@@ -6,6 +6,7 @@ Post v21.09.0
if not desired.
- Added Load_Balancer_Group support, which simplifies large scale
configurations of load balancers.
+ - Added Logical_Router_Port "gateway_mtu_bypass" option.
OVN v21.09.0 - 01 Oct 2021
--------------------------
@@ -10201,8 +10201,47 @@ build_adm_ctrl_flows_for_lrouter(
}
}
-static int
-build_check_pkt_len_action_string(struct ovn_port *op, struct ds *actions);
+/* All 'gateway_mtu' and 'gateway_mtu_bypass' flows should be built with this
+ * function.
+ */
+static void OVS_PRINTF_FORMAT(9, 10)
+build_gateway_mtu_flow(struct hmap *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,
+ const char *extra_actions_fmt, ...)
+{
+ int gw_mtu = smap_get_int(&op->nbrp->options, "gateway_mtu", 0);
+
+ va_list extra_actions_args;
+ va_start(extra_actions_args, extra_actions_fmt);
+
+ ds_clear(actions);
+ if (gw_mtu > 0) {
+ ds_put_format(actions, REGBIT_PKT_LARGER" = check_pkt_larger(%d); ",
+ gw_mtu + VLAN_ETH_HEADER_LEN);
+ }
+
+ 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);
+
+ if (gw_mtu > 0) {
+ const char *gw_mtu_bypass = smap_get(&op->nbrp->options,
+ "gateway_mtu_bypass");
+ if (gw_mtu_bypass) {
+ ds_clear(actions);
+ ds_put_format_valist(actions, extra_actions_fmt,
+ extra_actions_args);
+ 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);
+ }
+ }
+ va_end(extra_actions_args);
+}
/* Logical router ingress Table 0: L2 Admission Control
* This table drops packets that the router shouldn’t see at all based
@@ -10230,17 +10269,12 @@ build_adm_ctrl_flows_for_lrouter_port(
* This will save us from having to match on inport further down in
* the pipeline.
*/
- ds_clear(actions);
-
- build_check_pkt_len_action_string(op, actions);
- ds_put_format(actions, REG_INPORT_ETH_ADDR " = %s; next;",
- op->lrp_networks.ea_s);
-
ds_clear(match);
ds_put_format(match, "eth.mcast && inport == %s", op->json_key);
- ovn_lflow_add_with_hint(lflows, op->od, S_ROUTER_IN_ADMISSION, 50,
- ds_cstr(match), ds_cstr(actions),
- &op->nbrp->header_);
+ build_gateway_mtu_flow(lflows, op, S_ROUTER_IN_ADMISSION, 50, 55,
+ match, actions, &op->nbrp->header_,
+ REG_INPORT_ETH_ADDR " = %s; next;",
+ op->lrp_networks.ea_s);
ds_clear(match);
ds_put_format(match, "eth.dst == %s && inport == %s",
@@ -10251,9 +10285,10 @@ build_adm_ctrl_flows_for_lrouter_port(
ds_put_format(match, " && is_chassis_resident(%s)",
op->cr_port->json_key);
}
- ovn_lflow_add_with_hint(lflows, op->od, S_ROUTER_IN_ADMISSION, 50,
- ds_cstr(match), ds_cstr(actions),
- &op->nbrp->header_);
+ build_gateway_mtu_flow(lflows, op, S_ROUTER_IN_ADMISSION, 50, 55,
+ match, actions, &op->nbrp->header_,
+ REG_INPORT_ETH_ADDR " = %s; next;",
+ op->lrp_networks.ea_s);
}
}
@@ -11248,40 +11283,21 @@ build_icmperr_pkt_big_flows(struct ovn_port *op, int mtu, struct hmap *lflows,
free(outport_match);
}
-static int
-build_check_pkt_len_action_string(struct ovn_port *op, struct ds *actions)
-{
- int gw_mtu = smap_get_int(&op->nbrp->options, "gateway_mtu", 0);
-
- if (gw_mtu > 0) {
- /* Add the flows only if gateway_mtu is configured. */
- ds_put_format(actions,
- REGBIT_PKT_LARGER" = check_pkt_larger(%d); ",
- gw_mtu + VLAN_ETH_HEADER_LEN);
- }
- return gw_mtu;
-}
-
static void
