- Bird-users - Mailing Lists

rip is broken?
by Alexander V. Chernikov 23 Jul '12

23 Jul '12

Hello list! It seems that rip-ng changes introduced in f7615037 broke RIP so bird dumps core on first RIP packet: # gdb bird bird.core GNU gdb 6.1.1 [FreeBSD] Copyright 2004 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. Type "show copying" to see the conditions. There is absolutely no warranty for GDB. Type "show warranty" for details. This GDB was configured as "i386-marcel-freebsd"... Core was generated by `bird'. Program terminated with signal 6, Aborted. Reading symbols from /lib/libc.so.7...done. Loaded symbols for /lib/libc.so.7 Reading symbols from /libexec/ld-elf.so.1...done. Loaded symbols for /libexec/ld-elf.so.1 #0 0x281dbb97 in kill () from /lib/libc.so.7 (gdb) bt #0 0x281dbb97 in kill () from /lib/libc.so.7 #1 0x281dbaf6 in raise () from /lib/libc.so.7 #2 0x281da6ca in abort () from /lib/libc.so.7 #3 0x0808d5af in bug (msg=Could not find the frame base for "bug". ) at log.c:252 #4 0x080795df in advertise_entry (p=0x2831d310, b=0x283b8004, whotoldme=2886730113, iface=0x0) at ../../../proto/rip/rip.c:331 #5 0x080797f6 in process_block (p=0x2831d310, block=0x283b8004, whotoldme=2886730113, iface=0x0) at ../../../proto/rip/rip.c:387 #6 0x08079b00 in rip_process_packet (p=0x2831d310, packet=0x283b8000, num=2, whotoldme=2886730113, port=520, iface=0x0) at ../../../proto/rip/rip.c:450 #7 0x08079cd2 in rip_rx (s=0x283051c0, size=40) at ../../../proto/rip/rip.c:497 #8 0x08088cf9 in sk_read (s=0x283051c0) at io.c:1463 #9 0x080893eb in io_loop () at io.c:1676 #10 0x0808e7fe in main (argc=Cannot access memory at address 0x5 ) at main.c:699 rip_rx() now sends NULL interface name to rip_process_packet(). It seems that specifying interface makes rip working again: diff --git a/proto/rip/rip.c b/proto/rip/rip.c index 281296a..e4ccbff 100644 --- a/proto/rip/rip.c +++ b/proto/rip/rip.c @@ -477,9 +477,9 @@ rip_rx(sock *s, int size) #ifdef IPV6 if (! i->iface || s->lifindex != i->iface->index) return 1; +#endif iface = i->iface; -#endif CHK_MAGIC;

2 1

OSPF Fast Hello
by Javier Lorenzo Prieto 28 Jun '12

28 Jun '12

We are using BIRD to configure load balancing and redundancy between two sites. These sites are interconnected via two short-range RF links (each with 15 Mpbs effective throughput and delay of 5ms). Because we intend to send VoIP over these links and would like to avoid dropping calls in the event of interruptions in one of the links. Therefore, dynamic routes need to be updated very quickly. --------- SITE 1 -- R1 R2 -- SITE 2 --------- Does BIRD allow setting the parameters as follows? • dead 2 • hello 1 protocol ospf { import all; ecmp on; area 0.0.0.0 { interface "vlan200", "vlan201" { cost 5; check link on; type broadcast; hello 1; retransmit 4; wait 1; dead 2; authentication none; }; }; } When we try this configuration, our routers never reach the point of creating adjacency. However, everything works OK when setting dead=3 and hello=1. Furthermore the parameters dead=2 and hello=1 do work when using QUAGGA. Am I missing anything? Thanks in advance for your help!

