On Wed, Nov 18, 2026 at 2:19 PM Lucas Vargas Dias <[email protected]> wrote:
> Ok, it makes sense to use an ACL or policy and it works very well for > static scenarios. > However, both approaches need the CMS to keep a port group in sync with > each subnet and to tie the lifecycle of the policies/ACLs to the subnet. > > However, I had other idea, propose a Logical_Router option instead: > options:neigh_resolve_known_only=true > > For each router port whose peer switch is eligible, northd adds a > flow that drops packets whose next hop is in the port's networks but > was resolved neither statically nor via MAC_Binding, e.g.: > > table=lr_in_arp_request, priority=105, > match=(eth.dst == 00:00:00:00:00:00 && reg9[9] == 1 && > outport == "lrp-subnet" && reg0 == 192.168.30.0/24), > action=(drop;) > > Static entries and learned MAC_Bindings never reach this flow. > Router ports are skipped when the peer switch is a transit switch > (interconn-ts) or has "remote", "unknown" or "localnet" ports. > Router peers resolved dynamically (dynamic_neigh_routers) and > virtual IPs on the subnet keep a priority 110 arp/nd_ns flow. > > Since the flows are derived from the router port networks and the > switch ports, deleting or re-creating a subnet, with the same or a > different CIDR, needs no extra configuration and leaves nothing > behind. The option is set once, when the router is created, like > dynamic_neigh_routers. > > What do you think? > > I can send the new patch. > So, when someone from the community has time, they could review. > > My main point is to have an easier way to fix it (inappropriate packet > consuming CPU). > In my scenario, I have much more packets like this than packets that will > generate an entry > in the mac_binding table. > I see, I think it makes sense to send a new RFC patch and discuss it there. > > Regards, > Lucas > > > Em qua., 18 de nov. de 2026 às 09:14, Ales Musil <[email protected]> > escreveu: > >> >> >> On Wed, Nov 18, 2026 at 1:07 PM Lucas Vargas Dias >> <[email protected]> wrote: >> >>> Hi Ales, how are you? >>> >> >> Hi Lucas, >> >> >>> >>> Sorry, for talking about this again. >>> >> >> no worries. >> >> >>> >>> I'm using dynamic_neigh_routers in my scenario, so copp could drop valid >>> packets. >>> Or at least to have high latency to resolve the ARP. >>> >> >> This is indeed the downside of using CoPP. Looking back at the thread >> would it be feasible to create an ACL similar to the drop flow proposed >> by this patch? That still doesn't require a new option and doesn't >> have the CoPP downside. >> >> Thanks, >> Ales >> >>> >>> Following my example, a logical flow like could works: >>> ovn-nbctl show lr13 >>> router 3ee35ba1-605b-49cf-add4-9c0f2b521d84 (lr13) >>> port ts1001-lr13 >>> mac: "12:00:00:d9:b5:d3" >>> ipv6-lla: "fe80::1000:ff:fed9:b5d3" >>> networks: ["169.254.11.3/24"] >>> port rp13-bar >>> mac: "00:01:ff:22:00:03" >>> ipv6-lla: "fe80::201:ffff:fe22:3" >>> networks: ["192.168.30.1/24"] >>> >>> >>> port ts1001-lr13 is used for interconnect, rp13-bar is used for gw of >>> subnet. >>> >>> I'm thinking in something like: >>> table=27(lr_in_arp_request), priority=105, match=(eth.dst == >>> 00:00:00:00:00:00 && reg9[9] == 1 && outport == "rp13-bar" && reg0 == >>> 192.168.30.0/24), action=(drop;) >>> table=27(lr_in_arp_request ), priority=100 , match=(eth.dst == >>> 00:00:00:00:00:00 && reg9[9] == 0), action=(nd_ns { nd.target = xxreg0; >>> output; }; next;) >>> table=27(lr_in_arp_request ), priority=100 , match=(eth.dst == >>> 00:00:00:00:00:00 && reg9[9] == 1), action=(arp { eth.dst = >>> ff:ff:ff:ff:ff:ff; arp.spa = reg5; arp.tpa = reg0; arp.op = 1; output; }; >>> next;) >>> table=27(lr_in_arp_request ), priority=0 , match=(1), >>> action=(next;) >>> >>> I'll apply a config in rp13-bar (I'm still thinking of a good name for >>> config), and with config ovn-northd will generate lflow with prio 105. >>> >>> Does it make sense? >>> >>> Regards, >>> Lucas >>> >>> >>> >>> Em qui., 4 de dez. de 2025 às 09:43, Ales Musil <[email protected]> >>> escreveu: >>> >>>> >>>> >>>> On Thu, Dec 4, 2025 at 1:40 PM Lucas Vargas Dias < >>>> [email protected]> wrote: >>>> >>>>> Hi, >>>>> >>>>> We're testing the CoPP, in the first tests, it works correctly. >>>>> >>>>> I agree about avoiding new options. >>>>> >>>>> >>>>> Regards, >>>>> Lucas >>>>> >>>> >>>> Thank you, >>>> >>>> based on that I'll mark the series as deferred for now. >>>> >>>> Regards, >>>> Ales >>>> >>>> >>>>> >>>>> Em sex., 28 de nov. de 2025 às 07:29, Ales Musil <[email protected]> >>>>> escreveu: >>>>> >>>>>> >>>>>> >>>>>> On Mon, Oct 27, 2025 at 9:21 PM Lucas Vargas Dias via dev < >>>>>> [email protected]> wrote: >>>>>> >>>>>>> Consider the following scenario: >>>>>>> (192.168.10.10/24) VM1 -> LS -> LR -> TS -> LR -> LS -> VM2 ( >>>>>>> 192.168.20.20/24) >>>>>>> >>>>>>> Also, LSPs from LS have the addresses configured, it's not used >>>>>>> unknown addresses >>>>>>> in this case. >>>>>>> >>>>>>> Ping from VM1 to VM2 works correctly, it's ok. But, if VM1 try to >>>>>>> flood, >>>>>>> for example the IP 192.168.20.30 (an innexistent LSP), we'll see a >>>>>>> high CPU load in >>>>>>> ovn-controller from the network destination due to the get_arp >>>>>>> function. >>>>>>> For this case, ovn-controller have ovn-is-interconn = true. >>>>>>> >>>>>>> In the connection LR -> TS -> LR all addresses are known, so arp >>>>>>> resolve >>>>>>> could be disable. For this case, add the option >>>>>>> disable_arp_resolve=true >>>>>>> to logical_router_port drop the packets for unknown addresses. If the >>>>>>> communication is IPv6, configure the option disable_nd_resolve=true. >>>>>>> >>>>>>> >>>>>>> Co-authored-by: Tiago Matos Carvalho Reis <[email protected]> >>>>>>> Signed-off-by: Tiago Matos Carvalho Reis <[email protected]> >>>>>>> Signed-off-by: Lucas Vargas Dias <[email protected]> >>>>>>> --- >>>>>>> >>>>>> >>>>>> Hi Lucas, >>>>>> >>>>>> sorry for the late response, but I have to agree with Lorenzo and >>>>>> Alexander on this. Both of them mentioned CoPP as a solution which >>>>>> makes sense considering the problem seems to be high ovn-controller >>>>>> CPU usage in pinctrl. And it makes me wonder if there is any reason >>>>>> why couldn't CoPP be used? It was built exactly for this to