commit 1eab26bbc67b1dd36ac44f7fdd427871b3c6b59b
Author: Paul Jakma <[email protected]>
Date:   Sat Jun 7 20:44:30 2008 +0000

   [tests] BGP MP_(UN)REACH_NLRI unit tests (actual file..)

   2008-06-07 Paul Jakma <[email protected]

       * bgp_mp_attr_test.c: MP_(UN)REACH_NLRI unit tests

commit 3d52bb8085c32b54fa21d88588310716245291d4
Author: Paul Jakma <[email protected]>
Date:   Sat Jun 7 20:42:07 2008 +0000

   [lib] trivial: add const qualifier to stream_put/write

   2008-06-07 Paul Jakma <[email protected]>

       * stream.{c,h}: (stream_{put,write}) add const qualifier to source
         argument. Change u_char to void *.

commit 032928091c558d3b0940be079049009f47652711
Author: Paul Jakma <[email protected]>
Date:   Sat Jun 7 20:37:10 2008 +0000

   [bgpd] minor changes to bgp_mp_reach_parse

   2008-06-07 Paul Jakma <[email protected]>

       * bgp_attr.{c,h}: (bgp_mp_{un,}reach_parse) export, for unit tests.
       * bgp_attr.c: (bgp_mp_reach_parse) Add logging. Tighten length test
         to bounds check against the attribute length rather than the
         stream length..

commit 343f720b345dd377bdf4f53e3b02e0c2898f3f0b
Author: Paul Jakma <[email protected]>
Date:   Sat Jun 7 20:32:38 2008 +0000

   [tests] BGP MP_(UN)REACH_NLRI unit tests

   2008-06-07 Paul Jakma <[email protected]

       * bgp_mp_attr_test.c: MP_(UN)REACH_NLRI unit tests

commit 13e896ba2a449848791dd7d020c59fe75cbd562b
Author: Denis Ovsienko <[email protected]>
Date:   Sat Jun 7 17:54:38 2008 +0000

   + note the meta-queue in NEWS

commit e96f92034dad1a70c92ad4ad863a5259c122cecb
Author: Denis Ovsienko <[email protected]>
Date:   Mon Jun 2 12:03:22 2008 +0000

   + initial edition of meta-queue for RIB updates processing (bug #431)

commit a15cfd16fcdec39588ce2f780671ba7c6de0b919
Author: Paul Jakma <[email protected]>
Date:   Sun Jun 1 14:29:03 2008 +0000

   [bgpd] bug #419: partial aspath-limit incorrectly causes session reset

   2008-06-01 [email protected]

       * bgp_attr.c: (bgp_attr_aspathlimit) fix silly bug in flags check
         that was causing BGP to drop sessions if it received a
         aspath-limit with partial set. Fixes bug #419.

commit 0e7c124f870d4c95ff3edbde1a766717c696da7c
Author: Paul Jakma <[email protected]>
Date:   Sun Jun 1 14:26:48 2008 +0000

   [tools] multiple-bgpd.sh: make paths more configurable

commit fa93b16208c2e7ffb09bd5bf72fb1a70a1ad8f73
Author: Paul Jakma <[email protected]>
Date:   Thu May 29 19:03:08 2008 +0000

   [ripd] Fix mistaken empty string test

   2008-05-29 Stephen Hemminger <[email protected]>

       * ripd.c: (rip_auth_md5) fix bogus empty string test

commit 0d6b2ee249e9efba30da3236c413bf03c7b27ed9
Author: Paul Jakma <[email protected]>
Date:   Thu May 29 18:29:16 2008 +0000

   [daemons] Sanity check port number arguments before use

   2008-05-29 Martin Nagy <[email protected]>

       * */*main.c: Sanity check port numbers before using.

commit 768a27ea7ba25257c79689af83ab37945dc1cc3c
Author: Paul Jakma <[email protected]>
Date:   Thu May 29 18:23:08 2008 +0000

   [zebra/linux] Use BPF to filter out responses, to try avoid netlink overruns

   2008-05-29 Stephen Hemminger <[email protected]>

       * 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.

commit 311064ddc5c52f6859cf1e6e8636c19a2e705147
Author: Denis Ovsienko <[email protected]>
Date:   Thu Apr 10 11:53:33 2008 +0000

