filter.c
index 881ba420..3d7b5c9f 100644
--- a/filter/filter.c
+++ b/filter/filter.c
@@ -1241,6 +1241,20 @@ interpret(struct f_inst *what)
     default: runtime( "Prefix, path, clist or eclist expected" );
     }
     break;
+  case FI_SADR_SRC: 	/* Get SADR src prefix */
+    ONEARG;
+    if (v1.type != T_NET || !net_is_sadr(v1.val.net))
+      runtime( "SADR expected" );
+
+    {
+      net_addr_ip6_sadr *net = (void *) v1.val.net;
+      net_addr *src = lp_alloc(f_pool, sizeof(net_addr_ip6));
+      net_fill_ip6(src, net->src_prefix, net->src_pxlen);
+
+      res.type = T_NET;
+      res.val.net = src;
+    }
+    break;
   case FI_ROA_MAXLEN: 	/* Get ROA max prefix length */
     ONEARG;
     if (v1.type != T_NET || !net_is_roa(v1.val.net))
@@ -1714,6 +1728,7 @@ i_same(struct f_inst *f1, struct f_inst *f2)
   case FI_RETURN: ONEARG; break;
   case FI_ROA_MAXLEN: ONEARG; break;
   case FI_ROA_ASN: ONEARG; break;
+  case FI_SADR_SRC: ONEARG; break;
   case FI_IP: ONEARG; break;
   case FI_IS_V4: ONEARG; break;
   case FI_ROUTE_DISTINGUISHER: ONEARG; break;
diff --git a/filter/filter.h b/filter/filter.h
index d347924a..982276f0 100644
--- a/filter/filter.h
+++ b/filter/filter.h
@@ -55,6 +55,7 @@
   F(FI_LENGTH,			  0, 'L') \
   F(FI_ROA_MAXLEN,		'R', 'M') \
   F(FI_ROA_ASN,			'R', 'A') \
+  F(FI_SADR_SRC,		'n', 's') \
   F(FI_IP,			'c', 'p') \
   F(FI_ROUTE_DISTINGUISHER,	'R', 'D') \
   F(FI_AS_PATH_FIRST,		'a', 'f') \
diff --git a/lib/net.h b/lib/net.h
index ad4000fd..0cd5f735 100644
--- a/lib/net.h
+++ b/lib/net.h
@@ -268,6 +268,8 @@ static inline int net_is_roa(const net_addr *a)
 static inline int net_is_flow(const net_addr *a)
 { return (a->type == NET_FLOW4) || (a->type == NET_FLOW6); }
 
+static inline int net_is_sadr(const net_addr *a)
+{ return (a->type == NET_IP6_SADR); }
 
 static inline ip4_addr net4_prefix(const net_addr *a)
 { return ((net_addr_ip4 *) a)->prefix; }

From feae132e0f9bdc62d2b90bf676d12241af8e794c Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Thu, 24 May 2018 14:51:05 +0200
Subject: [PATCH 163/198] Do not initialize route metrics in import_control
 hook

During route export, the receiving protocol often initialized route
metrics to default value in its import_control hook before export filter
was executed. This is inconsistent with the expectation that an export
filter would process the same route as one in the routing table and it
breaks setting these metrics before (e.g. for static routes directly in
static protocol).

The patch removes the initialization of route metrics in import_control
hook, the default values are already handled in rt_notify hook called
after export filters.

The patch also changed the behavior of OSPF to keep metrics when a route
is reannounced between OSPF instances (to be consistent with other
protocols) and the behavior when both ospf_metric1 and ospf_metric2
are specified (to have more expected behavior).
---
 doc/bird.sgml         | 15 ++++++++++-----
 proto/babel/babel.c   | 12 +++---------
 proto/ospf/ospf.c     | 25 ++++++-------------------
 proto/ospf/topology.c | 34 +++++++++++++++++++++++++++-------
 proto/rip/rip.c       | 12 +-----------
 5 files changed, 47 insertions(+), 51 deletions(-)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index ae308d4c..b4d9ea4e 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -1618,9 +1618,7 @@ regarded as empty bgppath/*clist for most purposes.
 	The optional attribute that can be used to specify a distance to the
 	network for routes that do not have a native protocol metric attribute
 	(like <cf/ospf_metric1/ for OSPF routes). It is used mainly by BGP to
-	compare internal distances to boundary routers (see below). It is also
-	used when the route is exported to OSPF as a default value for OSPF type
-	1 metric.
+	compare internal distances to boundary routers (see below).
 </descrip>
 
 <p>There also exist protocol-specific attributes which are described in the
@@ -3536,8 +3534,15 @@ protocol ospf [v2|v3] &lt;name&gt; {
 with internal <cf/metric/, a <cf/metric of type 2/ is always longer than any
 <cf/metric of type 1/ or any <cf/internal metric/. <cf/Internal metric/ or
 <cf/metric of type 1/ is stored in attribute <cf/ospf_metric1/, <cf/metric type
-2/ is stored in attribute <cf/ospf_metric2/. If you specify both metrics only
-metric1 is used.
+2/ is stored in attribute <cf/ospf_metric2/.
+
+When both metrics are specified then <cf/metric of type 2/ is used. This is
+relevant e.g. when a type 2 external route is propagated from one OSPF domain to
+another and <cf/ospf_metric1/ is an internal distance to the original ASBR,
+while <cf/ospf_metric2/ stores the type 2 metric. Note that in such cases if
+<cf/ospf_metric1/ is non-zero then <cf/ospf_metric2/ is increased by one to
+ensure monotonicity of metric, as internal distance is reset to zero when an
+external route is announced.
 
 <p>Each external route can also carry attribute <cf/ospf_tag/ which is a 32-bit
 integer which is used when exporting routes to other protocols; otherwise, it
diff --git a/proto/babel/babel.c b/proto/babel/babel.c
index ce05191c..44c6adb8 100644
--- a/proto/babel/babel.c
+++ b/proto/babel/babel.c
@@ -2087,20 +2087,14 @@ babel_prepare_attrs(struct linpool *pool, ea_list *next, uint metric, u64 router
 
 
 static int
-babel_import_control(struct proto *P, struct rte **new, struct ea_list **attrs, struct linpool *pool)
+babel_import_control(struct proto *P, struct rte **new, struct ea_list **attrs UNUSED, struct linpool *pool UNUSED)
 {
-  struct babel_proto *p = (void *) P;
-  rte *rt = *new;
+  rte *e = *new;
 
   /* Reject our own unreachable routes */
-  if ((rt->attrs->dest == RTD_UNREACHABLE) && (rt->attrs->src->proto == P))
+  if ((e->attrs->dest == RTD_UNREACHABLE) && (e->attrs->src->proto == P))
     return -1;
 
-
-  /* Prepare attributes with initial values */
-  if (rt->attrs->source != RTS_BABEL)
-    *attrs = babel_prepare_attrs(pool, NULL, 0, p->router_id);
-
   return 0;
 }
 
diff --git a/proto/ospf/ospf.c b/proto/ospf/ospf.c
index df6c452e..0a9efd19 100644
--- a/proto/ospf/ospf.c
+++ b/proto/ospf/ospf.c
@@ -446,34 +446,21 @@ ospf_disp(timer * timer)
  * import to the filters.
  */
 static int
-ospf_import_control(struct proto *P, rte **new, ea_list **attrs, struct linpool *pool)
+ospf_import_control(struct proto *P, rte **new, ea_list **attrs UNUSED, struct linpool *pool UNUSED)
 {
   struct ospf_proto *p = (struct ospf_proto *) P;
   struct ospf_area *oa = ospf_main_area(p);
   rte *e = *new;
 
+  /* Reject our own routes */
   if (e->attrs->src->proto == P)
-    return -1;			/* Reject our own routes */
+    return -1;
 
+  /* Do not export routes to stub areas */
   if (oa_is_stub(oa))
-    return -1;			/* Do not export routes to stub areas */
+    return -1;
 
-  ea_list *ea = e->attrs->eattrs;
-  u32 m0 = ea_get_int(ea, EA_GEN_IGP_METRIC, LSINFINITY);
-  u32 m1 = MIN(m0, LSINFINITY);
-  u32 m2 = 10000;
-  u32 tag = 0;
-
-  /* Hack for setting attributes directly in static protocol */
-  if (e->attrs->source == RTS_STATIC)
-  {
-    m1 = ea_get_int(ea, EA_OSPF_METRIC1, m1);
-    m2 = ea_get_int(ea, EA_OSPF_METRIC2, 10000);
-    tag = ea_get_int(ea, EA_OSPF_TAG, 0);
-  }
-
-  *attrs = ospf_build_attrs(*attrs, pool, m1, m2, tag, 0);
-  return 0;			/* Leave decision to the filters */
+  return 0;
 }
 
 static struct ea_list *
diff --git a/proto/ospf/topology.c b/proto/ospf/topology.c
index 717c8280..e909bbe9 100644
--- a/proto/ospf/topology.c
+++ b/proto/ospf/topology.c
@@ -1281,14 +1281,34 @@ ospf_rt_notify(struct proto *P, struct channel *ch UNUSED, net *n, rte *new, rte
 
   /* Get route attributes */
   rta *a = new->attrs;
-  u32 m1 = ea_get_int(ea, EA_OSPF_METRIC1, LSINFINITY);
-  u32 m2 = ea_get_int(ea, EA_OSPF_METRIC2, 10000);
-  int ebit = (m1 == LSINFINITY);
-  u32 metric = ebit ? m2 : m1;
-  u32 tag = ea_get_int(ea, EA_OSPF_TAG, 0);
+  eattr *m1a = ea_find(ea, EA_OSPF_METRIC1);
+  eattr *m2a = ea_find(ea, EA_OSPF_METRIC2);
+  uint m1 = m1a ? m1a->u.data : 0;
+  uint m2 = m2a ? m2a->u.data : 10000;
+
+  if (m1 > LSINFINITY)
+  {
+    log(L_WARN "%s: Invalid ospf_metric1 value %u for route %N",
+	p->p.name, m1, n->n.addr);
+    m1 = LSINFINITY;
+  }
+
+  if (m2 > LSINFINITY)
+  {
+    log(L_WARN "%s: Invalid ospf_metric2 value %u for route %N",
+	p->p.name, m2, n->n.addr);
+    m2 = LSINFINITY;
+  }
+
+  /* Ensure monotonicity of metric if both m1 and m2 are used */
+  if ((m1 > 0) && (m2 < LSINFINITY))
+    m2++;
+
+  uint ebit = m2a || !m1a;
+  uint metric = ebit ? m2 : m1;
+  uint tag = ea_get_int(ea, EA_OSPF_TAG, 0);
+
   ip_addr fwd = IPA_NONE;
-
-
   if ((a->dest == RTD_UNICAST) && use_gw_for_fwaddr(p, a->nh.gw, a->nh.iface))
     fwd = a->nh.gw;
 
diff --git a/proto/rip/rip.c b/proto/rip/rip.c
index 85e37cea..baf98737 100644
--- a/proto/rip/rip.c
+++ b/proto/rip/rip.c
@@ -1020,16 +1020,6 @@ rip_prepare_attrs(struct linpool *pool, ea_list *next, u8 metric, u16 tag)
   return l;
 }
 
-static int
-rip_import_control(struct proto *P UNUSED, struct rte **rt, struct ea_list **attrs, struct linpool *pool)
-{
-  /* Prepare attributes with initial values */
-  if ((*rt)->attrs->source != RTS_RIP)
-    *attrs = rip_prepare_attrs(pool, *attrs, 1, 0);
-
-  return 0;
-}
-
 static void
 rip_reload_routes(struct channel *C)
 {
@@ -1091,7 +1081,7 @@ rip_init(struct proto_config *CF)
   P->if_notify = rip_if_notify;
   P->rt_notify = rip_rt_notify;
   P->neigh_notify = rip_neigh_notify;
-  P->import_control = rip_import_control;
+  // P->import_control = rip_import_control;
   P->reload_routes = rip_reload_routes;
   P->make_tmp_attrs = rip_make_tmp_attrs;
   P->store_tmp_attrs = rip_store_tmp_attrs;

From 906092534ba8479ca76723b7dd7ee233f5a70d1e Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Mon, 30 Apr 2018 12:20:04 +0200
Subject: [PATCH 164/198] Macro: Added a bunch of dirty C preprocessor tricks

Included are Makefile implicit rules to show the preprocessed source.
When debugging something around this, it may be handy.
---
 Makefile.in | 11 +++++-
 lib/macro.h | 98 +++++++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 108 insertions(+), 1 deletion(-)
 create mode 100644 lib/macro.h

diff --git a/Makefile.in b/Makefile.in
index eb6cc5c9..8f6c0c8b 100644
--- a/Makefile.in
+++ b/Makefile.in
@@ -95,7 +95,7 @@ clean = $(eval $(call clean_in,$(1)))
 include $(addsuffix /Makefile,$(addprefix $(srcdir)/,$(dirs)))
 
 # Generic rules
-
+# Object file rules
 $(objdir)/%.o: $(srcdir)/%.c $(objdir)/.dir-stamp $(objdir)/sysdep/paths.h
 	$(E)echo CC -o $@ -c $<
 	$(Q)$(CC) $(CFLAGS) -MMD -MP -o $@ -c $<
@@ -104,7 +104,16 @@ $(objdir)/%.o: $(objdir)/%.c $(objdir)/.dir-stamp $(objdir)/sysdep/paths.h
 	$(E)echo CC -o $@ -c $<
 	$(Q)$(CC) $(CFLAGS) -MMD -MP -o $@ -c $<
 
+# Debug: Preprocessed source rules
+$(objdir)/%.E: $(srcdir)/%.c $(objdir)/.dir-stamp $(objdir)/sysdep/paths.h
+	$(E)echo CC -o $@ -E $<
+	$(Q)$(CC) $(CFLAGS) -MMD -MP -o $@ -E $<
 
+$(objdir)/%.E: $(objdir)/%.c $(objdir)/.dir-stamp $(objdir)/sysdep/paths.h
+	$(E)echo CC -o $@ -E $<
+	$(Q)$(CC) $(CFLAGS) -MMD -MP -o $@ -E $<
+
+# Debug: Assembler object rules
 $(objdir)/%.S: $(srcdir)/%.c $(objdir)/.dir-stamp $(objdir)/sysdep/paths.h
 	$(E)echo CC -o $@ -S $<
 	$(Q)$(CC) $(CFLAGS) -MMD -MP -o $@ -S $<
diff --git a/lib/macro.h b/lib/macro.h
new file mode 100644
index 00000000..24fc3393
--- /dev/null
+++ b/lib/macro.h
@@ -0,0 +1,98 @@
+/* 
+ *	BIRD Macro Tricks
+ *
+ *	(c) 2018 Jan Maria Matejka <mq@jmq.cz>
+ *
+ *	Can be freely distributed and used under the terms of the GNU GPL.
+ *
+ *	Contains useful but dirty macro tricks:
+ *		MACRO_CONCAT(a, b)	-> concatenates a##b
+ *		MACRO_BOOL(x)		-> convert 0 to 0, anything else to 1
+ *		MACRO_IFELSE(b)(true-branch)(false-branch)
+ *					-> b shall be 0 or 1; expands to the appropriate branch
+ *		MACRO_ISEMPTY(...)	-> 1 for empty argument list, 0 otherwise
+ *		MACRO_FOREACH(func, ...)
+ *					-> calling FOREACH(func, a, b, c, d) expands to
+ *						func(a) func(b) func(c) func(d)
+ *		MACRO_RPACK(func, terminator, ...)
+ *					-> packs the list into recursive calls:
+ *						func(func(func(func(terminator, a), b), c), d)
+ */
+
+#ifndef _BIRD_MACRO_H_
+#define _BIRD_MACRO_H_
+
+/* What to do with args */
+#define MACRO_DROP(...)
+#define MACRO_UNPAREN(...) __VA_ARGS__
+#define MACRO_SEP(a, b, sep)  a sep b
+
+/* Aliases for some special chars */
+#define MACRO_COMMA ,
+#define MACRO_LPAREN (
+#define MACRO_RPAREN )
+#define MACRO_LPAREN_() (
+#define MACRO_RPAREN_() )
+
+/* Multiple expansion trick */
+#define MACRO_EXPAND0(...) __VA_ARGS__
+#define MACRO_EXPAND1(...) MACRO_EXPAND0(MACRO_EXPAND0(__VA_ARGS__))
+#define MACRO_EXPAND2(...) MACRO_EXPAND1(MACRO_EXPAND1(__VA_ARGS__))
+#define MACRO_EXPAND3(...) MACRO_EXPAND2(MACRO_EXPAND2(__VA_ARGS__))
+#define MACRO_EXPAND(...) MACRO_EXPAND3(MACRO_EXPAND3(__VA_ARGS__))
+
+/* Deferring expansion in the expansion trick */
+#define MACRO_EMPTY()
+#define MACRO_DEFER(t) t MACRO_EMPTY()
+#define MACRO_DEFER2(t) t MACRO_EMPTY MACRO_EMPTY()()
+#define MACRO_DEFER3(t) t MACRO_EMPTY MACRO_EMPTY MACRO_EMPTY()()()
+
+/* Token concatenation */
+#define MACRO_CONCAT(prefix, ...) prefix##__VA_ARGS__
+#define MACRO_CONCAT_AFTER(...) MACRO_CONCAT(__VA_ARGS__)
+
+/* Get first or second argument only */
+#define MACRO_FIRST(a, ...) a
+#define MACRO_SECOND(a, b, ...) b
+#define MACRO_SECOND_OR_ZERO(...) MACRO_SECOND(__VA_ARGS__, 0,)
+
+/* Macro Boolean auxiliary macros */
+#define MACRO_BOOL_CHECK_0 ~, 1
+#define MACRO_BOOL_NEG(x) MACRO_SECOND_OR_ZERO(MACRO_CONCAT(MACRO_BOOL_CHECK_, x))
+
+#define MACRO_BOOL_NOT_0  1
+#define MACRO_BOOL_NOT_1  0
+
+/* Macro Boolean negation */
+#define MACRO_NOT(x) MACRO_CONCAT(MACRO_BOOL_NOT_, x)
+
+/* Convert anything to bool (anything -> 1, 0 -> 0) */
+#define MACRO_BOOL(x) MACRO_NOT(MACRO_BOOL_NEG(x))
+
+/*
+ * Macro If/Else condition
+ * Usage: MACRO_IFELSE(condition)(true-branch)(false-branch)
+ * Expands to true-branch if condition is true, otherwise to false-branch.
+ */
+#define MACRO_IFELSE(b) MACRO_CONCAT(MACRO_IFELSE_, b)
+#define MACRO_IFELSE_0(...) MACRO_UNPAREN
+#define MACRO_IFELSE_1(...) __VA_ARGS__ MACRO_DROP
+
+/* Auxiliary macros for MACRO_FOREACH */
+#define MACRO_ISLAST(...) MACRO_BOOL_NEG(MACRO_FIRST(MACRO_ISLAST_CHECK __VA_ARGS__)())
+#define MACRO_ISLAST_CHECK() 0
+
+#define MACRO_FOREACH_EXPAND(call, a, ...) MACRO_IFELSE(MACRO_ISLAST(__VA_ARGS__))(call(a))(call(a) MACRO_DEFER2(MACRO_FOREACH_PAREN)()(call, __VA_ARGS__))
+#define MACRO_FOREACH_PAREN() MACRO_FOREACH_EXPAND
+
+#define MACRO_RPACK_EXPAND(call, terminator, a, ...) MACRO_IFELSE(MACRO_ISLAST(__VA_ARGS__))(call(terminator, a))(call(MACRO_DEFER2(MACRO_RPACK_PAREN)()(call, terminator, __VA_ARGS__), a))
+#define MACRO_RPACK_PAREN() MACRO_RPACK_EXPAND
+/*
+ * Call the first argument for each following:
+ * MACRO_FOREACH(func, a, b, c, d) expands to func(a) func(b) func(c) func(d).
+ * It supports also macros as func.
+ */
+#define MACRO_FOREACH(call, ...) MACRO_EXPAND(MACRO_FOREACH_EXPAND(call, __VA_ARGS__))
+#define MACRO_RPACK(call, terminator, ...) MACRO_EXPAND(MACRO_RPACK_EXPAND(call, terminator, __VA_ARGS__))
+
+#endif

From 31d6939cde071ab04d1da6b3ff6fb6cd579e164a Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Mon, 30 Apr 2018 12:39:32 +0200
Subject: [PATCH 165/198] Filter: Instruction codes linearized

---
 filter/filter.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/filter/filter.h b/filter/filter.h
index 982276f0..bc330aba 100644
--- a/filter/filter.h
+++ b/filter/filter.h
@@ -75,7 +75,7 @@
 
 enum f_instruction_code {
 #define F(c,a,b) \
-  c = FI__TWOCHAR(a,b),
+  c,
 FI__LIST
 #undef F
 } PACKED;

From cff9e937fd0ed42b88be1deb5e1aa9fe301caabd Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Mon, 30 Apr 2018 12:49:22 +0200
Subject: [PATCH 166/198] Filter: instruction names

---
 filter/f-util.c | 16 ++++++++++++++++
 filter/filter.h |  1 +
 2 files changed, 17 insertions(+)

diff --git a/filter/f-util.c b/filter/f-util.c
index 68aecd73..6170760b 100644
--- a/filter/f-util.c
+++ b/filter/f-util.c
@@ -77,6 +77,22 @@ f_generate_roa_check(struct rtable_config *table, struct f_inst *prefix, struct
   return &ret->i;
 }
 
+static const char * const f_instruction_name_str[] = {
+#define F(c,a,b) \
+  [c] = #c,
+FI__LIST
+#undef F
+};
+
+const char *
+f_instruction_name(enum f_instruction_code fi)
+{
+  if (fi < FI__MAX)
+    return f_instruction_name_str[fi];
+  else
+    bug("Got unknown instruction code: %d", fi);
+}
+
 char *
 filter_name(struct filter *filter)
 {
diff --git a/filter/filter.h b/filter/filter.h
index bc330aba..694bdfd7 100644
--- a/filter/filter.h
+++ b/filter/filter.h
@@ -78,6 +78,7 @@ enum f_instruction_code {
   c,
 FI__LIST
 #undef F
+  FI__MAX,
 } PACKED;
 
 struct f_inst {		/* Instruction */

From 478f9babed361f8df8a9c944f20bcbe116dc30aa Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Mon, 30 Apr 2018 13:29:05 +0200
Subject: [PATCH 167/198] Filter: Removing the third argument hack

Just to make the code a bit more clean and easier to maintain.
---
 filter/config.Y |  6 ++----
 filter/filter.c | 15 +++++----------
 filter/filter.h | 15 ++++-----------
 3 files changed, 11 insertions(+), 25 deletions(-)

diff --git a/filter/config.Y b/filter/config.Y
index e01e02ef..cf499b66 100644
--- a/filter/config.Y
+++ b/filter/config.Y
@@ -304,12 +304,10 @@ f_generate_lc(struct f_inst *t1, struct f_inst *t2, struct f_inst *t3)
   }
   else
   {
-    rv = cfg_allocz(sizeof(struct f_inst3));
-    rv->lineno = ifs->lino;
-    rv->fi_code = FI_LC_CONSTRUCT;
+    rv = f_new_inst(FI_LC_CONSTRUCT);
     rv->a1.p = t1;
     rv->a2.p = t2;
-    INST3(rv).p = t3;
+    rv->a3.p = t3;
   }
 
   return rv;
diff --git a/filter/filter.c b/filter/filter.c
index 3d7b5c9f..7117da48 100644
--- a/filter/filter.c
+++ b/filter/filter.c
@@ -601,6 +601,8 @@ static struct tbf rl_runtime_err = TBF_DEFAULT_LOG_LIMITS;
 #define TWOARGS_C TWOARGS \
                   if (v1.type != v2.type) \
 		    runtime( "Can't operate with values of incompatible types" );
+#define THREEARGS TWOARGS \
+		  ARG(v3, a3.p)
 #define ACCESS_RTE \
   do { if (!f_rte) runtime("No route to access"); } while (0)
 
@@ -628,7 +630,7 @@ static struct f_val
 interpret(struct f_inst *what)
 {
   struct symbol *sym;
-  struct f_val v1, v2, res = { .type = T_VOID }, *vp;
+  struct f_val v1, v2, v3, res = { .type = T_VOID }, *vp;
   unsigned u1, u2;
   int i;
   u32 as;
@@ -749,12 +751,7 @@ interpret(struct f_inst *what)
 
   case FI_LC_CONSTRUCT:
     {
-      TWOARGS;
-
-      /* Third argument hack */
-      struct f_val v3 = interpret(INST3(what).p);
-      if (v3.type & T_RETURN)
-	return v3;
+      THREEARGS;
 
       if ((v1.type != T_INT) || (v2.type != T_INT) || (v3.type != T_INT))
 	runtime( "Can't operate with value of non-integer type in LC constructor" );
@@ -1664,9 +1661,7 @@ i_same(struct f_inst *f1, struct f_inst *f2)
   case FI_TYPE: ONEARG; break;
 
   case FI_LC_CONSTRUCT:
-    TWOARGS;
-    if (!i_same(INST3(f1).p, INST3(f2).p))
-      return 0;
+    THREEARGS;
     break;
 
   case FI_SET:
diff --git a/filter/filter.h b/filter/filter.h
index 694bdfd7..1e5fc8ff 100644
--- a/filter/filter.h
+++ b/filter/filter.h
@@ -93,6 +93,10 @@ struct f_inst {		/* Instruction */
     uint i;
     void *p;
   } a2;			/* The second argument */
+  union {
+    int i;
+    void *p;
+  } a3;			/* The third argument */
   int lineno;
 };
 
@@ -105,17 +109,6 @@ struct f_inst_roa_check {
   struct rtable_config *rtc;
 };
 
-struct f_inst3 {
-  struct f_inst i;
-  union {
-    int i;
-    void *p;
-  } a3;
-};
-
-#define INST3(x) (((struct f_inst3 *) x)->a3)
-
-
 struct f_prefix {
   net_addr net;
   u8 lo, hi;

From 0ec6b5ecd37529d57079e13748c4ecbd336332c1 Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Mon, 30 Apr 2018 16:06:53 +0200
Subject: [PATCH 168/198] Filter: Simple type checks converted to ARG() macro

---
 filter/filter.c | 227 ++++++++++++++++++++----------------------------
 filter/filter.h |   2 +
 2 files changed, 95 insertions(+), 134 deletions(-)

diff --git a/filter/filter.c b/filter/filter.c
index 7117da48..2b22bfa3 100644
--- a/filter/filter.c
+++ b/filter/filter.c
@@ -590,19 +590,16 @@ static struct tbf rl_runtime_err = TBF_DEFAULT_LOG_LIMITS;
     return res; \
   } while(0)
 
-#define ARG(x,y) \
-	x = interpret(what->y); \
-	if (x.type & T_RETURN) \
-		return x;
+#define ARG_ANY(n) \
+    v##n = interpret(what->a##n.p); \
+    if (v##n.type & T_RETURN) \
+      return v##n;
+
+#define ARG(n,t) ARG_ANY(n) \
+    if (v##n.type != t) \
+      runtime("Argument %d of instruction %s must be of type %02x, got %02x", \
+	  n, f_instruction_name(what->fi_code), t, v##n.type);
 
-#define ONEARG ARG(v1, a1.p)
-#define TWOARGS ARG(v1, a1.p) \
-		ARG(v2, a2.p)
-#define TWOARGS_C TWOARGS \
-                  if (v1.type != v2.type) \
-		    runtime( "Can't operate with values of incompatible types" );
-#define THREEARGS TWOARGS \
-		  ARG(v3, a3.p)
 #define ACCESS_RTE \
   do { if (!f_rte) runtime("No route to access"); } while (0)
 
@@ -640,61 +637,44 @@ interpret(struct f_inst *what)
   switch(what->fi_code) {
 /* Binary operators */
   case FI_ADD:
-    TWOARGS_C;
-    switch (res.type = v1.type) {
-    case T_VOID: runtime( "Can't operate with values of type void" );
-    case T_INT: res.val.i = v1.val.i + v2.val.i; break;
-    default: runtime( "Usage of unknown type" );
-    }
+    ARG(1,T_INT);
+    ARG(2,T_INT);
+    res.type = T_INT;
+    res.val.i = v1.val.i + v2.val.i;
     break;
   case FI_SUBTRACT:
-    TWOARGS_C;
-    switch (res.type = v1.type) {
-    case T_VOID: runtime( "Can't operate with values of type void" );
-    case T_INT: res.val.i = v1.val.i - v2.val.i; break;
-    default: runtime( "Usage of unknown type" );
-    }
+    ARG(1,T_INT);
+    ARG(2,T_INT);
+    res.type = T_INT;
+    res.val.i = v1.val.i - v2.val.i;
     break;
   case FI_MULTIPLY:
-    TWOARGS_C;
-    switch (res.type = v1.type) {
-    case T_VOID: runtime( "Can't operate with values of type void" );
-    case T_INT: res.val.i = v1.val.i * v2.val.i; break;
-    default: runtime( "Usage of unknown type" );
-    }
+    ARG(1,T_INT);
+    ARG(2,T_INT);
+    res.type = T_INT;
+    res.val.i = v1.val.i * v2.val.i;
     break;
   case FI_DIVIDE:
-    TWOARGS_C;
-    switch (res.type = v1.type) {
-    case T_VOID: runtime( "Can't operate with values of type void" );
-    case T_INT: if (v2.val.i == 0) runtime( "Mother told me not to divide by 0" );
-      	        res.val.i = v1.val.i / v2.val.i; break;
-    default: runtime( "Usage of unknown type" );
-    }
+    ARG(1,T_INT);
+    ARG(2,T_INT);
+    res.type = T_INT;
+    if (v2.val.i == 0) runtime( "Mother told me not to divide by 0" );
+    res.val.i = v1.val.i / v2.val.i;
     break;
-
   case FI_AND:
   case FI_OR:
-    ARG(v1, a1.p);
-    if (v1.type != T_BOOL)
-      runtime( "Can't do boolean operation on non-booleans" );
+    ARG(1,T_BOOL);
     if (v1.val.i == (what->fi_code == FI_OR)) {
       res.type = T_BOOL;
       res.val.i = v1.val.i;
-      break;
+    } else {
+      ARG(2,T_BOOL);
+      res = v2;
     }
-
-    ARG(v2, a2.p);
-    if (v2.type != T_BOOL)
-      runtime( "Can't do boolean operation on non-booleans" );
-    res.type = T_BOOL;
-    res.val.i = v2.val.i;
     break;
-
   case FI_PAIR_CONSTRUCT:
-    TWOARGS;
-    if ((v1.type != T_INT) || (v2.type != T_INT))
-      runtime( "Can't operate with value of non-integer type in pair constructor" );
+    ARG(1,T_INT);
+    ARG(2,T_INT);
     u1 = v1.val.i;
     u2 = v2.val.i;
     if ((u1 > 0xFFFF) || (u2 > 0xFFFF))
@@ -705,7 +685,8 @@ interpret(struct f_inst *what)
 
   case FI_EC_CONSTRUCT:
     {
-      TWOARGS;
+      ARG_ANY(1);
+      ARG(2, T_INT);
 
       int check, ipv4_used;
       u32 key, val;
@@ -723,8 +704,6 @@ interpret(struct f_inst *what)
       else
 	runtime("Can't operate with key of non-integer/IPv4 type in EC constructor");
 
-      if (v2.type != T_INT)
-	runtime("Can't operate with value of non-integer type in EC constructor");
       val = v2.val.i;
 
       /* XXXX */
@@ -751,10 +730,9 @@ interpret(struct f_inst *what)
 
   case FI_LC_CONSTRUCT:
     {
-      THREEARGS;
-
-      if ((v1.type != T_INT) || (v2.type != T_INT) || (v3.type != T_INT))
-	runtime( "Can't operate with value of non-integer type in LC constructor" );
+      ARG(1, T_INT);
+      ARG(2, T_INT);
+      ARG(3, T_INT);
 
       res.type = T_LC;
       res.val.lc = (lcomm) { v1.val.i, v2.val.i, v3.val.i };
@@ -791,7 +769,8 @@ interpret(struct f_inst *what)
 /* Relational operators */
 
 #define COMPARE(x) \
-    TWOARGS; \
+    ARG_ANY(1); \
+    ARG_ANY(2); \
     i = val_compare(v1, v2); \
     if (i==CMP_ERROR) \
       runtime( "Can't compare values of incompatible types" ); \
@@ -800,7 +779,8 @@ interpret(struct f_inst *what)
     break;
 
 #define SAME(x) \
-    TWOARGS; \
+    ARG_ANY(1); \
+    ARG_ANY(2); \
     i = val_same(v1, v2); \
     res.type = T_BOOL; \
     res.val.i = (x); \
@@ -812,15 +792,14 @@ interpret(struct f_inst *what)
   case FI_LTE: COMPARE(i!=1);
 
   case FI_NOT:
-    ONEARG;
-    if (v1.type != T_BOOL)
-      runtime( "Not applied to non-boolean" );
+    ARG(1,T_BOOL);
     res = v1;
     res.val.i = !res.val.i;
     break;
 
   case FI_MATCH:
-    TWOARGS;
+    ARG_ANY(1);
+    ARG_ANY(2);
     res.type = T_BOOL;
     res.val.i = val_in_range(v1, v2);
     if (res.val.i == CMP_ERROR)
@@ -829,7 +808,8 @@ interpret(struct f_inst *what)
     break;
 
   case FI_NOT_MATCH:
-    TWOARGS;
+    ARG_ANY(1);
+    ARG_ANY(2);
     res.type = T_BOOL;
     res.val.i = val_in_range(v1, v2);
     if (res.val.i == CMP_ERROR)
@@ -838,12 +818,12 @@ interpret(struct f_inst *what)
     break;
 
   case FI_DEFINED:
-    ONEARG;
+    ARG_ANY(1);
     res.type = T_BOOL;
     res.val.i = (v1.type != T_VOID) && !undef_value(v1);
     break;
   case FI_TYPE:
-    ONEARG;
+    ARG_ANY(1); /* There may be more types supporting this operation */
     switch (v1.type)
     {
       case T_NET:
@@ -855,16 +835,14 @@ interpret(struct f_inst *what)
     }
     break;
   case FI_IS_V4:
-    ONEARG;
-    if (v1.type != T_IP)
-      runtime( "IP version check needs an IP address" );
+    ARG(1, T_IP);
     res.type = T_BOOL;
     res.val.i = ipa_is_ip4(v1.val.ip);
     break;
 
   /* Set to indirect value, a1 = variable, a2 = value */
   case FI_SET:
-    ARG(v2, a2.p);
+    ARG_ANY(2);
     sym = what->a1.p;
     vp = sym->def;
     if ((sym->class != (SYM_VARIABLE | v2.type)) && (v2.type != T_VOID))
@@ -899,24 +877,23 @@ interpret(struct f_inst *what)
     res = * ((struct f_val *) what->a1.p);
     break;
   case FI_PRINT:
-    ONEARG;
+    ARG_ANY(1);
     val_format(v1, &f_buf);
     break;
   case FI_CONDITION:	/* ? has really strange error value, so we can implement if ... else nicely :-) */
-    ONEARG;
-    if (v1.type != T_BOOL)
-      runtime( "If requires boolean expression" );
+    ARG(1, T_BOOL);
     if (v1.val.i) {
-      ARG(res,a2.p);
+      ARG_ANY(2);
       res.val.i = 0;
-    } else res.val.i = 1;
+    } else
+      res.val.i = 1;
     res.type = T_BOOL;
     break;
   case FI_NOP:
     debug( "No operation\n" );
     break;
   case FI_PRINT_AND_DIE:
-    ONEARG;
+    ARG_ANY(1);
     if ((what->a2.i == F_NOP || (what->a2.i != F_NONL && what->a1.p)) &&
 	!(f_flags & FF_SILENT))
       log_commit(*L_INFO, &f_buf);
@@ -963,7 +940,7 @@ interpret(struct f_inst *what)
     break;
   case FI_RTA_SET:
     ACCESS_RTE;
-    ONEARG;
+    ARG_ANY(1);
     if (what->aux != v1.type)
       runtime( "Attempt to set static attribute to incompatible type" );
 
@@ -1109,7 +1086,7 @@ interpret(struct f_inst *what)
     break;
   case FI_EA_SET:
     ACCESS_RTE;
-    ONEARG;
+    ARG_ANY(1);
     {
       struct ea_list *l = lp_alloc(f_pool, sizeof(struct ea_list) + sizeof(eattr));
       u16 code = what->a2.i;
@@ -1218,16 +1195,14 @@ interpret(struct f_inst *what)
     break;
   case FI_PREF_SET:
     ACCESS_RTE;
-    ONEARG;
-    if (v1.type != T_INT)
-      runtime( "Can't set preference to non-integer" );
+    ARG(1,T_INT);
     if (v1.val.i > 0xFFFF)
       runtime( "Setting preference value out of bounds" );
     f_rte_cow();
     (*f_rte)->pref = v1.val.i;
     break;
   case FI_LENGTH:	/* Get length of */
-    ONEARG;
+    ARG_ANY(1);
     res.type = T_INT;
     switch(v1.type) {
     case T_NET:    res.val.i = net_pxlen(v1.val.net); break;
@@ -1239,8 +1214,8 @@ interpret(struct f_inst *what)
     }
     break;
   case FI_SADR_SRC: 	/* Get SADR src prefix */
-    ONEARG;
-    if (v1.type != T_NET || !net_is_sadr(v1.val.net))
+    ARG(1, T_NET);
+    if (!net_is_sadr(v1.val.net))
       runtime( "SADR expected" );
 
     {
@@ -1253,8 +1228,8 @@ interpret(struct f_inst *what)
     }
     break;
   case FI_ROA_MAXLEN: 	/* Get ROA max prefix length */
-    ONEARG;
-    if (v1.type != T_NET || !net_is_roa(v1.val.net))
+    ARG(1, T_NET);
+    if (!net_is_roa(v1.val.net))
       runtime( "ROA expected" );
 
     res.type = T_INT;
@@ -1263,8 +1238,8 @@ interpret(struct f_inst *what)
       ((net_addr_roa6 *) v1.val.net)->max_pxlen;
     break;
   case FI_ROA_ASN: 	/* Get ROA ASN */
-    ONEARG;
-    if (v1.type != T_NET || !net_is_roa(v1.val.net))
+    ARG(1, T_NET);
+    if (!net_is_roa(v1.val.net))
       runtime( "ROA expected" );
 
     res.type = T_INT;
@@ -1273,25 +1248,20 @@ interpret(struct f_inst *what)
       ((net_addr_roa6 *) v1.val.net)->asn;
     break;
   case FI_IP:	/* Convert prefix to ... */
-    ONEARG;
-    if (v1.type != T_NET)
-      runtime( "Prefix expected" );
+    ARG(1, T_NET);
     res.type = T_IP;
     res.val.ip = net_prefix(v1.val.net);
     break;
   case FI_ROUTE_DISTINGUISHER:
-    ONEARG;
-    if (v1.type != T_NET)
-      runtime( "Prefix expected" );
+    ARG(1, T_NET);
+    res.type = T_IP;
     if (!net_is_vpn(v1.val.net))
       runtime( "VPN address expected" );
     res.type = T_RD;
     res.val.ec = net_rd(v1.val.net);
     break;
   case FI_AS_PATH_FIRST:	/* Get first ASN from AS PATH */
-    ONEARG;
-    if (v1.type != T_PATH)
-      runtime( "AS path expected" );
+    ARG(1, T_PATH);
 
     as = 0;
     as_path_get_first(v1.val.ad, &as);
@@ -1299,9 +1269,7 @@ interpret(struct f_inst *what)
     res.val.i = as;
     break;
   case FI_AS_PATH_LAST:	/* Get last ASN from AS PATH */
-    ONEARG;
-    if (v1.type != T_PATH)
-      runtime( "AS path expected" );
+    ARG(1, T_PATH);
 
     as = 0;
     as_path_get_last(v1.val.ad, &as);
@@ -1309,20 +1277,18 @@ interpret(struct f_inst *what)
     res.val.i = as;
     break;
   case FI_AS_PATH_LAST_NAG:	/* Get last ASN from non-aggregated part of AS PATH */
-    ONEARG;
-    if (v1.type != T_PATH)
-      runtime( "AS path expected" );
+    ARG(1, T_PATH);
 
     res.type = T_INT;
     res.val.i = as_path_get_last_nonaggregated(v1.val.ad);
     break;
   case FI_RETURN:
-    ONEARG;
+    ARG_ANY(1);
     res = v1;
     res.type |= T_RETURN;
     return res;
   case FI_CALL: /* CALL: this is special: if T_RETURN and returning some value, mask it out  */
-    ONEARG;
+    ARG_ANY(1);
     res = interpret(what->a2.p);
     if (res.type == T_RETURN)
       return res;
@@ -1333,7 +1299,7 @@ interpret(struct f_inst *what)
       ((struct f_val *) sym->def)->type = T_VOID;
     break;
   case FI_SWITCH:
-    ONEARG;
+    ARG_ANY(1);
     {
       struct f_tree *t = find_tree(what->a2.p, v1);
       if (!t) {
@@ -1352,11 +1318,8 @@ interpret(struct f_inst *what)
     }
     break;
   case FI_IP_MASK: /* IP.MASK(val) */
-    TWOARGS;
-    if (v2.type != T_INT)
-      runtime( "Integer expected");
-    if (v1.type != T_IP)
-      runtime( "You can mask only IP addresses" );
+    ARG(1, T_IP);
+    ARG(2, T_INT);
 
     res.type = T_IP;
     res.val.ip = ipa_is_ip4(v1.val.ip) ?
@@ -1369,18 +1332,16 @@ interpret(struct f_inst *what)
     res.val.ad = adata_empty(f_pool, 0);
     break;
   case FI_PATH_PREPEND:	/* Path prepend */
-    TWOARGS;
-    if (v1.type != T_PATH)
-      runtime("Can't prepend to non-path");
-    if (v2.type != T_INT)
-      runtime("Can't prepend non-integer");
+    ARG(1, T_PATH);
+    ARG(2, T_INT);
 
     res.type = T_PATH;
     res.val.ad = as_path_prepend(f_pool, v1.val.ad, v2.val.i);
     break;
 
   case FI_CLIST_ADD_DEL:	/* (Extended) Community list add or delete */
-    TWOARGS;
+    ARG_ANY(1);
+    ARG_ANY(2);
     if (v1.type == T_PATH)
     {
       struct f_tree *set = NULL;
@@ -1548,9 +1509,8 @@ interpret(struct f_inst *what)
   case FI_ROA_CHECK:	/* ROA Check */
     if (what->arg1)
     {
-      TWOARGS;
-      if ((v1.type != T_NET) || (v2.type != T_INT))
-	runtime("Invalid argument to roa_check()");
+      ARG(1, T_NET);
+      ARG(2, T_INT);
 
       as = v2.val.i;
     }
@@ -1586,17 +1546,14 @@ interpret(struct f_inst *what)
     break;
 
   case FI_FORMAT:	/* Format */
-    ONEARG;
+    ARG_ANY(1);
 
     res.type = T_STRING;
     res.val.s = val_format_str(v1);
     break;
 
   case FI_ASSERT:	/* Birdtest Assert */
-    ONEARG;
-
-    if (v1.type != T_BOOL)
-      runtime("Should be boolean value");
+    ARG(1, T_BOOL);
 
     res.type = v1.type;
     res.val = v1.val;
@@ -1611,13 +1568,15 @@ interpret(struct f_inst *what)
 }
 
 #undef ARG
-#define ARG(x,y) \
-	if (!i_same(f1->y, f2->y)) \
+#undef ARG_ANY
+
+#define ARG(n) \
+	if (!i_same(f1->a##n.p, f2->a##n.p)) \
 		return 0;
 
-#define ONEARG ARG(v1, a1.p)
-#define TWOARGS ARG(v1, a1.p) \
-		ARG(v2, a2.p)
+#define ONEARG		ARG(1);
+#define TWOARGS		ONEARG; ARG(2);
+#define THREEARGS	TWOARGS; ARG(3);
 
 #define A2_SAME if (f1->a2.i != f2->a2.i) return 0;
 
@@ -1665,7 +1624,7 @@ i_same(struct f_inst *f1, struct f_inst *f2)
     break;
 
   case FI_SET:
-    ARG(v2, a2.p);
+    ARG(2);
     {
       struct symbol *s1, *s2;
       s1 = f1->a1.p;
diff --git a/filter/filter.h b/filter/filter.h
index 1e5fc8ff..909d09b1 100644
--- a/filter/filter.h
+++ b/filter/filter.h
@@ -81,6 +81,8 @@ FI__LIST
   FI__MAX,
 } PACKED;
 
+const char *f_instruction_name(enum f_instruction_code fi);
+
 struct f_inst {		/* Instruction */
   struct f_inst *next;	/* Structure is 16 bytes, anyway */
   enum f_instruction_code fi_code;

From c3becfe1934da2dc2c0881a71eac8a26f810791f Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Wed, 2 May 2018 12:34:35 +0200
Subject: [PATCH 169/198] Filter: macro for recursive interpretation of
 instructions

---
 filter/filter.c | 17 +++++++++--------
 1 file changed, 9 insertions(+), 8 deletions(-)

diff --git a/filter/filter.c b/filter/filter.c
index 2b22bfa3..66a9e05e 100644
--- a/filter/filter.c
+++ b/filter/filter.c
@@ -590,16 +590,18 @@ static struct tbf rl_runtime_err = TBF_DEFAULT_LOG_LIMITS;
     return res; \
   } while(0)
 
-#define ARG_ANY(n) \
-    v##n = interpret(what->a##n.p); \
-    if (v##n.type & T_RETURN) \
-      return v##n;
+#define ARG_ANY(n) INTERPRET(v##n, what->a##n.p)
 
 #define ARG(n,t) ARG_ANY(n) \
     if (v##n.type != t) \
       runtime("Argument %d of instruction %s must be of type %02x, got %02x", \
 	  n, f_instruction_name(what->fi_code), t, v##n.type);
 
+#define INTERPRET(val, what_) \
+    val = interpret(what_); \
+    if (val.type & T_RETURN) \
+      return val;
+
 #define ACCESS_RTE \
   do { if (!f_rte) runtime("No route to access"); } while (0)
 
@@ -747,7 +749,8 @@ interpret(struct f_inst *what)
       while (tt) {
 	*vv = lp_alloc(f_pool, sizeof(struct f_path_mask));
 	if (tt->kind == PM_ASN_EXPR) {
-	  struct f_val res = interpret((struct f_inst *) tt->val);
+	  struct f_val res;
+	  INTERPRET(res, (struct f_inst *) tt->val);
 	  (*vv)->kind = PM_ASN;
 	  if (res.type != T_INT) {
 	    runtime( "Error resolving path mask template: value not an integer" );
@@ -1312,9 +1315,7 @@ interpret(struct f_inst *what)
       }
       /* It is actually possible to have t->data NULL */
 
-      res = interpret(t->data);
-      if (res.type & T_RETURN)
-	return res;
+      INTERPRET(res, t->data);
     }
     break;
   case FI_IP_MASK: /* IP.MASK(val) */

From ee7e2ffd265fd76dbc8c94d9c2d48da54c27ff76 Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Mon, 7 May 2018 14:47:00 +0200
Subject: [PATCH 170/198] Protocol: Introducing an enum protocol_class

This supersedes the EAP_* constants.
---
 filter/filter.c        |  2 +-
 nest/proto.c           |  9 +++----
 nest/protocol.h        | 21 +++++++++++++++-
 nest/route.h           | 16 ++----------
 nest/rt-attr.c         |  5 +---
 nest/rt-dev.c          |  1 +
 proto/babel/babel.c    |  2 +-
 proto/babel/babel.h    |  4 +--
 proto/bfd/bfd.c        |  1 +
 proto/bgp/attrs.c      | 56 +++++++++++++++++++++---------------------
 proto/bgp/bgp.c        |  2 +-
 proto/bgp/bgp.h        |  2 +-
 proto/bgp/config.Y     | 24 +++++++++---------
 proto/ospf/ospf.c      |  2 +-
 proto/ospf/ospf.h      |  8 +++---
 proto/pipe/pipe.c      |  1 +
 proto/radv/radv.c      |  2 +-
 proto/radv/radv.h      |  4 +--
 proto/rip/rip.c        |  2 +-
 proto/rip/rip.h        |  4 +--
 proto/rpki/rpki.c      |  1 +
 proto/static/static.c  |  1 +
 sysdep/linux/krt-sys.h | 38 ++++++++++++++--------------
 sysdep/linux/netlink.c |  2 +-
 sysdep/unix/krt.c      |  3 ++-
 sysdep/unix/krt.h      |  4 +--
 test/bt-utils.c        |  4 +--
 27 files changed, 115 insertions(+), 106 deletions(-)

diff --git a/filter/filter.c b/filter/filter.c
index 66a9e05e..6381550e 100644
--- a/filter/filter.c
+++ b/filter/filter.c
@@ -1522,7 +1522,7 @@ interpret(struct f_inst *what)
 
       /* We ignore temporary attributes, probably not a problem here */
       /* 0x02 is a value of BA_AS_PATH, we don't want to include BGP headers */
-      eattr *e = ea_find((*f_rte)->attrs->eattrs, EA_CODE(EAP_BGP, 0x02));
+      eattr *e = ea_find((*f_rte)->attrs->eattrs, EA_CODE(PROTOCOL_BGP, 0x02));
 
       if (!e || e->type != EAF_TYPE_AS_PATH)
 	runtime("Missing AS_PATH attribute");
diff --git a/nest/proto.c b/nest/proto.c
index 15d6f4de..49f71304 100644
--- a/nest/proto.c
+++ b/nest/proto.c
@@ -25,6 +25,7 @@ pool *proto_pool;
 list  proto_list;
 
 static list protocol_list;
+struct protocol *class_to_protocol[PROTOCOL__MAX];
 
 #define PD(pr, msg, args...) do { if (pr->debug & D_STATES) { log(L_TRACE "%s: " msg, pr->name , ## args); } } while(0)
 
@@ -1256,11 +1257,9 @@ void
 proto_build(struct protocol *p)
 {
   add_tail(&protocol_list, &p->n);
-  if (p->attr_class)
-    {
-      ASSERT(!attr_class_to_protocol[p->attr_class]);
-      attr_class_to_protocol[p->attr_class] = p;
-    }
+  ASSERT(p->class);
+  ASSERT(!class_to_protocol[p->class]);
+  class_to_protocol[p->class] = p;
 }
 
 /* FIXME: convert this call to some protocol hook */
diff --git a/nest/protocol.h b/nest/protocol.h
index 8a22d76b..d790e90e 100644
--- a/nest/protocol.h
+++ b/nest/protocol.h
@@ -37,12 +37,31 @@ struct symbol;
  *	Routing Protocol
  */
 
+enum protocol_class {
+  PROTOCOL_NONE,
+  PROTOCOL_BABEL,
+  PROTOCOL_BFD,
+  PROTOCOL_BGP,
+  PROTOCOL_DEVICE,
+  PROTOCOL_DIRECT,
+  PROTOCOL_KERNEL,
+  PROTOCOL_OSPF,
+  PROTOCOL_PIPE,
+  PROTOCOL_RADV,
+  PROTOCOL_RIP,
+  PROTOCOL_RPKI,
+  PROTOCOL_STATIC,
+  PROTOCOL__MAX
+};
+
+extern struct protocol *class_to_protocol[PROTOCOL__MAX];
+
 struct protocol {
   node n;
   char *name;
   char *template;			/* Template for automatic generation of names */
   int name_counter;			/* Counter for automatic name generation */
-  int attr_class;			/* Attribute class known to this protocol */
+  enum protocol_class class;		/* Machine readable protocol class */
   uint preference;			/* Default protocol preference */
   uint channel_mask;			/* Mask of accepted channel types (NB_*) */
   uint proto_size;			/* Size of protocol data structure */
diff --git a/nest/route.h b/nest/route.h
index 79127519..1391b357 100644
--- a/nest/route.h
+++ b/nest/route.h
@@ -457,7 +457,7 @@ static inline int rte_is_reachable(rte *r)
  */
 
 typedef struct eattr {
-  word id;				/* EA_CODE(EAP_..., protocol-dependent ID) */
+  word id;				/* EA_CODE(PROTOCOL_..., protocol-dependent ID) */
   byte flags;				/* Protocol-dependent flags */
   byte type;				/* Attribute type and several flags (EAF_...) */
   union {
@@ -466,20 +466,11 @@ typedef struct eattr {
   } u;
 } eattr;
 
-#define EAP_GENERIC 0			/* Generic attributes */
-#define EAP_BGP 1			/* BGP attributes */
-#define EAP_RIP 2			/* RIP */
-#define EAP_OSPF 3			/* OSPF */
-#define EAP_KRT 4			/* Kernel route attributes */
-#define EAP_BABEL 5			/* Babel attributes */
-#define EAP_RADV 6			/* Router advertisment attributes */
-#define EAP_MAX 7
-
 #define EA_CODE(proto,id) (((proto) << 8) | (id))
 #define EA_PROTO(ea) ((ea) >> 8)
 #define EA_ID(ea) ((ea) & 0xff)
 
-#define EA_GEN_IGP_METRIC EA_CODE(EAP_GENERIC, 0)
+#define EA_GEN_IGP_METRIC EA_CODE(PROTOCOL_NONE, 0)
 
 #define EA_CODE_MASK 0xffff
 #define EA_ALLOW_UNDEF 0x10000		/* ea_find: allow EAF_TYPE_UNDEF */
@@ -656,9 +647,6 @@ 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--; }
 
-
-extern struct protocol *attr_class_to_protocol[EAP_MAX];
-
 /*
  *	Default protocol preferences
  */
diff --git a/nest/rt-attr.c b/nest/rt-attr.c
index 881687de..f92efc2e 100644
--- a/nest/rt-attr.c
+++ b/nest/rt-attr.c
@@ -88,9 +88,6 @@ static struct idm src_ids;
 
 static HASH(struct rte_src) src_hash;
 
-struct protocol *attr_class_to_protocol[EAP_MAX];
-
-
 static void
 rte_src_init(void)
 {
@@ -851,7 +848,7 @@ ea_show(struct cli *c, eattr *e)
   byte buf[CLI_MSG_SIZE];
   byte *pos = buf, *end = buf + sizeof(buf);
 
-  if (p = attr_class_to_protocol[EA_PROTO(e->id)])
+  if (p = class_to_protocol[EA_PROTO(e->id)])
     {
       pos += bsprintf(pos, "%s.", p->name);
       if (p->get_attr)
diff --git a/nest/rt-dev.c b/nest/rt-dev.c
index 66f458e7..61f025ce 100644
--- a/nest/rt-dev.c
+++ b/nest/rt-dev.c
@@ -185,6 +185,7 @@ dev_copy_config(struct proto_config *dest, struct proto_config *src)
 struct protocol proto_device = {
   .name =		"Direct",
   .template =		"direct%d",
+  .class =		PROTOCOL_DIRECT,
   .preference =		DEF_PREF_DIRECT,
   .channel_mask =	NB_IP | NB_IP6_SADR,
   .proto_size =		sizeof(struct rt_dev_proto),
diff --git a/proto/babel/babel.c b/proto/babel/babel.c
index 44c6adb8..83986cb9 100644
--- a/proto/babel/babel.c
+++ b/proto/babel/babel.c
@@ -2312,7 +2312,7 @@ babel_reconfigure(struct proto *P, struct proto_config *CF)
 struct protocol proto_babel = {
   .name =		"Babel",
   .template =		"babel%d",
-  .attr_class =		EAP_BABEL,
+  .class =		PROTOCOL_BABEL,
   .preference =		DEF_PREF_BABEL,
   .channel_mask =	NB_IP | NB_IP6_SADR,
   .proto_size =		sizeof(struct babel_proto),
diff --git a/proto/babel/babel.h b/proto/babel/babel.h
index e5c9cd5b..14765c60 100644
--- a/proto/babel/babel.h
+++ b/proto/babel/babel.h
@@ -25,8 +25,8 @@
 #include "lib/string.h"
 #include "lib/timer.h"
 
-#define EA_BABEL_METRIC		EA_CODE(EAP_BABEL, 0)
-#define EA_BABEL_ROUTER_ID	EA_CODE(EAP_BABEL, 1)
+#define EA_BABEL_METRIC		EA_CODE(PROTOCOL_BABEL, 0)
+#define EA_BABEL_ROUTER_ID	EA_CODE(PROTOCOL_BABEL, 1)
 
 #define BABEL_MAGIC		42
 #define BABEL_VERSION		2
diff --git a/proto/bfd/bfd.c b/proto/bfd/bfd.c
index 67ec2270..64d5007b 100644
--- a/proto/bfd/bfd.c
+++ b/proto/bfd/bfd.c
@@ -1116,6 +1116,7 @@ bfd_show_sessions(struct proto *P)
 struct protocol proto_bfd = {
   .name =		"BFD",
   .template =		"bfd%d",
+  .class =		PROTOCOL_BFD,
   .proto_size =		sizeof(struct bfd_proto),
   .config_size =	sizeof(struct bfd_config),
   .init =		bfd_init,
diff --git a/proto/bgp/attrs.c b/proto/bgp/attrs.c
index 9003feb2..5695e1c1 100644
--- a/proto/bgp/attrs.c
+++ b/proto/bgp/attrs.c
@@ -92,7 +92,7 @@ bgp_set_attr(ea_list **attrs, struct linpool *pool, uint code, uint flags, uintp
   a->next = *attrs;
   *attrs = a;
 
-  e->id = EA_CODE(EAP_BGP, code);
+  e->id = EA_CODE(PROTOCOL_BGP, code);
   e->type = bgp_attr_table[code].type;
   e->flags = flags;
 
@@ -702,7 +702,7 @@ static inline void
 bgp_decode_unknown(struct bgp_parse_state *s, uint code, uint flags, byte *data, uint len, ea_list **to)
 {
   /* Cannot use bgp_set_attr_data() as it works on known attributes only */
-  ea_set_attr_data(to, s->pool, EA_CODE(EAP_BGP, code), flags, EAF_TYPE_OPAQUE, data, len);
+  ea_set_attr_data(to, s->pool, EA_CODE(PROTOCOL_BGP, code), flags, EAF_TYPE_OPAQUE, data, len);
 }
 
 
@@ -857,7 +857,7 @@ bgp_attr_known(uint code)
 static inline void
 bgp_export_attr(struct bgp_export_state *s, eattr *a, ea_list *to)
 {
-  if (EA_PROTO(a->id) != EAP_BGP)
+  if (EA_PROTO(a->id) != PROTOCOL_BGP)
     return;
 
   uint code = EA_ID(a->id);
@@ -937,7 +937,7 @@ bgp_export_attrs(struct bgp_export_state *s, ea_list *attrs)
 static inline int
 bgp_encode_attr(struct bgp_write_state *s, eattr *a, byte *buf, uint size)
 {
-  ASSERT(EA_PROTO(a->id) == EAP_BGP);
+  ASSERT(EA_PROTO(a->id) == PROTOCOL_BGP);
 
   uint code = EA_ID(a->id);
 