build_check_pkt_len_flows_for_lrp(struct ovn_port *op,
struct hmap *lflows, struct hmap *ports,
struct shash *meter_groups, struct ds *match,
struct ds *actions)
{
- ds_clear(actions);
- int gw_mtu = build_check_pkt_len_action_string(op, actions);
+ int gw_mtu = smap_get_int(&op->nbrp->options, "gateway_mtu", 0);
if (gw_mtu <= 0) {
return;
}
- ds_put_format(actions, "next;");
-
ds_clear(match);
ds_put_format(match, "outport == %s", op->json_key);
-
- ovn_lflow_add_with_hint(lflows, op->od, S_ROUTER_IN_CHK_PKT_LEN, 50,
- ds_cstr(match), ds_cstr(actions),
- &op->nbrp->header_);
+ build_gateway_mtu_flow(lflows, op, S_ROUTER_IN_CHK_PKT_LEN, 50, 55,
+ match, actions, &op->nbrp->header_, "next;");
/* ingress traffic */
build_icmperr_pkt_big_flows(op, gw_mtu, lflows, meter_groups,
@@ -12397,12 +12413,6 @@ build_lrouter_ingress_flow(struct hmap *lflows, struct ovn_datapath *od,
* This will save us from having to match on inport further
* down in the pipeline.
*/
- ds_clear(actions);
-
- build_check_pkt_len_action_string(od->l3dgw_ports[0], actions);
- ds_put_format(actions, REG_INPORT_ETH_ADDR " = %s; next;",
- od->l3dgw_ports[0]->lrp_networks.ea_s);
-
ds_clear(match);
ds_put_format(match,
"eth.dst == "ETH_ADDR_FMT" && inport == %s"
@@ -12410,9 +12420,11 @@ build_lrouter_ingress_flow(struct hmap *lflows, struct ovn_datapath *od,
ETH_ADDR_ARGS(mac),
od->l3dgw_ports[0]->json_key,
nat->logical_port);
- ovn_lflow_add_with_hint(lflows, od, S_ROUTER_IN_ADMISSION, 50,
- ds_cstr(match), ds_cstr(actions),
- &nat->header_);
+ build_gateway_mtu_flow(lflows, od->l3dgw_ports[0],
+ S_ROUTER_IN_ADMISSION, 50, 55,
+ match, actions, &nat->header_,
+ REG_INPORT_ETH_ADDR " = %s; next;",
+ od->l3dgw_ports[0]->lrp_networks.ea_s);
}
}
@@ -1984,6 +1984,13 @@ output;
<code>check_pkt_larger</code> in order to mark the packet
setting <code>REGBIT_PKT_LARGER</code> if the size is greater
than the MTU.
+
+ If the port is also configured with
+ <code>options:gateway_mtu_bypass</code> then another flow is
+ added, with priority-55, to bypass the <code>check_pkt_larger</code>
+ flow. This is useful for traffic that normally doesn't need to be
+ fragmented and for which check_pkt_larger, which might not be
+ offloadable, is not really needed. One such example is TCP traffic.
</p>
</li>
@@ -3784,6 +3791,13 @@ REGBIT_PKT_LARGER = check_pkt_larger(<var>L</var>); next;
<ref table="Logical_Router_Port" db="OVN_Northbound"/> row.
</p>
+ <p>
+ If the port is also configured with
+ <code>options:gateway_mtu_bypass</code> then another flow is
+ added, with priority-55, to bypass the <code>check_pkt_larger</code>
+ flow.
+ </p>
+
<p>
This table adds one priority-0 fallback flow that matches all packets
and advances to the next table.
@@ -2799,6 +2799,29 @@
prefix according to RFC3663
</p>
</column>
+
+ <column name="options" key="gateway_mtu">
+ <p>
+ Configures the maximum length, excluding VLAN and Ethernet header,
+ of packets that should be allowed to be received by distributed
+ and gateway router logical ports. For all packets with a larger
+ size than this value ICMP (v4/v6) error packets are generated
+ indicating that fragmentation should be performed before sending
+ packets to the OVN port.
