From 27b472533bc8bb0fdb37dcda7596c385d3b9f414 Mon Sep 17 00:00:00 2001
From: Paul Jakma <paul.jakma@sun.com>
Date: Sun, 2 Jul 2006 16:38:54 +0000
Subject: [zebra] Fix CID #104, check addr for null, and #18, check nexthop
 type args

2006-07-02 Paul Jakma <paul.jakma@sun.com>

	* rt_netlink.c: (netlink_interface_addr) Fix CID #104, can not
	  proceed if addr is NULL.
	* zebra_rib.c: (static_add_ipv6) Fix CID #18, double check
	  required arguments are supplied for the given nexthop type.
---
 zebra/rt_netlink.c | 7 +++++++
 1 file changed, 7 insertions(+)

(limited to 'zebra/rt_netlink.c')

diff --git a/zebra/rt_netlink.c b/zebra/rt_netlink.c
index f1c1a300..216625e9 100644
--- a/zebra/rt_netlink.c
+++ b/zebra/rt_netlink.c
@@ -673,6 +673,13 @@ netlink_interface_addr (struct sockaddr_nl *snl, struct nlmsghdr *h)
         broad = NULL;
     }
 
+  /* addr is primary key, SOL if we don't have one */
+  if (addr == NULL)
+    {
+      zlog_debug ("%s: NULL address", __func__);
+      return -1;
+    }
+
   /* Flags. */
   if (ifa->ifa_flags & IFA_F_SECONDARY)
     SET_FLAG (flags, ZEBRA_IFA_SECONDARY);
-- 
cgit v1.2.3


From e4529636b77124285cca96a62799d0ff6a7addeb Mon Sep 17 00:00:00 2001
From: "Andrew J. Schorr" <ajschorr@alumni.princeton.edu>
Date: Tue, 12 Dec 2006 19:18:21 +0000
Subject: [PtP over ethernet] New peer flag allows much more addressing
 flexibility

2006-12-12 Andrew J. Schorr <ajschorr@alumni.princeton.edu>

	* if.h: (struct connected) Add new ZEBRA_IFA_PEER flag indicating
	  whether a peer address has been configured.  Comment now shows
	  the new interpretation of the destination addr: if ZEBRA_IFA_PEER
	  is set, then it must contain the destination address, otherwise
	  it may contain the broadcast address or be NULL.
	  (CONNECTED_DEST_HOST,CONNECTED_POINTOPOINT_HOST) Remove obsolete
	  macros that were specific to IPv4 and not fully general.
	  (CONNECTED_PEER) New macro to check ZEBRA_IFA_PEER flag.
	  (CONNECTED_PREFIX) New macro giving the prefix to insert into
	  the RIB: if CONNECTED_PEER, then use the destination (peer) address,
	  else use the address field.
	  (CONNECTED_ID) New macro to come up with an identifying address
	  for the struct connected.
	* if.c: (if_lookup_address, connected_lookup_address) Streamline
	  logic with new CONNECTED_PREFIX macro.
	* prefix.h: (PREFIX_COPY_IPV4, PREFIX_COPY_IPV6) New macros
	  for better performance than the general prefix_copy function.
	* zclient.c: (zebra_interface_address_read) For non-null destination
	  addresses, set prefixlen to equal the address prefixlen.  This
	  is needed to get the new CONNECTED_PREFIX macro to work properly.
	* connected.c: (connected_up_ipv4, connected_down_ipv4,
	  connected_up_ipv6, connected_down_ipv6) Simplify logic using the
	  new CONNECTED_PREFIX macro.
	  (connected_add_ipv4) Set prefixlen in destination addresses (required
	  by the CONNECTED_PREFIX macro).  Use CONNECTED_PEER macro instead
	  of testing for IFF_POINTOPOINT.  Delete invalid warning message.
	  Warn about cases where the ZEBRA_IFA_PEER is set but no
	  destination address has been supplied (and turn off the flag).
	  (connected_add_ipv6) Add new flags argument so callers may set
	  the ZEBRA_IFA_PEER flag.  If peer/broadcast address satisfies
	  IN6_IS_ADDR_UNSPECIFIED, then reject it with a warning.
	  Set prefixlen in destination address so CONNECTED_PREFIX will work.
	* connected.h: (connected_add_ipv6) Add new flags argument so
	  callers may set the ZEBRA_IFA_PEER flag.
	* interface.c: (connected_dump_vty) Use CONNECTED_PEER macro
	  to decide whether the destination address is a peer or broadcast
	  address (instead of checking IFF_BROADCAST and IFF_POINTOPOINT).
	* if_ioctl.c: (if_getaddrs) Instead of setting a peer address
	  only when the IFF_POINTOPOINT is set, we now accept a peer
	  address whenever it is available and not the same as the local
	  address.  Otherwise (no peer address assigned), we check
	  for a broadcast address (regardless of the IFF_BROADCAST flag).
	  And must now pass a flags value of ZEBRA_IFA_PEER to
	  connected_add_ipv4 when a peer address is assigned.
	  The same new logic is used with the IPv6 code as well (and we
	  pass the new flags argument to connected_add_ipv6).
	  (if_get_addr) Do not bother to check IFF_POINTOPOINT: just
	  issue the SIOCGIFDSTADDR ioctl and see if we get back
	  a peer address not matching the local address (and set
	  the ZEBRA_IFA_PEER in that case).  If there's no peer address,
	  try to grab SIOCGIFBRDADDR regardless of whether IFF_BROADCAST is set.
	* if_ioctl_solaris.c: (if_get_addr) Just try the SIOCGLIFDSTADDR ioctl
	  without bothering to check the IFF_POINTOPOINT flag.  And if
	  no peer address was found, just try the SIOCGLIFBRDADDR ioctl
	  without checking the IFF_BROADCAST flag.  Call connected_add_ipv4
	  and connected_add_ipv6 with appropriate flags.
	* if_proc.c: (ifaddr_proc_ipv6) Must pass new flags argument to
	  connected_add_ipv6.
	* kernel_socket.c: (ifam_read) Must pass new flags argument to
	  connected_add_ipv6.
	* rt_netlink.c: (netlink_interface_addr) Copy logic from iproute2
	  to determine local and possible peer address (so there's no longer
	  a test for IFF_POINTOPOINT).  Set ZEBRA_IFA_PEER flag appropriately.
	  Pass new flags argument to connected_add_ipv6.
	  (netlink_address) Test !CONNECTED_PEER instead of if_is_broadcast
	  to determine whether the connected destination address is a
	  broadcast address.
	* bgp_nexthop.c: (bgp_connected_add, bgp_connected_delete)
	  Simplify logic by using new CONNECTED_PREFIX macro.
	* ospf_interface.c: (ospf_if_is_configured, ospf_if_lookup_by_prefix,
	  ospf_if_lookup_recv_if) Simplify logic using new CONNECTED_PREFIX
	  macro.
	* ospf_lsa.c: (lsa_link_ptop_set) Using the new CONNECTED_PREFIX
	  macro, both options collapse into the same code.
	* ospf_snmp.c: (ospf_snmp_if_update) Simplify logic using new
	  CONNECTED_ID macro.
	  (ospf_snmp_is_if_have_addr) Simplify logic using new CONNECTED_PREFIX
	  macro.
	* ospf_vty.c: (show_ip_ospf_interface_sub) Use new CONNECTED_PEER macro
	  instead of testing the IFF_POINTOPOINT flag.
	* ospfd.c: (ospf_network_match_iface) Use new CONNECTED_PEER macro
	  instead of testing with if_is_pointopoint.  And add commented-out
	  code to implement alternative (in my opinion) more elegant behavior
	  that has no special-case treatment for PtP addresses.
	  (ospf_network_run) Use new CONNECTED_ID macro to simplify logic.
	* rip_interface.c: (rip_interface_multicast_set) Use new CONNECTED_ID
	  macro to simplify logic.
	  (rip_request_interface_send) Fix minor bug: ipv4_broadcast_addr does
	  not give a useful result if prefixlen is 32 (we require a peer
	  address in such cases).
	* ripd.c: (rip_update_interface) Fix same bug as above.
---
 zebra/rt_netlink.c | 54 +++++++++++++++++++++---------------------------------
 1 file changed, 21 insertions(+), 33 deletions(-)

(limited to 'zebra/rt_netlink.c')

diff --git a/zebra/rt_netlink.c b/zebra/rt_netlink.c
index 216625e9..e2f1f9d9 100644
--- a/zebra/rt_netlink.c
+++ b/zebra/rt_netlink.c
@@ -577,8 +577,8 @@ netlink_interface_addr (struct sockaddr_nl *snl, struct nlmsghdr *h)
   struct ifaddrmsg *ifa;
   struct rtattr *tb[IFA_MAX + 1];
   struct interface *ifp;
-  void *addr = NULL;
-  void *broad = NULL;
+  void *addr;
+  void *broad;
   u_char flags = 0;
   char *label = NULL;
 