3 3

Route selection using MED
by xianglong han 28 Jun '12

28 Jun '12

Hi, all I am building a BGP and OSPF network specified in RFC 3345 which will lead to a BGP oscillation. see below: --------------------------------------------------------------- / -------------------- -------------------- \ | / \ / \ | | | Cluster 1 | | Cluster 2 | | | | | | | | | | | *1 | | | | | Ra(RR) . . . . . . . . . . . . . . Rd(RR) | | | | . . | | . | | | | .*5 .*4 | | .*12 | | | | . . | | . | | | | Rb(C) Rc(C) | | Re(C) | | | | . . | | . | | | \ . . / \ . / | | ---.------------.--- ---------.---------- | \ .(10) .(1) AS1 .(0) / -------.------------.---------------------------.-------------- . . . ------ . ------------ . / \ . / \ . | AS10 | | AS6 | \ / \ / ------ ------------ . . . . . -------------- . / \ | AS100 |- 10.0.0.0/8 \ / -------------- The problem is, the router Rd does not perform as expected in selecting the best route. 10.199.0.0/16 via 10.1.7.2 on eth2 [SR 14:30] * (100-/5) [AS65002i] Type: BGP unicast univ BGP.origin: IGP BGP.as_path: 65002 BGP.next_hop: 10.1.5.2 BGP.med: 1 BGP.local_pref: 100 BGP.originator_id: 10.1.4.2 BGP.cluster_list: 10.1.3.3 via 10.1.1.2 on eth3 [SC 14:29] (100/12) [AS65002i] Type: BGP unicast univ BGP.origin: IGP BGP.as_path: 65002 BGP.next_hop: 10.1.1.2 BGP.med: 0 Should the second route in the above route table be better? as it has a lower med value with the local preference, AS path length and orgins. Thanks

1 0

Routing Question
by dspazman＠epicup.com 28 Jun '12

28 Jun '12

Hi all, I figured you guys were a great place to ask this question, even though it's not bird specific. If you have 2 discrete sites, both with BGP, one announces a 100.100.100.0/20, and another announcing 100.100.100.0/24, will traffic for the /24 block always go to the site announcing the /24, even if it's closer to the site announcing the /20? I was thinking that yes, it always routes to the shortest prefix, no matter the ping or number hops. And I was thinking that held true for ipv6, as well. A /32 and a /28 that cover the same range of addresses (the /32 being a subset of the /28) being announced from two different BGP sessions on two different networks, even if you are originating traffic from the same network the /28 is being announced on, and the /32 is a different network, so has additional hops and latency, any traffic you have going to the addresses in the /32 will go to the other network to go to the specifically announced /32. Thanks!

2 2

Receiving routes with own AS?
by Tom van Leeuwen 24 Jun '12

24 Jun '12

Hi guys, We have a new location and have installed a bgp server with bird. We are already using our AS and bird at the main site, but want to use that one as well on the second site. I'm not sure if I want/need to run iBGP, but if I had and the link would fail, I should have at least a route through IPT to our main site. I'm already connected through IPT at our new site, but I'm not getting any routes for our subnets that are being announced at our main site. Our upstream provider says that they are not filtering the routes and that it should be my configuration that is filtering them. Well, I have 'import all' so that should not be the problem. Does bird have some built in logic that does not accept routes with my own AS? Or is the IPT incorrect? Kind regards, Tom van Leeuwen

3 4

Patch for renaming sk_new()
by Wes Hardaker 18 Jun '12

18 Jun '12

The sk_new() function within bird is identically named to an openssl function. We needed to link against openssl to support BGPSEC, which Michael Baer wrote you about back in March, so I wrote a short patch to rename sk_new() to sk_new_sock() instead. It's attached below and I'd love it if it could be applied to the primary repository. Thanks and let me know if this would cause a problem! -- Wes Hardaker My Pictures: http://capturedonearth.com/ My Thoughts: http://pontifications.hardakers.net/

2 2

Multiple interface per link support in BIRD OSPFv3
by Benjamin Paterson 16 Jun '12

16 Jun '12

Is this supported? (RFC5340 section 4.9) If not, any plans to support it? It'd be great for example to maintain a pointer in each ospf_iface to the active interface for the link, or at least have a way to tell if two interfaces are on the same link.

