| Message ID | 20240809024711.2169286-1-martin.kalcok@canonical.com |
|---|---|
| State | Superseded |
| Headers | show |
| Series | [ovs-dev,v5] northd: Routing protocol port redirection. | expand |
| Context | Check | Description |
|---|---|---|
| ovsrobot/apply-robot | fail | apply and check: fail |
On 8/9/24 04:45, Martin Kalcok wrote: > This change adds two new LRP options: > * routing-protocol-redirect > * routing-protocols > > These allow redirection of a routing protocol traffic to > an Logical Switch Port. This enables external routing daemons > to listen on an interface bound to an LSP and effectively act > as if they were listening on (and speaking from) LRP's IP address. > > Option 'routing-protocols' takes a comma-separated list of routing > protocols whose traffic should be redirected. Currently supported > are BGP (tcp 179) and BFD (udp 3784). > > Option 'routing-protocol-redirect' expects a string with an LSP > name. > > When both of these options are set, any traffic entering LS > that's destined for LRP's IP addresses (including IPv6 LLA) and > routing protocol's port number, is redirected to the LSP specified > in the 'routing-protocol-redirect' value. > > NOTE: this feature is experimental and may be subject to > removal/change in the future. > > Signed-off-by: Martin Kalcok <martin.kalcok@canonical.com> > --- Hi Martin, Thanks for v5! Unfortunately it needs a rebase. Also there were some minor things that need to be fixed in the man pages. While at it I have a few other small comments. > > v5 includes: > * change from pure BGP redirect to a configurable protocol redirection > * fixes issue with local routing daemon not being able to receive reply > traffic when contacting its peer. > * ARP and ND traffic is cloned to the redirected port, to properly populate > information about its neighbors. > > > northd/northd.c | 217 ++++++++++++++++++++++++++++++++++++++++ > northd/northd.h | 7 ++ > northd/ovn-northd.8.xml | 54 ++++++++++ > ovn-nb.xml | 42 ++++++++ > tests/ovn-northd.at | 93 +++++++++++++++++ > tests/system-ovn.at | 100 ++++++++++++++++++ > 6 files changed, 513 insertions(+) > > diff --git a/northd/northd.c b/northd/northd.c > index 4353df07d..39b1998fd 100644 > --- a/northd/northd.c > +++ b/northd/northd.c > @@ -13048,6 +13048,221 @@ build_arp_resolve_flows_for_lrp(struct ovn_port *op, > } > } > > +static void > +build_r_p_redirect_rule__( I think I prefer a more explicit name, e.g., build_routing_protocols_redirect_rule__(). > + const char *s_addr, const char *redirect_port_name, int protocol_port, > + const char *proto, bool is_ipv6, struct ovn_port *ls_peer, > + struct lflow_table *lflows, struct ds *match, struct ds *actions) > +{ > + int ip_ver = is_ipv6 ? 6 : 4; > + ds_clear(actions); > + ds_put_format(actions, "outport = \"%s\"; output;", redirect_port_name); > + > + /* Redirect packets in the input pipeline destined for LR's IP > + * and the routing protocol's port to the LSP specified in > + * 'routing-protocol-redirect' option.*/ > + ds_clear(match); > + ds_put_format(match, "ip%d.dst == %s && %s.dst == %d", ip_ver, s_addr, > + proto, protocol_port); > + ovn_lflow_add(lflows, ls_peer->od, S_SWITCH_IN_L2_LKUP, 100, > + ds_cstr(match), > + ds_cstr(actions), > + ls_peer->lflow_ref); > + > + /* To accomodate "peer" nature of the routing daemons, redirect also > + * replies to the daemons' client requests. */ > + ds_clear(match); > + ds_put_format(match, "ip%d.dst == %s && %s.src == %d", ip_ver, s_addr, > + proto, protocol_port); > + ovn_lflow_add(lflows, ls_peer->od, S_SWITCH_IN_L2_LKUP, 100, > + ds_cstr(match), > + ds_cstr(actions), > + ls_peer->lflow_ref); > +} > + > +static void > +apply_r_p_redirect__( Same here, I think I prefer apply_routing_protocols_redirect__(). > + const char *s_addr, const char *redirect_port_name, int protocol_flags, > + bool is_ipv6, struct ovn_port *ls_peer, struct lflow_table *lflows, > + struct ds *match, struct ds *actions) > +{ > + if (protocol_flags & REDIRECT_BGP) { > + build_r_p_redirect_rule__(s_addr, redirect_port_name, 179, "tcp", > + is_ipv6, ls_peer, lflows, match, actions); > + } > + > + if (protocol_flags & REDIRECT_BFD) { > + build_r_p_redirect_rule__(s_addr, redirect_port_name, 3784, "udp", > + is_ipv6, ls_peer, lflows, match, actions); > + } > + > + /* Because the redirected port shares IP and MAC addresses with the LRP, > + * special consideration needs to be given to the signaling protocols. */ > + if (is_ipv6) { > + /* Ensure that redirect port receives copy of NA messages destined to > + * its IP.*/ > + ds_clear(match); > + ds_clear(actions); > + ds_put_format(actions, "clone { outport = \"%s\"; output; };", > + redirect_port_name); > + ds_put_format(match, "ip6.dst == %s && nd_na", s_addr); > + ovn_lflow_add(lflows, ls_peer->od, S_SWITCH_IN_L2_LKUP, 100, > + ds_cstr(match), > + ds_cstr(actions), > + ls_peer->lflow_ref); > + } else { > + /* Ensure that redirect port receives copy of ARP replies destined to > + * its IP */ > + ds_clear(match); > + ds_clear(actions); > + ds_put_format(actions, "clone { outport = \"%s\"; output; };", > + redirect_port_name); > + ds_put_format(match, "arp.op == 2 && arp.tpa == %s", s_addr); > + ovn_lflow_add(lflows, ls_peer->od, S_SWITCH_IN_L2_LKUP, 100, > + ds_cstr(match), > + ds_cstr(actions), > + ls_peer->lflow_ref); > + Nit: no need for newline. > + } > +} > + > +static int > +parse_redirected_routing_protocols(struct ovn_port *lrp) { > + int redirected_protocol_flags = 0; > + const char *redirect_protocols = smap_get(&lrp->nbrp->options, > + "routing-protocols"); > + if (redirect_protocols == NULL) { > + return redirected_protocol_flags; > + } > + > + char *proto; > + char *save_ptr = NULL; > + char *tokstr = xstrdup(redirect_protocols); > + for (proto = strtok_r(tokstr, ",", &save_ptr); proto != NULL; > + proto = strtok_r(NULL, ",", &save_ptr)) { > + if (!strcmp(proto, "BGP")) { > + redirected_protocol_flags |= REDIRECT_BGP; > + continue; > + } > + > + if (!strcmp(proto, "BFD")) { > + redirected_protocol_flags |= REDIRECT_BFD; > + continue; > + } Nit: missing newlines. > + static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, 5); > + VLOG_WARN_RL(&rl, "Option 'routing-protocols' encountered unknown " > + "value %s", > + proto); > + } > + free(tokstr); > + return redirected_protocol_flags; > +} > + > +static void > +build_lrouter_routing_protocol_redirect( > + struct ovn_port *op, struct lflow_table *lflows, > + struct ds *match, struct ds *actions) > +{ > + /* LRP has to have a peer.*/ > + if (op->peer == NULL) { > + return; > + } Nit: missing newline. > + /* LRP has to have NB record.*/ > + if (op->nbrp == NULL) { > + return; > + } > + > + /* Proceed only for LRPs that have 'routing-protocol-redirect' option set. > + * Value of this option is the name of LSP to which the routing protocol > + * traffic will be redirected. */ > + const char *redirect_port = smap_get(&op->nbrp->options, > + "routing-protocol-redirect"); > + if (redirect_port == NULL) { > + return; > + } > + > + if (op->cr_port != NULL) { > + static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, 5); > + VLOG_WARN_RL(&rl, "Option 'routing-protocol-redirect' is not " > + "supported on Distributed Gateway Port '%s'", > + op->key); > + return; > + } > + > + /* Ensure that LSP, to which the routing protocol traffic is redirected, > + * exists.*/ Nit: missing space after '.' > + struct ovn_port *peer_lsp; > + bool redirect_port_exists = false; > + HMAP_FOR_EACH (peer_lsp, dp_node, &op->peer->od->ports) { > + size_t peer_lsp_s = strlen(peer_lsp->key); > + if (peer_lsp_s == strlen(redirect_port) > + && !strncmp(peer_lsp->key, redirect_port, peer_lsp_s)){ > + redirect_port_exists = true; > + break; > + } > + } > + > + if (!redirect_port_exists) { > + static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, 5); > + VLOG_WARN_RL(&rl, "Option 'routing-protocol-redirect' set on Logical " > + "Router Port '%s' refers to non-existent Logical " > + "Switch Port. Routing protocol redirecting won't be " > + "configured.", > + op->key); > + return; > + } > + > + Nit: too many new lines. > + int redirected_protocols = parse_redirected_routing_protocols(op); > + if (!redirected_protocols) { > + static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, 5); > + VLOG_WARN_RL(&rl, "Option 'routing-protocol-redirect' is set on " > + "Logical Router Port '%s' but no known protocols " > + "were set via 'routing-protocols' options. This " > + "configuration has no effect.", > + op->key); > + return; > + } > + > + /* Redirecte traffic destined for LRP's IPs and the specified routing Typo: redirecte > + * protocol ports to the port defined in 'routing-protocol-redirect' > + * option.*/ > + for (size_t i = 0; i < op->lrp_networks.n_ipv4_addrs; i++) { > + const char *ip_s = op->lrp_networks.ipv4_addrs[i].addr_s; > + apply_r_p_redirect__(ip_s, redirect_port, redirected_protocols, false, > + op->peer, lflows,match, actions); > + } > + for (size_t i = 0; i < op->lrp_networks.n_ipv6_addrs; i++) { > + const char *ip_s = op->lrp_networks.ipv6_addrs[i].addr_s; > + apply_r_p_redirect__(ip_s, redirect_port, redirected_protocols, true, > + op->peer, lflows,match, actions); > + } > + > + /* Drop ARP replies and IPv6 RA/NA packets originating from > + * 'routing-protocol-redirect' LSP. As this port shares IP and MAC > + * addresses with LRP, we don't want to create duplicates.*/ > + ds_clear(match); > + ds_put_format(match, "inport == \"%s\" && arp.op == 2", redirect_port); > + ovn_lflow_add(lflows, op->peer->od, S_SWITCH_IN_CHECK_PORT_SEC, 80, > + ds_cstr(match), > + REGBIT_PORT_SEC_DROP " = 1; next;", > + op->peer->lflow_ref); > + > + ds_clear(match); > + ds_put_format(match, "inport == \"%s\" && nd_na", redirect_port); > + ovn_lflow_add(lflows, op->peer->od, S_SWITCH_IN_CHECK_PORT_SEC, 80, > + ds_cstr(match), > + REGBIT_PORT_SEC_DROP " = 1; next;", > + op->peer->lflow_ref); > + > + ds_clear(match); > + ds_put_format(match, "inport == \"%s\" && nd_ra", redirect_port); > + ovn_lflow_add(lflows, op->peer->od, S_SWITCH_IN_CHECK_PORT_SEC, 80, > + ds_cstr(match), > + REGBIT_PORT_SEC_DROP " = 1; next;", > + op->peer->lflow_ref); > +} > + > /* This function adds ARP resolve flows related to a LSP. */ > static void > build_arp_resolve_flows_for_lsp( > @@ -16003,6 +16218,8 @@ build_lswitch_and_lrouter_iterate_by_lrp(struct ovn_port *op, > lsi->meter_groups, op->lflow_ref); > build_lrouter_icmp_packet_toobig_admin_flows(op, lsi->lflows, &lsi->match, > &lsi->actions, op->lflow_ref); > + build_lrouter_routing_protocol_redirect(op, lsi->lflows, > + &lsi->match, &lsi->actions); > } > > static void * > diff --git a/northd/northd.h b/northd/northd.h > index d4a8d75ab..691c23f01 100644 > --- a/northd/northd.h > +++ b/northd/northd.h > @@ -89,6 +89,13 @@ ods_size(const struct ovn_datapaths *datapaths) > > bool od_has_lb_vip(const struct ovn_datapath *od); > > +/* List of routing and routing-related protocols which > + * OVN is capable of redirecting from LRP to specific LSP. */ > +enum redirected_routing_protcol_flag_type { > + REDIRECT_BGP = (1 << 0), > + REDIRECT_BFD = (1 << 1), > +}; > + > struct tracked_ovn_ports { > /* tracked created ports. > * hmapx node data is 'struct ovn_port *' */ > diff --git a/northd/ovn-northd.8.xml b/northd/ovn-northd.8.xml > index b06b09ac5..250c602f2 100644 > --- a/northd/ovn-northd.8.xml > +++ b/northd/ovn-northd.8.xml > @@ -284,6 +284,32 @@ > dropped in the next stage. > </li> > > + <li> > + <p> > + For each logical port that's defined as a target of routing protocol > + redirecting (via <code>routing-protocol-redirect</code> option set on > + Logical Router Port), a filter is set in place that disallows > + following traffic exiting this port: > + </p> > + <ul> > + <li> > + ARP replies > + </li> > + <li> > + IPv6 Neighbor Discovery - Router Advertisements > + </li> > + <li> > + IPv6 Neighbor Discovery - Neighbor Advertisements > + </li> > + </ul> > + <p> > + Since this port shares IP and MAC addresses with the Logical Router > + Port, we wan't to prevent duplicate replies and advertisements. This > + is achieved by a rule with priority 80 that sets > + <code>REGBIT_PORT_SEC_DROP" = 1; next;"</code>. > + </p> > + </li> > + > <li> > For