On 5/8/25 4:15 PM, Serhat Rıfat Demircan via discuss wrote: > Hello Everyone, > > I've been testing OVS hardware offloaded ports with Openstack for a while. > I have ConnectX-5 and ConnectX-6 cards in my lab. Iperf tests are looking > promising, but when I try to use hardware offloaded ports for krbd traffic, > after some time, traffic falls back to the software path. Another interesting > thing for me is without changing anything, it can also switch back to the > hardware path. I could not find the real reason. > > I'm not sure if this is a configuration issue or if I'm hitting to cards > limit. > If any of you guys are experienced with this all ideas are more than welcome. > > Openstack version: Caracal > Operation System: Ubuntu 22.04 > OVS version: 3.3.0 > > I see tc flow change: > > *From Offloaded:* > $ tc filter show dev ens37f0npf0vf6 ingress > filter protocol ip pref 2 flower chain 0 > filter protocol ip pref 2 flower chain 0 handle 0x1 > dst_mac 00:1c:73:aa:bb:cc > src_mac fa:16:7e:2e:8c:f0 > eth_type ipv4 > ip_flags nofrag > skip_sw > in_hw in_hw_count 1 > action order 1: vlan push id 111 protocol 802.1Q priority 0 pipe > index 4 ref 1 bind 1 > no_percpu > used_hw_stats delayed > > action order 2: mirred (Egress Redirect to device openstack) stolen > index 4 ref 1 bind 1 > cookie 605b82e5b94ef3fa63c3e78349851a27 > no_percpu > used_hw_stats delayed > > *To software path:* > $ tc filter show dev ens37f0npf0vf6 ingress > filter protocol ip pref 2 flower chain 0 > filter protocol ip pref 2 flower chain 0 handle 0x1 > dst_mac 00:1c:73:aa:bb:cc > src_mac fa:16:7e:2e:8c:f0 > eth_type ipv4 > ip_flags nofrag > not_in_hw > action order 1: vlan push id 111 protocol 802.1Q priority 0 pipe > index 4 ref 1 bind 1 > no_percpu > > action order 2: skbedit ptype host pipe > index 11 ref 1 bind 1 > > action order 3: mirred (Ingress Mirror to device br-ex) pipe > index 13 ref 1 bind 1 > cookie 5037716e444e83b81c4ef9844df2377d > no_percpu > > action order 4: mirred (Egress Redirect to device openstack) stolen > index 15 ref 1 bind 1 > cookie 5037716e444e83b81c4ef9844df2377d > no_percpu
Here you can see that traffic is getting broadcasted on all ports in the br-ex bridge, and since some of these ports are virtual, this operation cannot be offloaded to hardware. I would guess, after some time the FDB entry expires and OVS no longer knows on which port the traffic supposed to be sent out. Is that a unidirection traffic? i.e. are you also sending traffic in the opposite direction? If not, are you sending some periodic ARP requests for the bridge to learn where the destination is? If the traffic is bidirectional, then it might be a NIC driver/firmware issue not providing updated stats. Best regards, Ilya Maximets. _______________________________________________ discuss mailing list disc...@openvswitch.org https://mail.openvswitch.org/mailman/listinfo/ovs-discuss