@@ -637,41 +637,29 @@ netlink_interface_addr (struct sockaddr_nl *snl, struct nlmsghdr *h)
         }
     }
   
+  /* logic copied from iproute2/ip/ipaddress.c:print_addrinfo() */
+  if (tb[IFA_LOCAL] == NULL)
+    tb[IFA_LOCAL] = tb[IFA_ADDRESS];
   if (tb[IFA_ADDRESS] == NULL)
     tb[IFA_ADDRESS] = tb[IFA_LOCAL];
   
-  if (ifp->flags & IFF_POINTOPOINT)
+  /* local interface address */
+  addr = (tb[IFA_LOCAL] ? RTA_DATA(tb[IFA_LOCAL]) : NULL);
+
+  /* is there a peer address? */
+  /* N.B. I do not understand why the memcmp compares 4 bytes regardless
+     of address family, but this is exactly how it appears in
+     print_addrinfo.  I wonder if it should be RTA_PAYLOAD(tb[IFA_ADDRESS])
+     instead of 4... */
+  if (tb[IFA_ADDRESS] &&
+      memcmp(RTA_DATA(tb[IFA_ADDRESS]), RTA_DATA(tb[IFA_LOCAL]), 4))
     {
-      if (tb[IFA_LOCAL])
-        {
-          addr = RTA_DATA (tb[IFA_LOCAL]);
-          if (tb[IFA_ADDRESS] &&
-	      memcmp(RTA_DATA(tb[IFA_ADDRESS]),RTA_DATA(tb[IFA_LOCAL]),4))
-	    /* if IFA_ADDRESS != IFA_LOCAL, then it's the peer address */
-            broad = RTA_DATA (tb[IFA_ADDRESS]);
-          else
-            broad = NULL;
-        }
-      else
-        {
-          if (tb[IFA_ADDRESS])
-            addr = RTA_DATA (tb[IFA_ADDRESS]);
-          else
-            addr = NULL;
-        }
+      broad = RTA_DATA(tb[IFA_ADDRESS]);
+      SET_FLAG (flags, ZEBRA_IFA_PEER);
     }
   else
-    {
-      if (tb[IFA_ADDRESS])
-        addr = RTA_DATA (tb[IFA_ADDRESS]);
-      else
-        addr = NULL;
-      
-      if (tb[IFA_BROADCAST])
-        broad = RTA_DATA(tb[IFA_BROADCAST]);
-      else
-        broad = NULL;
-    }
+    /* seeking a broadcast address */
+    broad = (tb[IFA_BROADCAST] ? RTA_DATA(tb[IFA_BROADCAST]) : NULL);
 
   /* addr is primary key, SOL if we don't have one */
   if (addr == NULL)
@@ -707,7 +695,7 @@ netlink_interface_addr (struct sockaddr_nl *snl, struct nlmsghdr *h)
   if (ifa->ifa_family == AF_INET6)
     {
       if (h->nlmsg_type == RTM_NEWADDR)
-        connected_add_ipv6 (ifp,
+        connected_add_ipv6 (ifp, flags,
                             (struct in6_addr *) addr, ifa->ifa_prefixlen,
                             (struct in6_addr *) broad, label);
       else
@@ -1849,7 +1837,7 @@ netlink_address (int cmd, int family, struct interface *ifp,
 
   if (family == AF_INET && cmd == RTM_NEWADDR)
     {
-      if (if_is_broadcast (ifp) && ifc->destination)
+      if (!CONNECTED_PEER(ifc) && ifc->destination)
         {
           p = ifc->destination;
           addattr_l (&req.n, sizeof req, IFA_BROADCAST, &p->u.prefix,
-- 
cgit v1.2.3


From 7514fb7739f74311830e9ddd1381d0d228224f61 Mon Sep 17 00:00:00 2001
From: Paul Jakma <paul.jakma@sun.com>
Date: Wed, 2 May 2007 16:05:35 +0000
Subject: [zebra] Routemap support on received routes, with 'set src' command
 (linux)

2007-05-01 David L Stevens <dlstevens@us.ibm.com>

	* (general) These changes collectively add route-map and
	  prefix-list support to zebra and fix a bug in "show
	  route-map" (with no argument).
	* doc/main.texi: added route-map, prefix-list, ip protocol
	  and set src documentation
	* lib/command.h: added PROTOCOL_NODE type
	* lib/log.c: (proto_name2num) new function, protocol name to
	  number translation.
	* lib/routemap.c: (vty_show_route_map) fixed "show route-map"
	  without route-map name
	* lib/routemap.h: added RMAP_ZEBRA type
	* lib/zebra.h: added proto_name2num() prototype
	* vtysh/extract.pl.in: added VTYSH_ZEBRA flag for route-map and
	  plist
	* vtysh/Makefile.am: added zebra_routemap.c
	* vtysh/vtysh.h: added VTYSH_ZEBRA flag to VTYSH_RMAP
	* zebra/connected.c: (connected_up_ipv4) added src preference argument
	  to rib_add_ipv4()
	* zebra/kernel_socket.c: (rtm_read) ditto
	* zebra/main.c: added prefix list initialization
	* zebra/Makefile.am: added zebra_routemap.c source file
	* zebra/rib.h: added generic address union "g_addr" and use in
	  existing places that had an explicit union.
	  Added "src" to struct nexthop.
	  Added preferred src arg to nexthop_ipv4_add and rib_add_ipv4.
	* zebra/rt_netlink.c: (netlink_routing_table) set preferred source on
	  netlink messages.
	  (netlink_route_change) ditto
	  (netlink_route_multipath) ditto.
	* zebra/rtread_getmsg.c: (handle_route_entry) added (NULL) src to
	  rib_add_ipv4() call.
	* zebra/rtread_proc.c: (proc_route_read) ditto
	* zebra/zebra_rib.c: (nexthop_ipv4_add) add src argument.
	  (nexthop_ipv4_ifindex_add) ditto
	  (rib_add_ipv4) ditto
	  (nexthop_active_check) Add route-map processing.
	* zebra/zebra_routemap.c: new file for zebra route-map commands.
	* zebra/zebra_vty.c: (ip_protocol_cmd) Apply route-map to protocol
	  (vty_show_ip_route_detail) added "src" printing
	  (vty_show_ip_route) ditto
	  (show_ip_protocol_cmd) new command, list routemaps.
	  (config_write_protocol) write out routemap protocl config.
	  (zebra_vty_init) Install the new routemap protocol commands.
	* zebra/zserv.c: (zread_ipv4_add) added (NULL) src arg
	  (zebra_init) init zebra route-maps.
	* zebra/zserv.h: add zebra_route_map_init
---
 zebra/rt_netlink.c | 84 +++++++++++++++++++++++++++++++++++++++++++++---------
 1 file changed, 71 insertions(+), 13 deletions(-)

(limited to 'zebra/rt_netlink.c')

diff --git a/zebra/rt_netlink.c b/zebra/rt_netlink.c
index e2f1f9d9..3b602c45 100644
--- a/zebra/rt_netlink.c
+++ b/zebra/rt_netlink.c
@@ -725,6 +725,7 @@ netlink_routing_table (struct sockaddr_nl *snl, struct nlmsghdr *h)
 
   void *dest;
   void *gate;
+  void *src;
 
   rtm = NLMSG_DATA (h);
 
@@ -764,6 +765,7 @@ netlink_routing_table (struct sockaddr_nl *snl, struct nlmsghdr *h)
   metric = 0;
   dest = NULL;
   gate = NULL;
+  src = NULL;
 
   if (tb[RTA_OIF])
     index = *(int *) RTA_DATA (tb[RTA_OIF]);
@@ -773,6 +775,9 @@ netlink_routing_table (struct sockaddr_nl *snl, struct nlmsghdr *h)
   else
     dest = anyaddr;
 
+  if (tb[RTA_PREFSRC])
+    src = RTA_DATA (tb[RTA_PREFSRC]);
+
   /* Multipath treatment is needed. */
   if (tb[RTA_GATEWAY])
     gate = RTA_DATA (tb[RTA_GATEWAY]);
@@ -787,7 +792,7 @@ netlink_routing_table (struct sockaddr_nl *snl, struct nlmsghdr *h)
       memcpy (&p.prefix, dest, 4);
       p.prefixlen = rtm->rtm_dst_len;
 
-      rib_add_ipv4 (ZEBRA_ROUTE_KERNEL, flags, &p, gate, index, table, metric, 0);
+      rib_add_ipv4 (ZEBRA_ROUTE_KERNEL, flags, &p, gate, src, index, table, metric, 0);
     }
 #ifdef HAVE_IPV6
   if (rtm->rtm_family == AF_INET6)
@@ -834,6 +839,7 @@ netlink_route_change (struct sockaddr_nl *snl, struct nlmsghdr *h)
   int table;
   void *dest;
   void *gate;
+  void *src;
 
   rtm = NLMSG_DATA (h);
 