each (enabled) vtep logical port, a priority 70 flow is added which > matches on all packets and applies the action > @@ -1949,6 +1975,34 @@ output; > on the logical switch. > </li> > > + <li> > + <p> > + For any logical port that's defined as a target of routing protocol > + redirecting (via <code>routing-protocol-redirect</code> option set on > + Logical Router Port), we redirect the traffic related to protocols > + specified in <code>routing-protocols</code> option. It's acoomplished > + with following priority-100 flows: > + <p> This should be </p>. > + <ul> > + <li> > + Flows that match Logical Router Port's IPs and destination port of > + the routing daemon are redirected to this port to allow external > + peers' connection to the daemon listening on this port. > + </li> > + <li> > + Flows that match Logical Router Port's IPs and source port of > + the routing daemon are redirected to this port to allow replies > + from the peers. > + </li> > + </ul> > + <p> > + In addition to this, we add priority-100 rules that > + <code>clone</code> ARP replies and IPv6 Neighbor Advertisements to > + this port as well. These allow to build proper ARP/IPv6 neighbor > + list on this port. > + </p> > + </li> > + > <li> > Priority-90 flows for transit switches that forward registered > IP multicast traffic to their corresponding multicast group , which > diff --git a/ovn-nb.xml b/ovn-nb.xml > index 9552534f6..092195473 100644 > --- a/ovn-nb.xml > +++ b/ovn-nb.xml > @@ -3440,6 +3440,48 @@ or > </p> > </column> > > + <column name="options" key="routing-protocol-redirect" > + type='{"type": "string"}'> > + <p> > + NOTE: this feature is experimental and may be subject to > + removal/change in the future. > + </p> > + <p> > + This option expects a name of a Logical Switch Port that's present > + in the peer's Logical Switch. If set, it causes any traffic > + that's destined for Logical Router Port's IP addresses (including > + its IPv6 LLA) and the ports associated with routing protocols defined > + ip <code>routing-protocols</code> option, to be redirected > + to the specified Logical Switch Port. > + > + This allows external routing daemons to be bound to a port in OVN's > + Logical Switch and act as if they were listening on Logical Router > + Port's IP addresses. > + </p> > + </column> > + > + <column name="options" key="routing-protocols" type='{"type": "string"}'> > + <p> > + NOTE: this feature is experimental and may be subject to > + removal/change in the future. > + </p> > + <p> > + This option expects a comma-separated list of routing, and > + routing-related protocols, whose control plane traffic will be > + redirected to a port specified in > + <code>routing-protocol-redirect</code> option. Currently supported > + options are: > + </p> > + <ul> > + <li> > + <code>BGP</code> (forwards TCP port 179) > + </li> > + <li> > + <code>BFD</code> (forwards UDP port 3748) > + </li> > + <ul> This should be </ul>. The rest looks OK to me. You can find the rebased version, including these minor nits fixed here: https://github.com/dceara/ovn/commits/review-pws418655-bgp-redirect-v5/ If the small changes I made look OK to you feel free to squash them in and post v6 with my: Acked-by: Dumitru Ceara <dceara@redhat.com> Thanks, Dumitru
Hi Dumitru, thank you for the fast review. I'll post v6 momentarily, I just want to do quick fresh end-to-end setup/test. I also want to add that you were correct about the ARP/ND. It didn't work properly and it was living off the entries populated by unicast traffic from the peer, however the entries were flapping wildly and it was discovered when I threw the BFD into the mix (real blessing in disguise). The slower BFD protocol was able to deal with it without disconnects, but BFD was disconnecting every once in a while. Hence the `clone`rules for the ARP replies and IPv6 NAs. On Fri, 2024-08-09 at 08:19 +0200, Dumitru Ceara wrote: > On 8/9/24 04:45, Martin Kalcok wrote: > > This change adds two new LRP options: > > * routing-protocol-redirect > > * routing-protocols > > > > These allow redirection of a routing protocol traffic to > > an Logical Switch Port. This enables external routing daemons > > to listen on an interface bound to an LSP and effectively act > > as if they were listening on (and speaking from) LRP's IP address. > > > > Option 'routing-protocols' takes a comma-separated list of routing > > protocols whose traffic should be redirected. Currently supported > > are BGP (tcp 179) and BFD (udp 3784). > > > > Option 'routing-protocol-redirect' expects a string with an LSP > > name. > > > > When both of these options are set, any traffic entering LS > > that's destined for LRP's IP addresses (including IPv6 LLA) and > > routing protocol's port number, is redirected to the LSP specified > > in the 'routing-protocol-redirect' value. > > > > NOTE: this feature is experimental and may be subject to > > removal/change in the future. > > > > Signed-off-by: Martin Kalcok <martin.kalcok@canonical.com> > > --- > > Hi Martin, > > Thanks for v5! Unfortunately it needs a rebase. > > Also there were some minor things that need to be fixed in the man > pages. While at it I have a few other small comments. > > > > > v5 includes: > > * change from pure BGP redirect to a configurable protocol > > redirection > > * fixes issue with local routing daemon not being able to receive > > reply > > traffic when contacting its peer. > > * ARP and ND traffic is cloned to the redirected port, to > > properly populate > > information about its neighbors. > > > > > > northd/northd.c | 217 > > ++++++++++++++++++++++++++++++++++++++++ > > northd/northd.h | 7 ++ > > northd/ovn-northd.8.xml | 54 ++++++++++ > > ovn-nb.xml | 42 ++++++++ > > tests/ovn-northd.at | 93 +++++++++++++++++ > > tests/system-ovn.at | 100 ++++++++++++++++++ > > 6 files changed, 513 insertions(+) > > > > diff --git a/northd/northd.c b/northd/northd.c > > index 4353df07d..39b1998fd 100644 > > --- a/northd/northd.c > > +++ b/northd/northd.c > > @@ -13048,6 +13048,221 @@ build_arp_resolve_flows_for_lrp(struct > > ovn_port *op, > > } > > } > > > > +static void > > +build_r_p_redirect_rule__( > > I think I prefer a more explicit name, e.g., > build_routing_protocols_redirect_rule__(). > > > + const char *s_addr, const char *redirect_port_name, int > > protocol_port, > > + const char *proto, bool is_ipv6, struct ovn_port *ls_peer, > > + struct lflow_table *lflows, struct ds *match, struct ds > > *actions) > > +{ > > + int ip_ver = is_ipv6 ? 6 : 4; > > + ds_clear(actions); > > + ds_put_format(actions, "outport = \"%s\"; output;", > > redirect_port_name); > > + > > + /* Redirect packets in the input pipeline destined for LR's IP > > + * and the routing protocol's port to the LSP specified in > > + * 'routing-protocol-redirect' option.*/ > > + ds_clear(match); > > + ds_put_format(match, "ip%d.dst == %s && %s.dst == %d", ip_ver, > > s_addr, > > + proto, protocol_port); > > + ovn_lflow_add(lflows, ls_peer->od, S_SWITCH_IN_L2_LKUP, 100, > > + ds_cstr(match), > > + ds_cstr(actions), > > + ls_peer->lflow_ref); > > + > > + /* To accomodate "peer" nature of the routing daemons, > > redirect also > > + * replies to the daemons' client requests. */ > > + ds_clear(match); > > + ds_put_format(match, "ip%d.dst == %s && %s.src == %d", ip_ver, > > s_addr, > > + proto, protocol_port); > > + ovn_lflow_add(lflows, ls_peer->od, S_SWITCH_IN_L2_LKUP, 100, > > + ds_cstr(match), > > + ds_cstr(actions), > > + ls_peer->lflow_ref); > > +} > > + > > +static void > > +apply_r_p_redirect__( > > Same here, I think I prefer apply_routing_protocols_redirect__(). > > > + const char *s_addr, const char *redirect_port_name, int > > protocol_flags, > > + bool is_ipv6, struct ovn_port *ls_peer, struct lflow_table > > *lflows, > > + struct ds *match, struct ds *actions) > > +{ > > + if (protocol_flags & REDIRECT_BGP) { > > + build_r_p_redirect_rule__(s_addr, redirect_port_name, 179, > > "tcp", > > + is_ipv6, ls_peer, lflows, match, > > actions); > > + } > > + > > + if (protocol_flags & REDIRECT_BFD) { > > + build_r_p_redirect_rule__(s_addr, redirect_port_name, > > 3784, "udp", > > + is_ipv6, ls_peer, lflows, match, > > actions); > > + } > > + > > + /* Because the redirected port shares IP and MAC addresses > > with the LRP, > > + * special consideration needs to be given to the signaling > > protocols. */ > > + if (is_ipv6) { > > + /* Ensure that redirect port receives copy of NA messages > > destined to > > + * its IP.*/ > > + ds_clear(match); > > + ds_clear(actions); > > + ds_put_format(actions, "clone { outport = \"%s\"; output; > > };", > > + redirect_port_name); > > + ds_put_format(match, "ip6.dst == %s && nd_na", s_addr); > > + ovn_lflow_add(lflows, ls_peer->od, S_SWITCH_IN_L2_LKUP, > > 100, > > + ds_cstr(match), > > + ds_cstr(actions), > > + ls_peer->lflow_ref); > > + } else { > > + /* Ensure that redirect port receives copy of ARP replies > > destined to > > + * its IP */ > > + ds_clear(match); > > + ds_clear(actions); > > + ds_put_format(actions, "clone { outport = \"%s\"; output; > > };", > > + redirect_port_name); > > + ds_put_format(match, "arp.op == 2 && arp.tpa == %s", > > s_addr); > > + ovn_lflow_add(lflows, ls_peer->od, S_SWITCH_IN_L2_LKUP, > > 100, > > + ds_cstr(match), > > + ds_cstr(actions), > > + ls_peer->lflow_ref); > > + > > Nit: no need for newline. > > > + } > > +} > > + > > +static int > > +parse_redirected_routing_protocols(struct ovn_port *lrp) { > > + int redirected_protocol_flags = 0; > > + const char *redirect_protocols = smap_get(&lrp->nbrp->options, > > + "routing- > > protocols"); > > + if (redirect_protocols == NULL) { > > + return redirected_protocol_flags; > > + } > > + > > + char *proto; > > + char *save_ptr = NULL; > > + char *tokstr = xstrdup(redirect_protocols); > > + for (proto = strtok_r(tokstr, ",", &save_ptr); proto != NULL; > > + proto = strtok_r(NULL, ",", &save_ptr)) { > > + if (!strcmp(proto, "BGP")) { > > + redirected_protocol_flags |= REDIRECT_BGP; > > + continue; > > + } > > + > > + if (!strcmp(proto, "BFD")) { > > + redirected_protocol_flags |= REDIRECT_BFD; > > + continue; > > + } > > Nit: missing newlines. > > > + static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, > > 5); > > + VLOG_WARN_RL(&rl, "Option 'routing-protocols' encountered > > unknown " > > + "value %s", > > + proto); > > + } > > + free(tokstr); > > + return redirected_protocol_flags; > > +} > > + > > +static void > > +build_lrouter_routing_protocol_redirect( > > + struct ovn_port *op, struct lflow_table *lflows, > > + struct ds *match, struct ds *actions) > > +{ > > + /* LRP has to have a peer.*/ > > + if (op->peer == NULL) { > > + return; > > + } > > Nit: missing newline. > > > + /* LRP has to have NB record.*/ > > + if (op->nbrp == NULL) { > > + return; > > + } > > + > > + /* Proceed only for LRPs that have 'routing-protocol-redirect' > > option set. > > + * Value of this option is the name of LSP to which the > > routing protocol > > + * traffic will be redirected. */ > > + const char *redirect_port = smap_get(&op->nbrp->options, > > + "routing-protocol- > > redirect"); > > + if (redirect_port == NULL) { > > + return; > > + } > > + > > + if (op->cr_port != NULL) { > > + static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, > > 5); > > + VLOG_WARN_RL(&rl, "Option 'routing-protocol-redirect' is > > not " > > + "supported on Distributed Gateway Port > > '%s'", > > + op->key); > > + return; > > + } > > + > > + /* Ensure that LSP, to which the routing protocol traffic is > > redirected, > > + * exists.