xref: /linux/tools/testing/selftests/net/pmtu.sh (revision 1b98f357dadd6ea613a435fbaef1a5dd7b35fd21)
1#!/bin/bash
2# SPDX-License-Identifier: GPL-2.0
3#
4# Check that route PMTU values match expectations, and that initial device MTU
5# values are assigned correctly
6#
7# Tests currently implemented:
8#
9# - pmtu_ipv4
10#	Set up two namespaces, A and B, with two paths between them over routers
11#	R1 and R2 (also implemented with namespaces), with different MTUs:
12#
13#	  segment a_r1    segment b_r1		a_r1: 2000
14#	.--------------R1--------------.	b_r1: 1400
15#	A                               B	a_r2: 2000
16#	'--------------R2--------------'	b_r2: 1500
17#	  segment a_r2    segment b_r2
18#
19#	Check that PMTU exceptions with the correct PMTU are created. Then
20#	decrease and increase the MTU of the local link for one of the paths,
21#	A to R1, checking that route exception PMTU changes accordingly over
22#	this path. Also check that locked exceptions are created when an ICMP
23#	message advertising a PMTU smaller than net.ipv4.route.min_pmtu is
24#	received
25#
26# - pmtu_ipv6
27#	Same as pmtu_ipv4, except for locked PMTU tests, using IPv6
28#
29# - pmtu_ipv4_dscp_icmp_exception
30#	Set up the same network topology as pmtu_ipv4, but use non-default
31#	routing table in A. A fib-rule is used to jump to this routing table
32#	based on DSCP. Send ICMPv4 packets with the expected DSCP value and
33#	verify that ECN doesn't interfere with the creation of PMTU exceptions.
34#
35# - pmtu_ipv4_dscp_udp_exception
36#	Same as pmtu_ipv4_dscp_icmp_exception, but use UDP instead of ICMP.
37#
38# - pmtu_ipv4_vxlan4_exception
39#	Set up the same network topology as pmtu_ipv4, create a VXLAN tunnel
40#	over IPv4 between A and B, routed via R1. On the link between R1 and B,
41#	set a MTU lower than the VXLAN MTU and the MTU on the link between A and
42#	R1. Send IPv4 packets, exceeding the MTU between R1 and B, over VXLAN
43#	from A to B and check that the PMTU exception is created with the right
44#	value on A
45#
46# - pmtu_ipv6_vxlan4_exception
47#	Same as pmtu_ipv4_vxlan4_exception, but send IPv6 packets from A to B
48#
49# - pmtu_ipv4_vxlan6_exception
50#	Same as pmtu_ipv4_vxlan4_exception, but use IPv6 transport from A to B
51#
52# - pmtu_ipv6_vxlan6_exception
53#	Same as pmtu_ipv4_vxlan6_exception, but send IPv6 packets from A to B
54#
55# - pmtu_ipv4_geneve4_exception
56#	Same as pmtu_ipv4_vxlan4_exception, but using a GENEVE tunnel instead of
57#	VXLAN
58#
59# - pmtu_ipv6_geneve4_exception
60#	Same as pmtu_ipv6_vxlan4_exception, but using a GENEVE tunnel instead of
61#	VXLAN
62#
63# - pmtu_ipv4_geneve6_exception
64#	Same as pmtu_ipv4_vxlan6_exception, but using a GENEVE tunnel instead of
65#	VXLAN
66#
67# - pmtu_ipv6_geneve6_exception
68#	Same as pmtu_ipv6_vxlan6_exception, but using a GENEVE tunnel instead of
69#	VXLAN
70#
71# - pmtu_ipv{4,6}_br_vxlan{4,6}_exception
72#	Set up three namespaces, A, B, and C, with routing between A and B over
73#	R1. R2 is unused in these tests. A has a veth connection to C, and is
74#	connected to B via a VXLAN endpoint, which is directly bridged to C.
75#	MTU on the B-R1 link is lower than other MTUs.
76#
77#	Check that both C and A are able to communicate with B over the VXLAN
78#	tunnel, and that PMTU exceptions with the correct values are created.