@@ -890,6 +896,7 @@ netlink_route_change (struct sockaddr_nl *snl, struct nlmsghdr *h)
   index = 0;
   dest = NULL;
   gate = NULL;
+  src = NULL;
 
   if (tb[RTA_OIF])
     index = *(int *) RTA_DATA (tb[RTA_OIF]);
@@ -902,6 +909,9 @@ netlink_route_change (struct sockaddr_nl *snl, struct nlmsghdr *h)
   if (tb[RTA_GATEWAY])
     gate = RTA_DATA (tb[RTA_GATEWAY]);
 
+  if (tb[RTA_PREFSRC])
+    src = RTA_DATA (tb[RTA_PREFSRC]);
+
   if (rtm->rtm_family == AF_INET)
     {
       struct prefix_ipv4 p;
@@ -920,7 +930,7 @@ netlink_route_change (struct sockaddr_nl *snl, struct nlmsghdr *h)
         }
 
       if (h->nlmsg_type == RTM_NEWROUTE)
-        rib_add_ipv4 (ZEBRA_ROUTE_KERNEL, 0, &p, gate, index, table, 0, 0);
+        rib_add_ipv4 (ZEBRA_ROUTE_KERNEL, 0, &p, gate, src, index, table, 0, 0);
       else
         rib_delete_ipv4 (ZEBRA_ROUTE_KERNEL, 0, &p, gate, index, table);
     }
@@ -1489,7 +1499,9 @@ netlink_route_multipath (int cmd, struct prefix *p, struct rib *rib,
 		    {
 		      addattr_l (&req.n, sizeof req, RTA_GATEWAY,
 				 &nexthop->rgate.ipv4, bytelen);
-
+                      if (nexthop->src.ipv4.s_addr)
+		          addattr_l(&req.n, sizeof req, RTA_PREFSRC,
+				     &nexthop->src.ipv4, bytelen);
 		      if (IS_ZEBRA_DEBUG_KERNEL)
 			zlog_debug("netlink_route_multipath() (recursive, "
 				   "1 hop): nexthop via %s if %u",
@@ -1519,6 +1531,11 @@ netlink_route_multipath (int cmd, struct prefix *p, struct rib *rib,
 		    {
 		      addattr32 (&req.n, sizeof req, RTA_OIF,
 				 nexthop->rifindex);
+                      if ((nexthop->rtype == NEXTHOP_TYPE_IPV4_IFINDEX
+                           || nexthop->rtype == NEXTHOP_TYPE_IFINDEX)
+                          && nexthop->src.ipv4.s_addr)
+                        addattr_l (&req.n, sizeof req, RTA_PREFSRC,
+				 &nexthop->src.ipv4, bytelen);
 
 		      if (IS_ZEBRA_DEBUG_KERNEL)
 			zlog_debug("netlink_route_multipath() (recursive, "
@@ -1547,6 +1564,9 @@ netlink_route_multipath (int cmd, struct prefix *p, struct rib *rib,
 		    {
 		      addattr_l (&req.n, sizeof req, RTA_GATEWAY,
 				 &nexthop->gate.ipv4, bytelen);
+		      if (nexthop->src.ipv4.s_addr)
+                        addattr_l (&req.n, sizeof req, RTA_PREFSRC,
+				 &nexthop->src.ipv4, bytelen);
 
 		      if (IS_ZEBRA_DEBUG_KERNEL)
 			zlog_debug("netlink_route_multipath() (single hop): "
@@ -1571,8 +1591,19 @@ netlink_route_multipath (int cmd, struct prefix *p, struct rib *rib,
 #endif /* HAVE_IPV6 */
                   if (nexthop->type == NEXTHOP_TYPE_IFINDEX
                       || nexthop->type == NEXTHOP_TYPE_IFNAME
-                      || nexthop->type == NEXTHOP_TYPE_IPV4_IFINDEX
-                      || nexthop->type == NEXTHOP_TYPE_IPV6_IFINDEX
+                      || nexthop->type == NEXTHOP_TYPE_IPV4_IFINDEX)
+		    {
+		      addattr32 (&req.n, sizeof req, RTA_OIF, nexthop->ifindex);
+
+		      if (nexthop->src.ipv4.s_addr)
+                        addattr_l (&req.n, sizeof req, RTA_PREFSRC,
+				 &nexthop->src.ipv4, bytelen);
+
+		      if (IS_ZEBRA_DEBUG_KERNEL)
+			zlog_debug("netlink_route_multipath() (single hop): "
+				   "nexthop via if %u", nexthop->ifindex);
+		    }
+                  else if (nexthop->type == NEXTHOP_TYPE_IPV6_IFINDEX
                       || nexthop->type == NEXTHOP_TYPE_IPV6_IFNAME)
 		    {
 		      addattr32 (&req.n, sizeof req, RTA_OIF, nexthop->ifindex);
@@ -1596,6 +1627,7 @@ netlink_route_multipath (int cmd, struct prefix *p, struct rib *rib,
       char buf[1024];
       struct rtattr *rta = (void *) buf;
       struct rtnexthop *rtnh;
+      union g_addr *src = NULL;
 
       rta->rta_type = RTA_MULTIPATH;
       rta->rta_len = RTA_LENGTH (0);
@@ -1640,6 +1672,9 @@ netlink_route_multipath (int cmd, struct prefix *p, struct rib *rib,
                                      &nexthop->rgate.ipv4, bytelen);
                       rtnh->rtnh_len += sizeof (struct rtattr) + 4;
 
+		      if (nexthop->src.ipv4.s_addr)
+                        src = &nexthop->src;
+
 		      if (IS_ZEBRA_DEBUG_KERNEL)
 			zlog_debug("netlink_route_multipath() (recursive, "
 				   "multihop): nexthop via %s if %u",
@@ -1662,10 +1697,20 @@ netlink_route_multipath (int cmd, struct prefix *p, struct rib *rib,
 		    }
 #endif /* HAVE_IPV6 */
                   /* ifindex */
-                  if (nexthop->rtype == NEXTHOP_TYPE_IFINDEX
-                      || nexthop->rtype == NEXTHOP_TYPE_IFNAME
-                      || nexthop->rtype == NEXTHOP_TYPE_IPV4_IFINDEX
-                      || nexthop->rtype == NEXTHOP_TYPE_IPV6_IFINDEX
+                  if (nexthop->rtype == NEXTHOP_TYPE_IPV4_IFINDEX
+		      || nexthop->rtype == NEXTHOP_TYPE_IFINDEX
+                      || nexthop->rtype == NEXTHOP_TYPE_IFNAME)
+		    {
+		      rtnh->rtnh_ifindex = nexthop->rifindex;
+                      if (nexthop->src.ipv4.s_addr)
+                        src = &nexthop->src;
+
+		      if (IS_ZEBRA_DEBUG_KERNEL)
+			zlog_debug("netlink_route_multipath() (recursive, "
+				   "multihop): nexthop via if %u",
+				   nexthop->rifindex);
+		    }
+		  else if (nexthop->rtype == NEXTHOP_TYPE_IPV6_IFINDEX
                       || nexthop->rtype == NEXTHOP_TYPE_IPV6_IFNAME)
 		    {
 		      rtnh->rtnh_ifindex = nexthop->rifindex;
@@ -1701,6 +1746,9 @@ netlink_route_multipath (int cmd, struct prefix *p, struct rib *rib,
 				     &nexthop->gate.ipv4, bytelen);
 		      rtnh->rtnh_len += sizeof (struct rtattr) + 4;
 
+		      if (nexthop->src.ipv4.s_addr)
+                        src = &nexthop->src;
+
                       if (IS_ZEBRA_DEBUG_KERNEL)
 			zlog_debug("netlink_route_multipath() (multihop): "
 				   "nexthop via %s if %u",
@@ -1723,10 +1771,18 @@ netlink_route_multipath (int cmd, struct prefix *p, struct rib *rib,
 		    }
 #endif /* HAVE_IPV6 */
                   /* ifindex */
-                  if (nexthop->type == NEXTHOP_TYPE_IFINDEX
-                      || nexthop->type == NEXTHOP_TYPE_IFNAME
-                      || nexthop->type == NEXTHOP_TYPE_IPV4_IFINDEX
-                      || nexthop->type == NEXTHOP_TYPE_IPV6_IFNAME
+                  if (nexthop->type == NEXTHOP_TYPE_IPV4_IFINDEX
+		      || nexthop->type == NEXTHOP_TYPE_IFINDEX
+                      || nexthop->type == NEXTHOP_TYPE_IFNAME)
+                    {
+		      rtnh->rtnh_ifindex = nexthop->ifindex;
+		      if (nexthop->src.ipv4.s_addr)
+			src = &nexthop->src;
+		      if (IS_ZEBRA_DEBUG_KERNEL)
+			zlog_debug("netlink_route_multipath() (multihop): "
+				   "nexthop via if %u", nexthop->ifindex);
+		    }
+                  else if (nexthop->type == NEXTHOP_TYPE_IPV6_IFNAME
                       || nexthop->type == NEXTHOP_TYPE_IPV6_IFINDEX)
 		    {
 		      rtnh->rtnh_ifindex = nexthop->ifindex;
@@ -1746,6 +1802,8 @@ netlink_route_multipath (int cmd, struct prefix *p, struct rib *rib,
                 SET_FLAG (nexthop->flags, NEXTHOP_FLAG_FIB);
             }
         }
+      if (src)
+        addattr_l (&req.n, sizeof req, RTA_PREFSRC, &src->ipv4, bytelen);
 
       if (rta->rta_len > RTA_LENGTH (0))
         addattr_l (&req.n, 1024, RTA_MULTIPATH, RTA_DATA (rta),
-- 
cgit v1.2.3


