On Mon, Jul 5, 2021, 12:09 PM Lorenzo Bianconi <[email protected]> wrote:
> Move stateful lb rules for logical switches in > build_lswitch_flows_for_lb routine in order to reduce cpu utilization > > Acked-by: Dumitru Ceara <[email protected]> > Signed-off-by: Lorenzo Bianconi <[email protected]> > Thanks Lorenzo for these improvement patches and Dumitru for the reviews. I applied all the patches of these series. Numan --- > northd/ovn-northd.c | 67 ++++++++++++++++++++------------------------- > 1 file changed, 29 insertions(+), 38 deletions(-) > > diff --git a/northd/ovn-northd.c b/northd/ovn-northd.c > index 22acbeed9..570c6a3ef 100644 > --- a/northd/ovn-northd.c > +++ b/northd/ovn-northd.c > @@ -6057,30 +6057,27 @@ build_qos(struct ovn_datapath *od, struct hmap > *lflows) { > } > > static void > -build_lb_rules(struct ovn_datapath *od, struct hmap *lflows, > - struct ovn_northd_lb *lb) > +build_lb_rules(struct hmap *lflows, struct ovn_northd_lb *lb, > + struct ds *match, struct ds *action) > { > - struct ds action = DS_EMPTY_INITIALIZER; > - struct ds match = DS_EMPTY_INITIALIZER; > - > for (size_t i = 0; i < lb->n_vips; i++) { > struct ovn_lb_vip *lb_vip = &lb->vips[i]; > struct ovn_northd_lb_vip *lb_vip_nb = &lb->vips_nb[i]; > const char *ip_match = NULL; > > - ds_clear(&action); > - ds_clear(&match); > + ds_clear(action); > + ds_clear(match); > > /* Store the original destination IP to be used when generating > * hairpin flows. > */ > if (IN6_IS_ADDR_V4MAPPED(&lb_vip->vip)) { > ip_match = "ip4"; > - ds_put_format(&action, REG_ORIG_DIP_IPV4 " = %s; ", > + ds_put_format(action, REG_ORIG_DIP_IPV4 " = %s; ", > lb_vip->vip_str); > } else { > ip_match = "ip6"; > - ds_put_format(&action, REG_ORIG_DIP_IPV6 " = %s; ", > + ds_put_format(action, REG_ORIG_DIP_IPV6 " = %s; ", > lb_vip->vip_str); > } > > @@ -6098,33 +6095,31 @@ build_lb_rules(struct ovn_datapath *od, struct > hmap *lflows, > /* Store the original destination port to be used when > generating > * hairpin flows. > */ > - ds_put_format(&action, REG_ORIG_TP_DPORT " = %"PRIu16"; ", > + ds_put_format(action, REG_ORIG_TP_DPORT " = %"PRIu16"; ", > lb_vip->vip_port); > } > > /* New connections in Ingress table. */ > - build_lb_vip_actions(lb_vip, lb_vip_nb, &action, > + build_lb_vip_actions(lb_vip, lb_vip_nb, action, > lb->selection_fields, true); > > - ds_put_format(&match, "ct.new && %s.dst == %s", ip_match, > + ds_put_format(match, "ct.new && %s.dst == %s", ip_match, > lb_vip->vip_str); > + int priority = 110; > if (lb_vip->vip_port) { > - ds_put_format(&match, " && %s.dst == %d", proto, > lb_vip->vip_port); > - ovn_lflow_add_with_hint(lflows, od, S_SWITCH_IN_STATEFUL, 120, > - ds_cstr(&match), ds_cstr(&action), > - &lb->nlb->header_); > - } else { > - ovn_lflow_add_with_hint(lflows, od, S_SWITCH_IN_STATEFUL, 110, > - ds_cstr(&match), ds_cstr(&action), > - &lb->nlb->header_); > + ds_put_format(match, " && %s.dst == %d", proto, > lb_vip->vip_port); > + priority = 120; > + } > + for (size_t j = 0; j < lb->n_nb_ls; j++) { > + ovn_lflow_add_with_hint(lflows, lb->nb_ls[j], > S_SWITCH_IN_STATEFUL, > + priority, ds_cstr(match), > + ds_cstr(action), &lb->nlb->header_); > } > } > - ds_destroy(&action); > - ds_destroy(&match); > } > > static void > -build_stateful(struct ovn_datapath *od, struct hmap *lflows, struct hmap > *lbs) > +build_stateful(struct ovn_datapath *od, struct hmap *lflows) > { > /* Ingress and Egress stateful Table (Priority 0): Packets are > * allowed by default. */ > @@ -6141,19 +6136,6 @@ build_stateful(struct ovn_datapath *od, struct hmap > *lflows, struct hmap *lbs) > ovn_lflow_add(lflows, od, S_SWITCH_OUT_STATEFUL, 100, > REGBIT_CONNTRACK_COMMIT" == 1", > "ct_commit { ct_label.blocked = 0; }; next;"); > - > - /* Load balancing rules for new connections get committed to conntrack > - * table. So even if REGBIT_CONNTRACK_COMMIT is set in a previous > table > - * a higher priority rule for load balancing below also commits the > - * connection, so it is okay if we do not hit the above match on > - * REGBIT_CONNTRACK_COMMIT. */ > - for (int i = 0; i < od->nbs->n_load_balancer; i++) { > - struct ovn_northd_lb *lb = > - ovn_northd_lb_find(lbs, > &od->nbs->load_balancer[i]->header_.uuid); > - > - ovs_assert(lb); > - build_lb_rules(od, lflows, lb); > - } > } > > static void > @@ -6921,7 +6903,7 @@ build_lswitch_lflows_pre_acl_and_acl(struct > ovn_datapath *od, > build_acl_hints(od, lflows); > build_acls(od, lflows, port_groups, meter_groups); > build_qos(od, lflows); > - build_stateful(od, lflows, lbs); > + build_stateful(od, lflows); > build_lb_hairpin(od, lflows); > } > } > @@ -8986,11 +8968,12 @@ build_lswitch_flows_for_lb(struct ovn_northd_lb > *lb, struct hmap *lflows, > for (size_t i = 0; i < lb->n_vips; i++) { > struct ovn_lb_vip *lb_vip = &lb->vips[i]; > > + /* pre-stateful lb */ > if (!build_empty_lb_event_flow(lb_vip, lb->nlb, meter_groups, > match, action)) { > continue; > } > - for (int j = 0; j < lb->n_nb_ls; j++) { > + for (size_t j = 0; j < lb->n_nb_ls; j++) { > ovn_lflow_add_with_hint(lflows, lb->nb_ls[j], > S_SWITCH_IN_PRE_LB, 130, > ds_cstr(match), > ds_cstr(action), &lb->nlb->header_); > @@ -9000,6 +8983,14 @@ build_lswitch_flows_for_lb(struct ovn_northd_lb > *lb, struct hmap *lflows, > * the packet through ct() action to de-fragment. In stateful > * table, we will eventually look at L4 information. */ > } > + > + /* stateful lb > + * Load balancing rules for new connections get committed to conntrack > + * table. So even if REGBIT_CONNTRACK_COMMIT is set in a previous > table > + * a higher priority rule for load balancing below also commits the > + * connection, so it is okay if we do not hit the above match on > + * REGBIT_CONNTRACK_COMMIT. */ > + build_lb_rules(lflows, lb, match, action); > } > > static void > -- > 2.31.1 > > _______________________________________________ > dev mailing list > [email protected] > https://mail.openvswitch.org/mailman/listinfo/ovs-dev > > _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