79#
80#	                  segment a_r1    segment b_r1            b_r1: 4000
81#	                .--------------R1--------------.    everything
82#	   C---veth     A                               B         else: 5000
83#	        ' bridge                                |
84#	            '---- - - - - - VXLAN - - - - - - - '
85#
86# - pmtu_ipv{4,6}_br_geneve{4,6}_exception
87#	Same as pmtu_ipv{4,6}_br_vxlan{4,6}_exception, with a GENEVE tunnel
88#	instead.
89#
90# - pmtu_ipv{4,6}_ovs_vxlan{4,6}_exception
91#	Set up two namespaces, B, and C, with routing between the init namespace
92#	and B over R1. A and R2 are unused in these tests. The init namespace
93#	has a veth connection to C, and is connected to B via a VXLAN endpoint,
94#	which is handled by Open vSwitch and bridged to C. MTU on the B-R1 link
95#	is lower than other MTUs.
96#
97#	Check that C is able to communicate with B over the VXLAN tunnel, and
98#	that PMTU exceptions with the correct values are created.
99#
100#	                  segment a_r1    segment b_r1            b_r1: 4000
101#	                .--------------R1--------------.    everything
102#	   C---veth    init                             B         else: 5000
103#	        '- ovs                                  |
104#	            '---- - - - - - VXLAN - - - - - - - '
105#
106# - pmtu_ipv{4,6}_ovs_geneve{4,6}_exception
107#	Same as pmtu_ipv{4,6}_ovs_vxlan{4,6}_exception, with a GENEVE tunnel
108#	instead.
109#
110# - pmtu_ipv{4,6}_fou{4,6}_exception
111#	Same as pmtu_ipv4_vxlan4, but using a direct IPv4/IPv6 encapsulation
112#	(FoU) over IPv4/IPv6, instead of VXLAN
113#
114# - pmtu_ipv{4,6}_fou{4,6}_exception
115#	Same as pmtu_ipv4_vxlan4, but using a generic UDP IPv4/IPv6
116#	encapsulation (GUE) over IPv4/IPv6, instead of VXLAN
117#
118# - pmtu_ipv{4,6}_ipv{4,6}_exception
119#	Same as pmtu_ipv4_vxlan4, but using a IPv4/IPv6 tunnel over IPv4/IPv6,
120#	instead of VXLAN
121#
122# - pmtu_vti4_exception
123#	Set up vti tunnel on top of veth, with xfrm states and policies, in two
124#	namespaces with matching endpoints. Check that route exception is not
125#	created if link layer MTU is not exceeded, then exceed it and check that
126#	exception is created with the expected PMTU. The approach described
127#	below for IPv6 doesn't apply here, because, on IPv4, administrative MTU
128#	changes alone won't affect PMTU
129#
130# - pmtu_vti4_udp_exception
131#       Same as pmtu_vti4_exception, but using ESP-in-UDP
132#
133# - pmtu_vti4_udp_routed_exception
134#       Set up vti tunnel on top of veth connected through routing namespace and
135#	add xfrm states and policies with ESP-in-UDP encapsulation. Check that
136#	route exception is not created if link layer MTU is not exceeded, then
137#	lower MTU on second part of routed environment and check that exception
138#	is created with the expected PMTU.