2 1

Protocol BGP with two tables
by Jan Markus 13 Jun '12

13 Jun '12

Hello, can I somehow say in "protocol bgp peer_xyz {}" to import routes to table1 but export routes from table2 ? Thank you for your time. -Jan

2 2

[PATCH] allow bird with openssl by renaming the conflicting sk_new() to sk_new_sock()
by Wes Hardaker 08 Jun '12

08 Jun '12

--- lib/socket.h | 2 +- proto/bgp/bgp.c | 4 ++-- proto/ospf/iface.c | 2 +- proto/radv/packets.c | 2 +- proto/rip/rip.c | 2 +- sysdep/bsd/krt-sock.c | 2 +- sysdep/linux/netlink.c | 2 +- sysdep/unix/io.c | 8 ++++---- sysdep/unix/main.c | 2 +- 9 files changed, 13 insertions(+), 13 deletions(-) diff --git a/lib/socket.h b/lib/socket.h index b0c3eda..5b01c6b 100644 --- a/lib/socket.h +++ b/lib/socket.h @@ -48,7 +48,7 @@ typedef struct birdsock { char *password; /* Password for MD5 authentication */ } sock; -sock *sk_new(pool *); /* Allocate new socket */ +sock *sk_new_sock(pool *); /* Allocate new socket */ int sk_open(sock *); /* Open socket */ int sk_send(sock *, unsigned len); /* Send data, <0=err, >0=ok, 0=sleep */ int sk_send_to(sock *, unsigned len, ip_addr to, unsigned port); /* sk_send to given destination */ diff --git a/proto/bgp/bgp.c b/proto/bgp/bgp.c index 3b9f7cc..2b78e14 100644 --- a/proto/bgp/bgp.c +++ b/proto/bgp/bgp.c @@ -558,7 +558,7 @@ bgp_connect(struct bgp_proto *p) /* Enter Connect state and start establishing c int hops = p->cf->multihop ? : 1; DBG("BGP: Connecting\n"); - s = sk_new(p->p.pool); + s = sk_new_sock(p->p.pool); s->type = SK_TCP_ACTIVE; s->saddr = p->source_addr; s->daddr = p->cf->remote_ip; @@ -674,7 +674,7 @@ bgp_listen_sock_err(sock *sk UNUSED, int err) static sock * bgp_setup_listen_sk(ip_addr addr, unsigned port, u32 flags) { - sock *s = sk_new(&root_pool); + sock *s = sk_new_sock(&root_pool); DBG("BGP: Creating listening socket\n"); s->type = SK_TCP_PASSIVE; s->ttl = 255; diff --git a/proto/ospf/iface.c b/proto/ospf/iface.c index a6a0c6c..3593d85 100644 --- a/proto/ospf/iface.c +++ b/proto/ospf/iface.c @@ -72,7 +72,7 @@ find_nbma_node_in(list *nnl, ip_addr ip) static int ospf_sk_open(struct ospf_iface *ifa) { - sock *sk = sk_new(ifa->pool); + sock *sk = sk_new_sock(ifa->pool); sk->type = SK_IP; sk->dport = OSPF_PROTO; sk->saddr = IPA_NONE; diff --git a/proto/radv/packets.c b/proto/radv/packets.c index ac59ce9..e19e116 100644 --- a/proto/radv/packets.c +++ b/proto/radv/packets.c @@ -224,7 +224,7 @@ 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_sock(ifa->ra->p.pool); sk->type = SK_IP; sk->dport = ICMPV6_PROTO; sk->saddr = IPA_NONE; diff --git a/proto/rip/rip.c b/proto/rip/rip.c index 281296a..df08ea9 100644 --- a/proto/rip/rip.c +++ b/proto/rip/rip.c @@ -691,7 +691,7 @@ new_iface(struct proto *p, struct iface *new, unsigned long flags, struct iface_ if (rif->multicast) DBG( "Doing multicasts!