From patchwork Fri Jan 5 03:23:17 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Numan Siddique X-Patchwork-Id: 1882707 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@legolas.ozlabs.org Authentication-Results: legolas.ozlabs.org; spf=pass (sender SPF authorized) smtp.mailfrom=openvswitch.org (client-ip=140.211.166.138; helo=smtp1.osuosl.org; envelope-from=ovs-dev-bounces@openvswitch.org; receiver=patchwork.ozlabs.org) Received: from smtp1.osuosl.org (smtp1.osuosl.org [140.211.166.138]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature ECDSA (secp384r1) server-digest SHA384) (No client certificate requested) by legolas.ozlabs.org (Postfix) with ESMTPS id 4T5phY0dWcz1ydd for ; Fri, 5 Jan 2024 14:23:57 +1100 (AEDT) Received: from localhost (localhost [127.0.0.1]) by smtp1.osuosl.org (Postfix) with ESMTP id 274B684660; Fri, 5 Jan 2024 03:23:55 +0000 (UTC) DKIM-Filter: OpenDKIM Filter v2.11.0 smtp1.osuosl.org 274B684660 X-Virus-Scanned: amavisd-new at osuosl.org Received: from smtp1.osuosl.org ([127.0.0.1]) by localhost (smtp1.osuosl.org [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id xHCg9ILtSYG9; Fri, 5 Jan 2024 03:23:51 +0000 (UTC) Received: from lists.linuxfoundation.org (lf-lists.osuosl.org [140.211.9.56]) by smtp1.osuosl.org (Postfix) with ESMTPS id 70CC580F27; Fri, 5 Jan 2024 03:23:50 +0000 (UTC) DKIM-Filter: OpenDKIM Filter v2.11.0 smtp1.osuosl.org 70CC580F27 Received: from lf-lists.osuosl.org (localhost [127.0.0.1]) by lists.linuxfoundation.org (Postfix) with ESMTP id 4DBE8C0077; Fri, 5 Jan 2024 03:23:50 +0000 (UTC) X-Original-To: dev@openvswitch.org Delivered-To: ovs-dev@lists.linuxfoundation.org Received: from smtp3.osuosl.org (smtp3.osuosl.org [IPv6:2605:bc80:3010::136]) by lists.linuxfoundation.org (Postfix) with ESMTP id 12E06C0077 for ; Fri, 5 Jan 2024 03:23:49 +0000 (UTC) Received: from localhost (localhost [127.0.0.1]) by smtp3.osuosl.org (Postfix) with ESMTP id 22BF5614F0 for ; Fri, 5 Jan 2024 03:23:37 +0000 (UTC) DKIM-Filter: OpenDKIM Filter v2.11.0 smtp3.osuosl.org 22BF5614F0 X-Virus-Scanned: amavisd-new at osuosl.org Received: from smtp3.osuosl.org ([127.0.0.1]) by localhost (smtp3.osuosl.org [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id n4X3RGygXR7s for ; Fri, 5 Jan 2024 03:23:34 +0000 (UTC) Received: from relay8-d.mail.gandi.net (relay8-d.mail.gandi.net [217.70.183.201]) by smtp3.osuosl.org (Postfix) with ESMTPS id 9190F614B5 for ; Fri, 5 Jan 2024 03:23:33 +0000 (UTC) DKIM-Filter: OpenDKIM Filter v2.11.0 smtp3.osuosl.org 9190F614B5 Received: by mail.gandi.net (Postfix) with ESMTPSA id 50C3D1BF204; Fri, 5 Jan 2024 03:23:28 +0000 (UTC) From: numans@ovn.org To: dev@openvswitch.org Date: Thu, 4 Jan 2024 22:23:17 -0500 Message-ID: <20240105032317.766887-1-numans@ovn.org> X-Mailer: git-send-email 2.43.0 In-Reply-To: <20240105032000.766473-1-numans@ovn.org> References: <20240105032000.766473-1-numans@ovn.org> MIME-Version: 1.0 X-GND-Sasl: numans@ovn.org Subject: [ovs-dev] [PATCH ovn v4 09/16] northd: Use lflow_ref when adding all logical flows. X-BeenThere: ovs-dev@openvswitch.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: ovs-dev-bounces@openvswitch.org Sender: "dev" From: Numan Siddique ovn_lflow_add* macros now expect 'lflow_ref' to be passed and it can be NULL too. References to the logical flows generated for router ports are now stored in op->lflow_ref and op->stateful_lflow_ref. Signed-off-by: Numan Siddique --- northd/lflow-mgr.h | 38 +- northd/northd.c | 1614 +++++++++++++++++++++++++++----------------- 2 files changed, 991 insertions(+), 661 deletions(-) diff --git a/northd/lflow-mgr.h b/northd/lflow-mgr.h index f9511df58b..ddeba9a354 100644 --- a/northd/lflow-mgr.h +++ b/northd/lflow-mgr.h @@ -88,41 +88,27 @@ void lflow_table_add_lflow_default_drop(struct lflow_table *, /* Adds a row with the specified contents to the Logical_Flow table. */ #define ovn_lflow_add_with_hint__(LFLOW_TABLE, OD, STAGE, PRIORITY, MATCH, \ ACTIONS, IN_OUT_PORT, CTRL_METER, \ - STAGE_HINT) \ - lflow_table_add_lflow(LFLOW_TABLE, OD, NULL, 0, STAGE, PRIORITY, MATCH, \ - ACTIONS, IN_OUT_PORT, CTRL_METER, STAGE_HINT, \ - OVS_SOURCE_LOCATOR, NULL) - -#define ovn_lflow_add_with_lflow_ref_hint__(LFLOW_TABLE, OD, STAGE, PRIORITY, \ - MATCH, ACTIONS, IN_OUT_PORT, \ - CTRL_METER, STAGE_HINT, LFLOW_REF)\ + STAGE_HINT, LFLOW_REF) \ lflow_table_add_lflow(LFLOW_TABLE, OD, NULL, 0, STAGE, PRIORITY, MATCH, \ ACTIONS, IN_OUT_PORT, CTRL_METER, STAGE_HINT, \ OVS_SOURCE_LOCATOR, LFLOW_REF) #define ovn_lflow_add_with_hint(LFLOW_TABLE, OD, STAGE, PRIORITY, MATCH, \ - ACTIONS, STAGE_HINT) \ - lflow_table_add_lflow(LFLOW_TABLE, OD, NULL, 0, STAGE, PRIORITY, MATCH, \ - ACTIONS, NULL, NULL, STAGE_HINT, \ - OVS_SOURCE_LOCATOR, NULL) - -#define ovn_lflow_add_with_lflow_ref_hint(LFLOW_TABLE, OD, STAGE, PRIORITY, \ - MATCH, ACTIONS, STAGE_HINT, \ - LFLOW_REF) \ + ACTIONS, STAGE_HINT, LFLOW_REF) \ lflow_table_add_lflow(LFLOW_TABLE, OD, NULL, 0, STAGE, PRIORITY, MATCH, \ ACTIONS, NULL, NULL, STAGE_HINT, \ OVS_SOURCE_LOCATOR, LFLOW_REF) #define ovn_lflow_add_with_dp_group(LFLOW_TABLE, DP_BITMAP, DP_BITMAP_LEN, \ STAGE, PRIORITY, MATCH, ACTIONS, \ - STAGE_HINT) \ + STAGE_HINT, LFLOW_REF) \ lflow_table_add_lflow(LFLOW_TABLE, NULL, DP_BITMAP, DP_BITMAP_LEN, STAGE, \ PRIORITY, MATCH, ACTIONS, NULL, NULL, STAGE_HINT, \ - OVS_SOURCE_LOCATOR, NULL) + OVS_SOURCE_LOCATOR, LFLOW_REF) -#define ovn_lflow_add_default_drop(LFLOW_TABLE, OD, STAGE) \ +#define ovn_lflow_add_default_drop(LFLOW_TABLE, OD, STAGE, LFLOW_REF) \ lflow_table_add_lflow_default_drop(LFLOW_TABLE, OD, STAGE, \ - OVS_SOURCE_LOCATOR, NULL) + OVS_SOURCE_LOCATOR, LFLOW_REF) /* This macro is similar to ovn_lflow_add_with_hint, except that it requires @@ -142,20 +128,16 @@ void lflow_table_add_lflow_default_drop(struct lflow_table *, ACTIONS, IN_OUT_PORT, NULL, STAGE_HINT, \ OVS_SOURCE_LOCATOR, LFLOW_REF) -#define ovn_lflow_add(LFLOW_TABLE, OD, STAGE, PRIORITY, MATCH, ACTIONS) \ - lflow_table_add_lflow(LFLOW_TABLE, OD, NULL, 0, STAGE, PRIORITY, MATCH, \ - ACTIONS, NULL, NULL, NULL, OVS_SOURCE_LOCATOR, NULL) - -#define ovn_lflow_add_with_lflow_ref(LFLOW_TABLE, OD, STAGE, PRIORITY, MATCH, \ - ACTIONS, LFLOW_REF) \ +#define ovn_lflow_add(LFLOW_TABLE, OD, STAGE, PRIORITY, MATCH, ACTIONS, \ + LFLOW_REF) \ lflow_table_add_lflow(LFLOW_TABLE, OD, NULL, 0, STAGE, PRIORITY, MATCH, \ ACTIONS, NULL, NULL, NULL, OVS_SOURCE_LOCATOR, \ LFLOW_REF) #define ovn_lflow_metered(LFLOW_TABLE, OD, STAGE, PRIORITY, MATCH, ACTIONS, \ - CTRL_METER) \ + CTRL_METER, LFLOW_REF) \ ovn_lflow_add_with_hint__(LFLOW_TABLE, OD, STAGE, PRIORITY, MATCH, \ - ACTIONS, NULL, CTRL_METER, NULL) + ACTIONS, NULL, CTRL_METER, NULL, LFLOW_REF) struct sbrec_logical_dp_group; diff --git a/northd/northd.c b/northd/northd.c index 3576199f4f..ace31ab806 100644 --- a/northd/northd.c +++ b/northd/northd.c @@ -6062,13 +6062,16 @@ build_lswitch_learn_fdb_op( static void build_lswitch_learn_fdb_od( - struct ovn_datapath *od, struct lflow_table *lflows) + struct ovn_datapath *od, struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbs); - ovn_lflow_add(lflows, od, S_SWITCH_IN_LOOKUP_FDB, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_SWITCH_IN_PUT_FDB, 0, "1", "next;"); + ovn_lflow_add(lflows, od, S_SWITCH_IN_LOOKUP_FDB, 0, "1", "next;", + lflow_ref); + ovn_lflow_add(lflows, od, S_SWITCH_IN_PUT_FDB, 0, "1", "next;", + lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_IN_L2_LKUP, 0, "1", - "outport = get_fdb(eth.dst); next;"); + "outport = get_fdb(eth.dst); next;", lflow_ref); } /* Egress tables 8: Egress port security - IP (priority 0) @@ -6076,25 +6079,30 @@ build_lswitch_learn_fdb_od( * (priority 100). */ static void build_lswitch_output_port_sec_od(struct ovn_datapath *od, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbs); ovn_lflow_add(lflows, od, S_SWITCH_OUT_CHECK_PORT_SEC, 100, - "eth.mcast", REGBIT_PORT_SEC_DROP" = 0; next;"); + "eth.mcast", REGBIT_PORT_SEC_DROP" = 0; next;", + lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_OUT_CHECK_PORT_SEC, 0, "1", - REGBIT_PORT_SEC_DROP" = check_out_port_sec(); next;"); + REGBIT_PORT_SEC_DROP" = check_out_port_sec(); next;", + lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_OUT_APPLY_PORT_SEC, 50, - REGBIT_PORT_SEC_DROP" == 1", debug_drop_action()); + REGBIT_PORT_SEC_DROP" == 1", debug_drop_action(), + lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_OUT_APPLY_PORT_SEC, 0, - "1", "output;"); + "1", "output;", lflow_ref); } static void skip_port_from_conntrack(const struct ovn_datapath *od, struct ovn_port *op, bool has_stateful_acl, enum ovn_stage in_stage, enum ovn_stage out_stage, uint16_t priority, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { /* Can't use ct() for router ports. Consider the following configuration: * lp1(10.0.0.2) on hostA--ls1--lr0--ls2--lp2(10.0.1.2) on hostB, For a @@ -6116,10 +6124,10 @@ skip_port_from_conntrack(const struct ovn_datapath *od, struct ovn_port *op, ovn_lflow_add_with_lport_and_hint(lflows, od, in_stage, priority, ingress_match, ingress_action, - op->key, &op->nbsp->header_, NULL); + op->key, &op->nbsp->header_, lflow_ref); ovn_lflow_add_with_lport_and_hint(lflows, od, out_stage, priority, egress_match, egress_action, - op->key, &op->nbsp->header_, NULL); + op->key, &op->nbsp->header_, lflow_ref); free(ingress_match); free(egress_match); @@ -6128,7 +6136,8 @@ skip_port_from_conntrack(const struct ovn_datapath *od, struct ovn_port *op, static void build_stateless_filter(const struct ovn_datapath *od, const struct nbrec_acl *acl, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { const char *action = REGBIT_ACL_STATELESS" = 1; next;"; if (!strcmp(acl->direction, "from-lport")) { @@ -6136,25 +6145,28 @@ build_stateless_filter(const struct ovn_datapath *od, acl->priority + OVN_ACL_PRI_OFFSET, acl->match, action, - &acl->header_); + &acl->header_, + lflow_ref); } else { ovn_lflow_add_with_hint(lflows, od, S_SWITCH_OUT_PRE_ACL, acl->priority + OVN_ACL_PRI_OFFSET, acl->match, action, - &acl->header_); + &acl->header_, + lflow_ref); } } static void build_stateless_filters(const struct ovn_datapath *od, const struct ls_port_group_table *ls_port_groups, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { for (size_t i = 0; i < od->nbs->n_acls; i++) { const struct nbrec_acl *acl = od->nbs->acls[i]; if (!strcmp(acl->action, "allow-stateless")) { - build_stateless_filter(od, acl, lflows); + build_stateless_filter(od, acl, lflows, lflow_ref); } } @@ -6170,32 +6182,38 @@ build_stateless_filters(const struct ovn_datapath *od, const struct nbrec_acl *acl = ls_pg_rec->nb_pg->acls[i]; if (!strcmp(acl->action, "allow-stateless")) { - build_stateless_filter(od, acl, lflows); + build_stateless_filter(od, acl, lflows, lflow_ref); } } } } static void -build_pre_acls(struct ovn_datapath *od, struct lflow_table *lflows) +build_pre_acls(struct ovn_datapath *od, struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { /* Ingress and Egress Pre-ACL Table (Priority 0): Packets are * allowed by default. */ - ovn_lflow_add(lflows, od, S_SWITCH_IN_PRE_ACL, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_SWITCH_OUT_PRE_ACL, 0, "1", "next;"); + ovn_lflow_add(lflows, od, S_SWITCH_IN_PRE_ACL, 0, "1", "next;", + lflow_ref); + ovn_lflow_add(lflows, od, S_SWITCH_OUT_PRE_ACL, 0, "1", "next;", + lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_IN_PRE_ACL, 110, - "eth.dst == $svc_monitor_mac", "next;"); + "eth.dst == $svc_monitor_mac", "next;", + lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_OUT_PRE_ACL, 110, - "eth.src == $svc_monitor_mac", "next;"); + "eth.src == $svc_monitor_mac", "next;", + lflow_ref); } static void build_ls_stateful_rec_pre_acls( const struct ls_stateful_record *ls_stateful_rec, const struct ls_port_group_table *ls_port_groups, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { const struct ovn_datapath *od = ls_stateful_rec->od; @@ -6206,17 +6224,17 @@ build_ls_stateful_rec_pre_acls( for (size_t i = 0; i < od->n_router_ports; i++) { skip_port_from_conntrack(od, od->router_ports[i], true, S_SWITCH_IN_PRE_ACL, S_SWITCH_OUT_PRE_ACL, - 110, lflows); + 110, lflows, lflow_ref); } for (size_t i = 0; i < od->n_localnet_ports; i++) { skip_port_from_conntrack(od, od->localnet_ports[i], true, S_SWITCH_IN_PRE_ACL, S_SWITCH_OUT_PRE_ACL, - 110, lflows); + 110, lflows, lflow_ref); } /* stateless filters always take precedence over stateful ACLs. */ - build_stateless_filters(od, ls_port_groups, lflows); + build_stateless_filters(od, ls_port_groups, lflows, lflow_ref); /* Ingress and Egress Pre-ACL Table (Priority 110). * @@ -6224,16 +6242,18 @@ build_ls_stateful_rec_pre_acls( * unreachable packets. */ ovn_lflow_add(lflows, od, S_SWITCH_IN_PRE_ACL, 110, "nd || nd_rs || nd_ra || mldv1 || mldv2 || " - "(udp && udp.src == 546 && udp.dst == 547)", "next;"); + "(udp && udp.src == 546 && udp.dst == 547)", "next;", + lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_OUT_PRE_ACL, 110, "nd || nd_rs || nd_ra || mldv1 || mldv2 || " - "(udp && udp.src == 546 && udp.dst == 547)", "next;"); + "(udp && udp.src == 546 && udp.dst == 547)", "next;", + lflow_ref); /* Do not send multicast packets to conntrack. */ ovn_lflow_add(lflows, od, S_SWITCH_IN_PRE_ACL, 110, "eth.mcast", - "next;"); + "next;", lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_OUT_PRE_ACL, 110, "eth.mcast", - "next;"); + "next;", lflow_ref); /* Ingress and Egress Pre-ACL Table (Priority 100). * @@ -6244,13 +6264,15 @@ build_ls_stateful_rec_pre_acls( * 'REGBIT_CONNTRACK_DEFRAG' is set to let the pre-stateful table send * it to conntrack for tracking and defragmentation. */ ovn_lflow_add(lflows, od, S_SWITCH_IN_PRE_ACL, 100, "ip", - REGBIT_CONNTRACK_DEFRAG" = 1; next;"); + REGBIT_CONNTRACK_DEFRAG" = 1; next;", + lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_OUT_PRE_ACL, 100, "ip", - REGBIT_CONNTRACK_DEFRAG" = 1; next;"); + REGBIT_CONNTRACK_DEFRAG" = 1; next;", + lflow_ref); } else if (ls_stateful_rec->has_lb_vip) { /* We'll build stateless filters if there are LB rules so that * the stateless flows are not tracked in pre-lb. */ - build_stateless_filters(od, ls_port_groups, lflows); + build_stateless_filters(od, ls_port_groups, lflows, lflow_ref); } } @@ -6314,7 +6336,8 @@ build_empty_lb_event_flow(struct ovn_lb_vip *lb_vip, static void build_interconn_mcast_snoop_flows(struct ovn_datapath *od, const struct shash *meter_groups, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { struct mcast_switch_info *mcast_sw_info = &od->mcast_info.sw; if (!mcast_sw_info->enabled @@ -6333,7 +6356,8 @@ build_interconn_mcast_snoop_flows(struct ovn_datapath *od, ovn_lflow_metered(lflows, od, S_SWITCH_OUT_PRE_LB, 120, match, "clone { igmp; }; next;", copp_meter_get(COPP_IGMP, od->nbs->copp, - meter_groups)); + meter_groups), + lflow_ref); free(match); /* Punt MLD traffic to controller. */ @@ -6341,50 +6365,54 @@ build_interconn_mcast_snoop_flows(struct ovn_datapath *od, ovn_lflow_metered(lflows, od, S_SWITCH_OUT_PRE_LB, 120, match, "clone { igmp; }; next;", copp_meter_get(COPP_IGMP, od->nbs->copp, - meter_groups)); + meter_groups), + lflow_ref); free(match); } } static void build_pre_lb(struct ovn_datapath *od, const struct shash *meter_groups, - struct lflow_table *lflows) + struct lflow_table *lflows, struct lflow_ref *lflow_ref) { /* Handle IGMP/MLD packets crossing AZs. */ - build_interconn_mcast_snoop_flows(od, meter_groups, lflows); + build_interconn_mcast_snoop_flows(od, meter_groups, lflows, lflow_ref); /* Do not send multicast packets to conntrack */ - ovn_lflow_add(lflows, od, S_SWITCH_IN_PRE_LB, 110, "eth.mcast", "next;"); - ovn_lflow_add(lflows, od, S_SWITCH_OUT_PRE_LB, 110, "eth.mcast", "next;"); + ovn_lflow_add(lflows, od, S_SWITCH_IN_PRE_LB, 110, "eth.mcast", "next;", + lflow_ref); + ovn_lflow_add(lflows, od, S_SWITCH_OUT_PRE_LB, 110, "eth.mcast", "next;", + lflow_ref); /* Do not send ND packets to conntrack */ ovn_lflow_add(lflows, od, S_SWITCH_IN_PRE_LB, 110, "nd || nd_rs || nd_ra || mldv1 || mldv2", - "next;"); + "next;", lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_OUT_PRE_LB, 110, "nd || nd_rs || nd_ra || mldv1 || mldv2", - "next;"); + "next;", lflow_ref); /* Do not send service monitor packets to conntrack. */ ovn_lflow_add(lflows, od, S_SWITCH_IN_PRE_LB, 110, - "eth.dst == $svc_monitor_mac", "next;"); + "eth.dst == $svc_monitor_mac", "next;", lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_OUT_PRE_LB, 110, - "eth.src == $svc_monitor_mac", "next;"); + "eth.src == $svc_monitor_mac", "next;", lflow_ref); /* Allow all packets to go to next tables by default. */ - ovn_lflow_add(lflows, od, S_SWITCH_IN_PRE_LB, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_SWITCH_OUT_PRE_LB, 0, "1", "next;"); + ovn_lflow_add(lflows, od, S_SWITCH_IN_PRE_LB, 0, "1", "next;", lflow_ref); + ovn_lflow_add(lflows, od, S_SWITCH_OUT_PRE_LB, 0, "1", "next;", lflow_ref); /* Do not send statless flows via conntrack */ ovn_lflow_add(lflows, od, S_SWITCH_IN_PRE_LB, 110, - REGBIT_ACL_STATELESS" == 1", "next;"); + REGBIT_ACL_STATELESS" == 1", "next;", lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_OUT_PRE_LB, 110, - REGBIT_ACL_STATELESS" == 1", "next;"); + REGBIT_ACL_STATELESS" == 1", "next;", lflow_ref); } static void build_ls_stateful_rec_pre_lb(const struct ls_stateful_record *ls_stateful_rec, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { const struct ovn_datapath *od = ls_stateful_rec->od; @@ -6392,7 +6420,7 @@ build_ls_stateful_rec_pre_lb(const struct ls_stateful_record *ls_stateful_rec, skip_port_from_conntrack(od, od->router_ports[i], ls_stateful_rec->has_stateful_acl, S_SWITCH_IN_PRE_LB, S_SWITCH_OUT_PRE_LB, - 110, lflows); + 110, lflows, lflow_ref); } /* Localnet ports have no need for going through conntrack, unless @@ -6405,7 +6433,7 @@ build_ls_stateful_rec_pre_lb(const struct ls_stateful_record *ls_stateful_rec, skip_port_from_conntrack(od, od->localnet_ports[i], ls_stateful_rec->has_stateful_acl, S_SWITCH_IN_PRE_LB, S_SWITCH_OUT_PRE_LB, - 110, lflows); + 110, lflows, lflow_ref); } } @@ -6441,21 +6469,26 @@ build_ls_stateful_rec_pre_lb(const struct ls_stateful_record *ls_stateful_rec, */ if (ls_stateful_rec->has_lb_vip) { ovn_lflow_add(lflows, od, S_SWITCH_IN_PRE_LB, - 100, "ip", REGBIT_CONNTRACK_NAT" = 1; next;"); + 100, "ip", REGBIT_CONNTRACK_NAT" = 1; next;", + lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_OUT_PRE_LB, - 100, "ip", REGBIT_CONNTRACK_NAT" = 1; next;"); + 100, "ip", REGBIT_CONNTRACK_NAT" = 1; next;", + lflow_ref); } } static void build_pre_stateful(struct ovn_datapath *od, const struct chassis_features *features, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { /* Ingress and Egress pre-stateful Table (Priority 0): Packets are * allowed by default. */ - ovn_lflow_add(lflows, od, S_SWITCH_IN_PRE_STATEFUL, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_SWITCH_OUT_PRE_STATEFUL, 0, "1", "next;"); + ovn_lflow_add(lflows, od, S_SWITCH_IN_PRE_STATEFUL, 0, "1", "next;", + lflow_ref); + ovn_lflow_add(lflows, od, S_SWITCH_OUT_PRE_STATEFUL, 0, "1", "next;", + lflow_ref); /* Note: priority-120 flows are added in build_lb_rules_pre_stateful(). */ @@ -6464,25 +6497,30 @@ build_pre_stateful(struct ovn_datapath *od, : "ct_lb;"; ovn_lflow_add(lflows, od, S_SWITCH_IN_PRE_STATEFUL, 110, - REGBIT_CONNTRACK_NAT" == 1", ct_lb_action); + REGBIT_CONNTRACK_NAT" == 1", ct_lb_action, + lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_OUT_PRE_STATEFUL, 110, - REGBIT_CONNTRACK_NAT" == 1", ct_lb_action); + REGBIT_CONNTRACK_NAT" == 1", ct_lb_action, + lflow_ref); /* If REGBIT_CONNTRACK_DEFRAG is set as 1, then the packets should be * sent to