+ </p>
+ </column>
+
+ <column name="options" key="gateway_mtu_bypass">
+ <p>
+ When configured, represents a match expression, in the same
+ expression language used for the <ref column="match"
+ table="Logical_Flow" db="OVN_Southbound"/> column in the OVN
+ Southbound database's <ref table="Logical_Flow" db="OVN_Southbound"/>
+ table. Packets matching this expression will bypass the length
+ check configured through the
+ <ref column="options" key="gateway_mtu"/> option.
+ </p>
+ </column>
</group>
<group title="Attachment">
@@ -5075,7 +5075,7 @@ check ovn-nbctl lsp-set-options public-lr0 router-port=lr0-public
check ovn-nbctl --wait=sb lrp-set-gateway-chassis lr0-public ch1
ovn-sbctl dump-flows lr0 > lr0flows
-AT_CAPTURE_FILE([sw0flows])
+AT_CAPTURE_FILE([lr0flows])
AT_CHECK([grep -e "chk_pkt_len" -e "lr_in_larger_pkts" lr0flows | sed 's/table=../table=??/' | sort], [0], [dnl
table=??(lr_in_chk_pkt_len ), priority=0 , match=(1), action=(next;)
@@ -5085,7 +5085,7 @@ AT_CHECK([grep -e "chk_pkt_len" -e "lr_in_larger_pkts" lr0flows | sed 's/table=.
check ovn-nbctl --wait=sb set logical_router_port lr0-public options:gateway_mtu=1500
ovn-sbctl dump-flows lr0 > lr0flows
-AT_CAPTURE_FILE([sw0flows])
+AT_CAPTURE_FILE([lr0flows])
AT_CHECK([grep -e "chk_pkt_len" -e "lr_in_larger_pkts" lr0flows | sed 's/table=../table=??/' | sort], [0], [dnl
table=??(lr_in_chk_pkt_len ), priority=0 , match=(1), action=(next;)
@@ -5116,7 +5116,7 @@ check ovn-nbctl --wait=sb clear logical_router_port lr0-public gateway_chassis
check ovn-nbctl --wait=sb set logical_router lr0 options:chassis=ch1
ovn-sbctl dump-flows lr0 > lr0flows
-AT_CAPTURE_FILE([sw0flows])
+AT_CAPTURE_FILE([lr0flows])
AT_CHECK([grep -e "chk_pkt_len" -e "lr_in_larger_pkts" lr0flows | sed 's/table=../table=??/' | sort], [0], [dnl
table=??(lr_in_chk_pkt_len ), priority=0 , match=(1), action=(next;)
@@ -5140,16 +5140,41 @@ AT_CHECK([grep -E "lr_in_ip_input.*icmp6_error" lr0flows | sort], [0], [dnl
table=3 (lr_in_ip_input ), priority=150 , match=(inport == "lr0-public" && ip6 && reg9[[1]] && reg9[[0]] == 0), action=(icmp6_error {reg9[[0]] = 1; reg9[[1]] = 0; eth.dst = 00:00:20:20:12:13; ip6.dst = ip6.src; ip6.src = fe80::200:20ff:fe20:1213; ip.ttl = 255; icmp6.type = 2; /* Packet Too Big. */ icmp6.code = 0; icmp6.frag_mtu = 1500; next(pipeline=ingress, table=0); };)
])
+# Set gateway_mtu_bypass to avoid check_pkt_larger() for tcp on lr0-public.
