- Bird-users - Mailing Lists

[PATCH] bird: treat former classe addresses as valid addresses
by Dave Taht 13 Dec '18

13 Dec '18

replying via gmail as taht.net's server mandates starttls and someone doesn't have that enabled for outgoing mail. >On Tue, Dec 11, 2018 at 03:26:08PM -0800, Dave Taht wrote: >> There is no reason to disallow the former class e address >> space, except in a bogon file. >Hi >> static inline int ip4_is_unicast(ip4_addr a) >> -{ return _I(a) < 0xe0000000; } >> +{ return _I(a) < 0xe0000000 || (_I(a) >= 0xf0000000 && _I(a) != 0xffffffff); } >Well, this summarises the reason why to disallow former class E address >space. If these addresses are not defined / reserved for future use, how >can we say that they are unicast and not multicast or for some complete >different but not yet defined purpose? As best as we can tell they have been treated as unicast by linux, freebsd, and OSX since at least about 2010, perhaps longer. >IMHO first should IETF redefine it as a regular unicast range, then >routing software should change its behavior. Please correct me if >that already happened. The deployment is all unicast in the 240/4 range. Try it - in linux: ip addr add 255.255.255.254/24 dev whatever Over the last few months we've tested apple IOS, OSX, Freebsd, most modern linux linuxes use the iproute suite per above, and just work, and those few that don't, like openwrt, use an ifconfig ioctl that we also have a few line patch awaiting approval to the linux kernel. The original patch in this series goes back to 2008 (https://lkml.org/lkml/2008/1/7/370 ). Multiple switches, APs, and devices "just work". Only windows, currently cannot route to or accept a 240/4 address, and we're hoping to address that. document in progress on that... So even if the IETF wanted to change the unicast interpretation, it would be unwise. as for how these addresses could be used, well, that's a subject for a new RFC (pending) but the technical feasibility appears bright at the moment. We (john gilmore, rob landley, others) behind this are trying to gain approval for a worldwide test opening up a few of these ranges, especially the "hard" one, 255/8, for analysis. >OTOH, perhaps we could allow all possible routes regardless of address >classification (even for prefixes from multicast range) in routing table. That's what the babeld patch does, currently. Under test. Many bogon files also filter out this range. one line changed, and voila! >OSes allow that and sometimes even uses that (e.g. route for multicast range in regular unicast table is used for selection of default multcast interface on clients). Thank you for pointing this out. We have been using things like searchcode for all the public sources in the world, and thus far... all but the IN_MULTICAST macros are essentially unused in userspace. I'll go look harder for other potential multicast/unicast conflicts like that, but the linux kernel, at least, looked clean. ... only the main routing daemons had issues. http://www.taht.net/classe/ has the tiny patch sets for everything so far We're in the FRR process over here: https://github.com/FRRouting/frr/pull/3462 does anyone still use xorp? -- Dave Täht CTO, TekLibre, LLC http://www.teklibre.com Tel: 1-831-205-9740

1 0

[PATCH] bird: treat former classe addresses as valid addresses
by Dave Taht 12 Dec '18