From 068fd77cff0ce09b6390641abd84fc09a89c49db Mon Sep 17 00:00:00 2001
From: vize <vize>
Date: Fri, 10 Aug 2007 06:25:20 +0000
Subject: Use the proper field length for the peer's address
 (netlink_interface_addr)

---
 zebra/rt_netlink.c | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)

(limited to 'zebra/rt_netlink.c')

diff --git a/zebra/rt_netlink.c b/zebra/rt_netlink.c
index 3b602c45..5b592f94 100644
--- a/zebra/rt_netlink.c
+++ b/zebra/rt_netlink.c
@@ -647,12 +647,8 @@ netlink_interface_addr (struct sockaddr_nl *snl, struct nlmsghdr *h)
   addr = (tb[IFA_LOCAL] ? RTA_DATA(tb[IFA_LOCAL]) : NULL);
 
   /* is there a peer address? */
-  /* N.B. I do not understand why the memcmp compares 4 bytes regardless
-     of address family, but this is exactly how it appears in
-     print_addrinfo.  I wonder if it should be RTA_PAYLOAD(tb[IFA_ADDRESS])
-     instead of 4... */
   if (tb[IFA_ADDRESS] &&
-      memcmp(RTA_DATA(tb[IFA_ADDRESS]), RTA_DATA(tb[IFA_LOCAL]), 4))
+      memcmp(RTA_DATA(tb[IFA_ADDRESS]), RTA_DATA(tb[IFA_LOCAL]), RTA_PAYLOAD(tb[IFA_ADDRESS])))
     {
       broad = RTA_DATA(tb[IFA_ADDRESS]);
       SET_FLAG (flags, ZEBRA_IFA_PEER);
-- 
cgit v1.2.3


From 768a27ea7ba25257c79689af83ab37945dc1cc3c Mon Sep 17 00:00:00 2001
From: Paul Jakma <paul.jakma@sun.com>
Date: Thu, 29 May 2008 18:23:08 +0000
Subject: [zebra/linux] Use BPF to filter out responses, to try avoid netlink
 overruns

2008-05-29 Stephen Hemminger <stephen.hemminger@vyatta.com>

	* rt_netlink.c: (netlink_install_filter) BPF filter to catch and
	  drop responses to zebra's own route messages.
	  (kernel_init) add BPF filter on the netlink socket.
---
 zebra/rt_netlink.c | 55 +++++++++++++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 54 insertions(+), 1 deletion(-)

(limited to 'zebra/rt_netlink.c')

diff --git a/zebra/rt_netlink.c b/zebra/rt_netlink.c
index 5b592f94..0bf2d9eb 100644
--- a/zebra/rt_netlink.c
+++ b/zebra/rt_netlink.c
@@ -1938,6 +1938,56 @@ kernel_read (struct thread *thread)
   return 0;
 }
 
+/* Filter out messages from self that occur on listener socket */
+static void netlink_install_filter (int sock)
+{
+  /*
+   * Filter is equivalent to netlink_route_change
+   *
+   * if (h->nlmsg_type == RTM_DELROUTE || h->nlmsg_type == RTM_NEWROUTE) {
+   *    if (rtm->rtm_type != RTM_UNICAST)
+   *    	return 0;
+   *    if (rtm->rtm_flags & RTM_F_CLONED)
+   *    	return 0;
+   *    if (rtm->rtm_protocol == RTPROT_REDIRECT)
+   *    	return 0;
+   *    if (rtm->rtm_protocol == RTPROT_KERNEL)
+   *        return 0;
+   *    if (rtm->rtm_protocol == RTPROT_ZEBRA && h->nlmsg_type == RTM_NEWROUTE)
+   * 	return 0;
+   * }
+   * return 0xffff;
+   */
+  struct sock_filter filter[] = {
+    /* 0*/ BPF_STMT(BPF_LD|BPF_ABS|BPF_H, offsetof(struct nlmsghdr, nlmsg_type)),
+    /* 1*/ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, htons(RTM_DELROUTE), 1, 0),
+    /* 2*/ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, htons(RTM_NEWROUTE), 0, 11),
+    /* 3*/ BPF_STMT(BPF_LD|BPF_ABS|BPF_B,
+		    sizeof(struct nlmsghdr) + offsetof(struct rtmsg, rtm_type)),
+    /* 4*/ BPF_JUMP(BPF_JMP|BPF_B, RTN_UNICAST, 0, 8),
+    /* 5*/ BPF_STMT(BPF_LD|BPF_ABS|BPF_B,
+		    sizeof(struct nlmsghdr) + offsetof(struct rtmsg, rtm_flags)),
+    /* 6*/ BPF_JUMP(BPF_JMP|BPF_JSET|BPF_K, RTM_F_CLONED, 6, 0),
+    /* 7*/ BPF_STMT(BPF_LD|BPF_ABS|BPF_B,
+		    sizeof(struct nlmsghdr) + offsetof(struct rtmsg, rtm_protocol)),
+    /* 8*/ BPF_JUMP(BPF_JMP+ BPF_B, RTPROT_REDIRECT, 4, 0),
+    /* 9*/ BPF_JUMP(BPF_JMP+ BPF_B, RTPROT_KERNEL, 0, 1),
+    /*10*/ BPF_JUMP(BPF_JMP+ BPF_B, RTPROT_ZEBRA, 0, 3),
+    /*11*/ BPF_STMT(BPF_LD|BPF_ABS|BPF_H, offsetof(struct nlmsghdr, nlmsg_type)),
+    /*12*/ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, htons(RTM_NEWROUTE), 0, 1),
+    /*13*/ BPF_STMT(BPF_RET|BPF_K, 0),		/* drop */
+    /*14*/ BPF_STMT(BPF_RET|BPF_K, 0xffff),	/* keep */
+  };
+
+  struct sock_fprog prog = {
+    .len = sizeof(filter) / sizeof(filter[0]),
+    .filter = filter,
+  };
+
+  if (setsockopt(sock, SOL_SOCKET, SO_ATTACH_FILTER, &prog, sizeof(prog)) < 0)
+    zlog_warn ("Can't install socket filter: %s\n", safe_strerror(errno));
+}
+
 /* Exported interface function.  This function simply calls
    netlink_socket (). */
 void
@@ -1954,5 +2004,8 @@ kernel_init (void)
 
   /* Register kernel socket. */
   if (netlink.sock > 0)
-    thread_add_read (zebrad.master, kernel_read, NULL, netlink.sock);
+    {
+      netlink_install_filter (netlink.sock);
+      thread_add_read (zebrad.master, kernel_read, NULL, netlink.sock);
+    }
 }
-- 
cgit v1.2.3


From 40da22166ff29753a65b7947ed5fa7261fee1d80 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Timo=20Ter=C3=A4s?= <timo.teras@iki.fi>
Date: Wed, 13 Aug 2008 17:37:14 +0100
Subject: [zebra:netlink] Set proto/scope on all route update messages
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

2008-08-13 Timo Teräs <timo.teras@iki.fi>

	* zebra/rt_netlink.c: (netlink_route{,_multipath) Set protocol and
	  scope appropriately for both NEW and DELROUTE messages, to avoid
	  wiping out non-zebra routes on delete.