*/ > > Nit: missing space after '.' > > > + struct ovn_port *peer_lsp; > > + bool redirect_port_exists = false; > > + HMAP_FOR_EACH (peer_lsp, dp_node, &op->peer->od->ports) { > > + size_t peer_lsp_s = strlen(peer_lsp->key); > > + if (peer_lsp_s == strlen(redirect_port) > > + && !strncmp(peer_lsp->key, redirect_port, > > peer_lsp_s)){ > > + redirect_port_exists = true; > > + break; > > + } > > + } > > + > > + if (!redirect_port_exists) { > > + static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, > > 5); > > + VLOG_WARN_RL(&rl, "Option 'routing-protocol-redirect' set > > on Logical " > > + "Router Port '%s' refers to non-existent > > Logical " > > + "Switch Port. Routing protocol > > redirecting won't be " > > + "configured.", > > + op->key); > > + return; > > + } > > + > > + > > Nit: too many new lines. > > > + int redirected_protocols = > > parse_redirected_routing_protocols(op); > > + if (!redirected_protocols) { > > + static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, > > 5); > > + VLOG_WARN_RL(&rl, "Option 'routing-protocol-redirect' is > > set on " > > + "Logical Router Port '%s' but no known > > protocols " > > + "were set via 'routing-protocols' > > options. This " > > + "configuration has no effect.", > > + op->key); > > + return; > > + } > > + > > + /* Redirecte traffic destined for LRP's IPs and the specified > > routing > > Typo: redirecte > > > + * protocol ports to the port defined in 'routing-protocol- > > redirect' > > + * option.*/ > > + for (size_t i = 0; i < op->lrp_networks.n_ipv4_addrs; i++) { > > + const char *ip_s = op->lrp_networks.ipv4_addrs[i].addr_s; > > + apply_r_p_redirect__(ip_s, redirect_port, > > redirected_protocols, false, > > + op->peer, lflows,match, actions); > > + } > > + for (size_t i = 0; i < op->lrp_networks.n_ipv6_addrs; i++) { > > + const char *ip_s = op->lrp_networks.ipv6_addrs[i].addr_s; > > + apply_r_p_redirect__(ip_s, redirect_port, > > redirected_protocols, true, > > + op->peer, lflows,match, > > actions); > > + } > > + > > + /* Drop ARP replies and IPv6 RA/NA packets originating from > > + * 'routing-protocol-redirect' LSP. As this port shares IP and > > MAC > > + * addresses with LRP, we don't want to create duplicates.*/ > > + ds_clear(match); > > + ds_put_format(match, "inport == \"%s\" && arp.op == 2", > > redirect_port); > > + ovn_lflow_add(lflows, op->peer->od, > > S_SWITCH_IN_CHECK_PORT_SEC, 80, > > + ds_cstr(match), > > + REGBIT_PORT_SEC_DROP " = 1; next;", > > + op->peer->lflow_ref); > > + > > + ds_clear(match); > > + ds_put_format(match, "inport == \"%s\" && nd_na", > > redirect_port); > > + ovn_lflow_add(lflows, op->peer->od, > > S_SWITCH_IN_CHECK_PORT_SEC, 80, > > + ds_cstr(match), > > + REGBIT_PORT_SEC_DROP " = 1; next;", > > + op->peer->lflow_ref); > > + > > + ds_clear(match); > > + ds_put_format(match, "inport == \"%s\" && nd_ra", > > redirect_port); > > + ovn_lflow_add(lflows, op->peer->od, > > S_SWITCH_IN_CHECK_PORT_SEC, 80, > > + ds_cstr(match), > > + REGBIT_PORT_SEC_DROP " = 1; next;", > > + op->peer->lflow_ref); > > +} > > + > > /* This function adds ARP resolve flows related to a LSP. */ > > static void > > build_arp_resolve_flows_for_lsp( > > @@ -16003,6 +16218,8 @@ > > build_lswitch_and_lrouter_iterate_by_lrp(struct ovn_port *op, > > lsi->meter_groups, op->lflow_ref); > > build_lrouter_icmp_packet_toobig_admin_flows(op, lsi->lflows, > > &lsi->match, > > &lsi->actions, > > op->lflow_ref); > > + build_lrouter_routing_protocol_redirect(op, lsi->lflows, > > + &lsi->match, &lsi- > > >actions); > > } > > > > static void * > > diff --git a/northd/northd.h b/northd/northd.h > > index d4a8d75ab..691c23f01 100644 > > --- a/northd/northd.h > > +++ b/northd/northd.h > > @@ -89,6 +89,13 @@ ods_size(const struct ovn_datapaths *datapaths) > > > > bool od_has_lb_vip(const struct ovn_datapath *od); > > > > +/* List of routing and routing-related protocols which > > + * OVN is capable of redirecting from LRP to specific LSP. */ > > +enum redirected_routing_protcol_flag_type { > > + REDIRECT_BGP = (1 << 0), > > + REDIRECT_BFD = (1 << 1), > > +}; > > + > > struct tracked_ovn_ports { > > /* tracked created ports. > > * hmapx node data is 'struct ovn_port *' */ > > diff --git a/northd/ovn-northd.8.xml b/northd/ovn-northd.8.xml > > index b06b09ac5..250c602f2 100644 > > --- a/northd/ovn-northd.8.xml > > +++ b/northd/ovn-northd.8.xml > > @@ -284,6 +284,32 @@ > > dropped in the next stage. > > </li> > > > > + <li> > > + <p> > > + For each logical port that's defined as a target of > > routing protocol > > + redirecting (via <code>routing-protocol-redirect</code> > > option set on > > + Logical Router Port), a filter is set in place that > > disallows > > + following traffic exiting this port: > > + </p> > > + <ul> > > + <li> > > + ARP replies > > + </li> > > + <li> > > + IPv6 Neighbor Discovery - Router Advertisements > > + </li> > > + <li> > > + IPv6 Neighbor Discovery - Neighbor Advertisements > > + </li> > > + </ul> > > + <p> > > + Since this port shares IP and MAC addresses with the > > Logical Router > > + Port, we wan't to prevent duplicate replies and > > advertisements. This > > + is achieved by a rule with priority 80 that sets > > + <code>REGBIT_PORT_SEC_DROP" = 1; next;"</code>. > > + </p> > > + </li> > > + > > <li> > > For each (enabled) vtep logical port, a priority 70 flow > > is added which > > matches on all packets and applies the action > > @@ -1949,6 +1975,34 @@ output; > > on the logical switch. > > </li> > > > > + <li> > > + <p> > > + For any logical port that's defined as a target of > > routing protocol > > + redirecting (via <code>routing-protocol-redirect</code> > > option set on > > + Logical Router Port), we redirect the traffic related to > > protocols > > + specified in <code>routing-protocols</code> option. It's > > acoomplished > > + with following priority-100 flows: > > + <p> > > This should be </p>. > > > + <ul> > > + <li> > > + Flows that match Logical Router Port's IPs and > > destination port of > > + the routing daemon are redirected to this port to > > allow external > > + peers' connection to the daemon listening on this > > port. > > + </li> > > + <li> > > + Flows that match Logical Router Port's IPs and source > > port of > > + the routing daemon are redirected to this port to > > allow replies > > + from the peers. > > + </li> > > + </ul> > > + <p> > > + In addition to this, we add priority-100 rules that > > + <code>clone</code> ARP replies and IPv6 Neighbor > > Advertisements to > > + this port as well. These allow to build proper ARP/IPv6 > > neighbor > > + list on this port. > > + </p> > > + </li> > > + > > <li> > > Priority-90 flows for transit switches that forward > > registered > > IP multicast traffic to their corresponding multicast > > group , which > > diff --git a/ovn-nb.xml b/ovn-nb.xml > > index 9552534f6..092195473 100644 > > --- a/ovn-nb.xml > > +++ b/ovn-nb.xml > > @@ -3440,6 +3440,48 @@ or > > </p> > > </column> > > > > + <column name="options" key="routing-protocol-redirect" > > + type='{"type": "string"}'> > > + <p> > > + NOTE: this feature is experimental and may be subject to > > + removal/change in the future. > > + </p> > > + <p> > > + This option expects a name of a Logical Switch Port > > that's present > > + in the peer's Logical Switch. If set, it causes any > > traffic > > + that's destined for Logical Router Port's IP addresses > > (including > > + its IPv6 LLA) and the ports associated with routing > > protocols defined > > + ip <code>routing-protocols</code> option, to be > > redirected > > + to the specified Logical Switch Port. > > + > > + This allows external routing daemons to be bound to a > > port in OVN's > > + Logical Switch and act as if they were listening on > > Logical Router > > + Port's IP addresses. > > + </p> > > + </column> > > + > > + <column name="options" key="routing-protocols" > > type='{"type": "string"}'> > > + <p> > > + NOTE: this feature is experimental and may be subject to > > + removal/change in the future. > > + </p> > > + <p> > > + This option expects a comma-separated list of routing, > > and > > + routing-related protocols, whose control plane traffic > > will be > > + redirected to a port specified in > > + <code>routing-protocol-redirect</code> option. Currently > > supported > > + options are: > > + </p> > > + <ul> > > + <li> > > + <code>BGP</code> (forwards TCP port 179) > > + </li> > > + <li> > > + <code>BFD</code> (forwards UDP port 3748) > > + </li> > > + <ul> > > This should be </ul>. > > The rest looks OK to me. You can find the rebased version, including > these minor nits fixed here: > > https://github.com/dceara/ovn/commits/review-pws418655-bgp-redirect-v5/ > > If the small changes I made look OK to you feel free to squash them > in > and post v6 with my: > > Acked-by: Dumitru Ceara <dceara@redhat.com> > > Thanks, > Dumitru >
Hi Martin, Dumitru, I'm gonna to perform a quick testing after rebase and get back with a feedback. Please see one question below. On 09.08.2024 15:25, martin.kalcok@canonical.com wrote: > Hi Dumitru, > thank you for the fast review. I'll post v6 momentarily, I just want to > do quick fresh end-to-end setup/test. > I also want to add that you were correct about the ARP/ND. It didn't > work properly and it was living off the entries populated by unicast > traffic from the peer, however the entries were flapping wildly and it > was discovered when I threw the BFD into the mix (real blessing in > disguise). The slower BFD protocol was able to deal with it without > disconnects, but BFD was disconnecting every once in a while. Hence the > `clone`rules for the ARP replies and IPv6 NAs. > > > On Fri, 2024-08-09 at 08:19 +0200, Dumitru Ceara wrote: >> On 8/9/24 04:45, Martin Kalcok wrote: >>> This change adds two new LRP options: >>> * routing-protocol-redirect >>> * routing-protocols >>> >>> These allow redirection of a routing protocol traffic to >>> an Logical Switch Port. This enables external routing daemons >>> to listen on an interface bound to an LSP and effectively act >>> as if they were listening on (and speaking from) LRP's IP address. >>> >>> Option 'routing-protocols' takes a comma-separated list of routing >>> protocols whose traffic should be redirected. Currently supported >>> are BGP (tcp 179) and BFD (udp 3784). >>> >>> Option 'routing-protocol-redirect' expects a string with an LSP >>> name. >>> >>> When both of these options are set, any traffic entering LS >>> that's destined for LRP's IP addresses (including IPv6 LLA) and >>> routing protocol's port number, is redirected to the LSP specified >>> in the 'routing-protocol-redirect' value. >>> >>> NOTE: this feature is experimental and may be subject to >>> removal/change in the future. >>> >>> Signed-off-by: Martin Kalcok<martin.kalcok@canonical.com> >>> --- >> Hi Martin, >> >> Thanks for v5! Unfortunately it needs a rebase. >> >> Also there were some minor things that need to be fixed in the man >> pages. While at it I have a few other small comments. >> >>> v5 includes: >>> * change from pure BGP redirect to a configurable protocol >>> redirection >>> * fixes issue with local routing daemon not being able to receive >>> reply >>> traffic when contacting its peer. >>> * ARP and ND traffic is cloned to the redirected port, to >>> properly populate >>> information about its neighbors. >>> >>> >>> northd/northd.c | 217 >>> ++++++++++++++++++++++++++++++++++++++++ >>> northd/northd.h | 7 ++ >>> northd/ovn-northd.8.xml | 54 ++++++++++ >>> ovn-nb.xml | 42 ++++++++ >>> tests/ovn-northd.at | 93 +++++++++++++++++ >>> tests/system-ovn.at | 100 ++++++++++++++++++ >>> 6 files changed, 513 insertions(+) >>> >>> diff --git a/northd/northd.c b/northd/northd.c >>> index 4353df07d..39b1998fd 100644 >>> --- a/northd/northd.c >>> +++ b/northd/northd.c >>> @@ -13048,6 +13048,221 @@ build_arp_resolve_flows_for_lrp(struct >>> ovn_port *op, >>> } >>> } >>> >>> +static void >>> +build_r_p_redirect_rule__( >> I think I prefer a more explicit name, e.g., >> build_routing_protocols_redirect_rule__(). >> >>> + const char *s_addr, const char *redirect_port_name, int >>> protocol_port, >>> + const char *proto, bool is_ipv6, struct ovn_port *ls_peer, >>> + struct lflow_table *lflows, struct ds *match, struct ds >>> *actions) >>> +{ >>> + int ip_ver = is_ipv6 ? 6 : 4; >>> + ds_clear(actions); >>> + ds_put_format(actions, "outport = \"%s\"; output;", >>> redirect_port_name); >>> + >>> + /* Redirect packets in the input pipeline destined for LR's IP >>> + * and the routing protocol's port to the LSP specified in >>> + * 'routing-protocol-redirect' option.