139#
140# - pmtu_vti6_exception
141#	Set up vti6 tunnel on top of veth, with xfrm states and policies, in two
142#	namespaces with matching endpoints. Check that route exception is
143#	created by exceeding link layer MTU with ping to other endpoint. Then
144#	decrease and increase MTU of tunnel, checking that route exception PMTU
145#	changes accordingly
146#
147# - pmtu_vti6_udp_exception
148#       Same as pmtu_vti6_exception, but using ESP-in-UDP
149#
150# - pmtu_vti6_udp_routed_exception
151#	Same as pmtu_vti6_udp_routed_exception but with routing between vti
152#	endpoints
153#
154# - pmtu_vti4_default_mtu
155#	Set up vti4 tunnel on top of veth, in two namespaces with matching
156#	endpoints. Check that MTU assigned to vti interface is the MTU of the
157#	lower layer (veth) minus additional lower layer headers (zero, for veth)
158#	minus IPv4 header length
159#
160# - pmtu_vti6_default_mtu
161#	Same as above, for IPv6
162#
163# - pmtu_vti4_link_add_mtu
164#	Set up vti4 interface passing MTU value at link creation, check MTU is
165#	configured, and that link is not created with invalid MTU values
166#
167# - pmtu_vti6_link_add_mtu
168#	Same as above, for IPv6
169#
170# - pmtu_vti6_link_change_mtu
171#	Set up two dummy interfaces with different MTUs, create a vti6 tunnel
172#	and check that configured MTU is used on link creation and changes, and
173#	that MTU is properly calculated instead when MTU is not configured from
174#	userspace
175#
176# - cleanup_ipv4_exception
177#	Similar to pmtu_ipv4_vxlan4_exception, but explicitly generate PMTU
178#	exceptions on multiple CPUs and check that the veth device tear-down
179# 	happens in a timely manner
180#
181# - cleanup_ipv6_exception
182#	Same as above, but use IPv6 transport from A to B
183#
184# - list_flush_ipv4_exception
185#	Using the same topology as in pmtu_ipv4, create exceptions, and check
186#	they are shown when listing exception caches, gone after flushing them
187#
188# - list_flush_ipv6_exception
189#	Using the same topology as in pmtu_ipv6, create exceptions, and check
190#	they are shown when listing exception caches, gone after flushing them
191#
192# - pmtu_ipv4_route_change
193#	Use the same topology as in pmtu_ipv4, but issue a route replacement
194#	command and delete the corresponding device afterward. This tests for
195#	proper cleanup of the PMTU exceptions by the route replacement path.
196#	Device unregistration should complete successfully
197#
198# - pmtu_ipv6_route_change
199#	Same as above but with IPv6
200#
201# - pmtu_ipv4_mp_exceptions
202#	Use the same topology as in pmtu_ipv4, but add routeable addresses
203#	on host A and B on lo reachable via both routers. Host A and B
204#	addresses have multipath routes to each other, b_r1 mtu = 1500.
205#	Check that PMTU exceptions are created for both paths.
206
207source lib.sh
208
209PAUSE_ON_FAIL=no
210VERBOSE=0
211TRACING=0
212
213# Some systems don't have a ping6 binary anymore
214which ping6 > /dev/null 2>&1 && ping6=$(which ping6) || ping6=$(which ping)
215
216#               Name                          Description                  re-run with nh
217tests="
218	pmtu_ipv4_exception		ipv4: PMTU exceptions			1
219	pmtu_ipv6_exception		ipv6: PMTU exceptions			1
220	pmtu_ipv4_dscp_icmp_exception	ICMPv4 with DSCP and ECN: PMTU exceptions	1
221	pmtu_ipv4_dscp_udp_exception	UDPv4 with DSCP and ECN: PMTU exceptions	1
222	pmtu_ipv4_vxlan4_exception	IPv4 over vxlan4: PMTU exceptions	1
223	pmtu_ipv6_vxlan4_exception	IPv6 over vxlan4: PMTU exceptions	1
224	pmtu_ipv4_vxlan6_exception	IPv4 over vxlan6: PMTU exceptions	1
225	pmtu_ipv6_vxlan6_exception	IPv6 over vxlan6: PMTU exceptions	1
