June 2020 - Bird-users - Mailing Lists

Master ready for 2.0.8?
by Kenth Eriksson 28 Sep '20

28 Sep '20

Hi! Is master ready for a 2.0.8 release? The 2.0.7 release has several bugs that I think are worth a release to a broader audience (seems like they are all in master) * OSPF authentication causes SEGV https://bird.network.cz/pipermail/bird-users/2020-January/014174.html * No LPM on static routes https://bird.network.cz/pipermail/bird-users/2020-May/014563.html * Primary route flag corrupt https://bird.network.cz/pipermail/bird-users/2019-October/013931.html * Incorrect CLI output (fix continuation lines) https://bird.network.cz/pipermail/bird-users/2019-December/014019.html When looking thru master I found the following bug fixes are missing; * Multiple routes to the same network with different preference Looks like this is maintained in branch mq-ordered. Will mq-ordered branch merge into master? https://bird.network.cz/pipermail/bird-users/2019-October/013897.html * Socket close https://bird.network.cz/pipermail/bird-users/2019-June/013465.html * CLI socket buffer full https://bird.network.cz/pipermail/bird-users/2019-June/013464.html Thanks, Kenth

6 10

Improper exports with BGP6 over IPv4
by Ave 31 Jul '20

31 Jul '20

Hello, I believe I stumbled upon a bug in BIRD 2.x while doing something that can only be summed up as "testing things on someone else's network with limited access provided to me". When peering over IPv4 yet exporting a full table, I am noticing that all exports, withdraw or update, are exported as withdrawals. Export side: Routes: 0 imported, 85839 exported, 0 preferred Route change stats: received rejected filtered ignored accepted Import updates: 0 0 0 0 0 Import withdraws: 0 0 --- 0 0 Export updates: 85840 0 0 --- 85840 Export withdraws: 0 --- --- --- 0 Import side: Routes: 0 imported, 0 exported, 0 preferred Route change stats: received rejected filtered ignored accepted Import updates: 0 0 0 0 0 Import withdraws: 85823 0 --- 85823 0 Export updates: 0 0 0 --- 0 Export withdraws: 0 --- --- --- 0 I've also tested with an instance where there were both updates and withdrawals exported, and as a result a sum of both were displayed as imported withdrawals. The configurations I've tested with are quite simple: Full import config (seems to import "correctly", but as exports seem to be faulty it all gets recorded as withdrawals): > log syslog all; > router id 192.168.4.22; > protocol device { > scan time 5; > } > > protocol bgp bird1 { > local 192.168.4.22 as 65529; > neighbor 192.168.4.1 as 65530; > multihop; > > ipv6 { > import all; > export none; > }; > } Trimmed export config that doesn't export correctly: > protocol bgp bird2 { > local 192.168.4.1 as 65530; > neighbor 192.168.4.22 as 65529; > multihop; > > ipv6 { > import none; > export all; > }; > } Things I've so far: - Doing BGP4 over IPv4 (renaming "ipv6" blocks to "ipv4" in configs). This had no issues, exports and imports were correct. - Doing BGP4 over IPv6 (doing same change as last point, plus changing IPv4 addresses to IPv6 ones in configs). This had no issues, exports and imports were correct. - Switching from IPv4 to IPv6 (changing IPv4 addresses to IPv6 ones in configs). This resulted in proper imports, hinting that problem only happens on BGP6 over IPv4. - Exporting to a non-bird2 BGP daemon. Specifically, I exported to a mikrotik router running ROS v6.46.6. It also classified these as withdrawals ("Withdrawn Received 85 979", with empty "Prefix Count" and "Updates Received"). Thanks and I hope I'm not just doing something wrong and wasting your time as a result, Ave