@@ -1405,7 +1405,7 @@ bgp_import_control(struct proto *P, rte **new, ea_list **attrs UNUSED, struct li
   /* Handle well-known communities, RFC 1997 */
   struct eattr *c;
   if (p->cf->interpret_communities &&
-      (c = ea_find(e->attrs->eattrs, EA_CODE(EAP_BGP, BA_COMMUNITY))))
+      (c = ea_find(e->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_COMMUNITY))))
   {
     struct adata *d = c->u.ptr;
 
@@ -1570,7 +1570,7 @@ bgp_rt_notify(struct proto *P, struct channel *C, net *n, rte *new, rte *old, ea
 static inline u32
 bgp_get_neighbor(rte *r)
 {
-  eattr *e = ea_find(r->attrs->eattrs, EA_CODE(EAP_BGP, BA_AS_PATH));
+  eattr *e = ea_find(r->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_AS_PATH));
   u32 as;
 
   if (e && as_path_get_first_regular(e->u.ptr, &as))
@@ -1612,8 +1612,8 @@ bgp_rte_better(rte *new, rte *old)
     return 0;
 
   /* Start with local preferences */
-  x = ea_find(new->attrs->eattrs, EA_CODE(EAP_BGP, BA_LOCAL_PREF));
-  y = ea_find(old->attrs->eattrs, EA_CODE(EAP_BGP, BA_LOCAL_PREF));
+  x = ea_find(new->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_LOCAL_PREF));
+  y = ea_find(old->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_LOCAL_PREF));
   n = x ? x->u.data : new_bgp->cf->default_local_pref;
   o = y ? y->u.data : old_bgp->cf->default_local_pref;
   if (n > o)
@@ -1624,8 +1624,8 @@ bgp_rte_better(rte *new, rte *old)
   /* RFC 4271 9.1.2.2. a)  Use AS path lengths */
   if (new_bgp->cf->compare_path_lengths || old_bgp->cf->compare_path_lengths)
   {
-    x = ea_find(new->attrs->eattrs, EA_CODE(EAP_BGP, BA_AS_PATH));
-    y = ea_find(old->attrs->eattrs, EA_CODE(EAP_BGP, BA_AS_PATH));
+    x = ea_find(new->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_AS_PATH));
+    y = ea_find(old->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_AS_PATH));
     n = x ? as_path_getlen(x->u.ptr) : AS_PATH_MAXLEN;
     o = y ? as_path_getlen(y->u.ptr) : AS_PATH_MAXLEN;
     if (n < o)
@@ -1635,8 +1635,8 @@ bgp_rte_better(rte *new, rte *old)
   }
 
   /* RFC 4271 9.1.2.2. b) Use origins */
-  x = ea_find(new->attrs->eattrs, EA_CODE(EAP_BGP, BA_ORIGIN));
-  y = ea_find(old->attrs->eattrs, EA_CODE(EAP_BGP, BA_ORIGIN));
+  x = ea_find(new->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_ORIGIN));
+  y = ea_find(old->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_ORIGIN));
   n = x ? x->u.data : ORIGIN_INCOMPLETE;
   o = y ? y->u.data : ORIGIN_INCOMPLETE;
   if (n < o)
@@ -1658,8 +1658,8 @@ bgp_rte_better(rte *new, rte *old)
   if (new_bgp->cf->med_metric || old_bgp->cf->med_metric ||
       (bgp_get_neighbor(new) == bgp_get_neighbor(old)))
   {
-    x = ea_find(new->attrs->eattrs, EA_CODE(EAP_BGP, BA_MULTI_EXIT_DISC));
-    y = ea_find(old->attrs->eattrs, EA_CODE(EAP_BGP, BA_MULTI_EXIT_DISC));
+    x = ea_find(new->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_MULTI_EXIT_DISC));
+    y = ea_find(old->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_MULTI_EXIT_DISC));
     n = x ? x->u.data : new_bgp->cf->default_med;
     o = y ? y->u.data : old_bgp->cf->default_med;
     if (n < o)
@@ -1684,8 +1684,8 @@ bgp_rte_better(rte *new, rte *old)
 
   /* RFC 4271 9.1.2.2. f) Compare BGP identifiers */
   /* RFC 4456 9. a) Use ORIGINATOR_ID instead of local neighbor ID */
-  x = ea_find(new->attrs->eattrs, EA_CODE(EAP_BGP, BA_ORIGINATOR_ID));
-  y = ea_find(old->attrs->eattrs, EA_CODE(EAP_BGP, BA_ORIGINATOR_ID));
+  x = ea_find(new->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_ORIGINATOR_ID));
+  y = ea_find(old->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_ORIGINATOR_ID));
   n = x ? x->u.data : new_bgp->remote_id;
   o = y ? y->u.data : old_bgp->remote_id;
 
@@ -1702,8 +1702,8 @@ bgp_rte_better(rte *new, rte *old)
     return 0;
 
   /* RFC 4456 9. b) Compare cluster list lengths */
-  x = ea_find(new->attrs->eattrs, EA_CODE(EAP_BGP, BA_CLUSTER_LIST));
-  y = ea_find(old->attrs->eattrs, EA_CODE(EAP_BGP, BA_CLUSTER_LIST));
+  x = ea_find(new->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_CLUSTER_LIST));
+  y = ea_find(old->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_CLUSTER_LIST));
   n = x ? int_set_get_size(x->u.ptr) : 0;
   o = y ? int_set_get_size(y->u.ptr) : 0;
   if (n < o)
@@ -1733,8 +1733,8 @@ bgp_rte_mergable(rte *pri, rte *sec)
     return 0;
 
   /* Start with local preferences */
-  x = ea_find(pri->attrs->eattrs, EA_CODE(EAP_BGP, BA_LOCAL_PREF));
-  y = ea_find(sec->attrs->eattrs, EA_CODE(EAP_BGP, BA_LOCAL_PREF));
+  x = ea_find(pri->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_LOCAL_PREF));
+  y = ea_find(sec->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_LOCAL_PREF));
   p = x ? x->u.data : pri_bgp->cf->default_local_pref;
   s = y ? y->u.data : sec_bgp->cf->default_local_pref;
   if (p != s)
@@ -1743,8 +1743,8 @@ bgp_rte_mergable(rte *pri, rte *sec)
   /* RFC 4271 9.1.2.2. a)  Use AS path lengths */
   if (pri_bgp->cf->compare_path_lengths || sec_bgp->cf->compare_path_lengths)
   {
-    x = ea_find(pri->attrs->eattrs, EA_CODE(EAP_BGP, BA_AS_PATH));
-    y = ea_find(sec->attrs->eattrs, EA_CODE(EAP_BGP, BA_AS_PATH));
+    x = ea_find(pri->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_AS_PATH));
+    y = ea_find(sec->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_AS_PATH));
     p = x ? as_path_getlen(x->u.ptr) : AS_PATH_MAXLEN;
     s = y ? as_path_getlen(y->u.ptr) : AS_PATH_MAXLEN;
 
@@ -1756,8 +1756,8 @@ bgp_rte_mergable(rte *pri, rte *sec)
   }
 
   /* RFC 4271 9.1.2.2. b) Use origins */
-  x = ea_find(pri->attrs->eattrs, EA_CODE(EAP_BGP, BA_ORIGIN));
-  y = ea_find(sec->attrs->eattrs, EA_CODE(EAP_BGP, BA_ORIGIN));
+  x = ea_find(pri->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_ORIGIN));
+  y = ea_find(sec->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_ORIGIN));
   p = x ? x->u.data : ORIGIN_INCOMPLETE;
   s = y ? y->u.data : ORIGIN_INCOMPLETE;
   if (p != s)
@@ -1767,8 +1767,8 @@ bgp_rte_mergable(rte *pri, rte *sec)
   if (pri_bgp->cf->med_metric || sec_bgp->cf->med_metric ||
       (bgp_get_neighbor(pri) == bgp_get_neighbor(sec)))
   {
-    x = ea_find(pri->attrs->eattrs, EA_CODE(EAP_BGP, BA_MULTI_EXIT_DISC));
-    y = ea_find(sec->attrs->eattrs, EA_CODE(EAP_BGP, BA_MULTI_EXIT_DISC));
+    x = ea_find(pri->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_MULTI_EXIT_DISC));
+    y = ea_find(sec->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_MULTI_EXIT_DISC));
     p = x ? x->u.data : pri_bgp->cf->default_med;
     s = y ? y->u.data : sec_bgp->cf->default_med;
     if (p != s)
@@ -2009,8 +2009,8 @@ bgp_get_attr(eattr *a, byte *buf, int buflen)
 void
 bgp_get_route_info(rte *e, byte *buf, ea_list *attrs)
 {
-  eattr *p = ea_find(attrs, EA_CODE(EAP_BGP, BA_AS_PATH));
-  eattr *o = ea_find(attrs, EA_CODE(EAP_BGP, BA_ORIGIN));
+  eattr *p = ea_find(attrs, EA_CODE(PROTOCOL_BGP, BA_AS_PATH));
+  eattr *o = ea_find(attrs, EA_CODE(PROTOCOL_BGP, BA_ORIGIN));
   u32 origas;
 
   buf += bsprintf(buf, " (%d", e->pref);
diff --git a/proto/bgp/bgp.c b/proto/bgp/bgp.c
index 9db26050..932ad9f3 100644
--- a/proto/bgp/bgp.c
+++ b/proto/bgp/bgp.c
@@ -2148,7 +2148,7 @@ struct channel_class channel_bgp = {
 struct protocol proto_bgp = {
   .name = 		"BGP",
   .template = 		"bgp%d",
-  .attr_class = 	EAP_BGP,
+  .class =		PROTOCOL_BGP,
   .preference = 	DEF_PREF_BGP,
   .channel_mask =	NB_IP | NB_VPN | NB_FLOW,
   .proto_size =		sizeof(struct bgp_proto),
diff --git a/proto/bgp/bgp.h b/proto/bgp/bgp.h
index 30424abb..de05dcfb 100644
--- a/proto/bgp/bgp.h
+++ b/proto/bgp/bgp.h
@@ -462,7 +462,7 @@ struct rte_source *bgp_get_source(struct bgp_proto *p, u32 path_id);
 static inline eattr *
 bgp_find_attr(ea_list *attrs, uint code)
 {
-  return ea_find(attrs, EA_CODE(EAP_BGP, code));
+  return ea_find(attrs, EA_CODE(PROTOCOL_BGP, code));
 }
 
 eattr *
diff --git a/proto/bgp/config.Y b/proto/bgp/config.Y
index 41eaa729..c731c4c7 100644
--- a/proto/bgp/config.Y
+++ b/proto/bgp/config.Y
@@ -253,29 +253,29 @@ bgp_proto_channel: bgp_channel_start bgp_channel_opt_list bgp_channel_end;
 
 
 CF_ADDTO(dynamic_attr, BGP_ORIGIN
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_ENUM_BGP_ORIGIN, EA_CODE(EAP_BGP, BA_ORIGIN)); })
+	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_ENUM_BGP_ORIGIN, EA_CODE(PROTOCOL_BGP, BA_ORIGIN)); })
 CF_ADDTO(dynamic_attr, BGP_PATH
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_AS_PATH, T_PATH, EA_CODE(EAP_BGP, BA_AS_PATH)); })
+	{ $$ = f_new_dynamic_attr(EAF_TYPE_AS_PATH, T_PATH, EA_CODE(PROTOCOL_BGP, BA_AS_PATH)); })
 CF_ADDTO(dynamic_attr, BGP_NEXT_HOP
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_IP_ADDRESS, T_IP, EA_CODE(EAP_BGP, BA_NEXT_HOP)); })
+	{ $$ = f_new_dynamic_attr(EAF_TYPE_IP_ADDRESS, T_IP, EA_CODE(PROTOCOL_BGP, BA_NEXT_HOP)); })
 CF_ADDTO(dynamic_attr, BGP_MED
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_CODE(EAP_BGP, BA_MULTI_EXIT_DISC)); })
+	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_CODE(PROTOCOL_BGP, BA_MULTI_EXIT_DISC)); })
 CF_ADDTO(dynamic_attr, BGP_LOCAL_PREF
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_CODE(EAP_BGP, BA_LOCAL_PREF)); })
+	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_CODE(PROTOCOL_BGP, BA_LOCAL_PREF)); })
 CF_ADDTO(dynamic_attr, BGP_ATOMIC_AGGR
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_OPAQUE, T_ENUM_EMPTY, EA_CODE(EAP_BGP, BA_ATOMIC_AGGR)); })
+	{ $$ = f_new_dynamic_attr(EAF_TYPE_OPAQUE, T_ENUM_EMPTY, EA_CODE(PROTOCOL_BGP, BA_ATOMIC_AGGR)); })
 CF_ADDTO(dynamic_attr, BGP_AGGREGATOR
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_CODE(EAP_BGP, BA_AGGREGATOR)); })
+	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_CODE(PROTOCOL_BGP, BA_AGGREGATOR)); })
 CF_ADDTO(dynamic_attr, BGP_COMMUNITY
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT_SET, T_CLIST, EA_CODE(EAP_BGP, BA_COMMUNITY)); })
+	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT_SET, T_CLIST, EA_CODE(PROTOCOL_BGP, BA_COMMUNITY)); })
 CF_ADDTO(dynamic_attr, BGP_ORIGINATOR_ID
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_ROUTER_ID, T_QUAD, EA_CODE(EAP_BGP, BA_ORIGINATOR_ID)); })
+	{ $$ = f_new_dynamic_attr(EAF_TYPE_ROUTER_ID, T_QUAD, EA_CODE(PROTOCOL_BGP, BA_ORIGINATOR_ID)); })
 CF_ADDTO(dynamic_attr, BGP_CLUSTER_LIST
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT_SET, T_CLIST, EA_CODE(EAP_BGP, BA_CLUSTER_LIST)); })
+	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT_SET, T_CLIST, EA_CODE(PROTOCOL_BGP, BA_CLUSTER_LIST)); })
 CF_ADDTO(dynamic_attr, BGP_EXT_COMMUNITY
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_EC_SET, T_ECLIST, EA_CODE(EAP_BGP, BA_EXT_COMMUNITY)); })
+	{ $$ = f_new_dynamic_attr(EAF_TYPE_EC_SET, T_ECLIST, EA_CODE(PROTOCOL_BGP, BA_EXT_COMMUNITY)); })
 CF_ADDTO(dynamic_attr, BGP_LARGE_COMMUNITY
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_LC_SET, T_LCLIST, EA_CODE(EAP_BGP, BA_LARGE_COMMUNITY)); })
+	{ $$ = f_new_dynamic_attr(EAF_TYPE_LC_SET, T_LCLIST, EA_CODE(PROTOCOL_BGP, BA_LARGE_COMMUNITY)); })
 
 
 
diff --git a/proto/ospf/ospf.c b/proto/ospf/ospf.c
index 0a9efd19..d9edc3e7 100644
--- a/proto/ospf/ospf.c
+++ b/proto/ospf/ospf.c
@@ -1463,7 +1463,7 @@ ospf_sh_lsadb(struct lsadb_show_data *ld)
 struct protocol proto_ospf = {
   .name =		"OSPF",
   .template =		"ospf%d",
-  .attr_class =		EAP_OSPF,
+  .class =		PROTOCOL_OSPF,
   .preference =		DEF_PREF_OSPF,
   .channel_mask =	NB_IP,
   .proto_size =		sizeof(struct ospf_proto),
diff --git a/proto/ospf/ospf.h b/proto/ospf/ospf.h
index f26ed99c..ff55621a 100644
--- a/proto/ospf/ospf.h
+++ b/proto/ospf/ospf.h
@@ -863,10 +863,10 @@ struct lsadb_show_data {
 };
 
 
-#define EA_OSPF_METRIC1	EA_CODE(EAP_OSPF, 0)
-#define EA_OSPF_METRIC2	EA_CODE(EAP_OSPF, 1)
-#define EA_OSPF_TAG	EA_CODE(EAP_OSPF, 2)
-#define EA_OSPF_ROUTER_ID EA_CODE(EAP_OSPF, 3)
+#define EA_OSPF_METRIC1	EA_CODE(PROTOCOL_OSPF, 0)
+#define EA_OSPF_METRIC2	EA_CODE(PROTOCOL_OSPF, 1)
+#define EA_OSPF_TAG	EA_CODE(PROTOCOL_OSPF, 2)
+#define EA_OSPF_ROUTER_ID EA_CODE(PROTOCOL_OSPF, 3)
 
 
 /* ospf.c */
diff --git a/proto/pipe/pipe.c b/proto/pipe/pipe.c
index 49ff52e2..ffc677f4 100644
--- a/proto/pipe/pipe.c
+++ b/proto/pipe/pipe.c
@@ -275,6 +275,7 @@ pipe_show_proto_info(struct proto *P)
 struct protocol proto_pipe = {
   .name =		"Pipe",
   .template =		"pipe%d",
+  .class =		PROTOCOL_PIPE,
   .proto_size =		sizeof(struct pipe_proto),
   .config_size =	sizeof(struct pipe_config),
   .postconfig =		pipe_postconfig,
diff --git a/proto/radv/radv.c b/proto/radv/radv.c
index 8a79dfaf..2f5f1c27 100644
--- a/proto/radv/radv.c
+++ b/proto/radv/radv.c
@@ -762,7 +762,7 @@ radv_get_attr(eattr *a, byte *buf, int buflen UNUSED)
 struct protocol proto_radv = {
   .name =		"RAdv",
   .template =		"radv%d",
-  .attr_class =		EAP_RADV,
+  .class =		PROTOCOL_RADV,
   .channel_mask =	NB_IP6,
   .proto_size =		sizeof(struct radv_proto),
   .config_size =	sizeof(struct radv_config),
diff --git a/proto/radv/radv.h b/proto/radv/radv.h
index 66f785a7..719948d5 100644
--- a/proto/radv/radv.h
+++ b/proto/radv/radv.h
@@ -192,8 +192,8 @@ struct radv_iface
 #define RA_PREF_MASK	0x18
 
 /* Attributes */
-#define EA_RA_PREFERENCE	EA_CODE(EAP_RADV, 0)
-#define EA_RA_LIFETIME		EA_CODE(EAP_RADV, 1)
+#define EA_RA_PREFERENCE	EA_CODE(PROTOCOL_RADV, 0)
+#define EA_RA_LIFETIME		EA_CODE(PROTOCOL_RADV, 1)
 
 #ifdef LOCAL_DEBUG
 #define RADV_FORCE_DEBUG 1
diff --git a/proto/rip/rip.c b/proto/rip/rip.c
index baf98737..adc2b471 100644
--- a/proto/rip/rip.c
+++ b/proto/rip/rip.c
@@ -1274,7 +1274,7 @@ rip_dump(struct proto *P)
 struct protocol proto_rip = {
   .name =		"RIP",
   .template =		"rip%d",
-  .attr_class =		EAP_RIP,
+  .class =		PROTOCOL_RIP,
   .preference =		DEF_PREF_RIP,
   .channel_mask =	NB_IP,
   .proto_size =		sizeof(struct rip_proto),
diff --git a/proto/rip/rip.h b/proto/rip/rip.h
index 55696333..2dc34862 100644
--- a/proto/rip/rip.h
+++ b/proto/rip/rip.h
@@ -182,8 +182,8 @@ struct rip_rte
 #define RIP_ENTRY_VALID		1	/* Valid outgoing route */
 #define RIP_ENTRY_STALE		2	/* Stale outgoing route, waiting for GC */
 
-#define EA_RIP_METRIC		EA_CODE(EAP_RIP, 0)
-#define EA_RIP_TAG		EA_CODE(EAP_RIP, 1)
+#define EA_RIP_METRIC		EA_CODE(PROTOCOL_RIP, 0)
+#define EA_RIP_TAG		EA_CODE(PROTOCOL_RIP, 1)
 
 static inline int rip_is_v2(struct rip_proto *p)
 { return p->rip2; }
diff --git a/proto/rpki/rpki.c b/proto/rpki/rpki.c
index 74860071..36097dc5 100644
--- a/proto/rpki/rpki.c
+++ b/proto/rpki/rpki.c
@@ -913,6 +913,7 @@ rpki_copy_config(struct proto_config *dest UNUSED, struct proto_config *src UNUS
 struct protocol proto_rpki = {
   .name = 		"RPKI",
   .template = 		"rpki%d",
+  .class =		PROTOCOL_RPKI,
   .preference = 	DEF_PREF_RPKI,
   .proto_size = 	sizeof(struct rpki_proto),
   .config_size =	sizeof(struct rpki_config),
diff --git a/proto/static/static.c b/proto/static/static.c
index ede4c734..8dfa6f35 100644
--- a/proto/static/static.c
+++ b/proto/static/static.c
@@ -656,6 +656,7 @@ static_show(struct proto *P)
 struct protocol proto_static = {
   .name =		"Static",
   .template =		"static%d",
+  .class =		PROTOCOL_STATIC,
   .preference =		DEF_PREF_STATIC,
   .channel_mask =	NB_ANY,
   .proto_size =		sizeof(struct static_proto),
diff --git a/sysdep/linux/krt-sys.h b/sysdep/linux/krt-sys.h
index 76ae29b7..2b8cdaa7 100644
--- a/sysdep/linux/krt-sys.h
+++ b/sysdep/linux/krt-sys.h
@@ -34,9 +34,9 @@ static inline struct ifa * kif_get_primary_ip(struct iface *i UNUSED) { return N
 
 #define KRT_ALLOW_MERGE_PATHS	1
 
-#define EA_KRT_PREFSRC		EA_CODE(EAP_KRT, 0x10)
-#define EA_KRT_REALM		EA_CODE(EAP_KRT, 0x11)
-#define EA_KRT_SCOPE		EA_CODE(EAP_KRT, 0x12)
+#define EA_KRT_PREFSRC		EA_CODE(PROTOCOL_KERNEL, 0x10)
+#define EA_KRT_REALM		EA_CODE(PROTOCOL_KERNEL, 0x11)
+#define EA_KRT_SCOPE		EA_CODE(PROTOCOL_KERNEL, 0x12)
 
 
 #define KRT_METRICS_MAX		0x10	/* RTAX_QUICKACK+1 */
@@ -48,22 +48,22 @@ static inline struct ifa * kif_get_primary_ip(struct iface *i UNUSED) { return N
  * Following attributes are parts of RTA_METRICS kernel route attribute, their
  * ids must be consistent with their RTAX_* constants (+ KRT_METRICS_OFFSET)
  */
-#define EA_KRT_METRICS		EA_CODE(EAP_KRT, 0x20)	/* Dummy one */
-#define EA_KRT_LOCK		EA_CODE(EAP_KRT, 0x21)
-#define EA_KRT_MTU		EA_CODE(EAP_KRT, 0x22)
-#define EA_KRT_WINDOW		EA_CODE(EAP_KRT, 0x23)
-#define EA_KRT_RTT		EA_CODE(EAP_KRT, 0x24)
-#define EA_KRT_RTTVAR		EA_CODE(EAP_KRT, 0x25)
-#define EA_KRT_SSTRESH		EA_CODE(EAP_KRT, 0x26)
-#define EA_KRT_CWND		EA_CODE(EAP_KRT, 0x27)
-#define EA_KRT_ADVMSS		EA_CODE(EAP_KRT, 0x28)
-#define EA_KRT_REORDERING 	EA_CODE(EAP_KRT, 0x29)
-#define EA_KRT_HOPLIMIT		EA_CODE(EAP_KRT, 0x2a)
-#define EA_KRT_INITCWND		EA_CODE(EAP_KRT, 0x2b)
-#define EA_KRT_FEATURES		EA_CODE(EAP_KRT, 0x2c)
-#define EA_KRT_RTO_MIN		EA_CODE(EAP_KRT, 0x2d)
-#define EA_KRT_INITRWND		EA_CODE(EAP_KRT, 0x2e)
-#define EA_KRT_QUICKACK		EA_CODE(EAP_KRT, 0x2f)
+#define EA_KRT_METRICS		EA_CODE(PROTOCOL_KERNEL, 0x20)	/* Dummy one */
+#define EA_KRT_LOCK		EA_CODE(PROTOCOL_KERNEL, 0x21)
+#define EA_KRT_MTU		EA_CODE(PROTOCOL_KERNEL, 0x22)
+#define EA_KRT_WINDOW		EA_CODE(PROTOCOL_KERNEL, 0x23)
+#define EA_KRT_RTT		EA_CODE(PROTOCOL_KERNEL, 0x24)
+#define EA_KRT_RTTVAR		EA_CODE(PROTOCOL_KERNEL, 0x25)
+#define EA_KRT_SSTRESH		EA_CODE(PROTOCOL_KERNEL, 0x26)
+#define EA_KRT_CWND		EA_CODE(PROTOCOL_KERNEL, 0x27)
+#define EA_KRT_ADVMSS		EA_CODE(PROTOCOL_KERNEL, 0x28)
+#define EA_KRT_REORDERING 	EA_CODE(PROTOCOL_KERNEL, 0x29)
+#define EA_KRT_HOPLIMIT		EA_CODE(PROTOCOL_KERNEL, 0x2a)
+#define EA_KRT_INITCWND		EA_CODE(PROTOCOL_KERNEL, 0x2b)
+#define EA_KRT_FEATURES		EA_CODE(PROTOCOL_KERNEL, 0x2c)
+#define EA_KRT_RTO_MIN		EA_CODE(PROTOCOL_KERNEL, 0x2d)
+#define EA_KRT_INITRWND		EA_CODE(PROTOCOL_KERNEL, 0x2e)
+#define EA_KRT_QUICKACK		EA_CODE(PROTOCOL_KERNEL, 0x2f)
 
 /* Bits of EA_KRT_LOCK, also based on RTAX_* constants */
 #define EA_KRT_LOCK_MTU		EA_KRT_LOCK | EA_BIT(0x2)
diff --git a/sysdep/linux/netlink.c b/sysdep/linux/netlink.c
index 84591eb2..f5db1575 100644
--- a/sysdep/linux/netlink.c
+++ b/sysdep/linux/netlink.c
@@ -1751,7 +1751,7 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h)
       for (t = 1; t < KRT_METRICS_MAX; t++)
 	if (metrics[0] & (1 << t))
 	  {
-	    ea->attrs[n].id = EA_CODE(EAP_KRT, KRT_METRICS_OFFSET + t);
+	    ea->attrs[n].id = EA_CODE(PROTOCOL_KERNEL, KRT_METRICS_OFFSET + t);
 	    ea->attrs[n].flags = 0;
 	    ea->attrs[n].type = EAF_TYPE_INT; /* FIXME: Some are EAF_TYPE_BITFIELD */
 	    ea->attrs[n].u.data = metrics[t];
diff --git a/sysdep/unix/krt.c b/sysdep/unix/krt.c
index b4fb1967..ae51927e 100644
--- a/sysdep/unix/krt.c
+++ b/sysdep/unix/krt.c
@@ -232,6 +232,7 @@ kif_copy_config(struct proto_config *dest, struct proto_config *src)
 struct protocol proto_unix_iface = {
   .name = 		"Device",
   .template = 		"device%d",
+  .class =		PROTOCOL_DEVICE,
   .proto_size =		sizeof(struct kif_proto),
   .config_size =	sizeof(struct kif_config),
   .preconfig =		kif_preconfig,
@@ -1235,7 +1236,7 @@ krt_get_attr(eattr *a, byte *buf, int buflen)
 struct protocol proto_unix_kernel = {
   .name =		"Kernel",
   .template =		"kernel%d",
-  .attr_class =		EAP_KRT,
+  .class =		PROTOCOL_KERNEL,
   .preference =		DEF_PREF_INHERITED,
   .channel_mask =	NB_IP | MAYBE_IP6_SADR | MAYBE_MPLS,
   .proto_size =		sizeof(struct krt_proto),
diff --git a/sysdep/unix/krt.h b/sysdep/unix/krt.h
index b627882d..ff27bcf5 100644
--- a/sysdep/unix/krt.h
+++ b/sysdep/unix/krt.h
@@ -30,8 +30,8 @@ struct kif_proto;
 
 #define KRT_DEFAULT_ECMP_LIMIT	16
 
-#define EA_KRT_SOURCE	EA_CODE(EAP_KRT, 0)
-#define EA_KRT_METRIC	EA_CODE(EAP_KRT, 1)
+#define EA_KRT_SOURCE	EA_CODE(PROTOCOL_KERNEL, 0)
+#define EA_KRT_METRIC	EA_CODE(PROTOCOL_KERNEL, 1)
 
 /* Whenever we recognize our own routes, we allow learing of foreign routes */
 
diff --git a/test/bt-utils.c b/test/bt-utils.c
index 571ef2fa..7653abf6 100644
--- a/test/bt-utils.c
+++ b/test/bt-utils.c
@@ -75,8 +75,8 @@ bt_bird_init(void)
 
 void bt_bird_cleanup(void)
 {
-  for (int i = 0; i < EAP_MAX; i++)
-    attr_class_to_protocol[i] = NULL;
+  for (int i = 0; i < PROTOCOL__MAX; i++)
+    class_to_protocol[i] = NULL;
 
   config = new_config = NULL;
 }

From 18b4f2082c30586890596988086fbc3e15336526 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 29 May 2018 14:23:14 +0200
Subject: [PATCH 171/198] OSPF: Fix invalid NSSA RFC references

---
 proto/ospf/rt.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/proto/ospf/rt.c b/proto/ospf/rt.c
index c0fe218a..b4f9b067 100644
--- a/proto/ospf/rt.c
+++ b/proto/ospf/rt.c
@@ -108,7 +108,7 @@ orta_pref(const orta *nf)
 static int
 orta_prio(const orta *nf)
 {
-  /* RFC 3103 2.5 (6e) priorities */
+  /* RFC 3101 2.5 (6e) priorities */
   u32 opts = nf->options & (ORTA_NSSA | ORTA_PROP);
 
   /* A Type-7 LSA with the P-bit set */
@@ -217,7 +217,7 @@ orta_compare_asbr(const struct ospf_proto *p, const orta *new, const orta *old)
 
 /*
  * Compare a routing table entry with a new one, for AS external routes
- * (RFC 2328 16.4) and NSSA routes (RFC 3103 2.5), Returns integer <, = or >
+ * (RFC 2328 16.4) and NSSA routes (RFC 3101 2.5), Returns integer <, = or >
  * than 0 if the new orta is less, equal or more preferred than the old orta.
  */
 static int
@@ -1078,7 +1078,7 @@ decide_nssa_lsa(struct ospf_proto *p, ort *nf, struct ospf_lsa_ext_local *rt)
   return 1;
 }
 
-/* RFC 3103 3.2 - translating Type-7 LSAs into Type-5 LSAs */
+/* RFC 3101 3.2 - translating Type-7 LSAs into Type-5 LSAs */
 static inline void
 check_nssa_lsa(struct ospf_proto *p, ort *nf)
 {
@@ -1101,12 +1101,12 @@ check_nssa_lsa(struct ospf_proto *p, ort *nf)
     }
   }
 
-  /* RFC 3103 3.2 (3) - originate the aggregated address range */
+  /* RFC 3101 3.2 (3) - originate the aggregated address range */
   if (anet && anet->active && !anet->hidden && oa->translate)
     ospf_originate_ext_lsa(p, NULL, nf, LSA_M_RTCALC, anet->metric,
 			   (anet->metric & LSA_EXT3_EBIT), IPA_NONE, anet->tag, 0);
 
-  /* RFC 3103 3.2 (2) - originate the same network */
+  /* RFC 3101 3.2 (2) - originate the same network */
   else if (decide_nssa_lsa(p, nf, &rt))
     ospf_originate_ext_lsa(p, NULL, nf, LSA_M_RTCALC, rt.metric, rt.ebit, rt.fwaddr, rt.tag, 0);
 }
@@ -1284,7 +1284,7 @@ ospf_rt_abr2(struct ospf_proto *p)
   struct ospf_area *oa;
   struct top_hash_entry *en;
 
-  /* RFC 3103 3.1 - type-7 translator election */
+  /* RFC 3101 3.1 - type-7 translator election */
   struct ospf_area *bb = p->backbone;
   WALK_LIST(oa, p->area_list)
     if (oa_is_nssa(oa))

From 13c0be19d3d2acc9c1636bbab9222aabdf27d7ac Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Tue, 29 May 2018 12:08:12 +0200
Subject: [PATCH 172/198] Nest: Removing separate tmpa from route propagation

This is a fundamental change of an original (1999) concept of route
processing inside BIRD. During import/export, there was a temporary
ea_list created which was to be used instead of the another one inside
the route itself.

This led to some confusion, quirks, and strange filter code that handled
extended route attributes. Dropping it now.

The protocol interface has changed in an uniform way -- the
`struct ea_list *attrs` argument has been removed from store_tmp_attrs(),
import_control(), rt_notify() and get_route_info().
---
 filter/filter.c        | 64 ++++++++++++----------------
 filter/filter.h        |  3 +-
 nest/protocol.h        | 24 ++++++-----
 nest/route.h           | 13 +++++-
 nest/rt-attr.c         | 60 +++++++++++++++-----------
 nest/rt-show.c         | 34 ++++++---------
 nest/rt-table.c        | 97 +++++++++++++++++-------------------------
 proto/babel/babel.c    | 17 ++++----
 proto/babel/config.Y   |  2 +-
 proto/bgp/attrs.c      | 12 +++---
 proto/bgp/bgp.h        |  6 +--
 proto/ospf/config.Y    |  8 ++--
 proto/ospf/ospf.c      | 18 ++++----
 proto/ospf/topology.c  |  3 +-
 proto/ospf/topology.h  |  2 +-
 proto/pipe/pipe.c      |  5 +--
 proto/radv/radv.c      |  8 ++--
 proto/rip/config.Y     |  4 +-
 proto/rip/rip.c        | 15 +++----
 sysdep/bsd/krt-sock.c  |  3 +-
 sysdep/linux/netlink.c | 21 ++++-----
 sysdep/unix/krt.Y      |  4 +-
 sysdep/unix/krt.c      | 37 +++++++---------
 sysdep/unix/krt.h      |  2 +-
 24 files changed, 220 insertions(+), 242 deletions(-)

diff --git a/filter/filter.c b/filter/filter.c
index 6381550e..6290e74a 100644
--- a/filter/filter.c
+++ b/filter/filter.c
@@ -538,14 +538,23 @@ val_format(struct f_val v, buffer *buf)
 
 static struct rte **f_rte;
 static struct rta *f_old_rta;
-static struct ea_list **f_tmp_attrs;
+static struct ea_list **f_eattrs;
 static struct linpool *f_pool;
 static struct buffer f_buf;
 static int f_flags;
 
+static inline void f_cache_eattrs(void)
+{
+  f_eattrs = &((*f_rte)->attrs->eattrs);
+}
+
 static inline void f_rte_cow(void)
 {
-  *f_rte = rte_cow(*f_rte);
+  if (!((*f_rte)->flags & REF_COW))
+    return;
+
+  *f_rte = rte_do_cow(*f_rte);
+  f_eattrs = NULL;
 }
 
 /*
@@ -570,6 +579,9 @@ f_rta_cow(void)
    * suppose hostentry is not changed by filters).
    */
   (*f_rte)->attrs = rta_do_cow((*f_rte)->attrs, f_pool);
+
+  /* Re-cache the ea_list */
+  f_cache_eattrs();
 }
 
 static char *
@@ -603,7 +615,10 @@ static struct tbf rl_runtime_err = TBF_DEFAULT_LOG_LIMITS;
       return val;
 
 #define ACCESS_RTE \
-  do { if (!f_rte) runtime("No route to access"); } while (0)
+  do { if (!f_rte) runtime("No route to access"); else f_cache_eattrs(); } while (0)
+
+#define ACCESS_EATTRS \
+  do { if (!f_eattrs) f_cache_eattrs(); } while (0)
 
 #define BITFIELD_MASK(what) \
   (1u << (what->a2.i >> 24))
@@ -995,17 +1010,11 @@ interpret(struct f_inst *what)
     break;
   case FI_EA_GET:	/* Access to extended attributes */
     ACCESS_RTE;
+    ACCESS_EATTRS;
     {
-      eattr *e = NULL;
       u16 code = what->a2.i;
       int f_type = what->aux >> 8;
-
-      if (!(f_flags & FF_FORCE_TMPATTR))
-	e = ea_find((*f_rte)->attrs->eattrs, code);
-      if (!e)
-	e = ea_find((*f_tmp_attrs), code);
-      if ((!e) && (f_flags & FF_FORCE_TMPATTR))
-	e = ea_find((*f_rte)->attrs->eattrs, code);
+      eattr *e = ea_find(*f_eattrs, code);
 
       if (!e) {
 	/* A special case: undefined as_path looks like empty as_path */
@@ -1089,6 +1098,7 @@ interpret(struct f_inst *what)
     break;
   case FI_EA_SET:
     ACCESS_RTE;
+    ACCESS_EATTRS;
     ARG_ANY(1);
     {
       struct ea_list *l = lp_alloc(f_pool, sizeof(struct ea_list) + sizeof(eattr));
@@ -1143,13 +1153,7 @@ interpret(struct f_inst *what)
 	  runtime( "Setting bit in bitfield attribute to non-bool value" );
 	{
 	  /* First, we have to find the old value */
-	  eattr *e = NULL;
-	  if (!(f_flags & FF_FORCE_TMPATTR))
-	    e = ea_find((*f_rte)->attrs->eattrs, code);
-	  if (!e)
-	    e = ea_find((*f_tmp_attrs), code);
-	  if ((!e) && (f_flags & FF_FORCE_TMPATTR))
-	    e = ea_find((*f_rte)->attrs->eattrs, code);
+	  eattr *e = ea_find(*f_eattrs, code);
 	  u32 data = e ? e->u.data : 0;
 
 	  if (v1.val.i)
@@ -1181,14 +1185,9 @@ interpret(struct f_inst *what)
       default: bug("Unknown type in e,S");
       }
 
-      if (!(what->aux & EAF_TEMP) && (!(f_flags & FF_FORCE_TMPATTR))) {
-	f_rta_cow();
-	l->next = (*f_rte)->attrs->eattrs;
-	(*f_rte)->attrs->eattrs = l;
-      } else {
-	l->next = (*f_tmp_attrs);
-	(*f_tmp_attrs) = l;
-      }
+      f_rta_cow();
+      l->next = *f_eattrs;
+      *f_eattrs = l;
     }
     break;
   case FI_PREF_GET:
@@ -1518,11 +1517,12 @@ interpret(struct f_inst *what)
     else
     {
       ACCESS_RTE;
+      ACCESS_EATTRS;
       v1.val.net = (*f_rte)->net->n.addr;
 
       /* We ignore temporary attributes, probably not a problem here */
       /* 0x02 is a value of BA_AS_PATH, we don't want to include BGP headers */
-      eattr *e = ea_find((*f_rte)->attrs->eattrs, EA_CODE(PROTOCOL_BGP, 0x02));
+      eattr *e = ea_find(*f_eattrs, EA_CODE(PROTOCOL_BGP, 0x02));
 
       if (!e || e->type != EAF_TYPE_AS_PATH)
 	runtime("Missing AS_PATH attribute");
@@ -1720,7 +1720,6 @@ i_same(struct f_inst *f1, struct f_inst *f2)
  * f_run - run a filter for a route
  * @filter: filter to run
  * @rte: route being filtered, may be modified
- * @tmp_attrs: temporary attributes, prepared by caller or generated by f_run()
  * @tmp_pool: all filter allocations go from this pool
  * @flags: flags
  *
@@ -1742,7 +1741,7 @@ i_same(struct f_inst *f1, struct f_inst *f2)
  * modified in place, old cached rta is possibly freed.
  */
 int
-f_run(struct filter *filter, struct rte **rte, struct ea_list **tmp_attrs, struct linpool *tmp_pool, int flags)
+f_run(struct filter *filter, struct rte **rte, struct linpool *tmp_pool, int flags)
 {
   if (filter == FILTER_ACCEPT)
     return F_ACCEPT;
@@ -1755,7 +1754,6 @@ f_run(struct filter *filter, struct rte **rte, struct ea_list **tmp_attrs, struc
 
   f_rte = rte;
   f_old_rta = NULL;
-  f_tmp_attrs = tmp_attrs;
   f_pool = tmp_pool;
   f_flags = flags;
 
@@ -1797,11 +1795,9 @@ f_run(struct filter *filter, struct rte **rte, struct ea_list **tmp_attrs, struc
 struct f_val
 f_eval_rte(struct f_inst *expr, struct rte **rte, struct linpool *tmp_pool)
 {
-  struct ea_list *tmp_attrs = NULL;
 
   f_rte = rte;
   f_old_rta = NULL;
-  f_tmp_attrs = &tmp_attrs;
   f_pool = tmp_pool;
   f_flags = 0;
 
@@ -1810,9 +1806,6 @@ f_eval_rte(struct f_inst *expr, struct rte **rte, struct linpool *tmp_pool)
   /* Note that in this function we assume that rte->attrs is private / uncached */
   struct f_val res = interpret(expr);
 
-  /* Hack to include EAF_TEMP attributes to the main list */
-  (*rte)->attrs->eattrs = ea_append(tmp_attrs, (*rte)->attrs->eattrs);
-
   return res;
 }
 
@@ -1820,7 +1813,6 @@ struct f_val
 f_eval(struct f_inst *expr, struct linpool *tmp_pool)
 {
   f_flags = 0;
-  f_tmp_attrs = NULL;
   f_rte = NULL;
   f_pool = tmp_pool;
 
diff --git a/filter/filter.h b/filter/filter.h
index 909d09b1..febfdc65 100644
--- a/filter/filter.h
+++ b/filter/filter.h
@@ -175,7 +175,7 @@ void trie_format(struct f_trie *t, buffer *buf);
 struct ea_list;
 struct rte;
 
-int f_run(struct filter *filter, struct rte **rte, struct ea_list **tmp_attrs, struct linpool *tmp_pool, int flags);
+int f_run(struct filter *filter, struct rte **rte, struct linpool *tmp_pool, int flags);
 struct f_val f_eval_rte(struct f_inst *expr, struct rte **rte, struct linpool *tmp_pool);
 struct f_val f_eval(struct f_inst *expr, struct linpool *tmp_pool);
 uint f_eval_int(struct f_inst *expr);
@@ -285,7 +285,6 @@ struct f_trie
 
 #define NEW_F_VAL struct f_val * val; val = cfg_alloc(sizeof(struct f_val));
 
-#define FF_FORCE_TMPATTR 1		/* Force all attributes to be temporary */
 #define FF_SILENT 2			/* Silent filter execution */
 
 /* Bird Tests */
diff --git a/nest/protocol.h b/nest/protocol.h
index d790e90e..dd942c10 100644
--- a/nest/protocol.h
+++ b/nest/protocol.h
@@ -77,7 +77,7 @@ struct protocol {
   int (*shutdown)(struct proto *);		/* Stop the instance */
   void (*cleanup)(struct proto *);		/* Called after shutdown when protocol became hungry/down */
   void (*get_status)(struct proto *, byte *buf); /* Get instance status (for `show protocols' command) */
-  void (*get_route_info)(struct rte *, byte *buf, struct ea_list *attrs); /* Get route information (for `show route' command) */
+  void (*get_route_info)(struct rte *, byte *buf); /* Get route information (for `show route' command) */
   int (*get_attr)(struct eattr *, byte *buf, int buflen);	/* ASCIIfy dynamic attribute (returns GA_*) */
   void (*show_proto_info)(struct proto *);	/* Show protocol info (for `show protocols all' command) */
   void (*copy_config)(struct proto_config *, struct proto_config *);	/* Copy config from given protocol instance */
@@ -191,7 +191,7 @@ struct proto {
    *	   rt_notify	Notify protocol about routing table updates.
    *	   neigh_notify	Notify protocol about neighbor cache events.
    *	   make_tmp_attrs  Construct ea_list from private attrs stored in rte.
-   *	   store_tmp_attrs Store private attrs back to the rte.
+   *	   store_tmp_attrs Store private attrs back to rta. The route MUST NOT be cached.
    *	   import_control  Called as the first step of the route importing process.
    *			It can construct a new rte, add private attributes and
    *			decide whether the route shall be imported: 1=yes, -1=no,
@@ -205,11 +205,11 @@ struct proto {
 
   void (*if_notify)(struct proto *, unsigned flags, struct iface *i);
   void (*ifa_notify)(struct proto *, unsigned flags, struct ifa *a);
-  void (*rt_notify)(struct proto *, struct channel *, struct network *net, struct rte *new, struct rte *old, struct ea_list *attrs);
+  void (*rt_notify)(struct proto *, struct channel *, struct network *net, struct rte *new, struct rte *old);
   void (*neigh_notify)(struct neighbor *neigh);
   struct ea_list *(*make_tmp_attrs)(struct rte *rt, struct linpool *pool);
-  void (*store_tmp_attrs)(struct rte *rt, struct ea_list *attrs);
-  int (*import_control)(struct proto *, struct rte **rt, struct ea_list **attrs, struct linpool *pool);
+  void (*store_tmp_attrs)(struct rte *rt);
+  int (*import_control)(struct proto *, struct rte **rt, struct linpool *pool);
   void (*reload_routes)(struct channel *);
   void (*feed_begin)(struct channel *, int initial);
   void (*feed_end)(struct channel *);
@@ -292,12 +292,16 @@ proto_get_router_id(struct proto_config *pc)
   return pc->router_id ? pc->router_id : pc->global->router_id;
 }
 
-static inline struct ea_list *
-rte_make_tmp_attrs(struct rte *rt, struct linpool *pool)
+static inline void
+rte_make_tmp_attrs(struct rte **rt, struct linpool *pool)
 {
   struct ea_list *(*mta)(struct rte *rt, struct linpool *pool);
-  mta = rt->attrs->src->proto->make_tmp_attrs;
-  return mta ? mta(rt, pool) : NULL;
+  mta = (*rt)->attrs->src->proto->make_tmp_attrs;
+  if (!mta) return;
+  *rt = rte_cow_rta(*rt, pool);
+  struct ea_list *ea = mta(*rt, pool);
+  ea->next = (*rt)->attrs->eattrs;
+  (*rt)->attrs->eattrs = ea;
 }
 
 /* Moved from route.h to avoid dependency conflicts */
@@ -466,7 +470,7 @@ struct channel_class {
   void (*dump_attrs)(struct rte *);		/* Dump protocol-dependent attributes */
 
   void (*get_status)(struct proto *, byte *buf); /* Get instance status (for `show protocols' command) */
-  void (*get_route_info)(struct rte *, byte *buf, struct ea_list *attrs); /* Get route information (for `show route' command) */
+  void (*get_route_info)(struct rte *, byte *buf); /* Get route information (for `show route' command) */
   int (*get_attr)(struct eattr *, byte *buf, int buflen);	/* ASCIIfy dynamic attribute (returns GA_*) */
   void (*show_proto_info)(struct proto *);	/* Show protocol info (for `show protocols all' command) */
 
diff --git a/nest/route.h b/nest/route.h
index 1391b357..cad15440 100644
--- a/nest/route.h
+++ b/nest/route.h
@@ -294,7 +294,7 @@ rte *rte_get_temp(struct rta *);
 void rte_update2(struct channel *c, const net_addr *n, rte *new, struct rte_src *src);
 /* rte_update() moved to protocol.h to avoid dependency conflicts */
 int rt_examine(rtable *t, net_addr *a, struct proto *p, struct filter *filter);
-rte *rt_export_merged(struct channel *c, net *net, rte **rt_free, struct ea_list **tmpa, linpool *pool, int silent);
+rte *rt_export_merged(struct channel *c, net *net, rte **rt_free, linpool *pool, int silent);
 void rt_refresh_begin(rtable *t, struct channel *c);
 void rt_refresh_end(rtable *t, struct channel *c);
 void rt_schedule_prune(rtable *t);
@@ -546,6 +546,15 @@ uint ea_hash(ea_list *e);	/* Calculate 16-bit hash value */
 ea_list *ea_append(ea_list *to, ea_list *what);
 void ea_format_bitfield(struct eattr *a, byte *buf, int bufsize, const char **names, int min, int max);
 
+#define ea_normalize(ea) do { \
+  if (ea->next) { \
+    ea_list *t = alloca(ea_scan(ea)); \
+    ea_merge(ea, t); \
+    ea = t; \
+  } \
+  ea_sort(ea); \
+} while(0) \
+
 static inline eattr *
 ea_set_attr(ea_list **to, struct linpool *pool, uint id, uint flags, uint type, uintptr_t val)
 {
@@ -611,7 +620,7 @@ rta *rta_do_cow(rta *o, linpool *lp);
 static inline rta * rta_cow(rta *r, linpool *lp) { return rta_is_cached(r) ? rta_do_cow(r, lp) : r; }
 void rta_dump(rta *);
 void rta_dump_all(void);
-void rta_show(struct cli *, rta *, ea_list *);
+void rta_show(struct cli *, rta *);
 
 struct hostentry * rt_get_hostentry(rtable *tab, ip_addr a, ip_addr ll, rtable *dep);
 void rta_apply_hostentry(rta *a, struct hostentry *he, mpls_label_stack *mls);
diff --git a/nest/rt-attr.c b/nest/rt-attr.c
index f92efc2e..73ca4748 100644
--- a/nest/rt-attr.c
+++ b/nest/rt-attr.c
@@ -550,29 +550,47 @@ ea_do_sort(ea_list *e)
   while (ss);
 }
 
+/**
+ * In place discard duplicates, undefs and temporary attributes in sorted
+ * ea_list. We use stable sort for this reason.
+ **/
 static inline void
 ea_do_prune(ea_list *e)
 {
   eattr *s, *d, *l, *s0;
   int i = 0;
 
-  /* Discard duplicates and undefs. Do you remember sorting was stable? */
-  s = d = e->attrs;
-  l = e->attrs + e->count;
+  s = d = e->attrs;	    /* Beginning of the list. @s is source, @d is destination. */
+  l = e->attrs + e->count;  /* End of the list */
+
+  /* Walk from begin to end. */
   while (s < l)
     {
       s0 = s++;
+      /* Find a consecutive block of the same attribute */
       while (s < l && s->id == s[-1].id)
 	s++;
-      /* s0 is the most recent version, s[-1] the oldest one */
-      if ((s0->type & EAF_TYPE_MASK) != EAF_TYPE_UNDEF)
-	{
-	  *d = *s0;
-	  d->type = (d->type & ~(EAF_ORIGINATED|EAF_FRESH)) | (s[-1].type & EAF_ORIGINATED);
-	  d++;
-	  i++;
-	}
+
+      /* Now s0 is the most recent version, s[-1] the oldest one */
+      /* Drop undefs */
+      if ((s0->type & EAF_TYPE_MASK) == EAF_TYPE_UNDEF)
+	continue;
+
+      /* Drop temporary attributes */
+      if (s0->type & EAF_TEMP)
+	continue;
+
+      /* Copy the newest version to destination */
+      *d = *s0;
+
+      /* Preserve info whether it originated locally */
+      d->type = (d->type & ~(EAF_ORIGINATED|EAF_FRESH)) | (s[-1].type & EAF_ORIGINATED);
+
+      /* Next destination */
+      d++;
+      i++;
     }
+
   e->count = i;
 }
 
@@ -1128,15 +1146,7 @@ rta_lookup(rta *o)
 
   ASSERT(!(o->aflags & RTAF_CACHED));
   if (o->eattrs)
-    {
-      if (o->eattrs->next)	/* Multiple ea_list's, need to merge them */
-	{
-	  ea_list *ml = alloca(ea_scan(o->eattrs));
-	  ea_merge(o->eattrs, ml);
-	  o->eattrs = ml;
-	}
-      ea_sort(o->eattrs);
-    }
+    ea_normalize(o->eattrs);
 
   h = rta_hash(o);
   for(r=rta_hash_table[h & rta_cache_mask]; r; r=r->next)
@@ -1250,17 +1260,15 @@ rta_dump_all(void)
 }
 
 void
-rta_show(struct cli *c, rta *a, ea_list *eal)
+rta_show(struct cli *c, rta *a)
 {
   static char *src_names[] = { "dummy", "static", "inherit", "device", "static-device", "redirect",
 			       "RIP", "OSPF", "OSPF-IA", "OSPF-E1", "OSPF-E2", "BGP", "pipe" };
-  int i;
 
   cli_printf(c, -1008, "\tType: %s %s", src_names[a->source], ip_scope_text(a->scope));
-  if (!eal)
-    eal = a->eattrs;
-  for(; eal; eal=eal->next)
-    for(i=0; i<eal->count; i++)
+
+  for(ea_list *eal = a->eattrs; eal; eal=eal->next)
+    for(int i=0; i<eal->count; i++)
       ea_show(c, &eal->attrs[i]);
 }
 
diff --git a/nest/rt-show.c b/nest/rt-show.c
index 1f1b73d2..90165c57 100644
--- a/nest/rt-show.c
+++ b/nest/rt-show.c
@@ -29,14 +29,14 @@ rt_show_table(struct cli *c, struct rt_show_data *d)
 }
 
 static void
-rt_show_rte(struct cli *c, byte *ia, rte *e, struct rt_show_data *d, ea_list *tmpa)
+rt_show_rte(struct cli *c, byte *ia, rte *e, struct rt_show_data *d)
 {
   byte from[IPA_MAX_TEXT_LENGTH+8];
   byte tm[TM_DATETIME_BUFFER_SIZE], info[256];
   rta *a = e->attrs;
   int primary = (e->net->routes == e);
   int sync_error = (e->net->n.flags & KRF_SYNC_ERROR);
-  void (*get_route_info)(struct rte *, byte *buf, struct ea_list *attrs);
+  void (*get_route_info)(struct rte *, byte *buf);
   struct nexthop *nh;
 
   tm_format_time(tm, &config->tf_route, e->lastmod);
@@ -46,17 +46,11 @@ rt_show_rte(struct cli *c, byte *ia, rte *e, struct rt_show_data *d, ea_list *tm
     from[0] = 0;
 
   get_route_info = a->src->proto->proto->get_route_info;
-  if (get_route_info || d->verbose)
-    {
-      /* Need to normalize the extended attributes */
-      ea_list *t = tmpa;
-      t = ea_append(t, a->eattrs);
-      tmpa = alloca(ea_scan(t));
-      ea_merge(t, tmpa);
-      ea_sort(tmpa);
-    }
+  /* Need to normalize the extended attributes */
+  if ((get_route_info || d->verbose) && !rta_is_cached(a))
+    ea_normalize(a->eattrs);
   if (get_route_info)
-    get_route_info(e, info, tmpa);
+    get_route_info(e, info);
   else
     bsprintf(info, " (%d)", e->pref);
 
@@ -93,7 +87,7 @@ rt_show_rte(struct cli *c, byte *ia, rte *e, struct rt_show_data *d, ea_list *tm
     }
 
   if (d->verbose)
-    rta_show(c, a, tmpa);
+    rta_show(c, a);
 }
 
 static void
@@ -101,7 +95,6 @@ rt_show_net(struct cli *c, net *n, struct rt_show_data *d)
 {
   rte *e, *ee;
   byte ia[NET_MAX_TEXT_LENGTH+1];
-  struct ea_list *tmpa;
   struct channel *ec = d->tab->export_channel;
   int first = 1;
   int pass = 0;
@@ -121,7 +114,7 @@ rt_show_net(struct cli *c, net *n, struct rt_show_data *d)
 	continue;
 
       ee = e;
-      tmpa = rte_make_tmp_attrs(e, c->show_pool);
+      rte_make_tmp_attrs(&e, c->show_pool);
 
       /* Export channel is down, do not try to export routes to it */
       if (ec && (ec->export_state == ES_DOWN))
@@ -131,7 +124,7 @@ rt_show_net(struct cli *c, net *n, struct rt_show_data *d)
       if ((d->export_mode == RSEM_EXPORT) && (ec->ra_mode == RA_MERGED))
 	{
 	  rte *rt_free;
-	  e = rt_export_merged(ec, n, &rt_free, &tmpa, c->show_pool, 1);
+	  e = rt_export_merged(ec, n, &rt_free, c->show_pool, 1);
 	  pass = 1;
 
 	  if (!e)
@@ -140,7 +133,7 @@ rt_show_net(struct cli *c, net *n, struct rt_show_data *d)
       else if (d->export_mode)
 	{
 	  struct proto *ep = ec->proto;
-	  int ic = ep->import_control ? ep->import_control(ep, &e, &tmpa, c->show_pool) : 0;
+	  int ic = ep->import_control ? ep->import_control(ep, &e, c->show_pool) : 0;
 
 	  if (ec->ra_mode == RA_OPTIMAL || ec->ra_mode == RA_MERGED)
 	    pass = 1;
@@ -156,8 +149,7 @@ rt_show_net(struct cli *c, net *n, struct rt_show_data *d)
 	       * command may change the export filter and do not update routes.
 	       */
 	      int do_export = (ic > 0) ||
-		(f_run(ec->out_filter, &e, &tmpa, c->show_pool,
-		       FF_FORCE_TMPATTR | FF_SILENT) <= F_ACCEPT);
+		(f_run(ec->out_filter, &e, c->show_pool, FF_SILENT) <= F_ACCEPT);
 
 	      if (do_export != (d->export_mode == RSEM_EXPORT))
 		goto skip;
@@ -170,11 +162,11 @@ rt_show_net(struct cli *c, net *n, struct rt_show_data *d)
       if (d->show_protocol && (d->show_protocol != e->attrs->src->proto))
 	goto skip;
 
-      if (f_run(d->filter, &e, &tmpa, c->show_pool, FF_FORCE_TMPATTR) > F_ACCEPT)
+      if (f_run(d->filter, &e, c->show_pool, 0) > F_ACCEPT)
 	goto skip;
 
       if (d->stats < 2)
-	rt_show_rte(c, ia, e, d, tmpa);
+	rt_show_rte(c, ia, e, d);
 
       d->show_counter++;
       ia[0] = 0;
diff --git a/nest/rt-table.c b/nest/rt-table.c
index b885c6e3..de7b05fc 100644
--- a/nest/rt-table.c
+++ b/nest/rt-table.c
@@ -317,11 +317,11 @@ rte_cow_rta(rte *r, linpool *lp)
   if (!rta_is_cached(r->attrs))
     return r;
 
-  rte *e = rte_cow(r);
+  r = rte_cow(r);
   rta *a = rta_do_cow(r->attrs, lp);
-  rta_free(e->attrs);
-  e->attrs = a;
-  return e;
+  rta_free(r->attrs);
+  r->attrs = a;
+  return r;
 }
 
 static int				/* Actually better or at least as good as */
@@ -393,24 +393,20 @@ rte_trace_out(uint flag, struct proto *p, rte *e, char *msg)
 }
 
 static rte *
-export_filter_(struct channel *c, rte *rt0, rte **rt_free, ea_list **tmpa, linpool *pool, int silent)
+export_filter_(struct channel *c, rte *rt0, rte **rt_free, linpool *pool, int silent)
 {
   struct proto *p = c->proto;
   struct filter *filter = c->out_filter;
   struct proto_stats *stats = &c->stats;
-  ea_list *tmpb = NULL;
   rte *rt;
   int v;
 
   rt = rt0;
   *rt_free = NULL;
 