conntrack for tracking and defragmentation. */ ovn_lflow_add(lflows, od, S_SWITCH_IN_PRE_STATEFUL, 100, - REGBIT_CONNTRACK_DEFRAG" == 1", "ct_next;"); + REGBIT_CONNTRACK_DEFRAG" == 1", "ct_next;", + lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_OUT_PRE_STATEFUL, 100, - REGBIT_CONNTRACK_DEFRAG" == 1", "ct_next;"); + REGBIT_CONNTRACK_DEFRAG" == 1", "ct_next;", + lflow_ref); } static void build_acl_hints(const struct ls_stateful_record *ls_stateful_rec, const struct chassis_features *features, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { const struct ovn_datapath *od = ls_stateful_rec->od; @@ -6510,9 +6548,10 @@ build_acl_hints(const struct ls_stateful_record *ls_stateful_rec, /* In any case, advance to the next stage. */ if (!ls_stateful_rec->has_acls && !ls_stateful_rec->has_lb_vip) { - ovn_lflow_add(lflows, od, stage, UINT16_MAX, "1", "next;"); + ovn_lflow_add(lflows, od, stage, UINT16_MAX, "1", "next;", + lflow_ref); } else { - ovn_lflow_add(lflows, od, stage, 0, "1", "next;"); + ovn_lflow_add(lflows, od, stage, 0, "1", "next;", lflow_ref); } if (!ls_stateful_rec->has_stateful_acl && @@ -6527,7 +6566,7 @@ build_acl_hints(const struct ls_stateful_record *ls_stateful_rec, ovn_lflow_add(lflows, od, stage, 7, "ct.new && !ct.est", REGBIT_ACL_HINT_ALLOW_NEW " = 1; " REGBIT_ACL_HINT_DROP " = 1; " - "next;"); + "next;", lflow_ref); /* Already established connections in the "request" direction that * are already marked as "blocked" may hit either: @@ -6543,13 +6582,13 @@ build_acl_hints(const struct ls_stateful_record *ls_stateful_rec, ovn_lflow_add(lflows, od, stage, 6, match, REGBIT_ACL_HINT_ALLOW_NEW " = 1; " REGBIT_ACL_HINT_DROP " = 1; " - "next;"); + "next;", lflow_ref); /* Not tracked traffic can either be allowed or dropped. */ ovn_lflow_add(lflows, od, stage, 5, "!ct.trk", REGBIT_ACL_HINT_ALLOW " = 1; " REGBIT_ACL_HINT_DROP " = 1; " - "next;"); + "next;", lflow_ref); /* Already established connections in the "request" direction may hit * either: @@ -6565,20 +6604,20 @@ build_acl_hints(const struct ls_stateful_record *ls_stateful_rec, ovn_lflow_add(lflows, od, stage, 4, match, REGBIT_ACL_HINT_ALLOW " = 1; " REGBIT_ACL_HINT_BLOCK " = 1; " - "next;"); + "next;", lflow_ref); /* Not established or established and already blocked connections may * hit drop ACLs. */ ovn_lflow_add(lflows, od, stage, 3, "!ct.est", REGBIT_ACL_HINT_DROP " = 1; " - "next;"); + "next;", lflow_ref); match = features->ct_no_masked_label ? "ct.est && ct_mark.blocked == 1" : "ct.est && ct_label.blocked == 1"; ovn_lflow_add(lflows, od, stage, 2, match, REGBIT_ACL_HINT_DROP " = 1; " - "next;"); + "next;", lflow_ref); /* Established connections that were previously allowed might hit * drop ACLs in which case the connection must be committed with @@ -6589,7 +6628,7 @@ build_acl_hints(const struct ls_stateful_record *ls_stateful_rec, : "ct.est && ct_label.blocked == 0"; ovn_lflow_add(lflows, od, stage, 1, match, REGBIT_ACL_HINT_BLOCK " = 1; " - "next;"); + "next;", lflow_ref); } } @@ -6655,7 +6694,8 @@ static void consider_acl(struct lflow_table *lflows, const struct ovn_datapath *od, const struct nbrec_acl *acl, bool has_stateful, bool ct_masked_mark, const struct shash *meter_groups, - uint64_t max_acl_tier, struct ds *match, struct ds *actions) + uint64_t max_acl_tier, struct ds *match, struct ds *actions, + struct lflow_ref *lflow_ref) { const char *ct_blocked_match = ct_masked_mark ? "ct_mark.blocked" @@ -6704,7 +6744,7 @@ consider_acl(struct lflow_table *lflows, const struct ovn_datapath *od, ds_put_format(match, "(%s)", acl->match); ovn_lflow_add_with_hint(lflows, od, stage, priority, ds_cstr(match), ds_cstr(actions), - &acl->header_); + &acl->header_, lflow_ref); return; } @@ -6741,7 +6781,7 @@ consider_acl(struct lflow_table *lflows, const struct ovn_datapath *od, ds_put_cstr(actions, "next;"); ovn_lflow_add_with_hint(lflows, od, stage, priority, ds_cstr(match), ds_cstr(actions), - &acl->header_); + &acl->header_, lflow_ref); /* Match on traffic in the request direction for an established * connection tracking entry that has not been marked for @@ -6763,7 +6803,7 @@ consider_acl(struct lflow_table *lflows, const struct ovn_datapath *od, ds_put_cstr(actions, "next;"); ovn_lflow_add_with_hint(lflows, od, stage, priority, ds_cstr(match), ds_cstr(actions), - &acl->header_); + &acl->header_, lflow_ref); } else if (!strcmp(acl->action, "drop") || !strcmp(acl->action, "reject")) { /* The implementation of "drop" differs if stateful ACLs are in @@ -6780,7 +6820,7 @@ consider_acl(struct lflow_table *lflows, const struct ovn_datapath *od, ds_put_cstr(actions, "next;"); ovn_lflow_add_with_hint(lflows, od, stage, priority, ds_cstr(match), ds_cstr(actions), - &acl->header_); + &acl->header_, lflow_ref); /* For an existing connection without ct_mark.blocked set, we've * encountered a policy change. ACLs previously allowed * this connection and we committed the connection tracking @@ -6801,7 +6841,7 @@ consider_acl(struct lflow_table *lflows, const struct ovn_datapath *od, ct_blocked_match); ovn_lflow_add_with_hint(lflows, od, stage, priority, ds_cstr(match), ds_cstr(actions), - &acl->header_); + &acl->header_, lflow_ref); } } @@ -6884,7 +6924,8 @@ build_acl_action_lflows(const struct ls_stateful_record *ls_stateful_rec, const char *default_acl_action, const struct shash *meter_groups, struct ds *match, - struct ds *actions) + struct ds *actions, + struct lflow_ref *lflow_ref) { const struct ovn_datapath *od = ls_stateful_rec->od; @@ -6907,18 +6948,20 @@ build_acl_action_lflows(const struct ls_stateful_record *ls_stateful_rec, for (size_t i = 0; i < ARRAY_SIZE(stages); i++) { enum ovn_stage stage = stages[i]; if (!ls_stateful_rec->has_acls) { - ovn_lflow_add(lflows, od, stage, 0, "1", "next;"); + ovn_lflow_add(lflows, od, stage, 0, "1", "next;", lflow_ref); continue; } ds_truncate(actions, verdict_len); ds_put_cstr(actions, "next;"); ovn_lflow_add(lflows, od, stage, 1000, - REGBIT_ACL_VERDICT_ALLOW " == 1", ds_cstr(actions)); + REGBIT_ACL_VERDICT_ALLOW " == 1", ds_cstr(actions), + lflow_ref); ds_truncate(actions, verdict_len); ds_put_cstr(actions, debug_implicit_drop_action()); ovn_lflow_add(lflows, od, stage, 1000, REGBIT_ACL_VERDICT_DROP " == 1", - ds_cstr(actions)); + ds_cstr(actions), + lflow_ref); bool ingress = ovn_stage_get_pipeline(stage) == P_IN; ds_truncate(actions, verdict_len); @@ -6934,11 +6977,11 @@ build_acl_action_lflows(const struct ls_stateful_record *ls_stateful_rec, ovn_lflow_metered(lflows, od, stage, 1000, REGBIT_ACL_VERDICT_REJECT " == 1", ds_cstr(actions), copp_meter_get(COPP_REJECT, od->nbs->copp, - meter_groups)); + meter_groups), lflow_ref); ds_truncate(actions, verdict_len); ds_put_cstr(actions, default_acl_action); - ovn_lflow_add(lflows, od, stage, 0, "1", ds_cstr(actions)); + ovn_lflow_add(lflows, od, stage, 0, "1", ds_cstr(actions), lflow_ref); struct ds tier_actions = DS_EMPTY_INITIALIZER; for (size_t j = 0; j < ls_stateful_rec->max_acl_tier; j++) { @@ -6950,7 +6993,7 @@ build_acl_action_lflows(const struct ls_stateful_record *ls_stateful_rec, j + 1, ingress ? "ingress" : "egress", ovn_stage_get_table(stage) - 1); ovn_lflow_add(lflows, od, stage, 500, ds_cstr(match), - ds_cstr(&tier_actions)); + ds_cstr(&tier_actions), lflow_ref); } ds_destroy(&tier_actions); } @@ -6962,7 +7005,8 @@ build_acl_log_related_flows(const struct ovn_datapath *od, const struct nbrec_acl *acl, bool has_stateful, bool ct_masked_mark, const struct shash *meter_groups, - struct ds *match, struct ds *actions) + struct ds *match, struct ds *actions, + struct lflow_ref *lflow_ref) { /* Related and reply traffic are universally allowed by priority * 65532 flows created in build_acls(). If logging is enabled on @@ -7016,7 +7060,7 @@ build_acl_log_related_flows(const struct ovn_datapath *od, ovn_lflow_add_with_hint(lflows, od, log_related_stage, UINT16_MAX - 2, ds_cstr(match), ds_cstr(actions), - &acl->header_); + &acl->header_, lflow_ref); ds_clear(match); ds_put_format(match, "!ct.est && ct.rel && !ct.new%s && " @@ -7027,7 +7071,7 @@ build_acl_log_related_flows(const struct ovn_datapath *od, ovn_lflow_add_with_hint(lflows, od, log_related_stage, UINT16_MAX - 2, ds_cstr(match), ds_cstr(actions), - &acl->header_); + &acl->header_, lflow_ref); } static void @@ -7035,7 +7079,8 @@ build_acls(const struct ls_stateful_record *ls_stateful_rec, const struct chassis_features *features, struct lflow_table *lflows, const struct ls_port_group_table *ls_port_groups, - const struct shash *meter_groups) + const struct shash *meter_groups, + struct lflow_ref *lflow_ref) { const struct ovn_datapath *od = ls_stateful_rec->od; @@ -7060,20 +7105,24 @@ build_acls(const struct ls_stateful_record *ls_stateful_rec, if (!ls_stateful_rec->has_acls) { if (!ls_stateful_rec->has_lb_vip) { ovn_lflow_add(lflows, od, S_SWITCH_IN_ACL_EVAL, UINT16_MAX, "1", - "next;"); + "next;", lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_OUT_ACL_EVAL, UINT16_MAX, "1", - "next;"); + "next;", lflow_ref); } else { - ovn_lflow_add(lflows, od, S_SWITCH_IN_ACL_EVAL, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_SWITCH_OUT_ACL_EVAL, 0, "1", "next;"); + ovn_lflow_add(lflows, od, S_SWITCH_IN_ACL_EVAL, 0, "1", "next;", + lflow_ref); + ovn_lflow_add(lflows, od, S_SWITCH_OUT_ACL_EVAL, 0, "1", "next;", + lflow_ref); } ovn_lflow_add(lflows, od, S_SWITCH_IN_ACL_AFTER_LB_EVAL, 0, "1", - "next;"); + "next;", lflow_ref); } else { - ovn_lflow_add(lflows, od, S_SWITCH_IN_ACL_EVAL, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_SWITCH_OUT_ACL_EVAL, 0, "1", "next;"); + ovn_lflow_add(lflows, od, S_SWITCH_IN_ACL_EVAL, 0, "1", "next;", + lflow_ref); + ovn_lflow_add(lflows, od, S_SWITCH_OUT_ACL_EVAL, 0, "1", "next;", + lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_IN_ACL_AFTER_LB_EVAL, 0, "1", - "next;"); + "next;", lflow_ref); } @@ -7105,19 +7154,21 @@ build_acls(const struct ls_stateful_record *ls_stateful_rec, ovn_lflow_add(lflows, od, S_SWITCH_IN_ACL_EVAL, 1, ds_cstr(&match), REGBIT_CONNTRACK_COMMIT" = 1; " - REGBIT_ACL_VERDICT_ALLOW" = 1; next;"); + REGBIT_ACL_VERDICT_ALLOW" = 1; next;", + lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_OUT_ACL_EVAL, 1, ds_cstr(&match), REGBIT_CONNTRACK_COMMIT" = 1; " - REGBIT_ACL_VERDICT_ALLOW" = 1; next;"); + REGBIT_ACL_VERDICT_ALLOW" = 1; next;", + lflow_ref); const char *next_action = default_acl_drop ? "next;" : REGBIT_CONNTRACK_COMMIT" = 1; next;"; ovn_lflow_add(lflows, od, S_SWITCH_IN_ACL_EVAL, 1, "ip && !ct.est", - next_action); + next_action, lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_OUT_ACL_EVAL, 1, "ip && !ct.est", - next_action); + next_action, lflow_ref); /* Ingress and Egress ACL Table (Priority 65532). * @@ -7131,9 +7182,11 @@ build_acls(const struct ls_stateful_record *ls_stateful_rec, use_ct_inv_match ? "ct.inv || " : "", ct_blocked_match); ovn_lflow_add(lflows, od, S_SWITCH_IN_ACL_EVAL, UINT16_MAX - 3, - ds_cstr(&match), REGBIT_ACL_VERDICT_DROP " = 1; next;"); + ds_cstr(&match), REGBIT_ACL_VERDICT_DROP " = 1; next;", + lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_OUT_ACL_EVAL, UINT16_MAX - 3, - ds_cstr(&match), REGBIT_ACL_VERDICT_DROP " = 1; next;"); + ds_cstr(&match), REGBIT_ACL_VERDICT_DROP " = 1; next;", + lflow_ref); /* Ingress and Egress ACL Table (Priority 65535 - 3). * @@ -7153,10 +7206,12 @@ build_acls(const struct ls_stateful_record *ls_stateful_rec, ds_cstr(&match), REGBIT_ACL_HINT_DROP" = 0; " REGBIT_ACL_HINT_BLOCK" = 0; " REGBIT_ACL_HINT_ALLOW_REL" = 1; " - REGBIT_ACL_VERDICT_ALLOW" = 1; next;"); + REGBIT_ACL_VERDICT_ALLOW" = 1; next;", + lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_OUT_ACL_EVAL, UINT16_MAX - 3, ds_cstr(&match), - REGBIT_ACL_VERDICT_ALLOW " = 1; next;"); + REGBIT_ACL_VERDICT_ALLOW " = 1; next;", + lflow_ref); /* Ingress and Egress ACL Table (Priority 65535). * @@ -7185,15 +7240,16 @@ build_acls(const struct ls_stateful_record *ls_stateful_rec, use_ct_inv_match ? " && !ct.inv" : "", ct_blocked_match); ovn_lflow_add(lflows, od, S_SWITCH_IN_ACL_EVAL, UINT16_MAX - 3, - ds_cstr(&match), ct_in_acl_action); + ds_cstr(&match), ct_in_acl_action, lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_OUT_ACL_EVAL, UINT16_MAX - 3, - ds_cstr(&match), ct_out_acl_action); + ds_cstr(&match), ct_out_acl_action, lflow_ref); /* Reply and related traffic matched by an "allow-related" ACL * should be allowed in the ls_in_acl_after_lb stage too. */ ovn_lflow_add(lflows, od, S_SWITCH_IN_ACL_AFTER_LB_EVAL, UINT16_MAX - 3, REGBIT_ACL_HINT_ALLOW_REL" == 1", - REGBIT_ACL_VERDICT_ALLOW " = 1; next;"); + REGBIT_ACL_VERDICT_ALLOW " = 1; next;", + lflow_ref); } /* Ingress and Egress ACL Table (Priority 65532). @@ -7205,24 +7261,28 @@ build_acls(const struct ls_stateful_record *ls_stateful_rec, * https://bugzilla.kernel.org/show_bug.cgi?id=11797. */ ovn_lflow_add(lflows, od, S_SWITCH_IN_ACL_EVAL, UINT16_MAX - 3, IPV6_CT_OMIT_MATCH, - REGBIT_ACL_VERDICT_ALLOW " = 1; next;"); + REGBIT_ACL_VERDICT_ALLOW " = 1; next;", + lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_OUT_ACL_EVAL, UINT16_MAX - 3, IPV6_CT_OMIT_MATCH, - REGBIT_ACL_VERDICT_ALLOW " = 1; next;"); + REGBIT_ACL_VERDICT_ALLOW " = 1; next;", + lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_IN_ACL_AFTER_LB_EVAL, UINT16_MAX - 3, IPV6_CT_OMIT_MATCH, - REGBIT_ACL_VERDICT_ALLOW " = 1; next;"); + REGBIT_ACL_VERDICT_ALLOW " = 1; next;", + lflow_ref); /* Ingress or Egress ACL Table (Various priorities). */ for (size_t i = 0; i < od->nbs->n_acls; i++) { struct nbrec_acl *acl = od->nbs->acls[i]; build_acl_log_related_flows(od, lflows, acl, has_stateful, features->ct_no_masked_label, - meter_groups, &match, &actions); + meter_groups, &match, &actions, + lflow_ref); consider_acl(lflows, od, acl, has_stateful, features->ct_no_masked_label, meter_groups, ls_stateful_rec->max_acl_tier, - &match, &actions); + &match, &actions, lflow_ref); } const struct ls_port_group *ls_pg = @@ -7235,11 +7295,12 @@ build_acls(const struct ls_stateful_record *ls_stateful_rec, build_acl_log_related_flows(od, lflows, acl, has_stateful, features->ct_no_masked_label, - meter_groups, &match, &actions); + meter_groups, &match, &actions, + lflow_ref); consider_acl(lflows, od, acl, has_stateful, features->ct_no_masked_label, meter_groups, ls_stateful_rec->max_acl_tier, - &match, &actions); + &match, &actions, lflow_ref); } } } @@ -7254,7 +7315,7 @@ build_acls(const struct ls_stateful_record *ls_stateful_rec, : REGBIT_ACL_VERDICT_ALLOW" = 1; next;"; ovn_lflow_add( lflows, od, S_SWITCH_OUT_ACL_EVAL, 34000, "udp.src == 53", - dns_actions); + dns_actions, lflow_ref); } if (ls_stateful_rec->has_acls || ls_stateful_rec->has_lb_vip) { @@ -7262,30 +7323,37 @@ build_acls(const struct ls_stateful_record *ls_stateful_rec, * packets to skip applying ingress ACLs. */ ovn_lflow_add(lflows, od, S_SWITCH_IN_ACL_EVAL, 34000, "eth.dst == $svc_monitor_mac", - REGBIT_ACL_VERDICT_ALLOW" = 1; next;"); + REGBIT_ACL_VERDICT_ALLOW" = 1; next;", + lflow_ref); /* Add a 34000 priority flow to advance the service monitor packets * generated by ovn-controller to skip applying egress ACLs. */ ovn_lflow_add(lflows, od, S_SWITCH_OUT_ACL_EVAL, 34000, "eth.src == $svc_monitor_mac", - REGBIT_ACL_VERDICT_ALLOW" = 1; next;"); + REGBIT_ACL_VERDICT_ALLOW" = 1; next;", + lflow_ref); } build_acl_action_lflows(ls_stateful_rec, lflows, default_acl_action, - meter_groups, &match, &actions); + meter_groups, &match, &actions, lflow_ref); ds_destroy(&match); ds_destroy(&actions); } static void -build_qos(struct ovn_datapath *od, struct lflow_table *lflows) { +build_qos(struct ovn_datapath *od, struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { struct ds action = DS_EMPTY_INITIALIZER; - ovn_lflow_add(lflows, od, S_SWITCH_IN_QOS_MARK, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_SWITCH_OUT_QOS_MARK, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_SWITCH_IN_QOS_METER, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_SWITCH_OUT_QOS_METER, 0, "1", "next;"); + ovn_lflow_add(lflows, od, S_SWITCH_IN_QOS_MARK, 0, "1", "next;", + lflow_ref); + ovn_lflow_add(lflows, od, S_SWITCH_OUT_QOS_MARK, 0, "1", "next;", + lflow_ref); + ovn_lflow_add(lflows, od, S_SWITCH_IN_QOS_METER, 0, "1", "next;", + lflow_ref); + ovn_lflow_add(lflows, od, S_SWITCH_OUT_QOS_METER, 0, "1", "next;", + lflow_ref); for (size_t i = 0; i < od->nbs->n_qos_rules; i++) { struct nbrec_qos *qos = od->nbs->qos_rules[i]; @@ -7302,7 +7370,7 @@ build_qos(struct ovn_datapath *od, struct lflow_table *lflows) { ovn_lflow_add_with_hint(lflows, od, stage, qos->priority, qos->match, ds_cstr(&action), - &qos->header_); + &qos->header_, lflow_ref); } } @@ -7333,7 +7401,7 @@ build_qos(struct ovn_datapath *od, struct lflow_table *lflows) { ovn_lflow_add_with_hint(lflows, od, stage, qos->priority, qos->match, ds_cstr(&action), - &qos->header_); + &qos->header_, lflow_ref); } } ds_destroy(&action); @@ -7398,7 +7466,7 @@ build_lb_rules_pre_stateful(struct lflow_table *lflows, ovn_lflow_add_with_dp_group( lflows, lb_dps->nb_ls_map, ods_size(ls_datapaths), S_SWITCH_IN_PRE_STATEFUL, 120, ds_cstr(match), ds_cstr(action), - &lb->nlb->header_); + &lb->nlb->header_, NULL); } } @@ -7446,7 +7514,8 @@ build_lb_affinity_lr_flows(struct lflow_table *lflows, const struct ovn_northd_lb *lb, struct ovn_lb_vip *lb_vip, char *new_lb_match, char *lb_action, const unsigned long *dp_bitmap, - const struct ovn_datapaths *lr_datapaths) + const struct ovn_datapaths *lr_datapaths, + struct lflow_ref *lflow_ref) { if (!lb->affinity_timeout || bitmap_is_all_zeros(dp_bitmap, ods_size(lr_datapaths))) { @@ -7485,7 +7554,8 @@ build_lb_affinity_lr_flows(struct lflow_table *lflows, ovn_lflow_add_with_dp_group( lflows, dp_bitmap, ods_size(lr_datapaths), S_ROUTER_IN_LB_AFF_CHECK, - 100, new_lb_match, ds_cstr(&aff_check_action), &lb->nlb->header_); + 100, new_lb_match, ds_cstr(&aff_check_action), &lb->nlb->header_, + lflow_ref); /* Prepare common part of affinity LB and affinity learn action. */ ds_put_format(&aff_action, "%s = %s; ", reg_vip, lb_vip->vip_str); @@ -7567,12 +7637,14 @@ build_lb_affinity_lr_flows(struct lflow_table *lflows, ovn_lflow_add_with_dp_group( lflows, dp_bitmap, ods_size(lr_datapaths), S_ROUTER_IN_LB_AFF_LEARN, 100, ds_cstr(&aff_match_learn), - ds_cstr(&aff_action_learn), &lb->nlb->header_); + ds_cstr(&aff_action_learn), &lb->nlb->header_, + lflow_ref); /* Use already selected backend within affinity timeslot. */ ovn_lflow_add_with_dp_group( lflows, dp_bitmap, ods_size(lr_datapaths), S_ROUTER_IN_DNAT, 150, - ds_cstr(&aff_match), ds_cstr(&aff_action), &lb->nlb->header_); + ds_cstr(&aff_match), ds_cstr(&aff_action), &lb->nlb->header_, + lflow_ref); ds_truncate(&aff_action, aff_action_len); ds_truncate(&aff_action_learn, aff_action_learn_len); @@ -7632,7 +7704,8 @@ static void build_lb_affinity_ls_flows(struct lflow_table *lflows, struct ovn_lb_datapaths *lb_dps, struct ovn_lb_vip *lb_vip, - const struct ovn_datapaths *ls_datapaths) + const struct ovn_datapaths *ls_datapaths, + struct lflow_ref *lflow_ref) { if (!lb_dps->lb->affinity_timeout || !lb_dps->n_nb_ls) { return; @@ -7660,7 +7733,7 @@ build_lb_affinity_ls_flows(struct lflow_table *lflows, ovn_lflow_add_with_dp_group( lflows, lb_dps->nb_ls_map, ods_size(ls_datapaths), S_SWITCH_IN_LB_AFF_CHECK, 100, ds_cstr(&new_lb_match), aff_check, - &lb_dps->lb->nlb->header_); + &lb_dps->lb->nlb->header_, lflow_ref); ds_destroy(&new_lb_match); struct ds aff_action = DS_EMPTY_INITIALIZER; @@ -7750,13 +7823,14 @@ build_lb_affinity_ls_flows(struct lflow_table *lflows, ovn_lflow_add_with_dp_group( lflows, lb_dps->nb_ls_map, ods_size(ls_datapaths), S_SWITCH_IN_LB_AFF_LEARN, 