*/ >>> + ds_clear(match); >>> + ds_put_format(match, "ip%d.dst == %s && %s.dst == %d", ip_ver, >>> s_addr, >>> + proto, protocol_port); >>> + ovn_lflow_add(lflows, ls_peer->od, S_SWITCH_IN_L2_LKUP, 100, >>> + ds_cstr(match), >>> + ds_cstr(actions), >>> + ls_peer->lflow_ref); >>> + >>> + /* To accomodate "peer" nature of the routing daemons, >>> redirect also >>> + * replies to the daemons' client requests. */ >>> + ds_clear(match); >>> + ds_put_format(match, "ip%d.dst == %s && %s.src == %d", ip_ver, >>> s_addr, >>> + proto, protocol_port); >>> + ovn_lflow_add(lflows, ls_peer->od, S_SWITCH_IN_L2_LKUP, 100, >>> + ds_cstr(match), >>> + ds_cstr(actions), >>> + ls_peer->lflow_ref); >>> +} Don't we want to merge these two conditions into one logical flow? E.g.: "(ip%d.dst == %s && (%s.dst == %d && %s.src == %d)" ? This will make one logical flow per LRP IP per protocol instead of two. >>> + >>> +static void >>> +apply_r_p_redirect__( >> Same here, I think I prefer apply_routing_protocols_redirect__(). >> >>> + const char *s_addr, const char *redirect_port_name, int >>> protocol_flags, >>> + bool is_ipv6, struct ovn_port *ls_peer, struct lflow_table >>> *lflows, >>> + struct ds *match, struct ds *actions) >>> +{ >>> + if (protocol_flags & REDIRECT_BGP) { >>> + build_r_p_redirect_rule__(s_addr, redirect_port_name, 179, >>> "tcp", >>> + is_ipv6, ls_peer, lflows, match, >>> actions); >>> + } >>> + >>> + if (protocol_flags & REDIRECT_BFD) { >>> + build_r_p_redirect_rule__(s_addr, redirect_port_name, >>> 3784, "udp", >>> + is_ipv6, ls_peer, lflows, match, >>> actions); >>> + } >>> + >>> + /* Because the redirected port shares IP and MAC addresses >>> with the LRP, >>> + * special consideration needs to be given to the signaling >>> protocols. */ >>> + if (is_ipv6) { >>> + /* Ensure that redirect port receives copy of NA messages >>> destined to >>> + * its IP.*/ >>> + ds_clear(match); >>> + ds_clear(actions); >>> + ds_put_format(actions, "clone { outport = \"%s\"; output; >>> };", >>> + redirect_port_name); >>> + ds_put_format(match, "ip6.dst == %s && nd_na", s_addr); >>> + ovn_lflow_add(lflows, ls_peer->od, S_SWITCH_IN_L2_LKUP, >>> 100, >>> + ds_cstr(match), >>> + ds_cstr(actions), >>> + ls_peer->lflow_ref); >>> + } else { >>> + /* Ensure that redirect port receives copy of ARP replies >>> destined to >>> + * its IP */ >>> + ds_clear(match); >>> + ds_clear(actions); >>> + ds_put_format(actions, "clone { outport = \"%s\"; output; >>> };", >>> + redirect_port_name); >>> + ds_put_format(match, "arp.op == 2 && arp.tpa == %s", >>> s_addr); >>> + ovn_lflow_add(lflows, ls_peer->od, S_SWITCH_IN_L2_LKUP, >>> 100, >>> + ds_cstr(match), >>> + ds_cstr(actions), >>> + ls_peer->lflow_ref); >>> + >> Nit: no need for newline. >> >>> + } >>> +} >>> + >>> +static int >>> +parse_redirected_routing_protocols(struct ovn_port *lrp) { >>> + int redirected_protocol_flags = 0; >>> + const char *redirect_protocols = smap_get(&lrp->nbrp->options, >>> + "routing- >>> protocols"); >>> + if (redirect_protocols == NULL) { >>> + return redirected_protocol_flags; >>> + } >>> + >>> + char *proto; >>> + char *save_ptr = NULL; >>> + char *tokstr = xstrdup(redirect_protocols); >>> + for (proto = strtok_r(tokstr, ",", &save_ptr); proto != NULL; >>> + proto = strtok_r(NULL, ",", &save_ptr)) { >>> + if (!strcmp(proto, "BGP")) { >>> + redirected_protocol_flags |= REDIRECT_BGP; >>> + continue; >>> + } >>> + >>> + if (!strcmp(proto, "BFD")) { >>> + redirected_protocol_flags |= REDIRECT_BFD; >>> + continue; >>> + } >> Nit: missing newlines. >> >>> + static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, >>> 5); >>> + VLOG_WARN_RL(&rl, "Option 'routing-protocols' encountered >>> unknown " >>> + "value %s", >>> + proto); >>> + } >>> + free(tokstr); >>> + return redirected_protocol_flags; >>> +} >>> + >>> +static void >>> +build_lrouter_routing_protocol_redirect( >>> + struct ovn_port *op, struct lflow_table *lflows, >>> + struct ds *match, struct ds *actions) >>> +{ >>> + /* LRP has to have a peer.*/ >>> + if (op->peer == NULL) { >>> + return; >>> + } >> Nit: missing newline. >> >>> + /* LRP has to have NB record.*/ >>> + if (op->nbrp == NULL) { >>> + return; >>> + } >>> + >>> + /* Proceed only for LRPs that have 'routing-protocol-redirect' >>> option set. >>> + * Value of this option is the name of LSP to which the >>> routing protocol >>> + * traffic will be redirected. */ >>> + const char *redirect_port = smap_get(&op->nbrp->options, >>> + "routing-protocol- >>> redirect"); >>> + if (redirect_port == NULL) { >>> + return; >>> + } >>> + >>> + if (op->cr_port != NULL) { >>> + static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, >>> 5); >>> + VLOG_WARN_RL(&rl, "Option 'routing-protocol-redirect' is >>> not " >>> + "supported on Distributed Gateway Port >>> '%s'", >>> + op->key); >>> + return; >>> + } >>> + >>> + /* Ensure that LSP, to which the routing protocol traffic is >>> redirected, >>> + * exists.*/ >> Nit: missing space after '.' >> >>> + struct ovn_port *peer_lsp; >>> + bool redirect_port_exists = false; >>> + HMAP_FOR_EACH (peer_lsp, dp_node, &op->peer->od->ports) { >>> + size_t peer_lsp_s = strlen(peer_lsp->key); >>> + if (peer_lsp_s == strlen(redirect_port) >>> + && !strncmp(peer_lsp->key, redirect_port, >>> peer_lsp_s)){ >>> + redirect_port_exists = true; >>> + break; >>> + } >>> + } >>> + >>> + if (!redirect_port_exists) { >>> + static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, >>> 5); >>> + VLOG_WARN_RL(&rl, "Option 'routing-protocol-redirect' set >>> on Logical " >>> + "Router Port '%s' refers to non-existent >>> Logical " >>> + "Switch Port. Routing protocol >>> redirecting won't be " >>> + "configured.", >>> + op->key); >>> + return; >>> + } >>> + >>> + >> Nit: too many new lines. >> >>> + int redirected_protocols = >>> parse_redirected_routing_protocols(op); >>> + if (!redirected_protocols) { >>> + static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, >>> 5); >>> + VLOG_WARN_RL(&rl, "Option 'routing-protocol-redirect' is >>> set on " >>> + "Logical Router Port '%s' but no known >>> protocols " >>> + "were set via 'routing-protocols' >>> options. This " >>> + "configuration has no effect.", >>> + op->key); >>> + return; >>> + } >>> + >>> + /* Redirecte traffic destined for LRP's IPs and the specified >>> routing >> Typo: redirecte >> >>> + * protocol ports to the port defined in 'routing-protocol- >>> redirect' >>> + * option.*/ >>> + for (size_t i = 0; i < op->lrp_networks.n_ipv4_addrs; i++) { >>> + const char *ip_s = op->lrp_networks.ipv4_addrs[i].addr_s; >>> + apply_r_p_redirect__(ip_s, redirect_port, >>> redirected_protocols, false, >>> + op->peer, lflows,match, actions); >>> + } >>> + for (size_t i = 0; i < op->lrp_networks.n_ipv6_addrs; i++) { >>> + const char *ip_s = op->lrp_networks.ipv6_addrs[i].addr_s; >>> + apply_r_p_redirect__(ip_s, redirect_port, >>> redirected_protocols, true, >>> + op->peer, lflows,match, >>> actions); >>> + } >>> + >>> + /* Drop ARP replies and IPv6 RA/NA packets originating from >>> + * 'routing-protocol-redirect' LSP. As this port shares IP and >>> MAC >>> + * addresses with LRP, we don't want to create duplicates.*/ >>> + ds_clear(match); >>> + ds_put_format(match, "inport == \"%s\" && arp.op == 2", >>> redirect_port); >>> + ovn_lflow_add(lflows, op->peer->od, >>> S_SWITCH_IN_CHECK_PORT_SEC, 80, >>> + ds_cstr(match), >>> + REGBIT_PORT_SEC_DROP " = 1; next;", >>> + op->peer->lflow_ref); >>> + >>> + ds_clear(match); >>> + ds_put_format(match, "inport == \"%s\" && nd_na", >>> redirect_port); >>> + ovn_lflow_add(lflows, op->peer->od, >>> S_SWITCH_IN_CHECK_PORT_SEC, 80, >>> + ds_cstr(match), >>> + REGBIT_PORT_SEC_DROP " = 1; next;", >>> + op->peer->lflow_ref); >>> + >>> + ds_clear(match); >>> + ds_put_format(match, "inport == \"%s\" && nd_ra", >>> redirect_port); >>> + ovn_lflow_add(lflows, op->peer->od, >>> S_SWITCH_IN_CHECK_PORT_SEC, 80, >>> + ds_cstr(match), >>> + REGBIT_PORT_SEC_DROP " = 1; next;", >>> + op->peer->lflow_ref); >>> +} >>> + >>> /* This function adds ARP resolve flows related to a LSP. */ >>> static void >>> build_arp_resolve_flows_for_lsp( >>> @@ -16003,6 +16218,8 @@ >>> build_lswitch_and_lrouter_iterate_by_lrp(struct ovn_port *op, >>> lsi->meter_groups, op->lflow_ref); >>> build_lrouter_icmp_packet_toobig_admin_flows(op, lsi->lflows, >>> &lsi->match, >>> &lsi->actions, >>> op->lflow_ref); >>> + build_lrouter_routing_protocol_redirect(op, lsi->lflows, >>> + &lsi->match, &lsi- >>>> actions); >>> } >>> >>> static void * >>> diff --git a/northd/northd.h b/northd/northd.h >>> index d4a8d75ab..691c23f01 100644 >>> --- a/northd/northd.h >>> +++ b/northd/northd.h >>> @@ -89,6 +89,13 @@ ods_size(const struct ovn_datapaths *datapaths) >>> >>> bool od_has_lb_vip(const struct ovn_datapath *od); >>> >>> +/* List of routing and routing-related protocols which >>> + * OVN is capable of redirecting from LRP to specific LSP. */ >>> +enum redirected_routing_protcol_flag_type { >>> + REDIRECT_BGP = (1 << 0), >>> + REDIRECT_BFD = (1 << 1), >>> +}; >>> + >>> struct tracked_ovn_ports { >>> /* tracked created ports. >>> * hmapx node data is 'struct ovn_port *' */ >>> diff --git a/northd/ovn-northd.8.xml b/northd/ovn-northd.8.xml >>> index b06b09ac5..250c602f2 100644 >>> --- a/northd/ovn-northd.8.xml >>> +++ b/northd/ovn-northd.8.xml >>> @@ -284,6 +284,32 @@ >>> dropped in the next stage. >>> </li> >>> >>> + <li> >>> + <p> >>> + For each logical port that's defined as a target of >>> routing protocol >>> + redirecting (via <code>routing-protocol-redirect</code> >>> option set on >>> + Logical Router Port), a filter is set in place that >>> disallows >>> + following traffic exiting this port: >>> + </p> >>> + <ul> >>> + <li> >>> + ARP replies >>> + </li> >>> + <li> >>> + IPv6 Neighbor Discovery - Router Advertisements >>> + </li> >>> + <li> >>> + IPv6 Neighbor Discovery - Neighbor Advertisements >>> + </li> >>> + </ul> >>> + <p> >>> + Since this port shares IP and MAC addresses with the >>> Logical Router >>> + Port, we wan't to prevent duplicate replies and >>> advertisements. This >>> + is achieved by a rule with priority 80 that sets >>> + <code>REGBIT_PORT_SEC_DROP" = 1; next;"</code>. >>> + </p> >>> + </li> >>> + >>> <li> >>> For each (enabled) vtep logical port, a priority 70 flow >>> is added which >>> matches on all packets and applies the action >>> @@ -1949,6 +1975,34 @@ output; >>> on the logical switch. >>> </li> >>> >>> + <li> >>> + <p> >>> + For any logical port that's defined as a target of >>> routing protocol >>> + redirecting (via <code>routing-protocol-redirect</code> >>> option set on >>> + Logical Router Port), we redirect the traffic related to >>> protocols >>> + specified in <code>routing-protocols</code> option. It's >>> acoomplished >>> + with following priority-100 flows: >>> + <p> >> This should be </p>. >> >>> + <ul> >>> + <li> >>> + Flows that match Logical Router Port's IPs and >>> destination port of >>> + the routing daemon are redirected to this port to >>> allow external >>> + peers' connection to the daemon listening on this >>> port. >>> + </li> >>> + <li> >>> + Flows that match Logical Router Port's IPs and source >>> port of >>> + the routing daemon are redirected to this port to >>> allow replies >>> + from the peers. >>> + </li> >>> + </ul> >>> + <p> >>> + In addition to this, we add priority-100 rules that >>> + <code>clone</code> ARP replies and IPv6 Neighbor >>> Advertisements to >>> + this port as well. These allow to build proper ARP/IPv6 >>> neighbor >>> + list on this port. >>> + </p> >>> + </li> >>> + >>> <li> >>> Priority-90 flows for transit switches that forward >>> registered >>> IP multicast traffic to their corresponding multicast >>> group , which >>> diff --git a/ovn-nb.xml b/ovn-nb.xml >>> index 9552534f6..092195473 100644 >>> --- a/ovn-nb.xml >>> +++ b/ovn-nb.xml >>> @@ -3440,6 +3440,48 @@ or >>> </p> >>> </column> >>> >>> + <column name="options" key="routing-protocol-redirect" >>> + type='{"type": "string"}'> >>> + <p> >>> + NOTE: this feature is experimental and may be subject to >>> + removal/change in the future. >>> + </p> >>> + <p> >>> + This option expects a name of a Logical Switch Port >>> that's present >>> + in the peer's Logical Switch. If set, it causes any >>> traffic >>> + that's destined for Logical Router Port's IP addresses >>> (including >>> + its IPv6 LLA) and the ports associated with routing >>> protocols defined >>> + ip <code>routing-protocols</code> option, to be >>> redirected >>> + to the specified Logical Switch Port. >>> + >>> + This allows external routing daemons to be bound to a >>> port in OVN's >>> + Logical Switch and act as if they were listening on >>> Logical Router >>> + Port's IP addresses. >>> + </p> >>> + </column> >>> + >>> + <column name="options" key="routing-protocols" >>> type='{"type": "string"}'> >>> + <p> >>> + NOTE: this feature is experimental and may be subject to >>> + removal/change in the future. >>> + </p> >>> + <p> >>> + This option expects a comma-separated list of routing, >>> and >>> + routing-related protocols, whose control plane traffic >>> will be >>> + redirected to a port specified in >>> + <code>routing-protocol-redirect</code> option. Currently >>> supported >>> + options are: >>> + </p> >>> + <ul> >>> + <li> >>> + <code>BGP</code> (forwards TCP port 179) >>> + </li> >>> + <li> >>> + <code>BFD</code> (forwards UDP port 3748) >>> + </li> >>> + <ul> >> This should be </ul>. >> >> The rest looks OK to me. You can find the rebased version, including >> these minor nits fixed here: >> >> https://github.com/dceara/ovn/commits/review-pws418655-bgp-redirect-v5/ >> >> If the small changes I made look OK to you feel free to squash them >> in >> and post v6 with my: >> >> Acked-by: Dumitru Ceara<dceara@redhat.com> >> >> Thanks, >> Dumitru >> > _______________________________________________ > dev mailing list > dev@openvswitch.org > https://mail.openvswitch.org/mailman/listinfo/ovs-dev