limit >>>>>> the traffic towards userspace, to protect the control plane. >>>>>> >>>>>> There might be another solution, that is ACLs. Would that be >>>>>> feasible in any capacity for you to create ACLs that would block >>>>>> this traffic? As you said in your case those IPs are known I think >>>>>> it should be possible to create ACL drop flow based on those IPs. >>>>>> >>>>>> While there isn't technically anything wrong with the option, >>>>>> it's the fact that we would be adding another option. I would >>>>>> love to avoid that if possible, especially if any of the solutions >>>>>> above would work. My bet would be on the CoPP, because as I said, >>>>>> it's specifically made for those scenarios. >>>>>> >>>>>> Please let me know what you think. >>>>>> >>>>>> Thanks, >>>>>> Ales >>>>>> >>>>>> >>>>>>> northd/northd.c | 54 ++++++++++++++++++++++++++++++ >>>>>>> ovn-nb.xml | 13 ++++++++ >>>>>>> tests/ovn-northd.at | 81 >>>>>>> +++++++++++++++++++++++++++++++++++++++++++++ >>>>>>> 3 files changed, 148 insertions(+) >>>>>>> >>>>>>> diff --git a/northd/northd.c b/northd/northd.c >>>>>>> index 9da995cea..356dbaa94 100644 >>>>>>> --- a/northd/northd.c >>>>>>> +++ b/northd/northd.c >>>>>>> @@ -14940,6 +14940,60 @@ build_arp_resolve_flows_for_lrp(struct >>>>>>> ovn_port *op, >>>>>>> &op->nbrp->header_, >>>>>>> lflow_ref); >>>>>>> } >>>>>>> + >>>>>>> + bool discard_arp_resolve = smap_get_bool(&op->nbrp->options, >>>>>>> + "disable_arp_resolve", false); >>>>>>> + bool discard_nd_resolve = smap_get_bool(&op->nbrp->options, >>>>>>> + "disable_nd_resolve", false); >>>>>>> + if (discard_arp_resolve || discard_nd_resolve) { >>>>>>> + ds_clear(match); >>>>>>> + >>>>>>> + ds_put_format(match, "inport == %s && >>>>>>> is_chassis_resident(%s) && " >>>>>>> + "(", op->json_key, op->cr_port->json_key); >>>>>>> + >>>>>>> + size_t match_len = match->length; >>>>>>> + for (size_t i = 0; i < op->od->nbr->n_ports; i++) { >>>>>>> + struct nbrec_logical_router_port *lrp = >>>>>>> op->od->nbr->ports[i]; >>>>>>> + struct lport_addresses lrp_networks; >>>>>>> + if (!extract_lrp_networks(lrp, &lrp_networks)) { >>>>>>> + continue; >>>>>>> + } >>>>>>> + >>>>>>> + for (int j = 0; j < lrp_networks.n_ipv4_addrs; j++) >>>>>>> { >>>>>>> + if (!discard_arp_resolve) { >>>>>>> + break; >>>>>>> + } >>>>>>> + struct ipv4_netaddr ipv4 = >>>>>>> lrp_networks.ipv4_addrs[j]; >>>>>>> + struct in6_addr ip = >>>>>>> in6_addr_mapped_ipv4(ipv4.addr); >>>>>>> + if (prefix_is_link_local(&ip, ipv4.plen)) { >>>>>>> + continue; >>>>>>> + } >>>>>>> + ds_put_format(match, "(ip4.dst == %s/%u) || ", >>>>>>> + ipv4.network_s, ipv4.plen); >>>>>>> + } >>>>>>> + for (int j = 0; j < lrp_networks.n_ipv6_addrs; j++) >>>>>>> { >>>>>>> + if (!discard_nd_resolve) { >>>>>>> + break; >>>>>>> + } >>>>>>> + struct ipv6_netaddr ipv6 = >>>>>>> lrp_networks.ipv6_addrs[j]; >>>>>>> + if (prefix_is_link_local(&ipv6.addr, >>>>>>> ipv6.plen)) { >>>>>>> + continue; >>>>>>> + } >>>>>>> + ds_put_format(match, "(ip6.dst == %s/%u) || ", >>>>>>> + ipv6.network_s, ipv6.plen); >>>>>>> + } >>>>>>> + destroy_lport_addresses(&lrp_networks); >>>>>>> + } >>>>>>> + >>>>>>> + if (match->length > match_len) { >>>>>>> + ds_truncate(match, match->length - 4); >>>>>>> + ds_put_format(match, ")"); >>>>>>> + ovn_lflow_add_drop_with_desc(lflows, op->od, >>>>>>> + S_ROUTER_IN_ARP_RESOLVE, 50, >>>>>>> + ds_cstr(match), "No L2 unknown", >>>>>>> + lflow_ref); >>>>>>> + } >>>>>>> + } >>>>>>> } >>>>>>> } >>>>>>> >>>>>>> diff --git a/ovn-nb.xml b/ovn-nb.xml >>>>>>> index 73b5f213f..872741a2f 100644 >>>>>>> --- a/ovn-nb.xml >>>>>>> +++ b/ovn-nb.xml >>>>>>> @@ -4499,6 +4499,19 @@ or >>>>>>> routes in <code>ovn-ic</code> daemon. >>>>>>> </p> >>>>>>> </column> >>>>>>> + >>>>>>> + <column name="options" key="disable_arp_resolve"> >>>>>>> + <p> >>>>>>> + If set to <code>true</code>, disable get_arp for >>>>>>> + known networks. >>>>>>> + </p> >>>>>>> + </column> >>>>>>> + <column name="options" key="disable_nd_resolve"> >>>>>>> + <p> >>>>>>> + If set to <code>true</code>, disable get_nd for >>>>>>> + known networks. >>>>>>> + </p> >>>>>>> + </column> >>>>>>> </group> >>>>>>> >>>>>>> <group title="Attachment"> >>>>>>> diff --git a/tests/ovn-northd.at b/tests/ovn-northd.at >>>>>>> index 5c9cc5cca..7c8237acd 100644 >>>>>>> --- a/tests/ovn-northd.at >>>>>>> +++ b/tests/ovn-northd.at >>>>>>> @@ -10111,6 +10111,87 @@ AT_CHECK([grep -e "ls_in_arp_rsp" S1flows | >>>>>>> ovn_strip_lflows], [0], [dnl >>>>>>> AT_CLEANUP >>>>>>> ]) >>>>>>> >>>>>>> + >>>>>>> +OVN_FOR_EACH_NORTHD_NO_HV([ >>>>>>> +AT_SETUP([check options:disable_arp_resolve and >>>>>>> options:disable_nd_resolve for LRP]) >>>>>>> +ovn_start NORTHD_TYPE >>>>>>> +check ovn-nbctl lr-add S1 >>>>>>> +check ovn-nbctl --wait=sb lrp-add S1 S1-vm1 50:54:00:00:00:010 >>>>>>> 192.168.0.10/24 fd00::2/64 >>>>>>> +check ovn-nbctl --wait=sb lrp-set-gateway-chassis S1-vm1 chassis-1 >>>>>>> + >>>>>>> +ovn-sbctl dump-flows S1 > S1flows >>>>>>> +AT_CAPTURE_FILE([S1flows]) >>>>>>> + >>>>>>> +ovn-sbctl dump-flows S1 > S1flows >>>>>>> +AT_CAPTURE_FILE([S1flows]) >>>>>>> + >>>>>>> +AT_CHECK([grep -e "lr_in_arp_resolve" S1flows | ovn_strip_lflows], >>>>>>> [0], [dnl >>>>>>> + table=??(lr_in_arp_resolve ), priority=0 , match=(1), >>>>>>> action=(drop;) >>>>>>> + table=??(lr_in_arp_resolve ), priority=1 , match=(reg9[[9]] >>>>>>> == 0), action=(get_nd(outport, xxreg0); next;) >>>>>>> + table=??(lr_in_arp_resolve ), priority=1 , match=(reg9[[9]] >>>>>>> == 1), action=(get_arp(outport, reg0); next;) >>>>>>> + table=??(lr_in_arp_resolve ), priority=500 , match=(ip4.mcast >>>>>>> || ip6.mcast), action=(next;) >>>>>>> +]) >>>>>>> + >>>>>>> + >>>>>>> +# Set the disable_arp_resolve option and verify the flow >>>>>>> +check ovn-nbctl --wait=sb set logical_router_port S1-vm1 >>>>>>> options:disable_arp_resolve=true >>>>>>> + >>>>>>> +ovn-sbctl dump-flows S1 > S1flows >>>>>>> +AT_CAPTURE_FILE([S1flows]) >>>>>>> + >>>>>>> +AT_CHECK([grep -e "lr_in_arp_resolve" S1flows | ovn_strip_lflows], >>>>>>> [0], [dnl >>>>>>> + table=??(lr_in_arp_resolve ), priority=0 , match=(1), >>>>>>> action=(drop;) >>>>>>> + table=??(lr_in_arp_resolve ), priority=1 , match=(reg9[[9]] >>>>>>> == 0), action=(get_nd(outport, xxreg0); next;) >>>>>>> + table=??