2 1

eclist matching question
by Alexander Shikov 14 Jul '20

14 Jul '20

Hello! I have some strange behavior of list matching operator applied to eclist in v.1.6.3. Example filter function: function test (int peer_as) { if net = 109.68.40.0/21 && peer_as = 28773 then { print "DEBUG2: ", bgp_ext_community; } if (rt,65001,peer_as) ~ bgp_ext_community then { print "DEBUG: Adding one prepend: ", bgp_path.first; bgp_path.prepend(bgp_path.first); } } In log file I see bgp_ext_community attribute which includes (rt,65001,28773) community: Jun 15 15:37:52 <daemon.info> rs1 bird: DEBUG2: (eclist (rt, 65001, 13188) (rt, 65001, 28773) (ro, 31210, 25372)) But the later matching never succeeds, and I don't understand why. Please advice. Thanks! -- Alexander Shikov Technical Staff, Digital Telecom IX Tel.: +380 44 201 14 07 Mob.: +380 50 410 30 57 http://dtel-ix.net/

2 5

Re: Pointers please - OSPFv3 vs Cisco IOS XR
by Elmar K. Bins 07 Jul '20

07 Jul '20

Hi Ondrej, sorry, I missed your response (T., thank you for forwarding): >> I am very certain somebody has experienced similar issues and solved them; I >> did not find any pointers - please throw them at me. If this is new or unsolved, >> please help me forward, I can get more output, naturally. > > Hi > > It might be useful if you enable 'debug all' for OSPFv3 in BIRD. Alright...sorry, I thought I had "all" on, but apparently didn't. Here's the output from that, shortened and edited to protect the innocent. (If you require the full output, I'll email it to you) Thank you for taking a look at this, Elmar. ======================================================================================= 2020-06-03 12:30:13.801 <TRACE> ospf2: HELLO packet received from nbr <router-1> on igb0 2020-06-03 12:30:16.899 <TRACE> ospf2: DBDES packet received from nbr <router-1> on igb0 2020-06-03 12:30:16.899 <TRACE> ospf2: length 28 2020-06-03 12:30:16.899 <TRACE> ospf2: router <router-1> 2020-06-03 12:30:16.899 <TRACE> ospf2: mtu 1500 2020-06-03 12:30:16.899 <TRACE> ospf2: imms I M MS 2020-06-03 12:30:16.899 <TRACE> ospf2: ddseq 147850133 2020-06-03 12:30:16.899 <TRACE> ospf2: Neighbor <router-1> on igb0 changed state from ExStart to Exchange 2020-06-03 12:30:16.899 <TRACE> ospf2: DBDES packet sent to nbr <router-1> on igb0 2020-06-03 12:30:16.899 <TRACE> ospf2: length 108 2020-06-03 12:30:16.899 <TRACE> ospf2: router <bird-ip> 2020-06-03 12:30:16.899 <TRACE> ospf2: mtu 1500 2020-06-03 12:30:16.899 <TRACE> ospf2: imms 2020-06-03 12:30:16.899 <TRACE> ospf2: ddseq 147850133 (a handful of LSAs) 2020-06-03 12:30:16.901 <TRACE> ospf2: DBDES packet received from nbr <router-1> on igb0 2020-06-03 12:30:16.901 <TRACE> ospf2: length 1228 2020-06-03 12:30:16.901 <TRACE> ospf2: router <router-1> 2020-06-03 12:30:16.901 <TRACE> ospf2: mtu 1500 2020-06-03 12:30:16.901 <TRACE> ospf2: imms M MS 2020-06-03 12:30:16.901 <TRACE> ospf2: ddseq 147850134 (more LSAs) 2020-06-03 12:30:16.901 <TRACE> ospf2: DBDES packet sent to nbr <router-1> on igb0 2020-06-03 12:30:16.901 <TRACE> ospf2: length 28 2020-06-03 12:30:16.901 <TRACE> ospf2: router <bird-ip> 2020-06-03 12:30:16.901 <TRACE> ospf2: mtu 1500 2020-06-03 12:30:16.901 <TRACE> ospf2: imms 2020-06-03 12:30:16.901 <TRACE> ospf2: ddseq 147850134 2020-06-03 12:30:16.903 <TRACE> ospf2: LSREQ packet sent to nbr <router-1> on igb0 2020-06-03 12:30:16.903 <TRACE> ospf2: length 736 2020-06-03 12:30:16.903 <TRACE> ospf2: router <bird-ip> (some LSRs) 2020-06-03 12:30:16.906 <TRACE> ospf2: LSREQ packet received from nbr <router-1> on igb0 2020-06-03 12:30:16.906 <TRACE> ospf2: length 52 2020-06-03 12:30:16.906 <TRACE> ospf2: router <router-1> (more LSRs) 2020-06-03 12:30:16.906 <TRACE> ospf2: LSUPD packet sent to nbr <router-1> on igb0 2020-06-03 12:30:16.906 <TRACE> ospf2: length 144 2020-06-03 12:30:16.906 <TRACE> ospf2: router <bird-ip> (some LSAs) 2020-06-03 12:30:16.906 <TRACE> ospf2: DBDES packet received from nbr <router-1> on igb0 2020-06-03 12:30:16.906 <TRACE> ospf2: length 28 2020-06-03 12:30:16.906 <TRACE> ospf2: router <router-1> 2020-06-03 12:30:16.906 <TRACE> ospf2: mtu 1500 2020-06-03 12:30:16.906 <TRACE> ospf2: imms I M MS 2020-06-03 12:30:16.906 <TRACE> ospf2: ddseq 147850135 2020-06-03 12:30:16.906 <RMT> ospf2: Bad DBDES packet from nbr <router-1> on igb0 - I-bit mismatch (7) =======================================================================================