Signed-off-by: Paul Jakma <paul@quagga.net>
---
 zebra/rt_netlink.c | 10 ++++------
 1 file changed, 4 insertions(+), 6 deletions(-)

(limited to 'zebra/rt_netlink.c')

diff --git a/zebra/rt_netlink.c b/zebra/rt_netlink.c
index 0bf2d9eb..4234a445 100644
--- a/zebra/rt_netlink.c
+++ b/zebra/rt_netlink.c
@@ -1348,6 +1348,8 @@ netlink_route (int cmd, int family, void *dest, int length, void *gate,
   req.r.rtm_family = family;
   req.r.rtm_table = table;
   req.r.rtm_dst_len = length;
+  req.r.rtm_protocol = RTPROT_ZEBRA;
+  req.r.rtm_scope = RT_SCOPE_UNIVERSE;
 
   if ((zebra_flags & ZEBRA_FLAG_BLACKHOLE)
       || (zebra_flags & ZEBRA_FLAG_REJECT))
@@ -1357,9 +1359,6 @@ netlink_route (int cmd, int family, void *dest, int length, void *gate,
 
   if (cmd == RTM_NEWROUTE)
     {
-      req.r.rtm_protocol = RTPROT_ZEBRA;
-      req.r.rtm_scope = RT_SCOPE_UNIVERSE;
-
       if (discard)
         {
           if (zebra_flags & ZEBRA_FLAG_BLACKHOLE)
@@ -1424,6 +1423,8 @@ netlink_route_multipath (int cmd, struct prefix *p, struct rib *rib,
   req.r.rtm_family = family;
   req.r.rtm_table = rib->table;
   req.r.rtm_dst_len = p->prefixlen;
+  req.r.rtm_protocol = RTPROT_ZEBRA;
+  req.r.rtm_scope = RT_SCOPE_UNIVERSE;
 
   if ((rib->flags & ZEBRA_FLAG_BLACKHOLE) || (rib->flags & ZEBRA_FLAG_REJECT))
     discard = 1;
@@ -1432,9 +1433,6 @@ netlink_route_multipath (int cmd, struct prefix *p, struct rib *rib,
 
   if (cmd == RTM_NEWROUTE)
     {
-      req.r.rtm_protocol = RTPROT_ZEBRA;
-      req.r.rtm_scope = RT_SCOPE_UNIVERSE;
-
       if (discard)
         {
           if (rib->flags & ZEBRA_FLAG_BLACKHOLE)
-- 
cgit v1.2.3


From 6072b24eb776e6486f076b1762d3650e803954c2 Mon Sep 17 00:00:00 2001
From: Stephen Hemminger <stephen.hemminger@vyatta.com>
Date: Thu, 14 Aug 2008 16:52:26 +0100
Subject: [zebra] mark some rt_netlink functions static

2008-08-14 Stephen Hemminger <stephen.hemminger@vyatta.com>

	* zebra/rt_netlink.c: local functions should be static.

Signed-off-by: Paul Jakma <paul@quagga.net>
---
 zebra/rt_netlink.c | 37 ++++++++++++++++++-------------------
 1 file changed, 18 insertions(+), 19 deletions(-)

(limited to 'zebra/rt_netlink.c')

diff --git a/zebra/rt_netlink.c b/zebra/rt_netlink.c
index 4234a445..0a6b8dde 100644
--- a/zebra/rt_netlink.c
+++ b/zebra/rt_netlink.c
@@ -52,7 +52,7 @@ struct nlsock
 } netlink      = { -1, 0, {0}, "netlink-listen"},     /* kernel messages */
   netlink_cmd  = { -1, 0, {0}, "netlink-cmd"};        /* command channel */
 
-struct message nlmsg_str[] = {
+static struct message nlmsg_str[] = {
   {RTM_NEWROUTE, "RTM_NEWROUTE"},
   {RTM_DELROUTE, "RTM_DELROUTE"},
   {RTM_GETROUTE, "RTM_GETROUTE"},
@@ -65,7 +65,7 @@ struct message nlmsg_str[] = {
   {0, NULL}
 };
 
-const char *nexthop_types_desc[] =  
+static const char *nexthop_types_desc[] =
 {
   "none",
   "Directly connected",
@@ -79,7 +79,6 @@ const char *nexthop_types_desc[] =
   "Null0 nexthop",
 };
 
-
 extern struct zebra_t zebrad;
 
 extern struct zebra_privs_t zserv_privs;
@@ -222,7 +221,7 @@ netlink_socket (struct nlsock *nl, unsigned long groups)
   return ret;
 }
 
-int
+static int
 set_netlink_blocking (struct nlsock *nl, int *flags)
 {
 
@@ -243,7 +242,7 @@ set_netlink_blocking (struct nlsock *nl, int *flags)
   return 0;
 }
 
-int
+static int
 set_netlink_nonblocking (struct nlsock *nl, int *flags)
 {
   /* Restore socket flags for nonblocking I/O */
@@ -493,7 +492,7 @@ netlink_parse_rtattr (struct rtattr **tb, int max, struct rtattr *rta,
 
 /* Called from interface_lookup_netlink().  This function is only used
    during bootstrap. */
-int
+static int
 netlink_interface (struct sockaddr_nl *snl, struct nlmsghdr *h)
 {
   int len;
@@ -570,7 +569,7 @@ netlink_interface (struct sockaddr_nl *snl, struct nlmsghdr *h)
 }
 
 /* Lookup interface IPv4/IPv6 address. */
-int
+static int
 netlink_interface_addr (struct sockaddr_nl *snl, struct nlmsghdr *h)
 {
   int len;
@@ -705,7 +704,7 @@ netlink_interface_addr (struct sockaddr_nl *snl, struct nlmsghdr *h)
 }
 
 /* Looking up routing table by netlink interface. */
-int
+static int
 netlink_routing_table (struct sockaddr_nl *snl, struct nlmsghdr *h)
 {
   int len;
@@ -822,7 +821,7 @@ struct message rtproto_str[] = {
 };
 
 /* Routing information change from the kernel. */
-int
+static int
 netlink_route_change (struct sockaddr_nl *snl, struct nlmsghdr *h)
 {
   int len;
@@ -963,7 +962,7 @@ netlink_route_change (struct sockaddr_nl *snl, struct nlmsghdr *h)
   return 0;
 }
 
-int
+static int
 netlink_link_change (struct sockaddr_nl *snl, struct nlmsghdr *h)
 {
   int len;
@@ -1064,7 +1063,7 @@ netlink_link_change (struct sockaddr_nl *snl, struct nlmsghdr *h)
   return 0;
 }
 
-int
+static int
 netlink_information_fetch (struct sockaddr_nl *snl, struct nlmsghdr *h)
 {
   switch (h->nlmsg_type)
@@ -1189,7 +1188,7 @@ netlink_route_read (void)
 
 /* Utility function  comes from iproute2. 
    Authors:	Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru> */
-int
+static int
 addattr_l (struct nlmsghdr *n, int maxlen, int type, void *data, int alen)
 {
   int len;
@@ -1209,7 +1208,7 @@ addattr_l (struct nlmsghdr *n, int maxlen, int type, void *data, int alen)
   return 0;
 }
 
-int
+static int
 rta_addattr_l (struct rtattr *rta, int maxlen, int type, void *data, int alen)
 {
   int len;
@@ -1231,7 +1230,7 @@ rta_addattr_l (struct rtattr *rta, int maxlen, int type, void *data, int alen)
 
 /* Utility function comes from iproute2. 
    Authors:	Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru> */
-int
+static int
 addattr32 (struct nlmsghdr *n, int maxlen, int type, int data)
 {
   int len;
@@ -1259,7 +1258,7 @@ netlink_talk_filter (struct sockaddr_nl *snl, struct nlmsghdr *h)
 }
 
 /* sendmsg() to netlink socket then recvmsg(). */
-int
+static int
 netlink_talk (struct nlmsghdr *n, struct nlsock *nl)
 {
   int status;
@@ -1322,7 +1321,7 @@ netlink_talk (struct nlmsghdr *n, struct nlsock *nl)
 }
 
 /* Routing table change via netlink interface. */
-int
+static int
 netlink_route (int cmd, int family, void *dest, int length, void *gate,
                int index, int zebra_flags, int table)
 {
@@ -1396,7 +1395,7 @@ netlink_route (int cmd, int family, void *dest, int length, void *gate,
 }
 
 /* Routing table change via netlink interface. */
-int
+static int
 netlink_route_multipath (int cmd, struct prefix *p, struct rib *rib,
                          int family)
 {
@@ -1858,7 +1857,7 @@ kernel_delete_ipv6_old (struct prefix_ipv6 *dest, struct in6_addr *gate,
 #endif /* HAVE_IPV6 */
 
 /* Interface address modification. */
-int
+static int
 netlink_address (int cmd, int family, struct interface *ifp,
                  struct connected *ifc)
 {
@@ -1923,7 +1922,7 @@ kernel_address_delete_ipv4 (struct interface *ifp, struct connected *ifc)
 extern struct thread_master *master;
 
 /* Kernel route reflection. */
-int
+static int
 kernel_read (struct thread *thread)
 {
   int ret;
-- 
cgit v1.2.3


From 1423c809cc4ddc2e013ba6264c49a11e5719c6f2 Mon Sep 17 00:00:00 2001
From: Stephen Hemminger <stephen.hemminger@vyatta.com>
Date: Thu, 14 Aug 2008 17:59:25 +0100
Subject: [lib] mes_lookup string lookup table argument should be marked const

2008-08-14 Stephen Hemminger <stephen.hemminger@vyatta.com>

	* lib/log.{c,h}: struct message argument should point to const
	* */*.c: adjust to suit,

Signed-off-by: Paul Jakma <paul@quagga.net>
---
 zebra/rt_netlink.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