(lr_in_arp_resolve ), priority=1 , match=(reg9[[9]] >>>>>>> == 1), action=(get_arp(outport, reg0); next;) >>>>>>> + table=??(lr_in_arp_resolve ), priority=50 , match=(inport == >>>>>>> "S1-vm1" && is_chassis_resident("cr-S1-vm1") && ((ip4.dst == >>>>>>> 192.168.0.0/24))), action=(drop;) >>>>>>> + table=??(lr_in_arp_resolve ), priority=500 , match=(ip4.mcast >>>>>>> || ip6.mcast), action=(next;) >>>>>>> +]) >>>>>>> + >>>>>>> +check ovn-nbctl --wait=sb set logical_router_port S1-vm1 >>>>>>> options:disable_nd_resolve=true >>>>>>> + >>>>>>> +ovn-sbctl dump-flows S1 > S1flows >>>>>>> +AT_CAPTURE_FILE([S1flows]) >>>>>>> + >>>>>>> +AT_CHECK([grep -e "lr_in_arp_resolve" S1flows | ovn_strip_lflows], >>>>>>> [0], [dnl >>>>>>> + table=??(lr_in_arp_resolve ), priority=0 , match=(1), >>>>>>> action=(drop;) >>>>>>> + table=??(lr_in_arp_resolve ), priority=1 , match=(reg9[[9]] >>>>>>> == 0), action=(get_nd(outport, xxreg0); next;) >>>>>>> + table=??(lr_in_arp_resolve ), priority=1 , match=(reg9[[9]] >>>>>>> == 1), action=(get_arp(outport, reg0); next;) >>>>>>> + table=??(lr_in_arp_resolve ), priority=50 , match=(inport == >>>>>>> "S1-vm1" && is_chassis_resident("cr-S1-vm1") && ((ip4.dst == >>>>>>> 192.168.0.0/24) || (ip6.dst == fd00::/64))), action=(drop;) >>>>>>> + table=??(lr_in_arp_resolve ), priority=500 , match=(ip4.mcast >>>>>>> || ip6.mcast), action=(next;) >>>>>>> +]) >>>>>>> + >>>>>>> +# Remove lrp from chassis >>>>>>> +check ovn-nbctl --wait=sb lrp-del-gateway-chassis S1-vm1 chassis-1 >>>>>>> + >>>>>>> +ovn-sbctl dump-flows S1 > S1flows >>>>>>> +AT_CAPTURE_FILE([S1flows]) >>>>>>> + >>>>>>> +AT_CHECK([grep -e "lr_in_arp_resolve" S1flows | ovn_strip_lflows], >>>>>>> [0], [dnl >>>>>>> + table=??(lr_in_arp_resolve ), priority=0 , match=(1), >>>>>>> action=(drop;) >>>>>>> + table=??(lr_in_arp_resolve ), priority=1 , match=(reg9[[9]] >>>>>>> == 0), action=(get_nd(outport, xxreg0); next;) >>>>>>> + table=??(lr_in_arp_resolve ), priority=1 , match=(reg9[[9]] >>>>>>> == 1), action=(get_arp(outport, reg0); next;) >>>>>>> + table=??(lr_in_arp_resolve ), priority=500 , match=(ip4.mcast >>>>>>> || ip6.mcast), action=(next;) >>>>>>> +]) >>>>>>> + >>>>>>> +check ovn-nbctl --wait=sb set logical_router_port S1-vm1 >>>>>>> options:disable_arp_resolve=false >>>>>>> +check ovn-nbctl --wait=sb lrp-set-gateway-chassis S1-vm1 chassis-1 >>>>>>> + >>>>>>> +ovn-sbctl dump-flows S1 > S1flows >>>>>>> +AT_CAPTURE_FILE([S1flows]) >>>>>>> + >>>>>>> +AT_CHECK([grep -e "lr_in_arp_resolve" S1flows | ovn_strip_lflows], >>>>>>> [0], [dnl >>>>>>> + table=??(lr_in_arp_resolve ), priority=0 , match=(1), >>>>>>> action=(drop;) >>>>>>> + table=??(lr_in_arp_resolve ), priority=1 , match=(reg9[[9]] >>>>>>> == 0), action=(get_nd(outport, xxreg0); next;) >>>>>>> + table=??(lr_in_arp_resolve ), priority=1 , match=(reg9[[9]] >>>>>>> == 1), action=(get_arp(outport, reg0); next;) >>>>>>> + table=??(lr_in_arp_resolve ), priority=50 , match=(inport == >>>>>>> "S1-vm1" && is_chassis_resident("cr-S1-vm1") && ((ip6.dst == >>>>>>> fd00::/64))), >>>>>>> action=(drop;) >>>>>>> + table=??