3 5

help bgp peer issue.
by seunghwan hwang 29 Jun '20

29 Jun '20

Hello. I'm using BIRD 1.6.3. Configurations(Peer is Juniper of Mellanox Switch): protocol 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 merge paths on; # Enable ECMP Multi path. } protocol direct { interface "lo"; } protocol device { scan time 10; # Scan interfaces every 10 seconds } ... protocol bgp peer1 { import all; export all; ... hold time 3; ... } protocol bgp peer2 { import all; export all; ... hold time 3; ... } In this situation, sometimes the BGP "Hold Time Expire" log remains and peers destroyed. But when I capture tcpdump, I can check the kernel sent an TCP ACK message for BGP keepalive packet. I think BGP Keepalive packets can't be transmitted to the BIRD(from the kernel). Is there an issue related to this in this version? And I would like to know if there are issues related to "scan time" and netlink. Please check it. Regards.

2 2

[PATCH 1/2] Filter: Fix comparison of BGP path mask
by Kazuki Yamaguchi 28 Jun '20

28 Jun '20

Add a missing return statement. Path masks with the same length were all considered the same. Comparing two with different length would cause out-of-bounds memory access. This bug was introduced by version 2.0.4, commit 4c553c5a5b40 ("Filter refactoring: dropped the recursion from the interpreter", 2018-12-27). --- filter/data.c | 1 + 1 file changed, 1 insertion(+) diff --git a/filter/data.c b/filter/data.c index 402202554c2f..9547a4c84833 100644 --- a/filter/data.c +++ b/filter/data.c @@ -230,6 +230,7 @@ static int pm_same(const struct f_path_mask *m1, const struct f_path_mask *m2) { if (m1->len != m2->len) + return 0; for (uint i=0; i<m1->len; i++) if (!pmi_same(&(m1->item[i]), &(m2->item[i]))) -- 2.27.0

2 2

Another newbie question: Creating BGP prefixes in a filter
by Elmar K. Bins 22 Jun '20

22 Jun '20

Hi guys, (status update: no solution for the OSPFv3 <--> IOS XR yet, no idea where to look, both sides pretend to be happy but are not) So.... I need to create a prefix to be BGP-advertised to peers and transits; this would look like: if (net = xxx/32) { remove it and replace by yyy/zz; } (more or less standard prefix aggregation on the border) I could not find a hint of documentation in regard of doing a thing like this, where should I look? (We are currently forced to configure the wrong netmasks on loopback addresses, so BIRD has something to export) Yours, Elmar.