-  if (!tmpa)
-    tmpa = &tmpb;
+  rte_make_tmp_attrs(&rt, pool);
 
-  *tmpa = rte_make_tmp_attrs(rt, pool);
-
-  v = p->import_control ? p->import_control(p, &rt, tmpa, pool) : 0;
+  v = p->import_control ? p->import_control(p, &rt, pool) : 0;
   if (v < 0)
     {
       if (silent)
@@ -429,8 +425,8 @@ export_filter_(struct channel *c, rte *rt0, rte **rt_free, ea_list **tmpa, linpo
     }
 
   v = filter && ((filter == FILTER_REJECT) ||
-		 (f_run(filter, &rt, tmpa, pool,
-			FF_FORCE_TMPATTR | (silent ? FF_SILENT : 0)) > F_ACCEPT));
+		 (f_run(filter, &rt, pool,
+			(silent ? FF_SILENT : 0)) > F_ACCEPT));
   if (v)
     {
       if (silent)
@@ -454,13 +450,13 @@ export_filter_(struct channel *c, rte *rt0, rte **rt_free, ea_list **tmpa, linpo
 }
 
 static inline rte *
-export_filter(struct channel *c, rte *rt0, rte **rt_free, ea_list **tmpa, int silent)
+export_filter(struct channel *c, rte *rt0, rte **rt_free, int silent)
 {
-  return export_filter_(c, rt0, rt_free, tmpa, rte_update_pool, silent);
+  return export_filter_(c, rt0, rt_free, rte_update_pool, silent);
 }
 
 static void
-do_rt_notify(struct channel *c, net *net, rte *new, rte *old, ea_list *tmpa, int refeed)
+do_rt_notify(struct channel *c, net *net, rte *new, rte *old, int refeed)
 {
   struct proto *p = c->proto;
   struct proto_stats *stats = &c->stats;
@@ -533,19 +529,7 @@ do_rt_notify(struct channel *c, net *net, rte *new, rte *old, ea_list *tmpa, int
       else if (old)
 	rte_trace_out(D_ROUTES, p, old, "removed");
     }
-  if (!new)
-    p->rt_notify(p, c, net, NULL, old, NULL);
-  else if (tmpa)
-    {
-      ea_list *t = tmpa;
-      while (t->next)
-	t = t->next;
-      t->next = new->attrs->eattrs;
-      p->rt_notify(p, c, net, new, old, tmpa);
-      t->next = NULL;
-    }
-  else
-    p->rt_notify(p, c, net, new, old, new->attrs->eattrs);
+  p->rt_notify(p, c, net, new, old);
 }
 
 static void
@@ -557,7 +541,6 @@ rt_notify_basic(struct channel *c, net *net, rte *new0, rte *old0, int refeed)
   rte *old = old0;
   rte *new_free = NULL;
   rte *old_free = NULL;
-  ea_list *tmpa = NULL;
 
   if (new)
     c->stats.exp_updates_received++;
@@ -585,10 +568,10 @@ rt_notify_basic(struct channel *c, net *net, rte *new0, rte *old0, int refeed)
    */
 
   if (new)
-    new = export_filter(c, new, &new_free, &tmpa, 0);
+    new = export_filter(c, new, &new_free, 0);
 
   if (old && !refeed)
-    old = export_filter(c, old, &old_free, NULL, 1);
+    old = export_filter(c, old, &old_free, 1);
 
   if (!new && !old)
   {
@@ -605,13 +588,13 @@ rt_notify_basic(struct channel *c, net *net, rte *new0, rte *old0, int refeed)
 
 #ifdef CONFIG_PIPE
     if ((p->proto == &proto_pipe) && !new0 && (p != old0->sender->proto))
-      p->rt_notify(p, c, net, NULL, old0, NULL);
+      p->rt_notify(p, c, net, NULL, old0);
 #endif
 
     return;
   }
 
-  do_rt_notify(c, net, new, old, tmpa, refeed);
+  do_rt_notify(c, net, new, old, refeed);
 
   /* Discard temporary rte's */
   if (new_free)
@@ -630,7 +613,6 @@ rt_notify_accepted(struct channel *c, net *net, rte *new_changed, rte *old_chang
   rte *old_best = NULL;
   rte *new_free = NULL;
   rte *old_free = NULL;
-  ea_list *tmpa = NULL;
 
   /* Used to track whether we met old_changed position. If before_old is NULL
      old_changed was the first and we met it implicitly before current best route. */
@@ -648,7 +630,7 @@ rt_notify_accepted(struct channel *c, net *net, rte *new_changed, rte *old_chang
   /* First, find the new_best route - first accepted by filters */
   for (r=net->routes; rte_is_valid(r); r=r->next)
     {
-      if (new_best = export_filter(c, r, &new_free, &tmpa, 0))
+      if (new_best = export_filter(c, r, &new_free, 0))
 	break;
 
       /* Note if we walked around the position of old_changed route */
@@ -699,7 +681,7 @@ rt_notify_accepted(struct channel *c, net *net, rte *new_changed, rte *old_chang
 
   /* First case */
   if (old_meet)
-    if (old_best = export_filter(c, old_changed, &old_free, NULL, 1))
+    if (old_best = export_filter(c, old_changed, &old_free, 1))
       goto found;
 
   /* Second case */
@@ -717,18 +699,18 @@ rt_notify_accepted(struct channel *c, net *net, rte *new_changed, rte *old_chang
   /* Fourth case */
   for (r=r->next; rte_is_valid(r); r=r->next)
     {
-      if (old_best = export_filter(c, r, &old_free, NULL, 1))
+      if (old_best = export_filter(c, r, &old_free, 1))
 	goto found;
 
       if (r == before_old)
-	if (old_best = export_filter(c, old_changed, &old_free, NULL, 1))
+	if (old_best = export_filter(c, old_changed, &old_free, 1))
 	  goto found;
     }
 
   /* Implicitly, old_best is NULL and new_best is non-NULL */
 
  found:
-  do_rt_notify(c, net, new_best, old_best, tmpa, (feed == 2));
+  do_rt_notify(c, net, new_best, old_best, (feed == 2));
 
   /* Discard temporary rte's */
   if (new_free)
@@ -745,7 +727,7 @@ nexthop_merge_rta(struct nexthop *nhs, rta *a, linpool *pool, int max)
 }
 
 rte *
-rt_export_merged(struct channel *c, net *net, rte **rt_free, ea_list **tmpa, linpool *pool, int silent)
+rt_export_merged(struct channel *c, net *net, rte **rt_free, linpool *pool, int silent)
 {
   // struct proto *p = c->proto;
   struct nexthop *nhs = NULL;
@@ -757,7 +739,7 @@ rt_export_merged(struct channel *c, net *net, rte **rt_free, ea_list **tmpa, lin
   if (!rte_is_valid(best0))
     return NULL;
 
-  best = export_filter_(c, best0, rt_free, tmpa, pool, silent);
+  best = export_filter_(c, best0, rt_free, pool, silent);
 
   if (!best || !rte_is_reachable(best))
     return best;
@@ -767,7 +749,7 @@ rt_export_merged(struct channel *c, net *net, rte **rt_free, ea_list **tmpa, lin
     if (!rte_mergable(best0, rt0))
       continue;
 
-    rt = export_filter_(c, rt0, &tmp, NULL, pool, 1);
+    rt = export_filter_(c, rt0, &tmp, pool, 1);
 
     if (!rt)
       continue;
@@ -807,7 +789,6 @@ rt_notify_merged(struct channel *c, net *net, rte *new_changed, rte *old_changed
   rte *old_best_free = NULL;
   rte *new_changed_free = NULL;
   rte *old_changed_free = NULL;
-  ea_list *tmpa = NULL;
 
   /* We assume that all rte arguments are either NULL or rte_is_valid() */
 
@@ -819,10 +800,10 @@ rt_notify_merged(struct channel *c, net *net, rte *new_changed, rte *old_changed
   if ((new_best == old_best) && !refeed)
   {
     new_changed = rte_mergable(new_best, new_changed) ?
-      export_filter(c, new_changed, &new_changed_free, NULL, 1) : NULL;
+      export_filter(c, new_changed, &new_changed_free, 1) : NULL;
 
     old_changed = rte_mergable(old_best, old_changed) ?
-      export_filter(c, old_changed, &old_changed_free, NULL, 1) : NULL;
+      export_filter(c, old_changed, &old_changed_free, 1) : NULL;
 
     if (!new_changed && !old_changed)
       return;
@@ -835,15 +816,15 @@ rt_notify_merged(struct channel *c, net *net, rte *new_changed, rte *old_changed
 
   /* Prepare new merged route */
   if (new_best)
-    new_best = rt_export_merged(c, net, &new_best_free, &tmpa, rte_update_pool, 0);
+    new_best = rt_export_merged(c, net, &new_best_free, rte_update_pool, 0);
 
   /* Prepare old merged route (without proper merged next hops) */
   /* There are some issues with running filter on old route - see rt_notify_basic() */
   if (old_best && !refeed)
-    old_best = export_filter(c, old_best, &old_best_free, NULL, 1);
+    old_best = export_filter(c, old_best, &old_best_free, 1);
 
   if (new_best || old_best)
-    do_rt_notify(c, net, new_best, old_best, tmpa, refeed);
+    do_rt_notify(c, net, new_best, old_best, refeed);
 
   /* Discard temporary rte's */
   if (new_best_free)
@@ -1341,7 +1322,6 @@ rte_update2(struct channel *c, const net_addr *n, rte *new, struct rte_src *src)
   struct proto *p = c->proto;
   struct proto_stats *stats = &c->stats;
   struct filter *filter = c->in_filter;
-  ea_list *tmpa = NULL;
   rte *dummy = NULL;
   net *nn;
 
@@ -1379,11 +1359,11 @@ rte_update2(struct channel *c, const net_addr *n, rte *new, struct rte_src *src)
 	}
       else
 	{
-	  tmpa = rte_make_tmp_attrs(new, rte_update_pool);
+	  rte_make_tmp_attrs(&new, rte_update_pool);
 	  if (filter && (filter != FILTER_REJECT))
 	    {
-	      ea_list *old_tmpa = tmpa;
-	      int fr = f_run(filter, &new, &tmpa, rte_update_pool, 0);
+	      ea_list *oldea = new->attrs->eattrs;
+	      int fr = f_run(filter, &new, rte_update_pool, 0);
 	      if (fr > F_ACCEPT)
 		{
 		  stats->imp_updates_filtered++;
@@ -1394,8 +1374,8 @@ rte_update2(struct channel *c, const net_addr *n, rte *new, struct rte_src *src)
 
 		  new->flags |= REF_FILTERED;
 		}
-	      if (tmpa != old_tmpa && src->proto->store_tmp_attrs)
-		src->proto->store_tmp_attrs(new, tmpa);
+	      if (new->attrs->eattrs != oldea && src->proto->store_tmp_attrs)
+		src->proto->store_tmp_attrs(new);
 	    }
 	}
       if (!rta_is_cached(new->attrs)) /* Need to copy attributes */
@@ -1459,11 +1439,10 @@ rt_examine(rtable *t, net_addr *a, struct proto *p, struct filter *filter)
   rte_update_lock();
 
   /* Rest is stripped down export_filter() */
-  ea_list *tmpa = rte_make_tmp_attrs(rt, rte_update_pool);
-  int v = p->import_control ? p->import_control(p, &rt, &tmpa, rte_update_pool) : 0;
+  rte_make_tmp_attrs(&rt, rte_update_pool);
+  int v = p->import_control ? p->import_control(p, &rt, rte_update_pool) : 0;
   if (v == RIC_PROCESS)
-    v = (f_run(filter, &rt, &tmpa, rte_update_pool,
-	       FF_FORCE_TMPATTR | FF_SILENT) <= F_ACCEPT);
+    v = (f_run(filter, &rt, rte_update_pool, FF_SILENT) <= F_ACCEPT);
 
    /* Discard temporary rte */
   if (rt != n->routes)
diff --git a/proto/babel/babel.c b/proto/babel/babel.c
index 83986cb9..d9e3aad2 100644
--- a/proto/babel/babel.c
+++ b/proto/babel/babel.c
@@ -1829,7 +1829,7 @@ babel_dump(struct proto *P)
 }
 
 static void
-babel_get_route_info(rte *rte, byte *buf, ea_list *attrs UNUSED)
+babel_get_route_info(rte *rte, byte *buf)
 {
   buf += bsprintf(buf, " (%d/%d) [%lR]", rte->pref, rte->u.babel.metric, rte->u.babel.router_id);
 }
@@ -2087,12 +2087,13 @@ babel_prepare_attrs(struct linpool *pool, ea_list *next, uint metric, u64 router
 
 
 static int
-babel_import_control(struct proto *P, struct rte **new, struct ea_list **attrs UNUSED, struct linpool *pool UNUSED)
+babel_import_control(struct proto *P, struct rte **new, struct linpool *pool)
 {
-  rte *e = *new;
+  struct babel_proto *p = (void *) P;
+  struct rta *a = (*new)->attrs;
 
   /* Reject our own unreachable routes */
-  if ((e->attrs->dest == RTD_UNREACHABLE) && (e->attrs->src->proto == P))
+  if ((a->dest == RTD_UNREACHABLE) && (a->src->proto == P))
     return -1;
 
   return 0;
@@ -2105,9 +2106,9 @@ babel_make_tmp_attrs(struct rte *rt, struct linpool *pool)
 }
 
 static void
-babel_store_tmp_attrs(struct rte *rt, struct ea_list *attrs)
+babel_store_tmp_attrs(struct rte *rt)
 {
-  rt->u.babel.metric = ea_get_int(attrs, EA_BABEL_METRIC, 0);
+  rt->u.babel.metric = ea_get_int(rt->attrs->eattrs, EA_BABEL_METRIC, 0);
 }
 
 /*
@@ -2116,7 +2117,7 @@ babel_store_tmp_attrs(struct rte *rt, struct ea_list *attrs)
  */
 static void
 babel_rt_notify(struct proto *P, struct channel *c UNUSED, struct network *net,
-		struct rte *new, struct rte *old UNUSED, struct ea_list *attrs UNUSED)
+		struct rte *new, struct rte *old UNUSED)
 {
   struct babel_proto *p = (void *) P;
   struct babel_entry *e;
@@ -2126,7 +2127,7 @@ babel_rt_notify(struct proto *P, struct channel *c UNUSED, struct network *net,
     /* Update */
     uint internal = (new->attrs->src->proto == P);
     uint rt_seqno = internal ? new->u.babel.seqno : p->update_seqno;
-    uint rt_metric = ea_get_int(attrs, EA_BABEL_METRIC, 0);
+    uint rt_metric = ea_get_int(new->attrs->eattrs, EA_BABEL_METRIC, 0);
     u64 rt_router_id = internal ? new->u.babel.router_id : p->router_id;
 
     if (rt_metric > BABEL_INFINITY)
diff --git a/proto/babel/config.Y b/proto/babel/config.Y
index 205b4e4f..2b20c43f 100644
--- a/proto/babel/config.Y
+++ b/proto/babel/config.Y
@@ -125,7 +125,7 @@ babel_iface_opt_list:
 babel_iface:
   babel_iface_start iface_patt_list_nopx babel_iface_opt_list babel_iface_finish;
 
-CF_ADDTO(dynamic_attr, BABEL_METRIC { $$ = f_new_dynamic_attr(EAF_TYPE_INT | EAF_TEMP, T_INT, EA_BABEL_METRIC); })
+CF_ADDTO(dynamic_attr, BABEL_METRIC { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_BABEL_METRIC); })
 
 CF_CLI_HELP(SHOW BABEL, ..., [[Show information about Babel protocol]]);
 
diff --git a/proto/bgp/attrs.c b/proto/bgp/attrs.c
index 5695e1c1..d65f6334 100644
--- a/proto/bgp/attrs.c
+++ b/proto/bgp/attrs.c
@@ -1372,7 +1372,7 @@ bgp_free_prefix(struct bgp_channel *c, struct bgp_prefix *px)
  */
 
 int
-bgp_import_control(struct proto *P, rte **new, ea_list **attrs UNUSED, struct linpool *pool UNUSED)
+bgp_import_control(struct proto *P, rte **new, struct linpool *pool UNUSED)
 {
   rte *e = *new;
   struct proto *SRC = e->attrs->src->proto;
@@ -1536,7 +1536,7 @@ bgp_update_attrs(struct bgp_proto *p, struct bgp_channel *c, rte *e, ea_list *at
 }
 
 void
-bgp_rt_notify(struct proto *P, struct channel *C, net *n, rte *new, rte *old, ea_list *attrs)
+bgp_rt_notify(struct proto *P, struct channel *C, net *n, rte *new, rte *old)
 {
   struct bgp_proto *p = (void *) P;
   struct bgp_channel *c = (void *) C;
@@ -1546,7 +1546,7 @@ bgp_rt_notify(struct proto *P, struct channel *C, net *n, rte *new, rte *old, ea
 
   if (new)
   {
-    attrs = bgp_update_attrs(p, c, new, attrs, bgp_linpool2);
+    struct ea_list *attrs = bgp_update_attrs(p, c, new, new->attrs->eattrs, bgp_linpool2);
 
     /* If attributes are invalid, we fail back to withdraw */
     buck = attrs ? bgp_get_bucket(c, attrs) : bgp_get_withdraw_bucket(c);
@@ -2007,10 +2007,10 @@ bgp_get_attr(eattr *a, byte *buf, int buflen)
 }
 
 void
-bgp_get_route_info(rte *e, byte *buf, ea_list *attrs)
+bgp_get_route_info(rte *e, byte *buf)
 {
-  eattr *p = ea_find(attrs, EA_CODE(PROTOCOL_BGP, BA_AS_PATH));
-  eattr *o = ea_find(attrs, EA_CODE(PROTOCOL_BGP, BA_ORIGIN));
+  eattr *p = ea_find(e->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_AS_PATH));
+  eattr *o = ea_find(e->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_ORIGIN));
   u32 origas;
 
   buf += bsprintf(buf, " (%d", e->pref);
diff --git a/proto/bgp/bgp.h b/proto/bgp/bgp.h
index de05dcfb..1235ee78 100644
--- a/proto/bgp/bgp.h
+++ b/proto/bgp/bgp.h
@@ -505,10 +505,10 @@ void bgp_free_prefix(struct bgp_channel *c, struct bgp_prefix *bp);
 int bgp_rte_better(struct rte *, struct rte *);
 int bgp_rte_mergable(rte *pri, rte *sec);
 int bgp_rte_recalculate(rtable *table, net *net, rte *new, rte *old, rte *old_best);
-void bgp_rt_notify(struct proto *P, struct channel *C, net *n, rte *new, rte *old, ea_list *attrs);
-int bgp_import_control(struct proto *, struct rte **, struct ea_list **, struct linpool *);
+void bgp_rt_notify(struct proto *P, struct channel *C, net *n, rte *new, rte *old);
+int bgp_import_control(struct proto *, struct rte **, struct linpool *);
 int bgp_get_attr(struct eattr *e, byte *buf, int buflen);
-void bgp_get_route_info(struct rte *, byte *buf, struct ea_list *attrs);
+void bgp_get_route_info(struct rte *, byte *buf);
 
 
 /* packets.c */
diff --git a/proto/ospf/config.Y b/proto/ospf/config.Y
index 0b09966d..c439a614 100644
--- a/proto/ospf/config.Y
+++ b/proto/ospf/config.Y
@@ -497,10 +497,10 @@ ospf_iface:
   ospf_iface_start ospf_iface_patt_list ospf_iface_opt_list { ospf_iface_finish(); }
  ;
 
-CF_ADDTO(dynamic_attr, OSPF_METRIC1 { $$ = f_new_dynamic_attr(EAF_TYPE_INT | EAF_TEMP, T_INT, EA_OSPF_METRIC1); })
-CF_ADDTO(dynamic_attr, OSPF_METRIC2 { $$ = f_new_dynamic_attr(EAF_TYPE_INT | EAF_TEMP, T_INT, EA_OSPF_METRIC2); })
-CF_ADDTO(dynamic_attr, OSPF_TAG { $$ = f_new_dynamic_attr(EAF_TYPE_INT | EAF_TEMP, T_INT, EA_OSPF_TAG); })
-CF_ADDTO(dynamic_attr, OSPF_ROUTER_ID { $$ = f_new_dynamic_attr(EAF_TYPE_ROUTER_ID | EAF_TEMP, T_QUAD, EA_OSPF_ROUTER_ID); })
+CF_ADDTO(dynamic_attr, OSPF_METRIC1 { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_OSPF_METRIC1); })
+CF_ADDTO(dynamic_attr, OSPF_METRIC2 { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_OSPF_METRIC2); })
+CF_ADDTO(dynamic_attr, OSPF_TAG { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_OSPF_TAG); })
+CF_ADDTO(dynamic_attr, OSPF_ROUTER_ID { $$ = f_new_dynamic_attr(EAF_TYPE_ROUTER_ID, T_QUAD, EA_OSPF_ROUTER_ID); })
 
 CF_CLI_HELP(SHOW OSPF, ..., [[Show information about OSPF protocol]]);
 CF_CLI(SHOW OSPF, optsym, [<name>], [[Show information about OSPF protocol]])
diff --git a/proto/ospf/ospf.c b/proto/ospf/ospf.c
index d9edc3e7..fa122f01 100644
--- a/proto/ospf/ospf.c
+++ b/proto/ospf/ospf.c
@@ -101,9 +101,9 @@
 #include <stdlib.h>
 #include "ospf.h"
 
-static int ospf_import_control(struct proto *P, rte **new, ea_list **attrs, struct linpool *pool);
+static int ospf_import_control(struct proto *P, rte **new, struct linpool *pool);
 static struct ea_list *ospf_make_tmp_attrs(struct rte *rt, struct linpool *pool);
-static void ospf_store_tmp_attrs(struct rte *rt, struct ea_list *attrs);
+static void ospf_store_tmp_attrs(struct rte *rt);
 static void ospf_reload_routes(struct channel *C);
 static int ospf_rte_better(struct rte *new, struct rte *old);
 static int ospf_rte_same(struct rte *new, struct rte *old);
@@ -446,7 +446,7 @@ ospf_disp(timer * timer)
  * import to the filters.
  */
 static int
-ospf_import_control(struct proto *P, rte **new, ea_list **attrs UNUSED, struct linpool *pool UNUSED)
+ospf_import_control(struct proto *P, rte **new, struct linpool *pool)
 {
   struct ospf_proto *p = (struct ospf_proto *) P;
   struct ospf_area *oa = ospf_main_area(p);
@@ -471,12 +471,12 @@ ospf_make_tmp_attrs(struct rte *rt, struct linpool *pool)
 }
 
 static void
-ospf_store_tmp_attrs(struct rte *rt, struct ea_list *attrs)
+ospf_store_tmp_attrs(struct rte *rt)
 {
-  rt->u.ospf.metric1 = ea_get_int(attrs, EA_OSPF_METRIC1, LSINFINITY);
-  rt->u.ospf.metric2 = ea_get_int(attrs, EA_OSPF_METRIC2, 10000);
-  rt->u.ospf.tag = ea_get_int(attrs, EA_OSPF_TAG, 0);
-  rt->u.ospf.router_id = ea_get_int(attrs, EA_OSPF_ROUTER_ID, 0);
+  rt->u.ospf.metric1 = ea_get_int(rt->attrs->eattrs, EA_OSPF_METRIC1, LSINFINITY);
+  rt->u.ospf.metric2 = ea_get_int(rt->attrs->eattrs, EA_OSPF_METRIC2, 10000);
+  rt->u.ospf.tag = ea_get_int(rt->attrs->eattrs, EA_OSPF_TAG, 0);
+  rt->u.ospf.router_id = ea_get_int(rt->attrs->eattrs, EA_OSPF_ROUTER_ID, 0);
 }
 
 /**
@@ -535,7 +535,7 @@ ospf_get_status(struct proto *P, byte * buf)
 }
 
 static void
-ospf_get_route_info(rte * rte, byte * buf, ea_list * attrs UNUSED)
+ospf_get_route_info(rte * rte, byte * buf)
 {
   char *type = "<bug>";
 
diff --git a/proto/ospf/topology.c b/proto/ospf/topology.c
index e909bbe9..54b255c9 100644
--- a/proto/ospf/topology.c
+++ b/proto/ospf/topology.c
@@ -1243,11 +1243,12 @@ find_surrogate_fwaddr(struct ospf_proto *p, struct ospf_area *oa)
 }
 
 void
-ospf_rt_notify(struct proto *P, struct channel *ch UNUSED, net *n, rte *new, rte *old UNUSED, ea_list *ea)
+ospf_rt_notify(struct proto *P, struct channel *ch UNUSED, net *n, rte *new, rte *old UNUSED)
 {
   struct ospf_proto *p = (struct ospf_proto *) P;
   struct ospf_area *oa = NULL;	/* non-NULL for NSSA-LSA */
   ort *nf;
+  struct ea_list *ea = new->attrs->eattrs;
 
   /*
    * There are several posibilities:
diff --git a/proto/ospf/topology.h b/proto/ospf/topology.h
index ac87334b..54ec9ccf 100644
--- a/proto/ospf/topology.h
+++ b/proto/ospf/topology.h
@@ -188,7 +188,7 @@ void ospf_originate_sum_net_lsa(struct ospf_proto *p, struct ospf_area *oa, ort
 void ospf_originate_sum_rt_lsa(struct ospf_proto *p, struct ospf_area *oa, u32 drid, int metric, u32 options);
 void ospf_originate_ext_lsa(struct ospf_proto *p, struct ospf_area *oa, ort *nf, u8 mode, u32 metric, u32 ebit, ip_addr fwaddr, u32 tag, int pbit);
 
-void ospf_rt_notify(struct proto *P, struct channel *ch, net *n, rte *new, rte *old, ea_list *attrs);
+void ospf_rt_notify(struct proto *P, struct channel *ch, net *n, rte *new, rte *old);
 void ospf_update_topology(struct ospf_proto *p);
 
 struct top_hash_entry *ospf_hash_find(struct top_graph *, u32 domain, u32 lsa, u32 rtr, u32 type);
diff --git a/proto/pipe/pipe.c b/proto/pipe/pipe.c
index ffc677f4..7aada37e 100644
--- a/proto/pipe/pipe.c
+++ b/proto/pipe/pipe.c
@@ -44,7 +44,7 @@
 #include "pipe.h"
 
 static void
-pipe_rt_notify(struct proto *P, struct channel *src_ch, net *n, rte *new, rte *old, ea_list *attrs)
+pipe_rt_notify(struct proto *P, struct channel *src_ch, net *n, rte *new, rte *old)
 {
   struct pipe_proto *p = (void *) P;
   struct channel *dst = (src_ch == p->pri) ? p->sec : p->pri;
@@ -69,7 +69,6 @@ pipe_rt_notify(struct proto *P, struct channel *src_ch, net *n, rte *new, rte *o
       memcpy(a, new->attrs, rta_size(new->attrs));
 
       a->aflags = 0;
-      a->eattrs = attrs;
       a->hostentry = NULL;
       e = rte_get_temp(a);
       e->pflags = 0;
@@ -93,7 +92,7 @@ pipe_rt_notify(struct proto *P, struct channel *src_ch, net *n, rte *new, rte *o
 }
 
 static int
-pipe_import_control(struct proto *P, rte **ee, ea_list **ea UNUSED, struct linpool *p UNUSED)
+pipe_import_control(struct proto *P, rte **ee, struct linpool *p UNUSED)
 {
   struct proto *pp = (*ee)->sender->proto;
 
diff --git a/proto/radv/radv.c b/proto/radv/radv.c
index 2f5f1c27..ce88c7cc 100644
--- a/proto/radv/radv.c
+++ b/proto/radv/radv.c
@@ -395,7 +395,7 @@ radv_net_match_trigger(struct radv_config *cf, net *n)
 }
 
 int
-radv_import_control(struct proto *P, rte **new, ea_list **attrs UNUSED, struct linpool *pool UNUSED)
+radv_import_control(struct proto *P, rte **new, struct linpool *pool UNUSED)
 {
   // struct radv_proto *p = (struct radv_proto *) P;
   struct radv_config *cf = (struct radv_config *) (P->cf);
@@ -410,7 +410,7 @@ radv_import_control(struct proto *P, rte **new, ea_list **attrs UNUSED, struct l
 }
 
 static void
-radv_rt_notify(struct proto *P, struct channel *ch UNUSED, net *n, rte *new, rte *old UNUSED, ea_list *attrs)
+radv_rt_notify(struct proto *P, struct channel *ch UNUSED, net *n, rte *new, rte *old UNUSED)
 {
   struct radv_proto *p = (struct radv_proto *) P;
   struct radv_config *cf = (struct radv_config *) (P->cf);
@@ -448,11 +448,11 @@ radv_rt_notify(struct proto *P, struct channel *ch UNUSED, net *n, rte *new, rte
   {
     /* Update */
 
-    ea = ea_find(attrs, EA_RA_PREFERENCE);
+    ea = ea_find(new->attrs->eattrs, EA_RA_PREFERENCE);
     uint preference = ea ? ea->u.data : RA_PREF_MEDIUM;
     uint preference_set = !!ea;
 
-    ea = ea_find(attrs, EA_RA_LIFETIME);
+    ea = ea_find(new->attrs->eattrs, EA_RA_LIFETIME);
     uint lifetime = ea ? ea->u.data : 0;
     uint lifetime_set = !!ea;
 
diff --git a/proto/rip/config.Y b/proto/rip/config.Y
index aff63f03..c46cf00c 100644
--- a/proto/rip/config.Y
+++ b/proto/rip/config.Y
@@ -186,8 +186,8 @@ rip_iface:
   rip_iface_start iface_patt_list_nopx rip_iface_opt_list rip_iface_finish;
 
 
-CF_ADDTO(dynamic_attr, RIP_METRIC { $$ = f_new_dynamic_attr(EAF_TYPE_INT | EAF_TEMP, T_INT, EA_RIP_METRIC); })
-CF_ADDTO(dynamic_attr, RIP_TAG { $$ = f_new_dynamic_attr(EAF_TYPE_INT | EAF_TEMP, T_INT, EA_RIP_TAG); })
+CF_ADDTO(dynamic_attr, RIP_METRIC { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_RIP_METRIC); })
+CF_ADDTO(dynamic_attr, RIP_TAG { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_RIP_TAG); })
 
 CF_CLI_HELP(SHOW RIP, ..., [[Show information about RIP protocol]]);
 
diff --git a/proto/rip/rip.c b/proto/rip/rip.c
index adc2b471..90bf8e5c 100644
--- a/proto/rip/rip.c
+++ b/proto/rip/rip.c
@@ -298,7 +298,7 @@ rip_withdraw_rte(struct rip_proto *p, net_addr *n, struct rip_neighbor *from)
  */
 static void
 rip_rt_notify(struct proto *P, struct channel *ch UNUSED, struct network *net, struct rte *new,
-	      struct rte *old UNUSED, struct ea_list *attrs)
+	      struct rte *old UNUSED)
 {
   struct rip_proto *p = (struct rip_proto *) P;
   struct rip_entry *en;
@@ -307,8 +307,8 @@ rip_rt_notify(struct proto *P, struct channel *ch UNUSED, struct network *net, s
   if (new)
   {
     /* Update */
-    u32 rt_metric = ea_get_int(attrs, EA_RIP_METRIC, 1);
-    u32 rt_tag = ea_get_int(attrs, EA_RIP_TAG, 0);
+    u32 rt_metric = ea_get_int(new->attrs->eattrs, EA_RIP_METRIC, 1);
+    u32 rt_tag = ea_get_int(new->attrs->eattrs, EA_RIP_TAG, 0);
 
     if (rt_metric > p->infinity)
     {
@@ -1040,10 +1040,10 @@ rip_make_tmp_attrs(struct rte *rt, struct linpool *pool)
 }
 
 static void
-rip_store_tmp_attrs(struct rte *rt, struct ea_list *attrs)
+rip_store_tmp_attrs(struct rte *rt)
 {
-  rt->u.rip.metric = ea_get_int(attrs, EA_RIP_METRIC, 1);
-  rt->u.rip.tag = ea_get_int(attrs, EA_RIP_TAG, 0);
+  rt->u.rip.metric = ea_get_int(rt->attrs->eattrs, EA_RIP_METRIC, 1);
+  rt->u.rip.tag = ea_get_int(rt->attrs->eattrs, EA_RIP_TAG, 0);
 }
 
 static int
@@ -1081,7 +1081,6 @@ rip_init(struct proto_config *CF)
   P->if_notify = rip_if_notify;
   P->rt_notify = rip_rt_notify;
   P->neigh_notify = rip_neigh_notify;
-  // P->import_control = rip_import_control;
   P->reload_routes = rip_reload_routes;
   P->make_tmp_attrs = rip_make_tmp_attrs;
   P->store_tmp_attrs = rip_store_tmp_attrs;
@@ -1145,7 +1144,7 @@ rip_reconfigure(struct proto *P, struct proto_config *CF)
 }
 
 static void
-rip_get_route_info(rte *rte, byte *buf, ea_list *attrs UNUSED)
+rip_get_route_info(rte *rte, byte *buf)
 {
   buf += bsprintf(buf, " (%d/%d)", rte->pref, rte->u.rip.metric);
 
diff --git a/sysdep/bsd/krt-sock.c b/sysdep/bsd/krt-sock.c
index e56dd616..8522e415 100644
--- a/sysdep/bsd/krt-sock.c
+++ b/sysdep/bsd/krt-sock.c
@@ -347,8 +347,7 @@ krt_send_route(struct krt_proto *p, int cmd, rte *e)
 }
 
 void
-krt_replace_rte(struct krt_proto *p, net *n, rte *new, rte *old,
-		struct ea_list *eattrs UNUSED)
+krt_replace_rte(struct krt_proto *p, net *n, rte *new, rte *old)
 {
   int err = 0;
 
diff --git a/sysdep/linux/netlink.c b/sysdep/linux/netlink.c
index f5db1575..73f77147 100644
--- a/sysdep/linux/netlink.c
+++ b/sysdep/linux/netlink.c
@@ -1177,11 +1177,12 @@ nh_bufsize(struct nexthop *nh)
 }
 
 static int
-nl_send_route(struct krt_proto *p, rte *e, struct ea_list *eattrs, int op, int dest, struct nexthop *nh)
+nl_send_route(struct krt_proto *p, rte *e, int op, int dest, struct nexthop *nh)
 {
   eattr *ea;
   net *net = e->net;
   rta *a = e->attrs;
+  ea_list *eattrs = a->eattrs;
   int bufsize = 128 + KRT_METRICS_MAX*8 + nh_bufsize(&(a->nh));
   u32 priority = 0;
 
@@ -1328,7 +1329,7 @@ dest:
 }
 
 static inline int
-nl_add_rte(struct krt_proto *p, rte *e, struct ea_list *eattrs)
+nl_add_rte(struct krt_proto *p, rte *e)
 {
   rta *a = e->attrs;
   int err = 0;
@@ -1337,34 +1338,34 @@ nl_add_rte(struct krt_proto *p, rte *e, struct ea_list *eattrs)
   {
     struct nexthop *nh = &(a->nh);
 
-    err = nl_send_route(p, e, eattrs, NL_OP_ADD, RTD_UNICAST, nh);
+    err = nl_send_route(p, e, NL_OP_ADD, RTD_UNICAST, nh);
     if (err < 0)
       return err;
 
     for (nh = nh->next; nh; nh = nh->next)
-      err += nl_send_route(p, e, eattrs, NL_OP_APPEND, RTD_UNICAST, nh);
+      err += nl_send_route(p, e, NL_OP_APPEND, RTD_UNICAST, nh);
 
     return err;
   }
 
-  return nl_send_route(p, e, eattrs, NL_OP_ADD, a->dest, &(a->nh));
+  return nl_send_route(p, e, NL_OP_ADD, a->dest, &(a->nh));
 }
 
 static inline int
-nl_delete_rte(struct krt_proto *p, rte *e, struct ea_list *eattrs)
+nl_delete_rte(struct krt_proto *p, rte *e)
 {
   int err = 0;
 
   /* For IPv6, we just repeatedly request DELETE until we get error */
   do
-    err = nl_send_route(p, e, eattrs, NL_OP_DELETE, RTD_NONE, NULL);
+    err = nl_send_route(p, e, NL_OP_DELETE, RTD_NONE, NULL);
   while (krt_ecmp6(p) && !err);
 
   return err;
 }
 
 void
-krt_replace_rte(struct krt_proto *p, net *n, rte *new, rte *old, struct ea_list *eattrs)
+krt_replace_rte(struct krt_proto *p, net *n, rte *new, rte *old)
 {
   int err = 0;
 
@@ -1380,10 +1381,10 @@ krt_replace_rte(struct krt_proto *p, net *n, rte *new, rte *old, struct ea_list
    */
 
   if (old)
-    nl_delete_rte(p, old, eattrs);
+    nl_delete_rte(p, old);
 
   if (new)
-    err = nl_add_rte(p, new, eattrs);
+    err = nl_add_rte(p, new);
 
   if (err < 0)
     n->n.flags |= KRF_SYNC_ERROR;
diff --git a/sysdep/unix/krt.Y b/sysdep/unix/krt.Y
index 9aac8668..98740b70 100644
--- a/sysdep/unix/krt.Y
+++ b/sysdep/unix/krt.Y
@@ -122,8 +122,8 @@ kif_iface:
   kif_iface_start iface_patt_list_nopx kif_iface_opt_list;
 
 
-CF_ADDTO(dynamic_attr, KRT_SOURCE { $$ = f_new_dynamic_attr(EAF_TYPE_INT | EAF_TEMP, T_INT, EA_KRT_SOURCE); })
-CF_ADDTO(dynamic_attr, KRT_METRIC { $$ = f_new_dynamic_attr(EAF_TYPE_INT | EAF_TEMP, T_INT, EA_KRT_METRIC); })
+CF_ADDTO(dynamic_attr, KRT_SOURCE { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_SOURCE); })
+CF_ADDTO(dynamic_attr, KRT_METRIC { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_METRIC); })
 
 CF_CODE
 
diff --git a/sysdep/unix/krt.c b/sysdep/unix/krt.c
index ae51927e..a79df54e 100644
--- a/sysdep/unix/krt.c
+++ b/sysdep/unix/krt.c
@@ -551,7 +551,7 @@ krt_flush_routes(struct krt_proto *p)
       if (rte_is_valid(e) && (n->n.flags & KRF_INSTALLED))
 	{
 	  /* FIXME: this does not work if gw is changed in export filter */
-	  krt_replace_rte(p, e->net, NULL, e, NULL);
+	  krt_replace_rte(p, e->net, NULL, e);
 	  n->n.flags &= ~KRF_INSTALLED;
 	}
     }
@@ -559,14 +559,14 @@ krt_flush_routes(struct krt_proto *p)
 }
 
 static struct rte *
-krt_export_net(struct krt_proto *p, net *net, rte **rt_free, ea_list **tmpa)
+krt_export_net(struct krt_proto *p, net *net, rte **rt_free)
 {
   struct channel *c = p->p.main_channel;
   struct filter *filter = c->out_filter;
   rte *rt;
 
   if (c->ra_mode == RA_MERGED)
-    return rt_export_merged(c, net, rt_free, tmpa, krt_filter_lp, 1);
+    return rt_export_merged(c, net, rt_free, krt_filter_lp, 1);
 
   rt = net->routes;
   *rt_free = NULL;
@@ -577,15 +577,14 @@ krt_export_net(struct krt_proto *p, net *net, rte **rt_free, ea_list **tmpa)
   if (filter == FILTER_REJECT)
     return NULL;
 
-  struct proto *src = rt->attrs->src->proto;
-  *tmpa = src->make_tmp_attrs ? src->make_tmp_attrs(rt, krt_filter_lp) : NULL;
+  rte_make_tmp_attrs(&rt, krt_filter_lp);
 
   /* We could run krt_import_control() here, but it is already handled by KRF_INSTALLED */
 
   if (filter == FILTER_ACCEPT)
     goto accept;
 
-  if (f_run(filter, &rt, tmpa, krt_filter_lp, FF_FORCE_TMPATTR | FF_SILENT) > F_ACCEPT)
+  if (f_run(filter, &rt, krt_filter_lp, FF_SILENT) > F_ACCEPT)
     goto reject;
 
 
@@ -667,9 +666,8 @@ krt_got_route(struct krt_proto *p, rte *e)
   if (net->n.flags & KRF_INSTALLED)
     {
       rte *new, *rt_free;
-      ea_list *tmpa;
 
-      new = krt_export_net(p, net, &rt_free, &tmpa);
+      new = krt_export_net(p, net, &rt_free);
 
       /* TODO: There also may be changes in route eattrs, we ignore that for now. */
 
@@ -714,7 +712,6 @@ krt_prune(struct krt_proto *p)
     {
       int verdict = n->n.flags & KRF_VERDICT_MASK;
       rte *new, *old, *rt_free = NULL;
-      ea_list *tmpa = NULL;
 
       if (verdict == KRF_UPDATE || verdict == KRF_DELETE)
 	{
@@ -728,12 +725,10 @@ krt_prune(struct krt_proto *p)
       if (verdict == KRF_CREATE || verdict == KRF_UPDATE)
 	{
 	  /* We have to run export filter to get proper 'new' route */
-	  new = krt_export_net(p, n, &rt_free, &tmpa);
+	  new = krt_export_net(p, n, &rt_free);
 
 	  if (!new)
 	    verdict = (verdict == KRF_CREATE) ? KRF_IGNORE : KRF_DELETE;
-	  else
-	    tmpa = ea_append(tmpa, new->attrs->eattrs);
 	}
       else
 	new = NULL;
@@ -744,7 +739,7 @@ krt_prune(struct krt_proto *p)
 	  if (new && (n->n.flags & KRF_INSTALLED))
 	    {
 	      krt_trace_in(p, new, "reinstalling");
-	      krt_replace_rte(p, n, new, NULL, tmpa);
+	      krt_replace_rte(p, n, new, NULL);
 	    }
 	  break;
 	case KRF_SEEN:
@@ -753,11 +748,11 @@ krt_prune(struct krt_proto *p)
 	  break;
 	case KRF_UPDATE:
 	  krt_trace_in(p, new, "updating");
-	  krt_replace_rte(p, n, new, old, tmpa);
+	  krt_replace_rte(p, n, new, old);
 	  break;
 	case KRF_DELETE:
 	  krt_trace_in(p, old, "deleting");
-	  krt_replace_rte(p, n, NULL, old, NULL);
+	  krt_replace_rte(p, n, NULL, old);
 	  break;
 	default:
 	  bug("krt_prune: invalid route status");
@@ -795,7 +790,7 @@ krt_got_route_async(struct krt_proto *p, rte *e, int new)
       if (new)
 	{
 	  krt_trace_in(p, e, "[redirect] deleting");
-	  krt_replace_rte(p, net, NULL, e, NULL);
+	  krt_replace_rte(p, net, NULL, e);
 	}
       /* If !new, it is probably echo of our deletion */
       break;
@@ -937,14 +932,14 @@ krt_make_tmp_attrs(rte *rt, struct linpool *pool)
 }
 
 static void
-krt_store_tmp_attrs(rte *rt, struct ea_list *attrs)
+krt_store_tmp_attrs(rte *rt)
 {
   /* EA_KRT_SOURCE is read-only */
-  rt->u.krt.metric = ea_get_int(attrs, EA_KRT_METRIC, 0);
+  rt->u.krt.metric = ea_get_int(rt->attrs->eattrs, EA_KRT_METRIC, 0);
 }
 
 static int
-krt_import_control(struct proto *P, rte **new, ea_list **attrs UNUSED, struct linpool *pool UNUSED)
+krt_import_control(struct proto *P, rte **new, struct linpool *pool UNUSED)
 {
   // struct krt_proto *p = (struct krt_proto *) P;
   rte *e = *new;
@@ -975,7 +970,7 @@ krt_import_control(struct proto *P, rte **new, ea_list **attrs UNUSED, struct li
 
 static void
 krt_rt_notify(struct proto *P, struct channel *ch UNUSED, net *net,
-	      rte *new, rte *old, struct ea_list *eattrs)
+	      rte *new, rte *old)
 {
   struct krt_proto *p = (struct krt_proto *) P;
 
@@ -988,7 +983,7 @@ krt_rt_notify(struct proto *P, struct channel *ch UNUSED, net *net,
   else
     net->n.flags &= ~KRF_INSTALLED;
   if (p->initialized)		/* Before first scan we don't touch the routes */
-    krt_replace_rte(p, net, new, old, eattrs);
+    krt_replace_rte(p, net, new, old);
 }
 
 static void
diff --git a/sysdep/unix/krt.h b/sysdep/unix/krt.h
index ff27bcf5..6ace2a86 100644
--- a/sysdep/unix/krt.h
+++ b/sysdep/unix/krt.h
@@ -139,7 +139,7 @@ void krt_sys_copy_config(struct krt_config *, struct krt_config *);
 
 int  krt_capable(rte *e);
 void krt_do_scan(struct krt_proto *);
-void krt_replace_rte(struct krt_proto *p, net *n, rte *new, rte *old, struct ea_list *eattrs);
+void krt_replace_rte(struct krt_proto *p, net *n, rte *new, rte *old);
 int krt_sys_get_attr(eattr *a, byte *buf, int buflen);
 
 

From 9c9050ff12c52762708dadda78a05108a5b533b2 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 13 Jun 2018 14:47:37 +0200
Subject: [PATCH 173/198] BGP: Handle missing NEXT_HOP attribute properly

RFC 7606 specifies handle-as-withdraw instead of session reset.
---
 proto/bgp/attrs.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/proto/bgp/attrs.c b/proto/bgp/attrs.c
index 9003feb2..d7a4e692 100644
--- a/proto/bgp/attrs.c
+++ b/proto/bgp/attrs.c
@@ -1109,6 +1109,9 @@ bgp_decode_attrs(struct bgp_parse_state *s, byte *data, uint len)
   if (!BIT32_TEST(s->attrs_seen, BA_AS_PATH))
   { REPORT(NO_MANDATORY, "AS_PATH"); goto withdraw; }
 
+  if (s->ip_reach_len && !BIT32_TEST(s->attrs_seen, BA_NEXT_HOP))
+  { REPORT(NO_MANDATORY, "NEXT_HOP"); goto withdraw; }
+
   /* When receiving attributes from non-AS4-aware BGP speaker, we have to
      reconstruct AS_PATH and AGGREGATOR attributes; RFC 6793 4.2.3 */
   if (!p->as4_session)

From caa9d03d65ce827ce536d54b26988e70767e032f Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 13 Jun 2018 15:22:29 +0200
Subject: [PATCH 174/198] Babel: Fix handling of missing IPv4 next hops

In case of missing IPv4 next hop, we should skip such routes
on transmit and ignore such routes on receive.

Thanks to Julian Schuh for the bugreport and Toke Hoiland-Jorgensen
for the original patch.
---
 proto/babel/babel.c   | 8 ++++++--
 proto/babel/packets.c | 2 +-
 2 files changed, 7 insertions(+), 3 deletions(-)

diff --git a/proto/babel/babel.c b/proto/babel/babel.c
index 44c6adb8..20953044 100644
--- a/proto/babel/babel.c
+++ b/proto/babel/babel.c
@@ -952,6 +952,10 @@ babel_send_update_(struct babel_iface *ifa, btime changed, struct fib *rtable)
     msg.update.next_hop = ((e->n.addr->type == NET_IP4) ?
 			   ifa->next_hop_ip4 : ifa->next_hop_ip6);
 
+    /* Do not send route if next hop is unknown, e.g. no configured IPv4 address */
+    if (ipa_zero(msg.update.next_hop))
+      continue;
+
     babel_enqueue(&msg, ifa);
 
     /* Update feasibility distance for redistributed routes */
@@ -1580,7 +1584,7 @@ babel_add_iface(struct babel_proto *p, struct iface *new, struct babel_iface_con
   ifa->next_hop_ip6 = ipa_nonzero(ic->next_hop_ip6) ? ic->next_hop_ip6 : ifa->addr;
 
   if (ipa_zero(ifa->next_hop_ip4) && p->ip4_channel)
-    log(L_WARN "%s: Cannot find IPv4 next hop addr on %s", p->p.name, new->name);
+    log(L_WARN "%s: Missing IPv4 next hop address for %s", p->p.name, new->name);
 
   init_list(&ifa->neigh_list);
   ifa->hello_seqno = 1;
@@ -1682,7 +1686,7 @@ babel_reconfigure_iface(struct babel_proto *p, struct babel_iface *ifa, struct b
   ifa->next_hop_ip6 = ipa_nonzero(new->next_hop_ip6) ? new->next_hop_ip6 : ifa->addr;
 
   if (ipa_zero(ifa->next_hop_ip4) && p->ip4_channel)
-    log(L_WARN "%s: Cannot find IPv4 next hop addr on %s", p->p.name, ifa->ifname);
+    log(L_WARN "%s: Missing IPv4 next hop address for %s", p->p.name, ifa->ifname);
 
   if (ifa->next_hello > (current_time() + new->hello_interval))
     ifa->next_hello = current_time() + (random() % new->hello_interval);
diff --git a/proto/babel/packets.c b/proto/babel/packets.c
index 59678678..d4ecf649 100644
--- a/proto/babel/packets.c
+++ b/proto/babel/packets.c
@@ -616,7 +616,7 @@ babel_read_update(struct babel_tlv *hdr, union babel_msg *m,
 
     /* Update must have next hop, unless it is retraction */
     if (ipa_zero(state->next_hop_ip4) && (msg->metric != BABEL_INFINITY))
-      return PARSE_ERROR;
+      return PARSE_IGNORE;
 
     /* Merge saved prefix and received prefix parts */
     memcpy(buf, state->def_ip4_prefix, tlv->omitted);

From 1771f70d7473b8c8e0c6bd47d5c35cc5fbe7eb19 Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Tue, 19 Jun 2018 16:16:08 +0200
Subject: [PATCH 175/198] Filter: fixed eattr cached pointer

Use ACCESS_RTE to guard **f_rte, use ACCESS_EATTRS to guard **f_eattrs.
Use f_rta_cow() before writing to rta or eattrs, use f_rte_cow() before
writing preference (stored in rte).

Do not access eattrs indirectly through (*f_rte)->attrs->eattrs, it is
way too slow. The cached pointer is faster.
---
 filter/filter.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/filter/filter.c b/filter/filter.c
index 6290e74a..439fd5bf 100644
--- a/filter/filter.c
+++ b/filter/filter.c
@@ -554,7 +554,6 @@ static inline void f_rte_cow(void)
     return;
 
   *f_rte = rte_do_cow(*f_rte);
-  f_eattrs = NULL;
 }
 
 /*
@@ -615,7 +614,7 @@ static struct tbf rl_runtime_err = TBF_DEFAULT_LOG_LIMITS;
       return val;
 
 #define ACCESS_RTE \
-  do { if (!f_rte) runtime("No route to access"); else f_cache_eattrs(); } while (0)
+  do { if (!f_rte) runtime("No route to access"); } while (0)
 
 #define ACCESS_EATTRS \
   do { if (!f_eattrs) f_cache_eattrs(); } while (0)
@@ -1753,6 +1752,7 @@ f_run(struct filter *filter, struct rte **rte, struct linpool *tmp_pool, int fla
   DBG( "Running filter `%s'...", filter->name );
 
   f_rte = rte;
+  f_eattrs = NULL;
   f_old_rta = NULL;
   f_pool = tmp_pool;
   f_flags = flags;
@@ -1797,6 +1797,7 @@ f_eval_rte(struct f_inst *expr, struct rte **rte, struct linpool *tmp_pool)
 {
 
   f_rte = rte;
+  f_eattrs = NULL;
   f_old_rta = NULL;
   f_pool = tmp_pool;
   f_flags = 0;
@@ -1813,6 +1814,7 @@ struct f_val
 f_eval(struct f_inst *expr, struct linpool *tmp_pool)
 {
   f_flags = 0;
+  f_eattrs = NULL;
   f_rte = NULL;
   f_pool = tmp_pool;
 

From f851f0d7e30cbd1e2ba20a5bf06a584acc136828 Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Tue, 26 Jun 2018 14:29:03 +0200
Subject: [PATCH 176/198] Config: Dropping CF_ADDTO.