(lr_in_arp_resolve ), priority=500 , match=(ip4.mcast >>>>>>> || ip6.mcast), action=(next;) >>>>>>> +]) >>>>>>> + >>>>>>> + >>>>>>> +AT_CLEANUP >>>>>>> +]) >>>>>>> + >>>>>>> + >>>>>>> OVN_FOR_EACH_NORTHD_NO_HV([ >>>>>>> AT_SETUP([Address set incremental processing]) >>>>>>> ovn_start >>>>>>> -- >>>>>>> 2.43.0 >>>>>>> >>>>>>> >>>>>>> -- >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> _'Esta mensagem é direcionada apenas para os endereços constantes no >>>>>>> cabeçalho inicial. Se você não está listado nos endereços constantes >>>>>>> no >>>>>>> cabeçalho, pedimos-lhe que desconsidere completamente o conteúdo >>>>>>> dessa >>>>>>> mensagem e cuja cópia, encaminhamento e/ou execução das ações >>>>>>> citadas estão >>>>>>> imediatamente anuladas e proibidas'._ >>>>>>> >>>>>>> >>>>>>> * **'Apesar do Magazine Luiza tomar >>>>>>> todas as precauções razoáveis para assegurar que nenhum vírus esteja >>>>>>> presente nesse e-mail, a empresa não poderá aceitar a >>>>>>> responsabilidade por >>>>>>> quaisquer perdas ou danos causados por esse e-mail ou por seus >>>>>>> anexos'.* >>>>>>> >>>>>>> >>>>>>> >>>>>>> _______________________________________________ >>>>>>> dev mailing list >>>>>>> [email protected] >>>>>>> https://mail.openvswitch.org/mailman/listinfo/ovs-dev >>>>>>> >>>>>>> >>>>> >>>>> *‘Esta mensagem é direcionada apenas para os endereços constantes no >>>>> cabeçalho inicial. Se você não está listado nos endereços constantes no >>>>> cabeçalho, pedimos-lhe que desconsidere completamente o conteúdo dessa >>>>> mensagem e cuja cópia, encaminhamento e/ou execução das ações citadas >>>>> estão >>>>> imediatamente anuladas e proibidas’.* >>>>> >>>>> *‘Apesar do Magazine Luiza tomar todas as precauções razoáveis para >>>>> assegurar que nenhum vírus esteja presente nesse e-mail, a empresa não >>>>> poderá aceitar a responsabilidade por quaisquer perdas ou danos causados >>>>> por esse e-mail ou por seus anexos’.* >>>>> >>>> >>> >>> *‘Esta mensagem é direcionada apenas para os endereços constantes no >>> cabeçalho inicial. Se você não está listado nos endereços constantes no >>> cabeçalho, pedimos-lhe que desconsidere completamente o conteúdo dessa >>> mensagem e cuja cópia, encaminhamento e/ou execução das ações citadas estão >>> imediatamente anuladas e proibidas’.* >>> >>> *‘Apesar do Magazine Luiza tomar todas as precauções razoáveis para >>> assegurar que nenhum vírus esteja presente nesse e-mail, a empresa não >>> poderá aceitar a responsabilidade por quaisquer perdas ou danos causados >>> por esse e-mail ou por seus anexos’.* >>> >> > > *‘Esta mensagem é direcionada apenas para os endereços constantes no > cabeçalho inicial. Se você não está listado nos endereços constantes no > cabeçalho, pedimos-lhe que desconsidere completamente o conteúdo dessa > mensagem e cuja cópia, encaminhamento e/ou execução das ações citadas estão > imediatamente anuladas e proibidas’.* > > *‘Apesar do Magazine Luiza tomar todas as precauções razoáveis para > assegurar que nenhum vírus esteja presente nesse e-mail, a empresa não > poderá aceitar a responsabilidade por quaisquer perdas ou danos causados > por esse e-mail ou por seus anexos’.* > _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