2 2

OSPF RFC 6860 Support
by Jake Mercer 21 Jun '20

21 Jun '20

Hi all, I was thinking of adding RFC 6860 support to the OSPF implementation (hiding transit-only networks); is there any appetite for this? Or has someone already looked at it? I have a hacky modification to make it work for PtP networks for my specific purpose, but I’m happy to implement it properly in all cases and submit a patch. Is this mailing list the correct place to submit code? Best regards, Jake Jake Mercer Senior Embedded Software Engineer [cid:logo_796e1f33-efd0-4657-bc9a-9b3e989e6bb3.png]<https://www.civica.com> [cid:instagram_99d052c0-ba99-4d14-9d03-2f8ac9ed66bd.png]<https://www.instagram.com/lifeatcivica> [cid:twitter_f5b6718e-d9bd-4350-a74c-f8d9b1774b3e.png] <https://twitter.com/CivicaUK> [cid:youtube_e9775c8c-7e04-4ef9-b8ca-4b19b46c1325.png] <https://www.youtube.com/user/CivicaUK> [cid:linkedin_92fd785b-4baa-4769-aba2-200d7d2cd7a8.png] <https://www.linkedin.com/company/civica> ------------------------------------------------------------------------------------------- This email is sent for and on behalf of Civica UK Ltd, Civica NI Ltd, Civica Group Ltd or Civica Election Services Ltd. Civica Group Ltd (no: 04968437), Civica UK Ltd (no: 01628868) and Civica Election Services Ltd (no: 02263092) are registered in England and have their registered office at Southbank Central, 30 Stamford Street, London, SE1 9LQ. Civica NI Ltd (no: NI043987) is registered in Northern Ireland and has its registered office at 10 Weavers Court, Belfast, BT12 5GH. See https://www.civica.com/en-gb/policies-and-statements/privacy-notice for our privacy notice. Confidentiality: This email and its attachments are intended for the above named only and may be confidential. If they have come to you in error you must take no action based on them, nor must you copy or show them to anyone. If you have received this in error, please advise the sender by replying to this email immediately highlighting the error and deleting it from your system. -------------------------------------------------------------------------------------------

2 1

[radv] Backport solicited ra unicast to 1.6.x?
by Boris Lytochkin 17 Jun '20

17 Jun '20

Hi. I am curious if commit 70a4320bdd44122d7a93bc71c77a9d684b3c9adc could be cherry-pick'ed into 1.6.x branch? I've tested it and it's pretty straightforward: the only thing that needs to be changed is "current_time()" with "now". Aside from that, it just works out-of-box. -- Boris Lytochkin Yandex NOC +7 (495) 739 70 00 ext 7671

1 0

Static routes BGP
by Fabiano D'Agostino 15 Jun '20

15 Jun '20

Hi everyone, I have this configuration: protocol static { ipv4 { preference 110;}; route 10.0.0.0/8 via 192.168.1.1; } protocol bgp { local as 1; neighbor x.x.x.x as x ipv4 { import all; export filter {if (net ~ [10.0.0.0/8] ) then accept; reject;}; #also tried with export all; }; } On route-server I have: ipv4 table test_4; protocol pipe Ptest_4 { table master4; peer table test_4; import all; export all; } protocol bgp Btest_4 from rsclient_4 { neighbor x.x.x.x as x multihop 15; ipv4 { table test_4; import all; export all; }; } I don’t understand why 10.0.0.0/8 is not in master4 and test_4 tables of the route-server. How could I announce a test prefix from a client to the route-server? Thanks in advance, Fabiano D'Agostino

1 0