---
 conf/confbase.Y        |  5 ++--
 conf/gen_parser.m4     | 11 +-------
 filter/config.Y        | 15 +++--------
 nest/config.Y          | 17 ++++++-------
 proto/babel/config.Y   |  4 +--
 proto/bfd/config.Y     |  2 +-
 proto/bgp/config.Y     | 50 ++++++++++++++++++------------------
 proto/ospf/config.Y    | 10 ++++----
 proto/pipe/config.Y    |  2 +-
 proto/radv/config.Y    |  6 ++---
 proto/rip/config.Y     |  6 ++---
 proto/rpki/config.Y    |  2 +-
 proto/static/config.Y  |  2 +-
 sysdep/bsd/krt-sock.Y  |  2 +-
 sysdep/linux/netlink.Y | 58 +++++++++++++++++++++---------------------
 sysdep/unix/config.Y   |  6 ++---
 sysdep/unix/krt.Y      | 16 ++++++------
 17 files changed, 99 insertions(+), 115 deletions(-)

diff --git a/conf/confbase.Y b/conf/confbase.Y
index 72f56f1e..8fd72e5e 100644
--- a/conf/confbase.Y
+++ b/conf/confbase.Y
@@ -113,12 +113,13 @@ conf_entries:
  | conf_entries conf
  ;
 
-CF_ADDTO(conf, ';')
+conf: ';' ;
 
 
 /* Constant expressions */
 
-CF_ADDTO(conf, definition)
+conf: definition ;
+
 definition:
    DEFINE SYM '=' term ';' {
      struct f_val *val = cfg_alloc(sizeof(struct f_val));
diff --git a/conf/gen_parser.m4 b/conf/gen_parser.m4
index 00b55023..af43d96f 100644
--- a/conf/gen_parser.m4
+++ b/conf/gen_parser.m4
@@ -34,15 +34,10 @@ m4_define(CF_keywd, `m4_ifdef([[CF_tok_$1]],,[[m4_define([[CF_tok_$1]],1)m4_defi
 m4_define(CF_KEYWORDS, `m4_define([[CF_toks]],[[]])CF_iterate([[CF_keywd]], [[$@]])m4_ifelse(CF_toks,,,%token[[]]CF_toks
 )DNL')
 
-# Dynamic syntax rules
-m4_define(CF_dyn_rules,)
-m4_define(CF_ADDTO, `m4_define([[CF_rule_$1]],m4_ifdef([[CF_rule_$1]],CF_rule_$1 | ,[[m4_define([[CF_dyn_rules]],CF_dyn_rules[[CF_RULE($1)
-]])]])$2)DNL')
-
 # CLI commands
 m4_define(CF_CLI, `m4_define([[CF_cmd]], cmd_[[]]m4_translit($1, [[ ]], _))DNL
 m4_divert(2)CF_KEYWORDS(m4_translit($1, [[ ]], [[,]]))
-m4_divert(3)CF_ADDTO(cli_cmd, CF_cmd)
+m4_divert(3)cli_cmd: CF_cmd
 CF_cmd: $1 $2 END')
 m4_define(CF_CLI_CMD, `')
 m4_define(CF_CLI_HELP, `')
@@ -62,10 +57,6 @@ m4_undivert(2)DNL
 %%
 m4_undivert(3)DNL
 
-/* Dynamic rules */
-
-m4_define(CF_RULE, [[$1: CF_rule_$1 ;]])
-CF_dyn_rules
 %%
 m4_undivert(4)DNL
 ')
diff --git a/filter/config.Y b/filter/config.Y
index cf499b66..bbb8a640 100644
--- a/filter/config.Y
+++ b/filter/config.Y
@@ -438,7 +438,7 @@ CF_KEYWORDS(FUNCTION, PRINT, PRINTN, UNSET, RETURN,
 
 CF_GRAMMAR
 
-CF_ADDTO(conf, filter_def)
+conf: filter_def ;
 filter_def:
    FILTER SYM { $2 = cf_define_symbol($2, SYM_FILTER, NULL); cf_push_scope( $2 ); }
      filter_body {
@@ -449,12 +449,12 @@ filter_def:
    }
  ;
 
-CF_ADDTO(conf, filter_eval)
+conf: filter_eval ;
 filter_eval:
    EVAL term { f_eval_int($2); }
  ;
 
-CF_ADDTO(conf, bt_test_suite)
+conf: bt_test_suite ;
 bt_test_suite:
  BT_TEST_SUITE '(' SYM ',' text ')' {
   if (!($3->class & SYM_FUNCTION))
@@ -588,7 +588,7 @@ function_body:
    }
  ;
 
-CF_ADDTO(conf, function_def)
+conf: function_def ;
 function_def:
    FUNCTION SYM { DBG( "Beginning of function %s\n", $2->name );
      $2 = cf_define_symbol($2, SYM_FUNCTION, NULL);
@@ -804,13 +804,6 @@ constructor:
  ;
 
 
-/*
- *  Maybe there are no dynamic attributes defined by protocols.
- *  For such cases, we force the dynamic_attr list to contain
- *  at least an invalid token, so it is syntantically correct.
- */
-CF_ADDTO(dynamic_attr, INVALID_TOKEN { $$ = (struct f_dynamic_attr) {}; })
-
 rtadot: /* EMPTY, we are not permitted RTA. prefix */
  ;
 
diff --git a/nest/config.Y b/nest/config.Y
index ab09a10c..88f74b96 100644
--- a/nest/config.Y
+++ b/nest/config.Y
@@ -103,7 +103,7 @@ CF_GRAMMAR
 
 /* Setting of router ID */
 
-CF_ADDTO(conf, rtrid)
+conf: rtrid ;
 
 rtrid:
    ROUTER ID idval ';' { new_config->router_id = $3; }
@@ -124,7 +124,7 @@ idval:
    }
  ;
 
-CF_ADDTO(conf, gr_opts)
+conf: gr_opts ;
 
 gr_opts: GRACEFUL RESTART WAIT expr ';' { new_config->gr_wait = $4; } ;
 
@@ -149,7 +149,7 @@ CF_ENUM(T_ENUM_NETTYPE, NET_, IP4, IP6, VPN4, VPN6, ROA4, ROA6, FLOW4, FLOW6, IP
 
 /* Creation of routing tables */
 
-CF_ADDTO(conf, table)
+conf: table ;
 
 table_sorted:
 	  { $$ = 0; }
@@ -166,7 +166,7 @@ table: net_type TABLE SYM table_sorted {
 
 /* Definition of protocols */
 
-CF_ADDTO(conf, proto { proto_postconfig(); })
+conf: proto { proto_postconfig(); } ;
 
 proto_start:
    PROTOCOL { $$ = SYM_PROTO; }
@@ -282,7 +282,7 @@ limit_spec:
  ;
 
 
-CF_ADDTO(conf, debug_default)
+conf: debug_default ;
 
 debug_default:
    DEBUG PROTOCOLS debug_mask { new_config->proto_default_debug = $3; }
@@ -291,7 +291,7 @@ debug_default:
 
 /* MRTDUMP PROTOCOLS is in systep/unix/config.Y */
 
-CF_ADDTO(conf, timeformat_base)
+conf: timeformat_base ;
 
 timeformat_which:
    ROUTE { $$ = &new_config->tf_route; }
@@ -367,7 +367,7 @@ tos:
 
 /* Direct device route protocol */
 
-CF_ADDTO(proto, dev_proto '}')
+proto: dev_proto '}' ;
 
 dev_proto_start: proto_start DIRECT {
      this_proto = proto_config_new(&proto_device, $1);
@@ -782,8 +782,7 @@ proto_patt2:
  | TEXT { $$.ptr = $1; $$.patt = 1; }
  ;
 
-CF_ADDTO(dynamic_attr, IGP_METRIC
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_GEN_IGP_METRIC); })
+dynamic_attr: IGP_METRIC { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_GEN_IGP_METRIC); } ;
 
 
 CF_CODE
diff --git a/proto/babel/config.Y b/proto/babel/config.Y
index 2b20c43f..3af79fd6 100644
--- a/proto/babel/config.Y
+++ b/proto/babel/config.Y
@@ -29,7 +29,7 @@ CF_KEYWORDS(BABEL, INTERFACE, METRIC, RXCOST, HELLO, UPDATE, INTERVAL, PORT,
 
 CF_GRAMMAR
 
-CF_ADDTO(proto, babel_proto)
+proto: babel_proto ;
 
 babel_proto_start: proto_start BABEL
 {
@@ -125,7 +125,7 @@ babel_iface_opt_list:
 babel_iface:
   babel_iface_start iface_patt_list_nopx babel_iface_opt_list babel_iface_finish;
 
-CF_ADDTO(dynamic_attr, BABEL_METRIC { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_BABEL_METRIC); })
+dynamic_attr: BABEL_METRIC { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_BABEL_METRIC); } ;
 
 CF_CLI_HELP(SHOW BABEL, ..., [[Show information about Babel protocol]]);
 
diff --git a/proto/bfd/config.Y b/proto/bfd/config.Y
index 73414362..3f5714fd 100644
--- a/proto/bfd/config.Y
+++ b/proto/bfd/config.Y
@@ -31,7 +31,7 @@ CF_KEYWORDS(BFD, MIN, IDLE, RX, TX, INTERVAL, MULTIPLIER, PASSIVE,
 
 CF_GRAMMAR
 
-CF_ADDTO(proto, bfd_proto)
+proto: bfd_proto ;
 
 bfd_proto_start: proto_start BFD
 {
diff --git a/proto/bgp/config.Y b/proto/bgp/config.Y
index c731c4c7..7583f7d1 100644
--- a/proto/bgp/config.Y
+++ b/proto/bgp/config.Y
@@ -40,7 +40,7 @@ CF_KEYWORDS(CEASE, PREFIX, LIMIT, HIT, ADMINISTRATIVE, SHUTDOWN, RESET, PEER,
 
 CF_GRAMMAR
 
-CF_ADDTO(proto, bgp_proto '}' )
+proto: bgp_proto '}'  ;
 
 bgp_proto_start: proto_start BGP {
      this_proto = proto_config_new(&proto_bgp, $1);
@@ -252,30 +252,30 @@ bgp_channel_end:
 bgp_proto_channel: bgp_channel_start bgp_channel_opt_list bgp_channel_end;
 
 
-CF_ADDTO(dynamic_attr, BGP_ORIGIN
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_ENUM_BGP_ORIGIN, EA_CODE(PROTOCOL_BGP, BA_ORIGIN)); })
-CF_ADDTO(dynamic_attr, BGP_PATH
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_AS_PATH, T_PATH, EA_CODE(PROTOCOL_BGP, BA_AS_PATH)); })
-CF_ADDTO(dynamic_attr, BGP_NEXT_HOP
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_IP_ADDRESS, T_IP, EA_CODE(PROTOCOL_BGP, BA_NEXT_HOP)); })
-CF_ADDTO(dynamic_attr, BGP_MED
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_CODE(PROTOCOL_BGP, BA_MULTI_EXIT_DISC)); })
-CF_ADDTO(dynamic_attr, BGP_LOCAL_PREF
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_CODE(PROTOCOL_BGP, BA_LOCAL_PREF)); })
-CF_ADDTO(dynamic_attr, BGP_ATOMIC_AGGR
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_OPAQUE, T_ENUM_EMPTY, EA_CODE(PROTOCOL_BGP, BA_ATOMIC_AGGR)); })
-CF_ADDTO(dynamic_attr, BGP_AGGREGATOR
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_CODE(PROTOCOL_BGP, BA_AGGREGATOR)); })
-CF_ADDTO(dynamic_attr, BGP_COMMUNITY
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT_SET, T_CLIST, EA_CODE(PROTOCOL_BGP, BA_COMMUNITY)); })
-CF_ADDTO(dynamic_attr, BGP_ORIGINATOR_ID
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_ROUTER_ID, T_QUAD, EA_CODE(PROTOCOL_BGP, BA_ORIGINATOR_ID)); })
-CF_ADDTO(dynamic_attr, BGP_CLUSTER_LIST
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT_SET, T_CLIST, EA_CODE(PROTOCOL_BGP, BA_CLUSTER_LIST)); })
-CF_ADDTO(dynamic_attr, BGP_EXT_COMMUNITY
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_EC_SET, T_ECLIST, EA_CODE(PROTOCOL_BGP, BA_EXT_COMMUNITY)); })
-CF_ADDTO(dynamic_attr, BGP_LARGE_COMMUNITY
-	{ $$ = f_new_dynamic_attr(EAF_TYPE_LC_SET, T_LCLIST, EA_CODE(PROTOCOL_BGP, BA_LARGE_COMMUNITY)); })
+dynamic_attr: BGP_ORIGIN
+	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_ENUM_BGP_ORIGIN, EA_CODE(PROTOCOL_BGP, BA_ORIGIN)); } ;
+dynamic_attr: BGP_PATH
+	{ $$ = f_new_dynamic_attr(EAF_TYPE_AS_PATH, T_PATH, EA_CODE(PROTOCOL_BGP, BA_AS_PATH)); } ;
+dynamic_attr: BGP_NEXT_HOP
+	{ $$ = f_new_dynamic_attr(EAF_TYPE_IP_ADDRESS, T_IP, EA_CODE(PROTOCOL_BGP, BA_NEXT_HOP)); } ;
+dynamic_attr: BGP_MED
+	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_CODE(PROTOCOL_BGP, BA_MULTI_EXIT_DISC)); } ;
+dynamic_attr: BGP_LOCAL_PREF
+	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_CODE(PROTOCOL_BGP, BA_LOCAL_PREF)); } ;
+dynamic_attr: BGP_ATOMIC_AGGR
+	{ $$ = f_new_dynamic_attr(EAF_TYPE_OPAQUE, T_ENUM_EMPTY, EA_CODE(PROTOCOL_BGP, BA_ATOMIC_AGGR)); } ;
+dynamic_attr: BGP_AGGREGATOR
+	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_CODE(PROTOCOL_BGP, BA_AGGREGATOR)); } ;
+dynamic_attr: BGP_COMMUNITY
+	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT_SET, T_CLIST, EA_CODE(PROTOCOL_BGP, BA_COMMUNITY)); } ;
+dynamic_attr: BGP_ORIGINATOR_ID
+	{ $$ = f_new_dynamic_attr(EAF_TYPE_ROUTER_ID, T_QUAD, EA_CODE(PROTOCOL_BGP, BA_ORIGINATOR_ID)); } ;
+dynamic_attr: BGP_CLUSTER_LIST
+	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT_SET, T_CLIST, EA_CODE(PROTOCOL_BGP, BA_CLUSTER_LIST)); } ;
+dynamic_attr: BGP_EXT_COMMUNITY
+	{ $$ = f_new_dynamic_attr(EAF_TYPE_EC_SET, T_ECLIST, EA_CODE(PROTOCOL_BGP, BA_EXT_COMMUNITY)); } ;
+dynamic_attr: BGP_LARGE_COMMUNITY
+	{ $$ = f_new_dynamic_attr(EAF_TYPE_LC_SET, T_LCLIST, EA_CODE(PROTOCOL_BGP, BA_LARGE_COMMUNITY)); } ;
 
 
 
diff --git a/proto/ospf/config.Y b/proto/ospf/config.Y
index c439a614..b89584e1 100644
--- a/proto/ospf/config.Y
+++ b/proto/ospf/config.Y
@@ -207,7 +207,7 @@ CF_KEYWORDS(SECONDARY, MERGE, LSA, SUPPRESSION, MULTICAST, RFC5838)
 
 CF_GRAMMAR
 
-CF_ADDTO(proto, ospf_proto '}' { ospf_proto_finish(); } )
+proto: ospf_proto '}' { ospf_proto_finish(); }  ;
 
 ospf_variant:
    OSPF    { $$ = 1; }
@@ -497,10 +497,10 @@ ospf_iface:
   ospf_iface_start ospf_iface_patt_list ospf_iface_opt_list { ospf_iface_finish(); }
  ;
 
-CF_ADDTO(dynamic_attr, OSPF_METRIC1 { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_OSPF_METRIC1); })
-CF_ADDTO(dynamic_attr, OSPF_METRIC2 { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_OSPF_METRIC2); })
-CF_ADDTO(dynamic_attr, OSPF_TAG { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_OSPF_TAG); })
-CF_ADDTO(dynamic_attr, OSPF_ROUTER_ID { $$ = f_new_dynamic_attr(EAF_TYPE_ROUTER_ID, T_QUAD, EA_OSPF_ROUTER_ID); })
+dynamic_attr: OSPF_METRIC1 { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_OSPF_METRIC1); } ;
+dynamic_attr: OSPF_METRIC2 { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_OSPF_METRIC2); } ;
+dynamic_attr: OSPF_TAG { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_OSPF_TAG); } ;
+dynamic_attr: OSPF_ROUTER_ID { $$ = f_new_dynamic_attr(EAF_TYPE_ROUTER_ID, T_QUAD, EA_OSPF_ROUTER_ID); } ;
 
 CF_CLI_HELP(SHOW OSPF, ..., [[Show information about OSPF protocol]]);
 CF_CLI(SHOW OSPF, optsym, [<name>], [[Show information about OSPF protocol]])
diff --git a/proto/pipe/config.Y b/proto/pipe/config.Y
index b3c332be..18e1d0fc 100644
--- a/proto/pipe/config.Y
+++ b/proto/pipe/config.Y
@@ -20,7 +20,7 @@ CF_KEYWORDS(PIPE, PEER, TABLE)
 
 CF_GRAMMAR
 
-CF_ADDTO(proto, pipe_proto '}' { this_channel = NULL; } )
+proto: pipe_proto '}' { this_channel = NULL; }  ;
 
 pipe_proto_start: proto_start PIPE
 {
diff --git a/proto/radv/config.Y b/proto/radv/config.Y
index 37815f0d..53715f77 100644
--- a/proto/radv/config.Y
+++ b/proto/radv/config.Y
@@ -39,7 +39,7 @@ CF_ENUM(T_ENUM_RA_PREFERENCE, RA_PREF_, LOW, MEDIUM, HIGH)
 
 CF_GRAMMAR
 
-CF_ADDTO(proto, radv_proto)
+proto: radv_proto ;
 
 radv_proto_start: proto_start RADV
 {
@@ -332,8 +332,8 @@ radv_sensitive:
  | SENSITIVE bool { $$ = $2; }
  ;
 
-CF_ADDTO(dynamic_attr, RA_PREFERENCE { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_ENUM_RA_PREFERENCE, EA_RA_PREFERENCE); })
-CF_ADDTO(dynamic_attr, RA_LIFETIME { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_RA_LIFETIME); })
+dynamic_attr: RA_PREFERENCE { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_ENUM_RA_PREFERENCE, EA_RA_PREFERENCE); } ;
+dynamic_attr: RA_LIFETIME { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_RA_LIFETIME); } ;
 
 CF_CODE
 
diff --git a/proto/rip/config.Y b/proto/rip/config.Y
index c46cf00c..172299d0 100644
--- a/proto/rip/config.Y
+++ b/proto/rip/config.Y
@@ -42,7 +42,7 @@ CF_KEYWORDS(RIP, NG, ECMP, LIMIT, WEIGHT, INFINITY, METRIC, UPDATE, TIMEOUT,
 
 CF_GRAMMAR
 
-CF_ADDTO(proto, rip_proto)
+proto: rip_proto ;
 
 rip_variant:
    RIP    { $$ = 1; }
@@ -186,8 +186,8 @@ rip_iface:
   rip_iface_start iface_patt_list_nopx rip_iface_opt_list rip_iface_finish;
 
 
-CF_ADDTO(dynamic_attr, RIP_METRIC { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_RIP_METRIC); })
-CF_ADDTO(dynamic_attr, RIP_TAG { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_RIP_TAG); })
+dynamic_attr: RIP_METRIC { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_RIP_METRIC); } ;
+dynamic_attr: RIP_TAG { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_RIP_TAG); } ;
 
 CF_CLI_HELP(SHOW RIP, ..., [[Show information about RIP protocol]]);
 
diff --git a/proto/rpki/config.Y b/proto/rpki/config.Y
index 39fdfd01..a88a29a1 100644
--- a/proto/rpki/config.Y
+++ b/proto/rpki/config.Y
@@ -38,7 +38,7 @@ CF_KEYWORDS(RPKI, REMOTE, BIRD, PRIVATE, PUBLIC, KEY, TCP, SSH, TRANSPORT, USER,
 
 CF_GRAMMAR
 
-CF_ADDTO(proto, rpki_proto)
+proto: rpki_proto ;
 
 rpki_proto_start: proto_start RPKI {
   this_proto = proto_config_new(&proto_rpki, $1);
diff --git a/proto/static/config.Y b/proto/static/config.Y
index 66e5ea4c..d7a02961 100644
--- a/proto/static/config.Y
+++ b/proto/static/config.Y
@@ -49,7 +49,7 @@ CF_KEYWORDS(ONLINK, WEIGHT, RECURSIVE, IGP, TABLE, BLACKHOLE, UNREACHABLE, BFD,
 
 CF_GRAMMAR
 
-CF_ADDTO(proto, static_proto '}')
+proto: static_proto '}' ;
 
 static_proto_start: proto_start STATIC
 {
diff --git a/sysdep/bsd/krt-sock.Y b/sysdep/bsd/krt-sock.Y
index 81422c79..8581bd43 100644
--- a/sysdep/bsd/krt-sock.Y
+++ b/sysdep/bsd/krt-sock.Y
@@ -14,7 +14,7 @@ CF_KEYWORDS(KERNEL, TABLE)
 
 CF_GRAMMAR
 
-CF_ADDTO(kern_proto, kern_proto kern_sys_item ';')
+kern_proto: kern_proto kern_sys_item ';' ;
 
 kern_sys_item:
    KERNEL TABLE expr {
diff --git a/sysdep/linux/netlink.Y b/sysdep/linux/netlink.Y
index f577244d..1030e5d4 100644
--- a/sysdep/linux/netlink.Y
+++ b/sysdep/linux/netlink.Y
@@ -19,44 +19,44 @@ CF_KEYWORDS(KERNEL, TABLE, METRIC, KRT_PREFSRC, KRT_REALM, KRT_SCOPE, KRT_MTU, K
 
 CF_GRAMMAR
 
-CF_ADDTO(kern_proto, kern_proto kern_sys_item ';')
+kern_proto: kern_proto kern_sys_item ';' ;
 
 kern_sys_item:
    KERNEL TABLE expr { THIS_KRT->sys.table_id = $3; }
  | METRIC expr { THIS_KRT->sys.metric = $2; }
  ;
 
-CF_ADDTO(dynamic_attr, KRT_PREFSRC	{ $$ = f_new_dynamic_attr(EAF_TYPE_IP_ADDRESS, T_IP, EA_KRT_PREFSRC); })
-CF_ADDTO(dynamic_attr, KRT_REALM	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_REALM); })
-CF_ADDTO(dynamic_attr, KRT_SCOPE	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_SCOPE); })
+dynamic_attr: KRT_PREFSRC	{ $$ = f_new_dynamic_attr(EAF_TYPE_IP_ADDRESS, T_IP, EA_KRT_PREFSRC); } ;
+dynamic_attr: KRT_REALM	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_REALM); } ;
+dynamic_attr: KRT_SCOPE	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_SCOPE); } ;
 
-CF_ADDTO(dynamic_attr, KRT_MTU		{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_MTU); })
-CF_ADDTO(dynamic_attr, KRT_WINDOW	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_WINDOW); })
-CF_ADDTO(dynamic_attr, KRT_RTT		{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_RTT); })
-CF_ADDTO(dynamic_attr, KRT_RTTVAR	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_RTTVAR); })
-CF_ADDTO(dynamic_attr, KRT_SSTRESH	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_SSTRESH); })
-CF_ADDTO(dynamic_attr, KRT_CWND		{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_CWND); })
-CF_ADDTO(dynamic_attr, KRT_ADVMSS	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_ADVMSS); })
-CF_ADDTO(dynamic_attr, KRT_REORDERING	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_REORDERING); })
-CF_ADDTO(dynamic_attr, KRT_HOPLIMIT	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_HOPLIMIT); })
-CF_ADDTO(dynamic_attr, KRT_INITCWND	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_INITCWND); })
-CF_ADDTO(dynamic_attr, KRT_RTO_MIN	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_RTO_MIN); })
-CF_ADDTO(dynamic_attr, KRT_INITRWND	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_INITRWND); })
-CF_ADDTO(dynamic_attr, KRT_QUICKACK	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_QUICKACK); })
+dynamic_attr: KRT_MTU		{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_MTU); } ;
+dynamic_attr: KRT_WINDOW	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_WINDOW); } ;
+dynamic_attr: KRT_RTT		{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_RTT); } ;
+dynamic_attr: KRT_RTTVAR	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_RTTVAR); } ;
+dynamic_attr: KRT_SSTRESH	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_SSTRESH); } ;
+dynamic_attr: KRT_CWND		{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_CWND); } ;
+dynamic_attr: KRT_ADVMSS	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_ADVMSS); } ;
+dynamic_attr: KRT_REORDERING	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_REORDERING); } ;
+dynamic_attr: KRT_HOPLIMIT	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_HOPLIMIT); } ;
+dynamic_attr: KRT_INITCWND	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_INITCWND); } ;
+dynamic_attr: KRT_RTO_MIN	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_RTO_MIN); } ;
+dynamic_attr: KRT_INITRWND	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_INITRWND); } ;
+dynamic_attr: KRT_QUICKACK	{ $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_QUICKACK); } ;
 
-CF_ADDTO(dynamic_attr, KRT_LOCK_MTU	{ $$ = f_new_dynamic_attr(EAF_TYPE_BITFIELD, T_BOOL, EA_KRT_LOCK_MTU); })
-CF_ADDTO(dynamic_attr, KRT_LOCK_WINDOW	{ $$ = f_new_dynamic_attr(EAF_TYPE_BITFIELD, T_BOOL, EA_KRT_LOCK_WINDOW); })
-CF_ADDTO(dynamic_attr, KRT_LOCK_RTT	{ $$ = f_new_dynamic_attr(EAF_TYPE_BITFIELD, T_BOOL, EA_KRT_LOCK_RTT); })
-CF_ADDTO(dynamic_attr, KRT_LOCK_RTTVAR	{ $$ = f_new_dynamic_attr(EAF_TYPE_BITFIELD, T_BOOL, EA_KRT_LOCK_RTTVAR); })
-CF_ADDTO(dynamic_attr, KRT_LOCK_SSTRESH	{ $$ = f_new_dynamic_attr(EAF_TYPE_BITFIELD, T_BOOL, EA_KRT_LOCK_SSTHRESH); })
-CF_ADDTO(dynamic_attr, KRT_LOCK_CWND	{ $$ = f_new_dynamic_attr(EAF_TYPE_BITFIELD, T_BOOL, EA_KRT_LOCK_CWND); })
-CF_ADDTO(dynamic_attr, KRT_LOCK_ADVMSS	{ $$ = f_new_dynamic_attr(EAF_TYPE_BITFIELD, T_BOOL, EA_KRT_LOCK_ADVMSS); })
-CF_ADDTO(dynamic_attr, KRT_LOCK_REORDERING { $$ = f_new_dynamic_attr(EAF_TYPE_BITFIELD, T_BOOL, EA_KRT_LOCK_REORDERING); })
-CF_ADDTO(dynamic_attr, KRT_LOCK_HOPLIMIT { $$ = f_new_dynamic_attr(EAF_TYPE_BITFIELD, T_BOOL, EA_KRT_LOCK_HOPLIMIT); })
-CF_ADDTO(dynamic_attr, KRT_LOCK_RTO_MIN { $$ = f_new_dynamic_attr(EAF_TYPE_BITFIELD, T_BOOL, EA_KRT_LOCK_RTO_MIN); })
+dynamic_attr: KRT_LOCK_MTU	{ $$ = f_new_dynamic_attr(EAF_TYPE_BITFIELD, T_BOOL, EA_KRT_LOCK_MTU); } ;
+dynamic_attr: KRT_LOCK_WINDOW	{ $$ = f_new_dynamic_attr(EAF_TYPE_BITFIELD, T_BOOL, EA_KRT_LOCK_WINDOW); } ;
+dynamic_attr: KRT_LOCK_RTT	{ $$ = f_new_dynamic_attr(EAF_TYPE_BITFIELD, T_BOOL, EA_KRT_LOCK_RTT); } ;
+dynamic_attr: KRT_LOCK_RTTVAR	{ $$ = f_new_dynamic_attr(EAF_TYPE_BITFIELD, T_BOOL, EA_KRT_LOCK_RTTVAR); } ;
+dynamic_attr: KRT_LOCK_SSTRESH	{ $$ = f_new_dynamic_attr(EAF_TYPE_BITFIELD, T_BOOL, EA_KRT_LOCK_SSTHRESH); } ;
+dynamic_attr: KRT_LOCK_CWND	{ $$ = f_new_dynamic_attr(EAF_TYPE_BITFIELD, T_BOOL, EA_KRT_LOCK_CWND); } ;
+dynamic_attr: KRT_LOCK_ADVMSS	{ $$ = f_new_dynamic_attr(EAF_TYPE_BITFIELD, T_BOOL, EA_KRT_LOCK_ADVMSS); } ;
+dynamic_attr: KRT_LOCK_REORDERING { $$ = f_new_dynamic_attr(EAF_TYPE_BITFIELD, T_BOOL, EA_KRT_LOCK_REORDERING); } ;
+dynamic_attr: KRT_LOCK_HOPLIMIT { $$ = f_new_dynamic_attr(EAF_TYPE_BITFIELD, T_BOOL, EA_KRT_LOCK_HOPLIMIT); } ;
+dynamic_attr: KRT_LOCK_RTO_MIN { $$ = f_new_dynamic_attr(EAF_TYPE_BITFIELD, T_BOOL, EA_KRT_LOCK_RTO_MIN); } ;
 
-CF_ADDTO(dynamic_attr, KRT_FEATURE_ECN	{ $$ = f_new_dynamic_attr(EAF_TYPE_BITFIELD, T_BOOL, EA_KRT_FEATURE_ECN); })
-CF_ADDTO(dynamic_attr, KRT_FEATURE_ALLFRAG { $$ = f_new_dynamic_attr(EAF_TYPE_BITFIELD, T_BOOL, EA_KRT_FEATURE_ALLFRAG); })
+dynamic_attr: KRT_FEATURE_ECN	{ $$ = f_new_dynamic_attr(EAF_TYPE_BITFIELD, T_BOOL, EA_KRT_FEATURE_ECN); } ;
+dynamic_attr: KRT_FEATURE_ALLFRAG { $$ = f_new_dynamic_attr(EAF_TYPE_BITFIELD, T_BOOL, EA_KRT_FEATURE_ALLFRAG); } ;
 
 
 CF_CODE
diff --git a/sysdep/unix/config.Y b/sysdep/unix/config.Y
index ccca4a62..1bffa322 100644
--- a/sysdep/unix/config.Y
+++ b/sysdep/unix/config.Y
@@ -24,7 +24,7 @@ CF_KEYWORDS(NAME, CONFIRM, UNDO, CHECK, TIMEOUT, DEBUG, LATENCY, LIMIT, WATCHDOG
 
 CF_GRAMMAR
 
-CF_ADDTO(conf, log_config)
+conf: log_config ;
 
 log_config: LOG log_file log_mask ';' {
     struct log_config *c = cfg_allocz(sizeof(struct log_config));
@@ -72,7 +72,7 @@ log_cat:
  ;
 
 
-CF_ADDTO(conf, mrtdump_base)
+conf: mrtdump_base ;
 
 mrtdump_base:
    MRTDUMP PROTOCOLS mrtdump_mask ';' { new_config->proto_default_mrtdump = $3; }
@@ -84,7 +84,7 @@ mrtdump_base:
  ;
 
 
-CF_ADDTO(conf, debug_unix)
+conf: debug_unix ;
 
 debug_unix:
    DEBUG LATENCY bool { new_config->latency_debug = $3; }
diff --git a/sysdep/unix/krt.Y b/sysdep/unix/krt.Y
index 98740b70..95b54d65 100644
--- a/sysdep/unix/krt.Y
+++ b/sysdep/unix/krt.Y
@@ -38,15 +38,15 @@ CF_GRAMMAR
 
 /* Kernel syncer protocol */
 
-CF_ADDTO(proto, kern_proto '}')
+proto: kern_proto '}' ;
 
 kern_proto_start: proto_start KERNEL {
      this_proto = krt_init_config($1);
 }
  ;
 
-CF_ADDTO(kern_proto, kern_proto_start proto_name '{')
-CF_ADDTO(kern_proto, kern_proto kern_item ';')
+kern_proto: kern_proto_start proto_name '{' ;
+kern_proto: kern_proto kern_item ';' ;
 
 kern_mp_limit:
    /* empty */ { $$ = KRT_DEFAULT_ECMP_LIMIT; }
@@ -80,13 +80,13 @@ kern_item:
 
 /* Kernel interface protocol */
 
-CF_ADDTO(proto, kif_proto '}')
+proto: kif_proto '}' ;
 
 kif_proto_start: proto_start DEVICE { this_proto = kif_init_config($1); }
  ;
 
-CF_ADDTO(kif_proto, kif_proto_start proto_name '{')
-CF_ADDTO(kif_proto, kif_proto kif_item ';')
+kif_proto: kif_proto_start proto_name '{' ;
+kif_proto: kif_proto kif_item ';' ;
 
 kif_item:
    proto_item
@@ -122,8 +122,8 @@ kif_iface:
   kif_iface_start iface_patt_list_nopx kif_iface_opt_list;
 
 
-CF_ADDTO(dynamic_attr, KRT_SOURCE { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_SOURCE); })
-CF_ADDTO(dynamic_attr, KRT_METRIC { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_METRIC); })
+dynamic_attr: KRT_SOURCE { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_SOURCE); } ;
+dynamic_attr: KRT_METRIC { $$ = f_new_dynamic_attr(EAF_TYPE_INT, T_INT, EA_KRT_METRIC); } ;
 
 CF_CODE
 

From c2fc4c10ac81b8815c1434aa2b0945aa937df4fa Mon Sep 17 00:00:00 2001
From: Maria Matejka <mq@jmq.cz>
Date: Mon, 23 Apr 2018 11:29:13 +0200
Subject: [PATCH 177/198] Doc: renamed progdoc files Doc -> progdoc to fix
 collision with doc/ folder on case-insensitive filesystems

---
 client/{Doc => progdoc}       | 0
 conf/{Doc => progdoc}         | 0
 doc/{Doc => progdoc}          | 0
 filter/{Doc => progdoc}       | 0
 lib/{Doc => progdoc}          | 0
 nest/{Doc => progdoc}         | 0
 Doc => progdoc                | 0
 proto/babel/{Doc => progdoc}  | 0
 proto/bfd/{Doc => progdoc}    | 0
 proto/bgp/{Doc => progdoc}    | 0
 proto/ospf/{Doc => progdoc}   | 0
 proto/pipe/{Doc => progdoc}   | 0
 proto/{Doc => progdoc}        | 0
 proto/radv/{Doc => progdoc}   | 0
 proto/rip/{Doc => progdoc}    | 0
 proto/static/{Doc => progdoc} | 0
 sysdep/{Doc => progdoc}       | 0
 sysdep/unix/{Doc => progdoc}  | 0
 tools/progdoc                 | 6 +++---
 19 files changed, 3 insertions(+), 3 deletions(-)
 rename client/{Doc => progdoc} (100%)
 rename conf/{Doc => progdoc} (100%)
 rename doc/{Doc => progdoc} (100%)
 rename filter/{Doc => progdoc} (100%)
 rename lib/{Doc => progdoc} (100%)
 rename nest/{Doc => progdoc} (100%)
 rename Doc => progdoc (100%)
 rename proto/babel/{Doc => progdoc} (100%)
 rename proto/bfd/{Doc => progdoc} (100%)
 rename proto/bgp/{Doc => progdoc} (100%)
 rename proto/ospf/{Doc => progdoc} (100%)
 rename proto/pipe/{Doc => progdoc} (100%)
 rename proto/{Doc => progdoc} (100%)
 rename proto/radv/{Doc => progdoc} (100%)
 rename proto/rip/{Doc => progdoc} (100%)
 rename proto/static/{Doc => progdoc} (100%)
 rename sysdep/{Doc => progdoc} (100%)
 rename sysdep/unix/{Doc => progdoc} (100%)

diff --git a/client/Doc b/client/progdoc
similarity index 100%
rename from client/Doc
rename to client/progdoc
diff --git a/conf/Doc b/conf/progdoc
similarity index 100%
rename from conf/Doc
rename to conf/progdoc
diff --git a/doc/Doc b/doc/progdoc
similarity index 100%
rename from doc/Doc
rename to doc/progdoc
diff --git a/filter/Doc b/filter/progdoc
similarity index 100%
rename from filter/Doc
rename to filter/progdoc
diff --git a/lib/Doc b/lib/progdoc
similarity index 100%
rename from lib/Doc
rename to lib/progdoc
diff --git a/nest/Doc b/nest/progdoc
similarity index 100%
rename from nest/Doc
rename to nest/progdoc
diff --git a/Doc b/progdoc
similarity index 100%
rename from Doc
rename to progdoc
diff --git a/proto/babel/Doc b/proto/babel/progdoc
similarity index 100%
rename from proto/babel/Doc
rename to proto/babel/progdoc
diff --git a/proto/bfd/Doc b/proto/bfd/progdoc
similarity index 100%
rename from proto/bfd/Doc
rename to proto/bfd/progdoc
diff --git a/proto/bgp/Doc b/proto/bgp/progdoc
similarity index 100%
rename from proto/bgp/Doc
rename to proto/bgp/progdoc
diff --git a/proto/ospf/Doc b/proto/ospf/progdoc
similarity index 100%
rename from proto/ospf/Doc
rename to proto/ospf/progdoc
diff --git a/proto/pipe/Doc b/proto/pipe/progdoc
similarity index 100%
rename from proto/pipe/Doc
rename to proto/pipe/progdoc
diff --git a/proto/Doc b/proto/progdoc
similarity index 100%
rename from proto/Doc
rename to proto/progdoc
diff --git a/proto/radv/Doc b/proto/radv/progdoc
similarity index 100%
rename from proto/radv/Doc
rename to proto/radv/progdoc
diff --git a/proto/rip/Doc b/proto/rip/progdoc
similarity index 100%
rename from proto/rip/Doc
rename to proto/rip/progdoc
diff --git a/proto/static/Doc b/proto/static/progdoc
similarity index 100%
rename from proto/static/Doc
rename to proto/static/progdoc
diff --git a/sysdep/Doc b/sysdep/progdoc
similarity index 100%
rename from sysdep/Doc
rename to sysdep/progdoc
diff --git a/sysdep/unix/Doc b/sysdep/unix/progdoc
similarity index 100%
rename from sysdep/unix/Doc
rename to sysdep/unix/progdoc
diff --git a/tools/progdoc b/tools/progdoc
index fc4024bf..9b27d14e 100755
--- a/tools/progdoc
+++ b/tools/progdoc
@@ -21,9 +21,9 @@ sub include {
 
 sub process {
   my $dir = shift @_;
-  print "$dir/Doc\n";
-  open(IN, "$dir/Doc") || die "Unable to read $dir/Doc";
-  push(@deps, "$dir/Doc");
+  print "$dir/progdoc\n";
+  open(IN, "$dir/progdoc") || die "Unable to read $dir/progdoc";
+  push(@deps, "$dir/progdoc");
   my @docfile = <IN>;
   close IN;
   foreach $_ (@docfile) {

From da16b33ab98602628d66ed4f434dc6fa76f338c6 Mon Sep 17 00:00:00 2001
From: Maria Matejka <mq@jmq.cz>
Date: Mon, 23 Apr 2018 15:54:20 +0200
Subject: [PATCH 178/198] Android: check for extra libs needed for build

---
 aclocal.m4   | 65 ++++++++++++++++++++++++++++++++++++++++++++++++++++
 configure.ac | 14 +++++++++++
 2 files changed, 79 insertions(+)

diff --git a/aclocal.m4 b/aclocal.m4
index 8160e539..dcf1bd8b 100644
--- a/aclocal.m4
+++ b/aclocal.m4
@@ -65,6 +65,71 @@ AC_DEFUN([BIRD_CHECK_MPLS_KERNEL],
   )
 ])
 
+AC_DEFUN([BIRD_CHECK_ANDROID_GLOB],
+[
+  AC_CACHE_CHECK(
+    [for glob.h],
+    [bird_cv_lib_glob],
+    AC_LINK_IFELSE([
+      AC_LANG_PROGRAM(
+        [
+	  #include <glob.h>
+	  #include <stdlib.h>
+	],
+        [ glob(NULL, 0, NULL, NULL); ]
+      )
+    ],
+    [bird_cv_lib_glob=yes],
+      [
+        bird_tmp_libs="$LIBS"
+        LIBS="$LIBS -landroid-glob"
+        AC_LINK_IFELSE([
+          AC_LANG_PROGRAM(
+            [
+	      #include <glob.h>
+	      #include <stdlib.h>
+	    ],
+            [ glob(NULL, 0, NULL, NULL); ]
+          )
+        ],
+        [bird_cv_lib_glob=-landroid-glob],
+        [bird_cv_lib_glob=no]
+        )
+        LIBS="$bird_tmp_libs"
+      ]
+    )
+  )
+])
+
+AC_DEFUN([BIRD_CHECK_ANDROID_LOG],
+[
+  AC_CACHE_CHECK(
+    [for syslog lib flags],
+    [bird_cv_lib_log],
+    AC_LINK_IFELSE([
+      AC_LANG_PROGRAM(
+        [ #include <sys/syslog.h> ],
+        [ syslog(0, ""); ]
+      )
+    ],
+    [bird_cv_lib_log=yes],
+      [
+        bird_tmp_libs="$LIBS"
+        LIBS="$LIBS -llog"
+        AC_LINK_IFELSE([
+          AC_LANG_PROGRAM(
+            [ #include <sys/syslog.h> ],
+            [ syslog(0, ""); ]
+          )
+        ],
+        [bird_cv_lib_log=-llog],
+        [bird_cv_lib_log=no]
+        )
+        LIBS="$bird_tmp_libs"
+      ]
+    )
+  )
+])
 
 AC_DEFUN([BIRD_CHECK_GCC_OPTION],
 [
diff --git a/configure.ac b/configure.ac
index f4de8f93..44184273 100644
--- a/configure.ac
+++ b/configure.ac
@@ -315,6 +315,20 @@ AC_C_BIGENDIAN(
   [AC_MSG_ERROR([Cannot determine CPU endianity.])]
 )
 
+BIRD_CHECK_ANDROID_GLOB
+if test "$bird_cv_lib_glob" = no ; then
+  AC_MSG_ERROR([glob.h not found.])
+elif test "$bird_cv_lib_glob" != yes ; then
+  LIBS="$LIBS $bird_cv_lib_glob"
+fi
+
+BIRD_CHECK_ANDROID_LOG
+if test "$bird_cv_lib_log" = no ; then
+  AC_MSG_ERROR([don't know how to link syslog.])
+elif test "$bird_cv_lib_log" != yes ; then
+  LIBS="$LIBS $bird_cv_lib_log"
+fi
+
 if test "$enable_debug" = yes ; then
   AC_DEFINE([DEBUGGING], [1], [Define to 1 if debugging is enabled])
   LDFLAGS="$LDFLAGS -rdynamic"

From 45f28d85818f79790968725a945063228989bae7 Mon Sep 17 00:00:00 2001
From: Maria Matejka <mq@jmq.cz>
Date: Mon, 23 Apr 2018 15:59:02 +0200
Subject: [PATCH 179/198] Autotools: updated config.guess and config.sub

Updated to version 63b4ce2e8c28aee6a32133e400436e4ca885215a
from git://git.savannah.gnu.org/config.git

Previous version was 93b5037172b15ad28952481933517f1ba93d125b
---
 tools/config.guess | 1233 +++++++++++++++++++++-----------------------
 tools/config.sub   |  583 ++++++++++++---------
 2 files changed, 925 insertions(+), 891 deletions(-)

diff --git a/tools/config.guess b/tools/config.guess
index da833146..256083a7 100755
--- a/tools/config.guess
+++ b/tools/config.guess
@@ -1,14 +1,12 @@
 #! /bin/sh
 # Attempt to guess a canonical system name.
-#   Copyright (C) 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
-#   2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008
-#   Free Software Foundation, Inc.
+#   Copyright 1992-2018 Free Software Foundation, Inc.
 
-timestamp='2009-04-27'
+timestamp='2018-03-08'
 
 # This file is free software; you can redistribute it and/or modify it
 # under the terms of the GNU General Public License as published by
-# the Free Software Foundation; either version 2 of the License, or
+# the Free Software Foundation; either version 3 of the License, or
 # (at your option) any later version.
 #
 # This program is distributed in the hope that it will be useful, but
@@ -17,26 +15,22 @@ timestamp='2009-04-27'
 # General Public License for more details.
 #
 # You should have received a copy of the GNU General Public License
-# along with this program; if not, write to the Free Software
-# Foundation, Inc., 51 Franklin Street - Fifth Floor, Boston, MA
-# 02110-1301, USA.
+# along with this program; if not, see <https://www.gnu.org/licenses/>.
 #
 # As a special exception to the GNU General Public License, if you
 # distribute this file as part of a program that contains a
 # configuration script generated by Autoconf, you may include it under
-# the same distribution terms that you use for the rest of that program.
-
-
-# Originally written by Per Bothner <per@bothner.com>.
-# Please send patches to <config-patches@gnu.org>.  Submit a context
-# diff and a properly formatted ChangeLog entry.
+# the same distribution terms that you use for the rest of that
+# program.  This Exception is an additional permission under section 7
+# of the GNU General Public License, version 3 ("GPLv3").
 #
-# This script attempts to guess a canonical system name similar to
-# config.sub.  If it succeeds, it prints the system name on stdout, and
-# exits with 0.  Otherwise, it exits with 1.
+# Originally written by Per Bothner; maintained since 2000 by Ben Elliston.
 #
-# The plan is that this can be called by configure scripts if you
-# don't specify an explicit build system type.
+# You can get the latest version of this script from:
+# https://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.guess
+#
+# Please send patches to <config-patches@gnu.org>.
+
 
 me=`echo "$0" | sed -e 's,.*/,,'`
 
@@ -45,7 +39,7 @@ Usage: $0 [OPTION]
 
 Output the configuration name of the system \`$me' is run on.
 
-Operation modes:
+Options:
   -h, --help         print this help, then exit
   -t, --time-stamp   print date of last modification, then exit
   -v, --version      print version number, then exit
@@ -56,8 +50,7 @@ version="\
 GNU config.guess ($timestamp)
 
 Originally written by Per Bothner.
-Copyright (C) 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001,
-2002, 2003, 2004, 2005, 2006, 2007, 2008 Free Software Foundation, Inc.
+Copyright 1992-2018 Free Software Foundation, Inc.
 
 This is free software; see the source for copying conditions.  There is NO
 warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE."
@@ -114,9 +107,9 @@ trap "rm -f \$tmpfiles 2>/dev/null; rmdir \$tmp 2>/dev/null; exit 1" 1 2 13 15 ;
 dummy=$tmp/dummy ;
 tmpfiles="$dummy.c $dummy.o $dummy.rel $dummy" ;
 case $CC_FOR_BUILD,$HOST_CC,$CC in
- ,,)    echo "int x;" > $dummy.c ;
+ ,,)    echo "int x;" > "$dummy.c" ;
 	for c in cc gcc c89 c99 ; do
-	  if ($c -c -o $dummy.o $dummy.c) >/dev/null 2>&1 ; then
+	  if ($c -c -o "$dummy.o" "$dummy.c") >/dev/null 2>&1 ; then
 	     CC_FOR_BUILD="$c"; break ;
 	  fi ;
 	done ;
@@ -139,12 +132,40 @@ UNAME_RELEASE=`(uname -r) 2>/dev/null` || UNAME_RELEASE=unknown
 UNAME_SYSTEM=`(uname -s) 2>/dev/null`  || UNAME_SYSTEM=unknown
 UNAME_VERSION=`(uname -v) 2>/dev/null` || UNAME_VERSION=unknown
 
+case "$UNAME_SYSTEM" in
+Linux|GNU|GNU/*)
+	# If the system lacks a compiler, then just pick glibc.
+	# We could probably try harder.
+	LIBC=gnu
+
+	eval "$set_cc_for_build"
+	cat <<-EOF > "$dummy.c"
+	#include <features.h>
+	#if defined(__UCLIBC__)
+	LIBC=uclibc
+	#elif defined(__dietlibc__)
+	LIBC=dietlibc
+	#else
+	LIBC=gnu
+	#endif
+	EOF
+	eval "`$CC_FOR_BUILD -E "$dummy.c" 2>/dev/null | grep '^LIBC' | sed 's, ,,g'`"
+
+	# If ldd exists, use it to detect musl libc.
+	if command -v ldd >/dev/null && \
+		ldd --version 2>&1 | grep -q ^musl
+	then
+	    LIBC=musl
+	fi
+	;;
+esac
+
 # Note: order is significant - the case branches are not exclusive.
 
-case "${UNAME_MACHINE}:${UNAME_SYSTEM}:${UNAME_RELEASE}:${UNAME_VERSION}" in
+case "$UNAME_MACHINE:$UNAME_SYSTEM:$UNAME_RELEASE:$UNAME_VERSION" in
     *:NetBSD:*:*)
 	# NetBSD (nbsd) targets should (where applicable) match one or
-	# more of the tupples: *-*-netbsdelf*, *-*-netbsdaout*,
+	# more of the tuples: *-*-netbsdelf*, *-*-netbsdaout*,
 	# *-*-netbsdecoff* and *-*-netbsd*.  For targets that recently
 	# switched to ELF, *-*-netbsd* would select the old
 	# object file format.  This provides both forward
@@ -154,23 +175,33 @@ case "${UNAME_MACHINE}:${UNAME_SYSTEM}:${UNAME_RELEASE}:${UNAME_VERSION}" in
 	# Note: NetBSD doesn't particularly care about the vendor
 	# portion of the name.  We always set it to "unknown".
 	sysctl="sysctl -n hw.machine_arch"
-	UNAME_MACHINE_ARCH=`(/sbin/$sysctl 2>/dev/null || \
-	    /usr/sbin/$sysctl 2>/dev/null || echo unknown)`
-	case "${UNAME_MACHINE_ARCH}" in
+	UNAME_MACHINE_ARCH=`(uname -p 2>/dev/null || \
+	    "/sbin/$sysctl" 2>/dev/null || \
+	    "/usr/sbin/$sysctl" 2>/dev/null || \
+	    echo unknown)`
+	case "$UNAME_MACHINE_ARCH" in
 	    armeb) machine=armeb-unknown ;;
 	    arm*) machine=arm-unknown ;;
 	    sh3el) machine=shl-unknown ;;
 	    sh3eb) machine=sh-unknown ;;
 	    sh5el) machine=sh5le-unknown ;;
-	    *) machine=${UNAME_MACHINE_ARCH}-unknown ;;
+	    earmv*)
+		arch=`echo "$UNAME_MACHINE_ARCH" | sed -e 's,^e\(armv[0-9]\).*$,\1,'`
+		endian=`echo "$UNAME_MACHINE_ARCH" | sed -ne 's,^.*\(eb\)$,\1,p'`
+		machine="${arch}${endian}"-unknown
+		;;
+	    *) machine="$UNAME_MACHINE_ARCH"-unknown ;;
 	esac
 	# The Operating System including object format, if it has switched
-	# to ELF recently, or will in the future.
-	case "${UNAME_MACHINE_ARCH}" in
+	# to ELF recently (or will in the future) and ABI.
+	case "$UNAME_MACHINE_ARCH" in
+	    earm*)
+		os=netbsdelf
+		;;
 	    arm*|i386|m68k|ns32k|sh3*|sparc|vax)
-		eval $set_cc_for_build
+		eval "$set_cc_for_build"
 		if echo __ELF__ | $CC_FOR_BUILD -E - 2>/dev/null \
-			| grep __ELF__ >/dev/null
+			| grep -q __ELF__
 		then
 		    # Once all utilities can be ECOFF (netbsdecoff) or a.out (netbsdaout).
 		    # Return netbsd for either.  FIX?
@@ -180,7 +211,14 @@ case "${UNAME_MACHINE}:${UNAME_SYSTEM}:${UNAME_RELEASE}:${UNAME_VERSION}" in
 		fi
 		;;
 	    *)
-	        os=netbsd
+		os=netbsd
+		;;
+	esac
+	# Determine ABI tags.
+	case "$UNAME_MACHINE_ARCH" in
+	    earm*)
+		expr='s/^earmv[0-9]/-eabi/;s/eb$//'
+		abi=`echo "$UNAME_MACHINE_ARCH" | sed -e "$expr"`
 		;;
 	esac
 	# The OS release
@@ -188,42 +226,62 @@ case "${UNAME_MACHINE}:${UNAME_SYSTEM}:${UNAME_RELEASE}:${UNAME_VERSION}" in
 	# thus, need a distinct triplet. However, they do not need
 	# kernel version information, so it can be replaced with a
 	# suitable tag, in the style of linux-gnu.
-	case "${UNAME_VERSION}" in
+	case "$UNAME_VERSION" in
 	    Debian*)
 		release='-gnu'
 		;;
 	    *)
-		release=`echo ${UNAME_RELEASE}|sed -e 's/[-_].*/\./'`
+		release=`echo "$UNAME_RELEASE" | sed -e 's/[-_].*//' | cut -d. -f1,2`
 		;;
 	esac
 	# Since CPU_TYPE-MANUFACTURER-KERNEL-OPERATING_SYSTEM:
 	# contains redundant information, the shorter form:
 	# CPU_TYPE-MANUFACTURER-OPERATING_SYSTEM is used.
-	echo "${machine}-${os}${release}"
+	echo "$machine-${os}${release}${abi}"
+	exit ;;
+    *:Bitrig:*:*)
+	UNAME_MACHINE_ARCH=`arch | sed 's/Bitrig.//'`
+	echo "$UNAME_MACHINE_ARCH"-unknown-bitrig"$UNAME_RELEASE"
 	exit ;;
     *:OpenBSD:*:*)
 	UNAME_MACHINE_ARCH=`arch | sed 's/OpenBSD.//'`
-	echo ${UNAME_MACHINE_ARCH}-unknown-openbsd${UNAME_RELEASE}
+	echo "$UNAME_MACHINE_ARCH"-unknown-openbsd"$UNAME_RELEASE"
+	exit ;;
+    *:LibertyBSD:*:*)
+	UNAME_MACHINE_ARCH=`arch | sed 's/^.*BSD\.//'`
+	echo "$UNAME_MACHINE_ARCH"-unknown-libertybsd"$UNAME_RELEASE"
+	exit ;;
+    *:MidnightBSD:*:*)
+	echo "$UNAME_MACHINE"-unknown-midnightbsd"$UNAME_RELEASE"
 	exit ;;
     *:ekkoBSD:*:*)
-	echo ${UNAME_MACHINE}-unknown-ekkobsd${UNAME_RELEASE}
+	echo "$UNAME_MACHINE"-unknown-ekkobsd"$UNAME_RELEASE"
 	exit ;;
     *:SolidBSD:*:*)
-	echo ${UNAME_MACHINE}-unknown-solidbsd${UNAME_RELEASE}
+	echo "$UNAME_MACHINE"-unknown-solidbsd"$UNAME_RELEASE"
 	exit ;;
     macppc:MirBSD:*:*)
-	echo powerpc-unknown-mirbsd${UNAME_RELEASE}
+	echo powerpc-unknown-mirbsd"$UNAME_RELEASE"
 	exit ;;
     *:MirBSD:*:*)
-	echo ${UNAME_MACHINE}-unknown-mirbsd${UNAME_RELEASE}
+	echo "$UNAME_MACHINE"-unknown-mirbsd"$UNAME_RELEASE"
 	exit ;;
+    *:Sortix:*:*)
+	echo "$UNAME_MACHINE"-unknown-sortix
+	exit ;;
+    *:Redox:*:*)
+	echo "$UNAME_MACHINE"-unknown-redox
+	exit ;;
+    mips:OSF1:*.*)
+        echo mips-dec-osf1
+        exit ;;
     alpha:OSF1:*:*)
 	case $UNAME_RELEASE in
 	*4.0)
 		UNAME_RELEASE=`/usr/sbin/sizer -v | awk '{print $3}'`
 		;;
 	*5.*)
-	        UNAME_RELEASE=`/usr/sbin/sizer -v | awk '{print $4}'`
+		UNAME_RELEASE=`/usr/sbin/sizer -v | awk '{print $4}'`
 		;;
 	esac
 	# According to Compaq, /usr/sbin/psrinfo has been available on
@@ -233,60 +291,54 @@ case "${UNAME_MACHINE}:${UNAME_SYSTEM}:${UNAME_RELEASE}:${UNAME_VERSION}" in
 	ALPHA_CPU_TYPE=`/usr/sbin/psrinfo -v | sed -n -e 's/^  The alpha \(.*\) processor.*$/\1/p' | head -n 1`
 	case "$ALPHA_CPU_TYPE" in
 	    "EV4 (21064)")
-		UNAME_MACHINE="alpha" ;;
+		UNAME_MACHINE=alpha ;;
 	    "EV4.5 (21064)")
-		UNAME_MACHINE="alpha" ;;
+		UNAME_MACHINE=alpha ;;
 	    "LCA4 (21066/21068)")
-		UNAME_MACHINE="alpha" ;;
+		UNAME_MACHINE=alpha ;;
 	    "EV5 (21164)")
-		UNAME_MACHINE="alphaev5" ;;
+		UNAME_MACHINE=alphaev5 ;;
 	    "EV5.6 (21164A)")
-		UNAME_MACHINE="alphaev56" ;;
+		UNAME_MACHINE=alphaev56 ;;
 	    "EV5.6 (21164PC)")
-		UNAME_MACHINE="alphapca56" ;;
+		UNAME_MACHINE=alphapca56 ;;
 	    "EV5.7 (21164PC)")
-		UNAME_MACHINE="alphapca57" ;;
+		UNAME_MACHINE=alphapca57 ;;
 	    "EV6 (21264)")
-		UNAME_MACHINE="alphaev6" ;;
+		UNAME_MACHINE=alphaev6 ;;
 	    "EV6.7 (21264A)")
-		UNAME_MACHINE="alphaev67" ;;
+		UNAME_MACHINE=alphaev67 ;;
 	    "EV6.8CB (21264C)")
-		UNAME_MACHINE="alphaev68" ;;
+		UNAME_MACHINE=alphaev68 ;;
 	    "EV6.8AL (21264B)")
-		UNAME_MACHINE="alphaev68" ;;
+		UNAME_MACHINE=alphaev68 ;;
 	    "EV6.8CX (21264D)")
-		UNAME_MACHINE="alphaev68" ;;
+		UNAME_MACHINE=alphaev68 ;;
 	    "EV6.9A (21264/EV69A)")
-		UNAME_MACHINE="alphaev69" ;;
+		UNAME_MACHINE=alphaev69 ;;
 	    "EV7 (21364)")
-		UNAME_MACHINE="alphaev7" ;;
+		UNAME_MACHINE=alphaev7 ;;
 	    "EV7.9 (21364A)")
-		UNAME_MACHINE="alphaev79" ;;
+		UNAME_MACHINE=alphaev79 ;;
 	esac
 	# A Pn.n version is a patched version.
 	# A Vn.n version is a released version.
 	# A Tn.n version is a released field test version.
 	# A Xn.n version is an unreleased experimental baselevel.
 	# 1.2 uses "1.2" for uname -r.
-	echo ${UNAME_MACHINE}-dec-osf`echo ${UNAME_RELEASE} | sed -e 's/^[PVTX]//' | tr 'ABCDEFGHIJKLMNOPQRSTUVWXYZ' 'abcdefghijklmnopqrstuvwxyz'`
-	exit ;;
-    Alpha\ *:Windows_NT*:*)
-	# How do we know it's Interix rather than the generic POSIX subsystem?
-	# Should we change UNAME_MACHINE based on the output of uname instead
-	# of the specific Alpha model?
-	echo alpha-pc-interix
-	exit ;;
-    21064:Windows_NT:50:3)
-	echo alpha-dec-winnt3.5
-	exit ;;
+	echo "$UNAME_MACHINE"-dec-osf"`echo "$UNAME_RELEASE" | sed -e 's/^[PVTX]//' | tr ABCDEFGHIJKLMNOPQRSTUVWXYZ abcdefghijklmnopqrstuvwxyz`"
+	# Reset EXIT trap before exiting to avoid spurious non-zero exit code.
+	exitcode=$?
+	trap '' 0
+	exit $exitcode ;;
     Amiga*:UNIX_System_V:4.0:*)
 	echo m68k-unknown-sysv4
 	exit ;;
     *:[Aa]miga[Oo][Ss]:*:*)
-	echo ${UNAME_MACHINE}-unknown-amigaos
+	echo "$UNAME_MACHINE"-unknown-amigaos
 	exit ;;
     *:[Mm]orph[Oo][Ss]:*:*)
-	echo ${UNAME_MACHINE}-unknown-morphos
+	echo "$UNAME_MACHINE"-unknown-morphos
 	exit ;;
     *:OS/390:*:*)
 	echo i370-ibm-openedition
@@ -295,12 +347,12 @@ case "${UNAME_MACHINE}:${UNAME_SYSTEM}:${UNAME_RELEASE}:${UNAME_VERSION}" in
 	echo s390-ibm-zvmoe
 	exit ;;
     *:OS400:*:*)
-        echo powerpc-ibm-os400
+	echo powerpc-ibm-os400
 	exit ;;
     arm:RISC*:1.[012]*:*|arm:riscix:1.[012]*:*)
-	echo arm-acorn-riscix${UNAME_RELEASE}
+	echo arm-acorn-riscix"$UNAME_RELEASE"
 	exit ;;
-    arm:riscos:*:*|arm:RISCOS:*:*)
+    arm*:riscos:*:*|arm*:RISCOS:*:*)
 	echo arm-unknown-riscos
 	exit ;;
     SR2?01:HI-UX/MPP:*:* | SR8000:HI-UX/MPP:*:*)
@@ -325,35 +377,38 @@ case "${UNAME_MACHINE}:${UNAME_SYSTEM}:${UNAME_RELEASE}:${UNAME_VERSION}" in
 	    sparc) echo sparc-icl-nx7; exit ;;
 	esac ;;
     s390x:SunOS:*:*)
-	echo ${UNAME_MACHINE}-ibm-solaris2`echo ${UNAME_RELEASE}|sed -e 's/[^.]*//'`
+	echo "$UNAME_MACHINE"-ibm-solaris2"`echo "$UNAME_RELEASE" | sed -e 's/[^.]*//'`"
 	exit ;;
     sun4H:SunOS:5.*:*)
-	echo sparc-hal-solaris2`echo ${UNAME_RELEASE}|sed -e 's/[^.]*//'`
+	echo sparc-hal-solaris2"`echo "$UNAME_RELEASE"|sed -e 's/[^.]*//'`"
 	exit ;;
     sun4*:SunOS:5.*:* | tadpole*:SunOS:5.*:*)
-	echo sparc-sun-solaris2`echo ${UNAME_RELEASE}|sed -e 's/[^.]*//'`
+	echo sparc-sun-solaris2"`echo "$UNAME_RELEASE" | sed -e 's/[^.]*//'`"
+	exit ;;
+    i86pc:AuroraUX:5.*:* | i86xen:AuroraUX:5.*:*)
+	echo i386-pc-auroraux"$UNAME_RELEASE"
 	exit ;;
     i86pc:SunOS:5.*:* | i86xen:SunOS:5.*:*)
-	eval $set_cc_for_build
-	SUN_ARCH="i386"
+	eval "$set_cc_for_build"
+	SUN_ARCH=i386
 	# If there is a compiler, see if it is configured for 64-bit objects.
 	# Note that the Sun cc does not turn __LP64__ into 1 like gcc does.
 	# This test works for both compilers.
-	if [ "$CC_FOR_BUILD" != 'no_compiler_found' ]; then
+	if [ "$CC_FOR_BUILD" != no_compiler_found ]; then
 	    if (echo '#ifdef __amd64'; echo IS_64BIT_ARCH; echo '#endif') | \
-		(CCOPTS= $CC_FOR_BUILD -E - 2>/dev/null) | \
+		(CCOPTS="" $CC_FOR_BUILD -E - 2>/dev/null) | \
 		grep IS_64BIT_ARCH >/dev/null
 	    then
-		SUN_ARCH="x86_64"
+		SUN_ARCH=x86_64
 	    fi
 	fi
-	echo ${SUN_ARCH}-pc-solaris2`echo ${UNAME_RELEASE}|sed -e 's/[^.]*//'`
+	echo "$SUN_ARCH"-pc-solaris2"`echo "$UNAME_RELEASE"|sed -e 's/[^.]*//'`"
 	exit ;;
     sun4*:SunOS:6*:*)
 	# According to config.sub, this is the proper way to canonicalize
 	# SunOS6.  Hard to guess exactly what SunOS6 will be like, but
 	# it's likely to be more like Solaris than SunOS4.
-	echo sparc-sun-solaris3`echo ${UNAME_RELEASE}|sed -e 's/[^.]*//'`
+	echo sparc-sun-solaris3"`echo "$UNAME_RELEASE"|sed -e 's/[^.]*//'`"
 	exit ;;
     sun4*:SunOS:*:*)
 	case "`/usr/bin/arch -k`" in
@@ -362,25 +417,25 @@ case "${UNAME_MACHINE}:${UNAME_SYSTEM}:${UNAME_RELEASE}:${UNAME_VERSION}" in
 		;;
 	esac
 	# Japanese Language versions have a version number like `4.1.3-JL'.
-	echo sparc-sun-sunos`echo ${UNAME_RELEASE}|sed -e 's/-/_/'`
+	echo sparc-sun-sunos"`echo "$UNAME_RELEASE"|sed -e 's/-/_/'`"
 	exit ;;
     sun3*:SunOS:*:*)
-	echo m68k-sun-sunos${UNAME_RELEASE}
+	echo m68k-sun-sunos"$UNAME_RELEASE"
 	exit ;;
     sun*:*:4.2BSD:*)
 	UNAME_RELEASE=`(sed 1q /etc/motd | awk '{print substr($5,1,3)}') 2>/dev/null`
-	test "x${UNAME_RELEASE}" = "x" && UNAME_RELEASE=3
+	test "x$UNAME_RELEASE" = x && UNAME_RELEASE=3
 	case "`/bin/arch`" in
 	    sun3)
-		echo m68k-sun-sunos${UNAME_RELEASE}
+		echo m68k-sun-sunos"$UNAME_RELEASE"
 		;;
 	    sun4)
-		echo sparc-sun-sunos${UNAME_RELEASE}
+		echo sparc-sun-sunos"$UNAME_RELEASE"
 		;;
 	esac
 	exit ;;
     aushp:SunOS:*:*)
-	echo sparc-auspex-sunos${UNAME_RELEASE}
+	echo sparc-auspex-sunos"$UNAME_RELEASE"
 	exit ;;
     # The situation for MiNT is a little confusing.  The machine name
     # can be virtually everything (everything which is not
@@ -391,44 +446,44 @@ case "${UNAME_MACHINE}:${UNAME_SYSTEM}:${UNAME_RELEASE}:${UNAME_VERSION}" in
     # MiNT.  But MiNT is downward compatible to TOS, so this should
     # be no problem.
     atarist[e]:*MiNT:*:* | atarist[e]:*mint:*:* | atarist[e]:*TOS:*:*)
-        echo m68k-atari-mint${UNAME_RELEASE}
+	echo m68k-atari-mint"$UNAME_RELEASE"
 	exit ;;
     atari*:*MiNT:*:* | atari*:*mint:*:* | atarist[e]:*TOS:*:*)
-	echo m68k-atari-mint${UNAME_RELEASE}
-        exit ;;
+	echo m68k-atari-mint"$UNAME_RELEASE"
+	exit ;;
     *falcon*:*MiNT:*:* | *falcon*:*mint:*:* | *falcon*:*TOS:*:*)
-        echo m68k-atari-mint${UNAME_RELEASE}
+	echo m68k-atari-mint"$UNAME_RELEASE"
 	exit ;;
     milan*:*MiNT:*:* | milan*:*mint:*:* | *milan*:*TOS:*:*)
-        echo m68k-milan-mint${UNAME_RELEASE}
-        exit ;;
+	echo m68k-milan-mint"$UNAME_RELEASE"
+	exit ;;
     hades*:*MiNT:*:* | hades*:*mint:*:* | *hades*:*TOS:*:*)
-        echo m68k-hades-mint${UNAME_RELEASE}
-        exit ;;
+	echo m68k-hades-mint"$UNAME_RELEASE"
+	exit ;;
     *:*MiNT:*:* | *:*mint:*:* | *:*TOS:*:*)
-        echo m68k-unknown-mint${UNAME_RELEASE}
-        exit ;;
+	echo m68k-unknown-mint"$UNAME_RELEASE"
+	exit ;;
     m68k:machten:*:*)
-	echo m68k-apple-machten${UNAME_RELEASE}
+	echo m68k-apple-machten"$UNAME_RELEASE"
 	exit ;;
     powerpc:machten:*:*)
-	echo powerpc-apple-machten${UNAME_RELEASE}
+	echo powerpc-apple-machten"$UNAME_RELEASE"
 	exit ;;
     RISC*:Mach:*:*)
 	echo mips-dec-mach_bsd4.3
 	exit ;;
     RISC*:ULTRIX:*:*)
-	echo mips-dec-ultrix${UNAME_RELEASE}
+	echo mips-dec-ultrix"$UNAME_RELEASE"
 	exit ;;
     VAX*:ULTRIX*:*:*)
-	echo vax-dec-ultrix${UNAME_RELEASE}
+	echo vax-dec-ultrix"$UNAME_RELEASE"
 	exit ;;
     2020:CLIX:*:* | 2430:CLIX:*:*)
-	echo clipper-intergraph-clix${UNAME_RELEASE}
+	echo clipper-intergraph-clix"$UNAME_RELEASE"
 	exit ;;
     mips:*:*:UMIPS | mips:*:*:RISCos)
-	eval $set_cc_for_build
-	sed 's/^	//' << EOF >$dummy.c
+	eval "$set_cc_for_build"
+	sed 's/^	//' << EOF > "$dummy.c"
 #ifdef __cplusplus
 #include <stdio.h>  /* for printf() prototype */
 	int main (int argc, char *argv[]) {
@@ -437,23 +492,23 @@ case "${UNAME_MACHINE}:${UNAME_SYSTEM}:${UNAME_RELEASE}:${UNAME_VERSION}" in
 #endif
 	#if defined (host_mips) && defined (MIPSEB)
 	#if defined (SYSTYPE_SYSV)
-	  printf ("mips-mips-riscos%ssysv\n", argv[1]); exit (0);
+	  printf ("mips-mips-riscos%ssysv\\n", argv[1]); exit (0);
 	#endif
 	#if defined (SYSTYPE_SVR4)
-	  printf ("mips-mips-riscos%ssvr4\n", argv[1]); exit (0);
+	  printf ("mips-mips-riscos%ssvr4\\n", argv[1]); exit (0);
 	#endif
 	#if defined (SYSTYPE_BSD43) || defined(SYSTYPE_BSD)
-	  printf ("mips-mips-riscos%sbsd\n", argv[1]); exit (0);
+	  printf ("mips-mips-riscos%sbsd\\n", argv[1]); exit (0);
 	#endif
 	#endif
 	  exit (-1);
 	}
 EOF
-	$CC_FOR_BUILD -o $dummy $dummy.c &&
-	  dummyarg=`echo "${UNAME_RELEASE}" | sed -n 's/\([0-9]*\).*/\1/p'` &&
-	  SYSTEM_NAME=`$dummy $dummyarg` &&
+	$CC_FOR_BUILD -o "$dummy" "$dummy.c" &&
+	  dummyarg=`echo "$UNAME_RELEASE" | sed -n 's/\([0-9]*\).*/\1/p'` &&
+	  SYSTEM_NAME=`"$dummy" "$dummyarg"` &&
 	    { echo "$SYSTEM_NAME"; exit; }
-	echo mips-mips-riscos${UNAME_RELEASE}
+	echo mips-mips-riscos"$UNAME_RELEASE"
 	exit ;;
     Motorola:PowerMAX_OS:*:*)
 	echo powerpc-motorola-powermax
@@ -477,21 +532,21 @@ EOF
 	echo m88k-motorola-sysv3
 	exit ;;
     AViiON:dgux:*:*)
-        # DG/UX returns AViiON for all architectures
-        UNAME_PROCESSOR=`/usr/bin/uname -p`
-	if [ $UNAME_PROCESSOR = mc88100 ] || [ $UNAME_PROCESSOR = mc88110 ]
+	# DG/UX returns AViiON for all architectures
+	UNAME_PROCESSOR=`/usr/bin/uname -p`
+	if [ "$UNAME_PROCESSOR" = mc88100 ] || [ "$UNAME_PROCESSOR" = mc88110 ]
 	then
-	    if [ ${TARGET_BINARY_INTERFACE}x = m88kdguxelfx ] || \
-	       [ ${TARGET_BINARY_INTERFACE}x = x ]
+	    if [ "$TARGET_BINARY_INTERFACE"x = m88kdguxelfx ] || \
+	       [ "$TARGET_BINARY_INTERFACE"x = x ]
 	    then
-		echo m88k-dg-dgux${UNAME_RELEASE}
+		echo m88k-dg-dgux"$UNAME_RELEASE"
 	    else
-		echo m88k-dg-dguxbcs${UNAME_RELEASE}
+		echo m88k-dg-dguxbcs"$UNAME_RELEASE"
 	    fi
 	else
-	    echo i586-dg-dgux${UNAME_RELEASE}
+	    echo i586-dg-dgux"$UNAME_RELEASE"
 	fi
- 	exit ;;
+	exit ;;
     M88*:DolphinOS:*:*)	# DolphinOS (SVR3)
 	echo m88k-dolphin-sysv3
 	exit ;;
@@ -506,7 +561,7 @@ EOF
 	echo m68k-tektronix-bsd
 	exit ;;
     *:IRIX*:*:*)
-	echo mips-sgi-irix`echo ${UNAME_RELEASE}|sed -e 's/-/_/g'`
+	echo mips-sgi-irix"`echo "$UNAME_RELEASE"|sed -e 's/-/_/g'`"
 	exit ;;
     ????????:AIX?:[12].1:2)   # AIX 2.2.1 or AIX 2.1.1 is RT/PC AIX.
 	echo romp-ibm-aix     # uname -m gives an 8 hex-code CPU id
@@ -518,14 +573,14 @@ EOF
 	if [ -x /usr/bin/oslevel ] ; then
 		IBM_REV=`/usr/bin/oslevel`
 	else
-		IBM_REV=${UNAME_VERSION}.${UNAME_RELEASE}
+		IBM_REV="$UNAME_VERSION.$UNAME_RELEASE"
 	fi
-	echo ${UNAME_MACHINE}-ibm-aix${IBM_REV}
+	echo "$UNAME_MACHINE"-ibm-aix"$IBM_REV"
 	exit ;;
     *:AIX:2:3)
 	if grep bos325 /usr/include/stdio.h >/dev/null 2>&1; then
-		eval $set_cc_for_build
-		sed 's/^		//' << EOF >$dummy.c
+		eval "$set_cc_for_build"
+		sed 's/^		//' << EOF > "$dummy.c"
 		#include <sys/systemcfg.h>
 