   + fix the set statement name

commit 841f7a57b13b8cba4efd51db4e7ac13fd403b17b
Author: Denis Ovsienko <[email protected]>
Date:   Thu Apr 10 11:47:45 2008 +0000

   + [bgpd] Added new route-map set statement: "as-path ignore"

commit 693b67b2b20510e0faee87a0950595832ce71054
Author: Paul Jakma <[email protected]>
Date:   Thu Mar 13 03:31:24 2008 +0000

   [bgpd] remove unnecessary 0 entries from struct message's

   2008-03-13 Paul Jakma <[email protected]>

       * (various) Remove 0 entries from struct message's, unneeded due to
         recent improvements in mes_lookup/LOOKUP.

commit ce6ab03a273beb903731621153722511910ebbe5
Author: Paul Jakma <[email protected]>
Date:   Thu Mar 13 03:28:13 2008 +0000

   [ripd] remove unnecessary 0 entries from struct message's

   2008-03-13 Paul Jakma <[email protected]>

       * ripd.c/rip_interface.c: Remove 0 entries from rip_msg
         ri_version_msg struct message's, not needed with recent fixes
         to mes_lookup.

commit 11486b5265b2e0e2cf8b140018c47bd9a35cba93
Author: Paul Jakma <[email protected]>
Date:   Thu Feb 28 23:26:02 2008 +0000

   [lib] Fix the struct message LOOKUP function to be more robust

   2008-02-28 Paul Jakma <[email protected]>

       * log.c: (mes_lookup) Sowmini Varadhan diagnosed a problem where
         this function can cause a NULL dereference, on lookups for unknown
         indices, or messages with NULL strings. Can occur, e.g., debug
         logging code when processing received messages. Fixed to accept a
         pointer to a default string to be used if there is no match.
       * log.h: LOOKUP adjusted to match

commit 5f56808431778fe8878ea8ea94225cca08884b48
Author: Paul Jakma <[email protected]>
Date:   Thu Feb 28 00:09:04 2008 +0000

   [lib/linklist] Enforce "nodes must have data" invariant more rigorously

   2008-02-28 Paul Jakma <[email protected]>

       * linklist.c: This implementation expects that the data pointer not
         be null, e.g. listgetdata() asserts this. The list add methods
         don't apply the same sanity check.

         Noted by Jim Carlson in bug #437.

commit 20e5ff0a8893ffad6d31739d68d224931c4a7992
Author: Denis Ovsienko <[email protected]>
Date:   Tue Feb 26 14:02:24 2008 +0000

   + fix bug#326 by rib_lookup_and_pushup()

commit dea04441fb51f74dc25f6ab4bd756b4159d961b6
Author: Paul Jakma <[email protected]>
Date:   Tue Feb 26 09:16:09 2008 +0000

   [c++] remove/rename some names in headers that clash with C++ reserved words

   2008-01-30 Peter Szilagyi <[email protected]>

       * lib/stream.h: Remove named 'new' parameter in prototype
         for c++ header compatibility.
       * ospfd/ospf_opaque.h: ditto
       * ospfd/ospfd.h: Renamed struct export to _export for c++
         header compatibility.
       * ospf6d/ospf6_area.h: ditto

commit d38e8d63fdfc0aefb78d0a51ffdf06d8b47993fd
Author: Paul Jakma <[email protected]>
Date:   Sat Feb 23 16:57:03 2008 +0000

   [tests:bgp] Further tests for 0 sized as_path segments

   2008-02-23 Paul Jakma <[email protected]>

       * aspath_test.c: Test for 0-ASN sequences that still have data.

commit c7350c48462e305f18f6fb02eceac170f892e450
Author: Paul Jakma <[email protected]>
Date:   Tue Jan 29 19:29:44 2008 +0000

   [isisd] Bug #437: fix ssert caused by bad list management

   2008-01-29 James Carlson <[email protected]>

       * Fix bug #437, assert due to bogus index management
       * isis_flags.c: (flags_initialize) new
       * (flags_get_index) fix off by one, leading to list assert
         on null node data.
         (flags_free_index) ditto.
       * isisd.c: (isis_area_create) use flags_initialize
         (isis_area_destroy) deconfigure circuits when
         taking down area.