+check ovn-nbctl --wait=sb set logical_router_port lr0-public options:gateway_mtu_bypass=tcp
+
+ovn-sbctl dump-flows lr0 > lr0flows
+AT_CAPTURE_FILE([lr0flows])
+
+AT_CHECK([grep -e "chk_pkt_len" -e "lr_in_larger_pkts" lr0flows | sed 's/table=../table=??/' | sort], [0], [dnl
+ table=??(lr_in_chk_pkt_len ), priority=0 , match=(1), action=(next;)
+ table=??(lr_in_chk_pkt_len ), priority=50 , match=(outport == "lr0-public"), action=(reg9[[1]] = check_pkt_larger(1518); next;)
+ table=??(lr_in_chk_pkt_len ), priority=55 , match=(outport == "lr0-public" && (tcp)), action=(next;)
+ table=??(lr_in_larger_pkts ), priority=0 , match=(1), action=(next;)
+ table=??(lr_in_larger_pkts ), priority=150 , match=(inport == "lr0-sw0" && outport == "lr0-public" && ip4 && reg9[[1]] && reg9[[0]] == 0), action=(icmp4_error {reg9[[0]] = 1; reg9[[1]] = 0; eth.dst = 00:00:00:00:ff:01; ip4.dst = ip4.src; ip4.src = 10.0.0.1; ip.ttl = 255; icmp4.type = 3; /* Destination Unreachable. */ icmp4.code = 4; /* Frag Needed and DF was Set. */ icmp4.frag_mtu = 1500; next(pipeline=ingress, table=0); };)
+ table=??(lr_in_larger_pkts ), priority=150 , match=(inport == "lr0-sw0" && outport == "lr0-public" && ip6 && reg9[[1]] && reg9[[0]] == 0), action=(icmp6_error {reg9[[0]] = 1; reg9[[1]] = 0; eth.dst = 00:00:00:00:ff:01; ip6.dst = ip6.src; ip6.src = fe80::200:ff:fe00:ff01; ip.ttl = 255; icmp6.type = 2; /* Packet Too Big. */ icmp6.code = 0; icmp6.frag_mtu = 1500; next(pipeline=ingress, table=0); };)
+ table=??(lr_in_larger_pkts ), priority=150 , match=(inport == "lr0-sw1" && outport == "lr0-public" && ip4 && reg9[[1]] && reg9[[0]] == 0), action=(icmp4_error {reg9[[0]] = 1; reg9[[1]] = 0; eth.dst = 00:00:00:00:ff:02; ip4.dst = ip4.src; ip4.src = 20.0.0.1; ip.ttl = 255; icmp4.type = 3; /* Destination Unreachable. */ icmp4.code = 4; /* Frag Needed and DF was Set. */ icmp4.frag_mtu = 1500; next(pipeline=ingress, table=0); };)
+ table=??(lr_in_larger_pkts ), priority=150 , match=(inport == "lr0-sw1" && outport == "lr0-public" && ip6 && reg9[[1]] && reg9[[0]] == 0), action=(icmp6_error {reg9[[0]] = 1; reg9[[1]] = 0; eth.dst = 00:00:00:00:ff:02; ip6.dst = ip6.src; ip6.src = fe80::200:ff:fe00:ff02; ip.ttl = 255; icmp6.type = 2; /* Packet Too Big. */ icmp6.code = 0; icmp6.frag_mtu = 1500; next(pipeline=ingress, table=0); };)
+])
+
+AT_CHECK([grep "lr_in_admission" lr0flows | grep -e "check_pkt_larger" -e "tcp" | sort], [0], [dnl
+ table=0 (lr_in_admission ), priority=50 , match=(eth.dst == 00:00:20:20:12:13 && inport == "lr0-public"), action=(reg9[[1]] = check_pkt_larger(1518); xreg0[[0..47]] = 00:00:20:20:12:13; next;)
+ table=0 (lr_in_admission ), priority=50 , match=(eth.mcast && inport == "lr0-public"), action=(reg9[[1]] = check_pkt_larger(1518); xreg0[[0..47]] = 00:00:20:20:12:13; next;)
+ table=0 (lr_in_admission ), priority=55 , match=(eth.dst == 00:00:20:20:12:13 && inport == "lr0-public" && (tcp)), action=(xreg0[[0..47]] = 00:00:20:20:12:13; next;)
+ table=0 (lr_in_admission ), priority=55 , match=(eth.mcast && inport == "lr0-public" && (tcp)), action=(xreg0[[0..47]] = 00:00:20:20:12:13; next;)
+])
+
# Set gateway_mtu option on lr0-sw0
check ovn-nbctl --wait=sb set logical_router_port lr0-sw0 options:gateway_mtu=1400
ovn-sbctl dump-flows lr0 > lr0flows