 		main()
@@ -536,7 +591,7 @@ EOF
 			exit(0);
 			}
 EOF
-		if $CC_FOR_BUILD -o $dummy $dummy.c && SYSTEM_NAME=`$dummy`
+		if $CC_FOR_BUILD -o "$dummy" "$dummy.c" && SYSTEM_NAME=`"$dummy"`
 		then
 			echo "$SYSTEM_NAME"
 		else
@@ -548,28 +603,29 @@ EOF
 		echo rs6000-ibm-aix3.2
 	fi
 	exit ;;
-    *:AIX:*:[456])
+    *:AIX:*:[4567])
 	IBM_CPU_ID=`/usr/sbin/lsdev -C -c processor -S available | sed 1q | awk '{ print $1 }'`
-	if /usr/sbin/lsattr -El ${IBM_CPU_ID} | grep ' POWER' >/dev/null 2>&1; then
+	if /usr/sbin/lsattr -El "$IBM_CPU_ID" | grep ' POWER' >/dev/null 2>&1; then
 		IBM_ARCH=rs6000
 	else
 		IBM_ARCH=powerpc
 	fi
-	if [ -x /usr/bin/oslevel ] ; then
-		IBM_REV=`/usr/bin/oslevel`
+	if [ -x /usr/bin/lslpp ] ; then
+		IBM_REV=`/usr/bin/lslpp -Lqc bos.rte.libc |
+			   awk -F: '{ print $3 }' | sed s/[0-9]*$/0/`
 	else
-		IBM_REV=${UNAME_VERSION}.${UNAME_RELEASE}
+		IBM_REV="$UNAME_VERSION.$UNAME_RELEASE"
 	fi
-	echo ${IBM_ARCH}-ibm-aix${IBM_REV}
+	echo "$IBM_ARCH"-ibm-aix"$IBM_REV"
 	exit ;;
     *:AIX:*:*)
 	echo rs6000-ibm-aix
 	exit ;;
-    ibmrt:4.4BSD:*|romp-ibm:BSD:*)
+    ibmrt:4.4BSD:*|romp-ibm:4.4BSD:*)
 	echo romp-ibm-bsd4.4
 	exit ;;
     ibmrt:*BSD:*|romp-ibm:BSD:*)            # covers RT/PC BSD and
-	echo romp-ibm-bsd${UNAME_RELEASE}   # 4.3 with uname added to
+	echo romp-ibm-bsd"$UNAME_RELEASE"   # 4.3 with uname added to
 	exit ;;                             # report: romp-ibm BSD 4.3
     *:BOSX:*:*)
 	echo rs6000-bull-bosx
@@ -584,67 +640,67 @@ EOF
 	echo m68k-hp-bsd4.4
 	exit ;;
     9000/[34678]??:HP-UX:*:*)
-	HPUX_REV=`echo ${UNAME_RELEASE}|sed -e 's/[^.]*.[0B]*//'`
-	case "${UNAME_MACHINE}" in
-	    9000/31? )            HP_ARCH=m68000 ;;
-	    9000/[34]?? )         HP_ARCH=m68k ;;
+	HPUX_REV=`echo "$UNAME_RELEASE"|sed -e 's/[^.]*.[0B]*//'`
+	case "$UNAME_MACHINE" in
+	    9000/31?)            HP_ARCH=m68000 ;;
+	    9000/[34]??)         HP_ARCH=m68k ;;
 	    9000/[678][0-9][0-9])
 		if [ -x /usr/bin/getconf ]; then
 		    sc_cpu_version=`/usr/bin/getconf SC_CPU_VERSION 2>/dev/null`
-                    sc_kernel_bits=`/usr/bin/getconf SC_KERNEL_BITS 2>/dev/null`
-                    case "${sc_cpu_version}" in
-                      523) HP_ARCH="hppa1.0" ;; # CPU_PA_RISC1_0
-                      528) HP_ARCH="hppa1.1" ;; # CPU_PA_RISC1_1
-                      532)                      # CPU_PA_RISC2_0
-                        case "${sc_kernel_bits}" in
-                          32) HP_ARCH="hppa2.0n" ;;
-                          64) HP_ARCH="hppa2.0w" ;;
-			  '') HP_ARCH="hppa2.0" ;;   # HP-UX 10.20
-                        esac ;;
-                    esac
+		    sc_kernel_bits=`/usr/bin/getconf SC_KERNEL_BITS 2>/dev/null`
+		    case "$sc_cpu_version" in
+		      523) HP_ARCH=hppa1.0 ;; # CPU_PA_RISC1_0
+		      528) HP_ARCH=hppa1.1 ;; # CPU_PA_RISC1_1
+		      532)                      # CPU_PA_RISC2_0
+			case "$sc_kernel_bits" in
+			  32) HP_ARCH=hppa2.0n ;;
+			  64) HP_ARCH=hppa2.0w ;;
+			  '') HP_ARCH=hppa2.0 ;;   # HP-UX 10.20
+			esac ;;
+		    esac
 		fi
-		if [ "${HP_ARCH}" = "" ]; then
-		    eval $set_cc_for_build
-		    sed 's/^              //' << EOF >$dummy.c
+		if [ "$HP_ARCH" = "" ]; then
+		    eval "$set_cc_for_build"
+		    sed 's/^		//' << EOF > "$dummy.c"
 
-              #define _HPUX_SOURCE
-              #include <stdlib.h>
-              #include <unistd.h>
+		#define _HPUX_SOURCE
+		#include <stdlib.h>
+		#include <unistd.h>
 
-              int main ()
-              {
-              #if defined(_SC_KERNEL_BITS)
-                  long bits = sysconf(_SC_KERNEL_BITS);
-              #endif
-                  long cpu  = sysconf (_SC_CPU_VERSION);
+		int main ()
+		{
+		#if defined(_SC_KERNEL_BITS)
+		    long bits = sysconf(_SC_KERNEL_BITS);
+		#endif
+		    long cpu  = sysconf (_SC_CPU_VERSION);
 
-                  switch (cpu)
-              	{
-              	case CPU_PA_RISC1_0: puts ("hppa1.0"); break;
-              	case CPU_PA_RISC1_1: puts ("hppa1.1"); break;
-              	case CPU_PA_RISC2_0:
-              #if defined(_SC_KERNEL_BITS)
-              	    switch (bits)
-              		{
-              		case 64: puts ("hppa2.0w"); break;
-              		case 32: puts ("hppa2.0n"); break;
-              		default: puts ("hppa2.0"); break;
-              		} break;
-              #else  /* !defined(_SC_KERNEL_BITS) */
-              	    puts ("hppa2.0"); break;
-              #endif
-              	default: puts ("hppa1.0"); break;
-              	}
-                  exit (0);
-              }
+		    switch (cpu)
+			{
+			case CPU_PA_RISC1_0: puts ("hppa1.0"); break;
+			case CPU_PA_RISC1_1: puts ("hppa1.1"); break;
+			case CPU_PA_RISC2_0:
+		#if defined(_SC_KERNEL_BITS)
+			    switch (bits)
+				{
+				case 64: puts ("hppa2.0w"); break;
+				case 32: puts ("hppa2.0n"); break;
+				default: puts ("hppa2.0"); break;
+				} break;
+		#else  /* !defined(_SC_KERNEL_BITS) */
+			    puts ("hppa2.0"); break;
+		#endif
+			default: puts ("hppa1.0"); break;
+			}
+		    exit (0);
+		}
 EOF
-		    (CCOPTS= $CC_FOR_BUILD -o $dummy $dummy.c 2>/dev/null) && HP_ARCH=`$dummy`
+		    (CCOPTS="" $CC_FOR_BUILD -o "$dummy" "$dummy.c" 2>/dev/null) && HP_ARCH=`"$dummy"`
 		    test -z "$HP_ARCH" && HP_ARCH=hppa
 		fi ;;
 	esac
-	if [ ${HP_ARCH} = "hppa2.0w" ]
+	if [ "$HP_ARCH" = hppa2.0w ]
 	then
-	    eval $set_cc_for_build
+	    eval "$set_cc_for_build"
 
 	    # hppa2.0w-hp-hpux* has a 64-bit kernel and a compiler generating
 	    # 32-bit code.  hppa64-hp-hpux* has the same kernel and a compiler
@@ -655,23 +711,23 @@ EOF
 	    # $ CC_FOR_BUILD="cc +DA2.0w" ./config.guess
 	    # => hppa64-hp-hpux11.23
 
-	    if echo __LP64__ | (CCOPTS= $CC_FOR_BUILD -E - 2>/dev/null) |
-		grep __LP64__ >/dev/null
+	    if echo __LP64__ | (CCOPTS="" $CC_FOR_BUILD -E - 2>/dev/null) |
+		grep -q __LP64__
 	    then
-		HP_ARCH="hppa2.0w"
+		HP_ARCH=hppa2.0w
 	    else
-		HP_ARCH="hppa64"
+		HP_ARCH=hppa64
 	    fi
 	fi
-	echo ${HP_ARCH}-hp-hpux${HPUX_REV}
+	echo "$HP_ARCH"-hp-hpux"$HPUX_REV"
 	exit ;;
     ia64:HP-UX:*:*)
-	HPUX_REV=`echo ${UNAME_RELEASE}|sed -e 's/[^.]*.[0B]*//'`
-	echo ia64-hp-hpux${HPUX_REV}
+	HPUX_REV=`echo "$UNAME_RELEASE"|sed -e 's/[^.]*.[0B]*//'`
+	echo ia64-hp-hpux"$HPUX_REV"
 	exit ;;
     3050*:HI-UX:*:*)
-	eval $set_cc_for_build
-	sed 's/^	//' << EOF >$dummy.c
+	eval "$set_cc_for_build"
+	sed 's/^	//' << EOF > "$dummy.c"
 	#include <unistd.h>
 	int
 	main ()
@@ -696,11 +752,11 @@ EOF
 	  exit (0);
 	}
 EOF
-	$CC_FOR_BUILD -o $dummy $dummy.c && SYSTEM_NAME=`$dummy` &&
+	$CC_FOR_BUILD -o "$dummy" "$dummy.c" && SYSTEM_NAME=`"$dummy"` &&
 		{ echo "$SYSTEM_NAME"; exit; }
 	echo unknown-hitachi-hiuxwe2
 	exit ;;
-    9000/7??:4.3bsd:*:* | 9000/8?[79]:4.3bsd:*:* )
+    9000/7??:4.3bsd:*:* | 9000/8?[79]:4.3bsd:*:*)
 	echo hppa1.1-hp-bsd
 	exit ;;
     9000/8??:4.3bsd:*:*)
@@ -709,7 +765,7 @@ EOF
     *9??*:MPE/iX:*:* | *3000*:MPE/iX:*:*)
 	echo hppa1.0-hp-mpeix
 	exit ;;
-    hp7??:OSF1:*:* | hp8?[79]:OSF1:*:* )
+    hp7??:OSF1:*:* | hp8?[79]:OSF1:*:*)
 	echo hppa1.1-hp-osf
 	exit ;;
     hp8??:OSF1:*:*)
@@ -717,9 +773,9 @@ EOF
 	exit ;;
     i*86:OSF1:*:*)
 	if [ -x /usr/sbin/sysversion ] ; then
-	    echo ${UNAME_MACHINE}-unknown-osf1mk
+	    echo "$UNAME_MACHINE"-unknown-osf1mk
 	else
-	    echo ${UNAME_MACHINE}-unknown-osf1
+	    echo "$UNAME_MACHINE"-unknown-osf1
 	fi
 	exit ;;
     parisc*:Lites*:*:*)
@@ -727,215 +783,126 @@ EOF
 	exit ;;
     C1*:ConvexOS:*:* | convex:ConvexOS:C1*:*)
 	echo c1-convex-bsd
-        exit ;;
+	exit ;;
     C2*:ConvexOS:*:* | convex:ConvexOS:C2*:*)
 	if getsysinfo -f scalar_acc
 	then echo c32-convex-bsd
 	else echo c2-convex-bsd
 	fi
-        exit ;;
+	exit ;;
     C34*:ConvexOS:*:* | convex:ConvexOS:C34*:*)
 	echo c34-convex-bsd
-        exit ;;
+	exit ;;
     C38*:ConvexOS:*:* | convex:ConvexOS:C38*:*)
 	echo c38-convex-bsd
-        exit ;;
+	exit ;;
     C4*:ConvexOS:*:* | convex:ConvexOS:C4*:*)
 	echo c4-convex-bsd
-        exit ;;
+	exit ;;
     CRAY*Y-MP:*:*:*)
-	echo ymp-cray-unicos${UNAME_RELEASE} | sed -e 's/\.[^.]*$/.X/'
+	echo ymp-cray-unicos"$UNAME_RELEASE" | sed -e 's/\.[^.]*$/.X/'
 	exit ;;
     CRAY*[A-Z]90:*:*:*)
-	echo ${UNAME_MACHINE}-cray-unicos${UNAME_RELEASE} \
+	echo "$UNAME_MACHINE"-cray-unicos"$UNAME_RELEASE" \
 	| sed -e 's/CRAY.*\([A-Z]90\)/\1/' \
 	      -e y/ABCDEFGHIJKLMNOPQRSTUVWXYZ/abcdefghijklmnopqrstuvwxyz/ \
 	      -e 's/\.[^.]*$/.X/'
 	exit ;;
     CRAY*TS:*:*:*)
-	echo t90-cray-unicos${UNAME_RELEASE} | sed -e 's/\.[^.]*$/.X/'
+	echo t90-cray-unicos"$UNAME_RELEASE" | sed -e 's/\.[^.]*$/.X/'
 	exit ;;
     CRAY*T3E:*:*:*)
-	echo alphaev5-cray-unicosmk${UNAME_RELEASE} | sed -e 's/\.[^.]*$/.X/'
+	echo alphaev5-cray-unicosmk"$UNAME_RELEASE" | sed -e 's/\.[^.]*$/.X/'
 	exit ;;
     CRAY*SV1:*:*:*)
-	echo sv1-cray-unicos${UNAME_RELEASE} | sed -e 's/\.[^.]*$/.X/'
+	echo sv1-cray-unicos"$UNAME_RELEASE" | sed -e 's/\.[^.]*$/.X/'
 	exit ;;
     *:UNICOS/mp:*:*)
-	echo craynv-cray-unicosmp${UNAME_RELEASE} | sed -e 's/\.[^.]*$/.X/'
+	echo craynv-cray-unicosmp"$UNAME_RELEASE" | sed -e 's/\.[^.]*$/.X/'
 	exit ;;
     F30[01]:UNIX_System_V:*:* | F700:UNIX_System_V:*:*)
-	FUJITSU_PROC=`uname -m | tr 'ABCDEFGHIJKLMNOPQRSTUVWXYZ' 'abcdefghijklmnopqrstuvwxyz'`
-        FUJITSU_SYS=`uname -p | tr 'ABCDEFGHIJKLMNOPQRSTUVWXYZ' 'abcdefghijklmnopqrstuvwxyz' | sed -e 's/\///'`
-        FUJITSU_REL=`echo ${UNAME_RELEASE} | sed -e 's/ /_/'`
-        echo "${FUJITSU_PROC}-fujitsu-${FUJITSU_SYS}${FUJITSU_REL}"
-        exit ;;
+	FUJITSU_PROC=`uname -m | tr ABCDEFGHIJKLMNOPQRSTUVWXYZ abcdefghijklmnopqrstuvwxyz`
+	FUJITSU_SYS=`uname -p | tr ABCDEFGHIJKLMNOPQRSTUVWXYZ abcdefghijklmnopqrstuvwxyz | sed -e 's/\///'`
+	FUJITSU_REL=`echo "$UNAME_RELEASE" | sed -e 's/ /_/'`
+	echo "${FUJITSU_PROC}-fujitsu-${FUJITSU_SYS}${FUJITSU_REL}"
+	exit ;;
     5000:UNIX_System_V:4.*:*)
-        FUJITSU_SYS=`uname -p | tr 'ABCDEFGHIJKLMNOPQRSTUVWXYZ' 'abcdefghijklmnopqrstuvwxyz' | sed -e 's/\///'`
-        FUJITSU_REL=`echo ${UNAME_RELEASE} | tr 'ABCDEFGHIJKLMNOPQRSTUVWXYZ' 'abcdefghijklmnopqrstuvwxyz' | sed -e 's/ /_/'`
-        echo "sparc-fujitsu-${FUJITSU_SYS}${FUJITSU_REL}"
+	FUJITSU_SYS=`uname -p | tr ABCDEFGHIJKLMNOPQRSTUVWXYZ abcdefghijklmnopqrstuvwxyz | sed -e 's/\///'`
+	FUJITSU_REL=`echo "$UNAME_RELEASE" | tr ABCDEFGHIJKLMNOPQRSTUVWXYZ abcdefghijklmnopqrstuvwxyz | sed -e 's/ /_/'`
+	echo "sparc-fujitsu-${FUJITSU_SYS}${FUJITSU_REL}"
 	exit ;;
     i*86:BSD/386:*:* | i*86:BSD/OS:*:* | *:Ascend\ Embedded/OS:*:*)
-	echo ${UNAME_MACHINE}-pc-bsdi${UNAME_RELEASE}
+	echo "$UNAME_MACHINE"-pc-bsdi"$UNAME_RELEASE"
 	exit ;;
     sparc*:BSD/OS:*:*)
-	echo sparc-unknown-bsdi${UNAME_RELEASE}
+	echo sparc-unknown-bsdi"$UNAME_RELEASE"
 	exit ;;
     *:BSD/OS:*:*)
-	echo ${UNAME_MACHINE}-unknown-bsdi${UNAME_RELEASE}
+	echo "$UNAME_MACHINE"-unknown-bsdi"$UNAME_RELEASE"
 	exit ;;
     *:FreeBSD:*:*)
-	case ${UNAME_MACHINE} in
-	    pc98)
-		echo i386-unknown-freebsd`echo ${UNAME_RELEASE}|sed -e 's/[-(].*//'` ;;
+	UNAME_PROCESSOR=`/usr/bin/uname -p`
+	case "$UNAME_PROCESSOR" in
 	    amd64)
-		echo x86_64-unknown-freebsd`echo ${UNAME_RELEASE}|sed -e 's/[-(].*//'` ;;
-	    *)
-		echo ${UNAME_MACHINE}-unknown-freebsd`echo ${UNAME_RELEASE}|sed -e 's/[-(].*//'` ;;
+		UNAME_PROCESSOR=x86_64 ;;
+	    i386)
+		UNAME_PROCESSOR=i586 ;;
 	esac
+	echo "$UNAME_PROCESSOR"-unknown-freebsd"`echo "$UNAME_RELEASE"|sed -e 's/[-(].*//'`"
 	exit ;;
     i*:CYGWIN*:*)
-	echo ${UNAME_MACHINE}-pc-cygwin
+	echo "$UNAME_MACHINE"-pc-cygwin
+	exit ;;
+    *:MINGW64*:*)
+	echo "$UNAME_MACHINE"-pc-mingw64
 	exit ;;
     *:MINGW*:*)
-	echo ${UNAME_MACHINE}-pc-mingw32
+	echo "$UNAME_MACHINE"-pc-mingw32
 	exit ;;
-    i*:windows32*:*)
-    	# uname -m includes "-pc" on this system.
-    	echo ${UNAME_MACHINE}-mingw32
+    *:MSYS*:*)
+	echo "$UNAME_MACHINE"-pc-msys
 	exit ;;
     i*:PW*:*)
-	echo ${UNAME_MACHINE}-pc-pw32
+	echo "$UNAME_MACHINE"-pc-pw32
 	exit ;;
-    *:Interix*:[3456]*)
-    	case ${UNAME_MACHINE} in
+    *:Interix*:*)
+	case "$UNAME_MACHINE" in
 	    x86)
-		echo i586-pc-interix${UNAME_RELEASE}
+		echo i586-pc-interix"$UNAME_RELEASE"
 		exit ;;
-	    EM64T | authenticamd | genuineintel)
-		echo x86_64-unknown-interix${UNAME_RELEASE}
+	    authenticamd | genuineintel | EM64T)
+		echo x86_64-unknown-interix"$UNAME_RELEASE"
 		exit ;;
 	    IA64)
-		echo ia64-unknown-interix${UNAME_RELEASE}
+		echo ia64-unknown-interix"$UNAME_RELEASE"
 		exit ;;
 	esac ;;
-    [345]86:Windows_95:* | [345]86:Windows_98:* | [345]86:Windows_NT:*)
-	echo i${UNAME_MACHINE}-pc-mks
-	exit ;;
-    i*:Windows_NT*:* | Pentium*:Windows_NT*:*)
-	# How do we know it's Interix rather than the generic POSIX subsystem?
-	# It also conflicts with pre-2.0 versions of AT&T UWIN. Should we
-	# UNAME_MACHINE based on the output of uname instead of i386?
-	echo i586-pc-interix
-	exit ;;
     i*:UWIN*:*)
-	echo ${UNAME_MACHINE}-pc-uwin
+	echo "$UNAME_MACHINE"-pc-uwin
 	exit ;;
     amd64:CYGWIN*:*:* | x86_64:CYGWIN*:*:*)
 	echo x86_64-unknown-cygwin
 	exit ;;
-    p*:CYGWIN*:*)
-	echo powerpcle-unknown-cygwin
-	exit ;;
     prep*:SunOS:5.*:*)
-	echo powerpcle-unknown-solaris2`echo ${UNAME_RELEASE}|sed -e 's/[^.]*//'`
+	echo powerpcle-unknown-solaris2"`echo "$UNAME_RELEASE"|sed -e 's/[^.]*//'`"
 	exit ;;
     *:GNU:*:*)
 	# the GNU system
-	echo `echo ${UNAME_MACHINE}|sed -e 's,[-/].*$,,'`-unknown-gnu`echo ${UNAME_RELEASE}|sed -e 's,/.*$,,'`
+	echo "`echo "$UNAME_MACHINE"|sed -e 's,[-/].*$,,'`-unknown-$LIBC`echo "$UNAME_RELEASE"|sed -e 's,/.*$,,'`"
 	exit ;;
     *:GNU/*:*:*)
 	# other systems with GNU libc and userland
-	echo ${UNAME_MACHINE}-unknown-`echo ${UNAME_SYSTEM} | sed 's,^[^/]*/,,' | tr '[A-Z]' '[a-z]'``echo ${UNAME_RELEASE}|sed -e 's/[-(].*//'`-gnu
+	echo "$UNAME_MACHINE-unknown-`echo "$UNAME_SYSTEM" | sed 's,^[^/]*/,,' | tr "[:upper:]" "[:lower:]"``echo "$UNAME_RELEASE"|sed -e 's/[-(].*//'`-$LIBC"
 	exit ;;
     i*86:Minix:*:*)
-	echo ${UNAME_MACHINE}-pc-minix
+	echo "$UNAME_MACHINE"-pc-minix
 	exit ;;
-    arm*:Linux:*:*)
-	eval $set_cc_for_build
-	if echo __ARM_EABI__ | $CC_FOR_BUILD -E - 2>/dev/null \
-	    | grep -q __ARM_EABI__
-	then
-	    echo ${UNAME_MACHINE}-unknown-linux-gnu
-	else
-	    echo ${UNAME_MACHINE}-unknown-linux-gnueabi
-	fi
+    aarch64:Linux:*:*)
+	echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"
 	exit ;;
-    avr32*:Linux:*:*)
-	echo ${UNAME_MACHINE}-unknown-linux-gnu
-	exit ;;
-    cris:Linux:*:*)
-	echo cris-axis-linux-gnu
-	exit ;;
-    crisv32:Linux:*:*)
-	echo crisv32-axis-linux-gnu
-	exit ;;
-    frv:Linux:*:*)
-    	echo frv-unknown-linux-gnu
-	exit ;;
-    ia64:Linux:*:*)
-	echo ${UNAME_MACHINE}-unknown-linux-gnu
-	exit ;;
-    m32r*:Linux:*:*)
-	echo ${UNAME_MACHINE}-unknown-linux-gnu
-	exit ;;
-    m68*:Linux:*:*)
-	echo ${UNAME_MACHINE}-unknown-linux-gnu
-	exit ;;
-    mips:Linux:*:*)
-	eval $set_cc_for_build
-	sed 's/^	//' << EOF >$dummy.c
-	#undef CPU
-	#undef mips
-	#undef mipsel
-	#if defined(__MIPSEL__) || defined(__MIPSEL) || defined(_MIPSEL) || defined(MIPSEL)
-	CPU=mipsel
-	#else
-	#if defined(__MIPSEB__) || defined(__MIPSEB) || defined(_MIPSEB) || defined(MIPSEB)
-	CPU=mips
-	#else
-	CPU=
-	#endif
-	#endif
-EOF
-	eval "`$CC_FOR_BUILD -E $dummy.c 2>/dev/null | sed -n '
-	    /^CPU/{
-		s: ::g
-		p
-	    }'`"
-	test x"${CPU}" != x && { echo "${CPU}-unknown-linux-gnu"; exit; }
-	;;
-    mips64:Linux:*:*)
-	eval $set_cc_for_build
-	sed 's/^	//' << EOF >$dummy.c
-	#undef CPU
-	#undef mips64
-	#undef mips64el
-	#if defined(__MIPSEL__) || defined(__MIPSEL) || defined(_MIPSEL) || defined(MIPSEL)
-	CPU=mips64el
-	#else
-	#if defined(__MIPSEB__) || defined(__MIPSEB) || defined(_MIPSEB) || defined(MIPSEB)
-	CPU=mips64
-	#else
-	CPU=
-	#endif
-	#endif
-EOF
-	eval "`$CC_FOR_BUILD -E $dummy.c 2>/dev/null | sed -n '
-	    /^CPU/{
-		s: ::g
-		p
-	    }'`"
-	test x"${CPU}" != x && { echo "${CPU}-unknown-linux-gnu"; exit; }
-	;;
-    or32:Linux:*:*)
-	echo or32-unknown-linux-gnu
-	exit ;;
-    ppc:Linux:*:*)
-	echo powerpc-unknown-linux-gnu
-	exit ;;
-    ppc64:Linux:*:*)
-	echo powerpc64-unknown-linux-gnu
+    aarch64_be:Linux:*:*)
+	UNAME_MACHINE=aarch64_be
+	echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"
 	exit ;;
     alpha:Linux:*:*)
 	case `sed -n '/^cpu model/s/^.*: \(.*\)/\1/p' < /proc/cpuinfo` in
@@ -946,106 +913,144 @@ EOF
 	  EV6)   UNAME_MACHINE=alphaev6 ;;
 	  EV67)  UNAME_MACHINE=alphaev67 ;;
 	  EV68*) UNAME_MACHINE=alphaev68 ;;
-        esac
-	objdump --private-headers /bin/sh | grep ld.so.1 >/dev/null
-	if test "$?" = 0 ; then LIBC="libc1" ; else LIBC="" ; fi
-	echo ${UNAME_MACHINE}-unknown-linux-gnu${LIBC}
+	esac
+	objdump --private-headers /bin/sh | grep -q ld.so.1
+	if test "$?" = 0 ; then LIBC=gnulibc1 ; fi
+	echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"
+	exit ;;
+    arc:Linux:*:* | arceb:Linux:*:*)
+	echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"
+	exit ;;
+    arm*:Linux:*:*)
+	eval "$set_cc_for_build"
+	if echo __ARM_EABI__ | $CC_FOR_BUILD -E - 2>/dev/null \
+	    | grep -q __ARM_EABI__
+	then
+	    echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"
+	else
+	    if echo __ARM_PCS_VFP | $CC_FOR_BUILD -E - 2>/dev/null \
+		| grep -q __ARM_PCS_VFP
+	    then
+		echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"eabi
+	    else
+		echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"eabihf
+	    fi
+	fi
+	exit ;;
+    avr32*:Linux:*:*)
+	echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"
+	exit ;;
+    cris:Linux:*:*)
+	echo "$UNAME_MACHINE"-axis-linux-"$LIBC"
+	exit ;;
+    crisv32:Linux:*:*)
+	echo "$UNAME_MACHINE"-axis-linux-"$LIBC"
+	exit ;;
+    e2k:Linux:*:*)
+	echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"
+	exit ;;
+    frv:Linux:*:*)
+	echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"
+	exit ;;
+    hexagon:Linux:*:*)
+	echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"
+	exit ;;
+    i*86:Linux:*:*)
+	echo "$UNAME_MACHINE"-pc-linux-"$LIBC"
+	exit ;;
+    ia64:Linux:*:*)
+	echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"
+	exit ;;
+    k1om:Linux:*:*)
+	echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"
+	exit ;;
+    m32r*:Linux:*:*)
+	echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"
+	exit ;;
+    m68*:Linux:*:*)
+	echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"
+	exit ;;
+    mips:Linux:*:* | mips64:Linux:*:*)
+	eval "$set_cc_for_build"
+	sed 's/^	//' << EOF > "$dummy.c"
+	#undef CPU
+	#undef ${UNAME_MACHINE}
+	#undef ${UNAME_MACHINE}el
+	#if defined(__MIPSEL__) || defined(__MIPSEL) || defined(_MIPSEL) || defined(MIPSEL)
+	CPU=${UNAME_MACHINE}el
+	#else
+	#if defined(__MIPSEB__) || defined(__MIPSEB) || defined(_MIPSEB) || defined(MIPSEB)
+	CPU=${UNAME_MACHINE}
+	#else
+	CPU=
+	#endif
+	#endif
+EOF
+	eval "`$CC_FOR_BUILD -E "$dummy.c" 2>/dev/null | grep '^CPU'`"
+	test "x$CPU" != x && { echo "$CPU-unknown-linux-$LIBC"; exit; }
+	;;
+    mips64el:Linux:*:*)
+	echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"
+	exit ;;
+    openrisc*:Linux:*:*)
+	echo or1k-unknown-linux-"$LIBC"
+	exit ;;
+    or32:Linux:*:* | or1k*:Linux:*:*)
+	echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"
 	exit ;;
     padre:Linux:*:*)
-	echo sparc-unknown-linux-gnu
+	echo sparc-unknown-linux-"$LIBC"
+	exit ;;
+    parisc64:Linux:*:* | hppa64:Linux:*:*)
+	echo hppa64-unknown-linux-"$LIBC"
 	exit ;;
     parisc:Linux:*:* | hppa:Linux:*:*)
 	# Look for CPU level
 	case `grep '^cpu[^a-z]*:' /proc/cpuinfo 2>/dev/null | cut -d' ' -f2` in
-	  PA7*) echo hppa1.1-unknown-linux-gnu ;;
-	  PA8*) echo hppa2.0-unknown-linux-gnu ;;
-	  *)    echo hppa-unknown-linux-gnu ;;
+	  PA7*) echo hppa1.1-unknown-linux-"$LIBC" ;;
+	  PA8*) echo hppa2.0-unknown-linux-"$LIBC" ;;
+	  *)    echo hppa-unknown-linux-"$LIBC" ;;
 	esac
 	exit ;;
-    parisc64:Linux:*:* | hppa64:Linux:*:*)
-	echo hppa64-unknown-linux-gnu
+    ppc64:Linux:*:*)
+	echo powerpc64-unknown-linux-"$LIBC"
+	exit ;;
+    ppc:Linux:*:*)
+	echo powerpc-unknown-linux-"$LIBC"
+	exit ;;
+    ppc64le:Linux:*:*)
+	echo powerpc64le-unknown-linux-"$LIBC"
+	exit ;;
+    ppcle:Linux:*:*)
+	echo powerpcle-unknown-linux-"$LIBC"
+	exit ;;
+    riscv32:Linux:*:* | riscv64:Linux:*:*)
+	echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"
 	exit ;;
     s390:Linux:*:* | s390x:Linux:*:*)
-	echo ${UNAME_MACHINE}-ibm-linux
+	echo "$UNAME_MACHINE"-ibm-linux-"$LIBC"
 	exit ;;
     sh64*:Linux:*:*)
-    	echo ${UNAME_MACHINE}-unknown-linux-gnu
+	echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"
 	exit ;;
     sh*:Linux:*:*)
-	echo ${UNAME_MACHINE}-unknown-linux-gnu
+	echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"
 	exit ;;
     sparc:Linux:*:* | sparc64:Linux:*:*)
-	echo ${UNAME_MACHINE}-unknown-linux-gnu
+	echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"
+	exit ;;
+    tile*:Linux:*:*)
+	echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"
 	exit ;;
     vax:Linux:*:*)
-	echo ${UNAME_MACHINE}-dec-linux-gnu
+	echo "$UNAME_MACHINE"-dec-linux-"$LIBC"
 	exit ;;
     x86_64:Linux:*:*)
-	echo x86_64-unknown-linux-gnu
+	echo "$UNAME_MACHINE"-pc-linux-"$LIBC"
 	exit ;;
     xtensa*:Linux:*:*)
-    	echo ${UNAME_MACHINE}-unknown-linux-gnu
+	echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"
 	exit ;;
-    i*86:Linux:*:*)
-	# The BFD linker knows what the default object file format is, so
-	# first see if it will tell us. cd to the root directory to prevent
-	# problems with other programs or directories called `ld' in the path.
-	# Set LC_ALL=C to ensure ld outputs messages in English.
-	ld_supported_targets=`cd /; LC_ALL=C ld --help 2>&1 \
-			 | sed -ne '/supported targets:/!d
-				    s/[ 	][ 	]*/ /g
-				    s/.*supported targets: *//
-				    s/ .*//
-				    p'`
-        case "$ld_supported_targets" in
-	  elf32-i386)
-		TENTATIVE="${UNAME_MACHINE}-pc-linux-gnu"
-		;;
-	  a.out-i386-linux)
-		echo "${UNAME_MACHINE}-pc-linux-gnuaout"
-		exit ;;
-	  "")
-		# Either a pre-BFD a.out linker (linux-gnuoldld) or
-		# one that does not give us useful --help.
-		echo "${UNAME_MACHINE}-pc-linux-gnuoldld"
-		exit ;;
-	esac
-	# Determine whether the default compiler is a.out or elf
-	eval $set_cc_for_build
-	sed 's/^	//' << EOF >$dummy.c
-	#include <features.h>
-	#ifdef __ELF__
-	# ifdef __GLIBC__
-	#  if __GLIBC__ >= 2
-	LIBC=gnu
-	#  else
-	LIBC=gnulibc1
-	#  endif
-	# else
-	LIBC=gnulibc1
-	# endif
-	#else
-	#if defined(__INTEL_COMPILER) || defined(__PGI) || defined(__SUNPRO_C) || defined(__SUNPRO_CC)
-	LIBC=gnu
-	#else
-	LIBC=gnuaout
-	#endif
-	#endif
-	#ifdef __dietlibc__
-	LIBC=dietlibc
-	#endif
-EOF
-	eval "`$CC_FOR_BUILD -E $dummy.c 2>/dev/null | sed -n '
-	    /^LIBC/{
-		s: ::g
-		p
-	    }'`"
-	test x"${LIBC}" != x && {
-		echo "${UNAME_MACHINE}-pc-linux-${LIBC}"
-		exit
-	}
-	test x"${TENTATIVE}" != x && { echo "${TENTATIVE}"; exit; }
-	;;
     i*86:DYNIX/ptx:4*:*)
 	# ptx 4.0 does uname -s correctly, with DYNIX/ptx in there.
 	# earlier versions are messed up and put the nodename in both
@@ -1053,54 +1058,54 @@ EOF
 	echo i386-sequent-sysv4
 	exit ;;
     i*86:UNIX_SV:4.2MP:2.*)
-        # Unixware is an offshoot of SVR4, but it has its own version
-        # number series starting with 2...
-        # I am not positive that other SVR4 systems won't match this,
+	# Unixware is an offshoot of SVR4, but it has its own version
+	# number series starting with 2...
+	# I am not positive that other SVR4 systems won't match this,
 	# I just have to hope.  -- rms.
-        # Use sysv4.2uw... so that sysv4* matches it.
-	echo ${UNAME_MACHINE}-pc-sysv4.2uw${UNAME_VERSION}
+	# Use sysv4.2uw... so that sysv4* matches it.
+	echo "$UNAME_MACHINE"-pc-sysv4.2uw"$UNAME_VERSION"
 	exit ;;
     i*86:OS/2:*:*)
 	# If we were able to find `uname', then EMX Unix compatibility
 	# is probably installed.
-	echo ${UNAME_MACHINE}-pc-os2-emx
+	echo "$UNAME_MACHINE"-pc-os2-emx
 	exit ;;
     i*86:XTS-300:*:STOP)
-	echo ${UNAME_MACHINE}-unknown-stop
+	echo "$UNAME_MACHINE"-unknown-stop
 	exit ;;
     i*86:atheos:*:*)
-	echo ${UNAME_MACHINE}-unknown-atheos
+	echo "$UNAME_MACHINE"-unknown-atheos
 	exit ;;
     i*86:syllable:*:*)
-	echo ${UNAME_MACHINE}-pc-syllable
+	echo "$UNAME_MACHINE"-pc-syllable
 	exit ;;
-    i*86:LynxOS:2.*:* | i*86:LynxOS:3.[01]*:* | i*86:LynxOS:4.0*:*)
-	echo i386-unknown-lynxos${UNAME_RELEASE}
+    i*86:LynxOS:2.*:* | i*86:LynxOS:3.[01]*:* | i*86:LynxOS:4.[02]*:*)
+	echo i386-unknown-lynxos"$UNAME_RELEASE"
 	exit ;;
     i*86:*DOS:*:*)
-	echo ${UNAME_MACHINE}-pc-msdosdjgpp
+	echo "$UNAME_MACHINE"-pc-msdosdjgpp
 	exit ;;
-    i*86:*:4.*:* | i*86:SYSTEM_V:4.*:*)
-	UNAME_REL=`echo ${UNAME_RELEASE} | sed 's/\/MP$//'`
+    i*86:*:4.*:*)
+	UNAME_REL=`echo "$UNAME_RELEASE" | sed 's/\/MP$//'`
 	if grep Novell /usr/include/link.h >/dev/null 2>/dev/null; then
-		echo ${UNAME_MACHINE}-univel-sysv${UNAME_REL}
+		echo "$UNAME_MACHINE"-univel-sysv"$UNAME_REL"
 	else
-		echo ${UNAME_MACHINE}-pc-sysv${UNAME_REL}
+		echo "$UNAME_MACHINE"-pc-sysv"$UNAME_REL"
 	fi
 	exit ;;
     i*86:*:5:[678]*)
-    	# UnixWare 7.x, OpenUNIX and OpenServer 6.
+	# UnixWare 7.x, OpenUNIX and OpenServer 6.
 	case `/bin/uname -X | grep "^Machine"` in
 	    *486*)	     UNAME_MACHINE=i486 ;;
 	    *Pentium)	     UNAME_MACHINE=i586 ;;
 	    *Pent*|*Celeron) UNAME_MACHINE=i686 ;;
 	esac
-	echo ${UNAME_MACHINE}-unknown-sysv${UNAME_RELEASE}${UNAME_SYSTEM}${UNAME_VERSION}
+	echo "$UNAME_MACHINE-unknown-sysv${UNAME_RELEASE}${UNAME_SYSTEM}{$UNAME_VERSION}"
 	exit ;;
     i*86:*:3.2:*)
 	if test -f /usr/options/cb.name; then
 		UNAME_REL=`sed -n 's/.*Version //p' </usr/options/cb.name`
-		echo ${UNAME_MACHINE}-pc-isc$UNAME_REL
+		echo "$UNAME_MACHINE"-pc-isc"$UNAME_REL"
 	elif /bin/uname -X 2>/dev/null >/dev/null ; then
 		UNAME_REL=`(/bin/uname -X|grep Release|sed -e 's/.*= //')`
 		(/bin/uname -X|grep i80486 >/dev/null) && UNAME_MACHINE=i486
@@ -1110,20 +1115,20 @@ EOF
 			&& UNAME_MACHINE=i686
 		(/bin/uname -X|grep '^Machine.*Pentium Pro' >/dev/null) \
 			&& UNAME_MACHINE=i686
-		echo ${UNAME_MACHINE}-pc-sco$UNAME_REL
+		echo "$UNAME_MACHINE"-pc-sco"$UNAME_REL"
 	else
-		echo ${UNAME_MACHINE}-pc-sysv32
+		echo "$UNAME_MACHINE"-pc-sysv32
 	fi
 	exit ;;
     pc:*:*:*)
 	# Left here for compatibility:
-        # uname -m prints for DJGPP always 'pc', but it prints nothing about
-        # the processor, so we play safe by assuming i586.
+	# uname -m prints for DJGPP always 'pc', but it prints nothing about
+	# the processor, so we play safe by assuming i586.
 	# Note: whatever this is, it MUST be the same as what config.sub
-	# prints for the "djgpp" host, or else GDB configury will decide that
+	# prints for the "djgpp" host, or else GDB configure will decide that
 	# this is a cross-build.
 	echo i586-pc-msdosdjgpp
-        exit ;;
+	exit ;;
     Intel:Mach:3*:*)
 	echo i386-pc-mach3
 	exit ;;
@@ -1132,9 +1137,9 @@ EOF
 	exit ;;
     i860:*:4.*:*) # i860-SVR4
 	if grep Stardent /usr/include/sys/uadmin.h >/dev/null 2>&1 ; then
-	  echo i860-stardent-sysv${UNAME_RELEASE} # Stardent Vistra i860-SVR4
+	  echo i860-stardent-sysv"$UNAME_RELEASE" # Stardent Vistra i860-SVR4
 	else # Add other i860-SVR4 vendors below as they are discovered.
-	  echo i860-unknown-sysv${UNAME_RELEASE}  # Unknown i860-SVR4
+	  echo i860-unknown-sysv"$UNAME_RELEASE"  # Unknown i860-SVR4
 	fi
 	exit ;;
     mini*:CTIX:SYS*5:*)
@@ -1154,39 +1159,39 @@ EOF
 	test -r /etc/.relid \
 	&& OS_REL=.`sed -n 's/[^ ]* [^ ]* \([0-9][0-9]\).*/\1/p' < /etc/.relid`
 	/bin/uname -p 2>/dev/null | grep 86 >/dev/null \
-	  && { echo i486-ncr-sysv4.3${OS_REL}; exit; }
+	  && { echo i486-ncr-sysv4.3"$OS_REL"; exit; }
 	/bin/uname -p 2>/dev/null | /bin/grep entium >/dev/null \
-	  && { echo i586-ncr-sysv4.3${OS_REL}; exit; } ;;
+	  && { echo i586-ncr-sysv4.3"$OS_REL"; exit; } ;;
     3[34]??:*:4.0:* | 3[34]??,*:*:4.0:*)
-        /bin/uname -p 2>/dev/null | grep 86 >/dev/null \
-          && { echo i486-ncr-sysv4; exit; } ;;
+	/bin/uname -p 2>/dev/null | grep 86 >/dev/null \
+	  && { echo i486-ncr-sysv4; exit; } ;;
     NCR*:*:4.2:* | MPRAS*:*:4.2:*)
 	OS_REL='.3'
 	test -r /etc/.relid \
 	    && OS_REL=.`sed -n 's/[^ ]* [^ ]* \([0-9][0-9]\).*/\1/p' < /etc/.relid`
 	/bin/uname -p 2>/dev/null | grep 86 >/dev/null \
-	    && { echo i486-ncr-sysv4.3${OS_REL}; exit; }
+	    && { echo i486-ncr-sysv4.3"$OS_REL"; exit; }
 	/bin/uname -p 2>/dev/null | /bin/grep entium >/dev/null \
-	    && { echo i586-ncr-sysv4.3${OS_REL}; exit; }
+	    && { echo i586-ncr-sysv4.3"$OS_REL"; exit; }
 	/bin/uname -p 2>/dev/null | /bin/grep pteron >/dev/null \
-	    && { echo i586-ncr-sysv4.3${OS_REL}; exit; } ;;
+	    && { echo i586-ncr-sysv4.3"$OS_REL"; exit; } ;;
     m68*:LynxOS:2.*:* | m68*:LynxOS:3.0*:*)
-	echo m68k-unknown-lynxos${UNAME_RELEASE}
+	echo m68k-unknown-lynxos"$UNAME_RELEASE"
 	exit ;;
     mc68030:UNIX_System_V:4.*:*)
 	echo m68k-atari-sysv4
 	exit ;;
     TSUNAMI:LynxOS:2.*:*)
-	echo sparc-unknown-lynxos${UNAME_RELEASE}
+	echo sparc-unknown-lynxos"$UNAME_RELEASE"
 	exit ;;
     rs6000:LynxOS:2.*:*)
-	echo rs6000-unknown-lynxos${UNAME_RELEASE}
+	echo rs6000-unknown-lynxos"$UNAME_RELEASE"
 	exit ;;
-    PowerPC:LynxOS:2.*:* | PowerPC:LynxOS:3.[01]*:* | PowerPC:LynxOS:4.0*:*)
-	echo powerpc-unknown-lynxos${UNAME_RELEASE}
+    PowerPC:LynxOS:2.*:* | PowerPC:LynxOS:3.[01]*:* | PowerPC:LynxOS:4.[02]*:*)
+	echo powerpc-unknown-lynxos"$UNAME_RELEASE"
 	exit ;;
     SM[BE]S:UNIX_SV:*:*)
-	echo mips-dde-sysv${UNAME_RELEASE}
+	echo mips-dde-sysv"$UNAME_RELEASE"
 	exit ;;
     RM*:ReliantUNIX-*:*:*)
 	echo mips-sni-sysv4
@@ -1197,15 +1202,15 @@ EOF
     *:SINIX-*:*:*)
 	if uname -p 2>/dev/null >/dev/null ; then
 		UNAME_MACHINE=`(uname -p) 2>/dev/null`
-		echo ${UNAME_MACHINE}-sni-sysv4
+		echo "$UNAME_MACHINE"-sni-sysv4
 	else
 		echo ns32k-sni-sysv
 	fi
 	exit ;;
-    PENTIUM:*:4.0*:*) # Unisys `ClearPath HMP IX 4000' SVR4/MP effort
-                      # says <Richard.M.Bartel@ccMail.Census.GOV>
-        echo i586-unisys-sysv4
-        exit ;;
+    PENTIUM:*:4.0*:*)	# Unisys `ClearPath HMP IX 4000' SVR4/MP effort
+			# says <Richard.M.Bartel@ccMail.Census.GOV>
+	echo i586-unisys-sysv4
+	exit ;;
     *:UNIX_System_V:4*:FTX*)
 	# From Gerald Hewes <hewes@openmarket.com>.
 	# How about differentiating between stratus architectures? -djm
@@ -1217,25 +1222,25 @@ EOF
 	exit ;;
     i*86:VOS:*:*)
 	# From Paul.Green@stratus.com.
-	echo ${UNAME_MACHINE}-stratus-vos
+	echo "$UNAME_MACHINE"-stratus-vos
 	exit ;;
     *:VOS:*:*)
 	# From Paul.Green@stratus.com.
 	echo hppa1.1-stratus-vos
 	exit ;;
     mc68*:A/UX:*:*)
-	echo m68k-apple-aux${UNAME_RELEASE}
+	echo m68k-apple-aux"$UNAME_RELEASE"
 	exit ;;
     news*:NEWS-OS:6*:*)
 	echo mips-sony-newsos6
 	exit ;;
     R[34]000:*System_V*:*:* | R4000:UNIX_SYSV:*:* | R*000:UNIX_SV:*:*)
 	if [ -d /usr/nec ]; then
-	        echo mips-nec-sysv${UNAME_RELEASE}
+		echo mips-nec-sysv"$UNAME_RELEASE"
 	else
-	        echo mips-unknown-sysv${UNAME_RELEASE}
+		echo mips-unknown-sysv"$UNAME_RELEASE"
 	fi
-        exit ;;
+	exit ;;
     BeBox:BeOS:*:*)	# BeOS running on hardware made by Be, PPC only.
 	echo powerpc-be-beos
 	exit ;;
@@ -1248,53 +1253,97 @@ EOF
     BePC:Haiku:*:*)	# Haiku running on Intel PC compatible.
 	echo i586-pc-haiku
 	exit ;;
+    x86_64:Haiku:*:*)
+	echo x86_64-unknown-haiku
+	exit ;;
     SX-4:SUPER-UX:*:*)
-	echo sx4-nec-superux${UNAME_RELEASE}
+	echo sx4-nec-superux"$UNAME_RELEASE"
 	exit ;;
     SX-5:SUPER-UX:*:*)
-	echo sx5-nec-superux${UNAME_RELEASE}
+	echo sx5-nec-superux"$UNAME_RELEASE"
 	exit ;;
     SX-6:SUPER-UX:*:*)
-	echo sx6-nec-superux${UNAME_RELEASE}
+	echo sx6-nec-superux"$UNAME_RELEASE"
 	exit ;;
     SX-7:SUPER-UX:*:*)
-	echo sx7-nec-superux${UNAME_RELEASE}
+	echo sx7-nec-superux"$UNAME_RELEASE"
 	exit ;;
     SX-8:SUPER-UX:*:*)
-	echo sx8-nec-superux${UNAME_RELEASE}
+	echo sx8-nec-superux"$UNAME_RELEASE"
 	exit ;;
     SX-8R:SUPER-UX:*:*)
-	echo sx8r-nec-superux${UNAME_RELEASE}
+	echo sx8r-nec-superux"$UNAME_RELEASE"
+	exit ;;
+    SX-ACE:SUPER-UX:*:*)
+	echo sxace-nec-superux"$UNAME_RELEASE"
 	exit ;;
     Power*:Rhapsody:*:*)
-	echo powerpc-apple-rhapsody${UNAME_RELEASE}
+	echo powerpc-apple-rhapsody"$UNAME_RELEASE"
 	exit ;;
     *:Rhapsody:*:*)
-	echo ${UNAME_MACHINE}-apple-rhapsody${UNAME_RELEASE}
+	echo "$UNAME_MACHINE"-apple-rhapsody"$UNAME_RELEASE"
 	exit ;;
     *:Darwin:*:*)
 	UNAME_PROCESSOR=`uname -p` || UNAME_PROCESSOR=unknown
-	case $UNAME_PROCESSOR in
-	    unknown) UNAME_PROCESSOR=powerpc ;;
-	esac
-	echo ${UNAME_PROCESSOR}-apple-darwin${UNAME_RELEASE}
+	eval "$set_cc_for_build"
+	if test "$UNAME_PROCESSOR" = unknown ; then
+	    UNAME_PROCESSOR=powerpc
+	fi
+	if test "`echo "$UNAME_RELEASE" | sed -e 's/\..*//'`" -le 10 ; then
+	    if [ "$CC_FOR_BUILD" != no_compiler_found ]; then
+		if (echo '#ifdef __LP64__'; echo IS_64BIT_ARCH; echo '#endif') | \
+		       (CCOPTS="" $CC_FOR_BUILD -E - 2>/dev/null) | \
+		       grep IS_64BIT_ARCH >/dev/null
+		then
+		    case $UNAME_PROCESSOR in
+			i386) UNAME_PROCESSOR=x86_64 ;;
+			powerpc) UNAME_PROCESSOR=powerpc64 ;;
+		    esac
+		fi
+		# On 10.4-10.6 one might compile for PowerPC via gcc -arch ppc
+		if (echo '#ifdef __POWERPC__'; echo IS_PPC; echo '#endif') | \
+		       (CCOPTS="" $CC_FOR_BUILD -E - 2>/dev/null) | \
+		       grep IS_PPC >/dev/null
+		then
+		    UNAME_PROCESSOR=powerpc
+		fi
+	    fi
+	elif test "$UNAME_PROCESSOR" = i386 ; then
+	    # Avoid executing cc on OS X 10.9, as it ships with a stub
+	    # that puts up a graphical alert prompting to install
+	    # developer tools.  Any system running Mac OS X 10.7 or
+	    # later (Darwin 11 and later) is required to have a 64-bit
+	    # processor. This is not true of the ARM version of Darwin
+	    # that Apple uses in portable devices.
+	    UNAME_PROCESSOR=x86_64
+	fi
+	echo "$UNAME_PROCESSOR"-apple-darwin"$UNAME_RELEASE"
 	exit ;;
     *:procnto*:*:* | *:QNX:[0123456789]*:*)
 	UNAME_PROCESSOR=`uname -p`
-	if test "$UNAME_PROCESSOR" = "x86"; then
+	if test "$UNAME_PROCESSOR" = x86; then
 		UNAME_PROCESSOR=i386
 		UNAME_MACHINE=pc
 	fi
-	echo ${UNAME_PROCESSOR}-${UNAME_MACHINE}-nto-qnx${UNAME_RELEASE}
+	echo "$UNAME_PROCESSOR"-"$UNAME_MACHINE"-nto-qnx"$UNAME_RELEASE"
 	exit ;;
     *:QNX:*:4*)
 	echo i386-pc-qnx
 	exit ;;
-    NSE-?:NONSTOP_KERNEL:*:*)
-	echo nse-tandem-nsk${UNAME_RELEASE}
+    NEO-*:NONSTOP_KERNEL:*:*)
+	echo neo-tandem-nsk"$UNAME_RELEASE"
 	exit ;;
-    NSR-?:NONSTOP_KERNEL:*:*)
-	echo nsr-tandem-nsk${UNAME_RELEASE}
+    NSE-*:NONSTOP_KERNEL:*:*)
+	echo nse-tandem-nsk"$UNAME_RELEASE"
+	exit ;;
+    NSR-*:NONSTOP_KERNEL:*:*)
+	echo nsr-tandem-nsk"$UNAME_RELEASE"
+	exit ;;
+    NSV-*:NONSTOP_KERNEL:*:*)
+	echo nsv-tandem-nsk"$UNAME_RELEASE"
+	exit ;;
+    NSX-*:NONSTOP_KERNEL:*:*)
+	echo nsx-tandem-nsk"$UNAME_RELEASE"
 	exit ;;
     *:NonStop-UX:*:*)
 	echo mips-compaq-nonstopux
@@ -1303,18 +1352,18 @@ EOF
 	echo bs2000-siemens-sysv
 	exit ;;
     DS/*:UNIX_System_V:*:*)
-	echo ${UNAME_MACHINE}-${UNAME_SYSTEM}-${UNAME_RELEASE}
+	echo "$UNAME_MACHINE"-"$UNAME_SYSTEM"-"$UNAME_RELEASE"
 	exit ;;
     *:Plan9:*:*)
 	# "uname -m" is not consistent, so use $cputype instead. 386
 	# is converted to i386 for consistency with other x86
 	# operating systems.
-	if test "$cputype" = "386"; then
+	if test "$cputype" = 386; then
 	    UNAME_MACHINE=i386
 	else
 	    UNAME_MACHINE="$cputype"
 	fi
-	echo ${UNAME_MACHINE}-unknown-plan9
+	echo "$UNAME_MACHINE"-unknown-plan9
 	exit ;;
     *:TOPS-10:*:*)
 	echo pdp10-unknown-tops10
@@ -1335,14 +1384,14 @@ EOF
 	echo pdp10-unknown-its
 	exit ;;
     SEI:*:*:SEIUX)
-        echo mips-sei-seiux${UNAME_RELEASE}
+	echo mips-sei-seiux"$UNAME_RELEASE"
 	exit ;;
     *:DragonFly:*:*)
-	echo ${UNAME_MACHINE}-unknown-dragonfly`echo ${UNAME_RELEASE}|sed -e 's/[-(].*//'`
+	echo "$UNAME_MACHINE"-unknown-dragonfly"`echo "$UNAME_RELEASE"|sed -e 's/[-(].*//'`"
 	exit ;;
     *:*VMS:*:*)
-    	UNAME_MACHINE=`(uname -p) 2>/dev/null`
-	case "${UNAME_MACHINE}" in
+	UNAME_MACHINE=`(uname -p) 2>/dev/null`
+	case "$UNAME_MACHINE" in
 	    A*) echo alpha-dec-vms ; exit ;;
 	    I*) echo ia64-dec-vms ; exit ;;
 	    V*) echo vax-dec-vms ; exit ;;
@@ -1351,182 +1400,48 @@ EOF
 	echo i386-pc-xenix
 	exit ;;
     i*86:skyos:*:*)
-	echo ${UNAME_MACHINE}-pc-skyos`echo ${UNAME_RELEASE}` | sed -e 's/ .*$//'
+	echo "$UNAME_MACHINE"-pc-skyos"`echo "$UNAME_RELEASE" | sed -e 's/ .*$//'`"
 	exit ;;
     i*86:rdos:*:*)
-	echo ${UNAME_MACHINE}-pc-rdos
+	echo "$UNAME_MACHINE"-pc-rdos
 	exit ;;
     i*86:AROS:*:*)
-	echo ${UNAME_MACHINE}-pc-aros
+	echo "$UNAME_MACHINE"-pc-aros
+	exit ;;
+    x86_64:VMkernel:*:*)
+	echo "$UNAME_MACHINE"-unknown-esx
+	exit ;;
+    amd64:Isilon\ OneFS:*:*)
+	echo x86_64-unknown-onefs
 	exit ;;
 esac
 