commit 90b68769a1a1d7623cae02be71b7a8e4094455ad
Author: Paul Jakma <[email protected]>
Date:   Tue Jan 29 17:26:34 2008 +0000

   [bgpd] fix crash on startup if compiled IPv4-only

   2008-01-29 Jorge Boncompte <[email protected]>

       * bgp_network.c: (bgp_socket) IPv4-only version crashes if -l is not
         used as address will be null.

commit 75f043990f9e3fe76a59605df8b0c1283a5b2ec6
Author: Paul Jakma <[email protected]>
Date:   Sun Jan 13 18:52:15 2008 +0000

   [solaris] add missing makefile.am dependency

   2008-01-13 Paul Jakma <[email protected]>

       * Makefile.am: pkg target should depend on the 'depend.%' files.
         Crops up now that solaris/ doesn't get descended into by
         general Quagga build.

commit 6418bedf29a5d573c4b39cc1ea4b0f5c91af22cd
Author: Andrew J. Schorr <[email protected]>
Date:   Fri Jan 11 16:33:59 2008 +0000

   [link-detect] Improve BSD support.
   2008-01-11 Ingo Flaschberger <[email protected]>

       * configure.ac: Improve HAVE_BSD_LINK_DETECT test.

commit 55edb0d446a272907715b0825ff008557174d34a
Author: Andrew J. Schorr <[email protected]>
Date:   Fri Jan 11 15:57:13 2008 +0000

   [link-detect] Improve BSD support.
   2008-01-11 Andrew J. Schorr <[email protected]>

       * lib/zebra.h: Revert previous change, no need to include
         <net/if_media.h> here.
       * zebra/ioctl.c: If HAVE_BSD_LINK_DETECT is defined, include
         <net/if_media.h>
         (if_get_flags) Remove debug messages about BSD link state.
       * zebra/kernel_socket.c: (bsd_linkdetect_translate) If link state
         is unknown, we should set the IFF_RUNNING flag.

commit c543a1737173fb438f1d8c06f650b2d0d479f45c
Author: Andrew J. Schorr <[email protected]>
Date:   Thu Jan 10 15:24:32 2008 +0000

   [link-detect] Try to get BSD link-detect to work properly.

   2008-01-10 Ingo Flaschberger <[email protected]>

       * configure.ac: Define HAVE_BSD_LINK_DETECT if <net/if_media.h> is
         present.
       * lib/zebra.h: If HAVE_BSD_LINK_DETECT is defined,
         include <net/if_media.h>.
       * zebra/ioctl.c: (if_get_flags) If HAVE_BSD_LINK_DETECT, use the
         SIOCGIFMEDIA ioctl to ascertain link state.
       * zebra/kernel_socket.c: (bsd_linkdetect_translate) New function to
         map the ifm_data.ifi_link_state value into the IFF_RUNNING flag.
         (ifm_read) Call bsd_linkdetect_translate to fix the IFF_RUNNING
         flag before calling if_flags_update.

commit 3f087670efa606021cde2f6a9615ac7e07aec2a5
Author: Andrew J. Schorr <[email protected]>
Date:   Tue Jan 8 20:12:46 2008 +0000

   [link-detect] Static interface routes should behave properly with link-detect.

   2008-01-08 Michael Larson <[email protected]>

       * zebra_rib.c: (nexthop_active_check) Replace if_is_up with
         if_is_operative to solve problems with static interface
         routes not behaving properly with link-detect.

commit 912df1e8ab61962c29eb4faff17a7d75c4905b84
Author: Paul Jakma <[email protected]>
Date:   Tue Jan 8 13:50:11 2008 +0000

   [lib] add mising UL qualifier to numerical constant

   2008-01-08 Pavol Rusnak <[email protected]>

       * memory.c: (mtype_memstr) Fix accidental shift past width of type,
         constant should have been forced to UL, rather than being left to
         default to int.

commit 370b64a2ad38e43b4bed028960481bbf4192becd
Author: Paul Jakma <[email protected]>
Date:   Sat Dec 22 16:49:52 2007 +0000

   [bgpd] Fix number of DoS security issues, restricted to configured peers.