On 09.08.2024 15:48, Vladislav Odintsov wrote: > Hi Martin, Dumitru, > > I'm gonna to perform a quick testing after rebase and get back with a > feedback. > > Please see one question below. > > On 09.08.2024 15:25, martin.kalcok@canonical.com wrote: >> Hi Dumitru, >> thank you for the fast review. I'll post v6 momentarily, I just want to >> do quick fresh end-to-end setup/test. >> I also want to add that you were correct about the ARP/ND. It didn't >> work properly and it was living off the entries populated by unicast >> traffic from the peer, however the entries were flapping wildly and it >> was discovered when I threw the BFD into the mix (real blessing in >> disguise). The slower BFD protocol was able to deal with it without >> disconnects, but BFD was disconnecting every once in a while. Hence the >> `clone`rules for the ARP replies and IPv6 NAs. >> >> >> On Fri, 2024-08-09 at 08:19 +0200, Dumitru Ceara wrote: >>> On 8/9/24 04:45, Martin Kalcok wrote: >>>> This change adds two new LRP options: >>>> * routing-protocol-redirect >>>> * routing-protocols >>>> >>>> These allow redirection of a routing protocol traffic to >>>> an Logical Switch Port. This enables external routing daemons >>>> to listen on an interface bound to an LSP and effectively act >>>> as if they were listening on (and speaking from) LRP's IP address. >>>> >>>> Option 'routing-protocols' takes a comma-separated list of routing >>>> protocols whose traffic should be redirected. Currently supported >>>> are BGP (tcp 179) and BFD (udp 3784). >>>> >>>> Option 'routing-protocol-redirect' expects a string with an LSP >>>> name. >>>> >>>> When both of these options are set, any traffic entering LS >>>> that's destined for LRP's IP addresses (including IPv6 LLA) and >>>> routing protocol's port number, is redirected to the LSP specified >>>> in the 'routing-protocol-redirect' value. >>>> >>>> NOTE: this feature is experimental and may be subject to >>>> removal/change in the future. >>>> >>>> Signed-off-by: Martin Kalcok<martin.kalcok@canonical.com> >>>> --- >>> Hi Martin, >>> >>> Thanks for v5! Unfortunately it needs a rebase. >>> >>> Also there were some minor things that need to be fixed in the man >>> pages. While at it I have a few other small comments. >>> >>>> v5 includes: >>>> * change from pure BGP redirect to a configurable protocol >>>> redirection >>>> * fixes issue with local routing daemon not being able to receive >>>> reply >>>> traffic when contacting its peer. >>>> * ARP and ND traffic is cloned to the redirected port, to >>>> properly populate >>>> information about its neighbors. >>>> >>>> >>>> northd/northd.c | 217 >>>> ++++++++++++++++++++++++++++++++++++++++ >>>> northd/northd.h | 7 ++ >>>> northd/ovn-northd.8.xml | 54 ++++++++++ >>>> ovn-nb.xml | 42 ++++++++ >>>> tests/ovn-northd.at | 93 +++++++++++++++++ >>>> tests/system-ovn.at | 100 ++++++++++++++++++ >>>> 6 files changed, 513 insertions(+) >>>> >>>> diff --git a/northd/northd.c b/northd/northd.c >>>> index 4353df07d..39b1998fd 100644 >>>> --- a/northd/northd.c >>>> +++ b/northd/northd.c >>>> @@ -13048,6 +13048,221 @@ build_arp_resolve_flows_for_lrp(struct >>>> ovn_port *op, >>>> } >>>> } >>>> +static void >>>> +build_r_p_redirect_rule__( >>> I think I prefer a more explicit name, e.g., >>> build_routing_protocols_redirect_rule__(). >>> >>>> + const char *s_addr, const char *redirect_port_name, int >>>> protocol_port, >>>> + const char *proto, bool is_ipv6, struct ovn_port *ls_peer, >>>> + struct lflow_table *lflows, struct ds *match, struct ds >>>> *actions) >>>> +{ >>>> + int ip_ver = is_ipv6 ? 6 : 4; >>>> + ds_clear(actions); >>>> + ds_put_format(actions, "outport = \"%s\"; output;", >>>> redirect_port_name); >>>> + >>>> + /* Redirect packets in the input pipeline destined for LR's IP >>>> + * and the routing protocol's port to the LSP specified in >>>> + * 'routing-protocol-redirect' option.*/ >>>> + ds_clear(match); >>>> + ds_put_format(match, "ip%d.dst == %s && %s.dst == %d", ip_ver, >>>> s_addr, >>>> + proto, protocol_port); >>>> + ovn_lflow_add(lflows, ls_peer->od, S_SWITCH_IN_L2_LKUP, 100, >>>> + ds_cstr(match), >>>> + ds_cstr(actions), >>>> + ls_peer->lflow_ref); >>>> + >>>> + /* To accomodate "peer" nature of the routing daemons, >>>> redirect also >>>> + * replies to the daemons' client requests. */ >>>> + ds_clear(match); >>>> + ds_put_format(match, "ip%d.dst == %s && %s.src == %d", ip_ver, >>>> s_addr, >>>> + proto, protocol_port); >>>> + ovn_lflow_add(lflows, ls_peer->od, S_SWITCH_IN_L2_LKUP, 100, >>>> + ds_cstr(match), >>>> + ds_cstr(actions), >>>> + ls_peer->lflow_ref); >>>> +} > > Don't we want to merge these two conditions into one logical flow? > > E.g.: > > "(ip%d.dst == %s && (%s.dst == %d && %s.src == %d)" Sorry, there is typo. It should be: "(ip%d.dst == %s && (%s.dst == %d || %s.src == %d)" > ? > > This will make one logical flow per LRP IP per protocol instead of two. > >>>> + >>>> +static void >>>> +apply_r_p_redirect__( >>> Same here, I think I prefer apply_routing_protocols_redirect__(). >>> >>>> + const char *s_addr, const char *redirect_port_name, int >>>> protocol_flags, >>>> + bool is_ipv6, struct ovn_port *ls_peer, struct lflow_table >>>> *lflows, >>>> + struct ds *match, struct ds *actions) >>>> +{ >>>> + if (protocol_flags & REDIRECT_BGP) { >>>> + build_r_p_redirect_rule__(s_addr, redirect_port_name, 179, >>>> "tcp", >>>> + is_ipv6, ls_peer, lflows, match, >>>> actions); >>>> + } >>>> + >>>> + if (protocol_flags & REDIRECT_BFD) { >>>> + build_r_p_redirect_rule__(s_addr, redirect_port_name, >>>> 3784, "udp", >>>> + is_ipv6, ls_peer, lflows, match, >>>> actions); >>>> + } >>>> + >>>> + /* Because the redirected port shares IP and MAC addresses >>>> with the LRP, >>>> + * special consideration needs to be given to the signaling >>>> protocols. */ >>>> + if (is_ipv6) { >>>> + /* Ensure that redirect port receives copy of NA messages >>>> destined to >>>> + * its IP.*/ >>>> + ds_clear(match); >>>> + ds_clear(actions); >>>> + ds_put_format(actions, "clone { outport = \"%s\"; output; >>>> };", >>>> + redirect_port_name); >>>> + ds_put_format(match, "ip6.dst == %s && nd_na", s_addr); >>>> + ovn_lflow_add(lflows, ls_peer->od, S_SWITCH_IN_L2_LKUP, >>>> 100, >>>> + ds_cstr(match), >>>> + ds_cstr(actions), >>>> + ls_peer->lflow_ref); >>>> + } else { >>>> + /* Ensure that redirect port receives copy of ARP replies >>>> destined to >>>> + * its IP */ >>>> + ds_clear(match); >>>> + ds_clear(actions); >>>> + ds_put_format(actions, "clone { outport = \"%s\"; output; >>>> };", >>>> + redirect_port_name); >>>> + ds_put_format(match, "arp.op == 2 && arp.tpa == %s", >>>> s_addr); >>>> + ovn_lflow_add(lflows, ls_peer->od, S_SWITCH_IN_L2_LKUP, >>>> 100, >>>> + ds_cstr(match), >>>> + ds_cstr(actions), >>>> + ls_peer->lflow_ref); >>>> + >>> Nit: no need for newline. >>> >>>> + } >>>> +} >>>> + >>>> +static int >>>> +parse_redirected_routing_protocols(struct ovn_port *lrp) { >>>> + int redirected_protocol_flags = 0; >>>> + const char *redirect_protocols = smap_get(&lrp->nbrp->options, >>>> + "routing- >>>> protocols"); >>>> + if (redirect_protocols == NULL) { >>>> + return redirected_protocol_flags; >>>> + } >>>> + >>>> + char *proto; >>>> + char *save_ptr = NULL; >>>> + char *tokstr = xstrdup(redirect_protocols); >>>> + for (proto = strtok_r(tokstr, ",", &save_ptr); proto != NULL; >>>> + proto = strtok_r(NULL, ",", &save_ptr)) { >>>> + if (!strcmp(proto, "BGP")) { >>>> + redirected_protocol_flags |= REDIRECT_BGP; >>>> + continue; >>>> + } >>>> + >>>> + if (!strcmp(proto, "BFD")) { >>>> + redirected_protocol_flags |= REDIRECT_BFD; >>>> + continue; >>>> + } >>> Nit: missing newlines. >>> >>>> + static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, >>>> 5); >>>> + VLOG_WARN_RL(&rl, "Option 'routing-protocols' encountered >>>> unknown " >>>> + "value %s", >>>> + proto); >>>> + } >>>> + free(tokstr); >>>> + return redirected_protocol_flags; >>>> +} >>>> + >>>> +static void >>>> +build_lrouter_routing_protocol_redirect( >>>> + struct ovn_port *op, struct lflow_table *lflows, >>>> + struct ds *match, struct ds *actions) >>>> +{ >>>> + /* LRP has to have a peer.*/ >>>> + if (op->peer == NULL) { >>>> + return; >>>> + } >>> Nit: missing newline. >>> >>>> + /* LRP has to have NB record.*/ >>>> + if (op->nbrp == NULL) { >>>> + return; >>>> + } >>>> + >>>> + /* Proceed only for LRPs that have 'routing-protocol-redirect' >>>> option set. >>>> + * Value of this option is the name of LSP to which the >>>> routing protocol >>>> + * traffic will be redirected. */ >>>> + const char *redirect_port = smap_get(&op->nbrp->options, >>>> + "routing-protocol- >>>> redirect"); >>>> + if (redirect_port == NULL) { >>>> + return; >>>> + } >>>> + >>>> + if (op->cr_port != NULL) { >>>> + static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, >>>> 5); >>>> + VLOG_WARN_RL(&rl, "Option 'routing-protocol-redirect' is >>>> not " >>>> + "supported on Distributed Gateway Port >>>> '%s'", >>>> + op->key); >>>> + return; >>>> + } >>>> + >>>> + /* Ensure that LSP, to which the routing protocol traffic is >>>> redirected, >>>> + * exists.*/ >>> Nit: missing space after '.' >>> >>>> + struct ovn_port *peer_lsp; >>>> + bool redirect_port_exists = false; >>>> + HMAP_FOR_EACH (peer_lsp, dp_node, &op->peer->od->ports) { >>>> + size_t peer_lsp_s = strlen(peer_lsp->key); >>>> + if (peer_lsp_s == strlen(redirect_port) >>>> + && !strncmp(peer_lsp->key, redirect_port, >>>> peer_lsp_s)){ >>>> + redirect_port_exists = true; >>>> + break; >>>> + } >>>> + } >>>> + >>>> + if (!redirect_port_exists) { >>>> + static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, >>>> 5); >>>> + VLOG_WARN_RL(&rl, "Option 'routing-protocol-redirect' set >>>> on Logical " >>>> + "Router Port '%s' refers to non-existent >>>> Logical " >>>> + "Switch Port. Routing protocol >>>> redirecting won't be " >>>> + "configured.", >>>> + op->key); >>>> + return; >>>> + } >>>> + >>>> + >>> Nit: too many new lines. >>> >>>> + int redirected_protocols = >>>> parse_redirected_routing_protocols(op); >>>> + if (!redirected_protocols) { >>>> + static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, >>>> 5); >>>> + VLOG_WARN_RL(&rl, "Option 'routing-protocol-redirect' is >>>> set on " >>>> + "Logical Router Port '%s' but no known >>>> protocols " >>>> + "were set via 'routing-protocols' >>>> options. This " >>>> + "configuration has no effect.", >>>> + op->key); >>>> + return; >>>> + } >>>> + >>>> + /* Redirecte traffic destined for LRP's IPs and the specified >>>> routing >>> Typo: redirecte >>> >>>> + * protocol ports to the port defined in 'routing-protocol- >>>> redirect' >>>> + * option.