226	pmtu_ipv4_geneve4_exception	IPv4 over geneve4: PMTU exceptions	1
227	pmtu_ipv6_geneve4_exception	IPv6 over geneve4: PMTU exceptions	1
228	pmtu_ipv4_geneve6_exception	IPv4 over geneve6: PMTU exceptions	1
229	pmtu_ipv6_geneve6_exception	IPv6 over geneve6: PMTU exceptions	1
230	pmtu_ipv4_br_vxlan4_exception	IPv4, bridged vxlan4: PMTU exceptions	1
231	pmtu_ipv6_br_vxlan4_exception	IPv6, bridged vxlan4: PMTU exceptions	1
232	pmtu_ipv4_br_vxlan6_exception	IPv4, bridged vxlan6: PMTU exceptions	1
233	pmtu_ipv6_br_vxlan6_exception	IPv6, bridged vxlan6: PMTU exceptions	1
234	pmtu_ipv4_br_geneve4_exception	IPv4, bridged geneve4: PMTU exceptions	1
235	pmtu_ipv6_br_geneve4_exception	IPv6, bridged geneve4: PMTU exceptions	1
236	pmtu_ipv4_br_geneve6_exception	IPv4, bridged geneve6: PMTU exceptions	1
237	pmtu_ipv6_br_geneve6_exception	IPv6, bridged geneve6: PMTU exceptions	1
238	pmtu_ipv4_ovs_vxlan4_exception	IPv4, OVS vxlan4: PMTU exceptions	1
239	pmtu_ipv6_ovs_vxlan4_exception	IPv6, OVS vxlan4: PMTU exceptions	1
240	pmtu_ipv4_ovs_vxlan6_exception	IPv4, OVS vxlan6: PMTU exceptions	1
241	pmtu_ipv6_ovs_vxlan6_exception	IPv6, OVS vxlan6: PMTU exceptions	1
242	pmtu_ipv4_ovs_geneve4_exception	IPv4, OVS geneve4: PMTU exceptions	1
243	pmtu_ipv6_ovs_geneve4_exception	IPv6, OVS geneve4: PMTU exceptions	1
244	pmtu_ipv4_ovs_geneve6_exception	IPv4, OVS geneve6: PMTU exceptions	1
245	pmtu_ipv6_ovs_geneve6_exception	IPv6, OVS geneve6: PMTU exceptions	1
246	pmtu_ipv4_fou4_exception	IPv4 over fou4: PMTU exceptions		1
247	pmtu_ipv6_fou4_exception	IPv6 over fou4: PMTU exceptions		1
248	pmtu_ipv4_fou6_exception	IPv4 over fou6: PMTU exceptions		1
249	pmtu_ipv6_fou6_exception	IPv6 over fou6: PMTU exceptions		1
250	pmtu_ipv4_gue4_exception	IPv4 over gue4: PMTU exceptions		1
251	pmtu_ipv6_gue4_exception	IPv6 over gue4: PMTU exceptions		1
252	pmtu_ipv4_gue6_exception	IPv4 over gue6: PMTU exceptions		1
253	pmtu_ipv6_gue6_exception	IPv6 over gue6: PMTU exceptions		1
254	pmtu_ipv4_ipv4_exception	IPv4 over IPv4: PMTU exceptions		1
255	pmtu_ipv6_ipv4_exception	IPv6 over IPv4: PMTU exceptions		1
256	pmtu_ipv4_ipv6_exception	IPv4 over IPv6: PMTU exceptions		1
257	pmtu_ipv6_ipv6_exception	IPv6 over IPv6: PMTU exceptions		1
258	pmtu_vti6_exception		vti6: PMTU exceptions			0
259	pmtu_vti4_exception		vti4: PMTU exceptions			0
260	pmtu_vti6_udp_exception		vti6: PMTU exceptions (ESP-in-UDP)	0
261	pmtu_vti4_udp_exception		vti4: PMTU exceptions (ESP-in-UDP)	0
262	pmtu_vti6_udp_routed_exception	vti6: PMTU exceptions, routed (ESP-in-UDP)	0
263	pmtu_vti4_udp_routed_exception	vti4: PMTU exceptions, routed (ESP-in-UDP)	0
264	pmtu_vti4_default_mtu		vti4: default MTU assignment		0
265	pmtu_vti6_default_mtu		vti6: default MTU assignment		0
266	pmtu_vti4_link_add_mtu		vti4: MTU setting on link creation	0
267	pmtu_vti6_link_add_mtu		vti6: MTU setting on link creation	0
268	pmtu_vti6_link_change_mtu	vti6: MTU changes on link changes	0
269	cleanup_ipv4_exception		ipv4: cleanup of cached exceptions	1
270	cleanup_ipv6_exception		ipv6: cleanup of cached exceptions	1
271	list_flush_ipv4_exception	ipv4: list and flush cached exceptions	1
272	list_flush_ipv6_exception	ipv6: list and flush cached exceptions	1
273	pmtu_ipv4_route_change		ipv4: PMTU exception w/route replace	1
274	pmtu_ipv6_route_change		ipv6: PMTU exception w/route replace	1
275	pmtu_ipv4_mp_exceptions		ipv4: PMTU multipath nh exceptions	1"
276
277# Addressing and routing for tests with routers: four network segments, with
278# index SEGMENT between 1 and 4, a common prefix (PREFIX4 or PREFIX6) and an
279# identifier ID, which is 1 for hosts (A and B), 2 for routers (R1 and R2).