   2007-12-22 Paul Jakma <[email protected]>

       * Fix series of vulnerabilities reported by "Mu Security
         Research Team", where bgpd can be made to crash by sending
         malformed packets - requires that bgpd be configured with a
         session to the peer.
       * bgp_attr.c: (bgp_attr_as4_path) aspath_parse may fail, only
         set the attribute flag indicating AS4_PATH if we actually managed
         to parse one.
         (bgp_attr_munge_as4_attrs) Assert was too general, it is possible
         to receive AS4_AGGREGATOR before AGGREGATOR.
         (bgp_attr_parse) Check that we have actually received the extra
         byte of header for Extended-Length attributes.
       * bgp_attr.h: Fix BGP_ATTR_MIN_LEN to account for the length byte.
       * bgp_open.c: (cap_minsizes) Fix size of CAPABILITY_CODE_RESTART,
         incorrect -2 left in place from a development version of as4-path
         patch.
       * bgp_packet.c: (bgp_route_refresh_receive) ORF length parameter
         needs to be properly sanity checked.
       * tests/bgp_capability_test.c: Test for empty capabilities.

commit a7f93f3e060fdb2dc7bf5ff4ed4563d4b689bc6c
Author: Denis Ovsienko <[email protected]>
Date:   Tue Dec 18 15:13:06 2007 +0000

   + fixed no_set_aspath_prepend() to correctly match existing "set" statement

commit b63e60899f2eba733fadbca020cdbe98d9d58b5f
Author: Paul Jakma <[email protected]>
Date:   Wed Nov 21 21:25:34 2007 +0000

   [bgpd] Fix typo in previous commit to bgp_main.c

commit 7e992e95e3fb9453efb6018dca9016139a8396f1
Author: Paul Jakma <[email protected]>
Date:   Tue Nov 13 09:32:23 2007 +0000

   [bgpd] small fix for crash if 'listenon' argument is not given

   2007-11-13 Paul Jakma <[email protected]>

       * bgp_main.c: Small fix, format string parameters should never
         be NULL - it's not strictly valid and some libc's don't let
         you away with.

commit 6ce80bdb2523ad82cef813f6d8a7e0d4daa8cfae
Author: Denis Ovsienko <[email protected]>
Date:   Mon Nov 12 14:55:01 2007 +0000

   + fixed bug #418 (changing address on an existing interface doesn't cause existing static routes to be revalidated)

commit 3a02d1f7fb778a1ea4f45d037f13dfcd126e2337
Author: Paul Jakma <[email protected]>
Date:   Thu Nov 1 14:29:11 2007 +0000

   2007-10-30 Nick Hilliard <[email protected]>

       * bgp_main.c: Add 'listenon' argument, to pass address to bind
         to.
       * bgp_network.c: (bgp_socket) Extend to take bind address.
       * bgpd.c: (bgp_init) Pass stored address.
       * bgpd.h: (struct bgp_master) storage for bind address

   2007-11-01 Paul Jakma <[email protected]>

       * tools/multiple-bgpd.sh: New, quick script to launch a bunch of
         bgpds.

commit 5619f56ba38b255f699f76b327cf82388c4a90cf
Author: Denis Ovsienko <[email protected]>
Date:   Wed Oct 24 13:13:21 2007 +0000

   + fix the bug reported by Milan Kocian (IPv6 route handling was broken by the RIB debug changeset).
   after

commit fbdbf1281660de877ef687a82145e82934f35744
Author: Paul Jakma <[email protected]>
Date:   Mon Oct 22 18:34:41 2007 +0000

   [ospf6d] Fix removal of defunct ASBR routes

   2007-10-22 Phil Spagnolo <[email protected]>

       * ospf6_asbr.c: (ospf6_asbr_lsentry_remove) Remove shortcut
         of LSDB search - it's based on assumption non-BEST routes
         can't have ASBR routes, which appears to be wrong. Safest to
         search.

commit a56ef88348be9f019631d6bc92801aae647be05b
Author: Paul Jakma <[email protected]>
Date:   Mon Oct 22 15:53:17 2007 +0000

   [snmp-smux] Fix problems if 'smux peer ...' is issued multiple times

   2007-10-22 Lorenzo Colitti <[email protected]>

       * smux.c: (smux_stop) Avoid cancelling a defunct thread pointer
         (smux_start) Stop smux before trying to start it, possibly again.