-AT_CAPTURE_FILE([sw0flows])
+AT_CAPTURE_FILE([lr0flows])
AT_CHECK([grep -e "chk_pkt_len" -e "lr_in_larger_pkts" lr0flows | sed 's/table=../table=??/' | sort], [0], [dnl
table=??(lr_in_chk_pkt_len ), priority=0 , match=(1), action=(next;)
table=??(lr_in_chk_pkt_len ), priority=50 , match=(outport == "lr0-public"), action=(reg9[[1]] = check_pkt_larger(1518); next;)
table=??(lr_in_chk_pkt_len ), priority=50 , match=(outport == "lr0-sw0"), action=(reg9[[1]] = check_pkt_larger(1418); next;)
+ table=??(lr_in_chk_pkt_len ), priority=55 , match=(outport == "lr0-public" && (tcp)), action=(next;)
table=??(lr_in_larger_pkts ), priority=0 , match=(1), action=(next;)
table=??(lr_in_larger_pkts ), priority=150 , match=(inport == "lr0-public" && outport == "lr0-sw0" && ip4 && reg9[[1]] && reg9[[0]] == 0), action=(icmp4_error {reg9[[0]] = 1; reg9[[1]] = 0; eth.dst = 00:00:20:20:12:13; ip4.dst = ip4.src; ip4.src = 172.168.0.100; ip.ttl = 255; icmp4.type = 3; /* Destination Unreachable. */ icmp4.code = 4; /* Frag Needed and DF was Set. */ icmp4.frag_mtu = 1400; next(pipeline=ingress, table=0); };)
table=??(lr_in_larger_pkts ), priority=150 , match=(inport == "lr0-public" && outport == "lr0-sw0" && ip6 && reg9[[1]] && reg9[[0]] == 0), action=(icmp6_error {reg9[[0]] = 1; reg9[[1]] = 0; eth.dst = 00:00:20:20:12:13; ip6.dst = ip6.src; ip6.src = fe80::200:20ff:fe20:1213; ip.ttl = 255; icmp6.type = 2; /* Packet Too Big. */ icmp6.code = 0; icmp6.frag_mtu = 1400; next(pipeline=ingress, table=0); };)
@@ -5161,7 +5186,7 @@ AT_CHECK([grep -e "chk_pkt_len" -e "lr_in_larger_pkts" lr0flows | sed 's/table=.
table=??(lr_in_larger_pkts ), priority=150 , match=(inport == "lr0-sw1" && outport == "lr0-sw0" && ip6 && reg9[[1]] && reg9[[0]] == 0), action=(icmp6_error {reg9[[0]] = 1; reg9[[1]] = 0; eth.dst = 00:00:00:00:ff:02; ip6.dst = ip6.src; ip6.src = fe80::200:ff:fe00:ff02; ip.ttl = 255; icmp6.type = 2; /* Packet Too Big. */ icmp6.code = 0; icmp6.frag_mtu = 1400; next(pipeline=ingress, table=0); };)
])
-AT_CHECK([grep -E "lr_in_admission.*check_pkt_larger" lr0flows | sort], [0], [dnl
+AT_CHECK([grep "lr_in_admission.*check_pkt_larger" lr0flows | sort], [0], [dnl
table=0 (lr_in_admission ), priority=50 , match=(eth.dst == 00:00:00:00:ff:01 && inport == "lr0-sw0"), action=(reg9[[1]] = check_pkt_larger(1418); xreg0[[0..47]] = 00:00:00:00:ff:01; next;)
table=0 (lr_in_admission ), priority=50 , match=(eth.dst == 00:00:20:20:12:13 && inport == "lr0-public"), action=(reg9[[1]] = check_pkt_larger(1518); xreg0[[0..47]] = 00:00:20:20:12:13; next;)
table=0 (lr_in_admission ), priority=50 , match=(eth.mcast && inport == "lr0-public"), action=(reg9[[1]] = check_pkt_larger(1518); xreg0[[0..47]] = 00:00:20:20:12:13; next;)
@@ -5177,14 +5202,49 @@ AT_CHECK([grep -E "lr_in_ip_input.*icmp6_error" lr0flows | sort], [0], [dnl
table=3 (lr_in_ip_input ), priority=150 , match=(inport == "lr0-sw0" && ip6 && reg9[[1]] && reg9[[0]] == 0), action=(icmp6_error {reg9[[0]] = 1; reg9[[1]] = 0; eth.dst = 00:00:00:00:ff:01; ip6.dst = ip6.src; ip6.src = fe80::200:ff:fe00:ff01; ip.ttl = 255; icmp6.type = 2; /* Packet Too Big. */ icmp6.code = 0; icmp6.frag_mtu = 1400; next(pipeline=ingress, table=0); };)
])
+# Set gateway_mtu_bypass to avoid check_pkt_larger() for tcp on lr0-sw0.