100, ds_cstr(&aff_match_learn), - ds_cstr(&aff_action_learn), &lb->nlb->header_); + ds_cstr(&aff_action_learn), &lb->nlb->header_, + lflow_ref); /* Use already selected backend within affinity timeslot. */ ovn_lflow_add_with_dp_group( lflows, lb_dps->nb_ls_map, ods_size(ls_datapaths), S_SWITCH_IN_LB, 150, ds_cstr(&aff_match), ds_cstr(&aff_action), - &lb->nlb->header_); + &lb->nlb->header_, lflow_ref); ds_truncate(&aff_action, aff_action_len); ds_truncate(&aff_action_learn, aff_action_learn_len); @@ -7772,20 +7846,26 @@ build_lb_affinity_ls_flows(struct lflow_table *lflows, static void build_lswitch_lb_affinity_default_flows(struct ovn_datapath *od, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbs); - ovn_lflow_add(lflows, od, S_SWITCH_IN_LB_AFF_CHECK, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_SWITCH_IN_LB_AFF_LEARN, 0, "1", "next;"); + ovn_lflow_add(lflows, od, S_SWITCH_IN_LB_AFF_CHECK, 0, "1", "next;", + lflow_ref); + ovn_lflow_add(lflows, od, S_SWITCH_IN_LB_AFF_LEARN, 0, "1", "next;", + lflow_ref); } static void build_lrouter_lb_affinity_default_flows(struct ovn_datapath *od, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbr); - ovn_lflow_add(lflows, od, S_ROUTER_IN_LB_AFF_CHECK, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_ROUTER_IN_LB_AFF_LEARN, 0, "1", "next;"); + ovn_lflow_add(lflows, od, S_ROUTER_IN_LB_AFF_CHECK, 0, "1", "next;", + lflow_ref); + ovn_lflow_add(lflows, od, S_ROUTER_IN_LB_AFF_LEARN, 0, "1", "next;", + lflow_ref); } static void @@ -7830,7 +7910,8 @@ build_lb_rules(struct lflow_table *lflows, struct ovn_lb_datapaths *lb_dps, priority = 120; } - build_lb_affinity_ls_flows(lflows, lb_dps, lb_vip, ls_datapaths); + build_lb_affinity_ls_flows(lflows, lb_dps, lb_vip, ls_datapaths, + NULL); unsigned long *dp_non_meter = NULL; bool build_non_meter = false; @@ -7853,14 +7934,16 @@ build_lb_rules(struct lflow_table *lflows, struct ovn_lb_datapaths *lb_dps, ovn_lflow_add_with_hint__( lflows, od, S_SWITCH_IN_LB, priority, ds_cstr(match), ds_cstr(action), - NULL, meter, &lb->nlb->header_); + NULL, meter, &lb->nlb->header_, + NULL); } } if (!reject || build_non_meter) { ovn_lflow_add_with_dp_group( lflows, dp_non_meter ? dp_non_meter : lb_dps->nb_ls_map, ods_size(ls_datapaths), S_SWITCH_IN_LB, priority, - ds_cstr(match), ds_cstr(action), &lb->nlb->header_); + ds_cstr(match), ds_cstr(action), &lb->nlb->header_, + NULL); } bitmap_free(dp_non_meter); } @@ -7869,7 +7952,8 @@ build_lb_rules(struct lflow_table *lflows, struct ovn_lb_datapaths *lb_dps, static void build_stateful(struct ovn_datapath *od, const struct chassis_features *features, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { const char *ct_block_action = features->ct_no_masked_label ? "ct_mark.blocked" @@ -7878,9 +7962,11 @@ build_stateful(struct ovn_datapath *od, /* Ingress LB, Ingress and Egress stateful Table (Priority 0): Packets are * allowed by default. */ - ovn_lflow_add(lflows, od, S_SWITCH_IN_LB, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_SWITCH_IN_STATEFUL, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_SWITCH_OUT_STATEFUL, 0, "1", "next;"); + ovn_lflow_add(lflows, od, S_SWITCH_IN_LB, 0, "1", "next;", lflow_ref); + ovn_lflow_add(lflows, od, S_SWITCH_IN_STATEFUL, 0, "1", "next;", + lflow_ref); + ovn_lflow_add(lflows, od, S_SWITCH_OUT_STATEFUL, 0, "1", "next;", + lflow_ref); /* If REGBIT_CONNTRACK_COMMIT is set as 1 and * REGBIT_CONNTRACK_SET_LABEL is set to 1, then the packets should be @@ -7894,11 +7980,13 @@ build_stateful(struct ovn_datapath *od, ovn_lflow_add(lflows, od, S_SWITCH_IN_STATEFUL, 100, REGBIT_CONNTRACK_COMMIT" == 1 && " REGBIT_ACL_LABEL" == 1", - ds_cstr(&actions)); + ds_cstr(&actions), + lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_OUT_STATEFUL, 100, REGBIT_CONNTRACK_COMMIT" == 1 && " REGBIT_ACL_LABEL" == 1", - ds_cstr(&actions)); + ds_cstr(&actions), + lflow_ref); /* If REGBIT_CONNTRACK_COMMIT is set as 1, then the packets should be * committed to conntrack. We always set ct_mark.blocked to 0 here as @@ -7909,26 +7997,32 @@ build_stateful(struct ovn_datapath *od, ovn_lflow_add(lflows, od, S_SWITCH_IN_STATEFUL, 100, REGBIT_CONNTRACK_COMMIT" == 1 && " REGBIT_ACL_LABEL" == 0", - ds_cstr(&actions)); + ds_cstr(&actions), + lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_OUT_STATEFUL, 100, REGBIT_CONNTRACK_COMMIT" == 1 && " REGBIT_ACL_LABEL" == 0", - ds_cstr(&actions)); + ds_cstr(&actions), + lflow_ref); ds_destroy(&actions); } static void build_lb_hairpin(const struct ls_stateful_record *ls_stateful_rec, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { const struct ovn_datapath *od = ls_stateful_rec->od; /* Ingress Pre-Hairpin/Nat-Hairpin/Hairpin tabled (Priority 0). * Packets that don't need hairpinning should continue processing. */ - ovn_lflow_add(lflows, od, S_SWITCH_IN_PRE_HAIRPIN, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_SWITCH_IN_NAT_HAIRPIN, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_SWITCH_IN_HAIRPIN, 0, "1", "next;"); + ovn_lflow_add(lflows, od, S_SWITCH_IN_PRE_HAIRPIN, 0, "1", "next;", + lflow_ref); + ovn_lflow_add(lflows, od, S_SWITCH_IN_NAT_HAIRPIN, 0, "1", "next;", + lflow_ref); + ovn_lflow_add(lflows, od, S_SWITCH_IN_HAIRPIN, 0, "1", "next;", + lflow_ref); if (ls_stateful_rec->has_lb_vip) { /* Check if the packet needs to be hairpinned. @@ -7940,7 +8034,8 @@ build_lb_hairpin(const struct ls_stateful_record *ls_stateful_rec, REGBIT_HAIRPIN " = chk_lb_hairpin(); " REGBIT_HAIRPIN_REPLY " = chk_lb_hairpin_reply(); " "next;", - &od->nbs->header_); + &od->nbs->header_, + lflow_ref); /* If packet needs to be hairpinned, snat the src ip with the VIP * for new sessions. */ @@ -7948,7 +8043,8 @@ build_lb_hairpin(const struct ls_stateful_record *ls_stateful_rec, "ip && ct.new && ct.trk" " && "REGBIT_HAIRPIN " == 1", "ct_snat_to_vip; next;", - &od->nbs->header_); + &od->nbs->header_, + lflow_ref); /* If packet needs to be hairpinned, for established sessions there * should already be an SNAT conntrack entry. @@ -7957,13 +8053,15 @@ build_lb_hairpin(const struct ls_stateful_record *ls_stateful_rec, "ip && ct.est && ct.trk" " && "REGBIT_HAIRPIN " == 1", "ct_snat;", - &od->nbs->header_); + &od->nbs->header_, + lflow_ref); /* For the reply of hairpinned traffic, snat the src ip to the VIP. */ ovn_lflow_add_with_hint(lflows, od, S_SWITCH_IN_NAT_HAIRPIN, 90, "ip && "REGBIT_HAIRPIN_REPLY " == 1", "ct_snat;", - &od->nbs->header_); + &od->nbs->header_, + lflow_ref); /* Ingress Hairpin table. * - Priority 1: Packets that were SNAT-ed for hairpinning should be @@ -7973,12 +8071,13 @@ build_lb_hairpin(const struct ls_stateful_record *ls_stateful_rec, lflows, od, S_SWITCH_IN_HAIRPIN, 1, "("REGBIT_HAIRPIN " == 1 || " REGBIT_HAIRPIN_REPLY " == 1)", "eth.dst <-> eth.src; outport = inport; flags.loopback = 1; " - "output;"); + "output;", lflow_ref); } } static void -build_vtep_hairpin(struct ovn_datapath *od, struct lflow_table *lflows) +build_vtep_hairpin(struct ovn_datapath *od, struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { if (!od->has_vtep_lports) { /* There is no need in these flows if datapath has no vtep lports. */ @@ -7991,7 +8090,7 @@ build_vtep_hairpin(struct ovn_datapath *od, struct lflow_table *lflows) char *action = xasprintf("next(pipeline=ingress, table=%d);", ovn_stage_get_table(S_SWITCH_IN_L2_LKUP)); ovn_lflow_add(lflows, od, S_SWITCH_IN_HAIRPIN, 1000, - REGBIT_FROM_RAMP" == 1", action); + REGBIT_FROM_RAMP" == 1", action, lflow_ref); free(action); /* Ingress pre-arp flow for traffic from VTEP (ramp) switch. @@ -8008,7 +8107,7 @@ build_vtep_hairpin(struct ovn_datapath *od, struct lflow_table *lflows) REGBIT_FROM_RAMP" == 1 && is_chassis_resident(%s)", op->cr_port->json_key); ovn_lflow_add(lflows, od, S_SWITCH_IN_HAIRPIN, 2000, - ds_cstr(&match), "next;"); + ds_cstr(&match), "next;", lflow_ref); } } @@ -8019,14 +8118,15 @@ build_vtep_hairpin(struct ovn_datapath *od, struct lflow_table *lflows) */ ovn_lflow_add(lflows, od, S_SWITCH_IN_ARP_ND_RSP, 65535, REGBIT_FROM_RAMP" == 1 && (arp || nd_ns)", - "flags.loopback = 1; next;"); + "flags.loopback = 1; next;", lflow_ref); ds_destroy(&match); } /* Build logical flows for the forwarding groups */ static void -build_fwd_group_lflows(struct ovn_datapath *od, struct lflow_table *lflows) +build_fwd_group_lflows(struct ovn_datapath *od, struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbs); if (!od->nbs->n_forwarding_groups) { @@ -8061,7 +8161,7 @@ build_fwd_group_lflows(struct ovn_datapath *od, struct lflow_table *lflows) ovn_lflow_add_with_hint(lflows, od, S_SWITCH_IN_ARP_ND_RSP, 50, ds_cstr(&match), ds_cstr(&actions), - &fwd_group->header_); + &fwd_group->header_, lflow_ref); /* L2 lookup for the forwarding group's virtual MAC */ ds_clear(&match); @@ -8084,7 +8184,7 @@ build_fwd_group_lflows(struct ovn_datapath *od, struct lflow_table *lflows) ds_put_format(&actions, "fwd_group(%s);", ds_cstr(&group_ports)); ovn_lflow_add_with_hint(lflows, od, S_SWITCH_IN_L2_LKUP, 50, ds_cstr(&match), ds_cstr(&actions), - &fwd_group->header_); + &fwd_group->header_, lflow_ref); } ds_destroy(&match); @@ -8232,10 +8332,8 @@ build_lswitch_rport_arp_req_self_orig_flow(struct ovn_port *op, ds_put_format(&match, "eth.src == %s && (arp.op == 1 || rarp.op == 3 || nd_ns)", ds_cstr(ð_src)); - ovn_lflow_add_with_lflow_ref(lflows, od, S_SWITCH_IN_L2_LKUP, priority, - ds_cstr(&match), - "outport = \""MC_FLOOD_L2"\"; output;", - lflow_ref); + ovn_lflow_add(lflows, od, S_SWITCH_IN_L2_LKUP, priority, ds_cstr(&match), + "outport = \""MC_FLOOD_L2"\"; output;", lflow_ref); ds_destroy(ð_src); ds_destroy(&match); @@ -8318,17 +8416,17 @@ build_lswitch_rport_arp_req_flow( ds_put_format(&actions, "clone {outport = %s; output; }; " "outport = \""MC_FLOOD_L2"\"; output;", patch_op->json_key); - ovn_lflow_add_with_lflow_ref_hint(lflows, od, S_SWITCH_IN_L2_LKUP, - priority, ds_cstr(&match), - ds_cstr(&actions), stage_hint, - lflow_ref); + ovn_lflow_add_with_hint(lflows, od, S_SWITCH_IN_L2_LKUP, + priority, ds_cstr(&match), + ds_cstr(&actions), stage_hint, + lflow_ref); } else { ds_put_format(&actions, "outport = %s; output;", patch_op->json_key); - ovn_lflow_add_with_lflow_ref_hint(lflows, od, S_SWITCH_IN_L2_LKUP, - priority, ds_cstr(&match), - ds_cstr(&actions), - stage_hint, - lflow_ref); + ovn_lflow_add_with_hint(lflows, od, S_SWITCH_IN_L2_LKUP, + priority, ds_cstr(&match), + ds_cstr(&actions), + stage_hint, + lflow_ref); } ds_destroy(&match); @@ -8518,7 +8616,8 @@ build_dhcpv4_options_flows(struct ovn_port *op, struct lport_addresses *lsp_addrs, struct ovn_port *inport, bool is_external, const struct shash *meter_groups, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { struct ds match = DS_EMPTY_INITIALIZER; @@ -8549,7 +8648,7 @@ build_dhcpv4_options_flows(struct ovn_port *op, op->json_key); } - ovn_lflow_add_with_lflow_ref_hint__(lflows, op->od, + ovn_lflow_add_with_hint__(lflows, op->od, S_SWITCH_IN_DHCP_OPTIONS, 100, ds_cstr(&match), ds_cstr(&options_action), @@ -8558,7 +8657,7 @@ build_dhcpv4_options_flows(struct ovn_port *op, op->od->nbs->copp, meter_groups), &op->nbsp->dhcpv4_options->header_, - op->lflow_ref); + lflow_ref); ds_clear(&match); /* If REGBIT_DHCP_OPTS_RESULT is set, it means the @@ -8578,7 +8677,7 @@ build_dhcpv4_options_flows(struct ovn_port *op, lflows, op->od, S_SWITCH_IN_DHCP_RESPONSE, 100, ds_cstr(&match), ds_cstr(&response_action), inport->key, &op->nbsp->dhcpv4_options->header_, - op->lflow_ref); + lflow_ref); ds_destroy(&options_action); ds_destroy(&response_action); ds_destroy(&ipv4_addr_match); @@ -8606,7 +8705,7 @@ build_dhcpv4_options_flows(struct ovn_port *op, lflows, op->od, S_SWITCH_OUT_ACL_EVAL, 34000, ds_cstr(&match),dhcp_actions, op->key, &op->nbsp->dhcpv4_options->header_, - op->lflow_ref); + lflow_ref); } break; } @@ -8619,7 +8718,8 @@ build_dhcpv6_options_flows(struct ovn_port *op, struct lport_addresses *lsp_addrs, struct ovn_port *inport, bool is_external, const struct shash *meter_groups, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { struct ds match = DS_EMPTY_INITIALIZER; @@ -8641,7 +8741,7 @@ build_dhcpv6_options_flows(struct ovn_port *op, op->json_key); } - ovn_lflow_add_with_lflow_ref_hint__(lflows, op->od, + ovn_lflow_add_with_hint__(lflows, op->od, S_SWITCH_IN_DHCP_OPTIONS, 100, ds_cstr(&match), ds_cstr(&options_action), @@ -8650,7 +8750,7 @@ build_dhcpv6_options_flows(struct ovn_port *op, op->od->nbs->copp, meter_groups), &op->nbsp->dhcpv6_options->header_, - op->lflow_ref); + lflow_ref); /* If REGBIT_DHCP_OPTS_RESULT is set to 1, it means the * put_dhcpv6_opts action is successful */ @@ -8658,7 +8758,7 @@ build_dhcpv6_options_flows(struct ovn_port *op, ovn_lflow_add_with_lport_and_hint( lflows, op->od, S_SWITCH_IN_DHCP_RESPONSE, 100, ds_cstr(&match), ds_cstr(&response_action), inport->key, - &op->nbsp->dhcpv6_options->header_, op->lflow_ref); + &op->nbsp->dhcpv6_options->header_, lflow_ref); ds_destroy(&options_action); ds_destroy(&response_action); @@ -8691,7 +8791,7 @@ build_dhcpv6_options_flows(struct ovn_port *op, lflows, op->od, S_SWITCH_OUT_ACL_EVAL, 34000, ds_cstr(&match),dhcp6_actions, op->key, &op->nbsp->dhcpv6_options->header_, - op->lflow_ref); + lflow_ref); } break; } @@ -8702,7 +8802,8 @@ build_dhcpv6_options_flows(struct ovn_port *op, static void build_drop_arp_nd_flows_for_unbound_router_ports(struct ovn_port *op, const struct ovn_port *port, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { struct ds match = DS_EMPTY_INITIALIZER; @@ -8722,7 +8823,7 @@ build_drop_arp_nd_flows_for_unbound_router_ports(struct ovn_port *op, ovn_lflow_add_with_lport_and_hint( lflows, op->od, S_SWITCH_IN_EXTERNAL_PORT, 100, ds_cstr(&match), debug_drop_action(), port->key, - &op->nbsp->header_, op->lflow_ref); + &op->nbsp->header_, lflow_ref); } for (size_t l = 0; l < rp->lsp_addrs[k].n_ipv6_addrs; l++) { ds_clear(&match); @@ -8738,7 +8839,7 @@ build_drop_arp_nd_flows_for_unbound_router_ports(struct ovn_port *op, ovn_lflow_add_with_lport_and_hint( lflows, op->od, S_SWITCH_IN_EXTERNAL_PORT, 100, ds_cstr(&match), debug_drop_action(), port->key, - &op->nbsp->header_, op->lflow_ref); + &op->nbsp->header_, lflow_ref); } ds_clear(&match); @@ -8755,7 +8856,7 @@ build_drop_arp_nd_flows_for_unbound_router_ports(struct ovn_port *op, debug_drop_action(), port->key, &op->nbsp->header_, - op->lflow_ref); + lflow_ref); } } } @@ -8770,19 +8871,22 @@ is_vlan_transparent(const struct ovn_datapath *od) static void build_lswitch_lflows_l2_unknown(struct ovn_datapath *od, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { /* Ingress table 25/26: Destination lookup for unknown MACs. */ if (od->has_unknown) { ovn_lflow_add(lflows, od, S_SWITCH_IN_L2_UNKNOWN, 50, "outport == \"none\"", - "outport = \""MC_UNKNOWN "\"; output;"); + "outport = \""MC_UNKNOWN "\"; output;", + lflow_ref); } else { ovn_lflow_add(lflows, od, S_SWITCH_IN_L2_UNKNOWN, 50, - "outport == \"none\"", debug_drop_action()); + "outport == \"none\"", debug_drop_action(), + lflow_ref); } ovn_lflow_add(lflows, od, S_SWITCH_IN_L2_UNKNOWN, 0, "1", - "output;"); + "output;", lflow_ref); } /* Build pre-ACL and ACL tables for both ingress and egress. @@ -8792,42 +8896,49 @@ build_lswitch_lflows_pre_acl_and_acl( struct ovn_datapath *od, const struct chassis_features *features, struct lflow_table *lflows, - const struct shash *meter_groups) + const struct shash *meter_groups, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbs); - build_pre_acls(od, lflows); - build_pre_lb(od, meter_groups, lflows); - build_pre_stateful(od, features, lflows); - build_qos(od, lflows); - build_stateful(od, features, lflows); - build_vtep_hairpin(od, lflows); + build_pre_acls(od, lflows, lflow_ref); + build_pre_lb(od, meter_groups, lflows, lflow_ref); + build_pre_stateful(od, features, lflows, lflow_ref); + build_qos(od, lflows, lflow_ref); + build_stateful(od, features, lflows, lflow_ref); + build_vtep_hairpin(od, lflows, lflow_ref); } /* Logical switch ingress table 0: Admission control framework (priority * 100). */ static void build_lswitch_lflows_admission_control(struct ovn_datapath *od, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbs); /* Logical VLANs not supported. */ if (!is_vlan_transparent(od)) { /* Block logical VLANs. */ ovn_lflow_add(lflows, od, S_SWITCH_IN_CHECK_PORT_SEC, 100, - "vlan.present", debug_drop_action()); + "vlan.present", debug_drop_action(), + lflow_ref); } /* Broadcast/multicast source address is invalid. */ ovn_lflow_add(lflows, od, S_SWITCH_IN_CHECK_PORT_SEC, 100, - "eth.src[40]", debug_drop_action()); + "eth.src[40]", debug_drop_action(), + lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_IN_CHECK_PORT_SEC, 50, "1", - REGBIT_PORT_SEC_DROP" = check_in_port_sec(); next;"); + REGBIT_PORT_SEC_DROP" = check_in_port_sec(); next;", + lflow_ref); ovn_lflow_add(lflows, od, S_SWITCH_IN_APPLY_PORT_SEC, 50, - REGBIT_PORT_SEC_DROP" == 1", debug_drop_action()); + REGBIT_PORT_SEC_DROP" == 1", debug_drop_action(), + lflow_ref); - ovn_lflow_add(lflows, od, S_SWITCH_IN_APPLY_PORT_SEC, 0, "1", "next;"); + ovn_lflow_add(lflows, od, S_SWITCH_IN_APPLY_PORT_SEC, 0, "1", "next;", + lflow_ref); } /* Ingress table 19: ARP/ND responder, skip requests coming from localnet @@ -8984,12 +9095,12 @@ build_lswitch_arp_nd_responder_known_ips(struct ovn_port *op, "output;", op->lsp_addrs[i].ea_s, op->lsp_addrs[i].ea_s, op->lsp_addrs[i].ipv4_addrs[j].addr_s); - ovn_lflow_add_with_lflow_ref_hint(lflows, op->od, - S_SWITCH_IN_ARP_ND_RSP, 50, - ds_cstr(match), - ds_cstr(actions), - &op->nbsp->header_, - op->lflow_ref); + ovn_lflow_add_with_hint(lflows, op->od, + S_SWITCH_IN_ARP_ND_RSP, 50, + ds_cstr(match), + ds_cstr(actions), + &op->nbsp->header_, + op->lflow_ref); /* Do not reply to an ARP request from the port that owns * the address (otherwise a DHCP client that ARPs to check @@ -9039,16 +9150,16 @@ build_lswitch_arp_nd_responder_known_ips(struct ovn_port *op, op->lsp_addrs[i].ipv6_addrs[j].addr_s, op->lsp_addrs[i].ipv6_addrs[j].addr_s, op->lsp_addrs[i].ea_s); - ovn_lflow_add_with_lflow_ref_hint__(lflows, op->od, - S_SWITCH_IN_ARP_ND_RSP, 50, - ds_cstr(match), - ds_cstr(actions), - NULL, - copp_meter_get(COPP_ND_NA, - op->od->nbs->copp, - meter_groups), - &op->nbsp->header_, - op->lflow_ref); + ovn_lflow_add_with_hint__(lflows, op->od, + S_SWITCH_IN_ARP_ND_RSP, 50, + ds_cstr(match), + ds_cstr(actions), + NULL, + copp_meter_get(COPP_ND_NA, + op->od->nbs->copp, + meter_groups), + &op->nbsp->header_, + op->lflow_ref); /* Do not reply to a solicitation from the port that owns * the address (otherwise DAD detection will fail). */ @@ -9104,12 +9215,12 @@ build_lswitch_arp_nd_responder_known_ips(struct ovn_port *op, ea_s, ea_s); - ovn_lflow_add_with_lflow_ref_hint(lflows, op->od, - S_SWITCH_IN_ARP_ND_RSP, - 30, ds_cstr(match), - ds_cstr(actions), - &op->nbsp->header_, - op->lflow_ref); + ovn_lflow_add_with_hint(lflows, op->od, + S_SWITCH_IN_ARP_ND_RSP, + 30, ds_cstr(match), + ds_cstr(actions), + &op->nbsp->header_, + op->lflow_ref); } /* Add IPv6 NDP responses. @@ -9152,16 +9263,16 @@ build_lswitch_arp_nd_responder_known_ips(struct ovn_port *op, lsp_is_router(op->nbsp) ? "nd_na_router" : "nd_na", ea_s, ea_s); - ovn_lflow_add_with_lflow_ref_hint__(lflows, op->od, - S_SWITCH_IN_ARP_ND_RSP, 30, - ds_cstr(match), - ds_cstr(actions), - NULL, - copp_meter_get(COPP_ND_NA, - op->od->nbs->copp, - meter_groups), - &op->nbsp->header_, - op->lflow_ref); + ovn_lflow_add_with_hint__(lflows, op->od, + S_SWITCH_IN_ARP_ND_RSP, 30, + ds_cstr(match), + ds_cstr(actions), + NULL, + copp_meter_get(COPP_ND_NA, + op->od->nbs->copp, + meter_groups), + &op->nbsp->header_, + op->lflow_ref); ds_destroy(&ip6_dst_match); ds_destroy(&nd_target_match); } @@ -9172,21 +9283,24 @@ build_lswitch_arp_nd_responder_known_ips(struct ovn_port *op, * (priority 0)*/ static void build_lswitch_arp_nd_responder_default(struct ovn_datapath *od, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbs); - ovn_lflow_add(lflows, od, S_SWITCH_IN_ARP_ND_RSP, 0, "1", "next;"); + ovn_lflow_add(lflows, od, S_SWITCH_IN_ARP_ND_RSP, 0, "1", "next;", + lflow_ref); } /* Ingress table 19: ARP/ND responder for service monitor source ip. * (priority 110)*/ static void -build_lswitch_arp_nd_service_monitor(const struct ovn_northd_lb *lb, +build_lswitch_arp_nd_service_monitor(const struct ovn_lb_datapaths *lb_dps, const struct hmap *ls_ports, struct lflow_table *lflows, struct ds *actions, struct ds *match) { + const struct ovn_northd_lb *lb = lb_dps->lb; for (size_t i = 0; i < lb->n_vips; i++) { struct ovn_northd_lb_vip *lb_vip_nb = &lb->vips_nb[i]; if (!lb_vip_nb->lb_health_check) { @@ -9249,7 +9363,8 @@ build_lswitch_arp_nd_service_monitor(const struct ovn_northd_lb *lb, op->od, S_SWITCH_IN_ARP_ND_RSP, 110, ds_cstr(match), ds_cstr(actions), - &lb->nlb->header_); + &lb->nlb->header_, + NULL); } } } @@ -9289,19 +9404,19 @@ build_lswitch_dhcp_options_and_response(struct ovn_port *op, build_dhcpv4_options_flows( op, &op->lsp_addrs[i], op->od->localnet_ports[j], is_external, - meter_groups, lflows); + meter_groups, lflows, op->lflow_ref); build_dhcpv6_options_flows( op, &op->lsp_addrs[i], op->od->localnet_ports[j], is_external, - meter_groups, lflows); + meter_groups, lflows, op->lflow_ref); } } else { build_dhcpv4_options_flows(op, &op->lsp_addrs[i], op, is_external, meter_groups, - lflows); + lflows, op->lflow_ref); build_dhcpv6_options_flows(op, &op->lsp_addrs[i], op, is_external, meter_groups, - lflows); + lflows, op->lflow_ref); } } } @@ -9314,14 +9429,20 @@ build_lswitch_dhcp_options_and_response(struct ovn_port *op, * (priority 0). */ static void build_lswitch_dhcp_and_dns_defaults(struct ovn_datapath *od, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbs); - ovn_lflow_add(lflows, od, S_SWITCH_IN_DHCP_OPTIONS, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_SWITCH_IN_DHCP_RESPONSE, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_SWITCH_IN_DNS_LOOKUP, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_SWITCH_IN_DNS_RESPONSE, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_SWITCH_IN_EXTERNAL_PORT, 0, "1", "next;"); + ovn_lflow_add(lflows, od, S_SWITCH_IN_DHCP_OPTIONS, 0, "1", "next;", + lflow_ref); + ovn_lflow_add(lflows, od, S_SWITCH_IN_DHCP_RESPONSE, 0, "1", "next;", + lflow_ref); + ovn_lflow_add(lflows, od, S_SWITCH_IN_DNS_LOOKUP, 0, "1", "next;", + lflow_ref); + ovn_lflow_add(lflows, od, S_SWITCH_IN_DNS_RESPONSE, 0, "1", "next;", + lflow_ref); + ovn_lflow_add(lflows, od, S_SWITCH_IN_EXTERNAL_PORT, 0, "1", "next;", + lflow_ref); } /* Logical switch ingress table 22 and 23: DNS lookup and response @@ -9330,7 +9451,8 @@ build_lswitch_dhcp_and_dns_defaults(struct ovn_datapath *od, static void build_lswitch_dns_lookup_and_response(struct ovn_datapath *od, struct lflow_table *lflows, - const struct shash *meter_groups) + const struct shash *meter_groups, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbs); if (!ls_has_dns_records(od->nbs)) { @@ -9340,18 +9462,18 @@ build_lswitch_dns_lookup_and_response(struct ovn_datapath *od, "udp.dst == 53", REGBIT_DNS_LOOKUP_RESULT" = dns_lookup(); next;", copp_meter_get(COPP_DNS, od->nbs->copp, - meter_groups)); + meter_groups), lflow_ref); const char *dns_action = "eth.dst <-> eth.src; ip4.src <-> ip4.dst; " "udp.dst = udp.src; udp.src = 53; outport = inport; " "flags.loopback = 1; output;"; const char *dns_match = "udp.dst == 53 && "REGBIT_DNS_LOOKUP_RESULT; ovn_lflow_add(lflows, od, S_SWITCH_IN_DNS_RESPONSE, 100, - dns_match, dns_action); + dns_match, dns_action, lflow_ref); dns_action = "eth.dst <-> eth.src; ip6.src <-> ip6.dst; " "udp.dst = udp.src; udp.src = 53; outport = inport; " "flags.loopback = 1; output;"; ovn_lflow_add(lflows, od, S_SWITCH_IN_DNS_RESPONSE, 100, - dns_match, dns_action); + dns_match, dns_action, lflow_ref); } /* Table 24: External port. Drop ARP request for router ips from @@ -9368,7 +9490,8 @@ build_lswitch_external_port(struct ovn_port *op, } for (size_t i = 0; i < op->od->n_localnet_ports; i++) { build_drop_arp_nd_flows_for_unbound_router_ports( - op, op->od->localnet_ports[i], lflows); + op, op->od->localnet_ports[i], lflows, + op->lflow_ref); } } @@ -9378,7 +9501,8 @@ static void build_lswitch_destination_lookup_bmcast(struct ovn_datapath *od, struct lflow_table *lflows, struct ds *actions, - const struct shash *meter_groups) + const struct shash *meter_groups, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbs); @@ -9386,7 +9510,7 @@ build_lswitch_destination_lookup_bmcast(struct ovn_datapath *od, "eth.dst == $svc_monitor_mac && (tcp || icmp || icmp6)", "handle_svc_check(inport);", copp_meter_get(COPP_SVC_MONITOR, od->nbs->copp, - meter_groups)); + meter_groups), lflow_ref); struct mcast_switch_info *mcast_sw_info = &od->mcast_info.sw; @@ -9397,27 +9521,31 @@ build_lswitch_destination_lookup_bmcast(struct ovn_datapath *od, ovn_lflow_metered(lflows, od, S_SWITCH_IN_L2_LKUP, 100, "igmp", ds_cstr(actions), copp_meter_get(COPP_IGMP, od->nbs->copp, - meter_groups)); + meter_groups), + lflow_ref); /* Punt MLD traffic to controller. */ ovn_lflow_metered(lflows, od, S_SWITCH_IN_L2_LKUP, 100, "mldv1 || mldv2", ds_cstr(actions), copp_meter_get(COPP_IGMP, od->nbs->copp, - meter_groups)); + meter_groups), + lflow_ref); /* Flood all IP multicast traffic destined to 224.0.0.X to all * ports - RFC 4541, section 2.1.2, item 2. */ ovn_lflow_add(lflows, od, S_SWITCH_IN_L2_LKUP, 85, "ip4.mcast && ip4.dst == 224.0.0.0/24", - "outport = \""MC_FLOOD_L2"\"; output;"); + "outport = \""MC_FLOOD_L2"\"; output;", + lflow_ref); /* Flood all IPv6 multicast traffic destined to reserved * multicast IPs (RFC 4291, 2.7.1). */ ovn_lflow_add(lflows, od, S_SWITCH_IN_L2_LKUP, 85, "ip6.mcast_flood", - "outport = \""MC_FLOOD"\"; output;"); + "outport = \""MC_FLOOD"\"; output;", + lflow_ref); /* Forward uregistered IP multicast to routers with relay enabled * and to any ports configured to flood IP multicast traffic. @@ -9449,7 +9577,7 @@ build_lswitch_destination_lookup_bmcast(struct ovn_datapath *od, ovn_lflow_add(lflows, od, S_SWITCH_IN_L2_LKUP, 80, "ip4.mcast || ip6.mcast", - ds_cstr(actions)); + ds_cstr(actions), lflow_ref); } } @@ -9457,11 +9585,12 @@ build_lswitch_destination_lookup_bmcast(struct ovn_datapath *od, "broadcast-arps-to-all-routers", true)) { ovn_lflow_add(lflows, od, S_SWITCH_IN_L2_LKUP, 72, "eth.mcast && (arp.op == 1 || nd_ns)", - "outport = \""MC_FLOOD_L2"\"; output;"); + "outport = \""MC_FLOOD_L2"\"; output;", + lflow_ref); } ovn_lflow_add(lflows, od, S_SWITCH_IN_L2_LKUP, 70, "eth.mcast", - "outport = \""MC_FLOOD"\"; output;"); + "outport = \""MC_FLOOD"\"; output;", lflow_ref); } @@ -9543,7 +9672,7 @@ build_lswitch_ip_mcast_igmp_mld(struct ovn_igmp_group *igmp_group, igmp_group->mcgroup.name); ovn_lflow_add(lflows, igmp_group->datapath, S_SWITCH_IN_L2_LKUP, - 90, ds_cstr(match), ds_cstr(actions)); + 90, ds_cstr(match), ds_cstr(actions), NULL); } } @@ -9583,12 +9712,12 @@ build_lswitch_ip_unicast_lookup(struct ovn_port *op, ds_clear(actions); ds_put_format(actions, action, op->json_key); - ovn_lflow_add_with_lflow_ref_hint(lflows, op->od, - S_SWITCH_IN_L2_LKUP, - 50, ds_cstr(match), - ds_cstr(actions), - &op->nbsp->header_, - op->lflow_ref); + ovn_lflow_add_with_hint(lflows, op->od, + S_SWITCH_IN_L2_LKUP, + 50, ds_cstr(match), + ds_cstr(actions), + &op->nbsp->header_, + op->lflow_ref); } else if (!strcmp(op->nbsp->addresses[i], "unknown")) { continue; } else if (is_dynamic_lsp_address(op->nbsp->addresses[i])) { @@ -9603,12 +9732,12 @@ build_lswitch_ip_unicast_lookup(struct ovn_port *op, ds_clear(actions); ds_put_format(actions, action, op->json_key); - ovn_lflow_add_with_lflow_ref_hint(lflows, op->od, - S_SWITCH_IN_L2_LKUP, - 50, ds_cstr(match), - ds_cstr(actions), - &op->nbsp->header_, - op->lflow_ref); + ovn_lflow_add_with_hint(lflows, op->od, + S_SWITCH_IN_L2_LKUP, + 50, ds_cstr(match), + ds_cstr(actions), + &op->nbsp->header_, + op->lflow_ref); } else if (!strcmp(op->nbsp->addresses[i], "router")) { if (!op->peer || !op->peer->nbrp || !ovs_scan(op->peer->nbrp->mac, @@ -9660,11 +9789,11 @@ build_lswitch_ip_unicast_lookup(struct ovn_port *op, ds_clear(actions); ds_put_format(actions, action, op->json_key); - ovn_lflow_add_with_lflow_ref_hint(lflows, op->od, - S_SWITCH_IN_L2_LKUP, 50, - ds_cstr(match), ds_cstr(actions), - &op->nbsp->header_, - op->lflow_ref); + ovn_lflow_add_with_hint(lflows, op->od, + S_SWITCH_IN_L2_LKUP, 50, + ds_cstr(match), ds_cstr(actions), + &op->nbsp->header_, + op->lflow_ref); } else { static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, 1); @@ -9713,7 +9842,7 @@ build_lswitch_ip_unicast_lookup_for_nats( ds_clear(actions); ds_put_format(actions, action, op->json_key); - ovn_lflow_add_with_lflow_ref_hint(lflows, op->od, + ovn_lflow_add_with_hint(lflows, op->od, S_SWITCH_IN_L2_LKUP, 50, ds_cstr(match), ds_cstr(actions), @@ -9961,7 +10090,8 @@ static void build_routing_policy_flow(struct lflow_table *lflows, struct ovn_datapath *od, const struct hmap *lr_ports, const struct nbrec_logical_router_policy *rule, - const struct ovsdb_idl_row *stage_hint) + const struct ovsdb_idl_row *stage_hint, + struct lflow_ref *lflow_ref) { struct ds match = DS_EMPTY_INITIALIZER; struct ds actions = DS_EMPTY_INITIALIZER; @@ -10017,7 +10147,8 @@ build_routing_policy_flow(struct lflow_table *lflows, struct ovn_datapath *od, ds_put_format(&match, "%s", rule->match); ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_POLICY, rule->priority, - ds_cstr(&match), ds_cstr(&actions), stage_hint); + ds_cstr(&match), ds_cstr(&actions), stage_hint, + lflow_ref); ds_destroy(&match); ds_destroy(&actions); } @@ -10027,7 +10158,8 @@ build_ecmp_routing_policy_flows(struct lflow_table *lflows, struct ovn_datapath *od, const struct hmap *lr_ports, const struct nbrec_logical_router_policy *rule, - uint16_t ecmp_group_id) + uint16_t ecmp_group_id, + struct lflow_ref *lflow_ref) { ovs_assert(rule->n_nexthops > 1); @@ -10099,7 +10231,8 @@ build_ecmp_routing_policy_flows(struct lflow_table *lflows, ecmp_group_id, i + 1); ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_POLICY_ECMP, 100, ds_cstr(&match), - ds_cstr(&actions), &rule->header_); + ds_cstr(&actions), &rule->header_, + lflow_ref); } ds_clear(&actions); @@ -10117,7 +10250,8 @@ build_ecmp_routing_policy_flows(struct lflow_table *lflows, ds_put_cstr(&actions, ");"); ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_POLICY, rule->priority, rule->match, - ds_cstr(&actions), &rule->header_); + ds_cstr(&actions), &rule->header_, + lflow_ref); cleanup: ds_destroy(&match); @@ -10162,7 +10296,8 @@ get_route_table_id(struct simap *route_tables, const char *route_table_name) static void build_route_table_lflow(struct ovn_datapath *od, struct lflow_table *lflows, struct nbrec_logical_router_port *lrp, - struct simap *route_tables) + struct simap *route_tables, + struct lflow_ref *lflow_ref) { struct ds match = DS_EMPTY_INITIALIZER; struct ds actions = DS_EMPTY_INITIALIZER; @@ -10178,7 +10313,7 @@ build_route_table_lflow(struct ovn_datapath *od, struct lflow_table *lflows, REG_ROUTE_TABLE_ID, rtb_id); ovn_lflow_add(lflows, od, S_ROUTER_IN_IP_ROUTING_PRE, 100, - ds_cstr(&match), ds_cstr(&actions)); + ds_cstr(&match), ds_cstr(&actions), lflow_ref); ds_destroy(&match); ds_destroy(&actions); @@ -10577,7 +10712,8 @@ add_ecmp_symmetric_reply_flows(struct lflow_table *lflows, const char *port_ip, struct ovn_port *out_port, const struct parsed_route *route, - struct ds *route_match) + struct ds *route_match, + struct lflow_ref *lflow_ref) { const struct nbrec_logical_router_static_route *st_route = route->route; struct ds base_match = DS_EMPTY_INITIALIZER; @@ -10600,12 +10736,12 @@ add_ecmp_symmetric_reply_flows(struct lflow_table *lflows, free(cidr); ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_DEFRAG, 100, ds_cstr(&base_match), "ct_next;", - &st_route->header_); + &st_route->header_, lflow_ref); /* And packets that go out over an ECMP route need conntrack */ ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_DEFRAG, 100, ds_cstr(route_match), "ct_next;", - &st_route->header_); + &st_route->header_, lflow_ref); /* Save src eth and inport in ct_label for packets that arrive over * an ECMP route. @@ -10622,7 +10758,8 @@ add_ecmp_symmetric_reply_flows(struct lflow_table *lflows, ct_ecmp_reply_port_match, out_port->sb->tunnel_key); ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_ECMP_STATEFUL, 100, ds_cstr(&match), ds_cstr(&actions), - &st_route->header_); + &st_route->header_, + lflow_ref); ds_clear(&match); ds_put_format(&match, "%s && (ct.new && !ct.est) && udp", ds_cstr(&base_match)); @@ -10634,7 +10771,8 @@ add_ecmp_symmetric_reply_flows(struct lflow_table *lflows, ct_ecmp_reply_port_match, out_port->sb->tunnel_key); ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_ECMP_STATEFUL, 100, ds_cstr(&match), ds_cstr(&actions), - &st_route->header_); + &st_route->header_, + lflow_ref); ds_clear(&match); ds_put_format(&match, "%s && (ct.new && !ct.est) && sctp", ds_cstr(&base_match)); @@ -10646,7 +10784,8 @@ add_ecmp_symmetric_reply_flows(struct lflow_table *lflows, ct_ecmp_reply_port_match, out_port->sb->tunnel_key); ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_ECMP_STATEFUL, 100, ds_cstr(&match), ds_cstr(&actions), - &st_route->header_); + &st_route->header_, + lflow_ref); ds_clear(&match); ds_put_format(&match, @@ -10660,7 +10799,8 @@ add_ecmp_symmetric_reply_flows(struct lflow_table *lflows, ct_ecmp_reply_port_match, out_port->sb->tunnel_key); ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_ECMP_STATEFUL, 100, ds_cstr(&match), ds_cstr(&actions), - &st_route->header_); + &st_route->header_, + lflow_ref); ds_clear(&match); ds_put_format(&match, @@ -10674,7 +10814,8 @@ add_ecmp_symmetric_reply_flows(struct lflow_table *lflows, ct_ecmp_reply_port_match, out_port->sb->tunnel_key); ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_ECMP_STATEFUL, 100, ds_cstr(&match), ds_cstr(&actions), - &st_route->header_); + &st_route->header_, + lflow_ref); ds_clear(&match); ds_put_format(&match, "%s && (!ct.rpl && ct.est) && sctp", @@ -10687,7 +10828,8 @@ add_ecmp_symmetric_reply_flows(struct lflow_table *lflows, ct_ecmp_reply_port_match, out_port->sb->tunnel_key); ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_ECMP_STATEFUL, 100, ds_cstr(&match), ds_cstr(&actions), - &st_route->header_); + &st_route->header_, + lflow_ref); /* Bypass ECMP selection if we already have ct_label information * for where to route the packet. @@ -10706,12 +10848,14 @@ add_ecmp_symmetric_reply_flows(struct lflow_table *lflows, port_ip, out_port->json_key); ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_IP_ROUTING, 10300, ds_cstr(&match), ds_cstr(&actions), - &st_route->header_); + &st_route->header_, + lflow_ref); /* Egress reply traffic for symmetric ECMP routes skips router policies. */ ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_POLICY, 65535, ds_cstr(&ecmp_reply), "next;", - &st_route->header_); + &st_route->header_, + lflow_ref); /* Use REG_ECMP_ETH_FULL to pass the eth field from ct_label to eth.dst to * avoid masked access to ct_label. Otherwise it may prevent OVS flow @@ -10727,7 +10871,8 @@ add_ecmp_symmetric_reply_flows(struct lflow_table *lflows, " pop(" REG_ECMP_ETH_FULL "); next;"; ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_ARP_RESOLVE, 200, ds_cstr(&ecmp_reply), - action, &st_route->header_); + action, &st_route->header_, + lflow_ref); ds_destroy(&base_match); ds_destroy(&match); @@ -10738,7 +10883,8 @@ add_ecmp_symmetric_reply_flows(struct lflow_table *lflows, static void build_ecmp_route_flow(struct lflow_table *lflows, struct ovn_datapath *od, bool ct_masked_mark, const struct hmap *lr_ports, - struct ecmp_groups_node *eg) + struct ecmp_groups_node *eg, + struct lflow_ref *lflow_ref) { bool is_ipv4 = IN6_IS_ADDR_V4MAPPED(&eg->prefix); @@ -10771,7 +10917,8 @@ build_ecmp_route_flow(struct lflow_table *lflows, struct ovn_datapath *od, ds_put_cstr(&actions, ");"); ovn_lflow_add(lflows, od, S_ROUTER_IN_IP_ROUTING, priority, - ds_cstr(&route_match), ds_cstr(&actions)); + ds_cstr(&route_match), ds_cstr(&actions), + lflow_ref); /* Add per member flow */ struct ds match = DS_EMPTY_INITIALIZER; @@ -10794,7 +10941,8 @@ build_ecmp_route_flow(struct lflow_table *lflows, struct ovn_datapath *od, out_port->key)) { add_ecmp_symmetric_reply_flows(lflows, od, ct_masked_mark, lrp_addr_s, out_port, - route_, &route_match); + route_, &route_match, + lflow_ref); } ds_clear(&match); ds_put_format(&match, REG_ECMP_GROUP_ID" == %"PRIu16" && " @@ -10814,7 +10962,7 @@ build_ecmp_route_flow(struct lflow_table *lflows, struct ovn_datapath *od, out_port->json_key); ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_IP_ROUTING_ECMP, 100, ds_cstr(&match), ds_cstr(&actions), - &route->header_); + &route->header_, lflow_ref); } sset_destroy(&visited_ports); ds_destroy(&match); @@ -10871,17 +11019,17 @@ add_route(struct lflow_table *lflows, struct ovn_datapath *od, ds_put_format(&actions, "ip.ttl--; %s", ds_cstr(&common_actions)); } - ovn_lflow_add_with_lflow_ref_hint(lflows, od, S_ROUTER_IN_IP_ROUTING, - priority, ds_cstr(&match), - ds_cstr(&actions), stage_hint, - lflow_ref); + ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_IP_ROUTING, + priority, ds_cstr(&match), + ds_cstr(&actions), stage_hint, + lflow_ref); if (op && op->has_bfd) { ds_put_format(&match, " && udp.dst == 3784"); - ovn_lflow_add_with_lflow_ref_hint(lflows, op->od, - S_ROUTER_IN_IP_ROUTING, - priority + 1, ds_cstr(&match), - ds_cstr(&common_actions),\ - stage_hint, lflow_ref); + ovn_lflow_add_with_hint(lflows, op->od, + S_ROUTER_IN_IP_ROUTING, + priority + 1, ds_cstr(&match), + ds_cstr(&common_actions),\ + stage_hint, lflow_ref); } ds_destroy(&match); ds_destroy(&common_actions); @@ -10891,7 +11039,8 @@ add_route(struct lflow_table *lflows, struct ovn_datapath *od, static void build_static_route_flow(struct lflow_table *lflows, struct ovn_datapath *od, const struct hmap *lr_ports, - const struct parsed_route *route_) + const struct parsed_route *route_, + struct lflow_ref *lflow_ref) { const char *lrp_addr_s = NULL; struct ovn_port *out_port = NULL; @@ -10915,7 +11064,7 @@ build_static_route_flow(struct lflow_table *lflows, struct ovn_datapath *od, add_route(lflows, route_->is_discard_route ? od : out_port->od, out_port, lrp_addr_s, prefix_s, route_->plen, route->nexthop, route_->is_src_route, route_->route_table_id, &route->header_, - route_->is_discard_route, ofs, NULL); + route_->is_discard_route, ofs, lflow_ref); free(prefix_s); } @@ -10991,7 +11140,8 @@ lrouter_use_common_zone(const struct ovn_datapath *od) static void build_distr_lrouter_nat_flows_for_lb(struct lrouter_nat_lb_flows_ctx *ctx, enum lrouter_nat_lb_flow_type type, - struct ovn_datapath *od) + struct ovn_datapath *od, + struct lflow_ref *lflow_ref) { struct ovn_port *dgp = od->l3dgw_ports[0]; @@ -11030,7 +11180,8 @@ build_distr_lrouter_nat_flows_for_lb(struct lrouter_nat_lb_flows_ctx *ctx, ovn_lflow_add_with_hint__(ctx->lflows, od, S_ROUTER_IN_DNAT, ctx->prio, ds_cstr(ctx->new_match), ctx->new_action[type], - NULL, meter, &ctx->lb->nlb->header_); + NULL, meter, &ctx->lb->nlb->header_, + lflow_ref); ds_truncate(ctx->new_match, new_match_len); @@ -11049,7 +11200,8 @@ build_distr_lrouter_nat_flows_for_lb(struct lrouter_nat_lb_flows_ctx *ctx, ovn_lflow_add_with_hint(ctx->lflows, od, S_ROUTER_IN_GW_REDIRECT, 200, ds_cstr(ctx->undnat_match), ds_cstr(ctx->gw_redir_action), - &ctx->lb->nlb->header_); + &ctx->lb->nlb->header_, + lflow_ref); ds_truncate(ctx->undnat_match, undnat_match_len); ds_put_format(ctx->undnat_match, ") && (inport == %s || outport == %s)" @@ -11057,7 +11209,8 @@ build_distr_lrouter_nat_flows_for_lb(struct lrouter_nat_lb_flows_ctx *ctx, dgp->cr_port->json_key); ovn_lflow_add_with_hint(ctx->lflows, od, S_ROUTER_OUT_UNDNAT, 120, ds_cstr(ctx->undnat_match), undnat_action, - &ctx->lb->nlb->header_); + &ctx->lb->nlb->header_, + lflow_ref); ds_truncate(ctx->undnat_match, undnat_match_len); } @@ -11065,7 +11218,8 @@ static void build_gw_lrouter_nat_flows_for_lb(struct lrouter_nat_lb_flows_ctx *ctx, enum lrouter_nat_lb_flow_type type, const struct ovn_datapaths *lr_datapaths, - const unsigned long *dp_bitmap) + const unsigned long *dp_bitmap, + struct lflow_ref *lflow_ref) { unsigned long *dp_non_meter = NULL; bool build_non_meter = false; @@ -11091,14 +11245,14 @@ build_gw_lrouter_nat_flows_for_lb(struct lrouter_nat_lb_flows_ctx *ctx, bitmap_set0(dp_non_meter, index); ovn_lflow_add_with_hint__(ctx->lflows, od, S_ROUTER_IN_DNAT, ctx->prio, ds_cstr(ctx->new_match), ctx->new_action[type], - NULL, meter, &ctx->lb->nlb->header_); + NULL, meter, &ctx->lb->nlb->header_, lflow_ref); } } if (!ctx->reject || build_non_meter) { ovn_lflow_add_with_dp_group(ctx->lflows, dp_non_meter ? dp_non_meter : dp_bitmap, ods_size(lr_datapaths), S_ROUTER_IN_DNAT, ctx->prio, ds_cstr(ctx->new_match), - ctx->new_action[type], &ctx->lb->nlb->header_); + ctx->new_action[type], &ctx->lb->nlb->header_, lflow_ref); } bitmap_free(dp_non_meter); } @@ -11232,7 +11386,7 @@ build_lrouter_nat_flows_for_lb( if (!od->n_l3dgw_ports) { bitmap_set1(gw_dp_bitmap[type], index); } else { - build_distr_lrouter_nat_flows_for_lb(&ctx, type, od); + build_distr_lrouter_nat_flows_for_lb(&ctx, type, od, NULL); } if (lb->affinity_timeout) { @@ -11253,16 +11407,17 @@ build_lrouter_nat_flows_for_lb( * S_ROUTER_IN_DNAT stage. */ ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_UNSNAT, 120, ds_cstr(&unsnat_match), "next;", - &lb->nlb->header_); + &lb->nlb->header_, NULL); } } for (size_t type = 0; type < LROUTER_NAT_LB_FLOW_MAX; type++) { build_gw_lrouter_nat_flows_for_lb(&ctx, type, lr_datapaths, - gw_dp_bitmap[type]); + gw_dp_bitmap[type], + NULL); build_lb_affinity_lr_flows(lflows, lb, lb_vip, ds_cstr(match), aff_action[type], aff_dp_bitmap[type], - lr_datapaths); + lr_datapaths, NULL); } ds_destroy(&unsnat_match); @@ -11310,7 +11465,8 @@ build_lswitch_flows_for_lb(struct ovn_lb_datapaths *lb_dps, copp_meter_get(COPP_EVENT_ELB, od->nbs->copp, meter_groups), - &lb->nlb->header_); + &lb->nlb->header_, + NULL); } /* Ignore L4 port information in the key because fragmented packets * may not have L4 information. The pre-stateful table will send @@ -11360,7 +11516,7 @@ build_lrouter_defrag_flows_for_lb(struct ovn_lb_datapaths *lb_dps, ovn_lflow_add_with_dp_group( lflows, lb_dps->nb_lr_map, ods_size(lr_datapaths), S_ROUTER_IN_DEFRAG, prio, ds_cstr(match), "ct_dnat;", - &lb_dps->lb->nlb->header_); + &lb_dps->lb->nlb->header_, NULL); } } @@ -11402,7 +11558,7 @@ build_lrouter_flows_for_lb(struct ovn_lb_datapaths *lb_dps, copp_meter_get(COPP_EVENT_ELB, od->nbr->copp, meter_groups), - &lb->nlb->header_); + &lb->nlb->header_, NULL); } } @@ -11411,7 +11567,8 @@ build_lrouter_flows_for_lb(struct ovn_lb_datapaths *lb_dps, struct ovn_datapath *od = lr_datapaths->array[index]; ovn_lflow_add(lflows, od, S_ROUTER_OUT_SNAT, 120, - "flags.skip_snat_for_lb == 1 && ip", "next;"); + "flags.skip_snat_for_lb == 1 && ip", "next;", + NULL); } } } @@ -11526,7 +11683,8 @@ static inline void lrouter_nat_add_ext_ip_match(const struct ovn_datapath *od, struct lflow_table *lflows, struct ds *match, const struct nbrec_nat *nat, - bool is_v6, bool is_src, int cidr_bits) + bool is_v6, bool is_src, int cidr_bits, + struct lflow_ref *lflow_ref) { struct nbrec_address_set *allowed_ext_ips = nat->allowed_ext_ips; struct nbrec_address_set *exempted_ext_ips = nat->exempted_ext_ips; @@ -11577,7 +11735,7 @@ lrouter_nat_add_ext_ip_match(const struct ovn_datapath *od, ovn_lflow_add_with_hint(lflows, od, stage, priority, ds_cstr(&match_exempt), "next;", - &nat->header_); + &nat->header_, lflow_ref); ds_destroy(&match_exempt); } } @@ -11591,7 +11749,8 @@ build_lrouter_arp_flow(const struct ovn_datapath *od, struct ovn_port *op, const char *ip_address, const char *eth_addr, struct ds *extra_match, bool drop, uint16_t priority, const struct ovsdb_idl_row *hint, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { struct ds match = DS_EMPTY_INITIALIZER; struct ds actions = DS_EMPTY_INITIALIZER; @@ -11623,7 +11782,8 @@ build_lrouter_arp_flow(const struct ovn_datapath *od, struct ovn_port *op, } ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_IP_INPUT, priority, - ds_cstr(&match), ds_cstr(&actions), hint); + ds_cstr(&match), ds_cstr(&actions), hint, + lflow_ref); ds_destroy(&match); ds_destroy(&actions); @@ -11642,7 +11802,8 @@ build_lrouter_nd_flow(const struct ovn_datapath *od, struct ovn_port *op, struct ds *extra_match, bool drop, uint16_t priority, const struct ovsdb_idl_row *hint, struct lflow_table *lflows, - const struct shash *meter_groups) + const struct shash *meter_groups, + struct lflow_ref *lflow_ref) { struct ds match = DS_EMPTY_INITIALIZER; struct ds actions = DS_EMPTY_INITIALIZER; @@ -11665,7 +11826,8 @@ build_lrouter_nd_flow(const struct ovn_datapath *od, struct ovn_port *op, if (drop) { ds_put_cstr(&actions, debug_drop_action()); ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_IP_INPUT, priority, - ds_cstr(&match), ds_cstr(&actions), hint); + ds_cstr(&match), ds_cstr(&actions), hint, + lflow_ref); } else { ds_put_format(&actions, "%s { " @@ -11683,7 +11845,7 @@ build_lrouter_nd_flow(const struct ovn_datapath *od, struct ovn_port *op, ds_cstr(&match), ds_cstr(&actions), NULL, copp_meter_get(COPP_ND_NA, od->nbr->copp, meter_groups), - hint); + hint, lflow_ref); } ds_destroy(&match); @@ -11694,7 +11856,8 @@ static void build_lrouter_nat_arp_nd_flow(const struct ovn_datapath *od, struct ovn_nat *nat_entry, struct lflow_table *lflows, - const struct shash *meter_groups) + const struct shash *meter_groups, + struct lflow_ref *lflow_ref) { struct lport_addresses *ext_addrs = &nat_entry->ext_addrs; const struct nbrec_nat *nat = nat_entry->nb; @@ -11704,12 +11867,14 @@ build_lrouter_nat_arp_nd_flow(const struct ovn_datapath *od, ext_addrs->ipv6_addrs[0].addr_s, ext_addrs->ipv6_addrs[0].sn_addr_s, REG_INPORT_ETH_ADDR, NULL, false, 90, - &nat->header_, lflows, meter_groups); + &nat->header_, lflows, meter_groups, + lflow_ref); } else { build_lrouter_arp_flow(od, NULL, ext_addrs->ipv4_addrs[0].addr_s, REG_INPORT_ETH_ADDR, NULL, false, 90, - &nat->header_, lflows); + &nat->header_, lflows, + lflow_ref); } } @@ -11717,7 +11882,8 @@ static void build_lrouter_port_nat_arp_nd_flow(struct ovn_port *op, struct ovn_nat *nat_entry, struct lflow_table *lflows, - const struct shash *meter_groups) + const struct shash *meter_groups, + struct lflow_ref *lflow_ref) { struct lport_addresses *ext_addrs = &nat_entry->ext_addrs; const struct nbrec_nat *nat = nat_entry->nb; @@ -11765,21 +11931,25 @@ build_lrouter_port_nat_arp_nd_flow(struct ovn_port *op, ext_addrs->ipv6_addrs[0].addr_s, ext_addrs->ipv6_addrs[0].sn_addr_s, mac_s, &match, false, 92, - &nat->header_, lflows, meter_groups); + &nat->header_, lflows, meter_groups, + lflow_ref); build_lrouter_nd_flow(op->od, op, "nd_na", ext_addrs->ipv6_addrs[0].addr_s, ext_addrs->ipv6_addrs[0].sn_addr_s, mac_s, NULL, true, 91, - &nat->header_, lflows, meter_groups); + &nat->header_, lflows, meter_groups, + lflow_ref); } else { build_lrouter_arp_flow(op->od, op, ext_addrs->ipv4_addrs[0].addr_s, mac_s, &match, false, 92, - &nat->header_, lflows); + &nat->header_, lflows, + lflow_ref); build_lrouter_arp_flow(op->od, op, ext_addrs->ipv4_addrs[0].addr_s, mac_s, NULL, true, 91, - &nat->header_, lflows); + &nat->header_, lflows, + lflow_ref); } ds_destroy(&match); @@ -11790,7 +11960,8 @@ build_lrouter_drop_own_dest(struct ovn_port *op, const struct lr_stateful_record *lr_stateful_rec, enum ovn_stage stage, uint16_t priority, bool drop_snat_ip, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { struct ds match_ips = DS_EMPTY_INITIALIZER; @@ -11817,7 +11988,8 @@ build_lrouter_drop_own_dest(struct ovn_port *op, char *match = xasprintf("ip4.dst == {%s}", ds_cstr(&match_ips)); ovn_lflow_add_with_hint(lflows, op->od, stage, priority, match, debug_drop_action(), - &op->nbrp->header_); + &op->nbrp->header_, + lflow_ref); free(match); } } @@ -11847,7 +12019,8 @@ build_lrouter_drop_own_dest(struct ovn_port *op, char *match = xasprintf("ip6.dst == {%s}", ds_cstr(&match_ips)); ovn_lflow_add_with_hint(lflows, op->od, stage, priority, match, debug_drop_action(), - &op->nbrp->header_); + &op->nbrp->header_, + lflow_ref); free(match); } } @@ -11858,14 +12031,15 @@ static void build_lrouter_force_snat_flows(struct lflow_table *lflows, const struct ovn_datapath *od, const char *ip_version, const char *ip_addr, - const char *context) + const char *context, + struct lflow_ref *lflow_ref) { struct ds match = DS_EMPTY_INITIALIZER; struct ds actions = DS_EMPTY_INITIALIZER; ds_put_format(&match, "ip%s && ip%s.dst == %s", ip_version, ip_version, ip_addr); ovn_lflow_add(lflows, od, S_ROUTER_IN_UNSNAT, 110, - ds_cstr(&match), "ct_snat;"); + ds_cstr(&match), "ct_snat;", lflow_ref); /* Higher priority rules to force SNAT with the IP addresses * configured in the Gateway router. This only takes effect @@ -11875,7 +12049,8 @@ build_lrouter_force_snat_flows(struct lflow_table *lflows, context, ip_version); ds_put_format(&actions, "ct_snat(%s);", ip_addr); ovn_lflow_add(lflows, od, S_ROUTER_OUT_SNAT, 100, - ds_cstr(&match), ds_cstr(&actions)); + ds_cstr(&match), ds_cstr(&actions), + lflow_ref); ds_destroy(&match); ds_destroy(&actions); @@ -11885,7 +12060,8 @@ static void build_lrouter_force_snat_flows_op(struct ovn_port *op, const struct lr_nat_record *lrnat_rec, struct lflow_table *lflows, - struct ds *match, struct ds *actions) + struct ds *match, struct ds *actions, + struct lflow_ref *lflow_ref) { ovs_assert(op->nbrp); if (!op->peer || !lrnat_rec->lb_force_snat_router_ip) { @@ -11899,7 +12075,7 @@ build_lrouter_force_snat_flows_op(struct ovn_port *op, ds_put_format(match, "inport == %s && ip4.dst == %s", op->json_key, op->lrp_networks.ipv4_addrs[0].addr_s); ovn_lflow_add(lflows, op->od, S_ROUTER_IN_UNSNAT, 110, - ds_cstr(match), "ct_snat;"); + ds_cstr(match), "ct_snat;", lflow_ref); ds_clear(match); @@ -11911,7 +12087,8 @@ build_lrouter_force_snat_flows_op(struct ovn_port *op, ds_put_format(actions, "ct_snat(%s);", op->lrp_networks.ipv4_addrs[0].addr_s); ovn_lflow_add(lflows, op->od, S_ROUTER_OUT_SNAT, 110, - ds_cstr(match), ds_cstr(actions)); + ds_cstr(match), ds_cstr(actions), + lflow_ref); if (op->lrp_networks.n_ipv4_addrs > 1) { static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, 5); VLOG_WARN_RL(&rl, "Logical router port %s is configured with " @@ -11931,7 +12108,7 @@ build_lrouter_force_snat_flows_op(struct ovn_port *op, ds_put_format(match, "inport == %s && ip6.dst == %s", op->json_key, op->lrp_networks.ipv6_addrs[0].addr_s); ovn_lflow_add(lflows, op->od, S_ROUTER_IN_UNSNAT, 110, - ds_cstr(match), "ct_snat;"); + ds_cstr(match), "ct_snat;", lflow_ref); ds_clear(match); @@ -11943,7 +12120,8 @@ build_lrouter_force_snat_flows_op(struct ovn_port *op, ds_put_format(actions, "ct_snat(%s);", op->lrp_networks.ipv6_addrs[0].addr_s); ovn_lflow_add(lflows, op->od, S_ROUTER_OUT_SNAT, 110, - ds_cstr(match), ds_cstr(actions)); + ds_cstr(match), ds_cstr(actions), + lflow_ref); if (op->lrp_networks.n_ipv6_addrs > 2) { static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, 5); VLOG_WARN_RL(&rl, "Logical router port %s is configured with " @@ -11957,7 +12135,8 @@ build_lrouter_force_snat_flows_op(struct ovn_port *op, static void build_lrouter_bfd_flows(struct lflow_table *lflows, struct ovn_port *op, - const struct shash *meter_groups) + const struct shash *meter_groups, + struct lflow_ref *lflow_ref) { if (!op->has_bfd) { return; @@ -11972,7 +12151,8 @@ build_lrouter_bfd_flows(struct lflow_table *lflows, struct ovn_port *op, ds_cstr(&ip_list)); ovn_lflow_add_with_hint(lflows, op->od, S_ROUTER_IN_IP_INPUT, 110, ds_cstr(&match), "next; ", - &op->nbrp->header_); + &op->nbrp->header_, + lflow_ref); ds_clear(&match); ds_put_format(&match, "ip4.dst == %s && udp.dst == 3784", ds_cstr(&ip_list)); @@ -11980,7 +12160,8 @@ build_lrouter_bfd_flows(struct lflow_table *lflows, struct ovn_port *op, ds_cstr(&match), "handle_bfd_msg(); ", NULL, copp_meter_get(COPP_BFD, op->od->nbr->copp, meter_groups), - &op->nbrp->header_); + &op->nbrp->header_, + lflow_ref); } if (op->lrp_networks.n_ipv6_addrs) { ds_clear(&ip_list); @@ -11991,7 +12172,8 @@ build_lrouter_bfd_flows(struct lflow_table *lflows, struct ovn_port *op, ds_cstr(&ip_list)); ovn_lflow_add_with_hint(lflows, op->od, S_ROUTER_IN_IP_INPUT, 110, ds_cstr(&match), "next; ", - &op->nbrp->header_); + &op->nbrp->header_, + lflow_ref); ds_clear(&match); ds_put_format(&match, "ip6.dst == %s && udp.dst == 3784", ds_cstr(&ip_list)); @@ -11999,7 +12181,8 @@ build_lrouter_bfd_flows(struct lflow_table *lflows, struct ovn_port *op, ds_cstr(&match), "handle_bfd_msg(); ", NULL, copp_meter_get(COPP_BFD, op->od->nbr->copp, meter_groups), - &op->nbrp->header_); + &op->nbrp->header_, + lflow_ref); } ds_destroy(&ip_list); @@ -12011,16 +12194,19 @@ build_lrouter_bfd_flows(struct lflow_table *lflows, struct ovn_port *op, */ static void build_adm_ctrl_flows_for_lrouter( - struct ovn_datapath *od, struct lflow_table *lflows) + struct ovn_datapath *od, struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbr); /* Logical VLANs not supported. * Broadcast/multicast source address is invalid. */ ovn_lflow_add(lflows, od, S_ROUTER_IN_ADMISSION, 100, - "vlan.present || eth.src[40]", debug_drop_action()); + "vlan.present || eth.src[40]", debug_drop_action(), + lflow_ref); /* Default action for L2 security is to drop. */ - ovn_lflow_add_default_drop(lflows, od, S_ROUTER_IN_ADMISSION); + ovn_lflow_add_default_drop(lflows, od, S_ROUTER_IN_ADMISSION, + lflow_ref); } static int @@ -12054,11 +12240,12 @@ build_gateway_get_l2_hdr_size(struct ovn_port *op) /* All 'gateway_mtu' and 'gateway_mtu_bypass' flows should be built with this * function. */ -static void OVS_PRINTF_FORMAT(9, 10) +static void OVS_PRINTF_FORMAT(10, 11) build_gateway_mtu_flow(struct lflow_table *lflows, struct ovn_port *op, enum ovn_stage stage, uint16_t prio_low, uint16_t prio_high, struct ds *match, struct ds *actions, const struct ovsdb_idl_row *hint, + struct lflow_ref *lflow_ref, const char *extra_actions_fmt, ...) { int gw_mtu = smap_get_int(&op->nbrp->options, "gateway_mtu", 0); @@ -12076,7 +12263,7 @@ build_gateway_mtu_flow(struct lflow_table *lflows, struct ovn_port *op, ds_put_format_valist(actions, extra_actions_fmt, extra_actions_args); ovn_lflow_add_with_hint(lflows, op->od, stage, prio_low, ds_cstr(match), ds_cstr(actions), - hint); + hint, lflow_ref); if (gw_mtu > 0) { const char *gw_mtu_bypass = smap_get(&op->nbrp->options, @@ -12088,7 +12275,7 @@ build_gateway_mtu_flow(struct lflow_table *lflows, struct ovn_port *op, ds_put_format(match, " && (%s)", gw_mtu_bypass); ovn_lflow_add_with_hint(lflows, op->od, stage, prio_high, ds_cstr(match), ds_cstr(actions), - hint); + hint, lflow_ref); } } va_end(extra_actions_args); @@ -12117,7 +12304,8 @@ consider_l3dgw_port_is_centralized(struct ovn_port *op) static void build_adm_ctrl_flows_for_lrouter_port( struct ovn_port *op, struct lflow_table *lflows, - struct ds *match, struct ds *actions) + struct ds *match, struct ds *actions, + struct lflow_ref *lflow_ref) { ovs_assert(op->nbrp); @@ -12141,6 +12329,7 @@ build_adm_ctrl_flows_for_lrouter_port( ds_put_format(match, "eth.mcast && inport == %s", op->json_key); build_gateway_mtu_flow(lflows, op, S_ROUTER_IN_ADMISSION, 50, 55, match, actions, &op->nbrp->header_, + lflow_ref, REG_INPORT_ETH_ADDR " = %s; next;", op->lrp_networks.ea_s); @@ -12161,6 +12350,7 @@ build_adm_ctrl_flows_for_lrouter_port( } build_gateway_mtu_flow(lflows, op, S_ROUTER_IN_ADMISSION, 50, 55, match, actions, &op->nbrp->header_, + lflow_ref, REG_INPORT_ETH_ADDR " = %s; next;", op->lrp_networks.ea_s); } @@ -12172,7 +12362,8 @@ static void build_neigh_learning_flows_for_lrouter( struct ovn_datapath *od, struct lflow_table *lflows, struct ds *match, struct ds *actions, - const struct shash *meter_groups) + const struct shash *meter_groups, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbr); @@ -12219,7 +12410,7 @@ build_neigh_learning_flows_for_lrouter( learn_from_arp_request ? "" : REGBIT_LOOKUP_NEIGHBOR_IP_RESULT" = 1; "); ovn_lflow_add(lflows, od, S_ROUTER_IN_LOOKUP_NEIGHBOR, 100, - "arp.op == 2", ds_cstr(actions)); + "arp.op == 2", ds_cstr(actions), lflow_ref); ds_clear(actions); ds_put_format(actions, REGBIT_LOOKUP_NEIGHBOR_RESULT @@ -12227,7 +12418,7 @@ build_neigh_learning_flows_for_lrouter( learn_from_arp_request ? "" : REGBIT_LOOKUP_NEIGHBOR_IP_RESULT" = 1; "); ovn_lflow_add(lflows, od, S_ROUTER_IN_LOOKUP_NEIGHBOR, 100, "nd_na", - ds_cstr(actions)); + ds_cstr(actions), lflow_ref); if (!learn_from_arp_request) { /* Add flow to skip GARP LLA if we don't know it already. @@ -12244,7 +12435,7 @@ build_neigh_learning_flows_for_lrouter( " = lookup_nd_ip(inport, ip6.src); next;"); ovn_lflow_add(lflows, od, S_ROUTER_IN_LOOKUP_NEIGHBOR, 110, "nd_na && ip6.src == fe80::/10 && ip6.dst == ff00::/8", - ds_cstr(actions)); + ds_cstr(actions), lflow_ref); } ds_clear(actions); @@ -12254,12 +12445,13 @@ build_neigh_learning_flows_for_lrouter( REGBIT_LOOKUP_NEIGHBOR_IP_RESULT " = lookup_nd_ip(inport, ip6.src); "); ovn_lflow_add(lflows, od, S_ROUTER_IN_LOOKUP_NEIGHBOR, 100, "nd_ns", - ds_cstr(actions)); + ds_cstr(actions), lflow_ref); /* For other packet types, we can skip neighbor learning. * So set REGBIT_LOOKUP_NEIGHBOR_RESULT to 1. */ ovn_lflow_add(lflows, od, S_ROUTER_IN_LOOKUP_NEIGHBOR, 0, "1", - REGBIT_LOOKUP_NEIGHBOR_RESULT" = 1; next;"); + REGBIT_LOOKUP_NEIGHBOR_RESULT" = 1; next;", + lflow_ref); /* Flows for LEARN_NEIGHBOR. */ /* Skip Neighbor learning if not required. */ @@ -12268,33 +12460,40 @@ build_neigh_learning_flows_for_lrouter( learn_from_arp_request ? "" : " || "REGBIT_LOOKUP_NEIGHBOR_IP_RESULT" == 0"); ovn_lflow_add(lflows, od, S_ROUTER_IN_LEARN_NEIGHBOR, 100, - ds_cstr(match), "mac_cache_use; next;"); + ds_cstr(match), "mac_cache_use; next;", + lflow_ref); ovn_lflow_metered(lflows, od, S_ROUTER_IN_LEARN_NEIGHBOR, 90, "arp", "put_arp(inport, arp.spa, arp.sha); next;", copp_meter_get(COPP_ARP, od->nbr->copp, - meter_groups)); + meter_groups), + lflow_ref); ovn_lflow_add(lflows, od, S_ROUTER_IN_LEARN_NEIGHBOR, 95, - "nd_ns && (ip6.src == 0 || nd.sll == 0)", "next;"); + "nd_ns && (ip6.src == 0 || nd.sll == 0)", "next;", + lflow_ref); ovn_lflow_metered(lflows, od, S_ROUTER_IN_LEARN_NEIGHBOR, 95, "nd_na && nd.tll == 0", "put_nd(inport, nd.target, eth.src); next;", copp_meter_get(COPP_ND_NA, od->nbr->copp, - meter_groups)); + meter_groups), + lflow_ref); ovn_lflow_metered(lflows, od, S_ROUTER_IN_LEARN_NEIGHBOR, 90, "nd_na", "put_nd(inport, nd.target, nd.tll); next;", copp_meter_get(COPP_ND_NA, od->nbr->copp, - meter_groups)); + meter_groups), + lflow_ref); ovn_lflow_metered(lflows, od, S_ROUTER_IN_LEARN_NEIGHBOR, 