-#echo '(No uname command or uname output not recognized.)' 1>&2
-#echo "${UNAME_MACHINE}:${UNAME_SYSTEM}:${UNAME_RELEASE}:${UNAME_VERSION}" 1>&2
+echo "$0: unable to guess system type" >&2
 
-eval $set_cc_for_build
-cat >$dummy.c <<EOF
-#ifdef _SEQUENT_
-# include <sys/types.h>
-# include <sys/utsname.h>
-#endif
-main ()
-{
-#if defined (sony)
-#if defined (MIPSEB)
-  /* BFD wants "bsd" instead of "newsos".  Perhaps BFD should be changed,
-     I don't know....  */
-  printf ("mips-sony-bsd\n"); exit (0);
-#else
-#include <sys/param.h>
-  printf ("m68k-sony-newsos%s\n",
-#ifdef NEWSOS4
-          "4"
-#else
-	  ""
-#endif
-         ); exit (0);
-#endif
-#endif
+case "$UNAME_MACHINE:$UNAME_SYSTEM" in
+    mips:Linux | mips64:Linux)
+	# If we got here on MIPS GNU/Linux, output extra information.
+	cat >&2 <<EOF
 
-#if defined (__arm) && defined (__acorn) && defined (__unix)
-  printf ("arm-acorn-riscix\n"); exit (0);
-#endif
-
-#if defined (hp300) && !defined (hpux)
-  printf ("m68k-hp-bsd\n"); exit (0);
-#endif
-
-#if defined (NeXT)
-#if !defined (__ARCHITECTURE__)
-#define __ARCHITECTURE__ "m68k"
-#endif
-  int version;
-  version=`(hostinfo | sed -n 's/.*NeXT Mach \([0-9]*\).*/\1/p') 2>/dev/null`;
-  if (version < 4)
-    printf ("%s-next-nextstep%d\n", __ARCHITECTURE__, version);
-  else
-    printf ("%s-next-openstep%d\n", __ARCHITECTURE__, version);
-  exit (0);
-#endif
-
-#if defined (MULTIMAX) || defined (n16)
-#if defined (UMAXV)
-  printf ("ns32k-encore-sysv\n"); exit (0);
-#else
-#if defined (CMU)
-  printf ("ns32k-encore-mach\n"); exit (0);
-#else
-  printf ("ns32k-encore-bsd\n"); exit (0);
-#endif
-#endif
-#endif
-
-#if defined (__386BSD__)
-  printf ("i386-pc-bsd\n"); exit (0);
-#endif
-
-#if defined (sequent)
-#if defined (i386)
-  printf ("i386-sequent-dynix\n"); exit (0);
-#endif
-#if defined (ns32000)
-  printf ("ns32k-sequent-dynix\n"); exit (0);
-#endif
-#endif
-
-#if defined (_SEQUENT_)
-    struct utsname un;
-
-    uname(&un);
-
-    if (strncmp(un.version, "V2", 2) == 0) {
-	printf ("i386-sequent-ptx2\n"); exit (0);
-    }
-    if (strncmp(un.version, "V1", 2) == 0) { /* XXX is V1 correct? */
-	printf ("i386-sequent-ptx1\n"); exit (0);
-    }
-    printf ("i386-sequent-ptx\n"); exit (0);
-
-#endif
-
-#if defined (vax)
-# if !defined (ultrix)
-#  include <sys/param.h>
-#  if defined (BSD)
-#   if BSD == 43
-      printf ("vax-dec-bsd4.3\n"); exit (0);
-#   else
-#    if BSD == 199006
-      printf ("vax-dec-bsd4.3reno\n"); exit (0);
-#    else
-      printf ("vax-dec-bsd\n"); exit (0);
-#    endif
-#   endif
-#  else
-    printf ("vax-dec-bsd\n"); exit (0);
-#  endif
-# else
-    printf ("vax-dec-ultrix\n"); exit (0);
-# endif
-#endif
-
-#if defined (alliant) && defined (i860)
-  printf ("i860-alliant-bsd\n"); exit (0);
-#endif
-
-  exit (1);
-}
+NOTE: MIPS GNU/Linux systems require a C compiler to fully recognize
+the system type. Please install a C compiler and try again.
 EOF
-
-$CC_FOR_BUILD -o $dummy $dummy.c 2>/dev/null && SYSTEM_NAME=`$dummy` &&
-	{ echo "$SYSTEM_NAME"; exit; }
-
-# Apollos put the system type in the environment.
-
-test -d /usr/apollo && { echo ${ISP}-apollo-${SYSTYPE}; exit; }
-
-# Convex versions that predate uname can use getsysinfo(1)
-
-if [ -x /usr/convex/getsysinfo ]
-then
-    case `getsysinfo -f cpu_type` in
-    c1*)
-	echo c1-convex-bsd
-	exit ;;
-    c2*)
-	if getsysinfo -f scalar_acc
-	then echo c32-convex-bsd
-	else echo c2-convex-bsd
-	fi
-	exit ;;
-    c34*)
-	echo c34-convex-bsd
-	exit ;;
-    c38*)
-	echo c38-convex-bsd
-	exit ;;
-    c4*)
-	echo c4-convex-bsd
-	exit ;;
-    esac
-fi
+	;;
+esac
 
 cat >&2 <<EOF
-$0: unable to guess system type
 
-This script, last modified $timestamp, has failed to recognize
-the operating system you are using. It is advised that you
-download the most up to date version of the config scripts from
+This script (version $timestamp), has failed to recognize the
+operating system you are using. If your script is old, overwrite *all*
+copies of config.guess and config.sub with the latest versions from:
 
-  http://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.guess;hb=HEAD
+  https://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.guess
 and
-  http://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.sub;hb=HEAD
+  https://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.sub
 
-If the version you run ($0) is already up to date, please
-send the following data and any information you think might be
-pertinent to <config-patches@gnu.org> in order to provide the needed
-information to handle your system.
+If $0 has already been updated, send the following data and any
+information you think might be pertinent to config-patches@gnu.org to
+provide the necessary information to handle your system.
 
 config.guess timestamp = $timestamp
 
@@ -1545,16 +1460,16 @@ hostinfo               = `(hostinfo) 2>/dev/null`
 /usr/bin/oslevel       = `(/usr/bin/oslevel) 2>/dev/null`
 /usr/convex/getsysinfo = `(/usr/convex/getsysinfo) 2>/dev/null`
 
-UNAME_MACHINE = ${UNAME_MACHINE}
-UNAME_RELEASE = ${UNAME_RELEASE}
-UNAME_SYSTEM  = ${UNAME_SYSTEM}
-UNAME_VERSION = ${UNAME_VERSION}
+UNAME_MACHINE = "$UNAME_MACHINE"
+UNAME_RELEASE = "$UNAME_RELEASE"
+UNAME_SYSTEM  = "$UNAME_SYSTEM"
+UNAME_VERSION = "$UNAME_VERSION"
 EOF
 
 exit 1
 
 # Local variables:
-# eval: (add-hook 'write-file-hooks 'time-stamp)
+# eval: (add-hook 'before-save-hook 'time-stamp)
 # time-stamp-start: "timestamp='"
 # time-stamp-format: "%:y-%02m-%02d"
 # time-stamp-end: "'"
diff --git a/tools/config.sub b/tools/config.sub
index a39437d0..657a8625 100755
--- a/tools/config.sub
+++ b/tools/config.sub
@@ -1,44 +1,40 @@
 #! /bin/sh
 # Configuration validation subroutine script.
-#   Copyright (C) 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
-#   2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008
-#   Free Software Foundation, Inc.
+#   Copyright 1992-2018 Free Software Foundation, Inc.
 
-timestamp='2009-04-17'
+timestamp='2018-04-16'
 
-# This file is (in principle) common to ALL GNU software.
-# The presence of a machine in this file suggests that SOME GNU software
-# can handle that machine.  It does not imply ALL GNU software can.
-#
-# This file is free software; you can redistribute it and/or modify
-# it under the terms of the GNU General Public License as published by
-# the Free Software Foundation; either version 2 of the License, or
+# This file is free software; you can redistribute it and/or modify it
+# under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
 # (at your option) any later version.
 #
-# This program is distributed in the hope that it will be useful,
-# but WITHOUT ANY WARRANTY; without even the implied warranty of
-# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-# GNU General Public License for more details.
+# This program is distributed in the hope that it will be useful, but
+# WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+# General Public License for more details.
 #
 # You should have received a copy of the GNU General Public License
-# along with this program; if not, write to the Free Software
-# Foundation, Inc., 51 Franklin Street - Fifth Floor, Boston, MA
-# 02110-1301, USA.
+# along with this program; if not, see <https://www.gnu.org/licenses/>.
 #
 # As a special exception to the GNU General Public License, if you
 # distribute this file as part of a program that contains a
 # configuration script generated by Autoconf, you may include it under
-# the same distribution terms that you use for the rest of that program.
+# the same distribution terms that you use for the rest of that
+# program.  This Exception is an additional permission under section 7
+# of the GNU General Public License, version 3 ("GPLv3").
 
 
-# Please send patches to <config-patches@gnu.org>.  Submit a context
-# diff and a properly formatted ChangeLog entry.
+# Please send patches to <config-patches@gnu.org>.
 #
 # Configuration subroutine to validate and canonicalize a configuration type.
 # Supply the specified configuration type as an argument.
 # If it is invalid, we print an error message on stderr and exit with code 1.
 # Otherwise, we print the canonical config type on stdout and succeed.
 
+# You can get the latest version of this script from:
+# https://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.sub
+
 # This file is supposed to be the same for all GNU packages
 # and recognize all the CPU types, system types and aliases
 # that are meaningful with *any* GNU software.
@@ -57,12 +53,11 @@ timestamp='2009-04-17'
 me=`echo "$0" | sed -e 's,.*/,,'`
 
 usage="\
-Usage: $0 [OPTION] CPU-MFR-OPSYS
-       $0 [OPTION] ALIAS
+Usage: $0 [OPTION] CPU-MFR-OPSYS or ALIAS
 
 Canonicalize a configuration name.
 
-Operation modes:
+Options:
   -h, --help         print this help, then exit
   -t, --time-stamp   print date of last modification, then exit
   -v, --version      print version number, then exit
@@ -72,8 +67,7 @@ Report bugs and patches to <config-patches@gnu.org>."
 version="\
 GNU config.sub ($timestamp)
 
-Copyright (C) 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001,
-2002, 2003, 2004, 2005, 2006, 2007, 2008 Free Software Foundation, Inc.
+Copyright 1992-2018 Free Software Foundation, Inc.
 
 This is free software; see the source for copying conditions.  There is NO
 warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE."
@@ -100,7 +94,7 @@ while test $# -gt 0 ; do
 
     *local*)
        # First pass through any local machine types.
-       echo $1
+       echo "$1"
        exit ;;
 
     * )
@@ -118,19 +112,24 @@ esac
 
 # Separate what the user gave into CPU-COMPANY and OS or KERNEL-OS (if any).
 # Here we must recognize all the valid KERNEL-OS combinations.
-maybe_os=`echo $1 | sed 's/^\(.*\)-\([^-]*-[^-]*\)$/\2/'`
+maybe_os=`echo "$1" | sed 's/^\(.*\)-\([^-]*-[^-]*\)$/\2/'`
 case $maybe_os in
-  nto-qnx* | linux-gnu* | linux-dietlibc | linux-newlib* | linux-uclibc* | \
-  uclinux-uclibc* | uclinux-gnu* | kfreebsd*-gnu* | knetbsd*-gnu* | netbsd*-gnu* | \
-  kopensolaris*-gnu* | \
+  nto-qnx* | linux-gnu* | linux-android* | linux-dietlibc | linux-newlib* | \
+  linux-musl* | linux-uclibc* | uclinux-uclibc* | uclinux-gnu* | kfreebsd*-gnu* | \
+  knetbsd*-gnu* | netbsd*-gnu* | netbsd*-eabi* | \
+  kopensolaris*-gnu* | cloudabi*-eabi* | \
   storm-chaos* | os2-emx* | rtmk-nova*)
     os=-$maybe_os
-    basic_machine=`echo $1 | sed 's/^\(.*\)-\([^-]*-[^-]*\)$/\1/'`
+    basic_machine=`echo "$1" | sed 's/^\(.*\)-\([^-]*-[^-]*\)$/\1/'`
+    ;;
+  android-linux)
+    os=-linux-android
+    basic_machine=`echo "$1" | sed 's/^\(.*\)-\([^-]*-[^-]*\)$/\1/'`-unknown
     ;;
   *)
-    basic_machine=`echo $1 | sed 's/-[^-]*$//'`
-    if [ $basic_machine != $1 ]
-    then os=`echo $1 | sed 's/.*-/-/'`
+    basic_machine=`echo "$1" | sed 's/-[^-]*$//'`
+    if [ "$basic_machine" != "$1" ]
+    then os=`echo "$1" | sed 's/.*-/-/'`
     else os=; fi
     ;;
 esac
@@ -149,10 +148,13 @@ case $os in
 	-convergent* | -ncr* | -news | -32* | -3600* | -3100* | -hitachi* |\
 	-c[123]* | -convex* | -sun | -crds | -omron* | -dg | -ultra | -tti* | \
 	-harris | -dolphin | -highlevel | -gould | -cbm | -ns | -masscomp | \
-	-apple | -axis | -knuth | -cray)
+	-apple | -axis | -knuth | -cray | -microblaze*)
 		os=
 		basic_machine=$1
 		;;
+	-bluegene*)
+		os=-cnk
+		;;
 	-sim | -cisco | -oki | -wec | -winbond)
 		os=
 		basic_machine=$1
@@ -167,62 +169,65 @@ case $os in
 		os=-chorusos
 		basic_machine=$1
 		;;
- 	-chorusrdb)
- 		os=-chorusrdb
+	-chorusrdb)
+		os=-chorusrdb
 		basic_machine=$1
- 		;;
+		;;
 	-hiux*)
 		os=-hiuxwe2
 		;;
 	-sco6)
 		os=-sco5v6
-		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
+		basic_machine=`echo "$1" | sed -e 's/86-.*/86-pc/'`
 		;;
 	-sco5)
 		os=-sco3.2v5
-		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
+		basic_machine=`echo "$1" | sed -e 's/86-.*/86-pc/'`
 		;;
 	-sco4)
 		os=-sco3.2v4
-		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
+		basic_machine=`echo "$1" | sed -e 's/86-.*/86-pc/'`
 		;;
 	-sco3.2.[4-9]*)
 		os=`echo $os | sed -e 's/sco3.2./sco3.2v/'`
-		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
+		basic_machine=`echo "$1" | sed -e 's/86-.*/86-pc/'`
 		;;
 	-sco3.2v[4-9]*)
 		# Don't forget version if it is 3.2v4 or newer.
-		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
+		basic_machine=`echo "$1" | sed -e 's/86-.*/86-pc/'`
 		;;
 	-sco5v6*)
 		# Don't forget version if it is 3.2v4 or newer.
-		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
+		basic_machine=`echo "$1" | sed -e 's/86-.*/86-pc/'`
 		;;
 	-sco*)
 		os=-sco3.2v2
-		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
+		basic_machine=`echo "$1" | sed -e 's/86-.*/86-pc/'`
 		;;
 	-udk*)
-		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
+		basic_machine=`echo "$1" | sed -e 's/86-.*/86-pc/'`
 		;;
 	-isc)
 		os=-isc2.2
-		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
+		basic_machine=`echo "$1" | sed -e 's/86-.*/86-pc/'`
 		;;
 	-clix*)
 		basic_machine=clipper-intergraph
 		;;
 	-isc*)
-		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
+		basic_machine=`echo "$1" | sed -e 's/86-.*/86-pc/'`
+		;;
+	-lynx*178)
+		os=-lynxos178
+		;;
+	-lynx*5)
+		os=-lynxos5
 		;;
 	-lynx*)
 		os=-lynxos
 		;;
 	-ptx*)
-		basic_machine=`echo $1 | sed -e 's/86-.*/86-sequent/'`
-		;;
-	-windowsnt*)
-		os=`echo $os | sed -e 's/windowsnt/winnt/'`
+		basic_machine=`echo "$1" | sed -e 's/86-.*/86-sequent/'`
 		;;
 	-psos*)
 		os=-psos
@@ -239,20 +244,29 @@ case $basic_machine in
 	# Some are omitted here because they have special meanings below.
 	1750a | 580 \
 	| a29k \
+	| aarch64 | aarch64_be \
 	| alpha | alphaev[4-8] | alphaev56 | alphaev6[78] | alphapca5[67] \
 	| alpha64 | alpha64ev[4-8] | alpha64ev56 | alpha64ev6[78] | alpha64pca5[67] \
 	| am33_2.0 \
-	| arc | arm | arm[bl]e | arme[lb] | armv[2345] | armv[345][lb] | avr | avr32 \
+	| arc | arceb \
+	| arm | arm[bl]e | arme[lb] | armv[2-8] | armv[3-8][lb] | armv7[arm] \
+	| avr | avr32 \
+	| ba \
+	| be32 | be64 \
 	| bfin \
-	| c4x | clipper \
+	| c4x | c8051 | clipper | csky \
 	| d10v | d30v | dlx | dsp16xx \
-	| fido | fr30 | frv \
+	| e2k | epiphany \
+	| fido | fr30 | frv | ft32 \
 	| h8300 | h8500 | hppa | hppa1.[01] | hppa2.0 | hppa2.0[nw] | hppa64 \
-	| i370 | i860 | i960 | ia64 \
+	| hexagon \
+	| i370 | i860 | i960 | ia16 | ia64 \
 	| ip2k | iq2000 \
+	| k1om \
+	| le32 | le64 \
 	| lm32 \
 	| m32c | m32r | m32rle | m68000 | m68k | m88k \
-	| maxq | mb | microblaze | mcore | mep | metag \
+	| maxq | mb | microblaze | microblazeel | mcore | mep | metag \
 	| mips | mipsbe | mipseb | mipsel | mipsle \
 	| mips16 \
 	| mips64 | mips64el \
@@ -266,45 +280,84 @@ case $basic_machine in
 	| mips64vr5900 | mips64vr5900el \
 	| mipsisa32 | mipsisa32el \
 	| mipsisa32r2 | mipsisa32r2el \
+	| mipsisa32r6 | mipsisa32r6el \
 	| mipsisa64 | mipsisa64el \
 	| mipsisa64r2 | mipsisa64r2el \
+	| mipsisa64r6 | mipsisa64r6el \
 	| mipsisa64sb1 | mipsisa64sb1el \
 	| mipsisa64sr71k | mipsisa64sr71kel \
+	| mipsr5900 | mipsr5900el \
 	| mipstx39 | mipstx39el \
 	| mn10200 | mn10300 \
 	| moxie \
 	| mt \
 	| msp430 \
-	| nios | nios2 \
+	| nds32 | nds32le | nds32be \
+	| nios | nios2 | nios2eb | nios2el \
 	| ns16k | ns32k \
-	| or32 \
-	| pdp10 | pdp11 | pj | pjl \
-	| powerpc | powerpc64 | powerpc64le | powerpcle | ppcbe \
+	| open8 | or1k | or1knd | or32 \
+	| pdp10 | pj | pjl \
+	| powerpc | powerpc64 | powerpc64le | powerpcle \
+	| pru \
 	| pyramid \
+	| riscv32 | riscv64 \
+	| rl78 | rx \
 	| score \
-	| sh | sh[1234] | sh[24]a | sh[24]aeb | sh[23]e | sh[34]eb | sheb | shbe | shle | sh[1234]le | sh3ele \
+	| sh | sh[1234] | sh[24]a | sh[24]aeb | sh[23]e | sh[234]eb | sheb | shbe | shle | sh[1234]le | sh3ele \
 	| sh64 | sh64le \
 	| sparc | sparc64 | sparc64b | sparc64v | sparc86x | sparclet | sparclite \
 	| sparcv8 | sparcv9 | sparcv9b | sparcv9v \
-	| spu | strongarm \
-	| tahoe | thumb | tic4x | tic80 | tron \
-	| v850 | v850e \
-	| we32k \
-	| x86 | xc16x | xscale | xscalee[bl] | xstormy16 | xtensa \
+	| spu \
+	| tahoe | tic4x | tic54x | tic55x | tic6x | tic80 | tron \
+	| ubicom32 \
+	| v850 | v850e | v850e1 | v850e2 | v850es | v850e2v3 \
+	| visium \
+	| wasm32 \
+	| x86 | xc16x | xstormy16 | xtensa \
 	| z8k | z80)
 		basic_machine=$basic_machine-unknown
 		;;
-	m6811 | m68hc11 | m6812 | m68hc12)
-		# Motorola 68HC11/12.
+	c54x)
+		basic_machine=tic54x-unknown
+		;;
+	c55x)
+		basic_machine=tic55x-unknown
+		;;
+	c6x)
+		basic_machine=tic6x-unknown
+		;;
+	leon|leon[3-9])
+		basic_machine=sparc-$basic_machine
+		;;
+	m6811 | m68hc11 | m6812 | m68hc12 | m68hcs12x | nvptx | picochip)
 		basic_machine=$basic_machine-unknown
 		os=-none
 		;;
-	m88110 | m680[12346]0 | m683?2 | m68360 | m5200 | v70 | w65 | z8k)
+	m88110 | m680[12346]0 | m683?2 | m68360 | m5200 | v70 | w65)
+		;;
+	m9s12z | m68hcs12z | hcs12z | s12z)
+		basic_machine=s12z-unknown
+		os=-none
 		;;
 	ms1)
 		basic_machine=mt-unknown
 		;;
 
+	strongarm | thumb | xscale)
+		basic_machine=arm-unknown
+		;;
+	xgate)
+		basic_machine=$basic_machine-unknown
+		os=-none
+		;;
+	xscaleeb)
+		basic_machine=armeb-unknown
+		;;
+
+	xscaleel)
+		basic_machine=armel-unknown
+		;;
+
 	# We use `pc' rather than `unknown'
 	# because (1) that's what they normally are, and
 	# (2) the word "unknown" tends to confuse beginning users.
@@ -313,31 +366,38 @@ case $basic_machine in
 	  ;;
 	# Object if more than one company name word.
 	*-*-*)
-		echo Invalid configuration \`$1\': machine \`$basic_machine\' not recognized 1>&2
+		echo Invalid configuration \`"$1"\': machine \`"$basic_machine"\' not recognized 1>&2
 		exit 1
 		;;
 	# Recognize the basic CPU types with company name.
 	580-* \
 	| a29k-* \
+	| aarch64-* | aarch64_be-* \
 	| alpha-* | alphaev[4-8]-* | alphaev56-* | alphaev6[78]-* \
 	| alpha64-* | alpha64ev[4-8]-* | alpha64ev56-* | alpha64ev6[78]-* \
-	| alphapca5[67]-* | alpha64pca5[67]-* | arc-* \
+	| alphapca5[67]-* | alpha64pca5[67]-* | arc-* | arceb-* \
 	| arm-*  | armbe-* | armle-* | armeb-* | armv*-* \
 	| avr-* | avr32-* \
+	| ba-* \
+	| be32-* | be64-* \
 	| bfin-* | bs2000-* \
-	| c[123]* | c30-* | [cjt]90-* | c4x-* | c54x-* | c55x-* | c6x-* \
-	| clipper-* | craynv-* | cydra-* \
+	| c[123]* | c30-* | [cjt]90-* | c4x-* \
+	| c8051-* | clipper-* | craynv-* | csky-* | cydra-* \
 	| d10v-* | d30v-* | dlx-* \
-	| elxsi-* \
+	| e2k-* | elxsi-* \
 	| f30[01]-* | f700-* | fido-* | fr30-* | frv-* | fx80-* \
 	| h8300-* | h8500-* \
 	| hppa-* | hppa1.[01]-* | hppa2.0-* | hppa2.0[nw]-* | hppa64-* \
-	| i*86-* | i860-* | i960-* | ia64-* \
+	| hexagon-* \
+	| i*86-* | i860-* | i960-* | ia16-* | ia64-* \
 	| ip2k-* | iq2000-* \
+	| k1om-* \
+	| le32-* | le64-* \
 	| lm32-* \
 	| m32c-* | m32r-* | m32rle-* \
 	| m68000-* | m680[012346]0-* | m68360-* | m683?2-* | m68k-* \
 	| m88110-* | m88k-* | maxq-* | mcore-* | metag-* \
+	| microblaze-* | microblazeel-* \
 	| mips-* | mipsbe-* | mipseb-* | mipsel-* | mipsle-* \
 	| mips16-* \
 	| mips64-* | mips64el-* \
@@ -351,32 +411,45 @@ case $basic_machine in
 	| mips64vr5900-* | mips64vr5900el-* \
 	| mipsisa32-* | mipsisa32el-* \
 	| mipsisa32r2-* | mipsisa32r2el-* \
+	| mipsisa32r6-* | mipsisa32r6el-* \
 	| mipsisa64-* | mipsisa64el-* \
 	| mipsisa64r2-* | mipsisa64r2el-* \
+	| mipsisa64r6-* | mipsisa64r6el-* \
 	| mipsisa64sb1-* | mipsisa64sb1el-* \
 	| mipsisa64sr71k-* | mipsisa64sr71kel-* \
+	| mipsr5900-* | mipsr5900el-* \
 	| mipstx39-* | mipstx39el-* \
 	| mmix-* \
 	| mt-* \
 	| msp430-* \
-	| nios-* | nios2-* \
+	| nds32-* | nds32le-* | nds32be-* \
+	| nios-* | nios2-* | nios2eb-* | nios2el-* \
 	| none-* | np1-* | ns16k-* | ns32k-* \
+	| open8-* \
+	| or1k*-* \
 	| orion-* \
 	| pdp10-* | pdp11-* | pj-* | pjl-* | pn-* | power-* \
-	| powerpc-* | powerpc64-* | powerpc64le-* | powerpcle-* | ppcbe-* \
+	| powerpc-* | powerpc64-* | powerpc64le-* | powerpcle-* \
+	| pru-* \
 	| pyramid-* \
-	| romp-* | rs6000-* \
+	| riscv32-* | riscv64-* \
+	| rl78-* | romp-* | rs6000-* | rx-* \
 	| sh-* | sh[1234]-* | sh[24]a-* | sh[24]aeb-* | sh[23]e-* | sh[34]eb-* | sheb-* | shbe-* \
 	| shle-* | sh[1234]le-* | sh3ele-* | sh64-* | sh64le-* \
 	| sparc-* | sparc64-* | sparc64b-* | sparc64v-* | sparc86x-* | sparclet-* \
 	| sparclite-* \
-	| sparcv8-* | sparcv9-* | sparcv9b-* | sparcv9v-* | strongarm-* | sv1-* | sx?-* \
-	| tahoe-* | thumb-* \
-	| tic30-* | tic4x-* | tic54x-* | tic55x-* | tic6x-* | tic80-* | tile-* \
+	| sparcv8-* | sparcv9-* | sparcv9b-* | sparcv9v-* | sv1-* | sx*-* \
+	| tahoe-* \
+	| tic30-* | tic4x-* | tic54x-* | tic55x-* | tic6x-* | tic80-* \
+	| tile*-* \
 	| tron-* \
-	| v850-* | v850e-* | vax-* \
+	| ubicom32-* \
+	| v850-* | v850e-* | v850e1-* | v850es-* | v850e2-* | v850e2v3-* \
+	| vax-* \
+	| visium-* \
+	| wasm32-* \
 	| we32k-* \
-	| x86-* | x86_64-* | xc16x-* | xps100-* | xscale-* | xscalee[bl]-* \
+	| x86-* | x86_64-* | xc16x-* | xps100-* \
 	| xstormy16-* | xtensa*-* \
 	| ymp-* \
 	| z8k-* | z80-*)
@@ -388,7 +461,7 @@ case $basic_machine in
 	# Recognize the various machine names and aliases which stand
 	# for a CPU type and a company and sometimes even an OS.
 	386bsd)
-		basic_machine=i386-unknown
+		basic_machine=i386-pc
 		os=-bsd
 		;;
 	3b1 | 7300 | 7300-att | att-7300 | pc7300 | safari | unixpc)
@@ -401,7 +474,7 @@ case $basic_machine in
 		basic_machine=a29k-amd
 		os=-udi
 		;;
-    	abacus)
+	abacus)
 		basic_machine=abacus-unknown
 		;;
 	adobe68k)
@@ -422,7 +495,7 @@ case $basic_machine in
 		basic_machine=x86_64-pc
 		;;
 	amd64-*)
-		basic_machine=x86_64-`echo $basic_machine | sed 's/^[^-]*-//'`
+		basic_machine=x86_64-`echo "$basic_machine" | sed 's/^[^-]*-//'`
 		;;
 	amdahl)
 		basic_machine=580-amdahl
@@ -451,6 +524,9 @@ case $basic_machine in
 		basic_machine=i386-pc
 		os=-aros
 		;;
+	asmjs)
+		basic_machine=asmjs-unknown
+		;;
 	aux)
 		basic_machine=m68k-apple
 		os=-aux
@@ -464,14 +540,27 @@ case $basic_machine in
 		os=-linux
 		;;
 	blackfin-*)
-		basic_machine=bfin-`echo $basic_machine | sed 's/^[^-]*-//'`
+		basic_machine=bfin-`echo "$basic_machine" | sed 's/^[^-]*-//'`
 		os=-linux
 		;;
+	bluegene*)
+		basic_machine=powerpc-ibm
+		os=-cnk
+		;;
+	c54x-*)
+		basic_machine=tic54x-`echo "$basic_machine" | sed 's/^[^-]*-//'`
+		;;
+	c55x-*)
+		basic_machine=tic55x-`echo "$basic_machine" | sed 's/^[^-]*-//'`
+		;;
+	c6x-*)
+		basic_machine=tic6x-`echo "$basic_machine" | sed 's/^[^-]*-//'`
+		;;
 	c90)
 		basic_machine=c90-cray
 		os=-unicos
 		;;
-        cegcc)
+	cegcc)
 		basic_machine=arm-unknown
 		os=-cegcc
 		;;
@@ -503,7 +592,7 @@ case $basic_machine in
 		basic_machine=craynv-cray
 		os=-unicosmp
 		;;
-	cr16)
+	cr16 | cr16-*)
 		basic_machine=cr16-unknown
 		os=-elf
 		;;
@@ -554,10 +643,18 @@ case $basic_machine in
 		basic_machine=rs6000-bull
 		os=-bosx
 		;;
-	dpx2* | dpx2*-bull)
+	dpx2*)
 		basic_machine=m68k-bull
 		os=-sysv3
 		;;
+	e500v[12])
+		basic_machine=powerpc-unknown
+		os=$os"spe"
+		;;
+	e500v[12]-*)
+		basic_machine=powerpc-`echo "$basic_machine" | sed 's/^[^-]*-//'`
+		os=$os"spe"
+		;;
 	ebmon29k)
 		basic_machine=a29k-amd
 		os=-ebmon
@@ -647,9 +744,6 @@ case $basic_machine in
 	hp9k8[0-9][0-9] | hp8[0-9][0-9])
 		basic_machine=hppa1.0-hp
 		;;
-	hppa-next)
-		os=-nextstep3
-		;;
 	hppaosf)
 		basic_machine=hppa1.1-hp
 		os=-osf
@@ -661,28 +755,27 @@ case $basic_machine in
 	i370-ibm* | ibm*)
 		basic_machine=i370-ibm
 		;;
-# I'm not sure what "Sysv32" means.  Should this be sysv3.2?
 	i*86v32)
-		basic_machine=`echo $1 | sed -e 's/86.*/86-pc/'`
+		basic_machine=`echo "$1" | sed -e 's/86.*/86-pc/'`
 		os=-sysv32
 		;;
 	i*86v4*)
-		basic_machine=`echo $1 | sed -e 's/86.*/86-pc/'`
+		basic_machine=`echo "$1" | sed -e 's/86.*/86-pc/'`
 		os=-sysv4
 		;;
 	i*86v)
-		basic_machine=`echo $1 | sed -e 's/86.*/86-pc/'`
+		basic_machine=`echo "$1" | sed -e 's/86.*/86-pc/'`
 		os=-sysv
 		;;
 	i*86sol2)
-		basic_machine=`echo $1 | sed -e 's/86.*/86-pc/'`
+		basic_machine=`echo "$1" | sed -e 's/86.*/86-pc/'`
 		os=-solaris2
 		;;
 	i386mach)
 		basic_machine=i386-mach
 		os=-mach
 		;;
-	i386-vsta | vsta)
+	vsta)
 		basic_machine=i386-unknown
 		os=-vsta
 		;;
@@ -700,17 +793,17 @@ case $basic_machine in
 		basic_machine=m68k-isi
 		os=-sysv
 		;;
+	leon-*|leon[3-9]-*)
+		basic_machine=sparc-`echo "$basic_machine" | sed 's/-.*//'`
+		;;
 	m68knommu)
 		basic_machine=m68k-unknown
 		os=-linux
 		;;
 	m68knommu-*)
-		basic_machine=m68k-`echo $basic_machine | sed 's/^[^-]*-//'`
+		basic_machine=m68k-`echo "$basic_machine" | sed 's/^[^-]*-//'`
 		os=-linux
 		;;
-	m88k-omron*)
-		basic_machine=m88k-omron
-		;;
 	magnum | m3230)
 		basic_machine=mips-mips
 		os=-sysv
@@ -719,8 +812,15 @@ case $basic_machine in
 		basic_machine=ns32k-utek
 		os=-sysv
 		;;
+	microblaze*)
+		basic_machine=microblaze-xilinx
+		;;
+	mingw64)
+		basic_machine=x86_64-pc
+		os=-mingw64
+		;;
 	mingw32)
-		basic_machine=i386-pc
+		basic_machine=i686-pc
 		os=-mingw32
 		;;
 	mingw32ce)
@@ -735,10 +835,10 @@ case $basic_machine in
 		os=-mint
 		;;
 	mips3*-*)
-		basic_machine=`echo $basic_machine | sed -e 's/mips3/mips64/'`
+		basic_machine=`echo "$basic_machine" | sed -e 's/mips3/mips64/'`
 		;;
 	mips3*)
-		basic_machine=`echo $basic_machine | sed -e 's/mips3/mips64/'`-unknown
+		basic_machine=`echo "$basic_machine" | sed -e 's/mips3/mips64/'`-unknown
 		;;
 	monitor)
 		basic_machine=m68k-rom68k
@@ -748,17 +848,29 @@ case $basic_machine in
 		basic_machine=powerpc-unknown
 		os=-morphos
 		;;
+	moxiebox)
+		basic_machine=moxie-unknown
+		os=-moxiebox
+		;;
 	msdos)
 		basic_machine=i386-pc
 		os=-msdos
 		;;
 	ms1-*)
-		basic_machine=`echo $basic_machine | sed -e 's/ms1-/mt-/'`
+		basic_machine=`echo "$basic_machine" | sed -e 's/ms1-/mt-/'`
+		;;
+	msys)
+		basic_machine=i686-pc
+		os=-msys
 		;;
 	mvs)
 		basic_machine=i370-ibm
 		os=-mvs
 		;;
+	nacl)
+		basic_machine=le32-unknown
+		os=-nacl
+		;;
 	ncr3000)
 		basic_machine=i486-ncr
 		os=-sysv4
@@ -787,7 +899,7 @@ case $basic_machine in
 		basic_machine=v70-nec
 		os=-sysv
 		;;
-	next | m*-next )
+	next | m*-next)
 		basic_machine=m68k-next
 		case $os in
 		    -nextstep* )
@@ -823,9 +935,21 @@ case $basic_machine in
 	np1)
 		basic_machine=np1-gould
 		;;
+	neo-tandem)
+		basic_machine=neo-tandem
+		;;
+	nse-tandem)
+		basic_machine=nse-tandem
+		;;
 	nsr-tandem)
 		basic_machine=nsr-tandem
 		;;
+	nsv-tandem)
+		basic_machine=nsv-tandem
+		;;
+	nsx-tandem)
+		basic_machine=nsx-tandem
+		;;
 	op50n-* | op60c-*)
 		basic_machine=hppa1.1-oki
 		os=-proelf
@@ -858,7 +982,7 @@ case $basic_machine in
 		os=-linux
 		;;
 	parisc-*)
-		basic_machine=hppa-`echo $basic_machine | sed 's/^[^-]*-//'`
+		basic_machine=hppa-`echo "$basic_machine" | sed 's/^[^-]*-//'`
 		os=-linux
 		;;
 	pbd)
@@ -874,7 +998,7 @@ case $basic_machine in
 		basic_machine=i386-pc
 		;;
 	pc98-*)
-		basic_machine=i386-`echo $basic_machine | sed 's/^[^-]*-//'`
+		basic_machine=i386-`echo "$basic_machine" | sed 's/^[^-]*-//'`
 		;;
 	pentium | p5 | k5 | k6 | nexgen | viac3)
 		basic_machine=i586-pc
@@ -889,41 +1013,42 @@ case $basic_machine in
 		basic_machine=i786-pc
 		;;
 	pentium-* | p5-* | k5-* | k6-* | nexgen-* | viac3-*)
-		basic_machine=i586-`echo $basic_machine | sed 's/^[^-]*-//'`
+		basic_machine=i586-`echo "$basic_machine" | sed 's/^[^-]*-//'`
 		;;
 	pentiumpro-* | p6-* | 6x86-* | athlon-*)
-		basic_machine=i686-`echo $basic_machine | sed 's/^[^-]*-//'`
+		basic_machine=i686-`echo "$basic_machine" | sed 's/^[^-]*-//'`
 		;;
 	pentiumii-* | pentium2-* | pentiumiii-* | pentium3-*)
-		basic_machine=i686-`echo $basic_machine | sed 's/^[^-]*-//'`
+		basic_machine=i686-`echo "$basic_machine" | sed 's/^[^-]*-//'`
 		;;
 	pentium4-*)
-		basic_machine=i786-`echo $basic_machine | sed 's/^[^-]*-//'`
+		basic_machine=i786-`echo "$basic_machine" | sed 's/^[^-]*-//'`
 		;;
 	pn)
 		basic_machine=pn-gould
 		;;
 	power)	basic_machine=power-ibm
 		;;
-	ppc)	basic_machine=powerpc-unknown
+	ppc | ppcbe)	basic_machine=powerpc-unknown
 		;;
-	ppc-*)	basic_machine=powerpc-`echo $basic_machine | sed 's/^[^-]*-//'`
+	ppc-* | ppcbe-*)
+		basic_machine=powerpc-`echo "$basic_machine" | sed 's/^[^-]*-//'`
 		;;
-	ppcle | powerpclittle | ppc-le | powerpc-little)
+	ppcle | powerpclittle)
 		basic_machine=powerpcle-unknown
 		;;
 	ppcle-* | powerpclittle-*)
-		basic_machine=powerpcle-`echo $basic_machine | sed 's/^[^-]*-//'`
+		basic_machine=powerpcle-`echo "$basic_machine" | sed 's/^[^-]*-//'`
 		;;
 	ppc64)	basic_machine=powerpc64-unknown
 		;;
-	ppc64-*) basic_machine=powerpc64-`echo $basic_machine | sed 's/^[^-]*-//'`
+	ppc64-*) basic_machine=powerpc64-`echo "$basic_machine" | sed 's/^[^-]*-//'`
 		;;
-	ppc64le | powerpc64little | ppc64-le | powerpc64-little)
+	ppc64le | powerpc64little)
 		basic_machine=powerpc64le-unknown
 		;;
 	ppc64le-* | powerpc64little-*)
-		basic_machine=powerpc64le-`echo $basic_machine | sed 's/^[^-]*-//'`
+		basic_machine=powerpc64le-`echo "$basic_machine" | sed 's/^[^-]*-//'`
 		;;
 	ps2)
 		basic_machine=i386-ibm
@@ -932,7 +1057,11 @@ case $basic_machine in
 		basic_machine=i586-unknown
 		os=-pw32
 		;;
-	rdos)
+	rdos | rdos64)
+		basic_machine=x86_64-pc
+		os=-rdos
+		;;
+	rdos32)
 		basic_machine=i386-pc
 		os=-rdos
 		;;
@@ -973,17 +1102,10 @@ case $basic_machine in
 	sequent)
 		basic_machine=i386-sequent
 		;;
-	sh)
-		basic_machine=sh-hitachi
-		os=-hms
-		;;
 	sh5el)
 		basic_machine=sh5le-unknown
 		;;
-	sh64)
-		basic_machine=sh64-unknown
-		;;
-	sparclite-wrs | simso-wrs)
+	simso-wrs)
 		basic_machine=sparclite-wrs
 		os=-vxworks
 		;;
@@ -1001,6 +1123,9 @@ case $basic_machine in
 		basic_machine=i860-stratus
 		os=-sysv4
 		;;
+	strongarm-* | thumb-*)
+		basic_machine=arm-`echo "$basic_machine" | sed 's/^[^-]*-//'`
+		;;
 	sun2)
 		basic_machine=m68000-sun
 		;;
@@ -1057,20 +1182,8 @@ case $basic_machine in
 		basic_machine=t90-cray
 		os=-unicos
 		;;
-	tic54x | c54x*)
-		basic_machine=tic54x-unknown
-		os=-coff
-		;;
-	tic55x | c55x*)
-		basic_machine=tic55x-unknown
-		os=-coff
-		;;
-	tic6x | c6x*)
-		basic_machine=tic6x-unknown
-		os=-coff
-		;;
 	tile*)
-		basic_machine=tile-unknown
+		basic_machine=$basic_machine-unknown
 		os=-linux-gnu
 		;;
 	tx39)
@@ -1133,6 +1246,9 @@ case $basic_machine in
 		basic_machine=hppa1.1-winbond
 		os=-proelf
 		;;
+	x64)
+		basic_machine=x86_64-pc
+		;;
 	xbox)
 		basic_machine=i686-pc
 		os=-mingw32
@@ -1140,18 +1256,13 @@ case $basic_machine in
 	xps | xps100)
 		basic_machine=xps100-honeywell
 		;;
+	xscale-* | xscalee[bl]-*)
+		basic_machine=`echo "$basic_machine" | sed 's/^xscale/arm/'`
+		;;
 	ymp)
 		basic_machine=ymp-cray
 		os=-unicos
 		;;
-	z8k-*-coff)
-		basic_machine=z8k-unknown
-		os=-sim
-		;;
-	z80-*-coff)
-		basic_machine=z80-unknown
-		os=-sim
-		;;
 	none)
 		basic_machine=none-none
 		os=-none
@@ -1180,10 +1291,6 @@ case $basic_machine in
 	vax)
 		basic_machine=vax-dec
 		;;
-	pdp10)
-		# there are many clones, so DEC is not a safe bet
-		basic_machine=pdp10-unknown
-		;;
 	pdp11)
 		basic_machine=pdp11-dec
 		;;
@@ -1193,9 +1300,6 @@ case $basic_machine in
 	sh[1234] | sh[24]a | sh[24]aeb | sh[34]eb | sh[1234]le | sh[23]ele)
 		basic_machine=sh-unknown
 		;;
-	sparc | sparcv8 | sparcv9 | sparcv9b | sparcv9v)
-		basic_machine=sparc-sun
-		;;
 	cydra)
 		basic_machine=cydra-cydrome
 		;;
@@ -1215,7 +1319,7 @@ case $basic_machine in
 		# Make sure to match an already-canonicalized machine name.
 		;;
 	*)
-		echo Invalid configuration \`$1\': machine \`$basic_machine\' not recognized 1>&2
+		echo Invalid configuration \`"$1"\': machine \`"$basic_machine"\' not recognized 1>&2
 		exit 1
 		;;
 esac
@@ -1223,10 +1327,10 @@ esac
 # Here we canonicalize certain aliases for manufacturers.
 case $basic_machine in
 	*-digital*)
-		basic_machine=`echo $basic_machine | sed 's/digital.*/dec/'`
+		basic_machine=`echo "$basic_machine" | sed 's/digital.*/dec/'`
 		;;
 	*-commodore*)
-		basic_machine=`echo $basic_machine | sed 's/commodore.*/cbm/'`
+		basic_machine=`echo "$basic_machine" | sed 's/commodore.*/cbm/'`
 		;;
 	*)
 		;;
@@ -1237,53 +1341,60 @@ esac
 if [ x"$os" != x"" ]
 then
 case $os in
-        # First match some system type aliases
-        # that might get confused with valid system types.
+	# First match some system type aliases that might get confused
+	# with valid system types.
 	# -solaris* is a basic system type, with this one exception.
+	-auroraux)
+		os=-auroraux
+		;;
 	-solaris1 | -solaris1.*)
 		os=`echo $os | sed -e 's|solaris1|sunos4|'`
 		;;
 	-solaris)
 		os=-solaris2
 		;;
-	-svr4*)
-		os=-sysv4
-		;;
 	-unixware*)
 		os=-sysv4.2uw
 		;;
 	-gnu/linux*)
 		os=`echo $os | sed -e 's|gnu/linux|linux-gnu|'`
 		;;
-	# First accept the basic system types.
+	# es1800 is here to avoid being matched by es* (a different OS)
+	-es1800*)
+		os=-ose
+		;;
+	# Now accept the basic system types.
 	# The portable systems comes first.
-	# Each alternative MUST END IN A *, to match a version number.
+	# Each alternative MUST end in a * to match a version number.
 	# -sysv* is not here because it comes later, after sysvr4.
 	-gnu* | -bsd* | -mach* | -minix* | -genix* | -ultrix* | -irix* \
-	      | -*vms* | -sco* | -esix* | -isc* | -aix* | -sunos | -sunos[34]*\
-	      | -hpux* | -unos* | -osf* | -luna* | -dgux* | -solaris* | -sym* \
-	      | -kopensolaris* \
+	      | -*vms* | -sco* | -esix* | -isc* | -aix* | -cnk* | -sunos | -sunos[34]*\
+	      | -hpux* | -unos* | -osf* | -luna* | -dgux* | -auroraux* | -solaris* \
+	      | -sym* | -kopensolaris* | -plan9* \
 	      | -amigaos* | -amigados* | -msdos* | -newsos* | -unicos* | -aof* \
-	      | -aos* | -aros* \
+	      | -aos* | -aros* | -cloudabi* | -sortix* \
 	      | -nindy* | -vxsim* | -vxworks* | -ebmon* | -hms* | -mvs* \
 	      | -clix* | -riscos* | -uniplus* | -iris* | -rtu* | -xenix* \
-	      | -hiux* | -386bsd* | -knetbsd* | -mirbsd* | -netbsd* \
-	      | -openbsd* | -solidbsd* \
+	      | -hiux* | -knetbsd* | -mirbsd* | -netbsd* \
+	      | -bitrig* | -openbsd* | -solidbsd* | -libertybsd* \
 	      | -ekkobsd* | -kfreebsd* | -freebsd* | -riscix* | -lynxos* \
 	      | -bosx* | -nextstep* | -cxux* | -aout* | -elf* | -oabi* \
 	      | -ptx* | -coff* | -ecoff* | -winnt* | -domain* | -vsta* \
-	      | -udi* | -eabi* | -lites* | -ieee* | -go32* | -aux* \
-	      | -chorusos* | -chorusrdb* | -cegcc* \
-	      | -cygwin* | -pe* | -psos* | -moss* | -proelf* | -rtems* \
-	      | -mingw32* | -linux-gnu* | -linux-newlib* | -linux-uclibc* \
-	      | -uxpv* | -beos* | -mpeix* | -udk* \
-	      | -interix* | -uwin* | -mks* | -rhapsody* | -darwin* | -opened* \
+	      | -udi* | -eabi* | -lites* | -ieee* | -go32* | -aux* | -hcos* \
+	      | -chorusos* | -chorusrdb* | -cegcc* | -glidix* \
+	      | -cygwin* | -msys* | -pe* | -psos* | -moss* | -proelf* | -rtems* \
+	      | -midipix* | -mingw32* | -mingw64* | -linux-gnu* | -linux-android* \
+	      | -linux-newlib* | -linux-musl* | -linux-uclibc* \
+	      | -uxpv* | -beos* | -mpeix* | -udk* | -moxiebox* \
+	      | -interix* | -uwin* | -mks* | -rhapsody* | -darwin* \
 	      | -openstep* | -oskit* | -conix* | -pw32* | -nonstopux* \
 	      | -storm-chaos* | -tops10* | -tenex* | -tops20* | -its* \
 	      | -os2* | -vos* | -palmos* | -uclinux* | -nucleus* \
-	      | -morphos* | -superux* | -rtmk* | -rtmk-nova* | -windiss* \
+	      | -morphos* | -superux* | -rtmk* | -windiss* \
 	      | -powermax* | -dnix* | -nx6 | -nx7 | -sei* | -dragonfly* \
-	      | -skyos* | -haiku* | -rdos* | -toppers* | -drops*)
+	      | -skyos* | -haiku* | -rdos* | -toppers* | -drops* | -es* \
+	      | -onefs* | -tirtos* | -phoenix* | -fuchsia* | -redox* | -bme* \
+	      | -midnightbsd*)
 	# Remember, each alternative MUST END IN *, to match a version number.
 		;;
 	-qnx*)
@@ -1300,12 +1411,12 @@ case $os in
 	-nto*)
 		os=`echo $os | sed -e 's|nto|nto-qnx|'`
 		;;
-	-sim | -es1800* | -hms* | -xray | -os68k* | -none* | -v88r* \
-	      | -windows* | -osx | -abug | -netware* | -os9* | -beos* | -haiku* \
+	-sim | -xray | -os68k* | -v88r* \
+	      | -windows* | -osx | -abug | -netware* | -os9* \
 	      | -macos* | -mpw* | -magic* | -mmixware* | -mon960* | -lnews*)
 		;;
 	-mac*)
-		os=`echo $os | sed -e 's|mac|macos|'`
+		os=`echo "$os" | sed -e 's|mac|macos|'`
 		;;
 	-linux-dietlibc)
 		os=-linux-dietlibc
@@ -1314,26 +1425,20 @@ case $os in
 		os=`echo $os | sed -e 's|linux|linux-gnu|'`
 		;;
 	-sunos5*)
-		os=`echo $os | sed -e 's|sunos5|solaris2|'`
+		os=`echo "$os" | sed -e 's|sunos5|solaris2|'`
 		;;
 	-sunos6*)
-		os=`echo $os | sed -e 's|sunos6|solaris3|'`
+		os=`echo "$os" | sed -e 's|sunos6|solaris3|'`
 		;;
 	-opened*)
 		os=-openedition
 		;;
-        -os400*)
+	-os400*)
 		os=-os400
 		;;
 	-wince*)
 		os=-wince
 		;;
-	-osfrose*)
-		os=-osfrose
-		;;
-	-osf*)
-		os=-osf
-		;;
 	-utek*)
 		os=-bsd
 		;;
@@ -1358,7 +1463,7 @@ case $os in
 	-nova*)
 		os=-rtmk-nova
 		;;
-	-ns2 )
+	-ns2)
 		os=-nextstep2
 		;;
 	-nsk*)
@@ -1371,7 +1476,7 @@ case $os in
 	-sinix*)
 		os=-sysv4
 		;;
-        -tpf*)
+	-tpf*)
 		os=-tpf
 		;;
 	-triton*)
@@ -1380,7 +1485,7 @@ case $os in
 	-oss*)
 		os=-sysv3
 		;;
-	-svr4)
+	-svr4*)
 		os=-sysv4
 		;;
 	-svr3)
@@ -1395,33 +1500,38 @@ case $os in
 	-ose*)
 		os=-ose
 		;;
-	-es1800*)
-		os=-ose
-		;;
-	-xenix)
-		os=-xenix
-		;;
 	-*mint | -mint[0-9]* | -*MiNT | -MiNT[0-9]*)
 		os=-mint
 		;;
-	-aros*)
-		os=-aros
-		;;
-	-kaos*)
-		os=-kaos
-		;;
 	-zvmoe)
 		os=-zvmoe
 		;;
 	-dicos*)
 		os=-dicos
 		;;
+	-pikeos*)
+		# Until real need of OS specific support for
+		# particular features comes up, bare metal
+		# configurations are quite functional.
+		case $basic_machine in
+		    arm*)
+			os=-eabi
+			;;
+		    *)
+			os=-elf
+			;;
+		esac
+		;;
+	-nacl*)
+		;;
+	-ios)
+		;;
 	-none)
 		;;
 	*)
 		# Get rid of the `-' at the beginning of $os.
 		os=`echo $os | sed 's/[^-]*-//'`
-		echo Invalid configuration \`$1\': system \`$os\' not recognized 1>&2
+		echo Invalid configuration \`"$1"\': system \`"$os"\' not recognized 1>&2
 		exit 1
 		;;
 esac
@@ -1438,10 +1548,10 @@ else
 # system, and we'll never get to this point.
 