+check ovn-nbctl --wait=sb set logical_router_port lr0-sw0 options:gateway_mtu_bypass=tcp
+
+ovn-sbctl dump-flows lr0 > lr0flows
+AT_CAPTURE_FILE([lr0flows])
+
+AT_CHECK([grep -e "chk_pkt_len" -e "lr_in_larger_pkts" lr0flows | sed 's/table=../table=??/' | sort], [0], [dnl
+ table=??(lr_in_chk_pkt_len ), priority=0 , match=(1), action=(next;)
+ table=??(lr_in_chk_pkt_len ), priority=50 , match=(outport == "lr0-public"), action=(reg9[[1]] = check_pkt_larger(1518); next;)
+ table=??(lr_in_chk_pkt_len ), priority=50 , match=(outport == "lr0-sw0"), action=(reg9[[1]] = check_pkt_larger(1418); next;)
+ table=??(lr_in_chk_pkt_len ), priority=55 , match=(outport == "lr0-public" && (tcp)), action=(next;)
+ table=??(lr_in_chk_pkt_len ), priority=55 , match=(outport == "lr0-sw0" && (tcp)), action=(next;)
+ table=??(lr_in_larger_pkts ), priority=0 , match=(1), action=(next;)
+ table=??(lr_in_larger_pkts ), priority=150 , match=(inport == "lr0-public" && outport == "lr0-sw0" && ip4 && reg9[[1]] && reg9[[0]] == 0), action=(icmp4_error {reg9[[0]] = 1; reg9[[1]] = 0; eth.dst = 00:00:20:20:12:13; ip4.dst = ip4.src; ip4.src = 172.168.0.100; ip.ttl = 255; icmp4.type = 3; /* Destination Unreachable. */ icmp4.code = 4; /* Frag Needed and DF was Set. */ icmp4.frag_mtu = 1400; next(pipeline=ingress, table=0); };)
+ table=??(lr_in_larger_pkts ), priority=150 , match=(inport == "lr0-public" && outport == "lr0-sw0" && ip6 && reg9[[1]] && reg9[[0]] == 0), action=(icmp6_error {reg9[[0]] = 1; reg9[[1]] = 0; eth.dst = 00:00:20:20:12:13; ip6.dst = ip6.src; ip6.src = fe80::200:20ff:fe20:1213; ip.ttl = 255; icmp6.type = 2; /* Packet Too Big. */ icmp6.code = 0; icmp6.frag_mtu = 1400; next(pipeline=ingress, table=0); };)
+ table=??(lr_in_larger_pkts ), priority=150 , match=(inport == "lr0-sw0" && outport == "lr0-public" && ip4 && reg9[[1]] && reg9[[0]] == 0), action=(icmp4_error {reg9[[0]] = 1; reg9[[1]] = 0; eth.dst = 00:00:00:00:ff:01; ip4.dst = ip4.src; ip4.src = 10.0.0.1; ip.ttl = 255; icmp4.type = 3; /* Destination Unreachable. */ icmp4.code = 4; /* Frag Needed and DF was Set. */ icmp4.frag_mtu = 1500; next(pipeline=ingress, table=0); };)
+ table=??(lr_in_larger_pkts ), priority=150 , match=(inport == "lr0-sw0" && outport == "lr0-public" && ip6 && reg9[[1]] && reg9[[0]] == 0), action=(icmp6_error {reg9[[0]] = 1; reg9[[1]] = 0; eth.dst = 00:00:00:00:ff:01; ip6.dst = ip6.src; ip6.src = fe80::200:ff:fe00:ff01; ip.ttl = 255; icmp6.type = 2; /* Packet Too Big. */ icmp6.code = 0; icmp6.frag_mtu = 1500; next(pipeline=ingress, table=0); };)
