janjanovicsk / quagga-eigrp Goto Github PK

There is a bug in eigrp_network.c:440.
if ((metric1->bandwith = metric2->bandwith)
I think you want to write a " == "?

I've merged EIGRP-Development to latest savannah master here:

https://github.com/donaldsharp/quagga/tree/EIGRP-Development

I've also fixed up a few of the warnings in a couple commits.

I haven't tested anything.

I have a very simple network (see attached)

Looking into the EIGRP topology on node n1 (upper left hand corner) shows:
n1# show ip eigrp topology

A.B.C.D Netwok to display information about
A.B.C.D/nn IP prefix /, e.g., 192.168.0.0/16
all-links Show all links in topology table
detail Show all links in topology table
summary Show a summary of the topology table
n1# show ip eigrp topology all-links

EIGRP Topology Table for AS(1)/ID(192.168.0.1)

Codes: P - Passive, A - Active, U - Update, Q - Query, R - Reply
r - reply Status, s - sia Status

P 192.168.0.0/24, 1 successors, FD is 256256, serno: 0
via Connected, eth0
P 192.168.1.0/24, 1 successors, FD is 256256, serno: 0
via Connected, eth1
via 192.168.0.2 (258256/257256), eth0
P 192.168.2.0/24, 1 successors, FD is 257256, serno: 0
via 192.168.1.2 (257256/256256), eth1 <-- FEASIBILITY CONDITION IS MET
via 192.168.0.2 (258256/257256), eth0
A 192.168.3.0/24, 1 successors, FD is 5000, serno: 0
via 192.168.1.2 (7000/6000), eth1
via 192.168.0.2 (9000/8000), eth0
P 192.168.4.0/24, 1 successors, FD is 1000, serno: 0
via 192.168.1.2 (1000/0), eth1 <-- FEASIBILITY CONDITION IS MET
A 192.168.5.0/24, 1 successors, FD is 5000, serno: 0
via 192.168.1.2 (7000/6000), eth1
via 192.168.0.2 (9000/8000), eth0

However, when I am looking at kernel routing table, not all the routes (discovered by EIGRP) are present (see below)
n1# exit
root@n1:/tmp/pycore.42477/n1.conf# route
Kernel IP routing table
Destination Gateway Genmask Flags Metric Ref Use Iface
192.168.0.0 * 255.255.255.0 U 0 0 0 eth0
192.168.1.0 * 255.255.255.0 U 0 0 0 eth1
192.168.2.0 192.168.1.2 255.255.255.0 UG 257256 0 0 eth1

The same is true for other nodes. Doing some google searches, I came across this blog entry
(http://brezular.com/2016/09/02/quagga-routing-software-with-eigrp-routing-protocol/) which discusses the same issue. Delaying the start of the EIGRP daemon did not solve the problem for me (as it seems to have done - in same cases - for the author of the blog); I even waited up to 1 minute after the start of the zebra daemon to start the EIGRP daemon.
For completeness, here is the EIGRP configuration of the n1 node:
n1# show running-config
Building configuration...

Current configuration:
!
log file /var/log/eiprg.log
!
service integrated-vtysh-config
!
interface eth0
ip hello-interval eigrp 1
ip hold-time eigrp 3
ipv6 nd suppress-ra
!
interface eth1
ip hello-interval eigrp 1
ip hold-time eigrp 3
ipv6 nd suppress-ra
!
interface lo
!
router eigrp 1
eigrp router-id 192.168.0.1
network 192.168.0.1/24
network 192.168.1.1/24
!
ip forwarding
ipv6 forwarding
!
line vty
!
end

Am I missing something?

thanks,
-silviu

Hello!
These days, I use the Eigrp protocol in dce-ns3, but it has some errors. I find the REPLY packet has wrong data. So I see the souce codes.
And I find there a function -- eigrp_fsm_event_keep_state in the file named eifrp_fsm.c. In here, one line code is eigrp_send_reply(msg->adv_router, msg->entry);. But I find the declaration of eigrp_send_reply is eigrp_send_reply (struct eigrp_neighbor *nbr, struct eigrp_prefix_entry *pe).
Obviously, the type of msg->entry is different from the type of pe.

If you want to convert the point, it will wrong. Because there is a pointer to struct eigrp_prefix_entry rather than a struct variable in struct eigrp_neighbor_entry.
So what does it means? Is it a bug？
If I have something wrong, please tell me. Thanks a lot!

It be a first update packet that has some tlvs to nbr.

when eigrp runing on many nodes, it will cause that the length of data written is more than MTU, and program will abort itself

I noticed that the sernos of the prefixes in EIGRP Topology Table are always zero,
and I checked the EIGRP code,and found the variable serno is defined in struct eigrp and eigrp_prefix_entry but not used.
Is there a plan to implement the serno?

Regards