90, "nd_ns", "put_nd(inport, ip6.src, nd.sll); next;", copp_meter_get(COPP_ND_NS, od->nbr->copp, - meter_groups)); + meter_groups), + lflow_ref); - ovn_lflow_add_default_drop(lflows, od, S_ROUTER_IN_LEARN_NEIGHBOR); + ovn_lflow_add_default_drop(lflows, od, S_ROUTER_IN_LEARN_NEIGHBOR, + lflow_ref); } /* Logical router ingress Table 1: Neighbor lookup lflows @@ -12302,7 +12501,8 @@ build_neigh_learning_flows_for_lrouter( static void build_neigh_learning_flows_for_lrouter_port( struct ovn_port *op, struct lflow_table *lflows, - struct ds *match, struct ds *actions) + struct ds *match, struct ds *actions, + struct lflow_ref *lflow_ref) { ovs_assert(op->nbrp); @@ -12334,7 +12534,8 @@ build_neigh_learning_flows_for_lrouter_port( ovn_lflow_add_with_hint(lflows, op->od, S_ROUTER_IN_LOOKUP_NEIGHBOR, 110, ds_cstr(match), actions_s, - &op->nbrp->header_); + &op->nbrp->header_, + lflow_ref); } ds_clear(match); ds_put_format(match, @@ -12355,7 +12556,8 @@ build_neigh_learning_flows_for_lrouter_port( ovn_lflow_add_with_hint(lflows, op->od, S_ROUTER_IN_LOOKUP_NEIGHBOR, 100, ds_cstr(match), ds_cstr(actions), - &op->nbrp->header_); + &op->nbrp->header_, + lflow_ref); } } @@ -12365,7 +12567,8 @@ static void build_ND_RA_flows_for_lrouter_port( struct ovn_port *op, struct lflow_table *lflows, struct ds *match, struct ds *actions, - const struct shash *meter_groups) + const struct shash *meter_groups, + struct lflow_ref *lflow_ref) { ovs_assert(op->nbrp); if (op->nbrp->peer || !op->peer) { @@ -12452,7 +12655,8 @@ build_ND_RA_flows_for_lrouter_port( copp_meter_get(COPP_ND_RA_OPTS, op->od->nbr->copp, meter_groups), - &op->nbrp->header_); + &op->nbrp->header_, + lflow_ref); ds_clear(actions); ds_clear(match); ds_put_format(match, "inport == %s && ip6.dst == ff02::2 && " @@ -12471,7 +12675,8 @@ build_ND_RA_flows_for_lrouter_port( ovn_lflow_add_with_hint(lflows, op->od, S_ROUTER_IN_ND_RA_RESPONSE, 50, ds_cstr(match), ds_cstr(actions), - &op->nbrp->header_); + &op->nbrp->header_, + lflow_ref); } } @@ -12479,22 +12684,26 @@ build_ND_RA_flows_for_lrouter_port( * responder, by default goto next. (priority 0). */ static void build_ND_RA_flows_for_lrouter(struct ovn_datapath *od, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbr); - ovn_lflow_add(lflows, od, S_ROUTER_IN_ND_RA_OPTIONS, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_ROUTER_IN_ND_RA_RESPONSE, 0, "1", "next;"); + ovn_lflow_add(lflows, od, S_ROUTER_IN_ND_RA_OPTIONS, 0, "1", "next;", + lflow_ref); + ovn_lflow_add(lflows, od, S_ROUTER_IN_ND_RA_RESPONSE, 0, "1", "next;", + lflow_ref); } /* Logical router ingress table IP_ROUTING_PRE: * by default goto next. (priority 0). */ static void build_ip_routing_pre_flows_for_lrouter(struct ovn_datapath *od, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbr); ovn_lflow_add(lflows, od, S_ROUTER_IN_IP_ROUTING_PRE, 0, "1", - REG_ROUTE_TABLE_ID" = 0; next;"); + REG_ROUTE_TABLE_ID" = 0; next;", lflow_ref); } /* Logical router ingress table IP_ROUTING : IP Routing. @@ -12518,7 +12727,8 @@ build_ip_routing_pre_flows_for_lrouter(struct ovn_datapath *od, */ static void build_ip_routing_flows_for_lrp( - struct ovn_port *op, struct lflow_table *lflows) + struct ovn_port *op, struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { ovs_assert(op->nbrp); for (int i = 0; i < op->lrp_networks.n_ipv4_addrs; i++) { @@ -12526,7 +12736,7 @@ build_ip_routing_flows_for_lrp( op->lrp_networks.ipv4_addrs[i].network_s, op->lrp_networks.ipv4_addrs[i].plen, NULL, false, 0, &op->nbrp->header_, false, ROUTE_PRIO_OFFSET_CONNECTED, - NULL); + lflow_ref); } for (int i = 0; i < op->lrp_networks.n_ipv6_addrs; i++) { @@ -12534,7 +12744,7 @@ build_ip_routing_flows_for_lrp( op->lrp_networks.ipv6_addrs[i].network_s, op->lrp_networks.ipv6_addrs[i].plen, NULL, false, 0, &op->nbrp->header_, false, ROUTE_PRIO_OFFSET_CONNECTED, - NULL); + lflow_ref); } } @@ -12599,13 +12809,17 @@ static void build_static_route_flows_for_lrouter( struct ovn_datapath *od, const struct chassis_features *features, struct lflow_table *lflows, const struct hmap *lr_ports, - const struct hmap *bfd_connections) + const struct hmap *bfd_connections, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbr); - ovn_lflow_add_default_drop(lflows, od, S_ROUTER_IN_IP_ROUTING_ECMP); - ovn_lflow_add_default_drop(lflows, od, S_ROUTER_IN_IP_ROUTING); + ovn_lflow_add_default_drop(lflows, od, S_ROUTER_IN_IP_ROUTING_ECMP, + lflow_ref); + ovn_lflow_add_default_drop(lflows, od, S_ROUTER_IN_IP_ROUTING, + lflow_ref); ovn_lflow_add(lflows, od, S_ROUTER_IN_IP_ROUTING_ECMP, 150, - REG_ECMP_GROUP_ID" == 0", "next;"); + REG_ECMP_GROUP_ID" == 0", "next;", + lflow_ref); struct hmap ecmp_groups = HMAP_INITIALIZER(&ecmp_groups); struct hmap unique_routes = HMAP_INITIALIZER(&unique_routes); @@ -12615,7 +12829,7 @@ build_static_route_flows_for_lrouter( for (int i = 0; i < od->nbr->n_ports; i++) { build_route_table_lflow(od, lflows, od->nbr->ports[i], - &route_tables); + &route_tables, lflow_ref); } for (int i = 0; i < od->nbr->n_static_routes; i++) { @@ -12645,11 +12859,11 @@ build_static_route_flows_for_lrouter( /* add a flow in IP_ROUTING, and one flow for each member in * IP_ROUTING_ECMP. */ build_ecmp_route_flow(lflows, od, features->ct_no_masked_label, - lr_ports, group); + lr_ports, group, lflow_ref); } const struct unique_routes_node *ur; HMAP_FOR_EACH (ur, hmap_node, &unique_routes) { - build_static_route_flow(lflows, od, lr_ports, ur->route); + build_static_route_flow(lflows, od, lr_ports, ur->route, lflow_ref); } ecmp_groups_destroy(&ecmp_groups); unique_routes_destroy(&unique_routes); @@ -12663,7 +12877,8 @@ build_static_route_flows_for_lrouter( static void build_mcast_lookup_flows_for_lrouter( struct ovn_datapath *od, struct lflow_table *lflows, - struct ds *match, struct ds *actions) + struct ds *match, struct ds *actions, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbr); @@ -12671,7 +12886,8 @@ build_mcast_lookup_flows_for_lrouter( * i.e., router solicitation and router advertisement. */ ovn_lflow_add(lflows, od, S_ROUTER_IN_IP_ROUTING, 10550, - "nd_rs || nd_ra", debug_drop_action()); + "nd_rs || nd_ra", debug_drop_action(), + lflow_ref); if (!od->mcast_info.rtr.relay) { return; } @@ -12699,7 +12915,8 @@ build_mcast_lookup_flows_for_lrouter( ds_put_format(actions, "outport = \"%s\"; ip.ttl--; next;", igmp_group->mcgroup.name); ovn_lflow_add(lflows, od, S_ROUTER_IN_IP_ROUTING, 10500, - ds_cstr(match), ds_cstr(actions)); + ds_cstr(match), ds_cstr(actions), + lflow_ref); } /* If needed, flood unregistered multicast on statically configured @@ -12718,13 +12935,15 @@ build_mcast_lookup_flows_for_lrouter( ds_put_format(match, "eth.src == %s && igmp", op->lrp_networks.ea_s); ovn_lflow_add(lflows, od, S_ROUTER_IN_IP_ROUTING, 10550, - ds_cstr(match), debug_drop_action()); + ds_cstr(match), debug_drop_action(), + lflow_ref); ds_clear(match); ds_put_format(match, "eth.src == %s && (mldv1 || mldv2)", op->lrp_networks.ea_s); ovn_lflow_add(lflows, od, S_ROUTER_IN_IP_ROUTING, 10550, - ds_cstr(match), debug_drop_action()); + ds_cstr(match), debug_drop_action(), + lflow_ref); } ovn_lflow_add(lflows, od, S_ROUTER_IN_IP_ROUTING, 10460, @@ -12732,23 +12951,27 @@ build_mcast_lookup_flows_for_lrouter( "clone { " "outport = \""MC_STATIC"\"; " "next; " - "};"); + "};", + lflow_ref); ovn_lflow_add(lflows, od, S_ROUTER_IN_IP_ROUTING, 10460, "mldv1 || mldv2", "clone { " "outport = \""MC_STATIC"\"; " "next; " - "};"); + "};", + lflow_ref); ovn_lflow_add(lflows, od, S_ROUTER_IN_IP_ROUTING, 10450, "ip4.mcast || ip6.mcast", "clone { " "outport = \""MC_STATIC"\"; " "ip.ttl--; " "next; " - "};"); + "};", + lflow_ref); } else { ovn_lflow_add(lflows, od, S_ROUTER_IN_IP_ROUTING, 10450, - "ip4.mcast || ip6.mcast", debug_drop_action()); + "ip4.mcast || ip6.mcast", debug_drop_action(), + lflow_ref); } } @@ -12764,16 +12987,20 @@ build_mcast_lookup_flows_for_lrouter( static void build_ingress_policy_flows_for_lrouter( struct ovn_datapath *od, struct lflow_table *lflows, - const struct hmap *lr_ports) + const struct hmap *lr_ports, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbr); /* This is a catch-all rule. It has the lowest priority (0) * does a match-all("1") and pass-through (next) */ ovn_lflow_add(lflows, od, S_ROUTER_IN_POLICY, 0, "1", - REG_ECMP_GROUP_ID" = 0; next;"); + REG_ECMP_GROUP_ID" = 0; next;", + lflow_ref); ovn_lflow_add(lflows, od, S_ROUTER_IN_POLICY_ECMP, 150, - REG_ECMP_GROUP_ID" == 0", "next;"); - ovn_lflow_add_default_drop(lflows, od, S_ROUTER_IN_POLICY_ECMP); + REG_ECMP_GROUP_ID" == 0", "next;", + lflow_ref); + ovn_lflow_add_default_drop(lflows, od, S_ROUTER_IN_POLICY_ECMP, + lflow_ref); /* Convert routing policies to flows. */ uint16_t ecmp_group_id = 1; @@ -12785,11 +13012,11 @@ build_ingress_policy_flows_for_lrouter( if (is_ecmp_reroute) { build_ecmp_routing_policy_flows(lflows, od, lr_ports, rule, - ecmp_group_id); + ecmp_group_id, lflow_ref); ecmp_group_id++; } else { build_routing_policy_flow(lflows, od, lr_ports, rule, - &rule->header_); + &rule->header_, lflow_ref); } } } @@ -12797,21 +13024,26 @@ build_ingress_policy_flows_for_lrouter( /* Local router ingress table ARP_RESOLVE: ARP Resolution. */ static void build_arp_resolve_flows_for_lrouter( - struct ovn_datapath *od, struct lflow_table *lflows) + struct ovn_datapath *od, struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbr); /* Multicast packets already have the outport set so just advance to * next table (priority 500). */ ovn_lflow_add(lflows, od, S_ROUTER_IN_ARP_RESOLVE, 500, - "ip4.mcast || ip6.mcast", "next;"); + "ip4.mcast || ip6.mcast", "next;", + lflow_ref); ovn_lflow_add(lflows, od, S_ROUTER_IN_ARP_RESOLVE, 1, "ip4", - "get_arp(outport, " REG_NEXT_HOP_IPV4 "); next;"); + "get_arp(outport, " REG_NEXT_HOP_IPV4 "); next;", + lflow_ref); ovn_lflow_add(lflows, od, S_ROUTER_IN_ARP_RESOLVE, 1, "ip6", - "get_nd(outport, " REG_NEXT_HOP_IPV6 "); next;"); + "get_nd(outport, " REG_NEXT_HOP_IPV6 "); next;", + lflow_ref); - ovn_lflow_add_default_drop(lflows, od, S_ROUTER_IN_ARP_RESOLVE); + ovn_lflow_add_default_drop(lflows, od, S_ROUTER_IN_ARP_RESOLVE, + lflow_ref); } static void @@ -12844,9 +13076,8 @@ routable_addresses_to_lflows(struct lflow_table *lflows, ds_clear(actions); ds_put_format(actions, "eth.dst = %s; next;", ra.laddrs[i].ea_s); - ovn_lflow_add_with_lflow_ref(lflows, peer->od, S_ROUTER_IN_ARP_RESOLVE, - 100, ds_cstr(match), ds_cstr(actions), - lflow_ref); + ovn_lflow_add(lflows, peer->od, S_ROUTER_IN_ARP_RESOLVE, 100, + ds_cstr(match), ds_cstr(actions), lflow_ref); } destroy_routable_addresses(&ra); } @@ -12865,7 +13096,8 @@ routable_addresses_to_lflows(struct lflow_table *lflows, static void build_arp_resolve_flows_for_lrp(struct ovn_port *op, struct lflow_table *lflows, - struct ds *match, struct ds *actions) + struct ds *match, struct ds *actions, + struct lflow_ref *lflow_ref) { ovs_assert(op->nbrp); /* This is a logical router port. If next-hop IP address in @@ -12890,7 +13122,8 @@ build_arp_resolve_flows_for_lrp(struct ovn_port *op, ovn_lflow_add_with_hint(lflows, op->peer->od, S_ROUTER_IN_ARP_RESOLVE, 100, ds_cstr(match), ds_cstr(actions), - &op->nbrp->header_); + &op->nbrp->header_, + lflow_ref); } if (op->lrp_networks.n_ipv6_addrs) { @@ -12906,7 +13139,8 @@ build_arp_resolve_flows_for_lrp(struct ovn_port *op, ovn_lflow_add_with_hint(lflows, op->peer->od, S_ROUTER_IN_ARP_RESOLVE, 100, ds_cstr(match), ds_cstr(actions), - &op->nbrp->header_); + &op->nbrp->header_, + lflow_ref); } } @@ -12931,7 +13165,8 @@ build_arp_resolve_flows_for_lrp(struct ovn_port *op, ovn_lflow_add_with_hint(lflows, op->od, S_ROUTER_IN_ARP_RESOLVE, 50, ds_cstr(match), ds_cstr(actions), - &op->nbrp->header_); + &op->nbrp->header_, + lflow_ref); } } } @@ -12981,7 +13216,7 @@ build_arp_resolve_flows_for_lsp( ds_clear(actions); ds_put_format(actions, "eth.dst = %s; next;", ea_s); - ovn_lflow_add_with_lflow_ref_hint(lflows, peer->od, + ovn_lflow_add_with_hint(lflows, peer->od, S_ROUTER_IN_ARP_RESOLVE, 100, ds_cstr(match), ds_cstr(actions), @@ -13013,7 +13248,7 @@ build_arp_resolve_flows_for_lsp( ds_clear(actions); ds_put_format(actions, "eth.dst = %s; next;", ea_s); - ovn_lflow_add_with_lflow_ref_hint(lflows, peer->od, + ovn_lflow_add_with_hint(lflows, peer->od, S_ROUTER_IN_ARP_RESOLVE, 100, ds_cstr(match), ds_cstr(actions), @@ -13062,7 +13297,7 @@ build_arp_resolve_flows_for_lsp( ds_clear(actions); ds_put_format(actions, "eth.dst = %s; next;", router_port->lrp_networks.ea_s); - ovn_lflow_add_with_lflow_ref_hint(lflows, peer->od, + ovn_lflow_add_with_hint(lflows, peer->od, S_ROUTER_IN_ARP_RESOLVE, 100, ds_cstr(match), ds_cstr(actions), &op->nbsp->header_, @@ -13079,7 +13314,7 @@ build_arp_resolve_flows_for_lsp( ds_clear(actions); ds_put_format(actions, "eth.dst = %s; next;", router_port->lrp_networks.ea_s); - ovn_lflow_add_with_lflow_ref_hint(lflows, peer->od, + ovn_lflow_add_with_hint(lflows, peer->od, S_ROUTER_IN_ARP_RESOLVE, 100, ds_cstr(match), ds_cstr(actions), &op->nbsp->header_, @@ -13140,7 +13375,8 @@ build_icmperr_pkt_big_flows(struct ovn_port *op, int mtu, struct lflow_table *lflows, const struct shash *meter_groups, struct ds *match, struct ds *actions, enum ovn_stage stage, - struct ovn_port *outport) + struct ovn_port *outport, + struct lflow_ref *lflow_ref) { char *outport_match = outport ? xasprintf("outport == %s && ", outport->json_key) @@ -13183,7 +13419,8 @@ build_icmperr_pkt_big_flows(struct ovn_port *op, int mtu, COPP_ICMP4_ERR, op->od->nbr->copp, meter_groups), - &op->nbrp->header_); + &op->nbrp->header_, + lflow_ref); } char *ip6_src = NULL; @@ -13223,7 +13460,8 @@ build_icmperr_pkt_big_flows(struct ovn_port *op, int mtu, COPP_ICMP6_ERR, op->od->nbr->copp, meter_groups), - &op->nbrp->header_); + &op->nbrp->header_, + lflow_ref); } free(outport_match); } @@ -13234,7 +13472,8 @@ build_check_pkt_len_flows_for_lrp(struct ovn_port *op, const struct hmap *lr_ports, const struct shash *meter_groups, struct ds *match, - struct ds *actions) + struct ds *actions, + struct lflow_ref *lflow_ref) { int gw_mtu = smap_get_int(&op->nbrp->options, "gateway_mtu", 0); if (gw_mtu <= 0) { @@ -13244,12 +13483,13 @@ build_check_pkt_len_flows_for_lrp(struct ovn_port *op, ds_clear(match); ds_put_format(match, "outport == %s", op->json_key); build_gateway_mtu_flow(lflows, op, S_ROUTER_IN_CHK_PKT_LEN, 50, 55, - match, actions, &op->nbrp->header_, "next;"); + match, actions, &op->nbrp->header_, + lflow_ref, "next;"); /* ingress traffic */ build_icmperr_pkt_big_flows(op, gw_mtu, lflows, meter_groups, match, actions, S_ROUTER_IN_IP_INPUT, - NULL); + NULL, lflow_ref); for (size_t i = 0; i < op->od->nbr->n_ports; i++) { struct ovn_port *rp = ovn_port_find(lr_ports, @@ -13261,7 +13501,7 @@ build_check_pkt_len_flows_for_lrp(struct ovn_port *op, /* egress traffic */ build_icmperr_pkt_big_flows(rp, gw_mtu, lflows, meter_groups, match, actions, S_ROUTER_IN_LARGER_PKTS, - op); + op, lflow_ref); } } @@ -13283,15 +13523,16 @@ build_check_pkt_len_flows_for_lrouter( struct ovn_datapath *od, struct lflow_table *lflows, const struct hmap *lr_ports, struct ds *match, struct ds *actions, - const struct shash *meter_groups) + const struct shash *meter_groups, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbr); /* Packets are allowed by default. */ ovn_lflow_add(lflows, od, S_ROUTER_IN_CHK_PKT_LEN, 0, "1", - "next;"); + "next;", lflow_ref); ovn_lflow_add(lflows, od, S_ROUTER_IN_LARGER_PKTS, 0, "1", - "next;"); + "next;", lflow_ref); for (size_t i = 0; i < od->nbr->n_ports; i++) { struct ovn_port *rp = ovn_port_find(lr_ports, @@ -13300,7 +13541,7 @@ build_check_pkt_len_flows_for_lrouter( continue; } build_check_pkt_len_flows_for_lrp(rp, lflows, lr_ports, meter_groups, - match, actions); + match, actions, lflow_ref); } } @@ -13308,7 +13549,8 @@ build_check_pkt_len_flows_for_lrouter( static void build_gateway_redirect_flows_for_lrouter( struct ovn_datapath *od, struct lflow_table *lflows, - struct ds *match, struct ds *actions) + struct ds *match, struct ds *actions, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbr); for (size_t i = 0; i < od->n_l3dgw_ports; i++) { @@ -13341,18 +13583,20 @@ build_gateway_redirect_flows_for_lrouter( od->l3dgw_ports[i]->cr_port->json_key); ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_GW_REDIRECT, 50, ds_cstr(match), ds_cstr(actions), - stage_hint); + stage_hint, lflow_ref); } /* Packets are allowed by default. */ - ovn_lflow_add(lflows, od, S_ROUTER_IN_GW_REDIRECT, 0, "1", "next;"); + ovn_lflow_add(lflows, od, S_ROUTER_IN_GW_REDIRECT, 0, "1", "next;", + lflow_ref); } /* Logical router ingress table GW_REDIRECT: Gateway redirect. */ static void build_lr_gateway_redirect_flows_for_nats( const struct ovn_datapath *od, const struct lr_nat_record *lrnat_rec, - struct lflow_table *lflows, struct ds *match, struct ds *actions) + struct lflow_table *lflows, struct ds *match, struct ds *actions, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbr); for (size_t i = 0; i < od->n_l3dgw_ports; i++) { @@ -13393,21 +13637,23 @@ build_lr_gateway_redirect_flows_for_nats( if (nat->nb->allowed_ext_ips) { ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_GW_REDIRECT, 75, ds_cstr(&match_ext), - ds_cstr(actions), stage_hint); + ds_cstr(actions), stage_hint, + lflow_ref); if (add_def_flow) { ds_clear(&match_ext); ds_put_format(&match_ext, "ip && ip%s.dst == %s", nat_entry_is_v6(nat) ? "6" : "4", nat->nb->external_ip); ovn_lflow_add(lflows, od, S_ROUTER_IN_GW_REDIRECT, 70, - ds_cstr(&match_ext), debug_drop_action()); + ds_cstr(&match_ext), debug_drop_action(), + lflow_ref); add_def_flow = false; } } else if (nat->nb->exempted_ext_ips) { ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_GW_REDIRECT, 75, ds_cstr(&match_ext), debug_drop_action(), - stage_hint); + stage_hint, lflow_ref); } ds_destroy(&match_ext); } @@ -13423,7 +13669,8 @@ static void build_arp_request_flows_for_lrouter( struct ovn_datapath *od, struct lflow_table *lflows, struct ds *match, struct ds *actions, - const struct shash *meter_groups) + const struct shash *meter_groups, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbr); for (int i = 0; i < od->nbr->n_static_routes; i++) { @@ -13465,7 +13712,8 @@ build_arp_request_flows_for_lrouter( copp_meter_get(COPP_ND_NS_RESOLVE, od->nbr->copp, meter_groups), - &route->header_); + &route->header_, + lflow_ref); } ovn_lflow_metered(lflows, od, S_ROUTER_IN_ARP_REQUEST, 100, @@ -13478,7 +13726,8 @@ build_arp_request_flows_for_lrouter( "output; " "};", copp_meter_get(COPP_ARP_RESOLVE, od->nbr->copp, - meter_groups)); + meter_groups), + lflow_ref); ovn_lflow_metered(lflows, od, S_ROUTER_IN_ARP_REQUEST, 100, "eth.dst == 00:00:00:00:00:00 && ip6", "nd_ns { " @@ -13486,8 +13735,10 @@ build_arp_request_flows_for_lrouter( "output; " "};", copp_meter_get(COPP_ND_NS_RESOLVE, od->nbr->copp, - meter_groups)); - ovn_lflow_add(lflows, od, S_ROUTER_IN_ARP_REQUEST, 0, "1", "output;"); + meter_groups), + lflow_ref); + ovn_lflow_add(lflows, od, S_ROUTER_IN_ARP_REQUEST, 0, "1", "output;", + lflow_ref); } /* Logical router egress table DELIVERY: Delivery (priority 100-110). @@ -13500,7 +13751,8 @@ build_arp_request_flows_for_lrouter( static void build_egress_delivery_flows_for_lrouter_port( struct ovn_port *op, struct lflow_table *lflows, - struct ds *match, struct ds *actions) + struct ds *match, struct ds *actions, + struct lflow_ref *lflow_ref) { ovs_assert(op->nbrp); if (!lrport_is_enabled(op->nbrp)) { @@ -13528,20 +13780,23 @@ build_egress_delivery_flows_for_lrouter_port( ds_put_format(actions, "eth.src = %s; output;", op->lrp_networks.ea_s); ovn_lflow_add(lflows, op->od, S_ROUTER_OUT_DELIVERY, 110, - ds_cstr(match), ds_cstr(actions)); + ds_cstr(match), ds_cstr(actions), + lflow_ref); } ds_clear(match); ds_put_format(match, "outport == %s", op->json_key); ovn_lflow_add(lflows, op->od, S_ROUTER_OUT_DELIVERY, 100, - ds_cstr(match), "output;"); + ds_cstr(match), "output;", lflow_ref); - ovn_lflow_add_default_drop(lflows, op->od, S_ROUTER_OUT_DELIVERY); + ovn_lflow_add_default_drop(lflows, op->od, S_ROUTER_OUT_DELIVERY, + lflow_ref); } static void build_misc_local_traffic_drop_flows_for_lrouter( - struct ovn_datapath *od, struct lflow_table *lflows) + struct ovn_datapath *od, struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbr); /* Allow IGMP and MLD packets (with TTL = 1) if the router is @@ -13549,9 +13804,11 @@ build_misc_local_traffic_drop_flows_for_lrouter( */ if (od->mcast_info.rtr.flood_static) { ovn_lflow_add(lflows, od, S_ROUTER_IN_IP_INPUT, 120, - "igmp && ip.ttl == 1", "next;"); + "igmp && ip.ttl == 1", "next;", + lflow_ref); ovn_lflow_add(lflows, od, S_ROUTER_IN_IP_INPUT, 120, - "(mldv1 || mldv2) && ip.ttl == 1", "next;"); + "(mldv1 || mldv2) && ip.ttl == 1", "next;", + lflow_ref); } /* L3 admission control: drop multicast and broadcast source, localhost @@ -13564,7 +13821,8 @@ build_misc_local_traffic_drop_flows_for_lrouter( "ip4.dst == 127.0.0.0/8 || " "ip4.src == 0.0.0.0/8 || " "ip4.dst == 0.0.0.0/8", - debug_drop_action()); + debug_drop_action(), + lflow_ref); /* Drop ARP packets (priority 85). ARP request packets for router's own * IPs are handled with priority-90 flows. @@ -13572,28 +13830,32 @@ build_misc_local_traffic_drop_flows_for_lrouter( * IPs are handled with priority-90 flows. */ ovn_lflow_add(lflows, od, S_ROUTER_IN_IP_INPUT, 85, - "arp || nd", debug_drop_action()); + "arp || nd", debug_drop_action(), + lflow_ref); /* Allow IPv6 multicast traffic that's supposed to reach the * router pipeline (e.g., router solicitations). */ ovn_lflow_add(lflows, od, S_ROUTER_IN_IP_INPUT, 84, "nd_rs || nd_ra", - "next;"); + "next;", lflow_ref); /* Drop other reserved multicast. */ ovn_lflow_add(lflows, od, S_ROUTER_IN_IP_INPUT, 83, - "ip6.mcast_rsvd", debug_drop_action()); + "ip6.mcast_rsvd", debug_drop_action(), + lflow_ref); /* Allow other multicast if relay enabled (priority 82). */ ovn_lflow_add(lflows, od, S_ROUTER_IN_IP_INPUT, 82, "ip4.mcast || ip6.mcast", (od->mcast_info.rtr.relay ? "next;" : - debug_drop_action())); + debug_drop_action()), + lflow_ref); /* Drop Ethernet local broadcast. By definition this traffic should * not be forwarded.