280# Addresses are:
281# - IPv4: PREFIX4.SEGMENT.ID (/24)
282# - IPv6: PREFIX6:SEGMENT::ID (/64)
283prefix4="10.0"
284prefix6="fc00"
285a_r1=1
286a_r2=2
287b_r1=3
288b_r2=4
289#	ns	peer	segment
290routing_addrs="
291	A	R1	${a_r1}
292	A	R2	${a_r2}
293	B	R1	${b_r1}
294	B	R2	${b_r2}
295"
296# Traffic from A to B goes through R1 by default, and through R2, if destined to
297# B's address on the b_r2 segment.
298# Traffic from B to A goes through R1.
299#	ns	destination		gateway
300routes="
301	A	default			${prefix4}.${a_r1}.2
302	A	${prefix4}.${b_r2}.1	${prefix4}.${a_r2}.2
303	B	default			${prefix4}.${b_r1}.2
304
305	A	default			${prefix6}:${a_r1}::2
306	A	${prefix6}:${b_r2}::1	${prefix6}:${a_r2}::2
307	B	default			${prefix6}:${b_r1}::2
308"
309USE_NH="no"
310#	ns	family	nh id	   destination		gateway
311nexthops="
312	A	4	41	${prefix4}.${a_r1}.2	veth_A-R1
313	A	4	42	${prefix4}.${a_r2}.2	veth_A-R2
314	B	4	41	${prefix4}.${b_r1}.2	veth_B-R1
315
316	A	6	61	${prefix6}:${a_r1}::2	veth_A-R1
317	A	6	62	${prefix6}:${a_r2}::2	veth_A-R2
318	B	6	61	${prefix6}:${b_r1}::2	veth_B-R1
319"
320
321# nexthop id correlates to id in nexthops config above
322#	ns    family	prefix			nh id
323routes_nh="
324	A	4	default			41
325	A	4	${prefix4}.${b_r2}.1	42
326	B	4	default			41
327
328	A	6	default			61
329	A	6	${prefix6}:${b_r2}::1	62
330	B	6	default			61
331"
332
333policy_mark=0x04
334rt_table=main
335
336veth4_a_addr="192.168.1.1"
337veth4_b_addr="192.168.1.2"
338veth4_c_addr="192.168.2.10"
339veth4_mask="24"
340veth6_a_addr="fd00:1::a"
341veth6_b_addr="fd00:1::b"
342veth6_c_addr="fd00:2::c"
343veth6_mask="64"
344
345tunnel4_a_addr="192.168.2.1"
346tunnel4_b_addr="192.168.2.2"
347tunnel4_mask="24"
348tunnel6_a_addr="fd00:2::a"
349tunnel6_b_addr="fd00:2::b"
350tunnel6_mask="64"
351
352host4_a_addr="192.168.99.99"
353host4_b_addr="192.168.88.88"
354
355dummy6_0_prefix="fc00:1000::"
356dummy6_1_prefix="fc00:1001::"
357dummy6_mask="64"
358
359err_buf=
360tcpdump_pids=
361nettest_pids=
362socat_pids=
363tmpoutfile=
364
365err() {
366	err_buf="${err_buf}${1}
367"
368}
369
370err_flush() {
371	echo -n "${err_buf}"
372	err_buf=
373}
374
375run_cmd() {
376	cmd="$*"
377
378	if [ "$VERBOSE" = "1" ]; then
379		printf "    COMMAND: $cmd\n"
380	fi
381
382	out="$($cmd 2>&1)"
383	rc=$?