12 Dec '18

There is no reason to disallow the former class e address space, except in a bogon file. --- lib/ip.c | 8 ++++++-- lib/ip.h | 2 +- 2 files changed, 7 insertions(+), 3 deletions(-) diff --git a/lib/ip.c b/lib/ip.c index 2050d19e..022c7d74 100644 --- a/lib/ip.c +++ b/lib/ip.c @@ -100,8 +100,12 @@ ip4_classify(ip4_addr ad) if (b >= 0xe0 && b <= 0xef) return IADDR_MULTICAST | SCOPE_UNIVERSE; - if (a == 0xffffffff) - return IADDR_BROADCAST | SCOPE_LINK; + if (b >= 0xf0) + if(a == 0xffffffff) { + return IADDR_BROADCAST | SCOPE_LINK; + } else { + return IADDR_HOST | SCOPE_UNIVERSE; + } return IADDR_INVALID; } diff --git a/lib/ip.h b/lib/ip.h index 4fbe4407..2d1b4cff 100644 --- a/lib/ip.h +++ b/lib/ip.h @@ -312,7 +312,7 @@ static inline int ip6_is_v4mapped(ip6_addr a) #endif static inline int ip4_is_unicast(ip4_addr a) -{ return _I(a) < 0xe0000000; } +{ return _I(a) < 0xe0000000 || (_I(a) >= 0xf0000000 && _I(a) != 0xffffffff); } static inline int ipa_classify_net(ip_addr a) { return ipa_zero2(a) ? (IADDR_HOST | SCOPE_UNIVERSE) : ipa_classify(a); } -- 2.17.1

2 1

Problems while experimenting some new features in BGP
by Mattia Milani 12 Dec '18

12 Dec '18

Hello all, I'm experimenting some new features in BGP but I'm stuck in a problem that I can't reach the solution. I want to trigger a Route Refresh every time after an update some constraints are violated. I thought that the right place to do that is inside this function: static void bgp_rx_update(struct bgp_conn *conn, byte *pkt, uint len) { .... if (Constraints are violated){ Trigger Route Refresh } return; } This was my idea, first of all, do you think this is the right place to trigger the route reresh? After that, I was trying to trigger the route refresh calling this function: void bgp_feed_begin(struct channel *C, int initial) But I didn't get the result I want, the node in the log wrote: Sending BEGIN-OF-RR But nothing else happened, so what is the right way to trigger a Route refresh? Thanks a lot for the help, I know that is a strange request but it could help a lot my experiments, Mattia

1 0

BIRD 2.0.2 documentation build failure
by Robert Scheck 11 Dec '18

11 Dec '18

Hello, I'm trying to build the documentation of BIRD 2.0.2 on Fedora and RHEL/ CentOS 7 which fails like this: --- snipp --- + /usr/bin/make -j4 -C doc make: Entering directory `/builddir/build/BUILD/bird-2.0.2/doc' make: Leaving directory `/builddir/build/BUILD/bird-2.0.2/doc' make: *** No rule to make target `/tools/progdoc', needed by `prog.sgml'. Stop. --- snapp --- When I decided to replace "$(srcdir)" in doc/Makefile by ".." it goes further, but still fails. Replacing "$(objdir)" by "../obj" and finally replacing "$(doc-srcdir)" by "." leads to the message that bird.sgml has a loop reference (which leads to an empty bird.sgml). And when removing this target, $(o)%.sgml: $(s)%.sgml $(objdir)/.dir-stamp cp $< $@ the documentation build even succeeds. In total that's then this diff: --- snipp --- --- bird-2.0.2/doc/Makefile 2018-01-16 10:46:06.000000000 +0100 +++ bird-2.0.2/doc/Makefile.doc 2018-11-25 16:38:22.988423717 +0100 @@ -1,7 +1,7 @@ # Force rebuilds .PHONY: progspell docs progdocs userdocs -doc-srcdir := $(shell cd $(s) && pwd) +doc-srcdir = . sgml2 := $(doc-srcdir)/sgml2 docs: progdocs userdocs @@ -16,11 +16,8 @@ userdocs: $(o)bird.html $(o)bird.pdf progspell: $(o)prog.spell -$(o)prog.sgml: $(srcdir)/tools/progdoc $(objdir)/.dir-stamp - $(srcdir)/tools/progdoc $(srcdir) $@ - -$(o)%.sgml: $(s)%.sgml $(objdir)/.dir-stamp - cp $< $@ +$(o)prog.sgml: ../tools/progdoc ../obj/.dir-stamp + ../tools/progdoc .. $@ $(o)%.html: $(o)%.sgml cd $(dir $@) && $(sgml2)html $(notdir $<) --- snapp --- Do you have any pointers? Is this a bug in BIRD or is this a Fedora/RHEL build system issue? Regards, Robert