(limited to 'zebra/rt_netlink.c')

diff --git a/zebra/rt_netlink.c b/zebra/rt_netlink.c
index 0a6b8dde..395f1ca6 100644
--- a/zebra/rt_netlink.c
+++ b/zebra/rt_netlink.c
@@ -52,7 +52,7 @@ struct nlsock
 } netlink      = { -1, 0, {0}, "netlink-listen"},     /* kernel messages */
   netlink_cmd  = { -1, 0, {0}, "netlink-cmd"};        /* command channel */
 
-static struct message nlmsg_str[] = {
+static const struct message nlmsg_str[] = {
   {RTM_NEWROUTE, "RTM_NEWROUTE"},
   {RTM_DELROUTE, "RTM_DELROUTE"},
   {RTM_GETROUTE, "RTM_GETROUTE"},
@@ -805,7 +805,7 @@ netlink_routing_table (struct sockaddr_nl *snl, struct nlmsghdr *h)
   return 0;
 }
 
-struct message rtproto_str[] = {
+static const struct message rtproto_str[] = {
   {RTPROT_REDIRECT, "redirect"},
   {RTPROT_KERNEL,   "kernel"},
   {RTPROT_BOOT,     "boot"},
-- 
cgit v1.2.3


From 3d265b4d9d748bf4c92aefebc2ca0c04fd607945 Mon Sep 17 00:00:00 2001
From: Stephen Hemminger <stephen.hemminger@vyatta.com>
Date: Sat, 16 Aug 2008 17:30:39 +0100
Subject: [zebra:linux] Filter out unwanted netlink messages by PID

2008-08-16 Stephen Hemminger <stephen.hemminger@vyatta.com>

	* zebra/rt_netlink.c: Update the netlink socket BPF to cancel out
	  unwanted ACKs on the listen socket by matching on the PID.

Signed-off-by: Paul Jakma <paul@quagga.net>
---
 zebra/rt_netlink.c | 73 ++++++++++++++++++++----------------------------------
 1 file changed, 27 insertions(+), 46 deletions(-)

(limited to 'zebra/rt_netlink.c')

diff --git a/zebra/rt_netlink.c b/zebra/rt_netlink.c
index 395f1ca6..592dfe2e 100644
--- a/zebra/rt_netlink.c
+++ b/zebra/rt_netlink.c
@@ -285,7 +285,7 @@ netlink_request (int family, int type, struct nlsock *nl)
   req.nlh.nlmsg_len = sizeof req;
   req.nlh.nlmsg_type = type;
   req.nlh.nlmsg_flags = NLM_F_ROOT | NLM_F_MATCH | NLM_F_REQUEST;
-  req.nlh.nlmsg_pid = 0;
+  req.nlh.nlmsg_pid = nl->snl.nl_pid;
   req.nlh.nlmsg_seq = ++nl->seq;
   req.g.rtgen_family = family;
 
@@ -367,13 +367,6 @@ netlink_parse_info (int (*filter) (struct sockaddr_nl *, struct nlmsghdr *),
           return -1;
         }
       
-      /* JF: Ignore messages that aren't from the kernel */
-      if ( snl.nl_pid != 0 )
-        {
-          zlog ( NULL, LOG_ERR, "Ignoring message from pid %u", snl.nl_pid );
-          continue;
-        }
-
       for (h = (struct nlmsghdr *) buf; NLMSG_OK (h, (unsigned int) status);
            h = NLMSG_NEXT (h, status))
         {
@@ -1066,6 +1059,13 @@ netlink_link_change (struct sockaddr_nl *snl, struct nlmsghdr *h)
 static int
 netlink_information_fetch (struct sockaddr_nl *snl, struct nlmsghdr *h)
 {
+  /* JF: Ignore messages that aren't from the kernel */
+  if ( snl->nl_pid != 0 )
+    {
+      zlog ( NULL, LOG_ERR, "Ignoring message from pid %u", snl->nl_pid );
+      return 0;
+    }
+
   switch (h->nlmsg_type)
     {
     case RTM_NEWROUTE:
@@ -1935,45 +1935,26 @@ kernel_read (struct thread *thread)
   return 0;
 }
 
-/* Filter out messages from self that occur on listener socket */
-static void netlink_install_filter (int sock)
+/* Filter out messages from self that occur on listener socket,
+   caused by our actions on the command socket
+ */
+static void netlink_install_filter (int sock, __u32 pid)
 {
-  /*
-   * Filter is equivalent to netlink_route_change
-   *
-   * if (h->nlmsg_type == RTM_DELROUTE || h->nlmsg_type == RTM_NEWROUTE) {
-   *    if (rtm->rtm_type != RTM_UNICAST)
-   *    	return 0;
-   *    if (rtm->rtm_flags & RTM_F_CLONED)
-   *    	return 0;
-   *    if (rtm->rtm_protocol == RTPROT_REDIRECT)
-   *    	return 0;
-   *    if (rtm->rtm_protocol == RTPROT_KERNEL)
-   *        return 0;
-   *    if (rtm->rtm_protocol == RTPROT_ZEBRA && h->nlmsg_type == RTM_NEWROUTE)
-   * 	return 0;
-   * }
-   * return 0xffff;
-   */
   struct sock_filter filter[] = {
-    /* 0*/ BPF_STMT(BPF_LD|BPF_ABS|BPF_H, offsetof(struct nlmsghdr, nlmsg_type)),
-    /* 1*/ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, htons(RTM_DELROUTE), 1, 0),
-    /* 2*/ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, htons(RTM_NEWROUTE), 0, 11),
-    /* 3*/ BPF_STMT(BPF_LD|BPF_ABS|BPF_B,
-		    sizeof(struct nlmsghdr) + offsetof(struct rtmsg, rtm_type)),
-    /* 4*/ BPF_JUMP(BPF_JMP|BPF_B, RTN_UNICAST, 0, 8),
-    /* 5*/ BPF_STMT(BPF_LD|BPF_ABS|BPF_B,
-		    sizeof(struct nlmsghdr) + offsetof(struct rtmsg, rtm_flags)),
-    /* 6*/ BPF_JUMP(BPF_JMP|BPF_JSET|BPF_K, RTM_F_CLONED, 6, 0),
-    /* 7*/ BPF_STMT(BPF_LD|BPF_ABS|BPF_B,
-		    sizeof(struct nlmsghdr) + offsetof(struct rtmsg, rtm_protocol)),
-    /* 8*/ BPF_JUMP(BPF_JMP+ BPF_B, RTPROT_REDIRECT, 4, 0),
-    /* 9*/ BPF_JUMP(BPF_JMP+ BPF_B, RTPROT_KERNEL, 0, 1),
-    /*10*/ BPF_JUMP(BPF_JMP+ BPF_B, RTPROT_ZEBRA, 0, 3),
-    /*11*/ BPF_STMT(BPF_LD|BPF_ABS|BPF_H, offsetof(struct nlmsghdr, nlmsg_type)),
-    /*12*/ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, htons(RTM_NEWROUTE), 0, 1),
-    /*13*/ BPF_STMT(BPF_RET|BPF_K, 0),		/* drop */
-    /*14*/ BPF_STMT(BPF_RET|BPF_K, 0xffff),	/* keep */
+    /* 0: ldh [4]	          */
+    BPF_STMT(BPF_LD|BPF_ABS|BPF_H, offsetof(struct nlmsghdr, nlmsg_type)),
+    /* 1: jeq 0x18 jt 3 jf 6  */
+    BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, htons(RTM_NEWROUTE), 1, 0),
+    /* 2: jeq 0x19 jt 3 jf 6  */
+    BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, htons(RTM_DELROUTE), 0, 3),
+    /* 3: ldw [12]		  */
+    BPF_STMT(BPF_LD|BPF_ABS|BPF_W, offsetof(struct nlmsghdr, nlmsg_pid)),
+    /* 4: jeq XX  jt 5 jf 6   */
+    BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, htonl(pid), 0, 1),
+    /* 5: ret 0    (skip)     */
+    BPF_STMT(BPF_RET|BPF_K, 0),
+    /* 6: ret 0xffff (keep)   */
+    BPF_STMT(BPF_RET|BPF_K, 0xffff),
   };
 
   struct sock_fprog prog = {
@@ -2002,7 +1983,7 @@ kernel_init (void)
   /* Register kernel socket. */
   if (netlink.sock > 0)
     {
-      netlink_install_filter (netlink.sock);
+      netlink_install_filter (netlink.sock, netlink_cmd.snl.nl_pid);
       thread_add_read (zebrad.master, kernel_read, NULL, netlink.sock);
     }
 }
-- 
cgit v1.2.3


From 4cde931e2d5a83708cf55e63eae7b629e8e4296a Mon Sep 17 00:00:00 2001
From: Stephen Hemminger <stephen.hemminger@vyatta.com>
Date: Sat, 16 Aug 2008 17:51:27 +0100
Subject: [zebra:linux] clean up blocking setting of netlink sockets

2008-08-16 Stephen Hemminger <stephen.hemminger@vyatta.com>

	* zebra/rt_netlink.c: This change removes the overhead of twiddling
          the socket twice on every route change. The netlink support has
          two sockets, one is used for request/response and should always
          block. The other is used for asynchronous notification and should
          always be non-blocking.