*/ ovn_lflow_add(lflows, od, S_ROUTER_IN_IP_INPUT, 50, - "eth.bcast", debug_drop_action()); + "eth.bcast", debug_drop_action(), + lflow_ref); /* Avoid ICMP time exceeded for multicast, silent drop instead. * See RFC1812 section 5.3.1: @@ -13610,21 +13872,25 @@ build_misc_local_traffic_drop_flows_for_lrouter( * (priority-31 flows will send ICMP time exceeded) */ ovn_lflow_add(lflows, od, S_ROUTER_IN_IP_INPUT, 32, "ip.ttl == {0, 1} && !ip.later_frag && " - "(ip4.mcast || ip6.mcast)", debug_drop_action()); + "(ip4.mcast || ip6.mcast)", debug_drop_action(), + lflow_ref); /* TTL discard */ ovn_lflow_add(lflows, od, S_ROUTER_IN_IP_INPUT, 30, - "ip.ttl == {0, 1}", debug_drop_action()); + "ip.ttl == {0, 1}", debug_drop_action(), + lflow_ref); /* Pass other traffic not already handled to the next table for * routing. */ - ovn_lflow_add(lflows, od, S_ROUTER_IN_IP_INPUT, 0, "1", "next;"); + ovn_lflow_add(lflows, od, S_ROUTER_IN_IP_INPUT, 0, "1", "next;", + lflow_ref); } static void build_dhcpv6_reply_flows_for_lrouter_port( struct ovn_port *op, struct lflow_table *lflows, - struct ds *match) + struct ds *match, + struct lflow_ref *lflow_ref) { ovs_assert(op->nbrp); if (op->l3dgw_port) { @@ -13637,7 +13903,8 @@ build_dhcpv6_reply_flows_for_lrouter_port( op->lrp_networks.ipv6_addrs[i].addr_s); ovn_lflow_add(lflows, op->od, S_ROUTER_IN_IP_INPUT, 100, ds_cstr(match), - "reg0 = 0; handle_dhcpv6_reply;"); + "reg0 = 0; handle_dhcpv6_reply;", + lflow_ref); } } @@ -13645,7 +13912,8 @@ static void build_ipv6_input_flows_for_lrouter_port( struct ovn_port *op, struct lflow_table *lflows, struct ds *match, struct ds *actions, - const struct shash *meter_groups) + const struct shash *meter_groups, + struct lflow_ref *lflow_ref) { ovs_assert(op->nbrp); if (is_cr_port(op)) { @@ -13669,7 +13937,7 @@ build_ipv6_input_flows_for_lrouter_port( "next; "; ovn_lflow_add_with_hint(lflows, op->od, S_ROUTER_IN_IP_INPUT, 90, ds_cstr(match), lrp_actions, - &op->nbrp->header_); + &op->nbrp->header_, lflow_ref); } /* ND reply. These flows reply to ND solicitations for the @@ -13690,7 +13958,8 @@ build_ipv6_input_flows_for_lrouter_port( op->lrp_networks.ipv6_addrs[i].addr_s, op->lrp_networks.ipv6_addrs[i].sn_addr_s, REG_INPORT_ETH_ADDR, match, false, 90, - &op->nbrp->header_, lflows, meter_groups); + &op->nbrp->header_, lflows, meter_groups, + lflow_ref); } /* UDP/TCP/SCTP port unreachable */ @@ -13710,7 +13979,8 @@ build_ipv6_input_flows_for_lrouter_port( COPP_TCP_RESET, op->od->nbr->copp, meter_groups), - &op->nbrp->header_); + &op->nbrp->header_, + lflow_ref); ds_clear(match); ds_put_format(match, @@ -13726,7 +13996,8 @@ build_ipv6_input_flows_for_lrouter_port( COPP_TCP_RESET, op->od->nbr->copp, meter_groups), - &op->nbrp->header_); + &op->nbrp->header_, + lflow_ref); ds_clear(match); ds_put_format(match, @@ -13745,7 +14016,8 @@ build_ipv6_input_flows_for_lrouter_port( COPP_ICMP6_ERR, op->od->nbr->copp, meter_groups), - &op->nbrp->header_); + &op->nbrp->header_, + lflow_ref); ds_clear(match); ds_put_format(match, @@ -13764,7 +14036,8 @@ build_ipv6_input_flows_for_lrouter_port( COPP_ICMP6_ERR, op->od->nbr->copp, meter_groups), - &op->nbrp->header_); + &op->nbrp->header_, + lflow_ref); } } @@ -13804,7 +14077,7 @@ build_ipv6_input_flows_for_lrouter_port( 31, ds_cstr(match), ds_cstr(actions), NULL, copp_meter_get(COPP_ICMP6_ERR, op->od->nbr->copp, meter_groups), - &op->nbrp->header_); + &op->nbrp->header_, lflow_ref); } ds_destroy(&ip_ds); } @@ -13813,7 +14086,8 @@ static void build_lrouter_arp_nd_for_datapath(const struct ovn_datapath *od, const struct lr_nat_record *lrnat_rec, struct lflow_table *lflows, - const struct shash *meter_groups) + const struct shash *meter_groups, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbr); if (!od->nbr->n_nat) { @@ -13842,7 +14116,8 @@ build_lrouter_arp_nd_for_datapath(const struct ovn_datapath *od, if (!strcmp(nat_entry->nb->type, "snat")) { continue; } - build_lrouter_nat_arp_nd_flow(od, nat_entry, lflows, meter_groups); + build_lrouter_nat_arp_nd_flow(od, nat_entry, lflows, meter_groups, + lflow_ref); } /* Now handle SNAT entries too, one per unique SNAT IP. */ @@ -13857,7 +14132,8 @@ build_lrouter_arp_nd_for_datapath(const struct ovn_datapath *od, struct ovn_nat *nat_entry = CONTAINER_OF(ovs_list_front(&snat_ip->snat_entries), struct ovn_nat, ext_addr_list_node); - build_lrouter_nat_arp_nd_flow(od, nat_entry, lflows, meter_groups); + build_lrouter_nat_arp_nd_flow(od, nat_entry, lflows, meter_groups, + lflow_ref); } } @@ -13866,7 +14142,8 @@ static void build_lrouter_ipv4_ip_input(struct ovn_port *op, struct lflow_table *lflows, struct ds *match, struct ds *actions, - const struct shash *meter_groups) + const struct shash *meter_groups, + struct lflow_ref *lflow_ref) { ovs_assert(op->nbrp); /* No ingress packets are accepted on a chassisredirect @@ -13884,7 +14161,7 @@ build_lrouter_ipv4_ip_input(struct ovn_port *op, ds_put_cstr(match, " && "REGBIT_EGRESS_LOOPBACK" == 0"); ovn_lflow_add_with_hint(lflows, op->od, S_ROUTER_IN_IP_INPUT, 100, ds_cstr(match), debug_drop_action(), - &op->nbrp->header_); + &op->nbrp->header_, lflow_ref); /* ICMP echo reply. These flows reply to ICMP echo requests * received for the router's IP address. Since packets only @@ -13903,11 +14180,11 @@ build_lrouter_ipv4_ip_input(struct ovn_port *op, "next; "; ovn_lflow_add_with_hint(lflows, op->od, S_ROUTER_IN_IP_INPUT, 90, ds_cstr(match), icmp_actions, - &op->nbrp->header_); + &op->nbrp->header_, lflow_ref); } /* BFD msg handling */ - build_lrouter_bfd_flows(lflows, op, meter_groups); + build_lrouter_bfd_flows(lflows, op, meter_groups, lflow_ref); /* ICMP time exceeded */ struct ds ip_ds = DS_EMPTY_INITIALIZER; @@ -13936,7 +14213,7 @@ build_lrouter_ipv4_ip_input(struct ovn_port *op, 31, ds_cstr(match), ds_cstr(actions), NULL, copp_meter_get(COPP_ICMP4_ERR, op->od->nbr->copp, meter_groups), - &op->nbrp->header_); + &op->nbrp->header_, lflow_ref); } ds_destroy(&ip_ds); @@ -13986,7 +14263,7 @@ build_lrouter_ipv4_ip_input(struct ovn_port *op, build_lrouter_arp_flow(op->od, op, op->lrp_networks.ipv4_addrs[i].addr_s, REG_INPORT_ETH_ADDR, match, false, 90, - &op->nbrp->header_, lflows); + &op->nbrp->header_, lflows, lflow_ref); } if (!op->od->is_gw_router && !op->od->n_l3dgw_ports) { @@ -14009,7 +14286,8 @@ build_lrouter_ipv4_ip_input(struct ovn_port *op, COPP_ICMP4_ERR, op->od->nbr->copp, meter_groups), - &op->nbrp->header_); + &op->nbrp->header_, + lflow_ref); ds_clear(match); ds_put_format(match, @@ -14025,7 +14303,8 @@ build_lrouter_ipv4_ip_input(struct ovn_port *op, COPP_TCP_RESET, op->od->nbr->copp, meter_groups), - &op->nbrp->header_); + &op->nbrp->header_, + lflow_ref); ds_clear(match); ds_put_format(match, @@ -14041,7 +14320,8 @@ build_lrouter_ipv4_ip_input(struct ovn_port *op, COPP_TCP_RESET, op->od->nbr->copp, meter_groups), - &op->nbrp->header_); + &op->nbrp->header_, + lflow_ref); ds_clear(match); ds_put_format(match, @@ -14060,7 +14340,8 @@ build_lrouter_ipv4_ip_input(struct ovn_port *op, COPP_ICMP4_ERR, op->od->nbr->copp, meter_groups), - &op->nbrp->header_); + &op->nbrp->header_, + lflow_ref); } } } @@ -14070,7 +14351,8 @@ static void build_lrouter_ipv4_ip_input_for_lbnats( struct ovn_port *op, struct lflow_table *lflows, const struct lr_stateful_record *lr_stateful_rec, - struct ds *match, const struct shash *meter_groups) + struct ds *match, const struct shash *meter_groups, + struct lflow_ref *lflow_ref) { ovs_assert(op->nbrp); /* No ingress packets are accepted on a chassisredirect @@ -14091,7 +14373,7 @@ build_lrouter_ipv4_ip_input_for_lbnats( AF_INET); build_lrouter_arp_flow(op->od, op, lb_ips_v4_as, REG_INPORT_ETH_ADDR, - match, false, 90, NULL, lflows); + match, false, 90, NULL, lflows, lflow_ref); free(lb_ips_v4_as); } @@ -14108,7 +14390,7 @@ build_lrouter_ipv4_ip_input_for_lbnats( AF_INET6); build_lrouter_nd_flow(op->od, op, "nd_na", lb_ips_v6_as, NULL, REG_INPORT_ETH_ADDR, match, false, 90, - NULL, lflows, meter_groups); + NULL, lflows, meter_groups, lflow_ref); free(lb_ips_v6_as); } @@ -14141,7 +14423,7 @@ build_lrouter_ipv4_ip_input_for_lbnats( continue; } build_lrouter_port_nat_arp_nd_flow(op, nat_entry, lflows, - meter_groups); + meter_groups, lflow_ref); } /* Now handle SNAT entries too, one per unique SNAT IP. */ @@ -14157,7 +14439,7 @@ build_lrouter_ipv4_ip_input_for_lbnats( CONTAINER_OF(ovs_list_front(&snat_ip->snat_entries), struct ovn_nat, ext_addr_list_node); build_lrouter_port_nat_arp_nd_flow(op, nat_entry, lflows, - meter_groups); + meter_groups, lflow_ref); } } @@ -14193,7 +14475,8 @@ build_lrouter_in_unsnat_stateless_flow(struct lflow_table *lflows, const struct nbrec_nat *nat, struct ds *match, bool distributed_nat, bool is_v6, - struct ovn_port *l3dgw_port) + struct ovn_port *l3dgw_port, + struct lflow_ref *lflow_ref) { if (strcmp(nat->type, "snat") && strcmp(nat->type, "dnat_and_snat")) { return; @@ -14206,7 +14489,7 @@ build_lrouter_in_unsnat_stateless_flow(struct lflow_table *lflows, ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_UNSNAT, priority, ds_cstr(match), "next;", - &nat->header_); + &nat->header_, lflow_ref); } static void @@ -14214,7 +14497,8 @@ build_lrouter_in_unsnat_in_czone_flow(struct lflow_table *lflows, const struct ovn_datapath *od, const struct nbrec_nat *nat, struct ds *match, bool distributed_nat, - bool is_v6, struct ovn_port *l3dgw_port) + bool is_v6, struct ovn_port *l3dgw_port, + struct lflow_ref *lflow_ref) { if (strcmp(nat->type, "snat") && strcmp(nat->type, "dnat_and_snat")) { return; @@ -14232,7 +14516,7 @@ build_lrouter_in_unsnat_in_czone_flow(struct lflow_table *lflows, ds_put_cstr(match, " && flags.loopback == 0"); ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_UNSNAT, 100, ds_cstr(match), "ct_snat_in_czone;", - &nat->header_); + &nat->header_, lflow_ref); ds_truncate(match, common_match_len); /* Update common zone match for the hairpin traffic. */ @@ -14240,7 +14524,7 @@ build_lrouter_in_unsnat_in_czone_flow(struct lflow_table *lflows, ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_UNSNAT, 100, ds_cstr(match), "ct_snat;", - &nat->header_); + &nat->header_, lflow_ref); } static void @@ -14248,7 +14532,8 @@ build_lrouter_in_unsnat_flow(struct lflow_table *lflows, const struct ovn_datapath *od, const struct nbrec_nat *nat, struct ds *match, bool distributed_nat, bool is_v6, - struct ovn_port *l3dgw_port) + struct ovn_port *l3dgw_port, + struct lflow_ref *lflow_ref) { if (strcmp(nat->type, "snat") && strcmp(nat->type, "dnat_and_snat")) { @@ -14262,7 +14547,7 @@ build_lrouter_in_unsnat_flow(struct lflow_table *lflows, ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_UNSNAT, priority, ds_cstr(match), "ct_snat;", - &nat->header_); + &nat->header_, lflow_ref); } static void @@ -14272,7 +14557,8 @@ build_lrouter_in_dnat_flow(struct lflow_table *lflows, const struct nbrec_nat *nat, struct ds *match, struct ds *actions, bool distributed_nat, int cidr_bits, bool is_v6, - struct ovn_port *l3dgw_port, bool stateless) + struct ovn_port *l3dgw_port, bool stateless, + struct lflow_ref *lflow_ref) { /* Ingress DNAT table: Packets enter the pipeline with destination * IP address that needs to be DNATted from a external IP address @@ -14318,7 +14604,8 @@ build_lrouter_in_dnat_flow(struct lflow_table *lflows, if (nat->allowed_ext_ips || nat->exempted_ext_ips) { lrouter_nat_add_ext_ip_match(od, lflows, match, nat, - is_v6, true, cidr_bits); + is_v6, true, cidr_bits, + lflow_ref); } if (stateless) { @@ -14334,7 +14621,7 @@ build_lrouter_in_dnat_flow(struct lflow_table *lflows, ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_DNAT, 100, ds_cstr(match), ds_cstr(actions), - &nat->header_); + &nat->header_, lflow_ref); } static void @@ -14343,7 +14630,8 @@ build_lrouter_out_undnat_flow(struct lflow_table *lflows, const struct nbrec_nat *nat, struct ds *match, struct ds *actions, bool distributed_nat, struct eth_addr mac, bool is_v6, - struct ovn_port *l3dgw_port, bool stateless) + struct ovn_port *l3dgw_port, bool stateless, + struct lflow_ref *lflow_ref) { /* Egress UNDNAT table: It is for already established connections' * reverse traffic. i.e., DNAT has already been done in ingress @@ -14385,7 +14673,7 @@ build_lrouter_out_undnat_flow(struct lflow_table *lflows, ovn_lflow_add_with_hint(lflows, od, S_ROUTER_OUT_UNDNAT, 100, ds_cstr(match), ds_cstr(actions), - &nat->header_); + &nat->header_, lflow_ref); } static void @@ -14393,7 +14681,8 @@ build_lrouter_out_is_dnat_local(struct lflow_table *lflows, const struct ovn_datapath *od, const struct nbrec_nat *nat, struct ds *match, struct ds *actions, bool distributed_nat, - bool is_v6, struct ovn_port *l3dgw_port) + bool is_v6, struct ovn_port *l3dgw_port, + struct lflow_ref *lflow_ref) { /* Note that this only applies for NAT on a distributed router. */ @@ -14416,7 +14705,7 @@ build_lrouter_out_is_dnat_local(struct lflow_table *lflows, ovn_lflow_add_with_hint(lflows, od, S_ROUTER_OUT_CHECK_DNAT_LOCAL, 50, ds_cstr(match), ds_cstr(actions), - &nat->header_); + &nat->header_, lflow_ref); } static void @@ -14424,7 +14713,8 @@ build_lrouter_out_snat_match(struct lflow_table *lflows, const struct ovn_datapath *od, const struct nbrec_nat *nat, struct ds *match, bool distributed_nat, int cidr_bits, bool is_v6, - struct ovn_port *l3dgw_port) + struct ovn_port *l3dgw_port, + struct lflow_ref *lflow_ref) { ds_clear(match); @@ -14444,7 +14734,8 @@ build_lrouter_out_snat_match(struct lflow_table *lflows, if (nat->allowed_ext_ips || nat->exempted_ext_ips) { lrouter_nat_add_ext_ip_match(od, lflows, match, nat, - is_v6, false, cidr_bits); + is_v6, false, cidr_bits, + lflow_ref); } } @@ -14455,7 +14746,8 @@ build_lrouter_out_snat_stateless_flow(struct lflow_table *lflows, struct ds *match, struct ds *actions, bool distributed_nat, struct eth_addr mac, int cidr_bits, - bool is_v6, struct ovn_port *l3dgw_port) + bool is_v6, struct ovn_port *l3dgw_port, + struct lflow_ref *lflow_ref) { if (strcmp(nat->type, "snat") && strcmp(nat->type, "dnat_and_snat")) { return; @@ -14469,7 +14761,7 @@ build_lrouter_out_snat_stateless_flow(struct lflow_table *lflows, uint16_t priority = cidr_bits + 1; build_lrouter_out_snat_match(lflows, od, nat, match, distributed_nat, - cidr_bits, is_v6, l3dgw_port); + cidr_bits, is_v6, l3dgw_port, lflow_ref); if (!od->is_gw_router) { /* Distributed router. */ @@ -14488,7 +14780,8 @@ build_lrouter_out_snat_stateless_flow(struct lflow_table *lflows, ovn_lflow_add_with_hint(lflows, od, S_ROUTER_OUT_SNAT, priority, ds_cstr(match), - ds_cstr(actions), &nat->header_); + ds_cstr(actions), &nat->header_, + lflow_ref); } static void @@ -14498,7 +14791,8 @@ build_lrouter_out_snat_in_czone_flow(struct lflow_table *lflows, struct ds *match, struct ds *actions, bool distributed_nat, struct eth_addr mac, int cidr_bits, - bool is_v6, struct ovn_port *l3dgw_port) + bool is_v6, struct ovn_port *l3dgw_port, + struct lflow_ref *lflow_ref) { if (strcmp(nat->type, "snat") && strcmp(nat->type, "dnat_and_snat")) { return; @@ -14513,7 +14807,8 @@ build_lrouter_out_snat_in_czone_flow(struct lflow_table *lflows, struct ds zone_actions = DS_EMPTY_INITIALIZER; build_lrouter_out_snat_match(lflows, od, nat, match, distributed_nat, - cidr_bits, is_v6, l3dgw_port); + cidr_bits, is_v6, l3dgw_port, + lflow_ref); if (od->n_l3dgw_ports) { priority += 128; @@ -14542,13 +14837,15 @@ build_lrouter_out_snat_in_czone_flow(struct lflow_table *lflows, ovn_lflow_add_with_hint(lflows, od, S_ROUTER_OUT_SNAT, priority, ds_cstr(match), - ds_cstr(actions), &nat->header_); + ds_cstr(actions), &nat->header_, + lflow_ref); ds_put_cstr(match, " && "REGBIT_DST_NAT_IP_LOCAL" == 1"); ovn_lflow_add_with_hint(lflows, od, S_ROUTER_OUT_SNAT, priority + 1, ds_cstr(match), - ds_cstr(&zone_actions), &nat->header_); + ds_cstr(&zone_actions), &nat->header_, + lflow_ref); ds_destroy(&zone_actions); } @@ -14559,7 +14856,8 @@ build_lrouter_out_snat_flow(struct lflow_table *lflows, const struct nbrec_nat *nat, struct ds *match, struct ds *actions, bool distributed_nat, struct eth_addr mac, int cidr_bits, bool is_v6, - struct ovn_port *l3dgw_port) + struct ovn_port *l3dgw_port, + struct lflow_ref *lflow_ref) { if (strcmp(nat->type, "snat") && strcmp(nat->type, "dnat_and_snat")) { return; @@ -14573,7 +14871,7 @@ build_lrouter_out_snat_flow(struct lflow_table *lflows, uint16_t priority = cidr_bits + 1; build_lrouter_out_snat_match(lflows, od, nat, match, distributed_nat, - cidr_bits, is_v6, l3dgw_port); + cidr_bits, is_v6, l3dgw_port, lflow_ref); if (!od->is_gw_router) { /* Distributed router. */ @@ -14596,7 +14894,8 @@ build_lrouter_out_snat_flow(struct lflow_table *lflows, ovn_lflow_add_with_hint(lflows, od, S_ROUTER_OUT_SNAT, priority, ds_cstr(match), - ds_cstr(actions), &nat->header_); + ds_cstr(actions), &nat->header_, + lflow_ref); } static void @@ -14606,7 +14905,8 @@ build_lrouter_ingress_nat_check_pkt_len(struct lflow_table *lflows, bool is_v6, struct ds *match, struct ds *actions, int mtu, struct ovn_port *l3dgw_port, - const struct shash *meter_groups) + const struct shash *meter_groups, + struct lflow_ref *lflow_ref) { ds_clear(match); ds_put_format(match, "inport == %s && "REGBIT_PKT_LARGER @@ -14640,7 +14940,8 @@ build_lrouter_ingress_nat_check_pkt_len(struct lflow_table *lflows, COPP_ICMP4_ERR, od->nbr->copp, meter_groups), - &nat->header_); + &nat->header_, + lflow_ref); } else { ds_put_format(match, " && ip6 && ip6.dst == %s", nat->external_ip); /* Set icmp6.frag_mtu to gw_mtu */ @@ -14667,7 +14968,8 @@ build_lrouter_ingress_nat_check_pkt_len(struct lflow_table *lflows, COPP_ICMP6_ERR, od->nbr->copp, meter_groups), - &nat->header_); + &nat->header_, + lflow_ref); } } @@ -14678,7 +14980,8 @@ build_lrouter_ingress_flow(struct lflow_table *lflows, struct ds *actions, struct eth_addr mac, bool distributed_nat, bool is_v6, struct ovn_port *l3dgw_port, - const struct shash *meter_groups) + const struct shash *meter_groups, + struct lflow_ref *lflow_ref) { if (od->n_l3dgw_ports && !strcmp(nat->type, "snat")) { ds_clear(match); @@ -14688,7 +14991,7 @@ build_lrouter_ingress_flow(struct lflow_table *lflows, is_v6 ? "ip6.src" : "ip4.src", nat->external_ip); ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_IP_INPUT, 120, ds_cstr(match), "next;", - &nat->header_); + &nat->header_, lflow_ref); } /* Logical router ingress table 0: * For NAT on a distributed router, add rules allowing @@ -14712,12 +15015,14 @@ build_lrouter_ingress_flow(struct lflow_table *lflows, build_gateway_mtu_flow(lflows, l3dgw_port, S_ROUTER_IN_ADMISSION, 50, 55, match, actions, &nat->header_, + lflow_ref, REG_INPORT_ETH_ADDR " = %s; next;", l3dgw_port->lrp_networks.ea_s); if (gw_mtu) { build_lrouter_ingress_nat_check_pkt_len(lflows, nat, od, is_v6, match, actions, gw_mtu, - l3dgw_port, meter_groups); + l3dgw_port, meter_groups, + lflow_ref); } } } @@ -14852,27 +15157,33 @@ lrouter_check_nat_entry(const struct ovn_datapath *od, /* NAT, Defrag and load balancing. */ static void build_lr_nat_defrag_and_lb_default_flows(struct ovn_datapath *od, - struct lflow_table *lflows) + struct lflow_table *lflows, + struct lflow_ref *lflow_ref) { ovs_assert(od->nbr); /* Packets are allowed by default. */ - ovn_lflow_add(lflows, od, S_ROUTER_IN_DEFRAG, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_ROUTER_IN_UNSNAT, 0, "1", "next;"); + ovn_lflow_add(lflows, od, S_ROUTER_IN_DEFRAG, 0, "1", "next;", lflow_ref); + ovn_lflow_add(lflows, od, S_ROUTER_IN_UNSNAT, 0, "1", "next;", lflow_ref); ovn_lflow_add(lflows, od, S_ROUTER_OUT_CHECK_DNAT_LOCAL, 0, "1", - REGBIT_DST_NAT_IP_LOCAL" = 0; next;"); - ovn_lflow_add(lflows, od, S_ROUTER_OUT_SNAT, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_ROUTER_IN_DNAT, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_ROUTER_OUT_UNDNAT, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_ROUTER_OUT_POST_UNDNAT, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_ROUTER_OUT_POST_SNAT, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_ROUTER_OUT_EGR_LOOP, 0, "1", "next;"); - ovn_lflow_add(lflows, od, S_ROUTER_IN_ECMP_STATEFUL, 0, "1", "next;"); + REGBIT_DST_NAT_IP_LOCAL" = 0; next;", lflow_ref); + ovn_lflow_add(lflows, od, S_ROUTER_OUT_SNAT, 0, "1", "next;", lflow_ref); + ovn_lflow_add(lflows, od, S_ROUTER_IN_DNAT, 0, "1", "next;", lflow_ref); + ovn_lflow_add(lflows, od, S_ROUTER_OUT_UNDNAT, 0, "1", "next;", lflow_ref); + ovn_lflow_add(lflows, od, S_ROUTER_OUT_POST_UNDNAT, 0, "1", "next;", + lflow_ref); + ovn_lflow_add(lflows, od, S_ROUTER_OUT_POST_SNAT, 0, "1", "next;", + lflow_ref); + ovn_lflow_add(lflows, od, S_ROUTER_OUT_EGR_LOOP, 0, "1", "next;", + lflow_ref); + ovn_lflow_add(lflows, od, S_ROUTER_IN_ECMP_STATEFUL, 0, "1", "next;", + lflow_ref); /* Send the IPv6 NS packets to next table. When ovn-controller * generates IPv6 NS (for the action - nd_ns{}), the injected * packet would go through conntrack - which is not required. */ - ovn_lflow_add(lflows, od, S_ROUTER_OUT_SNAT, 120, "nd_ns", "next;"); + ovn_lflow_add(lflows, od, S_ROUTER_OUT_SNAT, 120, "nd_ns", "next;", + lflow_ref); } static void @@ -14882,7 +15193,8 @@ build_lrouter_nat_defrag_and_lb( const struct hmap *ls_ports, const struct hmap *lr_ports, struct ds *match, struct ds *actions, const struct shash *meter_groups, - const struct chassis_features *features) + const struct chassis_features *features, + struct lflow_ref *lflow_ref) { const struct ovn_datapath *od = lr_stateful_rec->od; ovs_assert(od->nbr); @@ -14907,16 +15219,18 @@ build_lrouter_nat_defrag_and_lb( ds_put_format(match, " && %s.skip_snat == 1", ct_flag_reg); ovn_lflow_add(lflows, od, S_ROUTER_IN_DNAT, 70, ds_cstr(match), - "flags.skip_snat_for_lb = 1; ct_commit_nat;"); + "flags.skip_snat_for_lb = 1; ct_commit_nat;", + lflow_ref); ds_truncate(match, match_len); ds_put_format(match, " && %s.force_snat == 1", ct_flag_reg); ovn_lflow_add(lflows, od, S_ROUTER_IN_DNAT, 70, ds_cstr(match), - "flags.force_snat_for_lb = 1; ct_commit_nat;"); + "flags.force_snat_for_lb = 1; ct_commit_nat;", + lflow_ref); ds_truncate(match, match_len); ovn_lflow_add(lflows, od, S_ROUTER_IN_DNAT, 50, ds_cstr(match), - "ct_commit_nat;"); + "ct_commit_nat;", lflow_ref); } /* Ingress DNAT (Priority 50/70). @@ -14933,16 +15247,18 @@ build_lrouter_nat_defrag_and_lb( ds_put_format(match, " && %s.skip_snat == 1", ct_flag_reg); ovn_lflow_add(lflows, od, S_ROUTER_IN_DNAT, 70, ds_cstr(match), - "flags.skip_snat_for_lb = 1; next;"); + "flags.skip_snat_for_lb = 1; next;", + lflow_ref); ds_truncate(match, match_len); ds_put_format(match, " && %s.force_snat == 1", ct_flag_reg); ovn_lflow_add(lflows, od, S_ROUTER_IN_DNAT, 70, ds_cstr(match), - "flags.force_snat_for_lb = 1; next;"); + "flags.force_snat_for_lb = 1; next;", + lflow_ref); ds_truncate(match, match_len); ovn_lflow_add(lflows, od, S_ROUTER_IN_DNAT, 50, ds_cstr(match), - "next;"); + "next;", lflow_ref); } /* If the router has load balancer or DNAT rules, re-circulate every packet @@ -14957,11 +15273,14 @@ build_lrouter_nat_defrag_and_lb( if (od->is_gw_router && (od->nbr->n_nat || lr_stateful_rec->has_lb_vip)) { /* Do not send ND or ICMP packets to connection tracking. */ ovn_lflow_add(lflows, od, S_ROUTER_OUT_UNDNAT, 100, - "nd || nd_rs || nd_ra", "next;"); + "nd || nd_rs || nd_ra", "next;", + lflow_ref); ovn_lflow_add(lflows, od, S_ROUTER_OUT_UNDNAT, 50, - "ip", "flags.loopback = 1; ct_dnat;"); + "ip", "flags.loopback = 1; ct_dnat;", + lflow_ref); ovn_lflow_add(lflows, od, S_ROUTER_OUT_POST_UNDNAT, 50, - "ip && ct.new", "ct_commit { } ; next; "); + "ip && ct.new", "ct_commit { } ; next; ", + lflow_ref); } /* NAT rules are only valid on Gateway routers and routers with @@ -15010,19 +15329,20 @@ build_lrouter_nat_defrag_and_lb( if (stateless) { build_lrouter_in_unsnat_stateless_flow(lflows, od, nat, match, distributed_nat, is_v6, - l3dgw_port); + l3dgw_port, lflow_ref); } else if (lrouter_use_common_zone(od)) { build_lrouter_in_unsnat_in_czone_flow(lflows, od, nat, match, distributed_nat, is_v6, - l3dgw_port); + l3dgw_port, lflow_ref); } else { build_lrouter_in_unsnat_flow(lflows, od, nat, match, - distributed_nat, is_v6, l3dgw_port); + distributed_nat, is_v6, l3dgw_port, + lflow_ref); } /* S_ROUTER_IN_DNAT */ build_lrouter_in_dnat_flow(lflows, od, lrnat_rec, nat, match, actions, distributed_nat, cidr_bits, is_v6, - l3dgw_port, stateless); + l3dgw_port, stateless, lflow_ref); /* ARP resolve for NAT IPs. */ if (od->is_gw_router) { @@ -15045,7 +15365,8 @@ build_lrouter_nat_defrag_and_lb( S_ROUTER_IN_ARP_RESOLVE, 150, ds_cstr(match), debug_drop_action(), - &nat->header_); + &nat->header_, + lflow_ref); /* Now for packets coming from other (downlink) LRPs, allow ARP * resolve for the NAT IP, so that such packets can be * forwarded for E/W NAT. */ @@ -15064,7 +15385,8 @@ build_lrouter_nat_defrag_and_lb( S_ROUTER_IN_ARP_RESOLVE, 100, ds_cstr(match), ds_cstr(actions), - &nat->header_); + &nat->header_, + lflow_ref); if (od->redirect_bridged && distributed_nat) { ds_clear(match); ds_put_format( @@ -15085,7 +15407,8 @@ build_lrouter_nat_defrag_and_lb( ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_ARP_RESOLVE, 90, ds_cstr(match), ds_cstr(actions), - &nat->header_); + &nat->header_, + lflow_ref); } sset_add(&nat_entries, nat->external_ip); } @@ -15095,13 +15418,13 @@ build_lrouter_nat_defrag_and_lb( /* S_ROUTER_OUT_DNAT_LOCAL */ build_lrouter_out_is_dnat_local(lflows, od, nat, match, actions, distributed_nat, is_v6, - l3dgw_port); + l3dgw_port, lflow_ref); } /* S_ROUTER_OUT_UNDNAT */ build_lrouter_out_undnat_flow(lflows, od, nat, match, actions, distributed_nat, mac, is_v6, l3dgw_port, - stateless); + stateless, lflow_ref); /* S_ROUTER_OUT_SNAT * Egress SNAT table: Packets enter the egress pipeline with * source ip address that needs to be SNATted to a external ip @@ -15110,21 +15433,22 @@ build_lrouter_nat_defrag_and_lb( build_lrouter_out_snat_stateless_flow(lflows, od, nat, match, actions, distributed_nat, mac, cidr_bits, is_v6, - l3dgw_port); + l3dgw_port, lflow_ref); } else if (lrouter_use_common_zone(od)) { build_lrouter_out_snat_in_czone_flow(lflows, od, nat, match, actions, distributed_nat, mac, - cidr_bits, is_v6, l3dgw_port); + cidr_bits, is_v6, l3dgw_port, + lflow_ref); } else { build_lrouter_out_snat_flow(lflows, od, nat, match, actions, distributed_nat, mac, cidr_bits, is_v6, - l3dgw_port); + l3dgw_port, lflow_ref); } /* S_ROUTER_IN_ADMISSION - S_ROUTER_IN_IP_INPUT */ build_lrouter_ingress_flow(lflows, od, nat, match, actions, mac, distributed_nat, is_v6, l3dgw_port, - meter_groups); + meter_groups, lflow_ref); /* Ingress Gateway Redirect Table: For NAT on a distributed * router, add flows that are specific to a NAT rule. These @@ -15151,7 +15475,8 @@ build_lrouter_nat_defrag_and_lb( if (op && op->nbsp && !strcmp(op->nbsp->type, "virtual")) { ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_GW_REDIRECT, 80, ds_cstr(match), - debug_drop_action(), &nat->header_); + debug_drop_action(), &nat->header_, + lflow_ref); } ds_put_format(match, " && is_chassis_resident(\"%s\")", nat->logical_port); @@ -15161,7 +15486,8 @@ build_lrouter_nat_defrag_and_lb( nat->external_ip); ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_GW_REDIRECT, 100, ds_cstr(match), - ds_cstr(actions), &nat->header_); + ds_cstr(actions), &nat->header_, + lflow_ref); } /* Egress Loopback table: For NAT on a distributed router. @@ -15202,7 +15528,7 @@ build_lrouter_nat_defrag_and_lb( ovn_stage_get_table(S_ROUTER_IN_ADMISSION)); ovn_lflow_add_with_hint(lflows, od, S_ROUTER_OUT_EGR_LOOP, 100, ds_cstr(match), ds_cstr(actions), - &nat->header_); + &nat->header_, lflow_ref); } } @@ -15222,7 +15548,7 @@ build_lrouter_nat_defrag_and_lb( ds_put_cstr(actions, REGBIT_DST_NAT_IP_LOCAL" = 1; next;"); ovn_lflow_add_with_hint(lflows, od, S_ROUTER_OUT_CHECK_DNAT_LOCAL, 50, ds_cstr(match), ds_cstr(actions), - &od->nbr->header_); + &od->nbr->header_, lflow_ref); } @@ -15232,22 +15558,24 @@ build_lrouter_nat_defrag_and_lb( if (lrnat_rec->dnat_force_snat_addrs.n_ipv4_addrs) { build_lrouter_force_snat_flows(lflows, od, "4", lrnat_rec->dnat_force_snat_addrs.ipv4_addrs[0].addr_s, - "dnat"); + "dnat", lflow_ref); } if (lrnat_rec->dnat_force_snat_addrs.n_ipv6_addrs) { build_lrouter_force_snat_flows(lflows, od, "6", lrnat_rec->dnat_force_snat_addrs.ipv6_addrs[0].addr_s, - "dnat"); + "dnat", lflow_ref); } } if (lb_force_snat_ip) { if (lrnat_rec->lb_force_snat_addrs.n_ipv4_addrs) { build_lrouter_force_snat_flows(lflows, od, "4", - lrnat_rec->lb_force_snat_addrs.ipv4_addrs[0].addr_s, "lb"); + lrnat_rec->lb_force_snat_addrs.ipv4_addrs[0].addr_s, "lb", + lflow_ref); } if (lrnat_rec->lb_force_snat_addrs.n_ipv6_addrs) { build_lrouter_force_snat_flows(lflows, od, "6", - lrnat_rec->lb_force_snat_addrs.ipv6_addrs[0].addr_s, "lb"); + lrnat_rec->lb_force_snat_addrs.ipv6_addrs[0].addr_s, "lb", + lflow_ref); } } } @@ -15319,7 +15647,8 @@ build_lrp_lflows_for_lbnats(struct ovn_port *op, */ if (!lr_stateful_rec->lrnat_rec->lb_force_snat_router_ip) { build_lrouter_drop_own_dest(op, lr_stateful_rec, - S_ROUTER_IN_IP_INPUT, 60, false, lflows); + S_ROUTER_IN_IP_INPUT, 60, false, lflows, + op->stateful_lflow_ref); } /* Drop IP traffic destined to router owned IPs. Part of it is dropped @@ -15329,12 +15658,14 @@ build_lrp_lflows_for_lbnats(struct ovn_port *op, * Priority 2. */ build_lrouter_drop_own_dest(op, lr_stateful_rec, - S_ROUTER_IN_ARP_RESOLVE, 2, true, lflows); + S_ROUTER_IN_ARP_RESOLVE, 2, true, + lflows, op->stateful_lflow_ref); build_lrouter_ipv4_ip_input_for_lbnats(op, lflows, lr_stateful_rec, - match, meter_groups); + match, meter_groups, + op->stateful_lflow_ref); build_lrouter_force_snat_flows_op(op, lr_stateful_rec->lrnat_rec, lflows, - match, actions); + match, actions, op->stateful_lflow_ref); } static void @@ -15366,13 +15697,13 @@ build_lr_stateful_flows(const struct lr_stateful_record *lr_stateful_rec, { build_lrouter_nat_defrag_and_lb(lr_stateful_rec, lflows, ls_ports, lr_ports, match, actions, - meter_groups, features); + meter_groups, features, NULL); build_lr_gateway_redirect_flows_for_nats(lr_stateful_rec->od, lr_stateful_rec->lrnat_rec, - lflows, match, actions); + lflows, match, actions, NULL); build_lrouter_arp_nd_for_datapath(lr_stateful_rec->od, lr_stateful_rec->lrnat_rec, lflows, - meter_groups); + meter_groups, NULL); } static void @@ -15384,11 +15715,13 @@ build_ls_stateful_flows(const struct ls_stateful_record *ls_stateful_rec, { ovs_assert(ls_stateful_rec->od); - build_ls_stateful_rec_pre_acls(ls_stateful_rec, ls_pgs, lflows); - build_ls_stateful_rec_pre_lb(ls_stateful_rec, lflows); - build_acl_hints(ls_stateful_rec, features, lflows); - build_acls(ls_stateful_rec, features, lflows, ls_pgs, meter_groups); - build_lb_hairpin(ls_stateful_rec, lflows); + build_ls_stateful_rec_pre_acls(ls_stateful_rec, ls_pgs, lflows, + NULL); + build_ls_stateful_rec_pre_lb(ls_stateful_rec, lflows, NULL); + build_acl_hints(ls_stateful_rec, features, lflows, NULL); + build_acls(ls_stateful_rec, features, lflows, ls_pgs, meter_groups, + NULL); + build_lb_hairpin(ls_stateful_rec, lflows, NULL); } struct lswitch_flow_build_info { @@ -15424,19 +15757,20 @@ build_lswitch_and_lrouter_iterate_by_ls(struct ovn_datapath *od, { ovs_assert(od->nbs); build_lswitch_lflows_pre_acl_and_acl(od, lsi->features, lsi->lflows, - lsi->meter_groups); - - build_fwd_group_lflows(od, lsi->lflows); - build_lswitch_lflows_admission_control(od, lsi->lflows); - build_lswitch_learn_fdb_od(od, lsi->lflows); - build_lswitch_arp_nd_responder_default(od, lsi->lflows); - build_lswitch_dns_lookup_and_response(od, lsi->lflows, lsi->meter_groups); - build_lswitch_dhcp_and_dns_defaults(od, lsi->lflows); + lsi->meter_groups, NULL); + + build_fwd_group_lflows(od, lsi->lflows, NULL); + build_lswitch_lflows_admission_control(od, lsi->lflows, NULL); + build_lswitch_learn_fdb_od(od, lsi->lflows, NULL); + build_lswitch_arp_nd_responder_default(od, lsi->lflows, NULL); + build_lswitch_dns_lookup_and_response(od, lsi->lflows, lsi->meter_groups, + NULL); + build_lswitch_dhcp_and_dns_defaults(od, lsi->lflows, NULL); build_lswitch_destination_lookup_bmcast(od, lsi->lflows, &lsi->actions, - lsi->meter_groups); - build_lswitch_output_port_sec_od(od, lsi->lflows); - build_lswitch_lb_affinity_default_flows(od, lsi->lflows); - build_lswitch_lflows_l2_unknown(od, lsi->lflows); + lsi->meter_groups, NULL); + build_lswitch_output_port_sec_od(od, lsi->lflows, NULL); + build_lswitch_lb_affinity_default_flows(od, lsi->lflows, NULL); + build_lswitch_lflows_l2_unknown(od, lsi->lflows, NULL); } /* Helper function to combine all lflow generation which is iterated by @@ -15447,29 +15781,34 @@ build_lswitch_and_lrouter_iterate_by_lr(struct ovn_datapath *od, struct lswitch_flow_build_info *lsi) { ovs_assert(od->nbr); - build_adm_ctrl_flows_for_lrouter(od, lsi->lflows); + build_adm_ctrl_flows_for_lrouter(od, lsi->lflows, NULL); build_neigh_learning_flows_for_lrouter(od, lsi->lflows, &lsi->match, - &lsi->actions, lsi->meter_groups); - build_ND_RA_flows_for_lrouter(od, lsi->lflows); - build_ip_routing_pre_flows_for_lrouter(od, lsi->lflows); + &lsi->actions, + lsi->meter_groups, NULL); + build_ND_RA_flows_for_lrouter(od, lsi->lflows, NULL); + build_ip_routing_pre_flows_for_lrouter(od, lsi->lflows, NULL); build_static_route_flows_for_lrouter(od, lsi->features, lsi->lflows, lsi->lr_ports, - lsi->bfd_connections); + lsi->bfd_connections, + NULL); build_mcast_lookup_flows_for_lrouter(od, lsi->lflows, &lsi->match, - &lsi->actions); - build_ingress_policy_flows_for_lrouter(od, lsi->lflows, lsi->lr_ports); - build_arp_resolve_flows_for_lrouter(od, lsi->lflows); + &lsi->actions, NULL); + build_ingress_policy_flows_for_lrouter(od, lsi->lflows, lsi->lr_ports, + NULL); + build_arp_resolve_flows_for_lrouter(od, lsi->lflows, NULL); build_check_pkt_len_flows_for_lrouter(od, lsi->lflows, lsi->lr_ports, &lsi->match, &lsi->actions, - lsi->meter_groups); + lsi->meter_groups, NULL); build_gateway_redirect_flows_for_lrouter(od, lsi->lflows, &lsi->match, - &lsi->actions); + &lsi->actions, NULL); build_arp_request_flows_for_lrouter(od, lsi->lflows, &lsi->match, - &lsi->actions, lsi->meter_groups); - build_misc_local_traffic_drop_flows_for_lrouter(od, lsi->lflows); + &lsi->actions, + lsi->meter_groups, + NULL); + build_misc_local_traffic_drop_flows_for_lrouter(od, lsi->lflows, NULL); - build_lr_nat_defrag_and_lb_default_flows(od, lsi->lflows); - build_lrouter_lb_affinity_default_flows(od, lsi->lflows); + build_lr_nat_defrag_and_lb_default_flows(od, lsi->lflows, NULL); + build_lrouter_lb_affinity_default_flows(od, lsi->lflows, NULL); } /* Helper function to combine all lflow generation which is iterated by logical @@ -15511,22 +15850,26 @@ build_lswitch_and_lrouter_iterate_by_lrp(struct ovn_port *op, ovs_assert(op->nbrp); build_adm_ctrl_flows_for_lrouter_port(op, lsi->lflows, &lsi->match, - &lsi->actions); + &lsi->actions, op->lflow_ref); build_neigh_learning_flows_for_lrouter_port(op, lsi->lflows, &lsi->match, - &lsi->actions); - build_ip_routing_flows_for_lrp(op, lsi->lflows); + &lsi->actions, op->lflow_ref); + build_ip_routing_flows_for_lrp(op, lsi->lflows, op->lflow_ref); build_ND_RA_flows_for_lrouter_port(op, lsi->lflows, &lsi->match, - &lsi->actions, lsi->meter_groups); + &lsi->actions, lsi->meter_groups, + op->lflow_ref); build_arp_resolve_flows_for_lrp(op, lsi->lflows, - &lsi->match, &lsi->actions); + &lsi->match, &lsi->actions, op->lflow_ref); build_egress_delivery_flows_for_lrouter_port(op, lsi->lflows, &lsi->match, - &lsi->actions); - build_dhcpv6_reply_flows_for_lrouter_port(op, lsi->lflows, &lsi->match); + &lsi->actions, + op->lflow_ref); + build_dhcpv6_reply_flows_for_lrouter_port(op, lsi->lflows, &lsi->match, + op->lflow_ref); build_ipv6_input_flows_for_lrouter_port(op, lsi->lflows, &lsi->match, &lsi->actions, - lsi->meter_groups); + lsi->meter_groups, + op->lflow_ref); build_lrouter_ipv4_ip_input(op, lsi->lflows, &lsi->match, &lsi->actions, - lsi->meter_groups); + lsi->meter_groups, op->lflow_ref); } static void * @@ -15619,7 +15962,7 @@ build_lflows_thread(void *arg) if (stop_parallel_processing()) { return NULL; } - build_lswitch_arp_nd_service_monitor(lb_dps->lb, + build_lswitch_arp_nd_service_monitor(lb_dps, lsi->ls_ports, lsi->lflows, &lsi->match, @@ -15851,7 +16194,7 @@ build_lswitch_and_lrouter_flows( stopwatch_stop(LFLOWS_PORTS_STOPWATCH_NAME, time_msec()); stopwatch_start(LFLOWS_LBS_STOPWATCH_NAME, time_msec()); HMAP_FOR_EACH (lb_dps, hmap_node, lb_dps_map) { - build_lswitch_arp_nd_service_monitor(lb_dps->lb, lsi.ls_ports, + build_lswitch_arp_nd_service_monitor(lb_dps, lsi.ls_ports, lsi.lflows, &lsi.actions, &lsi.match); build_lrouter_defrag_flows_for_lb(lb_dps, lsi.lflows, @@ -16040,6 +16383,11 @@ lflow_reset_northd_refs(struct lflow_input *lflow_input) lflow_ref_clear(op->lflow_ref); lflow_ref_clear(op->stateful_lflow_ref); } + + HMAP_FOR_EACH (op, key_node, lflow_input->lr_ports) { + lflow_ref_clear(op->lflow_ref); + lflow_ref_clear(op->stateful_lflow_ref); + } } bool