+ table=??(lr_in_larger_pkts ), priority=150 , match=(inport == "lr0-sw1" && outport == "lr0-public" && ip4 && reg9[[1]] && reg9[[0]] == 0), action=(icmp4_error {reg9[[0]] = 1; reg9[[1]] = 0; eth.dst = 00:00:00:00:ff:02; ip4.dst = ip4.src; ip4.src = 20.0.0.1; ip.ttl = 255; icmp4.type = 3; /* Destination Unreachable. */ icmp4.code = 4; /* Frag Needed and DF was Set. */ icmp4.frag_mtu = 1500; next(pipeline=ingress, table=0); };)
+ table=??(lr_in_larger_pkts ), priority=150 , match=(inport == "lr0-sw1" && outport == "lr0-public" && ip6 && reg9[[1]] && reg9[[0]] == 0), action=(icmp6_error {reg9[[0]] = 1; reg9[[1]] = 0; eth.dst = 00:00:00:00:ff:02; ip6.dst = ip6.src; ip6.src = fe80::200:ff:fe00:ff02; ip.ttl = 255; icmp6.type = 2; /* Packet Too Big. */ icmp6.code = 0; icmp6.frag_mtu = 1500; next(pipeline=ingress, table=0); };)
+ table=??(lr_in_larger_pkts ), priority=150 , match=(inport == "lr0-sw1" && outport == "lr0-sw0" && ip4 && reg9[[1]] && reg9[[0]] == 0), action=(icmp4_error {reg9[[0]] = 1; reg9[[1]] = 0; eth.dst = 00:00:00:00:ff:02; ip4.dst = ip4.src; ip4.src = 20.0.0.1; ip.ttl = 255; icmp4.type = 3; /* Destination Unreachable. */ icmp4.code = 4; /* Frag Needed and DF was Set. */ icmp4.frag_mtu = 1400; next(pipeline=ingress, table=0); };)
+ table=??(lr_in_larger_pkts ), priority=150 , match=(inport == "lr0-sw1" && outport == "lr0-sw0" && ip6 && reg9[[1]] && reg9[[0]] == 0), action=(icmp6_error {reg9[[0]] = 1; reg9[[1]] = 0; eth.dst = 00:00:00:00:ff:02; ip6.dst = ip6.src; ip6.src = fe80::200:ff:fe00:ff02; ip.ttl = 255; icmp6.type = 2; /* Packet Too Big. */ icmp6.code = 0; icmp6.frag_mtu = 1400; next(pipeline=ingress, table=0); };)
+])
+
+AT_CHECK([grep "lr_in_admission" lr0flows | grep -e "check_pkt_larger" -e "tcp" | sort], [0], [dnl
+ table=0 (lr_in_admission ), priority=50 , match=(eth.dst == 00:00:00:00:ff:01 && inport == "lr0-sw0"), action=(reg9[[1]] = check_pkt_larger(1418); xreg0[[0..47]] = 00:00:00:00:ff:01; next;)
+ table=0 (lr_in_admission ), priority=50 , match=(eth.dst == 00:00:20:20:12:13 && inport == "lr0-public"), action=(reg9[[1]] = check_pkt_larger(1518); xreg0[[0..47]] = 00:00:20:20:12:13; next;)
+ table=0 (lr_in_admission ), priority=50 , match=(eth.mcast && inport == "lr0-public"), action=(reg9[[1]] = check_pkt_larger(1518); xreg0[[0..47]] = 00:00:20:20:12:13; next;)
+ table=0 (lr_in_admission ), priority=50 , match=(eth.mcast && inport == "lr0-sw0"), action=(reg9[[1]] = check_pkt_larger(1418); xreg0[[0..47]] = 00:00:00:00:ff:01; next;)
+ table=0 (lr_in_admission ), priority=55 , match=(eth.dst == 00:00:00:00:ff:01 && inport == "lr0-sw0" && (tcp)), action=(xreg0[[0..47]] = 00:00:00:00:ff:01; next;)