Signed-off-by: Paul Jakma <paul@quagga.net>
---
 zebra/rt_netlink.c | 97 ++++--------------------------------------------------
 1 file changed, 6 insertions(+), 91 deletions(-)

(limited to 'zebra/rt_netlink.c')

diff --git a/zebra/rt_netlink.c b/zebra/rt_netlink.c
index 592dfe2e..fe55d43e 100644
--- a/zebra/rt_netlink.c
+++ b/zebra/rt_netlink.c
@@ -130,14 +130,6 @@ netlink_socket (struct nlsock *nl, unsigned long groups)
       return -1;
     }
 
-  ret = fcntl (sock, F_SETFL, O_NONBLOCK);
-  if (ret < 0)
-    {
-      zlog (NULL, LOG_ERR, "Can't set %s socket flags: %s", nl->name,
-            safe_strerror (errno));
-      close (sock);
-      return -1;
-    }
 
   /* Set receive buffer size if it's set from command line */
   if (nl_rcvbufsize)
@@ -221,41 +213,6 @@ netlink_socket (struct nlsock *nl, unsigned long groups)
   return ret;
 }
 
-static int
-set_netlink_blocking (struct nlsock *nl, int *flags)
-{
-
-  /* Change socket flags for blocking I/O.  */
-  if ((*flags = fcntl (nl->sock, F_GETFL, 0)) < 0)
-    {
-      zlog (NULL, LOG_ERR, "%s:%i F_GETFL error: %s",
-            __FUNCTION__, __LINE__, safe_strerror (errno));
-      return -1;
-    }
-  *flags &= ~O_NONBLOCK;
-  if (fcntl (nl->sock, F_SETFL, *flags) < 0)
-    {
-      zlog (NULL, LOG_ERR, "%s:%i F_SETFL error: %s",
-            __FUNCTION__, __LINE__, safe_strerror (errno));
-      return -1;
-    }
-  return 0;
-}
-
-static int
-set_netlink_nonblocking (struct nlsock *nl, int *flags)
-{
-  /* Restore socket flags for nonblocking I/O */
-  *flags |= O_NONBLOCK;
-  if (fcntl (nl->sock, F_SETFL, *flags) < 0)
-    {
-      zlog (NULL, LOG_ERR, "%s:%i F_SETFL error: %s",
-            __FUNCTION__, __LINE__, safe_strerror (errno));
-      return -1;
-    }
-  return 0;
-}
-
 /* Get type specified information from netlink. */
 static int
 netlink_request (int family, int type, struct nlsock *nl)
@@ -1098,18 +1055,6 @@ int
 interface_lookup_netlink (void)
 {
   int ret;
-  int flags;
-  int snb_ret;
-
-  /* 
-   * Change netlink socket flags to blocking to ensure we get 
-   * a reply via nelink_parse_info
-   */
-  snb_ret = set_netlink_blocking (&netlink_cmd, &flags);
-  if (snb_ret < 0)
-    zlog (NULL, LOG_WARNING,
-          "%s:%i Warning: Could not set netlink socket to blocking.",
-          __FUNCTION__, __LINE__);
 
   /* Get interface information. */
   ret = netlink_request (AF_PACKET, RTM_GETLINK, &netlink_cmd);
@@ -1137,9 +1082,6 @@ interface_lookup_netlink (void)
     return ret;
 #endif /* HAVE_IPV6 */
 
-  /* restore socket flags */
-  if (snb_ret == 0)
-    set_netlink_nonblocking (&netlink_cmd, &flags);
   return 0;
 }
 
@@ -1149,18 +1091,6 @@ int
 netlink_route_read (void)
 {
   int ret;
-  int flags;
-  int snb_ret;
-
-  /* 
-   * Change netlink socket flags to blocking to ensure we get 
-   * a reply via nelink_parse_info
-   */
-  snb_ret = set_netlink_blocking (&netlink_cmd, &flags);
-  if (snb_ret < 0)
-    zlog (NULL, LOG_WARNING,
-          "%s:%i Warning: Could not set netlink socket to blocking.",
-          __FUNCTION__, __LINE__);
 
   /* Get IPv4 routing table. */
   ret = netlink_request (AF_INET, RTM_GETROUTE, &netlink_cmd);
@@ -1180,9 +1110,6 @@ netlink_route_read (void)
     return ret;
 #endif /* HAVE_IPV6 */
 
-  /* restore flags */
-  if (snb_ret == 0)
-    set_netlink_nonblocking (&netlink_cmd, &flags);
   return 0;
 }
 
@@ -1265,8 +1192,6 @@ netlink_talk (struct nlmsghdr *n, struct nlsock *nl)
   struct sockaddr_nl snl;
   struct iovec iov = { (void *) n, n->nlmsg_len };
   struct msghdr msg = { (void *) &snl, sizeof snl, &iov, 1, NULL, 0, 0 };
-  int flags = 0;
-  int snb_ret;
   int save_errno;
 
   memset (&snl, 0, sizeof snl);
@@ -1297,27 +1222,12 @@ netlink_talk (struct nlmsghdr *n, struct nlsock *nl)
       return -1;
     }
 
-  /* 
-   * Change socket flags for blocking I/O. 
-   * This ensures we wait for a reply in netlink_parse_info().
-   */
-  snb_ret = set_netlink_blocking (nl, &flags);
-  if (snb_ret < 0)
-    zlog (NULL, LOG_WARNING,
-          "%s:%i Warning: Could not set netlink socket to blocking.",
-          __FUNCTION__, __LINE__);
 
   /* 
    * Get reply from netlink socket. 
    * The reply should either be an acknowlegement or an error.
    */
-  status = netlink_parse_info (netlink_talk_filter, nl);
-
-  /* Restore socket flags for nonblocking I/O */
-  if (snb_ret == 0)
-    set_netlink_nonblocking (nl, &flags);
-
-  return status;
+  return netlink_parse_info (netlink_talk_filter, nl);
 }
 
 /* Routing table change via netlink interface. */
@@ -1983,6 +1893,11 @@ kernel_init (void)
   /* Register kernel socket. */
   if (netlink.sock > 0)
     {
+      /* Only want non-blocking on the netlink event socket */
+      if (fcntl (netlink.sock, F_SETFL, O_NONBLOCK) < 0)
+	zlog (NULL, LOG_ERR, "Can't set %s socket flags: %s", netlink.name,
+		safe_strerror (errno));
+
       netlink_install_filter (netlink.sock, netlink_cmd.snl.nl_pid);
       thread_add_read (zebrad.master, kernel_read, NULL, netlink.sock);
     }
-- 
cgit v1.2.3


From 30afea3b584515ba19b38fcde6f562a550c1fcab Mon Sep 17 00:00:00 2001
From: Stephen Hemminger <stephen.hemminger@vyatta.com>
Date: Sat, 16 Aug 2008 18:25:47 +0100
Subject: [zebra:linux] netlink buffer size tweaking only needed on receive
 socket

2008-08-16 Stephen Hemminger <stephen.hemminger@vyatta.com>

	* rt_netlink.c: (netlink_recvbuf) consolidate recvbuf setting
	  (netlink_socket) no need to set receive buffer on netlink sockets
	  generally
	  (kernel_init) set recvbuffer on the listen socket

Signed-off-by: Paul Jakma <paul@quagga.net>
---
 zebra/rt_netlink.c | 86 +++++++++++++++++++++++++++---------------------------
 1 file changed, 43 insertions(+), 43 deletions(-)

(limited to 'zebra/rt_netlink.c')

diff --git a/zebra/rt_netlink.c b/zebra/rt_netlink.c
index fe55d43e..a0dad997 100644
--- a/zebra/rt_netlink.c
+++ b/zebra/rt_netlink.c
@@ -112,6 +112,45 @@ set_ifindex(struct interface *ifp, unsigned int ifi_index)
   ifp->ifindex = ifi_index;
 }
 
+static int
+netlink_recvbuf (struct nlsock *nl, uint32_t newsize)
+{
+  u_int32_t oldsize;
+  socklen_t newlen = sizeof(newsize);
+  socklen_t oldlen = sizeof(oldsize);
+  int ret;
+
+  ret = getsockopt(nl->sock, SOL_SOCKET, SO_RCVBUF, &oldsize, &oldlen);
+  if (ret < 0)
+    {
+      zlog (NULL, LOG_ERR, "Can't get %s receive buffer size: %s", nl->name,
+	    safe_strerror (errno));
+      return -1;
+    }
+
+  ret = setsockopt(nl->sock, SOL_SOCKET, SO_RCVBUF, &nl_rcvbufsize,
+		   sizeof(nl_rcvbufsize));
+  if (ret < 0)
+    {
+      zlog (NULL, LOG_ERR, "Can't set %s receive buffer size: %s", nl->name,
+	    safe_strerror (errno));
+      return -1;
+    }
+
+  ret = getsockopt(nl->sock, SOL_SOCKET, SO_RCVBUF, &newsize, &newlen);
+  if (ret < 0)
+    {
+      zlog (NULL, LOG_ERR, "Can't get %s receive buffer size: %s", nl->name,
+	    safe_strerror (errno));
+      return -1;
+    }
+
+  zlog (NULL, LOG_INFO,
+	"Setting netlink socket receive buffer size: %u -> %u",
+	oldsize, newsize);
+  return 0;
+}
+
 /* Make socket for Linux netlink interface. */
 static int
 netlink_socket (struct nlsock *nl, unsigned long groups)
@@ -130,49 +169,6 @@ netlink_socket (struct nlsock *nl, unsigned long groups)
       return -1;
     }
 