384	if [ "$VERBOSE" = "1" -a -n "$out" ]; then
385		echo "    $out"
386		echo
387	fi
388
389	return $rc
390}
391
392run_cmd_bg() {
393	cmd="$*"
394
395	if [ "$VERBOSE" = "1" ]; then
396		printf "    COMMAND: %s &\n" "${cmd}"
397	fi
398
399	$cmd 2>&1 &
400}
401
402# Find the auto-generated name for this namespace
403nsname() {
404	eval echo \$NS_$1
405}
406
407setup_fou_or_gue() {
408	outer="${1}"
409	inner="${2}"
410	encap="${3}"
411
412	if [ "${outer}" = "4" ]; then
413		modprobe fou || return $ksft_skip
414		a_addr="${prefix4}.${a_r1}.1"
415		b_addr="${prefix4}.${b_r1}.1"
416		if [ "${inner}" = "4" ]; then
417			type="ipip"
418			ipproto="4"
419		else
420			type="sit"
421			ipproto="41"
422		fi
423	else
424		modprobe fou6 || return $ksft_skip
425		a_addr="${prefix6}:${a_r1}::1"
426		b_addr="${prefix6}:${b_r1}::1"
427		if [ "${inner}" = "4" ]; then
428			type="ip6tnl"
429			mode="mode ipip6"
430			ipproto="4 -6"
431		else
432			type="ip6tnl"
433			mode="mode ip6ip6"
434			ipproto="41 -6"
435		fi
436	fi
437
438	run_cmd ${ns_a} ip fou add port 5555 ipproto ${ipproto} || return $ksft_skip
439	run_cmd ${ns_a} ip link add ${encap}_a type ${type} ${mode} local ${a_addr} remote ${b_addr} encap ${encap} encap-sport auto encap-dport 5556 || return $ksft_skip
440
441	run_cmd ${ns_b} ip fou add port 5556 ipproto ${ipproto}
442	run_cmd ${ns_b} ip link add ${encap}_b type ${type} ${mode} local ${b_addr} remote ${a_addr} encap ${encap} encap-sport auto encap-dport 5555
443
444	if [ "${inner}" = "4" ]; then
445		run_cmd ${ns_a} ip addr add ${tunnel4_a_addr}/${tunnel4_mask} dev ${encap}_a
446		run_cmd ${ns_b} ip addr add ${tunnel4_b_addr}/${tunnel4_mask} dev ${encap}_b
447	else
448		run_cmd ${ns_a} ip addr add ${tunnel6_a_addr}/${tunnel6_mask} dev ${encap}_a
449		run_cmd ${ns_b} ip addr add ${tunnel6_b_addr}/${tunnel6_mask} dev ${encap}_b
450	fi
451
452	run_cmd ${ns_a} ip link set ${encap}_a up
453	run_cmd ${ns_b} ip link set ${encap}_b up
454}
455
456setup_fou44() {
457	setup_fou_or_gue 4 4 fou
458}
459
460setup_fou46() {
461	setup_fou_or_gue 4 6 fou
462}
463
464setup_fou64() {
465	setup_fou_or_gue 6 4 fou
466}
467
468setup_fou66() {
469	setup_fou_or_gue 6 6 fou
470}
471
472setup_gue44() {
473	setup_fou_or_gue 4 4 gue
474}
475
476setup_gue46() {
477	setup_fou_or_gue 4 6 gue
478}
479
480setup_gue64() {
481	setup_fou_or_gue 6 4 gue
482}
483
484setup_gue66() {
485	setup_fou_or_gue 6 6 gue
486}
487
488setup_ipvX_over_ipvY() {
489	inner=${1}
490	outer=${2}
491
492	if [ "${outer}" -eq 4 ]; then
493		a_addr="${prefix4}.${a_r1}.1"
494		b_addr="${prefix4}.${b_r1}.1"
495		if [ "${inner}" -eq 4 ]; then
496			type="ipip"
497			mode="ipip"
498		else
499			type="sit"
500			mode="ip6ip"
501		fi
502	else
503		a_addr="${prefix6}:${a_r1}::1"
504		b_addr="${prefix6}:${b_r1}::1"
