Commit graph

204 commits

Author SHA1 Message Date
Ondrej Zajicek
736e143fa5 Merge branch 'master' into add-path
Conflicts:

	filter/filter.c
	nest/proto.c
	nest/rt-table.c
	proto/bgp/bgp.h
	proto/bgp/config.Y
2013-11-23 11:50:34 +01:00
Ondrej Zajicek
77e43c8b72 Minor fixes. 2013-11-22 22:49:04 +01:00
Ondrej Zajicek
1ec522538f BFD protocol, ready for release.
Supports OSPF and BGP and also statically configured sessions.
2013-11-19 22:33:48 +01:00
Ondrej Zajicek
b21955e058 Fixes a bug related to mixed up neighbor events in BGP.
Neighbor events related to received route next hops got mixed up with
sticky neighbor node for an IP of the BGP peer. If a neighbor for a next
hop disappears, BGP session is shut down.
2013-08-13 20:42:43 +02:00
Ondrej Zajicek
48b15ef10f Fixes stuck connection during BGP session shutdown.
If TX buffers were full during BGP session shutdown
then a protocol waited indefinitely to be able to
send notification packet to close the session.
2013-07-13 01:39:41 +02:00
Ondrej Zajicek
9c99d753fd Fixes a problem with BGP neighbors, link-local addresses and locking.
Thanks to Fritz Grimpen for the bugfix.
2013-05-09 11:11:06 +02:00
Ondrej Zajicek
b662290f40 Separate import and receive limits.
They have different behavior w.r.t. filtered routes that are kept.
2013-01-10 13:07:33 +01:00
Ondrej Zajicek
79b4e12e60 Implements interface masks for choosing router id.
Router ID could be automatically determined based of subset of
ifaces/addresses specified by 'router id from' option. The patch also
does some minor changes related to router ID reconfiguration.

Thanks to Alexander V. Chernikov for most of the work.
2012-12-27 12:56:23 +01:00
Ondrej Zajicek
1555095795 Changes 'rejected' to 'filtered' in one of the last patches. 2012-11-15 01:29:01 +01:00
Ondrej Zajicek
cf98be7b67 Allows rejected routes to be kept and examined.
When 'import keep rejected' protocol option is activated, routes
rejected by the import filter are kept in the routing table, but they
are hidden and not propagated to other protocols. It is possible to
examine them using 'show route rejected'.
2012-11-10 14:26:13 +01:00
Ondrej Zajicek
094d2bdb79 Implements ADD-PATH extension for BGP.
Allows to send and receive multiple routes for one network by one BGP
session. Also contains necessary core changes to support this (routing
tables accepting several routes for one network from one protocol).
It needs some more cleanup before merging to the master branch.
2012-08-14 16:46:43 +02:00
Ondrej Zajicek
5400c0e7f9 Fixes BGP subcode during global shutdown. 2012-08-06 02:42:24 +02:00
Ondrej Zajicek
abced4a914 Merge branch 'rt-accepted'
Conflicts:

	nest/config.Y
	nest/rt-table.c
	proto/bgp/bgp.c
2012-07-16 14:44:45 +02:00
Ondrej Zajicek
26822d8fe1 Finalize RA_ACCEPTED handling. 2012-07-16 01:33:02 +02:00
Ondrej Zajicek
d9b77cc281 Implements generalized export limits.
And also fixes some minor bugs in limits.
2012-04-24 23:39:57 +02:00
Ondrej Zajicek
7d0a31deed Fixes in generalized import limits. 2012-04-21 21:05:36 +02:00
Ondrej Zajicek
ebecb6f6a1 Implements generalized import hooks.
Thanks to Alexander V. Chernikov for the original patch.
2012-04-15 15:28:29 +02:00
Ondrej Zajicek
00a09f3c36 Implement RA_ACCEPTED mode of route propagation. 2012-04-15 15:07:58 +02:00
Ondrej Zajicek
c0adf7e9fc Better support for multitable protocols.
The nest-protocol interaction is changed to better handle multitable
protocols. Multitable protocols now declare that by 'multitable' field,
which tells nest that a protocol handles things related to proto-rtable
interaction (table locking, announce hook adding, reconfiguration of
filters) itself.

Filters and stats are moved to announce hooks, a protocol could have
different filters and stats to different tables.