-
-  /* Set receive buffer size if it's set from command line */
-  if (nl_rcvbufsize)
-    {
-      u_int32_t oldsize, oldlen;
-      u_int32_t newsize, newlen;
-
-      oldlen = sizeof(oldsize);
-      newlen = sizeof(newsize);
-
-      ret = getsockopt(sock, SOL_SOCKET, SO_RCVBUF, &oldsize, &oldlen);
-      if (ret < 0)
-	{
-	  zlog (NULL, LOG_ERR, "Can't get %s receive buffer size: %s", nl->name,
-		safe_strerror (errno));
-	  close (sock);
-	  return -1;
-	}
-
-      ret = setsockopt(sock, SOL_SOCKET, SO_RCVBUF, &nl_rcvbufsize,
-		       sizeof(nl_rcvbufsize));
-      if (ret < 0)
-	{
-	  zlog (NULL, LOG_ERR, "Can't set %s receive buffer size: %s", nl->name,
-		safe_strerror (errno));
-	  close (sock);
-	  return -1;
-	}
-
-      ret = getsockopt(sock, SOL_SOCKET, SO_RCVBUF, &newsize, &newlen);
-      if (ret < 0)
-	{
-	  zlog (NULL, LOG_ERR, "Can't get %s receive buffer size: %s", nl->name,
-		safe_strerror (errno));
-	  close (sock);
-	  return -1;
-	}
-
-      zlog (NULL, LOG_INFO,
-	    "Setting netlink socket receive buffer size: %u -> %u",
-	    oldsize, newsize);
-    }
-
   memset (&snl, 0, sizeof snl);
   snl.nl_family = AF_NETLINK;
   snl.nl_groups = groups;
@@ -1898,6 +1894,10 @@ kernel_init (void)
 	zlog (NULL, LOG_ERR, "Can't set %s socket flags: %s", netlink.name,
 		safe_strerror (errno));
 
+      /* Set receive buffer size if it's set from command line */
+      if (nl_rcvbufsize)
+	netlink_recvbuf (&netlink, nl_rcvbufsize);
+
       netlink_install_filter (netlink.sock, netlink_cmd.snl.nl_pid);
       thread_add_read (zebrad.master, kernel_read, NULL, netlink.sock);
     }
-- 
cgit v1.2.3


From 898987e976e8f7a50a4549475ab4970a85895f31 Mon Sep 17 00:00:00 2001
From: Stephen Hemminger <stephen.hemminger@vyatta.com>
Date: Sun, 17 Aug 2008 16:56:15 +0100
Subject: [zebra:linux] netlink: quiet harmless errors

2008-06-17 Stephen Hemminger <stephen.hemminger@vyatta.com>

	* rt_netlink.c: (netlink_parse_info) There can be harmless races
	  such as adding an existing route, or deleting an already deleted
	  route. Don't bother logging these unless debugging is turned on.

Signed-off-by: Paul Jakma <paul@quagga.net>
---
 zebra/rt_netlink.c | 41 ++++++++++++++++++++---------------------
 1 file changed, 20 insertions(+), 21 deletions(-)

(limited to 'zebra/rt_netlink.c')

diff --git a/zebra/rt_netlink.c b/zebra/rt_netlink.c
index a0dad997..904367e0 100644
--- a/zebra/rt_netlink.c
+++ b/zebra/rt_netlink.c
@@ -331,6 +331,8 @@ netlink_parse_info (int (*filter) (struct sockaddr_nl *, struct nlmsghdr *),
           if (h->nlmsg_type == NLMSG_ERROR)
             {
               struct nlmsgerr *err = (struct nlmsgerr *) NLMSG_DATA (h);
+	      int errnum = err->error;
+	      int msg_type = err->msg.nlmsg_type;
 
               /* If the error field is zero, then this is an ACK */
               if (err->error == 0)
@@ -359,27 +361,24 @@ netlink_parse_info (int (*filter) (struct sockaddr_nl *, struct nlmsghdr *),
                   return -1;
                 }
 
-              /* Deal with Error Noise  - MAG */
-              {
-                int loglvl = LOG_ERR;
-                int errnum = err->error;
-                int msg_type = err->msg.nlmsg_type;
-
-                if (nl == &netlink_cmd
-                    && (-errnum == ENODEV || -errnum == ESRCH)
-                    && (msg_type == RTM_NEWROUTE || msg_type == RTM_DELROUTE))
-                  loglvl = LOG_DEBUG;
-
-                zlog (NULL, loglvl, "%s error: %s, type=%s(%u), "
-                      "seq=%u, pid=%u",
-                      nl->name, safe_strerror (-errnum),
-                      lookup (nlmsg_str, msg_type),
-                      msg_type, err->msg.nlmsg_seq, err->msg.nlmsg_pid);
-              }
-              /*
-                 ret = -1;
-                 continue;
-               */
+              /* Deal with errors that occur because of races in link handling */
+	      if (nl == &netlink_cmd
+		  && ((msg_type == RTM_DELROUTE &&
+		       (-errnum == ENODEV || -errnum == ESRCH))
+		      || (msg_type == RTM_NEWROUTE && -errnum == EEXIST)))
+		{
+		  if (IS_ZEBRA_DEBUG_KERNEL)
+		    zlog_debug ("%s: error: %s type=%s(%u), seq=%u, pid=%u",
+				nl->name, safe_strerror (-errnum),
+				lookup (nlmsg_str, msg_type),
+				msg_type, err->msg.nlmsg_seq, err->msg.nlmsg_pid);
+		  return 0;
+		}
+
+	      zlog_err ("%s error: %s, type=%s(%u), seq=%u, pid=%u",
+			nl->name, safe_strerror (-errnum),
+			lookup (nlmsg_str, msg_type),
+			msg_type, err->msg.nlmsg_seq, err->msg.nlmsg_pid);
               return -1;
             }
 
-- 
cgit v1.2.3


From 4c699477d24e7104893130d072963df5cea17214 Mon Sep 17 00:00:00 2001
From: Stephen Hemminger <stephen.hemminger@vyatta.com>
Date: Sun, 17 Aug 2008 17:01:44 +0100
Subject: [zebra:linux] netlink: no need to change privs for receive

2008-06-17 Stephen Hemminger <stephen.hemminger@vyatta.com>

	* rt_netlink.c: (netlink_parse_info) Linux doesn't check privledges
	  on receive, so avoid the overhead of lowering and raising the
	  priviledges on each received message

Signed-off-by: Paul Jakma <paul@quagga.net>
---
 zebra/rt_netlink.c | 15 +++------------
 1 file changed, 3 insertions(+), 12 deletions(-)

(limited to 'zebra/rt_netlink.c')

diff --git a/zebra/rt_netlink.c b/zebra/rt_netlink.c
index 904367e0..05254498 100644
--- a/zebra/rt_netlink.c
+++ b/zebra/rt_netlink.c
@@ -285,25 +285,16 @@ netlink_parse_info (int (*filter) (struct sockaddr_nl *, struct nlmsghdr *),
       struct sockaddr_nl snl;
       struct msghdr msg = { (void *) &snl, sizeof snl, &iov, 1, NULL, 0, 0 };
       struct nlmsghdr *h;
-      int save_errno;
-
-      if (zserv_privs.change (ZPRIVS_RAISE))
-        zlog (NULL, LOG_ERR, "Can't raise privileges");
 
       status = recvmsg (nl->sock, &msg, 0);
-      save_errno = errno;
-
-      if (zserv_privs.change (ZPRIVS_LOWER))
-        zlog (NULL, LOG_ERR, "Can't lower privileges");
-
       if (status < 0)
         {
-          if (save_errno == EINTR)
+          if (errno == EINTR)
             continue;
-          if (save_errno == EWOULDBLOCK || save_errno == EAGAIN)
+          if (errno == EWOULDBLOCK || errno == EAGAIN)
             break;
           zlog (NULL, LOG_ERR, "%s recvmsg overrun: %s",
-	  	nl->name, safe_strerror(save_errno));
+	  	nl->name, safe_strerror(errno));
           continue;
         }
 
-- 
cgit v1.2.3