\n" ); - rif->sock = sk_new( p->pool ); + rif->sock = sk_new_sock( p->pool ); rif->sock->type = SK_UDP; rif->sock->sport = P_CF->port; rif->sock->rx_hook = rip_rx; diff --git a/sysdep/bsd/krt-sock.c b/sysdep/bsd/krt-sock.c index e970d6b..561c018 100644 --- a/sysdep/bsd/krt-sock.c +++ b/sysdep/bsd/krt-sock.c @@ -723,7 +723,7 @@ krt_sys_start(struct krt_proto *x, int first UNUSED) if( (rt_sock = socket(PF_ROUTE, SOCK_RAW, AF_UNSPEC)) < 0) die("Cannot open kernel socket for routes"); - sk_rt = sk_new(krt_pool); + sk_rt = sk_new_sock(krt_pool); sk_rt->type = SK_MAGIC; sk_rt->rx_hook = krt_sock_hook; sk_rt->fd = rt_sock; diff --git a/sysdep/linux/netlink.c b/sysdep/linux/netlink.c index eaaf048..2198a9b 100644 --- a/sysdep/linux/netlink.c +++ b/sysdep/linux/netlink.c @@ -1066,7 +1066,7 @@ nl_open_async(void) return; } - sk = nl_async_sk = sk_new(krt_pool); + sk = nl_async_sk = sk_new_sock(krt_pool); sk->type = SK_MAGIC; sk->rx_hook = nl_async_hook; sk->fd = fd; diff --git a/sysdep/unix/io.c b/sysdep/unix/io.c index 475d660..5474b1a 100644 --- a/sysdep/unix/io.c +++ b/sysdep/unix/io.c @@ -582,7 +582,7 @@ static struct resclass sk_class = { }; /** - * sk_new - create a socket + * sk_new_sock - create a socket * @p: pool * * This function creates a new socket resource. If you want to use it, @@ -590,7 +590,7 @@ static struct resclass sk_class = { * call sk_open() to do the actual opening of the socket. */ sock * -sk_new(pool *p) +sk_new_sock(pool *p) { sock *s = ralloc(p, &sk_class); s->pool = p; @@ -1058,7 +1058,7 @@ sk_passive_connected(sock *s, struct sockaddr *sa, int al, int type) int fd = accept(s->fd, sa, &al); if (fd >= 0) { - sock *t = sk_new(s->pool); + sock *t = sk_new_sock(s->pool); char *err; t->type = type; t->fd = fd; @@ -1097,7 +1097,7 @@ sk_passive_connected(sock *s, struct sockaddr *sa, int al, int type) * sk_open - open a socket * @s: socket * - * This function takes a socket resource created by sk_new() and + * This function takes a socket resource created by sk_new_sock() and * initialized by the user and binds a corresponding network connection * to it. * diff --git a/sysdep/unix/main.c b/sysdep/unix/main.c index e0563aa..ed1fb6b 100644 --- a/sysdep/unix/main.c +++ b/sysdep/unix/main.c @@ -413,7 +413,7 @@ cli_init_unix(uid_t use_uid, gid_t use_gid) sock *s; cli_init(); - s = cli_sk = sk_new(cli_pool); + s = cli_sk = sk_new_sock(cli_pool); s->type = SK_UNIX_PASSIVE; s->rx_hook = cli_connect; s->rbsize = 1024; -- 1.7.7.6 --=-=-=--

1 0

[PATCH] include directive glob support
by Kelly Cochran 08 Jun '12

08 Jun '12

A quickie patch I whipped up a while ago to extend the include directive to add in glob support. include "/etc/bird.d/peers/*" just seemed like a useful thing. I'd be surprised if it didn't do something not quite right, but it's worked for what I've needed it to. -- -H U R R I C A N E - E L E C T R I C- Kelly Cochran Sr. Network Engineer 510-580-4100 http://www.he.net/ AS6939

2 1