commit 9c1deac7d6f68b14a130ad99d8a7e3329f570f11
Author: Paul Jakma <[email protected]>
Date:   Mon Oct 22 12:02:57 2007 +0000

   [news] Fix top-line version

commit 5b5bfb8c06d0dc444e2c3da8cc5c893519981834
Author: Paul Jakma <[email protected]>
Date:   Thu Oct 18 16:33:08 2007 +0000

   [tests] Forgot to commit ecommunity_tests.c

commit 0b2aa3a0a8b095bdef1eddda117d173af75dede2
Author: Paul Jakma <[email protected]>
Date:   Sun Oct 14 22:32:21 2007 +0000

   [bgpd] Merge AS4 support

   2007-10-14 Paul Jakma <[email protected]>

       * NEWS: Note that MRT dumps are now version 2
       * (general) Merge in Juergen Kammer's AS4 patch.

   2007-09-27 Paul Jakma <[email protected]>

       * bgp_aspath.c: (assegment_normalise) remove duplicates from
         from sets.
         (aspath_reconcile_as4) disregard a broken part of the RFC around
         error handling in path reconciliation.
       * aspath_test.c: Test dupe-weeding from sets.
         Test that reconciliation merges AS_PATH and AS4_PATH where
         former is shorter than latter.

   2007-09-26 Paul Jakma <[email protected]>

       * aspath_test.c: Test AS4_PATH reconcilation where length
         of AS_PATH and AS4_PATH is same.

   2007-09-25 Paul Jakma <[email protected]>

       * bgp_open.c: (peek_for_as4_capability) Fix to work.
       * bgp_packet.c: (bgp_open_receive) Fix sanity check of as4.
       * tests/bgp_capability_test.c: (general) Extend tests to validate
         peek_for_as4_capability.
         Add test of full OPEN Option block, with multiple capabilities,
         both as a series of Option, and a single option.
         Add some crap to beginning of stream, to prevent code depending
         on getp == 0.

   2007-09-18 Paul Jakma <[email protected]>

       * bgp_open.c: (bgp_capability_as4) debug printf inline with others.
         (peek_for_as4_capability) There's no need to signal failure, as
         failure is better dealt with through full capability parser -
         just return the AS4, simpler.
       * bgp_packet.c: (bgp_open_receive) Update to match
         peek_for_as4_capability change.
         Allow use of BGP_AS_TRANS by 2b speakers.
         Use NOTIFY_OPEN_ERR rather than CEASE for OPEN parsing errors.
         (bgp_capability_msg_parse) missing argument to debug print
         (bgp_capability_receive) missing return values.
       * tests/bgp_capability_test.c: (parse_test) update for changes to
         peek_for_as4_capability

   2007-07-25 Paul Jakma <[email protected]>

       * Remove 2-byte size macros, just make existing macros take
         argument to indicate which size to use.
         Adjust all users - typically they want '1'.
       * bgp_aspath.c: (aspath_has_as4) New, return 1 if there are any
         as4's in a path.
         (aspath_put) Return the number of bytes actually written, to
         fix the bug Juergen noted: Splitting of segments will change
         the number of bytes written from that already written to the
         AS_PATH header.
         (aspath_snmp_pathseg) Pass 2-byte flag to aspath_put. SNMP
         is still defined as 2b.
         (aspath_aggregate) fix latent bug.
         (aspath_reconcile_as4) AS_PATH+NEW_AS_PATH reconciliation
         function.
         (aspath_key_make) Hash the AS_PATH string, rather than
         just taking the addition of assegment ASes as the hash value,
         hopefully sligthly more collision resistant.
         (bgp_attr_munge_as4_attrs) Collide the NEW_ attributes
         together with the OLD 2-byte forms, code Juergen
         had in bgp_attr_parse but re-organised a bit.
         (bgp_attr_parse) Bunch of code from Juergen moves
         to previous function.
         (bgp_packet_attribute) Compact significantly by
         just /always/ using extended-length attr header.
         Fix bug Juergen noted, by using aspath_put's
         (new) returned size value for the attr header rather
         than the (guesstimate) of aspath_size() - the two could
         differ when aspath_put had to split large segments, unlikely
         this bug was ever hit in the 'wild'.
         (bgp_dump_routes_attr) Always use extended-len and
         use aspath_put return for header length. Output 4b ASN
         for AS_PATH and AGGREGATOR.
       * bgp_ecommunity.c: (ecommunity_{hash_make,cmp}) fix
         hash callback declarations to match prototypes.
         (ecommunity_gettoken) Updated for ECOMMUNITY_ENCODE_AS4,
         complete rewrite of Juergen's changes (no asdot support)
       * bgp_open.c: (bgp_capability_as4) New, does what it says
         on the tin.
         (peek_for_as4_capability) Rewritten to use streams and
         bgp_capability_as4.
       * bgp_packet.c: (bgp_open_send) minor edit
         checked (in the abstract at least) with Juergen.
         Changes are to be more accepting, e.g, allow AS_TRANS on
         a 2-byte session.
       * (general) Update all commands to use CMD_AS_RANGE.
       * bgp_vty.c: (bgp_clear) Fix return vals to use CMD_..
         Remove stuff replicated by VTY_GET_LONG
         (bgp_clear_vty) Return bgp_clear directly to vty.
       * tests/aspath_test.c: Exercise 32bit parsing. Test reconcile
         function.
       * tests/ecommunity_test.c: New, test AS4 ecommunity changes,
         positive test only at this time, error cases not tested yet.