 case $basic_machine in
-        score-*)
+	score-*)
 		os=-elf
 		;;
-        spu-*)
+	spu-*)
 		os=-elf
 		;;
 	*-acorn)
@@ -1453,8 +1563,23 @@ case $basic_machine in
 	arm*-semi)
 		os=-aout
 		;;
-        c4x-* | tic4x-*)
-        	os=-coff
+	c4x-* | tic4x-*)
+		os=-coff
+		;;
+	c8051-*)
+		os=-elf
+		;;
+	hexagon-*)
+		os=-elf
+		;;
+	tic54x-*)
+		os=-coff
+		;;
+	tic55x-*)
+		os=-coff
+		;;
+	tic6x-*)
+		os=-coff
 		;;
 	# This must come before the *-dec entry.
 	pdp10-*)
@@ -1474,14 +1599,11 @@ case $basic_machine in
 		;;
 	m68000-sun)
 		os=-sunos3
-		# This also exists in the configure program, but was not the
-		# default.
-		# os=-sunos4
 		;;
 	m68*-cisco)
 		os=-aout
 		;;
-        mep-*)
+	mep-*)
 		os=-elf
 		;;
 	mips*-cisco)
@@ -1499,16 +1621,16 @@ case $basic_machine in
 	sparc-* | *-sun)
 		os=-sunos4.1.1
 		;;
+	pru-*)
+		os=-elf
+		;;
 	*-be)
 		os=-beos
 		;;
-	*-haiku)
-		os=-haiku
-		;;
 	*-ibm)
 		os=-aix
 		;;
-    	*-knuth)
+	*-knuth)
 		os=-mmixware
 		;;
 	*-wec)
@@ -1544,7 +1666,7 @@ case $basic_machine in
 	m88k-omron*)
 		os=-luna
 		;;
-	*-next )
+	*-next)
 		os=-nextstep
 		;;
 	*-sequent)
@@ -1559,9 +1681,6 @@ case $basic_machine in
 	i370-*)
 		os=-mvs
 		;;
-	*-next)
-		os=-nextstep3
-		;;
 	*-gould)
 		os=-sysv
 		;;
@@ -1613,7 +1732,7 @@ case $basic_machine in
 			-sunos*)
 				vendor=sun
 				;;
-			-aix*)
+			-cnk*|-aix*)
 				vendor=ibm
 				;;
 			-beos*)
@@ -1671,15 +1790,15 @@ case $basic_machine in
 				vendor=stratus
 				;;
 		esac
-		basic_machine=`echo $basic_machine | sed "s/unknown/$vendor/"`
+		basic_machine=`echo "$basic_machine" | sed "s/unknown/$vendor/"`
 		;;
 esac
 
-echo $basic_machine$os
+echo "$basic_machine$os"
 exit
 
 # Local variables:
-# eval: (add-hook 'write-file-hooks 'time-stamp)
+# eval: (add-hook 'before-save-hook 'time-stamp)
 # time-stamp-start: "timestamp='"
 # time-stamp-format: "%:y-%02m-%02d"
 # time-stamp-end: "'"

From 586c1800c447ff099d34889b23647c4733876d9b Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 27 Jun 2018 16:51:53 +0200
Subject: [PATCH 180/198] Nest: Neighbor cache cleanups

Simplify neighbor cache code, fix several minor bugs, and improve
handling of ONLINK flag.
---
 filter/filter.c        |   2 +-
 lib/hash.h             |   7 +
 nest/iface.h           |  20 +-
 nest/neighbor.c        | 471 +++++++++++++++++++++--------------------
 nest/rt-table.c        |   8 +-
 proto/bfd/bfd.c        |   2 +-
 proto/bgp/bgp.c        |   4 +-
 proto/bgp/packets.c    |   4 +-
 proto/ospf/rt.c        |   2 +-
 proto/rip/packets.c    |   2 +-
 proto/rip/rip.c        |   2 +-
 proto/static/static.c  |   7 +-
 sysdep/linux/netlink.c |   8 +-
 13 files changed, 267 insertions(+), 272 deletions(-)

diff --git a/filter/filter.c b/filter/filter.c
index 439fd5bf..0a2c26b8 100644
--- a/filter/filter.c
+++ b/filter/filter.c
@@ -974,7 +974,7 @@ interpret(struct f_inst *what)
       case SA_GW:
 	{
 	  ip_addr ip = v1.val.ip;
-	  neighbor *n = neigh_find(rta->src->proto, &ip, 0);
+	  neighbor *n = neigh_find(rta->src->proto, ip, NULL, 0);
 	  if (!n || (n->scope == SCOPE_HOST))
 	    runtime( "Invalid gw address" );
 
diff --git a/lib/hash.h b/lib/hash.h
index 97d8f69c..8809aedc 100644
--- a/lib/hash.h
+++ b/lib/hash.h
@@ -230,4 +230,11 @@ mem_hash(void *p, uint s)
   return mem_hash_value(&h);
 }
 
+static inline uint
+ptr_hash(void *ptr)
+{
+  uintptr_t p = (uintptr_t) ptr;
+  return p ^ (p << 8) ^ (p >> 16);
+}
+
 #endif
diff --git a/nest/iface.h b/nest/iface.h
index ab3f8f35..bc241501 100644
--- a/nest/iface.h
+++ b/nest/iface.h
@@ -124,29 +124,21 @@ typedef struct neighbor {
   ip_addr addr;				/* Address of the neighbor */
   struct ifa *ifa;			/* Ifa on related iface */
   struct iface *iface;			/* Interface it's connected to */
+  struct iface *ifreq;			/* Requested iface, NULL for any */
   struct proto *proto;			/* Protocol this belongs to */
   void *data;				/* Protocol-specific data */
-  unsigned aux;				/* Protocol-specific data */
-  unsigned flags;
-  int scope;				/* Address scope, -1 for unreachable sticky neighbors,
+  uint aux;				/* Protocol-specific data */
+  u16 flags;				/* NEF_* flags */
+  s16 scope;				/* Address scope, -1 for unreachable neighbors,
 					   SCOPE_HOST when it's our own address */
 } neighbor;
 
 #define NEF_STICKY	1
 #define NEF_ONLINK	2
-#define NEF_BIND	4		/* Used internally for neighbors bound to an iface */
-#define NEF_IFACE	8		/* Neighbors bound to iface */
+#define NEF_IFACE	4		/* Entry for whole iface */
 
 
-neighbor *neigh_find(struct proto *, ip_addr *, unsigned flags);
-neighbor *neigh_find2(struct proto *p, ip_addr *a, struct iface *ifa, unsigned flags);
-neighbor *neigh_find_iface(struct proto *p, struct iface *ifa);
-
-static inline int neigh_connected_to(struct proto *p, ip_addr *a, struct iface *i)
-{
-  neighbor *n = neigh_find(p, a, 0);
-  return n && n->iface == i;
-}
+neighbor *neigh_find(struct proto *p, ip_addr a, struct iface *ifa, uint flags);
 
 void neigh_dump(neighbor *);
 void neigh_dump_all(void);
diff --git a/nest/neighbor.c b/nest/neighbor.c
index 4f93e29e..87253857 100644
--- a/nest/neighbor.c
+++ b/nest/neighbor.c
@@ -2,6 +2,8 @@
  *	BIRD -- Neighbor Cache
  *
  *	(c) 1998--2000 Martin Mares <mj@ucw.cz>
+ *	(c) 2008--2018 Ondrej Zajicek <santiago@crfreenet.org>
+ *	(c) 2008--2018 CZ.NIC z.s.p.o.
  *
  *	Can be freely distributed and used under the terms of the GNU GPL.
  */
@@ -10,32 +12,38 @@
  * DOC: Neighbor cache
  *
  * Most routing protocols need to associate their internal state data with
- * neighboring routers, check whether an address given as the next hop
- * attribute of a route is really an address of a directly connected host
- * and which interface is it connected through. Also, they often need to
- * be notified when a neighbor ceases to exist or when their long awaited
- * neighbor becomes connected. The neighbor cache is there to solve all
- * these problems.
+ * neighboring routers, check whether an address given as the next hop attribute
+ * of a route is really an address of a directly connected host and which
+ * interface is it connected through. Also, they often need to be notified when
+ * a neighbor ceases to exist or when their long awaited neighbor becomes
+ * connected. The neighbor cache is there to solve all these problems.
  *
- * The neighbor cache maintains a collection of neighbor entries. Each
- * entry represents one IP address corresponding to either our directly
- * connected neighbor or our own end of the link (when the scope of the
- * address is set to %SCOPE_HOST) together with per-neighbor data belonging to a
- * single protocol.
+ * The neighbor cache maintains a collection of neighbor entries. Each entry
+ * represents one IP address corresponding to either our directly connected
+ * neighbor or our own end of the link (when the scope of the address is set to
+ * %SCOPE_HOST) together with per-neighbor data belonging to a single protocol.
+ * A neighbor entry may be bound to a specific interface, which is required for
+ * link-local IP addresses and optional for global IP addresses.
  *
- * Active entries represent known neighbors and are stored in a hash
- * table (to allow fast retrieval based on the IP address of the node) and
- * two linked lists: one global and one per-interface (allowing quick
- * processing of interface change events). Inactive entries exist only
- * when the protocol has explicitly requested it via the %NEF_STICKY
- * flag because it wishes to be notified when the node will again become
- * a neighbor. Such entries are enqueued in a special list which is walked
- * whenever an interface changes its state to up. Neighbor entry VRF
- * association is implied by respective protocol.
+ * Neighbor cache entries are stored in a hash table, which is indexed by triple
+ * (protocol, IP, requested-iface), so if both regular and iface-bound neighbors
+ * are requested, they are represented by two neighbor cache entries. Active
+ * entries are also linked in per-interface list (allowing quick processing of
+ * interface change events). Inactive entries exist only when the protocol has
+ * explicitly requested it via the %NEF_STICKY flag because it wishes to be
+ * notified when the node will again become a neighbor. Such entries are instead
+ * linked in a special list, which is walked whenever an interface changes its
+ * state to up. Neighbor entry VRF association is implied by respective
+ * protocol.
+ *
+ * Besides the already mentioned %NEF_STICKY flag, there is also %NEF_ONLINK,
+ * which specifies that neighbor should be considered reachable on given iface
+ * regardless of associated address ranges, and %NEF_IFACE, which represents
+ * pseudo-neighbor entry for whole interface (and uses %IPA_NONE IP address).
  *
  * When a neighbor event occurs (a neighbor gets disconnected or a sticky
- * inactive neighbor becomes connected), the protocol hook neigh_notify()
- * is called to advertise the change.
+ * inactive neighbor becomes connected), the protocol hook neigh_notify() is
+ * called to advertise the change.
  */
 
 #undef LOCAL_DEBUG
@@ -43,126 +51,151 @@
 #include "nest/bird.h"
 #include "nest/iface.h"
 #include "nest/protocol.h"
+#include "lib/hash.h"
 #include "lib/resource.h"
 
 #define NEIGH_HASH_SIZE 256
 #define NEIGH_HASH_OFFSET 24
 
 static slab *neigh_slab;
-static list sticky_neigh_list, iface_neigh_list, neigh_hash_table[NEIGH_HASH_SIZE];
+static list neigh_hash_table[NEIGH_HASH_SIZE], sticky_neigh_list;
 
 static inline uint
-neigh_hash(struct proto *p, ip_addr *a)
+neigh_hash(struct proto *p, ip_addr a, struct iface *i)
 {
-  return (p->hash_key ^ ipa_hash(*a)) >> NEIGH_HASH_OFFSET;
+  return (p->hash_key ^ ipa_hash(a) ^ ptr_hash(i)) >> NEIGH_HASH_OFFSET;
 }
 
 static int
-if_connected(ip_addr *a, struct iface *i, struct ifa **ap)
+if_connected(ip_addr a, struct iface *i, struct ifa **ap, uint flags)
 {
   struct ifa *b;
 
+  /* Handle iface pseudo-neighbors */
+  if (flags & NEF_IFACE)
+    return *ap = NULL, (i->flags & IF_UP) ? SCOPE_HOST : -1;
+
+  /* Host addresses match even if iface is down */
+  WALK_LIST(b, i->addrs)
+    if (ipa_equal(a, b->ip))
+      return *ap = b, SCOPE_HOST;
+
+  /* Rest do not match if iface is down */
   if (!(i->flags & IF_UP))
+    return *ap = NULL, -1;
+
+  /* Regular neighbors */
+  WALK_LIST(b, i->addrs)
   {
-    *ap = NULL;
-    return -1;
+    if (b->flags & IA_PEER)
+    {
+      if (ipa_equal(a, b->opposite))
+	return *ap = b, b->scope;
+    }
+    else
+    {
+      if (ipa_in_netX(a, &b->prefix))
+      {
+	/* Do not allow IPv4 network and broadcast addresses */
+	if (ipa_is_ip4(a) &&
+	    (net_pxlen(&b->prefix) < (IP4_MAX_PREFIX_LENGTH - 1)) &&
+	    (ipa_equal(a, net_prefix(&b->prefix)) ||	/* Network address */
+	     ipa_equal(a, b->brd)))			/* Broadcast */
+	  return *ap = NULL, -1;
+
+	return *ap = b, b->scope;
+      }
+    }
   }
 
-  WALK_LIST(b, i->addrs)
-    {
-      *ap = b;
+  /* Handle ONLINK flag */
+  if (flags & NEF_ONLINK)
+    return *ap = NULL, ipa_classify(a) & IADDR_SCOPE_MASK;
 
-      if (ipa_equal(*a, b->ip))
-	return SCOPE_HOST;
-      if (b->flags & IA_PEER)
-	{
-	  if (ipa_equal(*a, b->opposite))
-	    return b->scope;
-	}
-      else
-	{
-	  if (ipa_in_netX(*a, &b->prefix))
-	    {
-	      /* Do not allow IPv4 network and broadcast addresses */
-	      if (ipa_is_ip4(*a) &&
-		  (net_pxlen(&b->prefix) < (IP4_MAX_PREFIX_LENGTH - 1)) &&
-		  (ipa_equal(*a, net_prefix(&b->prefix)) ||	/* Network address */
-		   ipa_equal(*a, b->brd)))	/* Broadcast */
-	      {
-		*ap = NULL;
-		return -1;
-	      }
+  return *ap = NULL, -1;
+}
 
-	      return b->scope;
-	    }
-	}
+static inline int
+if_connected_any(ip_addr a, struct iface *vrf, struct iface **iface, struct ifa **addr, uint flags)
+{
+  struct iface *i;
+  struct ifa *b;
+  int s, scope = -1;
+
+  *iface = NULL;
+  *addr = NULL;
+
+  /* Get first match, but prefer SCOPE_HOST to other matches */
+  WALK_LIST(i, iface_list)
+    if ((!vrf || vrf == i->master) && ((s = if_connected(a, i, &b, flags)) >= 0))
+      if ((scope < 0) || ((scope > SCOPE_HOST) && (s == SCOPE_HOST)))
+      {
+	*iface = i;
+	*addr = b;
+	scope = s;
       }
 
-  *ap = NULL;
-  return -1;
+  return scope;
 }
 
 /**
- * neigh_find - find or create a neighbor entry.
- * @p: protocol which asks for the entry.
- * @a: pointer to IP address of the node to be searched for.
- * @flags: 0 or %NEF_STICKY if you want to create a sticky entry.
+ * neigh_find - find or create a neighbor entry
+ * @p: protocol which asks for the entry
+ * @a: IP address of the node to be searched for
+ * @iface: optionally bound neighbor to this iface (may be NULL)
+ * @flags: %NEF_STICKY for sticky entry, %NEF_ONLINK for onlink entry
  *
- * Search the neighbor cache for a node with given IP address. If
- * it's found, a pointer to the neighbor entry is returned. If no
- * such entry exists and the node is directly connected on
- * one of our active interfaces, a new entry is created and returned
- * to the caller with protocol-dependent fields initialized to zero.
- * If the node is not connected directly or *@a is not a valid unicast
- * IP address, neigh_find() returns %NULL.
+ * Search the neighbor cache for a node with given IP address. Iface can be
+ * specified for link-local addresses or for cases, where neighbor is expected
+ * on given interface. If it is found, a pointer to the neighbor entry is
+ * returned. If no such entry exists and the node is directly connected on one
+ * of our active interfaces, a new entry is created and returned to the caller
+ * with protocol-dependent fields initialized to zero.  If the node is not
+ * connected directly or *@a is not a valid unicast IP address, neigh_find()
+ * returns %NULL.
  */
 neighbor *
-neigh_find(struct proto *p, ip_addr *a, unsigned flags)
-{
-  return neigh_find2(p, a, NULL, flags);
-}
-
-
-neighbor *
-neigh_find2(struct proto *p, ip_addr *a, struct iface *ifa, unsigned flags)
+neigh_find(struct proto *p, ip_addr a, struct iface *iface, uint flags)
 {
   neighbor *n;
   int class, scope = -1;
-  uint h = neigh_hash(p, a);
-  struct iface *i;
-  struct ifa *addr;
+  uint h = neigh_hash(p, a, iface);
+  struct iface *ifreq = iface;
+  struct ifa *addr = NULL;
 
   WALK_LIST(n, neigh_hash_table[h])	/* Search the cache */
-    if (n->proto == p && ipa_equal(*a, n->addr) && (!ifa || (ifa == n->iface)))
+    if ((n->proto == p) && ipa_equal(n->addr, a) && (n->ifreq == iface))
       return n;
 
-  class = ipa_classify(*a);
-  if (class < 0)			/* Invalid address */
-    return NULL;
-  if (((class & IADDR_SCOPE_MASK) == SCOPE_HOST) ||
-      (((class & IADDR_SCOPE_MASK) == SCOPE_LINK) && (ifa == NULL)) ||
-      !(class & IADDR_HOST))
-    return NULL;			/* Bad scope or a somecast */
-
-  if (ifa)
-    {
-      scope = if_connected(a, ifa, &addr);
-      flags |= NEF_BIND;
-
-      if ((scope < 0) && (flags & NEF_ONLINK))
-	scope = class & IADDR_SCOPE_MASK;
-    }
+  if (flags & NEF_IFACE)
+  {
+    if (ipa_nonzero(a) || !iface)
+      return NULL;
+  }
   else
-    WALK_LIST(i, iface_list)
-      if ((!p->vrf || p->vrf == i->master) &&
-	  ((scope = if_connected(a, i, &addr)) >= 0))
-	{
-	  ifa = i;
-	  break;
-	}
+  {
+    class = ipa_classify(a);
+    if (class < 0)			/* Invalid address */
+      return NULL;
+    if (((class & IADDR_SCOPE_MASK) == SCOPE_HOST) ||
+	(((class & IADDR_SCOPE_MASK) == SCOPE_LINK) && !iface) ||
+	!(class & IADDR_HOST))
+      return NULL;			/* Bad scope or a somecast */
+  }
+
+  if ((flags & NEF_ONLINK) && !iface)
+      return NULL;
+
+  if (iface)
+  {
+    scope = if_connected(a, iface, &addr, flags);
+    iface = (scope < 0) ? NULL : iface;
+  }
+  else
+    scope = if_connected_any(a, p->vrf, &iface, &addr, flags);
 
   /* scope < 0 means i don't know neighbor */
-  /* scope >= 0 implies ifa != NULL */
+  /* scope >= 0  <=>  iface != NULL */
 
   if ((scope < 0) && !(flags & NEF_STICKY))
     return NULL;
@@ -170,52 +203,15 @@ neigh_find2(struct proto *p, ip_addr *a, struct iface *ifa, unsigned flags)
   n = sl_alloc(neigh_slab);
   memset(n, 0, sizeof(neighbor));
 
-  n->addr = *a;
-  if (scope >= 0)
-    {
-      add_tail(&neigh_hash_table[h], &n->n);
-      add_tail(&ifa->neighbors, &n->if_n);
-    }
-  else
-    {
-      add_tail(&sticky_neigh_list, &n->n);
-      scope = -1;
-    }
-  n->iface = ifa;
+  add_tail(&neigh_hash_table[h], &n->n);
+  add_tail((scope >= 0) ? &iface->neighbors : &sticky_neigh_list, &n->if_n);
+  n->addr = a;
   n->ifa = addr;
+  n->iface = iface;
+  n->ifreq = ifreq;
   n->proto = p;
-  n->data = NULL;
-  n->aux = 0;
   n->flags = flags;
   n->scope = scope;
-  return n;
-}
-
-neighbor *
-neigh_find_iface(struct proto *p, struct iface *ifa)
-{
-  neighbor *n;
-  node *nn;
-
-  /* We keep neighbors with NEF_IFACE foremost in ifa->neighbors list */
-  WALK_LIST2(n, nn, ifa->neighbors, if_n)
-  {
-    if (! (n->flags & NEF_IFACE))
-      break;
-
-    if (n->proto == p)
-      return n;
-  }
-
-  n = sl_alloc(neigh_slab);
-  memset(n, 0, sizeof(neighbor));
-
-  add_tail(&iface_neigh_list, &n->n);
-  add_head(&ifa->neighbors, &n->if_n);
-  n->iface = ifa;
-  n->proto = p;
-  n->flags = NEF_IFACE;
-  n->scope = (ifa->flags & IF_UP) ? SCOPE_HOST : -1;
 
   return n;
 }
@@ -224,30 +220,26 @@ neigh_find_iface(struct proto *p, struct iface *ifa)
  * neigh_dump - dump specified neighbor entry.
  * @n: the entry to dump
  *
- * This functions dumps the contents of a given neighbor entry
- * to debug output.
+ * This functions dumps the contents of a given neighbor entry to debug output.
  */
 void
 neigh_dump(neighbor *n)
 {
-  debug("%p %I ", n, n->addr);
-  if (n->iface)
-    debug("%s ", n->iface->name);
-  else
-    debug("[] ");
+  debug("%p %I %s %s ", n, n->addr,
+	n->iface ? n->iface->name : "[]",
+	n->ifreq ? n->ifreq->name : "[]");
   debug("%s %p %08x scope %s", n->proto->name, n->data, n->aux, ip_scope_text(n->scope));
   if (n->flags & NEF_STICKY)
     debug(" STICKY");
-  if (n->flags & NEF_IFACE)
-    debug(" IFACE");
+  if (n->flags & NEF_ONLINK)
+    debug(" ONLINK");
   debug("\n");
 }
 
 /**
  * neigh_dump_all - dump all neighbor entries.
  *
- * This function dumps the contents of the neighbor cache to
- * debug output.
+ * This function dumps the contents of the neighbor cache to debug output.
  */
 void
 neigh_dump_all(void)
@@ -256,73 +248,109 @@ neigh_dump_all(void)
   int i;
 
   debug("Known neighbors:\n");
-  WALK_LIST(n, sticky_neigh_list)
-    neigh_dump(n);
-  WALK_LIST(n, iface_neigh_list)
-    neigh_dump(n);
   for(i=0; i<NEIGH_HASH_SIZE; i++)
     WALK_LIST(n, neigh_hash_table[i])
       neigh_dump(n);
   debug("\n");
 }
 
+static inline void
+neigh_notify(neighbor *n)
+{
+  if (n->proto->neigh_notify && (n->proto->proto_state != PS_STOP))
+    n->proto->neigh_notify(n);
+}
+
 static void
-neigh_up(neighbor *n, struct iface *i, int scope, struct ifa *a)
+neigh_up(neighbor *n, struct iface *i, struct ifa *a, int scope)
 {
   DBG("Waking up sticky neighbor %I\n", n->addr);
   n->iface = i;
   n->ifa = a;
   n->scope = scope;
 
-  if (! (n->flags & NEF_IFACE))
-  {
-    add_tail(&i->neighbors, &n->if_n);
-    rem_node(&n->n);
-    add_tail(&neigh_hash_table[neigh_hash(n->proto, &n->addr)], &n->n);
-  }
+  rem_node(&n->if_n);
+  add_tail(&i->neighbors, &n->if_n);
 
-  if (n->proto->neigh_notify && (n->proto->proto_state != PS_STOP))
-    n->proto->neigh_notify(n);
+  neigh_notify(n);
 }
 
 static void
 neigh_down(neighbor *n)
 {
   DBG("Flushing neighbor %I on %s\n", n->addr, n->iface->name);
-  if (! (n->flags & (NEF_BIND | NEF_IFACE)))
-    n->iface = NULL;
+  n->iface = NULL;
   n->ifa = NULL;
   n->scope = -1;
 
-  if (! (n->flags & NEF_IFACE))
+  rem_node(&n->if_n);
+  add_tail(&sticky_neigh_list, &n->if_n);
+
+  neigh_notify(n);
+}
+
+static inline void
+neigh_free(neighbor *n)
+{
+  rem_node(&n->n);
+  rem_node(&n->if_n);
+  sl_free(neigh_slab, n);
+}
+
+/**
+ * neigh_update: update neighbor entry w.r.t. change on specific iface
+ * @n: neighbor to update
+ * @iface: changed iface
+ *
+ * The function recalculates state of the neighbor entry @n assuming that only
+ * the interface @iface may changed its state or addresses. Then, appropriate
+ * actions are executed (the neighbor goes up, down, up-down, or just notified).
+ */
+void
+neigh_update(neighbor *n, struct iface *iface)
+{
+  struct ifa *ifa = NULL;
+  int scope = -1;
+
+  /* Iface-bound neighbors ignore other ifaces */
+  if (n->ifreq && (n->ifreq != iface))
+    return;
+
+  /* VRF-bound neighbors ignore changes in other VRFs */
+  if (n->proto->vrf && (n->proto->vrf != iface->master))
+    return;
+
+  scope = if_connected(n->addr, iface, &ifa, n->flags);
+
+  /* When neighbor is going down, try to respawn it on other ifaces */
+  if ((scope < 0) && (n->scope >= 0) && !n->ifreq && (n->flags & NEF_STICKY))
+    scope = if_connected_any(n->addr, n->proto->vrf, &iface, &ifa, n->flags);
+
+  /* No change or minor change - ignore or notify */
+  if ((scope == n->scope) && (iface == n->iface))
+  {
+    if (ifa != n->ifa)
     {
-      rem_node(&n->if_n);
-      rem_node(&n->n);
+      n->ifa = ifa;
+      neigh_notify(n);
     }
 
-  if (n->proto->neigh_notify && (n->proto->proto_state != PS_STOP))
-    n->proto->neigh_notify(n);
+    return;
+  }
 
-  if (n->flags & NEF_STICKY)
-    {
-      add_tail(&sticky_neigh_list, &n->n);
+  /* Major change - going down and/or going up */
 
-      /* Respawn neighbor if there is another matching prefix */
-      struct iface *i;
-      struct ifa *a;
-      int scope;
+  if (n->scope >= 0)
+    neigh_down(n);
 
-      if (!n->iface)
-	WALK_LIST(i, iface_list)
-	  if ((scope = if_connected(&n->addr, i, &a)) >= 0)
-	    {
-	      neigh_up(n, i, scope, a);
-	      return;
-	    }
-    }
+  if ((n->scope < 0) && !(n->flags & NEF_STICKY))
+  {
+    neigh_free(n);
+    return;
+  }
 
-  if (! (n->flags & (NEF_STICKY | NEF_IFACE)))
-    sl_free(neigh_slab, n);
+  if (scope >= 0)
+    neigh_up(n, iface, ifa, scope);
 }
 
 
@@ -338,21 +366,11 @@ neigh_down(neighbor *n)
 void
 neigh_if_up(struct iface *i)
 {
-  struct ifa *a;
   neighbor *n;
   node *x, *y;
-  int scope;
 
-  /* Wake up all iface neighbors */
-  WALK_LIST2_DELSAFE(n, x, y, i->neighbors, if_n)
-    if ((n->scope < 0) && (n->flags & NEF_IFACE))
-      neigh_up(n, i, SCOPE_HOST, NULL);
-
-  /* Wake up appropriate sticky neighbors */
-  WALK_LIST_DELSAFE(n, x, sticky_neigh_list)
-    if ((!n->iface || n->iface == i) &&
-	((scope = if_connected(&n->addr, i, &a)) >= 0))
-      neigh_up(n, i, scope, a);
+  WALK_LIST2_DELSAFE(n, x, y, sticky_neigh_list, if_n)
+    neigh_update(n, i);
 }
 
 /**
@@ -361,8 +379,7 @@ neigh_if_up(struct iface *i)
  *
  * Notify the neighbor cache that an interface has ceased to exist.
  *
- * It causes all entries belonging to neighbors connected to this interface
- * to be flushed.
+ * It causes all neighbors connected to this interface to be updated or removed.
  */
 void
 neigh_if_down(struct iface *i)
@@ -371,16 +388,15 @@ neigh_if_down(struct iface *i)
   node *x, *y;
 
   WALK_LIST2_DELSAFE(n, x, y, i->neighbors, if_n)
-    neigh_down(n);
+    neigh_update(n, i);
 }
 
 /**
  * neigh_if_link - notify neighbor cache about interface link change
  * @i: the interface in question
  *
- * Notify the neighbor cache that an interface changed link state.
- * All owners of neighbor entries connected to this interface are
- * notified.
+ * Notify the neighbor cache that an interface changed link state. All owners of
+ * neighbor entries connected to this interface are notified.
  */
 void
 neigh_if_link(struct iface *i)
@@ -389,8 +405,7 @@ neigh_if_link(struct iface *i)
   node *x, *y;
 
   WALK_LIST2_DELSAFE(n, x, y, i->neighbors, if_n)
-    if (n->proto->neigh_notify && (n->proto->proto_state != PS_STOP))
-      n->proto->neigh_notify(n);
+    neigh_notify(n);
 }
 
 /**
@@ -407,21 +422,16 @@ void
 neigh_ifa_update(struct ifa *a)
 {
   struct iface *i = a->iface;
-  struct ifa *aa;
-  node *x, *y;
   neighbor *n;
-  int scope;
+  node *x, *y;
 
-  /* Remove all neighbors whose scope has changed */
+  /* Update all neighbors whose scope has changed */
   WALK_LIST2_DELSAFE(n, x, y, i->neighbors, if_n)
-    if (n->ifa && (if_connected(&n->addr, i, &aa) != n->scope))
-      neigh_down(n);
+    neigh_update(n, i);
 
   /* Wake up all sticky neighbors that are reachable now */
-  WALK_LIST_DELSAFE(n, x, sticky_neigh_list)
-    if ((!n->iface || n->iface == i) &&
-	((scope = if_connected(&n->addr, i, &aa)) >= 0))
-      neigh_up(n, i, scope, aa);
+  WALK_LIST2_DELSAFE(n, x, y, sticky_neigh_list, if_n)
+    neigh_update(n, i);
 }
 
 static inline void
@@ -429,10 +439,8 @@ neigh_prune_one(neighbor *n)
 {
   if (n->proto->proto_state != PS_DOWN)
     return;
-  rem_node(&n->n);
-  if (n->if_n.next)
-    rem_node(&n->if_n);
-  sl_free(neigh_slab, n);
+
+  neigh_free(n);
 }
 
 /**
@@ -453,10 +461,6 @@ neigh_prune(void)
   for(i=0; i<NEIGH_HASH_SIZE; i++)
     WALK_LIST_DELSAFE(n, m, neigh_hash_table[i])
       neigh_prune_one(n);
-  WALK_LIST_DELSAFE(n, m, sticky_neigh_list)
-    neigh_prune_one(n);
-  WALK_LIST_DELSAFE(n, m, iface_neigh_list)
-    neigh_prune_one(n);
 }
 
 /**
@@ -471,9 +475,8 @@ neigh_init(pool *if_pool)
 {
   neigh_slab = sl_new(if_pool, sizeof(neighbor));
 
-  init_list(&sticky_neigh_list);
-  init_list(&iface_neigh_list);
-
   for(int i = 0; i < NEIGH_HASH_SIZE; i++)
     init_list(&neigh_hash_table[i]);
+
+  init_list(&sticky_neigh_list);
 }
diff --git a/nest/rt-table.c b/nest/rt-table.c
index de7b05fc..0c0e365e 100644
--- a/nest/rt-table.c
+++ b/nest/rt-table.c
@@ -39,6 +39,7 @@
 #include "lib/string.h"
 #include "conf/conf.h"
 #include "filter/filter.h"
+#include "lib/hash.h"
 #include "lib/string.h"
 #include "lib/alloca.h"
 
@@ -2220,13 +2221,6 @@ rt_feed_channel_abort(struct channel *c)
     }
 }
 
-static inline unsigned
-ptr_hash(void *ptr)
-{
-  uintptr_t p = (uintptr_t) ptr;
-  return p ^ (p << 8) ^ (p >> 16);
-}
-
 static inline u32
 hc_hash(ip_addr a, rtable *dep)
 {
diff --git a/proto/bfd/bfd.c b/proto/bfd/bfd.c
index 64d5007b..fdcd7225 100644
--- a/proto/bfd/bfd.c
+++ b/proto/bfd/bfd.c
@@ -775,7 +775,7 @@ bfd_start_neighbor(struct bfd_proto *p, struct bfd_neighbor *n)
     return;
   }
 
-  struct neighbor *nb = neigh_find2(&p->p, &n->addr, n->iface, NEF_STICKY);
+  struct neighbor *nb = neigh_find(&p->p, n->addr, n->iface, NEF_STICKY);
   if (!nb)
   {
     log(L_ERR "%s: Invalid remote address %I%J", p->p.name, n->addr, n->iface);
diff --git a/proto/bgp/bgp.c b/proto/bgp/bgp.c
index 932ad9f3..ced83c5c 100644
--- a/proto/bgp/bgp.c
+++ b/proto/bgp/bgp.c
@@ -1277,7 +1277,7 @@ bgp_start_locked(struct object_lock *lock)
     return;
   }
 
-  neighbor *n = neigh_find2(&p->p, &cf->remote_ip, cf->iface, NEF_STICKY);
+  neighbor *n = neigh_find(&p->p, cf->remote_ip, cf->iface, NEF_STICKY);
   if (!n)
   {
     log(L_ERR "%s: Invalid remote address %I%J", p->p.name, cf->remote_ip, cf->iface);
@@ -1521,7 +1521,7 @@ bgp_channel_start(struct channel *C)
   if (ipa_zero(c->next_hop_addr))
   {
     /* We know the iface for single-hop, we make lookup for multihop */
-    struct neighbor *nbr = p->neigh ?: neigh_find2(&p->p, &src, NULL, 0);
+    struct neighbor *nbr = p->neigh ?: neigh_find(&p->p, src, NULL, 0);
     struct iface *iface = nbr ? nbr->iface : NULL;
 
     if (bgp_channel_is_ipv4(c) && iface && iface->addr4)
diff --git a/proto/bgp/packets.c b/proto/bgp/packets.c
index aa08732d..190fd6fe 100644
--- a/proto/bgp/packets.c
+++ b/proto/bgp/packets.c
@@ -744,9 +744,9 @@ bgp_apply_next_hop(struct bgp_parse_state *s, rta *a, ip_addr gw, ip_addr ll)
 
     /* GW_DIRECT -> single_hop -> p->neigh != NULL */
     if (ipa_nonzero(gw))
-      nbr = neigh_find2(&p->p, &gw, NULL, 0);
+      nbr = neigh_find(&p->p, gw, NULL, 0);
     else if (ipa_nonzero(ll))
-      nbr = neigh_find2(&p->p, &ll, p->neigh->iface, 0);
+      nbr = neigh_find(&p->p, ll, p->neigh->iface, 0);
 
     if (!nbr || (nbr->scope == SCOPE_HOST))
       WITHDRAW(BAD_NEXT_HOP);
diff --git a/proto/ospf/rt.c b/proto/ospf/rt.c
index b4f9b067..4549ce2a 100644
--- a/proto/ospf/rt.c
+++ b/proto/ospf/rt.c
@@ -1987,7 +1987,7 @@ again1:
       for (nh = nf->n.nhs; nh; nh = nh->next)
 	if (ipa_nonzero(nh->gw))
 	{
-	  neighbor *ng = neigh_find2(&p->p, &nh->gw, nh->iface, 0);
+	  neighbor *ng = neigh_find(&p->p, nh->gw, nh->iface, 0);
 	  if (!ng || (ng->scope == SCOPE_HOST))
 	    { reset_ri(nf); break; }
 	}
diff --git a/proto/rip/packets.c b/proto/rip/packets.c
index 891f454f..1b65362f 100644
--- a/proto/rip/packets.c
+++ b/proto/rip/packets.c
@@ -627,7 +627,7 @@ rip_receive_response(struct rip_proto *p, struct rip_iface *ifa, struct rip_pack
 
     if (ipa_nonzero(rte.next_hop))
     {
-      neighbor *nbr = neigh_find2(&p->p, &rte.next_hop, ifa->iface, 0);
+      neighbor *nbr = neigh_find(&p->p, rte.next_hop, ifa->iface, 0);
       if (!nbr || (nbr->scope <= 0))
 	rte.next_hop = IPA_NONE;
     }
diff --git a/proto/rip/rip.c b/proto/rip/rip.c
index 90bf8e5c..2838d425 100644
--- a/proto/rip/rip.c
+++ b/proto/rip/rip.c
@@ -377,7 +377,7 @@ rip_rt_notify(struct proto *P, struct channel *ch UNUSED, struct network *net, s
 struct rip_neighbor *
 rip_get_neighbor(struct rip_proto *p, ip_addr *a, struct rip_iface *ifa)
 {
-  neighbor *nbr = neigh_find2(&p->p, a, ifa->iface, 0);
+  neighbor *nbr = neigh_find(&p->p, *a, ifa->iface, 0);
 
   if (!nbr || (nbr->scope == SCOPE_HOST) || !rip_iface_link_up(ifa))
     return NULL;
diff --git a/proto/static/static.c b/proto/static/static.c
index 8dfa6f35..40096c16 100644
--- a/proto/static/static.c
+++ b/proto/static/static.c
@@ -205,10 +205,9 @@ static_add_rte(struct static_proto *p, struct static_route *r)
 
     for (r2 = r; r2; r2 = r2->mp_next)
     {
-      n = ipa_nonzero(r2->via) ?
-	neigh_find2(&p->p, &r2->via, r2->iface,
-		    NEF_STICKY | (r2->onlink ? NEF_ONLINK : 0)) :
-	neigh_find_iface(&p->p, r2->iface);
+      n = neigh_find(&p->p, r2->via, r2->iface, NEF_STICKY |
+		     (r2->onlink ? NEF_ONLINK : 0) |
+		     (ipa_zero(r2->via) ? NEF_IFACE : 0));
 
       if (!n)
       {
diff --git a/sysdep/linux/netlink.c b/sysdep/linux/netlink.c
index 73f77147..834504d0 100644
--- a/sysdep/linux/netlink.c
+++ b/sysdep/linux/netlink.c
@@ -696,8 +696,8 @@ nl_parse_multipath(struct nl_parse_state *s, struct krt_proto *p, struct rtattr
 	    rv->flags |= RNF_ONLINK;
 
 	  neighbor *nbr;
-	  nbr = neigh_find2(&p->p, &rv->gw, rv->iface,
-			    (rv->flags & RNF_ONLINK) ? NEF_ONLINK : 0);
+	  nbr = neigh_find(&p->p, rv->gw, rv->iface,
+			   (rv->flags & RNF_ONLINK) ? NEF_ONLINK : 0);
 	  if (!nbr || (nbr->scope == SCOPE_HOST))
 	    return NULL;
 	}
@@ -1636,8 +1636,8 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h)
 	    ra->nh.flags |= RNF_ONLINK;
 
 	  neighbor *nbr;
-	  nbr = neigh_find2(&p->p, &(ra->nh.gw), ra->nh.iface,
-			    (ra->nh.flags & RNF_ONLINK) ? NEF_ONLINK : 0);
+	  nbr = neigh_find(&p->p, ra->nh.gw, ra->nh.iface,
+			   (ra->nh.flags & RNF_ONLINK) ? NEF_ONLINK : 0);
 	  if (!nbr || (nbr->scope == SCOPE_HOST))
 	    {
 	      log(L_ERR "KRT: Received route %N with strange next-hop %I", net->n.addr,

From d8e816c150ebad08ff75ef34eb459a67a09a34d0 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Wed, 27 Jun 2018 17:08:59 +0200
Subject: [PATCH 181/198] BSD: Fix of the previous commit

---
 sysdep/bsd/krt-sock.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sysdep/bsd/krt-sock.c b/sysdep/bsd/krt-sock.c
index 8522e415..e646c414 100644
--- a/sysdep/bsd/krt-sock.c
+++ b/sysdep/bsd/krt-sock.c
@@ -532,7 +532,7 @@ krt_read_route(struct ks_msg *msg, struct krt_proto *p, int scan)
     if (ipa_is_link_local(a.nh.gw))
       _I0(a.nh.gw) = 0xfe800000;
 
-    ng = neigh_find2(&p->p, &a.nh.gw, a.nh.iface, 0);
+    ng = neigh_find(&p->p, a.nh.gw, a.nh.iface, 0);
     if (!ng || (ng->scope == SCOPE_HOST))
       {
 	/* Ignore routes with next-hop 127.0.0.1, host routes with such

From daf113ac66fb9dc83a904ce9bcc68e90830624d0 Mon Sep 17 00:00:00 2001
From: Maria Jan Matejka <mq@jmq.cz>
Date: Fri, 29 Jun 2018 17:23:27 +0200
Subject: [PATCH 182/198] BGP: Attribute set function merged with its common
 counterpart

---
 proto/bgp/attrs.c | 26 ++++++++------------------
 1 file changed, 8 insertions(+), 18 deletions(-)

diff --git a/proto/bgp/attrs.c b/proto/bgp/attrs.c
index 8ecfaaa2..6580db57 100644
--- a/proto/bgp/attrs.c
+++ b/proto/bgp/attrs.c
@@ -84,24 +84,14 @@ bgp_set_attr(ea_list **attrs, struct linpool *pool, uint code, uint flags, uintp
 {
   ASSERT(bgp_attr_known(code));
 
-  ea_list *a = lp_alloc(pool, sizeof(ea_list) + sizeof(eattr));
-  eattr *e = &a->attrs[0];
-
-  a->flags = EALF_SORTED;
-  a->count = 1;
-  a->next = *attrs;
-  *attrs = a;
-
-  e->id = EA_CODE(PROTOCOL_BGP, code);
-  e->type = bgp_attr_table[code].type;
-  e->flags = flags;
-
-  if (e->type & EAF_EMBEDDED)
-    e->u.data = (u32) val;
-  else
-    e->u.ptr = (struct adata *) val;
-
-  return e;
+  return ea_set_attr(
+      attrs,
+      pool,
+      EA_CODE(PROTOCOL_BGP, code),
+      flags,
+      bgp_attr_table[code].type,
+      val
+  );
 }
 
 

From 93c1defdb03729ae8b41752d2155dff65964df5b Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Sun, 1 Jul 2018 00:43:24 +0200
Subject: [PATCH 183/198] BGP: Fix parsing of MPLS withdrawals

RFC 3107 was bit vague with regard to labeled withdrawals, RFC 8277
clarified that. The old code was incompatible with some implementations,
namely with Juniper.

Thanks to Vadim Fedorenko for the original patch.
---
 proto/bgp/packets.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/proto/bgp/packets.c b/proto/bgp/packets.c
index 190fd6fe..d67ef0b2 100644
--- a/proto/bgp/packets.c
+++ b/proto/bgp/packets.c
@@ -1165,9 +1165,10 @@ bgp_decode_mpls_labels(struct bgp_parse_state *s, byte **pos, uint *len, uint *p
     ADVANCE(*pos, *len, 3);
     *pxlen -= 24;
 
-    /* Withdraw: Magic label stack value 0x800000 according to RFC 3107, section 3, last paragraph */
-    if (!a && !s->err_withdraw && (lnum == 1) && (label == BGP_MPLS_MAGIC))
-      break;
+    /* RFC 8277 2.4 - withdraw does not have variable-size MPLS stack but
+       fixed-size 24-bit Compatibility field, which MUST be ignored */
+    if (!a && !s->err_withdraw)
+      return;
   }
   while (!(label & BGP_MPLS_BOS));
 

From 8e86ffce8251f4e48f61b6d8e89966d037ef8e59 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Sun, 1 Jul 2018 01:03:16 +0200
Subject: [PATCH 184/198] BGP: Use implicit-NULL label when announcing MPLS
 routes with local next-hop

We currently cannot assing local labels, but we can still be LSP egress
router. Therefore when we announce labeled route with local next-hop, we
should announce implicit-NULL label instead of rejecting it completely.
---
 proto/bgp/packets.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/proto/bgp/packets.c b/proto/bgp/packets.c
index d67ef0b2..c40b8ec2 100644
--- a/proto/bgp/packets.c
+++ b/proto/bgp/packets.c
@@ -875,7 +875,10 @@ bgp_update_next_hop_ip(struct bgp_export_state *s, eattr *a, ea_list **to)
 
       /* TODO: Use local MPLS assigned label */
       if (s->mpls)
-	bgp_unset_attr(to, s->pool, BA_MPLS_LABEL_STACK);
+      {
+	u32 implicit_null = BGP_MPLS_NULL;
+	bgp_set_attr_data(to, s->pool, BA_MPLS_LABEL_STACK, 0, &implicit_null, 4);
+      }
     }
   }
 

From a81e18da254ddd7cccff82feab61aa943a277805 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 3 Jul 2018 17:52:51 +0200
Subject: [PATCH 185/198] Nest: Fix race condition during reconfiguration

If export filter is changed during reconfiguration and a route disappears
between reconfiguration and refeed (e.g., if the route is a static route
also removed during the reconfiguration), the route is not withdrawn.
The patch fixes that by adding tx reconfiguration timestamp.
---
 nest/proto.c    |  4 ++++
 nest/protocol.h |  1 +
 nest/rt-table.c | 30 +++++++++++++++---------------
 3 files changed, 20 insertions(+), 15 deletions(-)

diff --git a/nest/proto.c b/nest/proto.c
index 49f71304..8a8221a8 100644
--- a/nest/proto.c
+++ b/nest/proto.c
@@ -165,6 +165,7 @@ proto_add_channel(struct proto *p, struct channel_config *cf)
   c->channel_state = CS_DOWN;
   c->export_state = ES_DOWN;
   c->last_state_change = current_time();
+  c->last_tx_filter_change = current_time();
   c->reloadable = 1;
 
   CALL(c->channel->init, c, cf);
@@ -557,6 +558,9 @@ channel_reconfigure(struct channel *c, struct channel_config *cf)
 
   channel_verify_limits(c);
 
+  if (export_changed)
+    c->last_tx_filter_change = current_time();
+
   /* Execute channel-specific reconfigure hook */
   if (c->channel->reconfigure && !c->channel->reconfigure(c, cf))
     return 0;
diff --git a/nest/protocol.h b/nest/protocol.h
index dd942c10..f0b65598 100644
--- a/nest/protocol.h
+++ b/nest/protocol.h
@@ -536,6 +536,7 @@ struct channel {
   u8 gr_wait;				/* Route export to channel is postponed until graceful restart */
 
   btime last_state_change;		/* Time of last state transition */
+  btime last_tx_filter_change;
 };
 
 
diff --git a/nest/rt-table.c b/nest/rt-table.c
index 0c0e365e..9ce52428 100644
--- a/nest/rt-table.c
+++ b/nest/rt-table.c
@@ -549,29 +549,29 @@ rt_notify_basic(struct channel *c, net *net, rte *new0, rte *old0, int refeed)
     c->stats.exp_withdraws_received++;
 
   /*
-   * This is a tricky part - we don't know whether route 'old' was
-   * exported to protocol 'p' or was filtered by the export filter.
-   * We try to run the export filter to know this to have a correct
-   * value in 'old' argument of rte_update (and proper filter value)
+   * This is a tricky part - we don't know whether route 'old' was exported to
+   * protocol 'p' or was filtered by the export filter. We try to run the export
+   * filter to know this to have a correct value in 'old' argument of rte_update
+   * (and proper filter value).
    *
-   * FIXME - this is broken because 'configure soft' may change
-   * filters but keep routes. Refeed is expected to be called after
-   * change of the filters and with old == new, therefore we do not
-   * even try to run the filter on an old route, This may lead to
-   * 'spurious withdraws' but ensure that there are no 'missing
+   * This is broken because 'configure soft' may change filters but keep routes.
+   * Refeed cycle is expected to be called after change of the filters and with
+   * old == new, therefore we do not even try to run the filter on an old route.
+   * This may lead to 'spurious withdraws' but ensure that there are no 'missing
    * withdraws'.
    *
-   * This is not completely safe as there is a window between
-   * reconfiguration and the end of refeed - if a newly filtered
-   * route disappears during this period, proper withdraw is not
-   * sent (because old would be also filtered) and the route is
-   * not refeeded (because it disappeared before that).
+   * This is not completely safe as there is a window between reconfiguration
+   * and the end of refeed - if a newly filtered route disappears during this
+   * period, proper withdraw is not sent (because old would be also filtered)
+   * and the route is not refeeded (because it disappeared before that).
+   * Therefore, we also do not try to run the filter on old routes that are
+   * older than the last filter change.
    */
 
   if (new)
     new = export_filter(c, new, &new_free, 0);
 
-  if (old && !refeed)
+  if (old && !(refeed || (old->lastmod <= c->last_tx_filter_change)))
     old = export_filter(c, old, &old_free, 1);
 
   if (!new && !old)

From 7b9b0c0a0087def6a3fc11824a891be5940a257b Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 3 Jul 2018 18:08:35 +0200
Subject: [PATCH 186/198] Cleanup some warnings

---
 proto/babel/babel.c | 3 +--
 proto/ospf/ospf.c   | 2 +-
 2 files changed, 2 insertions(+), 3 deletions(-)

diff --git a/proto/babel/babel.c b/proto/babel/babel.c
index b4404f45..afa482bb 100644
--- a/proto/babel/babel.c
+++ b/proto/babel/babel.c
@@ -2091,9 +2091,8 @@ babel_prepare_attrs(struct linpool *pool, ea_list *next, uint metric, u64 router
 
 
 static int
-babel_import_control(struct proto *P, struct rte **new, struct linpool *pool)
+babel_import_control(struct proto *P, struct rte **new, struct linpool *pool UNUSED)
 {
-  struct babel_proto *p = (void *) P;
   struct rta *a = (*new)->attrs;
 
   /* Reject our own unreachable routes */
diff --git a/proto/ospf/ospf.c b/proto/ospf/ospf.c
index fa122f01..150f519f 100644
--- a/proto/ospf/ospf.c
+++ b/proto/ospf/ospf.c
@@ -446,7 +446,7 @@ ospf_disp(timer * timer)
  * import to the filters.
  */
 static int
-ospf_import_control(struct proto *P, rte **new, struct linpool *pool)
+ospf_import_control(struct proto *P, rte **new, struct linpool *pool UNUSED)
 {
   struct ospf_proto *p = (struct ospf_proto *) P;
   struct ospf_area *oa = ospf_main_area(p);

From 092c4930277c0f0f0dcb0c4013ff642314f90842 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Fri, 6 Jul 2018 02:04:45 +0200
Subject: [PATCH 187/198] Nest: Fix race condition during reconfiguration, part
 2

If export filter is changed during reconfiguration and a route disappears
between reconfiguration and refeed (e.g., if the route is a static route
also removed during the reconfiguration), the route is not withdrawn.

The issue was fixed for regular channels by an earlier patch. This patch
fixes the issue for channels in RA_ACCEPTED mode (first-pass-the-filter),
used by BGP with 'secondary' option.
---
 nest/rt-table.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/nest/rt-table.c b/nest/rt-table.c
index 9ce52428..9dcbda05 100644
--- a/nest/rt-table.c
+++ b/nest/rt-table.c
@@ -678,8 +678,18 @@ rt_notify_accepted(struct channel *c, net *net, rte *new_changed, rte *old_chang
    *
    * - We found new_best the same as new_changed, therefore it cannot
    *   be old_best and we have to continue search for old_best.
+   *
+   * There is also a hack to ensure consistency in case of changed filters.
+   * It does not find the proper old_best, just selects a non-NULL route.
    */
 
+  /* Hack for changed filters */
+  if (old_changed && (old_changed->lastmod <= c->last_tx_filter_change))
+    {
+      old_best = old_changed;
+      goto found;
+    }
+
   /* First case */
   if (old_meet)
     if (old_best = export_filter(c, old_changed, &old_free, 1))

From 8bd718b3ba34cee95a5443f3d20d6f16c2d4c946 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Thu, 19 Jul 2018 20:48:13 +0200
Subject: [PATCH 188/198] OSPF: Fix crash during route removal

The bug was introduced by an earler patch which removed additional eattr
argument to rt_notify hook.
---
 proto/ospf/topology.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

diff --git a/proto/ospf/topology.c b/proto/ospf/topology.c
index 54b255c9..03b77349 100644
--- a/proto/ospf/topology.c
+++ b/proto/ospf/topology.c
@@ -1248,7 +1248,6 @@ ospf_rt_notify(struct proto *P, struct channel *ch UNUSED, net *n, rte *new, rte
   struct ospf_proto *p = (struct ospf_proto *) P;
   struct ospf_area *oa = NULL;	/* non-NULL for NSSA-LSA */
   ort *nf;
-  struct ea_list *ea = new->attrs->eattrs;
 
   /*
    * There are several posibilities:
@@ -1282,8 +1281,8 @@ ospf_rt_notify(struct proto *P, struct channel *ch UNUSED, net *n, rte *new, rte
 
   /* Get route attributes */
   rta *a = new->attrs;
-  eattr *m1a = ea_find(ea, EA_OSPF_METRIC1);
-  eattr *m2a = ea_find(ea, EA_OSPF_METRIC2);
+  eattr *m1a = ea_find(a->eattrs, EA_OSPF_METRIC1);
+  eattr *m2a = ea_find(a->eattrs, EA_OSPF_METRIC2);
   uint m1 = m1a ? m1a->u.data : 0;
   uint m2 = m2a ? m2a->u.data : 10000;
 
@@ -1307,7 +1306,7 @@ ospf_rt_notify(struct proto *P, struct channel *ch UNUSED, net *n, rte *new, rte
 
   uint ebit = m2a || !m1a;
   uint metric = ebit ? m2 : m1;
-  uint tag = ea_get_int(ea, EA_OSPF_TAG, 0);
+  uint tag = ea_get_int(a->eattrs, EA_OSPF_TAG, 0);
 
   ip_addr fwd = IPA_NONE;
   if ((a->dest == RTD_UNICAST) && use_gw_for_fwaddr(p, a->nh.gw, a->nh.iface))

From 0ed3129f6b0a80afea877340d940e45f1a5d3000 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Thu, 19 Jul 2018 20:54:10 +0200
Subject: [PATCH 189/198] RAdv: Fix crash during prefix change

Thanks to Julian Schuh for the bugreport.
---
 proto/radv/radv.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/proto/radv/radv.c b/proto/radv/radv.c
index ce88c7cc..389f598c 100644
--- a/proto/radv/radv.c
+++ b/proto/radv/radv.c
@@ -174,7 +174,7 @@ radv_prepare_prefixes(struct radv_iface *ifa)
     if (pfx->valid && !pfx->mark)
     {
       RADV_TRACE(D_EVENTS, "Invalidating prefix %N on %s",
-		 pfx->prefix, ifa->iface->name);
+		 &pfx->prefix, ifa->iface->name);
 
       pfx->valid = 0;
       pfx->changed = now;
@@ -201,7 +201,7 @@ radv_prune_prefixes(struct radv_iface *ifa)
       if (expires <= now)
       {
 	RADV_TRACE(D_EVENTS, "Removing prefix %N on %s",
-		   px->prefix, ifa->iface->name);
+		   &px->prefix, ifa->iface->name);
 
 	rem_node(NODE px);
 	mb_free(px);

From 318acb0f6cb77a32aad5d7f79e06f3c5065ac702 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Sat, 28 Jul 2018 16:54:06 +0200
Subject: [PATCH 190/198] BSD: Use MSG_DONTROUTE for unicast packets on FreeBSD

BSD systems cannot use SO_DONTROUTE, because it does not work properly
with multicast packets (perhaps it tries to find iface based on multicast
group address). But we can use MSG_DONTROUTE sendmsg() flag for unicast
packets. Works on FreeBSD, is ignored on OpenBSD and is broken on NetBSD
(i guess due to integrated routing table and ARP table).
---
 lib/ip.h           | 3 +++
 sysdep/bsd/sysio.h | 5 +++++
 sysdep/cf/README   | 1 +
 sysdep/unix/io.c   | 8 +++++++-
 4 files changed, 16 insertions(+), 1 deletion(-)

diff --git a/lib/ip.h b/lib/ip.h
index 5cfce1f1..c3f3e905 100644
--- a/lib/ip.h
+++ b/lib/ip.h
@@ -241,6 +241,9 @@ static inline int ip6_is_v4mapped(ip6_addr a)
 #define ipa_classify(x) ip6_classify(&(x))
 #define ipa_is_link_local(x) ip6_is_link_local(x)
 
+static inline int ip4_is_unicast(ip4_addr a)
+{ return _I(a) < 0xe0000000; }
+
 /* XXXX remove */
 static inline int ipa_classify_net(ip_addr a)
 { return ipa_zero2(a) ? (IADDR_HOST | SCOPE_UNIVERSE) : ipa_classify(a); }
diff --git a/sysdep/bsd/sysio.h b/sysdep/bsd/sysio.h
index 68296e65..a5d161ba 100644
--- a/sysdep/bsd/sysio.h
+++ b/sysdep/bsd/sysio.h
@@ -12,6 +12,11 @@
 #include <sys/param.h>
 
 
+#ifdef __FreeBSD__
+/* Should be defined in sysdep/cf/bsd.h, but it is flavor-specific */
+#define CONFIG_DONTROUTE_UNICAST
+#endif
+
 #ifdef __NetBSD__
 
 #ifndef IP_RECVTTL
diff --git a/sysdep/cf/README b/sysdep/cf/README
index e62c3481..9a7a4afa 100644
--- a/sysdep/cf/README
+++ b/sysdep/cf/README
@@ -11,6 +11,7 @@ CONFIG_MC_PROPER_SRC	Multicast packets have source address according to socket s
 CONFIG_SKIP_MC_BIND	Don't call bind on multicast socket (def for *BSD)
 CONFIG_NO_IFACE_BIND	Bind to iface is not available, use workarounds (def for *BSD)
 CONFIG_UNIX_DONTROUTE	Use setsockopts DONTROUTE (undef for *BSD)
+CONFIG_DONTROUTE_UNICAST Use MSG_DONTROUTE flag for unicast packets (def for FreeBSD)
 CONFIG_USE_HDRINCL	Use IP_HDRINCL instead of control messages for source address on raw IP sockets.
 