*/ >>>> + for (size_t i = 0; i < op->lrp_networks.n_ipv4_addrs; i++) { >>>> + const char *ip_s = op->lrp_networks.ipv4_addrs[i].addr_s; >>>> + apply_r_p_redirect__(ip_s, redirect_port, >>>> redirected_protocols, false, >>>> + op->peer, lflows,match, actions); >>>> + } >>>> + for (size_t i = 0; i < op->lrp_networks.n_ipv6_addrs; i++) { >>>> + const char *ip_s = op->lrp_networks.ipv6_addrs[i].addr_s; >>>> + apply_r_p_redirect__(ip_s, redirect_port, >>>> redirected_protocols, true, >>>> + op->peer, lflows,match, >>>> actions); >>>> + } >>>> + >>>> + /* Drop ARP replies and IPv6 RA/NA packets originating from >>>> + * 'routing-protocol-redirect' LSP. As this port shares IP and >>>> MAC >>>> + * addresses with LRP, we don't want to create duplicates.*/ >>>> + ds_clear(match); >>>> + ds_put_format(match, "inport == \"%s\" && arp.op == 2", >>>> redirect_port); >>>> + ovn_lflow_add(lflows, op->peer->od, >>>> S_SWITCH_IN_CHECK_PORT_SEC, 80, >>>> + ds_cstr(match), >>>> + REGBIT_PORT_SEC_DROP " = 1; next;", >>>> + op->peer->lflow_ref); >>>> + >>>> + ds_clear(match); >>>> + ds_put_format(match, "inport == \"%s\" && nd_na", >>>> redirect_port); >>>> + ovn_lflow_add(lflows, op->peer->od, >>>> S_SWITCH_IN_CHECK_PORT_SEC, 80, >>>> + ds_cstr(match), >>>> + REGBIT_PORT_SEC_DROP " = 1; next;", >>>> + op->peer->lflow_ref); >>>> + >>>> + ds_clear(match); >>>> + ds_put_format(match, "inport == \"%s\" && nd_ra", >>>> redirect_port); >>>> + ovn_lflow_add(lflows, op->peer->od, >>>> S_SWITCH_IN_CHECK_PORT_SEC, 80, >>>> + ds_cstr(match), >>>> + REGBIT_PORT_SEC_DROP " = 1; next;", >>>> + op->peer->lflow_ref); >>>> +} >>>> + >>>> /* This function adds ARP resolve flows related to a LSP. */ >>>> static void >>>> build_arp_resolve_flows_for_lsp( >>>> @@ -16003,6 +16218,8 @@ >>>> build_lswitch_and_lrouter_iterate_by_lrp(struct ovn_port *op, >>>> lsi->meter_groups, op->lflow_ref); >>>> build_lrouter_icmp_packet_toobig_admin_flows(op, lsi->lflows, >>>> &lsi->match, >>>> &lsi->actions, >>>> op->lflow_ref); >>>> + build_lrouter_routing_protocol_redirect(op, lsi->lflows, >>>> + &lsi->match, &lsi- >>>>> actions); >>>> } >>>> static void * >>>> diff --git a/northd/northd.h b/northd/northd.h >>>> index d4a8d75ab..691c23f01 100644 >>>> --- a/northd/northd.h >>>> +++ b/northd/northd.h >>>> @@ -89,6 +89,13 @@ ods_size(const struct ovn_datapaths *datapaths) >>>> bool od_has_lb_vip(const struct ovn_datapath *od); >>>> +/* List of routing and routing-related protocols which >>>> + * OVN is capable of redirecting from LRP to specific LSP. */ >>>> +enum redirected_routing_protcol_flag_type { >>>> + REDIRECT_BGP = (1 << 0), >>>> + REDIRECT_BFD = (1 << 1), >>>> +}; >>>> + >>>> struct tracked_ovn_ports { >>>> /* tracked created ports. >>>> * hmapx node data is 'struct ovn_port *' */ >>>> diff --git a/northd/ovn-northd.8.xml b/northd/ovn-northd.8.xml >>>> index b06b09ac5..250c602f2 100644 >>>> --- a/northd/ovn-northd.8.xml >>>> +++ b/northd/ovn-northd.8.xml >>>> @@ -284,6 +284,32 @@ >>>> dropped in the next stage. >>>> </li> >>>> + <li> >>>> + <p> >>>> + For each logical port that's defined as a target of >>>> routing protocol >>>> + redirecting (via <code>routing-protocol-redirect</code> >>>> option set on >>>> + Logical Router Port), a filter is set in place that >>>> disallows >>>> + following traffic exiting this port: >>>> + </p> >>>> + <ul> >>>> + <li> >>>> + ARP replies >>>> + </li> >>>> + <li> >>>> + IPv6 Neighbor Discovery - Router Advertisements >>>> + </li> >>>> + <li> >>>> + IPv6 Neighbor Discovery - Neighbor Advertisements >>>> + </li> >>>> + </ul> >>>> + <p> >>>> + Since this port shares IP and MAC addresses with the >>>> Logical Router >>>> + Port, we wan't to prevent duplicate replies and >>>> advertisements. This >>>> + is achieved by a rule with priority 80 that sets >>>> + <code>REGBIT_PORT_SEC_DROP" = 1; next;"</code>. >>>> + </p> >>>> + </li> >>>> + >>>> <li> >>>> For each (enabled) vtep logical port, a priority 70 flow >>>> is added which >>>> matches on all packets and applies the action >>>> @@ -1949,6 +1975,34 @@ output; >>>> on the logical switch. >>>> </li> >>>> + <li> >>>> + <p> >>>> + For any logical port that's defined as a target of >>>> routing protocol >>>> + redirecting (via <code>routing-protocol-redirect</code> >>>> option set on >>>> + Logical Router Port), we redirect the traffic related to >>>> protocols >>>> + specified in <code>routing-protocols</code> option. It's >>>> acoomplished >>>> + with following priority-100 flows: >>>> + <p> >>> This should be </p>. >>> >>>> + <ul> >>>> + <li> >>>> + Flows that match Logical Router Port's IPs and >>>> destination port of >>>> + the routing daemon are redirected to this port to >>>> allow external >>>> + peers' connection to the daemon listening on this >>>> port. >>>> + </li> >>>> + <li> >>>> + Flows that match Logical Router Port's IPs and source >>>> port of >>>> + the routing daemon are redirected to this port to >>>> allow replies >>>> + from the peers. >>>> + </li> >>>> + </ul> >>>> + <p> >>>> + In addition to this, we add priority-100 rules that >>>> + <code>clone</code> ARP replies and IPv6 Neighbor >>>> Advertisements to >>>> + this port as well. These allow to build proper ARP/IPv6 >>>> neighbor >>>> + list on this port. >>>> + </p> >>>> + </li> >>>> + >>>> <li> >>>> Priority-90 flows for transit switches that forward >>>> registered >>>> IP multicast traffic to their corresponding multicast >>>> group , which >>>> diff --git a/ovn-nb.xml b/ovn-nb.xml >>>> index 9552534f6..092195473 100644 >>>> --- a/ovn-nb.xml >>>> +++ b/ovn-nb.xml >>>> @@ -3440,6 +3440,48 @@ or >>>> </p> >>>> </column> >>>> + <column name="options" key="routing-protocol-redirect" >>>> + type='{"type": "string"}'> >>>> + <p> >>>> + NOTE: this feature is experimental and may be subject to >>>> + removal/change in the future. >>>> + </p> >>>> + <p> >>>> + This option expects a name of a Logical Switch Port >>>> that's present >>>> + in the peer's Logical Switch. If set, it causes any >>>> traffic >>>> + that's destined for Logical Router Port's IP addresses >>>> (including >>>> + its IPv6 LLA) and the ports associated with routing >>>> protocols defined >>>> + ip <code>routing-protocols</code> option, to be >>>> redirected >>>> + to the specified Logical Switch Port. >>>> + >>>> + This allows external routing daemons to be bound to a >>>> port in OVN's >>>> + Logical Switch and act as if they were listening on >>>> Logical Router >>>> + Port's IP addresses. >>>> + </p> >>>> + </column> >>>> + >>>> + <column name="options" key="routing-protocols" >>>> type='{"type": "string"}'> >>>> + <p> >>>> + NOTE: this feature is experimental and may be subject to >>>> + removal/change in the future. >>>> + </p> >>>> + <p> >>>> + This option expects a comma-separated list of routing, >>>> and >>>> + routing-related protocols, whose control plane traffic >>>> will be >>>> + redirected to a port specified in >>>> + <code>routing-protocol-redirect</code> option. Currently >>>> supported >>>> + options are: >>>> + </p> >>>> + <ul> >>>> + <li> >>>> + <code>BGP</code> (forwards TCP port 179) >>>> + </li> >>>> + <li> >>>> + <code>BFD</code> (forwards UDP port 3748) >>>> + </li> >>>> + <ul> >>> This should be </ul>. >>> >>> The rest looks OK to me. You can find the rebased version, including >>> these minor nits fixed here: >>> >>> https://github.com/dceara/ovn/commits/review-pws418655-bgp-redirect-v5/ >>> >>> If the small changes I made look OK to you feel free to squash them >>> in >>> and post v6 with my: >>> >>> Acked-by: Dumitru Ceara<dceara@redhat.com> >>> >>> Thanks, >>> Dumitru >>> >> _______________________________________________ >> dev mailing list >> dev@openvswitch.org >> https://mail.openvswitch.org/mailman/listinfo/ovs-dev > _______________________________________________ > dev mailing list > dev@openvswitch.org > https://mail.openvswitch.org/mailman/listinfo/ovs-dev
On Friday, August 9, 2024, Vladislav Odintsov <odivlad@gmail.com> wrote: > > Don't we want to merge these two conditions into one logical flow? > > E.g.: > > "(ip%d.dst == %s && (%s.dst == %d && %s.src == %d)" > > Sorry, there is typo. It should be: > > "(ip%d.dst == %s && (%s.dst == %d || %s.src == %d)" > > ? > > This will make one logical flow per LRP IP per protocol instead of two. > > I didn’t test this but I think that looks ok. Regards, Dumitru
I don't mind merging them, but while we are at the topic. Are there benefits in processing performance in "fewer, more complex rules" vs "more, less complex rules"? Or is it just to improve readability? Martin. On Fri, 2024-08-09 at 11:08 +0200, Dumitru Ceara wrote: > On Friday, August 9, 2024, Vladislav Odintsov <odivlad@gmail.com> > wrote: > > > Don't we want to merge these two conditions into one logical > > > flow? > > > > > > E.g.: > > > > > > "(ip%d.dst == %s && (%s.dst == %d && %s.src == %d)" > > > > Sorry, there is typo. It should be: > > > > "(ip%d.dst == %s && (%s.dst == %d || %s.src == %d)" > > > > > ? > > > > > > This will make one logical flow per LRP IP per protocol instead > > > of two. > > > > I didn’t test this but I think that looks ok. > > Regards, > Dumitru
On Friday, August 9, 2024, <martin.kalcok@canonical.com> wrote: > I don't mind merging them, but while we are at the topic. Are there > benefits in processing performance in "fewer, more complex rules" vs "more, > less complex rules"? Or is it just to improve readability? > > In this case I don’t expect we’ll have a lot of ports with redirect enabled so, from my perspective, it’s just readability (I was ok with the 2 flow version too). From an openflow perspective there’s no difference. Thanks, Dumitru > Martin. > > On Fri, 2024-08-09 at 11:08 +0200, Dumitru Ceara wrote: > > On Friday, August 9, 2024, Vladislav Odintsov <odivlad@gmail.com> wrote: > > Don't we want to merge these two conditions into one logical flow? > > E.g.: > > "(ip%d.dst == %s && (%s.dst == %d && %s.src == %d)" > > Sorry, there is typo. It should be: > > "(ip%d.dst == %s && (%s.dst == %d || %s.src == %d)" > > ? > > This will make one logical flow per LRP IP per protocol instead of two. > > > > I didn’t test this but I think that looks ok. > > Regards, > Dumitru > > >
On Fri, 2024-08-09 at 11:28 +0200, Dumitru Ceara wrote: > On Friday, August 9, 2024, <martin.kalcok@canonical.com> wrote: > > I don't mind merging them, but while we are at the topic. Are there > > benefits in processing performance in "fewer, more complex rules" > > vs "more, less complex rules"? Or is it just to improve > > readability? > > > > > > In this case I don’t expect we’ll have a lot of ports with redirect > enabled so, from my perspective, it’s just readability (I was ok with > the 2 flow version too). From an openflow perspective there’s no > difference. Thanks, I think I'll have to keep the current 2-line implementation in and add this into the list of future improvements. @Vladislav I do appreciate the review and the feedback, and I don't want to look like I'm just ignoring it, but due to the time pressure (freeze today) and this change resulting in tests/docs change (which always takes me way more than I expect) i don't think I'll fit it in. It doesn't look like it but the overhead adds up. Martin. > > Thanks, > Dumitru > > > Martin. > > > > On Fri, 2024-08-09 at 11:08 +0200, Dumitru Ceara wrote: > > > On Friday, August 9, 2024, Vladislav Odintsov <odivlad@gmail.com> > > > wrote: > > > > > Don't we want to merge these two conditions into one logical > > > > > flow? > > > > > > > > > > E.g.: > > > > > > > > > > "(ip%d.dst == %s && (%s.dst == %d && %s.src == %d)" > > > > > > > > Sorry, there is typo. It should be: > > > > > > > > "(ip%d.dst == %s && (%s.dst == %d || %s.src == %d)" > > > > > > > > > ? > > > > > > > > > > This will make one logical flow per LRP IP per protocol > > > > > instead of two. > > > > > > > > > > I didn’t test this but I think that looks ok. > > > > > > Regards, > > > Dumitru > > > > > >