2 5

ospf with bfd ecmp conf
by Alon Dotan 11 Dec '18

11 Dec '18

Hey All, Im trying to configure the folowing topology: lns box (multiple ppp devices that connect and disconnect) -> multiple default gw (gw load balance) the issue is if one of the multiple gw crash I got routing blackhole... Im trying to configurefd to delete dead peers from the osp routing table, this is my conf: debug protocols all; log "/var/log/bird.log" all; router id 10.10.11.10; protocol device { scan time 1; # Scan interfaces every 10 seconds export all; } protocol direct { interface "ppp*"; export all; } protocol kernel { scan time 1; # Scan kernel routing table every 20 seconds export all; # Default is export none import all; kernel table 100; learn; persist; } protocol bfd { interface "eth1" { min rx interval 20 ms; min tx interval 50 ms; idle tx interval 300 ms; }; } protocol ospf MyOSPF { ecmp yes; export filter { if (net ~ 10.10.11.0/24) then reject; accept;}; # export all; import all; merge external; area 0 { interface "eth1" { hello 5; bfd yes; }; }; } protocol static { import all; } routs learned by ospf: [root@lns-testing-01 centos]# ip r show table 100 default proto bird nexthop via 10.10.11.11 dev eth1 weight 1 nexthop via 10.10.11.12 dev eth1 weight 1 [root@lns-testing-01 centos]# bfd status: [root@lns-testing-01 centos]# birdc show bfd sessions BIRD 1.4.5 ready. bfd1: IP address Interface State Since Interval Timeout 10.10.11.11 eth1 Up 08:09:55 0.050 0.250 10.10.11.12 eth1 Down 08:10:41 0.300 0.000 what Im missing? Thanks,

1 0

segfault when adding OSPF virtual link on Bird 1.6.4
by Cybertinus 10 Dec '18

10 Dec '18

Hello everybody, Today I was experimenting with a Bird setup. All the Bird machines are CentOS 7.6 VMs running in Virtualbox on my laptop. When I add an OSPF virtual link to hook up two area 0's together I get a segfault of Bird 1.6.4 itself. All the links between the VMs are defined as "Internal Network" links. Each connection between routers got it's own internal network, so there should be no overlap between those connections. Steps to reproduce: 1. Create a bird.conf with the following contents specified in the attached bird.conf 2. Start Bird 3. Add the following line to the area 1 config: virtual link 2.2.2.2; 4. Run `birdc configure` 5. Observe the following error in your syslog: Dec 8 21:01:35 r3 kernel: bird[3469]: segfault at 32 ip 0000000000429fe0 sp 00007fff599d3350 error 4 in bird[400000+73 I've ran Bird 1.6.4 in a gdb session with the debug symbols installed and I created a stack trace after the crash. This gives the output which can be found in the attached gdb.txt. What is interesting to note is that the segfault only occurs when I reconfigure Bird. I can start it just fine right after the crash and the OSPF sessions come back online and everything. I hope this gives some useful information about what is happening. If you need any more information please don't hesitate to ask. It is a test network, so I can tell you absolutely everything about it, nothing is a company secret or something like that :). Kind regards, Cybertinus

2 3

FLAPPING PEER
by Rwakagabo Janvier 10 Dec '18

10 Dec '18

Hello bird user I am using bird 1.4.0 version on route server (RS) to peer on IXP players, One of the peer is flapping with this error Can you help to understand the cause. Regards, Janvier R.