The patch is based on one from Alexander V. Chernikov, thanks.
2012-03-15 12:13:04 +01:00
Ondrej Zajicek
39c028e9e9 Assign default protocol preference via proto_config_new().
The patch from Alexander V. Chernikov.
2012-01-24 11:31:00 +01:00
Ondrej Zajicek
53ffbff39f Implements support for link-local addresses in BGP.
Thanks Matthias Schiffer for the original patch.
2012-01-08 15:31:34 +01:00
Ondrej Zajicek
69a8259c5e Allows sticky link-local neighbors.
Allows using NEF_STICKY neighbors with link-local addresses. This is
used for static route nexthops, they can be specified like fe80::1%eth0
.
2012-01-01 12:14:42 +01:00
Ondrej Zajicek
be4cd99a36 Implements deterministic MED handling.
Thanks to Alexander V. Chernikov for many suggestions.
2011-12-22 13:20:29 +01:00
Ondrej Zajicek
a7f23f581f Implements protocol templates.
Based on the patch from Alexander V. Chernikov.
Extended to support almost all protocols.
Uses 'protocol bgp NAME from TEMPLATE { ... }' syntax.
2011-11-07 00:31:23 +01:00
Ondrej Zajicek
51947659ab Minor changes in BGP protocol info.
Shows neighbor IP and ASN even if protocol is down.
2011-09-04 00:56:02 +02:00
Ondrej Zajicek
b1b1943360 The generalized TTL security mechanism (RFC 5082) support.
Thanks to Alexander V. Chernikov for the patch.
2011-08-16 23:13:05 +02:00
Ondrej Zajicek
35f8c731ea Fixes output of BGP timers in 'show protocols all'. 2011-06-25 14:51:16 +02:00
Ondrej Zajicek
4ef0950603 Fixes a problem with BGP protocols and implicit router IDs. 2011-04-01 13:55:52 +02:00
Ondrej Zajicek
fcf5a4f4b3 Change default for BGP IPv6 socket to V6ONLY.
Also add a new option 'dual' for the old behavior.
2010-11-03 10:02:24 +01:00
Ondrej Zajicek
ac4a1eedfc Change default for handling missing link-local addresses on route servers. 2010-08-13 14:30:36 +02:00
Ondrej Zajicek
087cecd0e2 Adds some options and documentation related to recursive next hops. 2010-07-13 12:48:23 +02:00
Ondrej Zajicek
9be9a26413 Implements proper multihop BGP.
Also does some incompatible changes to config file syntax,
like removing 'via IP' from multihop option.
2010-07-12 17:39:39 +02:00
Ondrej Zajicek
cfe34a316e Implements hostcache and recursive next hops.
Hostcache is a structure for monitoring changes in a routing table that
is used for routes with dynamic/recursive next hops. This is needed for
proper iBGP next hop handling.
2010-07-05 17:50:19 +02:00
Ondrej Zajicek
a34b09349e Disable BGP protocol when cannot open listening socket. 2010-05-31 11:35:29 +02:00
Ondrej Zajicek
475977242a Handle EPIPE as a common connection close event. 2010-05-14 16:54:39 +02:00
Ondrej Zajicek
b8113a5e92 Implements BGP 'show protocols' info details. 2010-04-07 00:19:23 +02:00
Ondrej Zajicek
e81b440f68 Fix configure to enable warnings and fix most of them. 2010-02-21 14:34:53 +01:00
Ondrej Zajicek
a2ea1bac60 Moves errno.h include. 2010-02-11 21:19:20 +01:00
Ondrej Zajicek
2af25a971a Fixes a crash caused by missing error hook on BGP listening socket.
Error happened when too many BGP connections arrived in one moment
(ECONNABORTED).
2010-02-11 11:12:58 +01:00
Ondrej Zajicek
cf31112f0d Implements MRTdump feature. 2010-01-03 12:17:52 +01:00
Ondrej Zajicek
43c1ceccb9 Remove bgp_as4_support variable. 2009-12-14 23:31:25 +01:00
Ondrej Zajicek
bf47fe4b2e Implements BGP route refresh. 2009-11-26 20:47:59 +01:00
Ondrej Zajicek
3f9b7bfe9f Implements option that controls IPv6 BGP next hops when lladdr is missing. 2009-11-18 20:32:36 +01:00
Ondrej Zajicek
4cdd078453 Implements protocol-specific router id and updates documentation. 2009-10-12 23:31:42 +02:00
Ondrej Zajicek
ea89da381f Workaround for stupid callback scheduler.
There is no reak callback scheduler and previous behavior causes
bad things during hard congestion (like BGP hold timeouts).