505		type="ip6tnl"
506		if [ "${inner}" -eq 4 ]; then
507			mode="ipip6"
508		else
509			mode="ip6ip6"
510		fi
511	fi
512
513	run_cmd ${ns_a} ip link add ip_a type ${type} local ${a_addr} remote ${b_addr} mode ${mode} || return $ksft_skip
514	run_cmd ${ns_b} ip link add ip_b type ${type} local ${b_addr} remote ${a_addr} mode ${mode}
515
516	run_cmd ${ns_a} ip link set ip_a up
517	run_cmd ${ns_b} ip link set ip_b up
518
519	if [ "${inner}" = "4" ]; then
520		run_cmd ${ns_a} ip addr add ${tunnel4_a_addr}/${tunnel4_mask} dev ip_a
521		run_cmd ${ns_b} ip addr add ${tunnel4_b_addr}/${tunnel4_mask} dev ip_b
522	else
523		run_cmd ${ns_a} ip addr add ${tunnel6_a_addr}/${tunnel6_mask} dev ip_a
524		run_cmd ${ns_b} ip addr add ${tunnel6_b_addr}/${tunnel6_mask} dev ip_b
525	fi
526}
527
528setup_ip4ip4() {
529	setup_ipvX_over_ipvY 4 4
530}
531
532setup_ip6ip4() {
533	setup_ipvX_over_ipvY 6 4
534}
535
536setup_ip4ip6() {
537	setup_ipvX_over_ipvY 4 6
538}
539
540setup_ip6ip6() {
541	setup_ipvX_over_ipvY 6 6
542}
543
544setup_namespaces() {
545	setup_ns NS_A NS_B NS_C NS_R1 NS_R2
546	for n in ${NS_A} ${NS_B} ${NS_C} ${NS_R1} ${NS_R2}; do
547		# Disable DAD, so that we don't have to wait to use the
548		# configured IPv6 addresses
549		ip netns exec ${n} sysctl -q net/ipv6/conf/default/accept_dad=0
550	done
551	ns_a="ip netns exec ${NS_A}"
552	ns_b="ip netns exec ${NS_B}"
553	ns_c="ip netns exec ${NS_C}"
554	ns_r1="ip netns exec ${NS_R1}"
555	ns_r2="ip netns exec ${NS_R2}"
556}
557
558setup_veth() {
559	run_cmd ${ns_a} ip link add veth_a type veth peer name veth_b || return 1
560	run_cmd ${ns_a} ip link set veth_b netns ${NS_B}
561
562	run_cmd ${ns_a} ip addr add ${veth4_a_addr}/${veth4_mask} dev veth_a
563	run_cmd ${ns_b} ip addr add ${veth4_b_addr}/${veth4_mask} dev veth_b
564
565	run_cmd ${ns_a} ip addr add ${veth6_a_addr}/${veth6_mask} dev veth_a
566	run_cmd ${ns_b} ip addr add ${veth6_b_addr}/${veth6_mask} dev veth_b
567
568	run_cmd ${ns_a} ip link set veth_a up
569	run_cmd ${ns_b} ip link set veth_b up
570}
571
572setup_vti() {
573	proto=${1}
574	veth_a_addr="${2}"
575	veth_b_addr="${3}"
576	vti_a_addr="${4}"
577	vti_b_addr="${5}"
578	vti_mask=${6}
579
580	[ ${proto} -eq 6 ] && vti_type="vti6" || vti_type="vti"
581
582	run_cmd ${ns_a} ip link add vti${proto}_a type ${vti_type} local ${veth_a_addr} remote ${veth_b_addr} key 10 || return 1
583	run_cmd ${ns_b} ip link add vti${proto}_b type ${vti_type} local ${veth_b_addr} remote ${veth_a_addr} key 10
584
585	run_cmd ${ns_a} ip addr add ${vti_a_addr}/${vti_mask} dev vti${proto}_a
586	run_cmd ${ns_b} ip addr add ${vti_b_addr}/${vti_mask} dev vti${proto}_b
587
588	run_cmd ${ns_a} ip link set vti${proto}_a up
589	run_cmd ${ns_b} ip link set vti${proto}_b up
590}
591
592setup_vti4() {
