Hi Claudio & Co.,
I am running 4.3-current. I am using a very basic bgpd config

file that worked for 4.2 very well. Unfortunatelly, with

-current my router does not want to announce our prefix anymore.
Any ideas?
Thanks,

- Christian
Details: (my as: 303, transit: 304)
I am able to receive prefixes from my transit provider.

bgpctl shows the following (77.X.Y.0/21 being our prefix):
# bgpctl show rib 77.X.Y.0

flags: * = Valid, > = Selected, I = via IBGP, A = Announced

origin: i = IGP, e = EGP, ? = Incomplete
flags destination         gateway          lpref   med aspath origin

       77.X.Y.0/21       A.B.C.D 105     0 304 [SOMEAS] 303 i

I*    77.X.Y.0/21       77.X.Y.2           100     0 i

AI    77.X.Y.0/21       0.0.0.0            100     0 i
So there are two funny things:
- I get my own prefix (announced by our other router, 77.X.Y.2)

via the transit provider.

- The line with my announced prefix ("AI") is missing the "*".
2# bgpctl network show

flags: S = Static

flags destination

*     77.X.Y.0/21
/etc/bgpd.conf:
AS 303
router-id 77.X.Z.3
network 77.X.Y.0/21
group "Peering XXXXXXXXXXX" {
         remote-as 304
         neighbor A.B.C.D {

                 tcp md5sig password "xxxxxxx"

                 announce self

                 set localpref 105

         }

}
group IBPG {
group IBPG {
         remote-as 303
         neighbor 77.X.Y.2 {

                 tcp md5sig password "xxxxxxxxx"

         }
     [...some other internal IBGP sessions...]

}
deny from any
allow quick from group IBPG

allow from any inet prefixlen 8 - 24

allow from any prefix 0.0.0.0/0
# filter bogus networks

deny from any prefix 10.0.0.0/8 prefixlen >= 8

deny from any prefix 172.16.0.0/12 prefixlen >= 12

deny from any prefix 192.168.0.0/16 prefixlen >= 16

deny from any prefix 169.254.0.0/16 prefixlen >= 16

deny from any prefix 192.0.2.0/24 prefixlen >= 24

deny from any prefix 224.0.0.0/4 prefixlen >= 4

deny from...
[ message continues ]

The missing * in tha AI output is an error when composing the control

message to bgpctl. The problem with the looped AS path is a bit more evil.

In the decision process the wrong flags field was inspected and because of

that loops remained undetected.

Having the loop in the table with a higher lpref will result in missing

anouncements. That's why your own route is not selected (missing ">").
I hope the attached diff fixes all your issues. At least it seems to work

for me.
Index: rde.c

===================================================================

RCS file: /cvs/src/usr.sbin/bgpd/rde.c,v

retrieving revision 1.230

diff -u -p -r1.230 rde.c

--- rde.c	26 Feb 2008 19:58:51 -0000	1.230

+++ rde.c	16 Apr 2008 14:04:10 -0000

@@ -1717,7 +1717,7 @@ rde_dump_rib_as(struct prefix *p, struct

 		rib.flags |= F_RIB_INTERNAL;

 	if (asp->flags & F_PREFIX_ANNOUNCED)

 		rib.flags |= F_RIB_ANNOUNCE;

-	if (asp->nexthop != NULL && asp->nexthop->state == NEXTHOP_REACH)

+	if (asp->nexthop == NULL || asp->nexthop->state == NEXTHOP_REACH)

 		rib.flags |= F_RIB_ELIGIBLE;

 	if (asp->flags & F_ATTR_LOOP)

 		rib.flags &= ~F_RIB_ELIGIBLE;

@@ -2846,4 +2846,3 @@ sa_cmp(struct bgpd_addr *a, struct socka
 	return (0);

 }

-

Index: rde_decide.c

===================================================================

RCS file: /cvs/src/usr.sbin/bgpd/rde_decide.c,v

retrieving revision 1.49

diff -u -p -r1.49 rde_decide.c

--- rde_decide.c	27 Nov 2007 01:13:54 -0000	1.49

+++ rde_decide.c	16 Apr 2008 14:43:48 -0000

@@ -121,14 +121,14 @@ prefix_cmp(struct prefix *p1, struct pre

 	if (!(p2->flags & F_LOCAL))

 		return (1);
+	asp1 = p1->aspath;

+	asp2 = p2->aspath;

+

 	/* only loop free pathes are eligible */

-	if (p1->flags & F_ATTR_LOOP)

+	if (asp1->flags & F_ATTR_LOOP)

 		return (-1);

-	if (p2->flags & F_ATTR_LOOP)

+	if (asp2->flags & F_ATTR_LOOP)

 		return (1);

-

-	asp1 = p1->aspath;

-	asp2 = p2->aspat...
[ message continues ]

Since we discussed this offline, here for the record:

The patch works fine for me, no problems during the last few days.
Thanks,

  Christian