Hi Dear VPP There is a difference between VPP and Kernel's behavior in route management. For instance, consider that a route like "ip route add x.x.x.x/x via y.y.y.y dev ethX" is added. It would be deleted when ethx down performed in Kernel. While The VPP behavior is that forwarding of that route entry in unicast-ip4-chain will be changed to dpo-drop. Now what will happen if we change the state of the interface to up again?! defined route associated with the kernel interface will be disappeared while VPP will make previously added route lives again. Which behavior is acceptable? and is it necessary for VPP to operate the same as kernel anyway?
Kernel senario: given: ethX interface has ip and is up when: "ip route add x.x.x.x/x via y.y.y.y dev ethX" and "ifconfig ethX down" then: any route dose not exist for ethX in "route -n" output. then: "ifconfig ethX up" and any route dose not exist for ethX in "route -n" output again. VPP senario: given: GigabitEthernetX/0/0 interface has ip and is up when: "ip route add x.x.x.x/x via y.y.y.y GigabitEthernetX/0/0" and "set interface state GigabitEthernetX/0/0 down" then: "show ip fib x.x.x.x/x detail" => forwarding: unicast-ip4-chain ... dpo-drop then: "set interface state GigabitEthernetX/0/0 up" and "show ip fib x.x.x.x/x detail" => forwarding: unicast-ip4-chain ... arp-ipv4: via y.y.y.y GigabitEthernetX/0/0
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#14481): https://lists.fd.io/g/vpp-dev/message/14481 Mute This Topic: https://lists.fd.io/mt/41180423/21656 Group Owner: [email protected] Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [[email protected]] -=-=-=-=-=-=-=-=-=-=-=-