1 0

Recommended way to shutdown a working BGP session -- ordered bgp session stop -- bgp session shutdown
by Francis Brosnan Blázquez 10 Dec '18

10 Dec '18

Hello. We are using bird with several upstream providers, all of them with a share of traffic. We are in the process of shutting down one of them but we are unsure how to proceed to minimize loss of traffic. We have been reading and looking for general recommendations but it is not clear (besides using graceful shutdown which is not supported by the upstream we want to shutdown). We have been looking at mailing list but we haven't found anything treating this matter. So far, solutions we have come up are: 1) Use AS-path prepend to increase metric on the uptstream to be shutted down and once nearly no traffic comes in through that link, shutdown BGP and unplug. Something like: export filter { 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); bgp_path.prepend(65000); accept; } reject; }; 2) Another solution is to just shutdown BGP session but leave upstream connected and configured (so outdated routers we still reach us...). 3) And the obvious, just shutdown BGP session and unplug the cable. It would be great to know your opinion and what's the recommended way to proceed. What do you think? Many thanks Best Regards. -- Francis Brosnan Blázquez <francis.brosnan(a)aspl.es> ASPL -- Advanced Software Production Line, S.L.

2 2

iBGP stuck in connect state
by Brooks Swinnerton 06 Dec '18

06 Dec '18

Hello, I'm having trouble configuring two BIRD instances to talk to one another over a wireguard tunnel with iBGP. They both stay in a `Connect` state. With debug <protocol> all, I only see: ``` Dec 05 23:31:32 border bird[3460]: foo: Connecting to 169.254.2.2 from local address 169.254.3.3 Dec 05 23:33:19 border bird[3460]: foo: Connecting to 169.254.2.2 from local address 169.254.3.3 Dec 05 23:35:13 border bird[3460]: foo: Connecting to 169.254.2.2 from local address 169.254.3.3 ``` The wireguard tunnel is up, and each side can ping the other: ``` PING 169.254.2.2 (169.254.2.2) 56(84) bytes of data. 64 bytes from 169.254.2.2: icmp_seq=1 ttl=64 time=66.7 ms 64 bytes from 169.254.2.2: icmp_seq=2 ttl=64 time=66.6 ms 64 bytes from 169.254.2.2: icmp_seq=3 ttl=64 time=66.8 ms ``` ``` PING 169.254.3.3 (169.254.3.3) 56(84) bytes of data. 64 bytes from 169.254.3.3: icmp_seq=1 ttl=64 time=66.5 ms 64 bytes from 169.254.3.3: icmp_seq=2 ttl=64 time=66.2 ms 64 bytes from 169.254.3.3: icmp_seq=3 ttl=64 time=66.3 ms ``` I can even see some BGP traffic when doing a tcpdump: ``` 23:44:16.487852 ip: (tos 0xc0, ttl 64, id 40780, offset 0, flags [DF], proto TCP (6), length 60) 169.254.3.3.53585 > 169.254.2.2.179: Flags [S], cksum 0x5930 (incorrect -> 0x61e7), seq 4023371188, win 27600, options [mss 1380,sackOK,TS val 891392 ecr 0,nop,wscale 7], length 0 23:44:16.989732 ip: (tos 0xc0, ttl 64, id 29307, offset 0, flags [DF], proto TCP (6), length 60) 169.254.2.2.46081 > 169.254.3.3.179: Flags [S], cksum 0x3ae5 (correct), seq 156366469, win 27600, options [mss 1380,sackOK,TS val 889856 ecr 0,nop,wscale 7], length 0 23:44:18.535863 ip: (class 0xc0, flowlabel 0xe9696, hlim 64, next-header TCP (6) payload length: 40) fd00:169:254:3::3.47019 > fd00:169:254:2::2.179: Flags [S], cksum 0x01b4 (incorrect -> 0x5b17), seq 1562824692, win 27200, options [mss 1360,sackOK,TS val 891904 ecr 0,nop,wscale 7], length 0 23:44:33.117481 ip: (tos 0xc0, ttl 64, id 29308, offset 0, flags [DF], proto TCP (6), length 60) 169.254.2.2.46081 > 169.254.3.3.179: Flags [S], cksum 0x2b25 (correct), seq 156366469, win 27600, options [mss 1380,sackOK,TS val 893888 ecr 0,nop,wscale 7], length 0 23:44:37.610488 ip: (class 0xc0, flowlabel 0xeaf06, hlim 64, next-header TCP (6) payload length: 40) fd00:169:254:2::2.36355 > fd00:169:254:3::3.179: Flags [S], cksum 0x243b (correct), seq 476210459, win 27200, options [mss 1360,sackOK,TS val 895010 ecr 0,nop,wscale 7], length 0 23:44:38.621468 ip: (class 0xc0, flowlabel 0xd3b4f, hlim 64, next-header TCP (6) payload length: 40) fd00:169:254:2::2.36355 > fd00:169:254:3::3.179: Flags [S], cksum 0x233d (correct), seq 476210459, win 27200, options [mss 1360,sackOK,TS val 895264 ecr 0,nop,wscale 7], length 0 23:44:40.637418 ip: (class 0xc0, flowlabel 0xd41b0, hlim 64, next-header TCP (6) payload length: 40) fd00:169:254:2::2.36355 > fd00:169:254:3::3.179: Flags [S], cksum 0x2145 (correct), seq 476210459, win 27200, options [mss 1360,sackOK,TS val 895768 ecr 0,nop,wscale 7], length 0 23:44:44.893305 ip: (class 0xc0, flowlabel 0xcc08e, hlim 64, next-header TCP (6) payload length: 40) fd00:169:254:2::2.36355 > fd00:169:254:3::3.179: Flags [S], cksum 0x1d1d (correct), seq 476210459, win 27200, options [mss 1360,sackOK,TS val 896832 ecr 0,nop,wscale 7], length 0 ``` The bird.conf file is identical between hosts can be found in: https://gist.github.com/bswinnerton/9ffa236a55f120ba9491658ae74a841a. Does anything seem amiss? What's odd is that I can establish iBGP connections with other BGP routers that are not running BIRD with the exact same iBGP peer configuration. $ sudo birdc -v 0001 BIRD 2.0.2 ready.

