Matching on ct.dnat creates openflows that often are not offloadable to hardware. ovn-northd uses ct.dnat only for load balancer hairpin traffic handling and it turns out we don't really need to match on ct.dnat.
Signed-off-by: Dumitru Ceara <[email protected]> --- northd/ovn-northd.8.xml | 32 +++++++++++++++-------------- northd/ovn-northd.c | 52 +++++++++++++++++++++++++++++------------------ 2 files changed, 49 insertions(+), 35 deletions(-) diff --git a/northd/ovn-northd.8.xml b/northd/ovn-northd.8.xml index deffe8c..a16937a 100644 --- a/northd/ovn-northd.8.xml +++ b/northd/ovn-northd.8.xml @@ -813,19 +813,12 @@ <li> If the logical switch has load balancer(s) configured, then a priority-100 flow is added with the match - <code>ip && ct.trk&& ct.dnat</code> to check if the + <code>ip && ct.trk</code> to check if the packet needs to be hairpinned (if after load balancing the destination - IP matches the source IP) or not by executing the action - <code>reg0[6] = chk_lb_hairpin();</code> and advances the packet to - the next table. - </li> - - <li> - If the logical switch has load balancer(s) configured, then a - priority-90 flow is added with the match <code>ip</code> to check if - the packet is a reply for a hairpinned connection or not by executing - the action <code>reg0[6] = chk_lb_hairpin_reply();</code> and advances - the packet to the next table. + IP matches the source IP) or not by executing the actions + <code>reg0[6] = chk_lb_hairpin();</code> and + <code>reg0[12] = chk_lb_hairpin_reply();</code> and advances the packet + to the next table. </li> <li> @@ -838,16 +831,25 @@ <li> If the logical switch has load balancer(s) configured, then a priority-100 flow is added with the match - <code>ip && (ct.new || ct.est) && ct.trk && - ct.dnat && reg0[6] == 1</code> which hairpins the traffic by + <code>ip && ct.new && ct.trk && + reg0[6] == 1</code> which hairpins the traffic by NATting source IP to the load balancer VIP by executing the action <code>ct_snat_to_vip</code> and advances the packet to the next table. </li> <li> If the logical switch has load balancer(s) configured, then a + priority-100 flow is added with the match + <code>ip && ct.est && ct.trk && + reg0[6] == 1</code> which hairpins the traffic by + NATting source IP to the load balancer VIP by executing the action + <code>ct_snat</code> and advances the packet to the next table. + </li> + + <li> + If the logical switch has load balancer(s) configured, then a priority-90 flow is added with the match - <code>ip && reg0[6] == 1</code> which matches on the replies + <code>ip && reg0[12] == 1</code> which matches on the replies of hairpinned traffic (i.e., destination IP is VIP, source IP is the backend IP and source L4 port is backend port for L4 load balancers) and executes <code>ct_snat</code> and advances the diff --git a/northd/ovn-northd.c b/northd/ovn-northd.c index 18e4cac..f66bdea 100644 --- a/northd/ovn-northd.c +++ b/northd/ovn-northd.c @@ -227,6 +227,7 @@ enum ovn_stage { #define REGBIT_ACL_HINT_DROP "reg0[9]" #define REGBIT_ACL_HINT_BLOCK "reg0[10]" #define REGBIT_LKUP_FDB "reg0[11]" +#define REGBIT_HAIRPIN_REPLY "reg0[12]" #define REG_ORIG_DIP_IPV4 "reg1" #define REG_ORIG_DIP_IPV6 "xxreg1" @@ -266,7 +267,8 @@ enum ovn_stage { * * Logical Switch pipeline: * +----+----------------------------------------------+---+------------------+ - * | R0 | REGBIT_{CONNTRACK/DHCP/DNS/HAIRPIN} | | | + * | R0 | REGBIT_{CONNTRACK/DHCP/DNS} | | | + * | | REGBIT_{HAIRPIN/HAIRPIN_REPLY} | X | | * | | REGBIT_ACL_HINT_{ALLOW_NEW/ALLOW/DROP/BLOCK} | X | | * +----+----------------------------------------------+ X | | * | R1 | ORIG_DIP_IPV4 (>= IN_STATEFUL) | R | | @@ -6036,39 +6038,49 @@ build_lb_hairpin(struct ovn_datapath *od, struct hmap *lflows) ovn_lflow_add(lflows, od, S_SWITCH_IN_HAIRPIN, 0, "1", "next;"); if (od->has_lb_vip) { - /* Check if the packet needs to be hairpinned. */ - ovn_lflow_add_with_hint(lflows, od, S_SWITCH_IN_PRE_HAIRPIN, 100, - "ip && ct.trk && ct.dnat", - REGBIT_HAIRPIN " = chk_lb_hairpin(); next;", + /* Check if the packet needs to be hairpinned. + * Set REGBIT_HAIRPIN in the original direction and + * REGBIT_HAIRPIN_REPLY in the reply direction. + */ + ovn_lflow_add_with_hint( + lflows, od, S_SWITCH_IN_PRE_HAIRPIN, 100, "ip && ct.trk", + REGBIT_HAIRPIN " = chk_lb_hairpin(); " + REGBIT_HAIRPIN_REPLY " = chk_lb_hairpin_reply(); " + "next;", + &od->nbs->header_); + + /* If packet needs to be hairpinned, snat the src ip with the VIP + * for new sessions. */ + ovn_lflow_add_with_hint(lflows, od, S_SWITCH_IN_NAT_HAIRPIN, 100, + "ip && ct.new && ct.trk" + " && "REGBIT_HAIRPIN " == 1", + "ct_snat_to_vip; next;", &od->nbs->header_); - /* Check if the packet is a reply of hairpinned traffic. */ - ovn_lflow_add_with_hint(lflows, od, S_SWITCH_IN_PRE_HAIRPIN, 90, "ip", - REGBIT_HAIRPIN " = chk_lb_hairpin_reply(); " - "next;", &od->nbs->header_); - - /* If packet needs to be hairpinned, snat the src ip with the VIP. */ + /* If packet needs to be hairpinned, for established sessions there + * should already be an SNAT conntrack entry. + */ ovn_lflow_add_with_hint(lflows, od, S_SWITCH_IN_NAT_HAIRPIN, 100, - "ip && (ct.new || ct.est) && ct.trk && ct.dnat" + "ip && ct.est && ct.trk" " && "REGBIT_HAIRPIN " == 1", - "ct_snat_to_vip; next;", + "ct_snat;", &od->nbs->header_); /* 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 " == 1", "ct_snat;", + "ip && "REGBIT_HAIRPIN_REPLY " == 1", + "ct_snat;", &od->nbs->header_); /* Ingress Hairpin table. * - Priority 1: Packets that were SNAT-ed for hairpinning should be * looped back (i.e., swap ETH addresses and send back on inport). */ - ovn_lflow_add(lflows, od, S_SWITCH_IN_HAIRPIN, 1, - REGBIT_HAIRPIN " == 1", - "eth.dst <-> eth.src;" - "outport = inport;" - "flags.loopback = 1;" - "output;"); + ovn_lflow_add( + lflows, od, S_SWITCH_IN_HAIRPIN, 1, + "("REGBIT_HAIRPIN " == 1 || " REGBIT_HAIRPIN_REPLY " == 1)", + "eth.dst <-> eth.src; outport = inport; flags.loopback = 1; " + "output;"); } } _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