+ table=0 (lr_in_admission ), priority=55 , match=(eth.dst == 00:00:20:20:12:13 && inport == "lr0-public" && (tcp)), action=(xreg0[[0..47]] = 00:00:20:20:12:13; next;)
+ table=0 (lr_in_admission ), priority=55 , match=(eth.mcast && inport == "lr0-public" && (tcp)), action=(xreg0[[0..47]] = 00:00:20:20:12:13; next;)
+ table=0 (lr_in_admission ), priority=55 , match=(eth.mcast && inport == "lr0-sw0" && (tcp)), action=(xreg0[[0..47]] = 00:00:00:00:ff:01; next;)
+])
+
# Clear gateway_mtu option on lr0-public
check ovn-nbctl --wait=sb clear logical_router_port lr0-public options
ovn-sbctl dump-flows lr0 > lr0flows
-AT_CAPTURE_FILE([sw0flows])
+AT_CAPTURE_FILE([lr0flows])
AT_CHECK([grep -e "chk_pkt_len" -e "lr_in_larger_pkts" lr0flows | sed 's/table=../table=??/' | sort], [0], [dnl
table=??(lr_in_chk_pkt_len ), priority=0 , match=(1), action=(next;)
table=??(lr_in_chk_pkt_len ), priority=50 , match=(outport == "lr0-sw0"), action=(reg9[[1]] = check_pkt_larger(1418); next;)
+ table=??(lr_in_chk_pkt_len ), priority=55 , match=(outport == "lr0-sw0" && (tcp)), action=(next;)
table=??(lr_in_larger_pkts ), priority=0 , match=(1), action=(next;)
table=??(lr_in_larger_pkts ), priority=150 , match=(inport == "lr0-public" && outport == "lr0-sw0" && ip4 && reg9[[1]] && reg9[[0]] == 0), action=(icmp4_error {reg9[[0]] = 1; reg9[[1]] = 0; eth.dst = 00:00:20:20:12:13; ip4.dst = ip4.src; ip4.src = 172.168.0.100; ip.ttl = 255; icmp4.type = 3; /* Destination Unreachable. */ icmp4.code = 4; /* Frag Needed and DF was Set. */ icmp4.frag_mtu = 1400; next(pipeline=ingress, table=0); };)
table=??(lr_in_larger_pkts ), priority=150 , match=(inport == "lr0-public" && outport == "lr0-sw0" && ip6 && reg9[[1]] && reg9[[0]] == 0), action=(icmp6_error {reg9[[0]] = 1; reg9[[1]] = 0; eth.dst = 00:00:20:20:12:13; ip6.dst = ip6.src; ip6.src = fe80::200:20ff:fe20:1213; ip.ttl = 255; icmp6.type = 2; /* Packet Too Big. */ icmp6.code = 0; icmp6.frag_mtu = 1400; next(pipeline=ingress, table=0); };)
There are various costs (e.g., not being able to perform hardware offload in some cases) when using check_pkt_larger() so the CMS can now limit the impact by bypassing the packet length checks for specific types of traffic (e.g., TCP). Reported-at: https://bugzilla.redhat.com/show_bug.cgi?id=2011779 Signed-off-by: Dumitru Ceara <dceara@redhat.com> --- NEWS | 1 + northd/northd.c | 102 ++++++++++++++++++++++------------------ northd/ovn-northd.8.xml | 14 ++++++ ovn-nb.xml | 23 +++++++++ tests/ovn-northd.at | 72 +++++++++++++++++++++++++--- 5 files changed, 161 insertions(+), 51 deletions(-)