Smart callback scheduler is still missing, but main loop was
changed such that it first processes all tx callbacks (which
are fast enough) (but max 4* per socket) + rx callbacks for CLI,
and in the second phase it processes one rx callback per
socket up to four sockets (as rx callback can be slow when
there are too many protocols, because route redistribution
is done synchronously inside rx callback). If there is event
callback ready, second phase is skipped in 90% of iterations
(to speed up CLI during congestion).
2009-10-11 18:56:16 +02:00
Ondrej Zajicek
be6e39ebbf Passive option. 2009-09-24 22:12:11 +02:00
Ondrej Zajicek
70670bf317 Fixed bug related to reconfiguration of BGP with MD5 passwords. 2009-07-09 15:42:39 +02:00
Ondrej Zajicek
d72cdff411 Replace 'bind' option with 'listen' option.
To be consistent with other daemons.
2009-06-23 11:08:30 +02:00
Ondrej Zajicek
789772ed45 Implements option that changes BGP listening socket parametres. 2009-06-18 19:20:07 +02:00
Ondrej Zajicek
b99d378698 Minor BGP changes related to error wait time. 2009-06-07 00:38:38 +02:00
Ondrej Zajicek
7238262671 Fixes bug related to startup delay change. 2009-06-06 18:56:33 +02:00
Ondrej Zajicek
dd91e46765 Differentiate between error delay and connect/reconnect delay.
The difference is here to reject incoming connections in the
first case.
2009-06-06 18:16:22 +02:00
Ondrej Zajicek
72b28a041d Implements import route limits. 2009-06-04 13:31:09 +02:00
Ondrej Zajicek
23ac9e9a9e Changes pipes to transfer all routes between routing table, not just optimal routes. 2009-05-31 15:24:27 +02:00
Ondrej Zajicek
4827b69ff4 Fixes BGP IPv6 link local next hop handling.
When sending 'third party' BGP update, Bird used bogus link local
addresses instead of addresses it received before.
2009-04-29 18:58:24 +02:00
Ondrej Zajicek
ad440a570b Fixes handling of 'next hop self' and 'source address' configuration
options.
2009-04-28 18:11:56 +02:00
Ondrej Zajicek
48d79d521c Better handling of AS4 optional attribute errors
AS4 optional attribute errors were handled by session
drop (according to BGP RFC). This patch implements
error handling according to new BGP AS4 draft (*)
 - ignoring invalid AS4 optional attributes.

(*) http://www.ietf.org/internet-drafts/draft-chen-rfc4893bis-02.txt
2009-03-18 20:30:21 +01:00
Ondrej Zajicek
e8ba557c7f Update capability handshake options
Add 'capabilities' option, change default behavior to advertise ipv4,
add some checks and ignore incoming capabilities when capabilities
are disabled.
2009-03-13 12:49:44 +01:00
Ondrej Zajicek
165a622727 Adds support for fallback to capabilityless BGP connect
When capability related error is received, next connect will be
without capabilities. Also cease error subcodes descriptions
(according to [RFC4486]) are added.
2008-12-24 17:24:41 +01:00
Ondrej Zajicek
591211557f Fixes bug related to reconfiguration of BGP.
BGP keeps its copy of configuration ptr and didn't update it during
reconfiguration. But old configuration is freed during reconfiguration.
That leads to unnecessary reset of BGP connection during reconfiguration
(old conf is corrupted and therefore different) and possibly other strange
behavior.
2008-12-24 12:18:10 +01:00
Ondrej Zajicek
11b32d9117 Major changes to BGP
Fixes two race conditions causing crash of Bird, several unhandled
cases during BGP initialization, and some other bugs. Also changes
handling of startup delay to be more useful and implement
reporting of last error in 'show protocols' command.
2008-12-19 01:34:39 +01:00
Ondrej Zajicek
a39b165e45 Multihop BGP was completely broken, because listening socket has always
ttl 1.
2008-11-01 16:58:40 +01:00
Ondrej Zajicek
a92fe60717 Implementation of route server. 2008-11-01 12:55:43 +01:00
Ondrej Zajicek
ba5ed6f3e4 Implementation of an option for disabling AS4 support per BGP instance. 2008-10-26 22:48:02 +01:00
Ondrej Zajicek
4847a894bf Implementation of route reflection for BGP 2008-10-26 22:45:09 +01:00
Ondrej Zajicek
d51aa28190 Implementation of MD5 authentication of BGP sessions. 2008-10-26 22:42:39 +01:00
Ondrej Zajicek
11cb620266 Implementation of 4B ASN support for BGP 2008-10-26 22:36:08 +01:00
Martin Mares
9831e5916f Staticized lots of local functions. 2004-06-05 09:58:23 +00:00
Martin Mares
e21423bab8 ... in BGP. 2004-06-05 09:27:17 +00:00
Martin Mares
38a608c55a Rewritten the I/O loop. All socket operations are now safe, meaning that
you can delete the socket from anywhere in the hooks and nothing should break.
Also, the receive/transmit buffers are now regular xmalloc()'ed buffers,
not separate resources which would need shuffling around between pools.