 CONFIG_RESTRICTED_PRIVILEGES	Implements restricted privileges using drop_uid()
diff --git a/sysdep/unix/io.c b/sysdep/unix/io.c
index 11a0d6f1..a960b7f8 100644
--- a/sysdep/unix/io.c
+++ b/sysdep/unix/io.c
@@ -1541,6 +1541,7 @@ sk_sendmsg(sock *s)
   struct iovec iov = {s->tbuf, s->tpos - s->tbuf};
   byte cmsg_buf[CMSG_TX_SPACE];
   sockaddr dst;
+  int flags = 0;
 
   sockaddr_fill(&dst, s->af, s->daddr, s->iface, s->dport);
 
@@ -1551,6 +1552,11 @@ sk_sendmsg(sock *s)
     .msg_iovlen = 1
   };
 
+#ifdef CONFIG_DONTROUTE_UNICAST
+  if (ipa_is_ip4(s->daddr) && ip4_is_unicast(ipa_to_ip4(s->daddr)))
+    flags = MSG_DONTROUTE;
+#endif
+
 #ifdef CONFIG_USE_HDRINCL
   byte hdr[20];
   struct iovec iov2[2] = { {hdr, 20}, iov };
@@ -1566,7 +1572,7 @@ sk_sendmsg(sock *s)
   if (s->flags & SKF_PKTINFO)
     sk_prepare_cmsgs(s, &msg, cmsg_buf, sizeof(cmsg_buf));
 
-  return sendmsg(s->fd, &msg, 0);
+  return sendmsg(s->fd, &msg, flags);
 }
 
 static inline int

From 5bd734317c05008a66eefaa14fc98a6d533cf9ef Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 31 Jul 2018 18:40:38 +0200
Subject: [PATCH 191/198] BGP: Long-lived graceful restart

The patch implements long-lived graceful restart for BGP, namely
draft-uttaro-idr-bgp-persistence-03.
---
 doc/bird.sgml       |  42 +++++++++--
 nest/protocol.h     |   1 +
 nest/route.h        |   3 +
 nest/rt-table.c     |  59 +++++++++++++++
 proto/bgp/attrs.c   |  55 +++++++++++++-
 proto/bgp/bgp.c     | 172 +++++++++++++++++++++++++++++++++++++++-----
 proto/bgp/bgp.h     |  34 ++++++++-
 proto/bgp/config.Y  |  13 +++-
 proto/bgp/packets.c |  75 ++++++++++++++++++-
 proto/pipe/pipe.c   |   6 ++
 10 files changed, 432 insertions(+), 28 deletions(-)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index b4d9ea4e..44fef1c5 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -2220,13 +2220,16 @@ using the following configuration parameters:
 	immediately shut down. Note that this option cannot be used with
 	multihop BGP. Default: enabled for direct BGP, disabled otherwise.
 
-	<tag><label id="bgp-bfd">bfd <M>switch</M></tag>
+	<tag><label id="bgp-bfd">bfd <M>switch</M>|graceful</tag>
 	BGP could use BFD protocol as an advisory mechanism for neighbor
 	liveness and failure detection. If enabled, BIRD setups a BFD session
 	for the BGP neighbor and tracks its liveness by it. This has an
 	advantage of an order of magnitude lower detection times in case of
-	failure. Note that BFD protocol also has to be configured, see
-	<ref id="bfd" name="BFD"> section for details. Default: disabled.
+	failure. When a neighbor failure is detected, the BGP session is
+	restarted. Optionally, it can be configured (by <cf/graceful/ argument)
+	to trigger graceful restart instead of regular restart. Note that BFD
+	protocol also has to be configured, see <ref id="bfd" name="BFD">
+	section for details. Default: disabled.
 
 	<tag><label id="bgp-ttl-security">ttl security <m/switch/</tag>
 	Use GTSM (<rfc id="5082"> - the generalized TTL security mechanism). GTSM
@@ -2348,6 +2351,25 @@ using the following configuration parameters:
 	re-establish after a restart before deleting stale routes. Default:
 	120 seconds.
 
+	<tag><label id="bgp-long-lived-graceful-restart">long lived graceful restart <m/switch/|aware</tag>
+	The long-lived graceful restart is an extension of the traditional
+	<ref id="bgp-graceful-restart" name="BGP graceful restart">, where stale
+	routes are kept even after the <ref id="bgp-graceful-restart-time"
+	name="restart time"> expires for additional long-lived stale time, but
+	they are marked with the LLGR_STALE community, depreferenced, and
+	withdrawn from routers not supporting LLGR. Like traditional BGP
+	graceful restart, it has three states: disabled, aware (receiving-only),
+	and enabled. Note that long-lived graceful restart requires at least
+	aware level of traditional BGP graceful restart. Default: aware, unless
+	graceful restart is disabled.
+
+	<tag><label id="bgp-long-lived-stale-time">long lived stale time <m/number/</tag>
+	The long-lived stale time is announced in the BGP long-lived graceful
+	restart capability and specifies how long the neighbor would keep stale
+	routes depreferenced during long-lived graceful restart until either the
+	session is re-stablished and synchronized or the stale time expires and
+	routes are removed. Default: 3600 seconds.
+
 	<tag><label id="bgp-interpret-communities">interpret communities <m/switch/</tag>
 	<rfc id="1997"> demands that BGP speaker should process well-known
 	communities like no-export (65535, 65281) or no-advertise (65535,
@@ -2607,6 +2629,19 @@ be used in explicit configuration.
 	configure restarting role per AFI/SAFI pair by this channel option.
 	The option is ignored if graceful restart is disabled by protocol-wide
 	option. Default: off in aware mode, on in full mode.
+
+	<tag><label id="bgp-long-lived-graceful-restart-c">long lived graceful restart <m/switch/</tag>
+	BGP long-lived graceful restart is configured mainly by protocol-wide
+	<ref id="bgp-long-lived-graceful-restart" name="options">, but the
+	restarting role can be set per AFI/SAFI pair by this channel option.
+	The option is ignored if long-lived graceful restart is disabled by
+	protocol-wide option. Default: off in aware mode, on in full mode.
+
+	<tag><label id="bgp-long-lived-stale-time-c">long lived stale time <m/number/</tag>
+	Like previous graceful restart channel options, this option allows to
+	set <ref id="bgp-long-lived-stale-time" name="long lived stale time">
+	per AFI/SAFI pair instead of per protocol. Default: set by protocol-wide
+	option.
 </descrip>
 
 <sect1>Attributes
@@ -2761,7 +2796,6 @@ interfaces to be defined for them to work with.
 	so the default time is set to a large value.
 
 	<tag><label id="device-iface">interface <m/pattern/ [, <m/.../]</tag>
-
 	By default, the Device protocol handles all interfaces without any
 	configuration. Interface definitions allow to specify optional
 	parameters for specific interfaces. See <ref id="proto-iface"
diff --git a/nest/protocol.h b/nest/protocol.h
index f0b65598..343ce523 100644
--- a/nest/protocol.h
+++ b/nest/protocol.h
@@ -229,6 +229,7 @@ struct proto {
   int (*rte_better)(struct rte *, struct rte *);
   int (*rte_same)(struct rte *, struct rte *);
   int (*rte_mergable)(struct rte *, struct rte *);
+  struct rte * (*rte_modify)(struct rte *, struct linpool *);
   void (*rte_insert)(struct network *, struct rte *);
   void (*rte_remove)(struct network *, struct rte *);
 
diff --git a/nest/route.h b/nest/route.h
index cad15440..2ec1dc7f 100644
--- a/nest/route.h
+++ b/nest/route.h
@@ -231,6 +231,7 @@ typedef struct rte {
 #ifdef CONFIG_BGP
     struct {
       u8 suppressed;			/* Used for deterministic MED comparison */
+      s8 stale;				/* Route is LLGR_STALE, -1 if unknown */
     } bgp;
 #endif
 #ifdef CONFIG_BABEL
@@ -254,6 +255,7 @@ typedef struct rte {
 #define REF_FILTERED	2		/* Route is rejected by import filter */
 #define REF_STALE	4		/* Route is stale in a refresh cycle */
 #define REF_DISCARD	8		/* Route is scheduled for discard */
+#define REF_MODIFY	16		/* Route is scheduled for modify */
 
 /* Route is valid for propagation (may depend on other flags in the future), accepts NULL */
 static inline int rte_is_valid(rte *r) { return r && !(r->flags & REF_FILTERED); }
@@ -297,6 +299,7 @@ int rt_examine(rtable *t, net_addr *a, struct proto *p, struct filter *filter);
 rte *rt_export_merged(struct channel *c, net *net, rte **rt_free, linpool *pool, int silent);
 void rt_refresh_begin(rtable *t, struct channel *c);
 void rt_refresh_end(rtable *t, struct channel *c);
+void rt_modify_stale(rtable *t, struct channel *c);
 void rt_schedule_prune(rtable *t);
 void rte_dump(rte *);
 void rte_free(rte *);
diff --git a/nest/rt-table.c b/nest/rt-table.c
index 9dcbda05..036b34c4 100644
--- a/nest/rt-table.c
+++ b/nest/rt-table.c
@@ -1437,6 +1437,28 @@ rte_discard(rte *old)	/* Non-filtered route deletion, used during garbage collec
   rte_update_unlock();
 }
 
+/* Modify existing route by protocol hook, used for long-lived graceful restart */
+static inline void
+rte_modify(rte *old)
+{
+  rte_update_lock();
+
+  rte *new = old->sender->proto->rte_modify(old, rte_update_pool);
+  if (new != old)
+  {
+    if (new)
+    {
+      if (!rta_is_cached(new->attrs))
+	new->attrs = rta_lookup(new->attrs);
+      new->flags = (old->flags & ~REF_MODIFY) | REF_COW;
+    }
+
+    rte_recalculate(old->sender, old->net, new, old->attrs->src);
+  }
+
+  rte_update_unlock();
+}
+
 /* Check rtable for best route to given net whether it would be exported do p */
 int
 rt_examine(rtable *t, net_addr *a, struct proto *p, struct filter *filter)
@@ -1521,6 +1543,26 @@ rt_refresh_end(rtable *t, struct channel *c)
     rt_schedule_prune(t);
 }
 
+void
+rt_modify_stale(rtable *t, struct channel *c)
+{
+  int prune = 0;
+
+  FIB_WALK(&t->fib, net, n)
+    {
+      rte *e;
+      for (e = n->routes; e; e = e->next)
+	if ((e->sender == c) && (e->flags & REF_STALE) && !(e->flags & REF_FILTERED))
+	  {
+	    e->flags |= REF_MODIFY;
+	    prune = 1;
+	  }
+    }
+  FIB_WALK_END;
+
+  if (prune)
+    rt_schedule_prune(t);
+}
 
 /**
  * rte_dump - dump a route
@@ -1712,6 +1754,7 @@ again:
 
     rescan:
       for (e=n->routes; e; e=e->next)
+      {
 	if (e->sender->flush_active || (e->flags & REF_DISCARD))
 	  {
 	    if (limit <= 0)
@@ -1727,6 +1770,22 @@ again:
 	    goto rescan;
 	  }
 
+	if (e->flags & REF_MODIFY)
+	  {
+	    if (limit <= 0)
+	      {
+		FIB_ITERATE_PUT(fit);
+		ev_schedule(tab->rt_event);
+		return;
+	      }
+
+	    rte_modify(e);
+	    limit--;
+
+	    goto rescan;
+	  }
+      }
+
       if (!n->routes)		/* Orphaned FIB entry */
 	{
 	  FIB_ITERATE_PUT(fit);
diff --git a/proto/bgp/attrs.c b/proto/bgp/attrs.c
index 6580db57..dc267fdb 100644
--- a/proto/bgp/attrs.c
+++ b/proto/bgp/attrs.c
@@ -1413,6 +1413,10 @@ bgp_import_control(struct proto *P, rte **new, struct linpool *pool UNUSED)
     /* Do not export outside of AS (or confederation) */
     if (!p->is_interior && int_set_contains(d, BGP_COMM_NO_EXPORT))
       return -1;
+
+    /* Do not export LLGR_STALE routes to LLGR-ignorant peers */
+    if (!p->conn->remote_caps->llgr_aware && int_set_contains(d, BGP_COMM_LLGR_STALE))
+      return -1;
   }
 
   return 0;
@@ -1580,6 +1584,19 @@ rte_resolvable(rte *rt)
   return rt->attrs->dest == RTD_UNICAST;
 }
 
+static inline int
+rte_stale(rte *r)
+{
+  if (r->u.bgp.stale < 0)
+  {
+    /* If staleness is unknown, compute and cache it */
+    eattr *a = ea_find(r->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_COMMUNITY));
+    r->u.bgp.stale = a && int_set_contains(a->u.ptr, BGP_COMM_LLGR_STALE);
+  }
+
+  return r->u.bgp.stale;
+}
+
 int
 bgp_rte_better(rte *new, rte *old)
 {
@@ -1604,7 +1621,15 @@ bgp_rte_better(rte *new, rte *old)
   if (n < o)
     return 0;
 
-  /* Start with local preferences */
+  /* LLGR draft - depreference stale routes */
+  n = rte_stale(new);
+  o = rte_stale(old);
+  if (n > o)
+    return 0;
+  if (n < o)
+    return 1;
+
+ /* Start with local preferences */
   x = ea_find(new->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_LOCAL_PREF));
   y = ea_find(old->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_LOCAL_PREF));
   n = x ? x->u.data : new_bgp->cf->default_local_pref;
@@ -1725,6 +1750,10 @@ bgp_rte_mergable(rte *pri, rte *sec)
   if (!rte_resolvable(sec))
     return 0;
 
+  /* LLGR draft - depreference stale routes */
+  if (rte_stale(pri) != rte_stale(sec))
+    return 0;
+
   /* Start with local preferences */
   x = ea_find(pri->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_LOCAL_PREF));
   y = ea_find(sec->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_LOCAL_PREF));
@@ -1926,6 +1955,27 @@ bgp_rte_recalculate(rtable *table, net *net, rte *new, rte *old, rte *old_best)
     return old_is_group_best;
 }
 
+struct rte *
+bgp_rte_modify_stale(struct rte *r, struct linpool *pool)
+{
+  eattr *a = ea_find(r->attrs->eattrs, EA_CODE(PROTOCOL_BGP, BA_COMMUNITY));
+  struct adata *ad = a ? a->u.ptr : NULL;
+  uint flags = a ? a->flags : BAF_PARTIAL;
+
+  if (ad && int_set_contains(ad, BGP_COMM_NO_LLGR))
+    return NULL;
+
+  if (ad && int_set_contains(ad, BGP_COMM_LLGR_STALE))
+    return r;
+
+  r = rte_cow_rta(r, pool);
+  bgp_set_attr_ptr(&(r->attrs->eattrs), pool, BA_COMMUNITY, flags,
+		   int_set_add(pool, ad, BGP_COMM_LLGR_STALE));
+  r->u.bgp.stale = 1;
+
+  return r;
+}
+
 
 /*
  * Reconstruct AS_PATH and AGGREGATOR according to RFC 6793 4.2.3
@@ -2011,6 +2061,9 @@ bgp_get_route_info(rte *e, byte *buf)
   if (e->u.bgp.suppressed)
     buf += bsprintf(buf, "-");
 
+  if (rte_stale(e))
+    buf += bsprintf(buf, "s");
+
   if (e->attrs->hostentry)
   {
     if (!rte_resolvable(e))
diff --git a/proto/bgp/bgp.c b/proto/bgp/bgp.c
index ced83c5c..6dea88c8 100644
--- a/proto/bgp/bgp.c
+++ b/proto/bgp/bgp.c
@@ -513,8 +513,8 @@ bgp_conn_enter_established_state(struct bgp_conn *conn)
   p->route_refresh = peer->route_refresh;
   p->enhanced_refresh = local->enhanced_refresh && peer->enhanced_refresh;
 
-  /* Whether we may handle possible GR of peer (it has some AF GR-able) */
-  p->gr_ready = 0;	/* Updated later */
+  /* Whether we may handle possible GR/LLGR of peer (it has some AF GR-able) */
+  p->gr_ready = p->llgr_ready = 0;	/* Updated later */
 
   /* Whether peer is ready to handle our GR recovery */
   int peer_gr_ready = peer->gr_aware && !(peer->gr_flags & BGP_GRF_RESTART);
@@ -547,8 +547,15 @@ bgp_conn_enter_established_state(struct bgp_conn *conn)
     c->load_state = BFS_NONE;
 
     /* Channels where peer may do GR */
-    c->gr_ready = active && local->gr_aware && rem->gr_able;
+    uint gr_ready = active && local->gr_aware && rem->gr_able;
+    uint llgr_ready = active && local->llgr_aware && rem->llgr_able;
+
+    c->gr_ready = gr_ready || llgr_ready;
     p->gr_ready = p->gr_ready || c->gr_ready;
+    p->llgr_ready = p->llgr_ready || llgr_ready;
+
+    /* Remember last LLGR stale time */
+    c->stale_time = local->llgr_aware ? rem->llgr_time : 0;
 
     /* Channels not able to recover gracefully */
     if (p->p.gr_recovery && (!active || !peer_gr_ready))
@@ -558,8 +565,14 @@ bgp_conn_enter_established_state(struct bgp_conn *conn)
     if (p->p.gr_recovery && loc->gr_able && peer_gr_ready)
       c->c.gr_wait = 1;
 
-    /* Channels where peer is not able to recover gracefully */
-    if (c->gr_active && ! (c->gr_ready && (rem->gr_af_flags & BGP_GRF_FORWARDING)))
+    /* Channels where regular graceful restart failed */
+    if ((c->gr_active == BGP_GRS_ACTIVE) &&
+	!(active && rem->gr_able && (rem->gr_af_flags & BGP_GRF_FORWARDING)))
+      bgp_graceful_restart_done(c);
+
+    /* Channels where regular long-lived restart failed */
+    if ((c->gr_active == BGP_GRS_LLGR) &&
+	!(active && rem->llgr_able && (rem->gr_af_flags & BGP_LLGRF_FORWARDING)))
       bgp_graceful_restart_done(c);
 
     /* GR capability implies that neighbor will send End-of-RIB */
@@ -669,12 +682,25 @@ bgp_handle_graceful_restart(struct bgp_proto *p)
 
     if (c->gr_ready)
     {
-      if (c->gr_active)
-	rt_refresh_end(c->c.table, &c->c);
-
-      c->gr_active = 1;
       p->gr_active_num++;
-      rt_refresh_begin(c->c.table, &c->c);
+
+      switch (c->gr_active)
+      {
+      case BGP_GRS_NONE:
+	c->gr_active = BGP_GRS_ACTIVE;
+	rt_refresh_begin(c->c.table, &c->c);
+	break;
+
+      case BGP_GRS_ACTIVE:
+	rt_refresh_end(c->c.table, &c->c);
+	rt_refresh_begin(c->c.table, &c->c);
+	break;
+
+      case BGP_GRS_LLGR:
+	rt_refresh_begin(c->c.table, &c->c);
+	rt_modify_stale(c->c.table, &c->c);
+	break;
+      }
     }
     else
     {
@@ -695,7 +721,7 @@ bgp_handle_graceful_restart(struct bgp_proto *p)
   ASSERT(p->gr_active_num > 0);
 
   proto_notify_state(&p->p, PS_START);
-  bgp_start_timer(p->gr_timer, p->conn->remote_caps->gr_time);
+  tm_start(p->gr_timer, p->conn->remote_caps->gr_time S);
 }
 
 /**
@@ -720,6 +746,7 @@ bgp_graceful_restart_done(struct bgp_channel *c)
   if (!p->gr_active_num)
     BGP_TRACE(D_EVENTS, "Neighbor graceful restart done");
 
+  tm_stop(c->stale_timer);
   rt_refresh_end(c->c.table, &c->c);
 }
 
@@ -738,7 +765,46 @@ bgp_graceful_restart_timeout(timer *t)
   struct bgp_proto *p = t->data;
 
   BGP_TRACE(D_EVENTS, "Neighbor graceful restart timeout");
-  bgp_stop(p, 0, NULL, 0);
+
+  if (p->llgr_ready)
+  {
+    struct bgp_channel *c;
+    WALK_LIST(c, p->p.channels)
+    {
+      /* Channel is not in GR and is already flushed */
+      if (!c->gr_active)
+	continue;
+
+      /* Channel is already in LLGR from past restart */
+      if (c->gr_active == BGP_GRS_LLGR)
+	continue;
+
+      /* Channel is in GR, but does not support LLGR -> stop GR */
+      if (!c->stale_time)
+      {
+	bgp_graceful_restart_done(c);
+	continue;
+      }
+
+      /* Channel is in GR, and supports LLGR -> start LLGR */
+      c->gr_active = BGP_GRS_LLGR;
+      tm_start(c->stale_timer, c->stale_time S);
+      rt_modify_stale(c->c.table, &c->c);
+    }
+  }
+  else
+    bgp_stop(p, 0, NULL, 0);
+}
+
+static void
+bgp_long_lived_stale_timeout(timer *t)
+{
+  struct bgp_channel *c = t->data;
+  struct bgp_proto *p = (void *) c->c.proto;
+
+  BGP_TRACE(D_EVENTS, "Long-lived stale timeout");
+
+  bgp_graceful_restart_done(c);
 }
 
 
@@ -873,6 +939,12 @@ bgp_hold_timeout(timer *t)
 
   if (sk_rx_ready(conn->sk) > 0)
     bgp_start_timer(conn->hold_timer, 10);
+  else if ((conn->state == BS_ESTABLISHED) && p->llgr_ready)
+  {
+    BGP_TRACE(D_EVENTS, "Hold timer expired");
+    bgp_handle_graceful_restart(p);
+    bgp_conn_enter_idle_state(conn);
+  }
   else
     bgp_error(conn, 4, 0, NULL, 0);
 }
@@ -1172,9 +1244,26 @@ bgp_bfd_notify(struct bfd_request *req)
   {
     BGP_TRACE(D_EVENTS, "BFD session down");
     bgp_store_error(p, NULL, BE_MISC, BEM_BFD_DOWN);
-    if (ps == PS_UP)
-      bgp_update_startup_delay(p);
-    bgp_stop(p, 0, NULL, 0);
+
+    if (p->cf->bfd == BGP_BFD_GRACEFUL)
+    {
+      /* Trigger graceful restart */
+      if (p->conn && (p->conn->state == BS_ESTABLISHED) && p->gr_ready)
+	bgp_handle_graceful_restart(p);
+
+      if (p->incoming_conn.state > BS_IDLE)
+	bgp_conn_enter_idle_state(&p->incoming_conn);
+
+      if (p->outgoing_conn.state > BS_IDLE)
+	bgp_conn_enter_idle_state(&p->outgoing_conn);
+    }
+    else
+    {
+      /* Trigger session down */
+      if (ps == PS_UP)
+	bgp_update_startup_delay(p);
+      bgp_stop(p, 0, NULL, 0);
+    }
   }
 }
 
@@ -1447,6 +1536,7 @@ bgp_init(struct proto_config *CF)
   P->rte_better = bgp_rte_better;
   P->rte_mergable = bgp_rte_mergable;
   P->rte_recalculate = cf->deterministic_med ? bgp_rte_recalculate : NULL;
+  P->rte_modify = bgp_rte_modify_stale;
 
   p->cf = cf;
   p->local_as = cf->local_as;
@@ -1503,6 +1593,8 @@ bgp_channel_start(struct channel *C)
   bgp_init_bucket_table(c);
   bgp_init_prefix_table(c);
 
+  c->stale_timer = tm_new_init(c->pool, bgp_long_lived_stale_timeout, c, 0, 0);
+
   c->next_hop_addr = c->cf->next_hop_addr;
   c->link_addr = IPA_NONE;
   c->packets_to_send = 0;
@@ -1634,6 +1726,10 @@ bgp_postconfig(struct proto_config *CF)
   if (cf->multihop < 0)
     cf->multihop = internal ? 64 : 0;
 
+  /* LLGR mode default based on GR mode */
+  if (cf->llgr_mode < 0)
+    cf->llgr_mode = cf->gr_mode ? BGP_LLGR_AWARE : 0;
+
   /* Link check for single-hop BGP by default */
   if (cf->check_link < 0)
     cf->check_link = !cf->multihop;
@@ -1676,6 +1772,9 @@ bgp_postconfig(struct proto_config *CF)
   if (cf->multihop && cf->bfd && ipa_zero(cf->local_ip))
     cf_error("Multihop BGP with BFD requires specified local address");
 
+  if (!cf->gr_mode && cf->llgr_mode)
+    cf_error("Long-lived graceful restart requires basic graceful restart");
+
 
   struct bgp_channel_config *cc;
   WALK_LIST(cc, CF->channels)
@@ -1706,10 +1805,16 @@ bgp_postconfig(struct proto_config *CF)
     if (!cc->gw_mode)
       cc->gw_mode = cf->multihop ? GW_RECURSIVE : GW_DIRECT;
 
-    /* Default based on proto config */
+    /* Defaults based on proto config */
     if (cc->gr_able == 0xff)
       cc->gr_able = (cf->gr_mode == BGP_GR_ABLE);
 
+    if (cc->llgr_able == 0xff)
+      cc->llgr_able = (cf->llgr_mode == BGP_LLGR_ABLE);
+
+    if (cc->llgr_time == ~0U)
+      cc->llgr_time = cf->llgr_time;
+
     /* Default values of IGP tables */
     if ((cc->gw_mode == GW_RECURSIVE) && !cc->desc->no_igp)
     {
@@ -1885,6 +1990,7 @@ static char *bgp_state_names[] = { "Idle", "Connect", "Active", "OpenSent", "Ope
 static char *bgp_err_classes[] = { "", "Error: ", "Socket: ", "Received: ", "BGP Error: ", "Automatic shutdown: ", ""};
 static char *bgp_misc_errors[] = { "", "Neighbor lost", "Invalid next hop", "Kernel MD5 auth failed", "No listening socket", "Link down", "BFD session down", "Graceful restart"};
 static char *bgp_auto_errors[] = { "", "Route limit exceeded"};
+static char *bgp_gr_states[] = { "None", "Regular", "Long-lived"};
 
 static const char *
 bgp_last_errmsg(struct bgp_proto *p)
@@ -1963,6 +2069,7 @@ bgp_show_capabilities(struct bgp_proto *p UNUSED, struct bgp_caps *caps)
   uint any_gr_able = 0;
   uint any_add_path = 0;
   uint any_ext_next_hop = 0;
+  uint any_llgr_able = 0;
   u32 *afl1 = alloca(caps->af_count * sizeof(u32));
   u32 *afl2 = alloca(caps->af_count * sizeof(u32));
   uint afn1, afn2;
@@ -1973,6 +2080,7 @@ bgp_show_capabilities(struct bgp_proto *p UNUSED, struct bgp_caps *caps)
     any_gr_able |= ac->gr_able;
     any_add_path |= ac->add_path;
     any_ext_next_hop |= ac->ext_next_hop;
+    any_llgr_able |= ac->llgr_able;
   }
 
   if (any_mp_bgp)
@@ -2052,6 +2160,32 @@ bgp_show_capabilities(struct bgp_proto *p UNUSED, struct bgp_caps *caps)
 
   if (caps->enhanced_refresh)
     cli_msg(-1006, "      Enhanced refresh");
+
+  if (caps->llgr_aware)
+    cli_msg(-1006, "      Long-lived graceful restart");
+
+  if (any_llgr_able)
+  {
+    u32 stale_time = 0;
+
+    afn1 = afn2 = 0;
+    WALK_AF_CAPS(caps, ac)
+    {
+      stale_time = MAX(stale_time, ac->llgr_time);
+
+      if (ac->llgr_able && ac->llgr_time)
+	afl1[afn1++] = ac->afi;
+
+      if (ac->llgr_flags & BGP_GRF_FORWARDING)
+	afl2[afn2++] = ac->afi;
+    }
+
+    /* Continues from llgr_aware */
+    cli_msg(-1006, "        LL stale time: %u", stale_time);
+
+    bgp_show_afis(-1006, "        AF supported:", afl1, afn1);
+    bgp_show_afis(-1006, "        AF preserved:", afl2, afn2);
+  }
 }
 
 static void
@@ -2118,6 +2252,12 @@ bgp_show_proto_info(struct proto *P)
     {
       channel_show_info(&c->c);
 
+      if (p->gr_active_num)
+	cli_msg(-1006, "    Neighbor GR:    %s", bgp_gr_states[c->gr_active]);
+
+      if (tm_active(c->stale_timer))
+	cli_msg(-1006, "    LL stale timer: %t/-", tm_remains(c->stale_timer));
+
       if (c->c.channel_state == CS_UP)
       {
 	if (ipa_zero(c->link_addr))
diff --git a/proto/bgp/bgp.h b/proto/bgp/bgp.h
index 1235ee78..6f0a5587 100644
--- a/proto/bgp/bgp.h
+++ b/proto/bgp/bgp.h
@@ -107,9 +107,11 @@ struct bgp_config {
   int allow_local_as;			/* Allow that number of local ASNs in incoming AS_PATHs */
   int allow_local_pref;			/* Allow LOCAL_PREF in EBGP sessions */
   int gr_mode;				/* Graceful restart mode (BGP_GR_*) */
+  int llgr_mode;			/* Long-lived graceful restart mode (BGP_LLGR_*) */
   int setkey;				/* Set MD5 password to system SA/SP database */
   /* Times below are in seconds */
   unsigned gr_time;			/* Graceful restart timeout */
+  unsigned llgr_time;			/* Long-lived graceful restart stale time */
   unsigned connect_delay_time;		/* Minimum delay between connect attempts */
   unsigned connect_retry_time;		/* Timeout for connect attempts */
   unsigned hold_time, initial_hold_time;
@@ -138,6 +140,8 @@ struct bgp_channel_config {
   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 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 */
   u8 ext_next_hop;			/* Allow both IPv4 and IPv6 next hops */
   u8 add_path;				/* Use ADD-PATH extension [RFC 7911] */
 
@@ -166,12 +170,26 @@ struct bgp_channel_config {
 /* For GR capability per-AF flags */
 #define BGP_GRF_FORWARDING 0x80
 
+#define BGP_LLGR_ABLE		1
+#define BGP_LLGR_AWARE		2
+
+#define BGP_LLGRF_FORWARDING 0x80
+
+#define BGP_GRS_NONE		0	/* No GR  */
+#define BGP_GRS_ACTIVE		1	/* Graceful restart per RFC 4724 */
+#define BGP_GRS_LLGR		2	/* Long-lived GR phase (stale timer active) */
+
+#define BGP_BFD_GRACEFUL	2	/* BFD down triggers graceful restart */
+
 
 struct bgp_af_caps {
   u32 afi;
   u8 ready;				/* Multiprotocol capability, RFC 4760 */
   u8 gr_able;				/* Graceful restart support, RFC 4724 */
   u8 gr_af_flags;			/* Graceful restart per-AF flags */
+  u8 llgr_able;				/* Long-lived GR, RFC draft */
+  u32 llgr_time;			/* Long-lived GR stale time */
+  u8 llgr_flags;			/* Long-lived GR per-AF flags */
   u8 ext_next_hop;			/* Extended IPv6 next hop,   RFC 5549 */
   u8 add_path;				/* Multiple paths support,   RFC 7911 */
 };
@@ -188,6 +206,8 @@ struct bgp_caps {
   u8 gr_flags;				/* Graceful restart flags */
   u16 gr_time;				/* Graceful restart time in seconds */
 
+  u8 llgr_aware;			/* Long-lived GR capability, RFC draft */
+
   u16 af_count;				/* Number of af_data items */
 
   struct bgp_af_caps af_data[0];	/* Per-AF capability data */
@@ -243,6 +263,7 @@ struct bgp_proto {
   u8 route_refresh;			/* Route refresh allowed to send [RFC 2918] */
   u8 enhanced_refresh;			/* Enhanced refresh is negotiated [RFC 7313] */
   u8 gr_ready;				/* Neighbor could do graceful restart */
+  u8 llgr_ready;			/* Neighbor could do Long-lived GR, implies gr_ready */
   u8 gr_active_num;			/* Neighbor is doing GR, number of active channels */
   u8 channel_count;			/* Number of active channels */
   u32 *afi_map;				/* Map channel index -> AFI */
@@ -291,11 +312,14 @@ struct bgp_channel {
 
   u32 packets_to_send;			/* Bitmap of packet types to be sent */
 
-  u8 gr_ready;				/* Neighbor could do GR on this AF */
-  u8 gr_active;				/* Neighbor is doing GR and keeping fwd state */
-
   u8 ext_next_hop;			/* Session allows both IPv4 and IPv6 next hops */
 
+  u8 gr_ready;				/* Neighbor could do GR on this AF */
+  u8 gr_active;				/* Neighbor is doing GR (BGP_GRS_*) */
+
+  timer *stale_timer;			/* Long-lived stale timer for LLGR */
+  u32 stale_time;			/* Stored LLGR stale time from last session */
+
   u8 add_path_rx;			/* Session expects receive of ADD-PATH extended NLRI */
   u8 add_path_tx;			/* Session expects transmit of ADD-PATH extended NLRI */
 
@@ -505,6 +529,7 @@ void bgp_free_prefix(struct bgp_channel *c, struct bgp_prefix *bp);
 int bgp_rte_better(struct rte *, struct rte *);
 int bgp_rte_mergable(rte *pri, rte *sec);
 int bgp_rte_recalculate(rtable *table, net *net, rte *new, rte *old, rte *old_best);
+struct rte *bgp_rte_modify_stale(struct rte *r, struct linpool *pool);
 void bgp_rt_notify(struct proto *P, struct channel *C, net *n, rte *new, rte *old);
 int bgp_import_control(struct proto *, struct rte **, struct linpool *);
 int bgp_get_attr(struct eattr *e, byte *buf, int buflen);
@@ -645,6 +670,9 @@ void bgp_update_next_hop(struct bgp_export_state *s, eattr *a, ea_list **to);
 #define BGP_COMM_NO_ADVERTISE		0xffffff02	/* Don't export at all */
 #define BGP_COMM_NO_EXPORT_SUBCONFED	0xffffff03	/* NO_EXPORT even in local confederation */
 
+#define BGP_COMM_LLGR_STALE		0xffff0006	/* Route is stale according to LLGR */
+#define BGP_COMM_NO_LLGR		0xffff0007	/* Do not treat the route according to LLGR */
+
 /* Origins */
 
 #define ORIGIN_IGP		0
diff --git a/proto/bgp/config.Y b/proto/bgp/config.Y
index 7583f7d1..120b1e88 100644
--- a/proto/bgp/config.Y
+++ b/proto/bgp/config.Y
@@ -28,7 +28,8 @@ CF_KEYWORDS(BGP, LOCAL, NEIGHBOR, AS, HOLD, TIME, CONNECT, RETRY, KEEPALIVE,
 	BGP_CLUSTER_LIST, IGP, TABLE, GATEWAY, DIRECT, RECURSIVE, MED, TTL,
 	SECURITY, DETERMINISTIC, SECONDARY, ALLOW, BFD, ADD, PATHS, RX, TX,
 	GRACEFUL, RESTART, AWARE, CHECK, LINK, PORT, EXTENDED, MESSAGES, SETKEY,
-	STRICT, BIND, CONFEDERATION, MEMBER, MULTICAST, FLOW4, FLOW6)
+	STRICT, BIND, CONFEDERATION, MEMBER, MULTICAST, FLOW4, FLOW6, LONG,
+	LIVED, STALE)
 
 %type <i32> bgp_afi
 
@@ -63,6 +64,8 @@ bgp_proto_start: proto_start BGP {
      BGP_CFG->default_local_pref = 100;
      BGP_CFG->gr_mode = BGP_GR_AWARE;
      BGP_CFG->gr_time = 120;
+     BGP_CFG->llgr_mode = -1;
+     BGP_CFG->llgr_time = 3600;
      BGP_CFG->setkey = 1;
      BGP_CFG->check_link = -1;
    }
@@ -161,9 +164,13 @@ bgp_proto:
  | bgp_proto GRACEFUL RESTART bool ';' { BGP_CFG->gr_mode = $4; }
  | bgp_proto GRACEFUL RESTART AWARE ';' { BGP_CFG->gr_mode = BGP_GR_AWARE; }
  | bgp_proto GRACEFUL RESTART TIME expr ';' { BGP_CFG->gr_time = $5; }
+ | bgp_proto LONG LIVED GRACEFUL RESTART bool ';' { BGP_CFG->llgr_mode = $6; }
+ | bgp_proto LONG LIVED GRACEFUL RESTART AWARE ';' { BGP_CFG->llgr_mode = BGP_LLGR_AWARE; }
+ | bgp_proto LONG LIVED STALE TIME expr ';' { BGP_CFG->llgr_time = $6; }
  | bgp_proto TTL SECURITY bool ';' { BGP_CFG->ttl_security = $4; }
  | bgp_proto CHECK LINK bool ';' { BGP_CFG->check_link = $4; }
  | bgp_proto BFD bool ';' { BGP_CFG->bfd = $3; cf_check_bfd($3); }
+ | bgp_proto BFD GRACEFUL ';' { BGP_CFG->bfd = BGP_BFD_GRACEFUL; cf_check_bfd(1); }
  ;
 
 bgp_afi:
@@ -199,6 +206,8 @@ bgp_channel_start: bgp_afi
     BGP_CC->afi = $1;
     BGP_CC->desc = desc;
     BGP_CC->gr_able = 0xff;	/* undefined */
+    BGP_CC->llgr_able = 0xff;	/* undefined */
+    BGP_CC->llgr_time = ~0U;	/* undefined */
   }
 };
 
@@ -214,6 +223,8 @@ bgp_channel_item:
  | GATEWAY RECURSIVE { BGP_CC->gw_mode = GW_RECURSIVE; }
  | SECONDARY bool { BGP_CC->secondary = $2; }
  | 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; }
  | EXTENDED NEXT HOP bool { BGP_CC->ext_next_hop = $4; }
  | ADD PATHS RX { BGP_CC->add_path = BGP_ADD_PATH_RX; }
  | ADD PATHS TX { BGP_CC->add_path = BGP_ADD_PATH_TX; }
diff --git a/proto/bgp/packets.c b/proto/bgp/packets.c
index c40b8ec2..ed1db04b 100644
--- a/proto/bgp/packets.c
+++ b/proto/bgp/packets.c
@@ -260,6 +260,9 @@ bgp_write_capabilities(struct bgp_conn *conn, byte *buf)
     caps->gr_flags = p->p.gr_recovery ? BGP_GRF_RESTART : 0;
   }
 
+  if (p->cf->llgr_mode)
+    caps->llgr_aware = 1;
+
   /* Allocate and fill per-AF fields */
   WALK_LIST(c, p->p.channels)
   {
@@ -280,6 +283,15 @@ bgp_write_capabilities(struct bgp_conn *conn, byte *buf)
       if (p->p.gr_recovery)
 	ac->gr_af_flags |= BGP_GRF_FORWARDING;
     }
+
+    if (c->cf->llgr_able)
+    {
+      ac->llgr_able = 1;
+      ac->llgr_time = c->cf->llgr_time;
+
+      if (p->p.gr_recovery)
+	ac->llgr_flags |= BGP_LLGRF_FORWARDING;
+    }
   }
 
   /* Sort capability fields by AFI/SAFI */
@@ -289,9 +301,9 @@ bgp_write_capabilities(struct bgp_conn *conn, byte *buf)
   /* Create capability list in buffer */
 
   /*
-   * Note that max length is ~ 20+14*af_count. With max 12 channels that is
-   * 188. Option limit is 253 and buffer size is 4096, so we cannot overflow
-   * unless we add new capabilities or more AFs.
+   * Note that max length is ~ 22+21*af_count. With max 12 channels that is
+   * 274. Option limit is 253 and buffer size is 4096, so we cannot overflow
+   * unless we add new capabilities or more AFs. XXXXX
    */
 
   WALK_AF_CAPS(caps, ac)
@@ -384,6 +396,24 @@ bgp_write_capabilities(struct bgp_conn *conn, byte *buf)
     *buf++ = 0;			/* Capability data length */
   }
 
+  if (caps->llgr_aware)
+  {
+    *buf++ = 71;		/* Capability 71: Support for long-lived graceful restart */
+    *buf++ = 0;			/* Capability data length, will be fixed later */
+    data = buf;
+
+    WALK_AF_CAPS(caps, ac)
+      if (ac->llgr_able)
+      {
+	put_af3(buf, ac->afi);
+	buf[3] = ac->llgr_flags;
+	put_u24(buf+4, ac->llgr_time);
+	buf += 7;
+      }
+
+    data[-1] = buf - data;
+  }
+
   return buf;
 }
 
@@ -508,11 +538,49 @@ bgp_read_capabilities(struct bgp_conn *conn, struct bgp_caps *caps, byte *pos, i
       caps->enhanced_refresh = 1;
       break;
 
+    case 71: /* Long lived graceful restart capability, RFC draft */
+      if (cl % 7)
+	goto err;
+
+      /* Presumably, only the last instance is valid */
+      WALK_AF_CAPS(caps, ac)
+      {
+	ac->llgr_able = 0;
+	ac->llgr_flags = 0;
+	ac->llgr_time = 0;
+      }
+
+      caps->llgr_aware = 1;
+
+      for (i = 0; i < cl; i += 7)
+      {
+	af = get_af3(pos+2+i);
+	ac = bgp_get_af_caps(caps, af);
+	ac->llgr_able = 1;
+	ac->llgr_flags = pos[2+i+3];
+	ac->llgr_time = get_u24(pos + 2+i+4);
+      }
+      break;
+
       /* We can safely ignore all other capabilities */
     }
 
     ADVANCE(pos, len, 2 + cl);
   }
+
+  /* The LLGR capability must be advertised together with the GR capability,
+     otherwise it must be disregarded */
+  if (!caps->gr_aware && caps->llgr_aware)
+  {
+    caps->llgr_aware = 0;
+    WALK_AF_CAPS(caps, ac)
+    {
+      ac->llgr_able = 0;
+      ac->llgr_flags = 0;
+      ac->llgr_time = 0;
+    }
+  }
+
   return;
 
 err:
@@ -1131,6 +1199,7 @@ bgp_rte_update(struct bgp_parse_state *s, net_addr *n, u32 path_id, rta *a0)
 
   e->pflags = 0;
   e->u.bgp.suppressed = 0;
+  e->u.bgp.stale = -1;
   rte_update2(&s->channel->c, n, e, s->last_src);
 }
 
diff --git a/proto/pipe/pipe.c b/proto/pipe/pipe.c
index 7aada37e..82ccf38a 100644
--- a/proto/pipe/pipe.c
+++ b/proto/pipe/pipe.c
@@ -78,6 +78,12 @@ pipe_rt_notify(struct proto *P, struct channel *src_ch, net *n, rte *new, rte *o
       e->pref = new->pref;
       e->pflags = new->pflags;
 
+#ifdef CONFIG_BGP
+      /* Hack to cleanup cached value */
+      if (e->attrs->src->proto->proto == &proto_bgp)
+	e->u.bgp.stale = -1;
+#endif
+
       src = a->src;
     }
   else

From d33cf3f4c3a92f895e8b52fc19ed8a88f350a32b Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 7 Aug 2018 14:46:24 +0200
Subject: [PATCH 192/198] Doc: Fix notes related to obsolete option

Thanks to Julien Dessaux for the report.
---
 doc/bird.sgml | 20 ++++----------------
 1 file changed, 4 insertions(+), 16 deletions(-)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index 44fef1c5..93b701d3 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -2848,14 +2848,6 @@ itself and BGP protocol is usually used for exporting aggregate routes. But the
 Direct protocol is necessary for distance-vector protocols like RIP or Babel to
 announce local networks.
 
-<p>There is one notable case when you definitely want to use the direct protocol
--- running BIRD on BSD systems. Having high priority device routes for directly
-connected networks from the direct protocol protects kernel device routes from
-being overwritten or removed by IGP routes during some transient network
-conditions, because a lower priority IGP route for the same network is not
-exported to the kernel routing table. This is an issue on BSD systems only, as
-on Linux systems BIRD cannot change non-BIRD route in the kernel routing table.
-
 <p>There are just few configuration options for the Direct protocol:
 
 <p><descrip>
@@ -2900,14 +2892,10 @@ interface) or whether an `alien' route has been added by someone else (depending
 on the <cf/learn/ switch, such routes are either ignored or accepted to our
 table).
 
-<p>Unfortunately, there is one thing that makes the routing table synchronization
-a bit more complicated. In the kernel routing table there are also device routes
-for directly connected networks. These routes are usually managed by OS itself
-(as a part of IP address configuration) and we don't want to touch that. They
-are completely ignored during the scan of the kernel tables and also the export
-of device routes from BIRD tables to kernel routing tables is restricted to
-prevent accidental interference. This restriction can be disabled using
-<cf/device routes/ switch.
+<p>Note that routes created by OS kernel itself, namely direct routes
+representing IP subnets of associated interfaces, are not imported even with
+<cf/learn/ enabled. You can use <ref id="direct" name="Direct protocol"> to
+generate these direct routes.
 
 <p>If your OS supports only a single routing table, you can configure only one
 instance of the Kernel protocol. If it supports multiple tables (in order to

From 1279a83103262950ab99e8a6fe3c6cc2da8d42a8 Mon Sep 17 00:00:00 2001
From: Pavel Tvrdik <pawel.tvrdik@gmail.com>
Date: Mon, 14 Nov 2016 14:53:10 +0100
Subject: [PATCH 193/198] sysdep/unix/main.c: Remove trailing spaces

---
 sysdep/unix/main.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/sysdep/unix/main.c b/sysdep/unix/main.c
index 2251d3fb..ea1476e7 100644
--- a/sysdep/unix/main.c
+++ b/sysdep/unix/main.c
@@ -132,7 +132,7 @@ read_iproute_table(char *file, char *prefix, int max)
 
     if (*p == '#' || *p == '\n' || *p == 0)
       continue;
-   
+
     if (sscanf(p, "0x%x %s\n", &val, name) != 2 &&
 	sscanf(p, "0x%x %s #", &val, name) != 2 &&
 	sscanf(p, "%d %s\n", &val, name) != 2 &&
@@ -522,7 +522,7 @@ write_pid_file(void)
   rv = ftruncate(pid_fd, 0);
   if (rv < 0)
     die("fruncate: %m");
-    
+
   rv = write(pid_fd, ps, pl);
   if(rv < 0)
     die("write: %m");

From 86b9e8e39a0b42407c95921ca8262b0a75cad5f2 Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Tue, 17 Jul 2018 15:30:59 +0200
Subject: [PATCH 194/198] M4: generate synchronization lines

This also includes Bison version check. Versions before 3.0 don't
support them in a reliable way and we don't promise to work with
versions older than 2.4.
---
 Makefile.in   |  1 +
 aclocal.m4    | 19 +++++++++++++++++++
 conf/Makefile |  2 +-
 configure.ac  |  9 +++++++++
 4 files changed, 30 insertions(+), 1 deletion(-)

diff --git a/Makefile.in b/Makefile.in
index 8f6c0c8b..5665f1d3 100644
--- a/Makefile.in
+++ b/Makefile.in
@@ -9,6 +9,7 @@ MAKEFLAGS += -r
 CPPFLAGS=-I$(objdir) -I$(srcdir) @CPPFLAGS@
 CFLAGS=$(CPPFLAGS) @CFLAGS@
 LDFLAGS=@LDFLAGS@
+M4FLAGS=@M4FLAGS@
 LIBS=@LIBS@
 DAEMON_LIBS=@DAEMON_LIBS@
 CLIENT_LIBS=@CLIENT_LIBS@
diff --git a/aclocal.m4 b/aclocal.m4
index dcf1bd8b..28a632c8 100644
--- a/aclocal.m4
+++ b/aclocal.m4
@@ -175,3 +175,22 @@ AC_DEFUN([BIRD_CHECK_PROG_FLAVOR_GNU],
   )
   esac
 ])
+
+AC_DEFUN([BIRD_CHECK_BISON_VERSION],
+[
+  $1=`bison --version | ( read line; echo ${line##* } )`
+  case "$$1" in
+    1.* | 2.0* | 2.1* | 2.2* | 2.3*)
+      AC_MSG_ERROR([Provided Bison version $$1 is too old, need at least 2.4])
+      ;;
+    2.*)
+      bird_bison_synclines=no
+      ;;
+    3.* | 4.* | 5.* | 6.* | 7.* | 8.* | 9.*)
+      bird_bison_synclines=yes
+      ;;
+    *)
+      AC_MSG_ERROR([Couldn't parse Bison version $$1. Call the developers for help.])
+      ;;
+  esac
+])
diff --git a/conf/Makefile b/conf/Makefile
index fb3dd052..47024d26 100644
--- a/conf/Makefile
+++ b/conf/Makefile
@@ -11,7 +11,7 @@ BISON_DEBUG=-t
 endif
 
 $(conf-y-targets): $(s)confbase.Y $(s)flowspec.Y
-	$(M4) -P $| $^ >$@
+	$(M4) $(M4FLAGS) -P $| $^ >$@
 
 $(o)cf-parse.y: | $(s)gen_parser.m4
 $(o)keywords.h: | $(s)gen_keywords.m4
diff --git a/configure.ac b/configure.ac
index 44184273..9a2e55e2 100644
--- a/configure.ac
+++ b/configure.ac
@@ -157,6 +157,15 @@ test -z "$FLEX"	 && AC_MSG_ERROR([Flex is missing.])
 test -z "$BISON" && AC_MSG_ERROR([Bison is missing.])
 test -z "$M4"	 && AC_MSG_ERROR([M4 is missing.])
 
+AC_MSG_CHECKING([bison version])
+BIRD_CHECK_BISON_VERSION(BISON_VERSION)
+AC_MSG_RESULT([$BISON_VERSION])
+if test "$bird_bison_synclines" = yes; then
+  M4FLAGS="$M4FLAGS -s"
+fi
+
+AC_SUBST([M4FLAGS])
+
 BIRD_CHECK_PROG_FLAVOR_GNU([$M4],
   [],
   [AC_MSG_ERROR([Provided M4 is not GNU M4.])]

From 78ca6ea8f0b8ab312b1e0ef56629a18e3efbce42 Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Tue, 17 Jul 2018 15:39:06 +0200
Subject: [PATCH 195/198] Make: Add option to force colors in compiler output

---
 Makefile.in | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/Makefile.in b/Makefile.in
index 5665f1d3..6e281cfc 100644
--- a/Makefile.in
+++ b/Makefile.in
@@ -55,6 +55,10 @@ else
   Q:=
 endif
 
+ifneq ($(COLOR),)
+  CFLAGS += -fdiagnostics-color=always
+endif
+
 # Meta rules
 docgoals := docs userdocs progdocs
 testgoals := check test tests tests_run

From 7ffc0a6534fb66ad27d22984e25bd0789f4404b6 Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Tue, 14 Aug 2018 14:36:44 +0200
Subject: [PATCH 196/198] Bison: A bit more verbose error messages in config.

---
 Makefile.in   | 1 +
 aclocal.m4    | 2 ++
 conf/Makefile | 2 +-
 configure.ac  | 5 +++++
 4 files changed, 9 insertions(+), 1 deletion(-)

diff --git a/Makefile.in b/Makefile.in
index 6e281cfc..b755df44 100644
--- a/Makefile.in
+++ b/Makefile.in
@@ -10,6 +10,7 @@ CPPFLAGS=-I$(objdir) -I$(srcdir) @CPPFLAGS@
 CFLAGS=$(CPPFLAGS) @CFLAGS@
 LDFLAGS=@LDFLAGS@
 M4FLAGS=@M4FLAGS@
+BISONFLAGS=@BISONFLAGS@
 LIBS=@LIBS@
 DAEMON_LIBS=@DAEMON_LIBS@
 CLIENT_LIBS=@CLIENT_LIBS@
diff --git a/aclocal.m4 b/aclocal.m4
index 28a632c8..c401d447 100644
--- a/aclocal.m4
+++ b/aclocal.m4
@@ -185,9 +185,11 @@ AC_DEFUN([BIRD_CHECK_BISON_VERSION],
       ;;
     2.*)
       bird_bison_synclines=no
+      bird_bison_enhanced_error=no
       ;;
     3.* | 4.* | 5.* | 6.* | 7.* | 8.* | 9.*)
       bird_bison_synclines=yes
+      bird_bison_enhanced_error=yes
       ;;
     *)
       AC_MSG_ERROR([Couldn't parse Bison version $$1. Call the developers for help.])
diff --git a/conf/Makefile b/conf/Makefile
index 47024d26..39628bff 100644
--- a/conf/Makefile
+++ b/conf/Makefile
@@ -20,7 +20,7 @@ $(o)commands.h: | $(s)gen_commands.m4 $(srcdir)/client/cmds.m4
 $(o)cf-parse.tab.h: $(o)cf-parse.tab.c
 
 $(o)cf-parse.tab.c: $(o)cf-parse.y
-	$(BISON) $(BISON_DEBUG) -dv -pcf_ -b $(@:.tab.c=) $<
+	$(BISON) $(BISON_DEBUG) $(BISONFLAGS) -dv -pcf_ -b $(@:.tab.c=) $<
 
 $(o)cf-lex.c: $(s)cf-lex.l
 	$(FLEX) $(FLEX_DEBUG) -s -B -8 -Pcf_ -o$@ $<
diff --git a/configure.ac b/configure.ac
index 9a2e55e2..943d3a7b 100644
--- a/configure.ac
+++ b/configure.ac
@@ -164,7 +164,12 @@ if test "$bird_bison_synclines" = yes; then
   M4FLAGS="$M4FLAGS -s"
 fi
 
+if test "$bird_bison_enhanced_error" = yes; then
+  BISONFLAGS="$BISONFLAGS -Dparse.lac=full -Dparse.error=verbose"
+fi
+
 AC_SUBST([M4FLAGS])
+AC_SUBST([BISONFLAGS])
 
 BIRD_CHECK_PROG_FLAVOR_GNU([$M4],
   [],

From 765f400f6b7ac054ddb4fcc4f5bec58f94d1bdd4 Mon Sep 17 00:00:00 2001
From: "Ondrej Zajicek (work)" <santiago@crfreenet.org>
Date: Tue, 21 Aug 2018 15:24:55 +0200
Subject: [PATCH 197/198] DOC: Remove pipe mode reference

Pipe mode was removed in 2.0, remove reference to it in the documentation.

Thanks to Piotr Wydrych for the bugreport.
---
 doc/bird.sgml | 15 +++------------
 1 file changed, 3 insertions(+), 12 deletions(-)

diff --git a/doc/bird.sgml b/doc/bird.sgml
index 93b701d3..0057d2c5 100644
--- a/doc/bird.sgml
+++ b/doc/bird.sgml
@@ -3656,18 +3656,9 @@ the filters. Export filters control export of routes from the primary table to
 the secondary one, import filters control the opposite direction. Both tables
 must be of the same nettype.
 
-<p>The Pipe protocol may work in the transparent mode mode or in the opaque
-mode. In the transparent mode, the Pipe protocol retransmits all routes from
-one table to the other table, retaining their original source and attributes.
-If import and export filters are set to accept, then both tables would have
-the same content. The transparent mode is the default mode.
-
-<p>In the opaque mode, the Pipe protocol retransmits optimal route from one
-table to the other table in a similar way like other protocols send and receive
-routes. Retransmitted route will have the source set to the Pipe protocol, which
-may limit access to protocol specific route attributes. This mode is mainly for
-compatibility, it is not suggested for new configs. The mode can be changed by
-<tt/mode/ option.
+<p>The Pipe protocol retransmits all routes from one table to the other table,
+retaining their original source and attributes. If import and export filters
+are set to accept, then both tables would have the same content.
 
 <p>The primary use of multiple routing tables and the Pipe protocol is for
 policy routing, where handling of a single packet doesn't depend only on its

From 64c5ad58d276d8a0463aa9ad2b34f75b7d1f4108 Mon Sep 17 00:00:00 2001
From: Jan Maria Matejka <mq@ucw.cz>
Date: Wed, 22 Aug 2018 14:58:53 +0200
Subject: [PATCH 198/198] Lib: recursive printf

Use like this:

void func(const char *msg, va_list args) {
  ...
  bvsnprintf(buf, len, "file %s, line %d: %V (foo %d, bar %d)", file, line, msg, &args, foo, bar);
  ...
}
---
 lib/printf.c | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/lib/printf.c b/lib/printf.c
index 533a1300..cafeeb93 100644
--- a/lib/printf.c
+++ b/lib/printf.c
@@ -270,6 +270,17 @@ int bvsnprintf(char *buf, int size, const char *fmt, va_list args)
 				*str++ = ' ';
 			continue;
 
+		case 'V': {
+			const char *vfmt = va_arg(args, const char *);
+			va_list *vargs = va_arg(args, va_list *);
+			int res = bvsnprintf(str, size, vfmt, *vargs);
+			if (res < 0)
+				return -1;
+			str += res;
+			size -= res;
+			continue;
+			}
+
 		case 'p':
 			if (field_width == -1) {
 				field_width = 2*sizeof(void *);