2 2

Static routes after `configure` CLI command
by Vladimir Osmolovskiy 04 Dec '18

04 Dec '18

Hello! In CLI command line, after 'configure' command, pointer to static routes in neighbor structure not valid. This patch delete heighbors. Perhaps there is a better solution. diff --git a/nest/neighbor.c b/nest/neighbor.c index f8159d3..7fc449a 100644 --- a/nest/neighbor.c +++ b/nest/neighbor.c @@ -379,6 +379,37 @@ neigh_prune_one(neighbor *n) sl_free(neigh_slab, n); } . + +void +neigh_prune_proto(struct proto *p) +{ + neighbor *n; + node *m; + int i; + + for(i=0; i<NEIGH_HASH_SIZE; i++) + WALK_LIST_DELSAFE(n, m, neigh_hash_table[i]) + if (n->proto == p ) + { + rem_node(&n->n); + if (n->scope >= 0) + rem_node(&n->if_n); + sl_free(neigh_slab, n); + } + WALK_LIST_DELSAFE(n, m, sticky_neigh_list) + if (n->proto == p ) + { + rem_node(&n->n); + if (n->scope >= 0) + rem_node(&n->if_n); + sl_free(neigh_slab, n); + } +} + + + + + /** * neigh_prune - prune neighbor cache * diff --git a/proto/static/static.c b/proto/static/static.c index 849067b..8a7d1a0 100644 --- a/proto/static/static.c +++ b/proto/static/static.c @@ -559,6 +559,8 @@ static_reconfigure(struct proto *p, struct proto_config *new) WALK_LIST(r, o->other_routes) static_match(p, r, n); . + neigh_prune_proto(p); + /* Now add all new routes, those not changed will be ignored by static_install() */ WALK_LIST(r, n->iface_routes) {

2 4