On 09.08.2024 16:51, martin.kalcok@canonical.com wrote: > On Fri, 2024-08-09 at 11:28 +0200, Dumitru Ceara wrote: >> On Friday, August 9, 2024,<martin.kalcok@canonical.com> wrote: >>> I don't mind merging them, but while we are at the topic. Are there >>> benefits in processing performance in "fewer, more complex rules" >>> vs "more, less complex rules"? Or is it just to improve >>> readability? >>> >>> >> In this case I don’t expect we’ll have a lot of ports with redirect >> enabled so, from my perspective, it’s just readability (I was ok with >> the 2 flow version too). From an openflow perspective there’s no >> difference. > Thanks, I think I'll have to keep the current 2-line implementation in > and add this into the list of future improvements. @Vladislav I do > appreciate the review and the feedback, and I don't want to look like > I'm just ignoring it, but due to the time pressure (freeze today) and > this change resulting in tests/docs change (which always takes me way > more than I expect) i don't think I'll fit it in. It doesn't look like > it but the overhead adds up. Sure, no problem. I'm going to test v6. Will return with feedback soon. Stay tuned :) > > Martin. > >> Thanks, >> Dumitru >> >>> Martin. >>> >>> On Fri, 2024-08-09 at 11:08 +0200, Dumitru Ceara wrote: >>>> On Friday, August 9, 2024, Vladislav Odintsov<odivlad@gmail.com> >>>> wrote: >>>>>> Don't we want to merge these two conditions into one logical >>>>>> flow? >>>>>> >>>>>> E.g.: >>>>>> >>>>>> "(ip%d.dst == %s && (%s.dst == %d && %s.src == %d)" >>>>> >>>>> Sorry, there is typo. It should be: >>>>> >>>>> "(ip%d.dst == %s && (%s.dst == %d || %s.src == %d)" >>>>> >>>>>> ? >>>>>> >>>>>> This will make one logical flow per LRP IP per protocol >>>>>> instead of two. >>>> I didn’t test this but I think that looks ok. >>>> >>>> Regards, >>>> Dumitru >>> >>> > _______________________________________________ > dev mailing list > dev@openvswitch.org > https://mail.openvswitch.org/mailman/listinfo/ovs-dev
diff --git a/northd/northd.c b/northd/northd.c index 4353df07d..39b1998fd 100644 --- a/northd/northd.c +++ b/northd/northd.c @@ -13048,6 +13048,221 @@ build_arp_resolve_flows_for_lrp(struct ovn_port *op, } } +static void +build_r_p_redirect_rule__( + const char *s_addr, const char *redirect_port_name, int protocol_port, + const char *proto, bool is_ipv6, struct ovn_port *ls_peer, + struct lflow_table *lflows, struct ds *match, struct ds *actions) +{ + int ip_ver = is_ipv6 ? 6 : 4; + ds_clear(actions); + ds_put_format(actions, "outport = \"%s\"; output;", redirect_port_name); + + /* Redirect packets in the input pipeline destined for LR's IP + * and the routing protocol's port to the LSP specified in + * 'routing-protocol-redirect' option.*/ + ds_clear(match); + ds_put_format(match, "ip%d.dst == %s && %s.dst == %d", ip_ver, s_addr, + proto, protocol_port); + ovn_lflow_add(lflows, ls_peer->od, S_SWITCH_IN_L2_LKUP, 100, + ds_cstr(match), + ds_cstr(actions), + ls_peer->lflow_ref); + + /* To accomodate "peer" nature of the routing daemons, redirect also + * replies to the daemons' client requests. */ + ds_clear(match); + ds_put_format(match, "ip%d.dst == %s && %s.src == %d", ip_ver, s_addr, + proto, protocol_port); + ovn_lflow_add(lflows, ls_peer->od, S_SWITCH_IN_L2_LKUP, 100, + ds_cstr(match), + ds_cstr(actions), + ls_peer->lflow_ref); +} + +static void +apply_r_p_redirect__( + const char *s_addr, const char *redirect_port_name, int protocol_flags, + bool is_ipv6, struct ovn_port *ls_peer, struct lflow_table *lflows, + struct ds *match, struct ds *actions) +{ + if (protocol_flags & REDIRECT_BGP) { + build_r_p_redirect_rule__(s_addr, redirect_port_name, 179, "tcp", + is_ipv6, ls_peer, lflows, match, actions); + } + + if (protocol_flags & REDIRECT_BFD) { + build_r_p_redirect_rule__(s_addr, redirect_port_name, 3784, "udp", + is_ipv6, ls_peer, lflows, match, actions); + } + + /* Because the redirected port shares IP and MAC addresses with the LRP, + * special consideration needs to be given to the signaling protocols. */ + if (is_ipv6) { + /* Ensure that redirect port receives copy of NA messages destined to + * its IP.*/ + ds_clear(match); + ds_clear(actions); + ds_put_format(actions, "clone { outport = \"%s\"; output; };", + redirect_port_name); + ds_put_format(match, "ip6.dst == %s && nd_na", s_addr); + ovn_lflow_add(lflows, ls_peer->od, S_SWITCH_IN_L2_LKUP, 100, + ds_cstr(match), + ds_cstr(actions), + ls_peer->lflow_ref); + } else { + /* Ensure that redirect port receives copy of ARP replies destined to + * its IP */ + ds_clear(match); + ds_clear(actions); + ds_put_format(actions, "clone { outport = \"%s\"; output; };", + redirect_port_name); + ds_put_format(match, "arp.op == 2 && arp.tpa == %s", s_addr); + ovn_lflow_add(lflows, ls_peer->od, S_SWITCH_IN_L2_LKUP, 100, + ds_cstr(match), + ds_cstr(actions), + ls_peer->lflow_ref); + + } +} + +static int +parse_redirected_routing_protocols(struct ovn_port *lrp) { + int redirected_protocol_flags = 0; + const char *redirect_protocols = smap_get(&lrp->nbrp->options, + "routing-protocols"); + if (redirect_protocols == NULL) { + return redirected_protocol_flags; + } + + char *proto; + char *save_ptr = NULL; + char *tokstr = xstrdup(redirect_protocols); + for (proto = strtok_r(tokstr, ",", &save_ptr); proto != NULL; + proto = strtok_r(NULL, ",", &save_ptr)) { + if (!strcmp(proto, "BGP")) { + redirected_protocol_flags |= REDIRECT_BGP; + continue; + } + + if (!strcmp(proto, "BFD")) { + redirected_protocol_flags |= REDIRECT_BFD; + continue; + } + static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, 5); + VLOG_WARN_RL(&rl, "Option 'routing-protocols' encountered unknown " + "value %s", + proto); + } + free(tokstr); + return redirected_protocol_flags; +} + +static void +build_lrouter_routing_protocol_redirect( + struct ovn_port *op, struct lflow_table *lflows, + struct ds *match, struct ds *actions) +{ + /* LRP has to have a peer.*/ + if (op->peer == NULL) { + return; + } + /* LRP has to have NB record.*/ + if (op->nbrp == NULL) { + return; + } + + /* Proceed only for LRPs that have 'routing-protocol-redirect' option set. + * Value of this option is the name of LSP to which the routing protocol + * traffic will be redirected. */ + const char *redirect_port = smap_get(&op->nbrp->options, + "routing-protocol-redirect"); + if (redirect_port == NULL) { + return; + } + + if (op->cr_port != NULL) { + static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, 5); + VLOG_WARN_RL(&rl, "Option 'routing-protocol-redirect' is not " + "supported on Distributed Gateway Port '%s'", + op->key); + return; + } + + /* Ensure that LSP, to which the routing protocol traffic is redirected, + * exists.*/ + struct ovn_port *peer_lsp; + bool redirect_port_exists = false; + HMAP_FOR_EACH (peer_lsp, dp_node, &op->peer->od->ports) { + size_t peer_lsp_s = strlen(peer_lsp->key); + if (peer_lsp_s == strlen(redirect_port) + && !strncmp(peer_lsp->key, redirect_port, peer_lsp_s)){ + redirect_port_exists = true; + break; + } + } + + if (!redirect_port_exists) { + static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, 5); + VLOG_WARN_RL(&rl, "Option 'routing-protocol-redirect' set on Logical " + "Router Port '%s' refers to non-existent Logical " + "Switch Port. Routing protocol redirecting won't be " + "configured.", + op->key); + return; + } + + + int redirected_protocols = parse_redirected_routing_protocols(op); + if (!redirected_protocols) { + static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, 5); + VLOG_WARN_RL(&rl, "Option 'routing-protocol-redirect' is set on " + "Logical Router Port '%s' but no known protocols " + "were set via 'routing-protocols' options. This " + "configuration has no effect.", + op->key); + return; + } + + /* Redirecte traffic destined for LRP's IPs and the specified routing + * protocol ports to the port defined in 'routing-protocol-redirect' + * option.*/ + for (size_t i = 0; i < op->lrp_networks.n_ipv4_addrs; i++) { + const char *ip_s = op->lrp_networks.ipv4_addrs[i].addr_s; + apply_r_p_redirect__(ip_s, redirect_port, redirected_protocols, false, + op->peer, lflows,match, actions); + } + for (size_t i = 0; i < op->lrp_networks.n_ipv6_addrs; i++) { + const char *ip_s = op->lrp_networks.ipv6_addrs[i].addr_s; + apply_r_p_redirect__(ip_s, redirect_port, redirected_protocols, true, + op->peer, lflows,match, actions); + } + + /* Drop ARP replies and IPv6 RA/NA packets originating from + * 'routing-protocol-redirect' LSP. As this port shares IP and MAC + * addresses with LRP, we don't want to create duplicates.*/ + ds_clear(match); + ds_put_format(match, "inport == \"%s\" && arp.op == 2", redirect_port); + ovn_lflow_add(lflows, op->peer->od, S_SWITCH_IN_CHECK_PORT_SEC, 80, + ds_cstr(match), + REGBIT_PORT_SEC_DROP " = 1; next;", + op->peer->lflow_ref); + + ds_clear(match); + ds_put_format(match, "inport == \"%s\" && nd_na", redirect_port); + ovn_lflow_add(lflows, op->peer->od, S_SWITCH_IN_CHECK_PORT_SEC, 80, + ds_cstr(match), + REGBIT_PORT_SEC_DROP " = 1; next;", + op->peer->lflow_ref); + + ds_clear(match); + ds_put_format(match, "inport == \"%s\" && nd_ra", redirect_port); + ovn_lflow_add(lflows, op->peer->od, S_SWITCH_IN_CHECK_PORT_SEC, 80, + ds_cstr(match), + REGBIT_PORT_SEC_DROP " = 1; next;", + op->peer->lflow_ref); +} + /* This function adds ARP resolve flows related to a LSP. */ static void build_arp_resolve_flows_for_lsp( @@ -16003,6 +16218,8 @@ build_lswitch_and_lrouter_iterate_by_lrp(struct ovn_port *op, lsi->meter_groups, op->lflow_ref); build_lrouter_icmp_packet_toobig_admin_flows(op, lsi->lflows, &lsi->match, &lsi->actions, op->lflow_ref); + build_lrouter_routing_protocol_redirect(op, lsi->lflows, + &lsi->match, &lsi->actions); } static void * diff --git a/northd/northd.h b/northd/northd.h index d4a8d75ab..691c23f01 100644 --- a/northd/northd.h +++ b/northd/northd.h @@ -89,6 +89,13 @@ ods_size(const struct ovn_datapaths *datapaths) bool od_has_lb_vip(const struct ovn_datapath *od); +/* List of routing and routing-related protocols which + * OVN is capable of redirecting from LRP to specific LSP. */ +enum redirected_routing_protcol_flag_type { + REDIRECT_BGP = (1 << 0), + REDIRECT_BFD = (1 << 1), +}; + struct tracked_ovn_ports { /* tracked created ports. * hmapx node data is 'struct ovn_port *' */ diff --git a/northd/ovn-northd.8.xml b/northd/ovn-northd.8.xml index b06b09ac5..250c602f2 100644 --- a/northd/ovn-northd.8.xml +++ b/northd/ovn-northd.8.xml @@ -284,6 +284,32 @@ dropped in the next stage. </li> + <li> + <p> + For each logical port that's defined as a target of routing protocol + redirecting (via <code>routing-protocol-redirect</code> option set on + Logical Router Port), a filter is set in place that disallows + following traffic exiting this port: + </p> + <ul> + <li> + ARP replies + </li> + <li> + IPv6 Neighbor Discovery - Router Advertisements + </li> + <li> + IPv6 Neighbor Discovery - Neighbor Advertisements + </li> + </ul> + <p> + Since this port shares IP and MAC addresses with the Logical Router + Port, we wan't to prevent duplicate replies and advertisements. This + is achieved by a rule with priority 80 that sets + <code>REGBIT_PORT_SEC_DROP" = 1; next;"</code>. + </p> + </li> + <li> For each (enabled) vtep logical port, a priority 70 flow is added which matches on all packets and applies the action @@ -1949,6 +1975,34 @@ output; on the logical switch. </li> + <li> + <p> + For any logical port that's defined as a target of routing protocol + redirecting (via <code>routing-protocol-redirect</code> option set on + Logical Router Port), we redirect the traffic related to protocols + specified in <code>routing-protocols</code> option. It's acoomplished + with following priority-100 flows: + <p> + <ul> + <li> + Flows that match Logical Router Port's IPs and destination port of + the routing daemon are redirected to this port to allow external + peers' connection to the daemon listening on this port. + </li> + <li> + Flows that match Logical Router Port's IPs and source port of + the routing daemon are redirected to this port to allow replies + from the peers. + </li> + </ul> + <p> + In addition to this, we add priority-100 rules that + <code>clone</code> ARP replies and IPv6 Neighbor Advertisements to + this port as well. These allow to build proper ARP/IPv6 neighbor + list on this port. + </p> + </li> + <li> Priority-90 flows for transit switches that forward registered IP multicast traffic to their corresponding multicast group , which diff --git a/ovn-nb.xml b/ovn-nb.xml index 9552534f6..092195473 100644 --- a/ovn-nb.xml +++ b/ovn-nb.xml @@ -3440,6 +3440,48 @@ or </p> </column> + <column name="options" key="routing-protocol-redirect" + type='{"type": "string"}'> + <p> + NOTE: this feature is experimental and may be subject to + removal/change in the future. + </p> + <p> + This option expects a name of a Logical Switch Port that's present + in the peer's Logical Switch. If set, it causes any traffic + that's destined for Logical Router Port's IP addresses (including + its IPv6 LLA) and the ports associated with routing protocols defined + ip <code>routing-protocols</code> option, to be redirected + to the specified Logical Switch Port. + + This allows external routing daemons to be bound to a port in OVN's + Logical Switch and act as if they were listening on Logical Router + Port's IP addresses. + </p> + </column> + + <column name="options" key="routing-protocols" type='{"type": "string"}'> + <p> + NOTE: this feature is experimental and may be subject to + removal/change in the future. + </p> + <p> + This option expects a comma-separated list of routing, and + routing-related protocols, whose control plane traffic will be + redirected to a port specified in + <code>routing-protocol-redirect</code> option. Currently supported + options are: + </p> + <ul> + <li> + <code>BGP</code> (forwards TCP port 179) + </li> + <li> + <code>BFD</code> (forwards UDP port 3748) + </li> + <ul> + </column> + <column name="options" key="gateway_mtu_bypass"> <p> When configured, represents a match expression, in the same diff --git a/tests/ovn-northd.at b/tests/ovn-northd.at index a389d1988..faedaf1d4 100644 --- a/tests/ovn-northd.at +++ b/tests/ovn-northd.at @@ -12721,3 +12721,96 @@ AT_CHECK([ovn-sbctl dump-flows lr | grep lr_in_dnat | ovn_strip_lflows], [0], [d AT_CLEANUP ]) + +OVN_FOR_EACH_NORTHD_NO_HV([ +AT_SETUP([Routing protocol control plane redirect]) +ovn_start + +check ovn-sbctl chassis-add hv1 geneve 127.0.0.1 + +check ovn-nbctl lr-add lr -- \ + lrp-add lr lr-ls 02:ac:10:01:00:01 172.16.1.1/24 +check ovn-nbctl --wait=sb set logical_router lr options:chassis=hv1 + +check ovn-nbctl ls-add ls -- \ + lsp-add ls ls-lr -- \ + lsp-set-type ls-lr router -- \ + lsp-set-addresses ls-lr router -- \ + lsp-set-options ls-lr router-port=lr-ls + +check ovn-nbctl lsp-add ls lsp-bgp -- \ + lsp-set-addresses lsp-bgp unknown + +# Function that ensures that no redirect rules are installed. +check_no_redirect() { + AT_CHECK([ovn-sbctl dump-flows ls | grep ls_in_l2_lkup | grep -E "tcp.dst == 179|tcp.src == 179" | wc -l], [0], [0 +]) + + AT_CHECK([ovn-sbctl dump-flows ls | grep ls_in_check_port_sec | grep -E "priority=80" | wc -l], [0], [0 +]) + check_no_bfd_redirect +} + +# Function that ensures that no BFD redirect rules are installed. +check_no_bfd_redirect() { + AT_CHECK([ovn-sbctl dump-flows ls | grep ls_in_l2_lkup | grep -E "udp.dst == 3784|udp.src == 3784" | wc -l], [0], [0 +]) +} + +# By default, no rules related to routing protocol redirect are present +check_no_redirect + +# Set "lsp-bgp" port as target of BGP control plane redirected traffic +check ovn-nbctl --wait=sb set logical_router_port lr-ls options:routing-protocol-redirect=lsp-bgp +check ovn-nbctl --wait=sb set logical_router_port lr-ls options:routing-protocols=BGP + +# Check that BGP control plane traffic is redirected "lsp-bgp" +AT_CHECK([ovn-sbctl dump-flows ls | grep ls_in_l2_lkup | grep -E "tcp.dst == 179|tcp.src == 179" | ovn_strip_lflows], [0], [dnl + table=??(ls_in_l2_lkup ), priority=100 , match=(ip4.dst == 172.16.1.1 && tcp.dst == 179), action=(outport = "lsp-bgp"; output;) + table=??(ls_in_l2_lkup ), priority=100 , match=(ip4.dst == 172.16.1.1 && tcp.src == 179), action=(outport = "lsp-bgp"; output;) + table=??(ls_in_l2_lkup ), priority=100 , match=(ip6.dst == fe80::ac:10ff:fe01:1 && tcp.dst == 179), action=(outport = "lsp-bgp"; output;) + table=??(ls_in_l2_lkup ), priority=100 , match=(ip6.dst == fe80::ac:10ff:fe01:1 && tcp.src == 179), action=(outport = "lsp-bgp"; output;) +]) + +# Check that ARP/ND traffic is cloned to the "lsp-bgp" +AT_CHECK([ovn-sbctl dump-flows ls | grep ls_in_l2_lkup | grep "arp.op == 2 && arp.tpa == 172.16.1.1" | ovn_strip_lflows], [0], [dnl + table=??(ls_in_l2_lkup ), priority=100 , match=(arp.op == 2 && arp.tpa == 172.16.1.1), action=(clone { outport = "lsp-bgp"; output; };) +]) +AT_CHECK([ovn-sbctl dump-flows ls | grep ls_in_l2_lkup | grep "&& nd_na" | ovn_strip_lflows], [0], [dnl + table=??(ls_in_l2_lkup ), priority=100 , match=(ip6.dst == fe80::ac:10ff:fe01:1 && nd_na), action=(clone { outport = "lsp-bgp"; output; };) +]) + +# Check that at this point no BFD redirecting is present +check_no_bfd_redirect + +# Add BFD traffic redirect +check ovn-nbctl --wait=sb set logical_router_port lr-ls options:routing-protocols=BGP,BFD + +# Check that BFD traffic is redirected to "lsp-bgp" +AT_CHECK([ovn-sbctl dump-flows ls | grep ls_in_l2_lkup | grep -E "udp.dst == 3784|udp.src == 3784" | ovn_strip_lflows], [0], [dnl + table=??(ls_in_l2_lkup ), priority=100 , match=(ip4.dst == 172.16.1.1 && udp.dst == 3784), action=(outport = "lsp-bgp"; output;) + table=??(ls_in_l2_lkup ), priority=100 , match=(ip4.dst == 172.16.1.1 && udp.src == 3784), action=(outport = "lsp-bgp"; output;) + table=??(ls_in_l2_lkup ), priority=100 , match=(ip6.dst == fe80::ac:10ff:fe01:1 && udp.dst == 3784), action=(outport = "lsp-bgp"; output;) + table=??(ls_in_l2_lkup ), priority=100 , match=(ip6.dst == fe80::ac:10ff:fe01:1 && udp.src == 3784), action=(outport = "lsp-bgp"; output;) +]) + + +# Check that ARP replies and ND advertisements are blocked from exiting "lsp-bgp" +AT_CHECK([ovn-sbctl dump-flows ls | grep ls_in_check_port_sec | grep "priority=80" | ovn_strip_lflows], [0], [dnl + table=??(ls_in_check_port_sec), priority=80 , match=(inport == "lsp-bgp" && arp.op == 2), action=(reg0[[15]] = 1; next;) + table=??(ls_in_check_port_sec), priority=80 , match=(inport == "lsp-bgp" && nd_na), action=(reg0[[15]] = 1; next;) + table=??(ls_in_check_port_sec), priority=80 , match=(inport == "lsp-bgp" && nd_ra), action=(reg0[[15]] = 1; next;) +]) + +# Remove 'bgp-redirect' option from LRP and check that rules are removed +check ovn-nbctl --wait=sb remove logical_router_port lr-ls options routing-protocol-redirect +check ovn-nbctl --wait=sb remove logical_router_port lr-ls options routing-protocols +check_no_redirect + +# Set non-existent LSP as target of 'bgp-redirect' and check that no rules are added +check ovn-nbctl --wait=sb set logical_router_port lr-ls options:routing-protocol-redirect=lsp-foo +check ovn-nbctl --wait=sb set logical_router_port lr-ls options:routing-protocols=BGP,BFD +check_no_redirect + +AT_CLEANUP +]) diff --git a/tests/system-ovn.at b/tests/system-ovn.at index c24ede7c5..1012a92c4 100644 --- a/tests/system-ovn.at +++ b/tests/system-ovn.at @@ -13027,3 +13027,103 @@ OVS_TRAFFIC_VSWITCHD_STOP(["/failed to query port patch-.*/d /connection dropped.*/d"]) AT_CLEANUP ]) + +OVN_FOR_EACH_NORTHD([ +AT_SETUP([Routing protocol redirect]) +AT_SKIP_IF([test $HAVE_NC = no]) + +ovn_start +OVS_TRAFFIC_VSWITCHD_START() + +ADD_BR([br-int]) +ADD_BR([br-ext]) + +check ovs-ofctl add-flow br-ext action=normal +# Set external-ids in br-int needed for ovn-controller +check ovs-vsctl \ + -- set Open_vSwitch . external-ids:system-id=hv1 \ + -- set Open_vSwitch . external-ids:ovn-remote=unix:$ovs_base/ovn-sb/ovn-sb.sock \ + -- set Open_vSwitch . external-ids:ovn-encap-type=geneve \ + -- set Open_vSwitch . external-ids:ovn-encap-ip=169.0.0.1 \ + -- set bridge br-int fail-mode=secure other-config:disable-in-band=true + +# Start ovn-controller +start_daemon ovn-controller + +check ovn-nbctl lr-add R1 \ + -- set Logical_Router R1 options:chassis=hv1 + +check ovn-nbctl ls-add public + +check ovn-nbctl lrp-add R1 rp-public 00:00:02:01:02:03 172.16.1.1/24 + +check ovn-nbctl lsp-add public public-rp -- set Logical_Switch_Port public-rp \ + type=router options:router-port=rp-public \ + -- lsp-set-addresses public-rp router + +check ovn-nbctl lsp-add public bgp-daemon \ + -- lsp-set-addresses bgp-daemon unknown + +AT_CHECK([ovs-vsctl set Open_vSwitch . external-ids:ovn-bridge-mappings=phynet:br-ext]) +check ovn-nbctl lsp-add public public1 \ + -- lsp-set-addresses public1 unknown \ + -- lsp-set-type public1 localnet \ + -- lsp-set-options public1 network_name=phynet + +check ovn-nbctl --wait=hv sync + +# Set option that redirects BGP traffic to a LSP "bgp-daemon" +check ovn-nbctl --wait=sb set logical_router_port rp-public options:routing-protocol-redirect=bgp-daemon +check ovn-nbctl --wait=sb set logical_router_port rp-public options:routing-protocols=BGP + +# Create "bgp-daemon" interface in a namespace with IP and MAC matching LRP "rp-public" +ADD_NAMESPACES(bgp-daemon) +ADD_VETH(bgp-daemon, bgp-daemon, br-int, "172.16.1.1/24", "00:00:02:01:02:03") + +ADD_NAMESPACES(ext-foo) +ADD_VETH(ext-foo, ext-foo, br-ext, "172.16.1.100/24", "00:10:10:01:02:13", \ + "172.16.1.1") + +# Flip the interface down/up to get proper IPv6 LLA +NS_EXEC([bgp-daemon], [ip link set down bgp-daemon]) +NS_EXEC([bgp-daemon], [ip link set up bgp-daemon]) +NS_EXEC([ext-foo], [ip link set down ext-foo]) +NS_EXEC([ext-foo], [ip link set up ext-foo]) + +# Wait until IPv6 LLA loses the "tentative" flag otherwise it can't be bound to. +OVS_WAIT_UNTIL([NS_EXEC([bgp-daemon], [ip a show dev bgp-daemon | grep "fe80::" | grep -v tentative])]) +OVS_WAIT_UNTIL([NS_EXEC([ext-foo], [ip a show dev ext-foo | grep "fe80::" | grep -v tentative])]) + +# Verify that BGP control plane traffic is delivered to the "bgp-daemon" +# interface on both IPv4 and IPv6 LLA addresses +NETNS_DAEMONIZE([bgp-daemon], [nc -l -k 172.16.1.1 179], [bgp_v4.pid]) +NS_CHECK_EXEC([ext-foo], [echo "TCP test" | nc --send-only 172.16.1.1 179]) + +NETNS_DAEMONIZE([bgp-daemon], [nc -l -6 -k fe80::200:2ff:fe01:203%bgp-daemon 179], [bgp_v6.pid]) +NS_CHECK_EXEC([ext-foo], [echo "TCP test" | nc --send-only -6 fe80::200:2ff:fe01:203%ext-foo 179]) + +# Verify connection in other direction. i.e when daemon running on "bgp-daemon" port +# makes a client connection to its peer +NETNS_DAEMONIZE([ext-foo], [nc -l -k 172.16.1.100 179], [reply_bgp_v4.pid]) +NS_CHECK_EXEC([bgp-daemon], [echo "TCP test" | nc --send-only 172.16.1.100 179]) + +NETNS_DAEMONIZE([ext-foo], [nc -l -6 -k fe80::210:10ff:fe01:213%ext-foo 179], [reply_bgp_v6.pid]) +NS_CHECK_EXEC([bgp-daemon], [echo "TCP test" | nc --send-only -6 fe80::210:10ff:fe01:213%bgp-daemon 179]) + + +OVS_APP_EXIT_AND_WAIT([ovn-controller]) + +as ovn-sb +OVS_APP_EXIT_AND_WAIT([ovsdb-server]) + +as ovn-nb +OVS_APP_EXIT_AND_WAIT([ovsdb-server]) + +as northd +OVS_APP_EXIT_AND_WAIT([ovn-northd]) + +as +OVS_TRAFFIC_VSWITCHD_STOP(["/.*error receiving.*/d +/.*terminating with signal 15.*/d"]) +AT_CLEANUP +])
This change adds two new LRP options: * routing-protocol-redirect * routing-protocols These allow redirection of a routing protocol traffic to an Logical Switch Port. This enables external routing daemons to listen on an interface bound to an LSP and effectively act as if they were listening on (and speaking from) LRP's IP address. Option 'routing-protocols' takes a comma-separated list of routing protocols whose traffic should be redirected. Currently supported are BGP (tcp 179) and BFD (udp 3784). Option 'routing-protocol-redirect' expects a string with an LSP name. When both of these options are set, any traffic entering LS that's destined for LRP's IP addresses (including IPv6 LLA) and routing protocol's port number, is redirected to the LSP specified in the 'routing-protocol-redirect' value. NOTE: this feature is experimental and may be subject to removal/change in the future. Signed-off-by: Martin Kalcok <martin.kalcok@canonical.com> --- v5 includes: * change from pure BGP redirect to a configurable protocol redirection * fixes issue with local routing daemon not being able to receive reply traffic when contacting its peer. * ARP and ND traffic is cloned to the redirected port, to properly populate information about its neighbors. northd/northd.c | 217 ++++++++++++++++++++++++++++++++++++++++ northd/northd.h | 7 ++ northd/ovn-northd.8.xml | 54 ++++++++++ ovn-nb.xml | 42 ++++++++ tests/ovn-northd.at | 93 +++++++++++++++++ tests/system-ovn.at | 100 ++++++++++++++++++ 6 files changed, 513 insertions(+)