sk_close() is gone, use rfree() instead.
2004-05-31 21:48:19 +00:00
Martin Mares
6a57bb3110 Killed a couple of unused variables.
We really should compile with warnings enabled.
2004-05-31 20:57:38 +00:00
Martin Mares
11d4474c17 Better selection of link-local NLRI addresses, at least for our own
address. Need to do it better for the other neighbors -- the current
solution works only if they use the standard 64+64 global addresses
and the interface identifier in lower 64 bits is the same as for the
link-scope addresses.
2003-02-22 22:47:45 +00:00
Martin Mares
7d875e094b Added missing includes. 2002-11-13 08:47:06 +00:00
Martin Mares
bcbdcbb6ae Don't accept incoming connections when the neighbor is not up. 2000-06-20 07:49:08 +00:00
Martin Mares
58f7d004fd Fixes to the progdoc. 2000-06-07 13:25:53 +00:00
Martin Mares
2e9b24211a Spelling fixes to progdoc. 2000-06-07 12:29:08 +00:00
Martin Mares
54e55169da BGP documented. 2000-06-04 17:06:18 +00:00
Martin Mares
5e88d73025 BGP now reports originating AS and origin type in get_route_info(). 2000-05-19 11:01:41 +00:00
Martin Mares
f4ab231746 bgp_get_status: If protocol is down, don't print BGP state. 2000-05-15 12:19:28 +00:00
Martin Mares
93d6bf38a6 The bgp_list is gone. Incomming connections are now handled in a much
more straightforward manner by scanning the active configuration for
matching protocols.
2000-05-08 13:26:30 +00:00
Martin Mares
79681f4a19 Link the instance to the global BGP list as soon as possible. 2000-05-08 11:04:22 +00:00
Martin Mares
acfce55c86 Setup of incoming connection is now a separate function. 2000-05-07 10:41:45 +00:00
Martin Mares
53943a0022 Defined format specifier %M' which behaves as %m', but takes the
error code as an argument.

Use it in socket hooks where we really shouldn't rely on errno containing
the right value or even existing.
2000-05-02 16:10:04 +00:00
Martin Mares
85368cd4b7 Full protocol tracing. 2000-05-02 16:07:41 +00:00
Martin Mares
9165888ad2 Handle connect errors correctly. 2000-04-27 19:41:10 +00:00
Martin Mares
5db9bae286 IBGP fixes. 2000-04-26 13:26:31 +00:00
Martin Mares
6fd766c17e Implemented automatic restart after error with all the timers needed. 2000-04-25 23:08:31 +00:00
Martin Mares
b3155b3399 Randomize timers properly. 2000-04-25 22:01:19 +00:00
Martin Mares
42532f0846 Support dynamic reconfiguration. 2000-04-25 21:58:17 +00:00
Martin Mares
efcece2da3 Better reporting of both local and remote errors. 2000-04-25 21:13:25 +00:00
Martin Mares
8b258e4e65 LOCAL_PREF is now always present and exported over all ibgp connections [draft]
Allow setting of address of the local end of the TCP connection.

Several bug fixes.
2000-04-21 13:01:28 +00:00
Martin Mares
e0d6a7bda4 Delay fetching of router ID. 2000-04-19 13:28:56 +00:00
Martin Mares
035044b1d9 Select the right source address and don't check port numbers. 2000-04-19 12:51:14 +00:00
Martin Mares
48e842cc98 Use neighbor cache to track direct route to the peer or multihop destination.
Calculate next_hop properly based on the local address we get from the
neighbor entry.
2000-04-10 11:21:40 +00:00
Martin Mares
ef2c708dfa More BGP progress...
For Pavel: You can use bgp_path_prepend() for prepending AS numbers to AS paths.
2000-04-09 22:05:02 +00:00
Martin Mares
10be74da20 Formatting of dynamic attributes (except for paths and communities which
will be added soon).
2000-04-01 10:21:11 +00:00
Martin Mares
f880924990 BGP now handles incoming routes (IPv4 only). 2000-04-01 09:17:33 +00:00
Martin Mares
c00d31befa Parsing of BGP attributes. 2000-03-31 23:21:37 +00:00
Martin Mares
973399ae2c Basic analysis of UPDATE packets. 2000-03-30 18:44:23 +00:00
Martin Mares
b552ecc4d7 Connection state machine works. 2000-03-30 17:39:48 +00:00
Martin Mares
3fdbafb6f4 More BGP. This time it connects, but the state machine still isn't complete. 2000-03-30 10:44:20 +00:00
Martin Mares
72a6ef11fe Construction of BGP packets. 2000-03-21 15:53:50 +00:00
Martin Mares
c01e37416d Started work on BGP. Wrote main part of the connection handling code. 2000-03-20 21:50:17 +00:00
Martin Mares
2638249d34 Bare skeleton of the BGP. 2000-03-19 22:09:07 +00:00