   2007-07-25 Juergen Kammer <[email protected]>

       * (general) AS4 support.
       * bgpd.h: as_t changes to 4-bytes.
       * bgp_aspath.h: Add BGP_AS4_MAX and BGP_AS_TRANS defines.
       * bgp_aspath.c: AS_VALUE_SIZE becomes 4-byte, AS16_VALUE_SIZE
         added for 2-byte.
         Add AS16 versions of length calc macros.
         (aspath_count_numas) New, count number of ASes.
         (aspath_has_as4) New, return 1 if there are any as4's in a
         path.
         (assegments_parse) Interpret assegment as 4 or 2 byte,
         according to how the caller instructs us, with a new
         argument.
         (aspath_parse) Add use32bit argument to pass to
         assegments_parse. Adjust all its callers to pass 1, unless
         otherwise noted.
         (assegment_data_put) Adjust to be able to write 2 or 4 byte
         AS, according to new use32bit argument.
         (aspath_put) Adjust to write 2 or 4.
         (aspath_gettoken) Use a long for passed in asno.
       * bgp_attr.c: (attr_str) Add BGP_ATTR_AS4_PATH and
         BGP_ATTR_AS4_AGGREGATOR.
         (bgp_attr_aspath) Call aspath_parse with right 2/4 arg, as
         determined by received-capability flag.
         (bgp_attr_aspath_check) New, code previously in attr_aspath
         but moved to new func so it can be run after NEW_AS_PATH
         reconciliation.
         (bgp_attr_as4_path) New, handle NEW_AS_PATH.
         (bgp_attr_aggregator) Adjust to cope with 2/4 byte ASes.
         (bgp_attr_as4_aggregator) New, read NEW_AGGREGATOR.
         (bgp_attr_parse) Add handoffs to previous parsers for the two
         new AS4 NEW_ attributes.
         Various checks added for NEW/OLD reconciliation.
         (bgp_packet_attribute) Support 2/4 for AS_PATH and
         AGGREGATOR, detect when NEW_ attrs need to be sent.
       * bgp_debug.{c,h}: Add 'debug bgp as4'.
       * bgp_dump.c: MRTv2 support, unconditionally enabled, which
         supports AS4. Based on patches from Erik (RIPE?).
       * bgp_ecommunity.c: (ecommunity_ecom2str) ECOMMUNITY_ENCODE_AS4
         support.
       * bgp_open.c: (peek_for_as4_capability) New, peek for AS4
         capability prior to full capability parsing, so we know which
         ASN to use for struct peer lookup.
         (bgp_open_capability) Always send AS4 capability.
       * bgp_packet.c: (bgp_open_send) AS4 handling for AS field
         (bgp_open_receive) Peek for AS4 capability first, and figure
         out which AS to believe.
       * bgp_vty.c: (bgp_show_peer) Print AS4 cap
       * tests/aspath_test.c: Support asn32 changes, call aspath_parse
         with 16 bit.
       * vtysh/extract.pl: AS4 compatibility for router bgp ASNUMBER
       * vtysh/extract.pl.in: AS4 compatibility for router bgp ASNUMBER
       * vtysh/vtysh.c: AS4 compatibility for router bgp ASNUMBER

commit 7593fddfa1558d086e3142ad96cf2790007f5d84
Author: Denis Ovsienko <[email protected]>
Date:   Thu Oct 4 15:09:19 2007 +0000

   + pidfiles are now always created with 0644 perms instead if LOGFILE_MASK (0600)

commit a3aaf5b0ceb8c5b3821793333b4057df872fbedd
Author: Denis Ovsienko <[email protected]>
Date:   Thu Oct 4 10:49:21 2007 +0000

   + rib_process() speedup for multi-nexthop route nodes

commit f304cb48f0d7d0ff8f36e7aca8293141a9fa9e60
Author: Denis Ovsienko <[email protected]>
Date:   Wed Oct 3 12:27:16 2007 +0000

   + Minor bugfix: IPv6 prefixes were logged incorrectly in RIB debugging calls. Fixed.

commit 6b148faa291f68dd43bba37fbc0e6b75a2e1e078
Author: Paul Jakma <[email protected]>
Date:   Tue Sep 18 18:07:18 2007 +0000

   [privs/Solaris] Quagga should work in zones with IP instances

   2007-09-18  Paul Jakma <[email protected]>

       * privs.c: definition of ZCAP_NET_ADMIN on Solaris should
         be PRIV_SYS_IP_CONFIG, when that's available. Thus allowing
         Quagga to work with in Solaris zones with exclusive IP
         instances.

commit e08286bcf8a73d1922b0464bef52c1b27eda7c43
Author: Paul Jakma <[email protected]>
Date:   Tue Sep 18 12:11:26 2007 +0000

   [bgpd] Fix typo, which prevented advertisement of MP (non-IPv4) prefixes

   2007-09-17 Paul Jakma <[email protected]>

       * bgp_open.c: (bgp_capability_mp) We were setting
         afc_nego[safi][safi] rather than afc_nego[afi][safi], thus
         failling to announce any non-IPv4 prefixes. Remove the extra,
         typo-ed character.
       * bgp_capability_test.c: Test that peer's adv_recv and adv_nego get
         set correctly for MP capability and given AFI/SAFI.
         Colour OK/failed result so it's easier to find them.

commit fb31c0fea38c863ddcda56f46c36f065340b9e32
Author: Denis Ovsienko <[email protected]>
Date:   Tue Sep 18 09:03:13 2007 +0000

   + fix missing arg to zlog_warn()

commit f102e75f613af740241dfa7253a2362c7935b9a8
Author: Denis Ovsienko <[email protected]>
Date:   Tue Sep 18 09:01:13 2007 +0000

   + fix minor regression in OSPF sending buffer adjustment logic

commit 96934e6ac6ec9ff6e67131cc8324741b771d5e0d
Author: Denis Ovsienko <[email protected]>
Date:   Fri Sep 14 14:56:28 2007 +0000

   + sayonara old_pid!

commit 91b7351d4e150bbc60786f0facb911fa2c598553
Author: Denis Ovsienko <[email protected]>
Date:   Fri Sep 14 13:31:52 2007 +0000

   + fixed bug #402: now the second zebra process doesn't destroy routes
     of the first one before dying
   + we are not going to receive routing messages originated by old_pid,
     because rib_sweep_route() is called after damon() now. This will allow
     to drop old_pid completely soon.

commit 2d844524ad9431ea1e53336dfcfebd3ee6520cb7
Author: Denis Ovsienko <[email protected]>
Date:   Fri Sep 14 11:31:55 2007 +0000

   Switch from LOOKUP() to lookup() for rtm_type (see bug #401 for details).

commit bd6c86d3a30d27ed4960863a653376c54e5e5e34
Author: Denis Ovsienko <[email protected]>
Date:   Wed Sep 12 15:24:27 2007 +0000

   * rt_socket.c: (kernel_rtm_ipv4) prefix_buf could be passed
     to zlog_err() uninitialized with debug disabled. Fixed.