593	setup_vti 4 ${veth4_a_addr} ${veth4_b_addr} ${tunnel4_a_addr} ${tunnel4_b_addr} ${tunnel4_mask}
594}
595
596setup_vti6() {
597	setup_vti 6 ${veth6_a_addr} ${veth6_b_addr} ${tunnel6_a_addr} ${tunnel6_b_addr} ${tunnel6_mask}
598}
599
600setup_vti4routed() {
601	setup_vti 4 ${prefix4}.${a_r1}.1 ${prefix4}.${b_r1}.1 ${tunnel4_a_addr} ${tunnel4_b_addr} ${tunnel4_mask}
602}
603
604setup_vti6routed() {
605	setup_vti 6 ${prefix6}:${a_r1}::1 ${prefix6}:${b_r1}::1 ${tunnel6_a_addr} ${tunnel6_b_addr} ${tunnel6_mask}
606}
607
608setup_vxlan_or_geneve() {
609	type="${1}"
610	a_addr="${2}"
611	b_addr="${3}"
612	opts="${4}"
613	br_if_a="${5}"
614
615	if [ "${type}" = "vxlan" ]; then
616		opts="${opts} ttl 64 dstport 4789"
617		opts_a="local ${a_addr}"
618		opts_b="local ${b_addr}"
619	else
620		opts_a=""
621		opts_b=""
622	fi
623
624	run_cmd ${ns_a} ip link add ${type}_a type ${type} id 1 ${opts_a} remote ${b_addr} ${opts} || return 1
625	run_cmd ${ns_b} ip link add ${type}_b type ${type} id 1 ${opts_b} remote ${a_addr} ${opts}
626
627	if [ -n "${br_if_a}" ]; then
628		run_cmd ${ns_a} ip addr add ${tunnel4_a_addr}/${tunnel4_mask} dev ${br_if_a}
629		run_cmd ${ns_a} ip addr add ${tunnel6_a_addr}/${tunnel6_mask} dev ${br_if_a}
630		run_cmd ${ns_a} ip link set ${type}_a master ${br_if_a}
631	else
632		run_cmd ${ns_a} ip addr add ${tunnel4_a_addr}/${tunnel4_mask} dev ${type}_a
633		run_cmd ${ns_a} ip addr add ${tunnel6_a_addr}/${tunnel6_mask} dev ${type}_a
634	fi
635
636	run_cmd ${ns_b} ip addr add ${tunnel4_b_addr}/${tunnel4_mask} dev ${type}_b
637	run_cmd ${ns_b} ip addr add ${tunnel6_b_addr}/${tunnel6_mask} dev ${type}_b
638
639	run_cmd ${ns_a} ip link set ${type}_a up
640	run_cmd ${ns_b} ip link set ${type}_b up
641}
642
643setup_geneve4() {
644	setup_vxlan_or_geneve geneve ${prefix4}.${a_r1}.1  ${prefix4}.${b_r1}.1  "df set"
645}
646
647setup_vxlan4() {
648	setup_vxlan_or_geneve vxlan  ${prefix4}.${a_r1}.1  ${prefix4}.${b_r1}.1  "df set"
649}
650
651setup_geneve6() {
652	setup_vxlan_or_geneve geneve ${prefix6}:${a_r1}::1 ${prefix6}:${b_r1}::1 ""
653}
654
655setup_vxlan6() {
656	setup_vxlan_or_geneve vxlan  ${prefix6}:${a_r1}::1 ${prefix6}:${b_r1}::1 ""
657}
658
659setup_bridged_geneve4() {
660	setup_vxlan_or_geneve geneve ${prefix4}.${a_r1}.1  ${prefix4}.${b_r1}.1  "df set" "br0"
661}
662
663setup_bridged_vxlan4() {
664	setup_vxlan_or_geneve vxlan  ${prefix4}.${a_r1}.1  ${prefix4}.${b_r1}.1  "df set" "br0"
665}
666
667setup_bridged_geneve6() {
668	setup_vxlan_or_geneve geneve ${prefix6}:${a_r1}::1 ${prefix6}:${b_r1}::1 "" "br0"
669}
670
671setup_bridged_vxlan6() {
672	setup_vxlan_or_geneve vxlan  ${prefix6}:${a_r1}::1 ${prefix6}:${b_r1}::1 "" "br0"
673}
674
675setup_xfrm() {
676	proto=${1}
677	veth_a_addr="${2}"
678	veth_b_addr="${3}"
679	encap=${4}
680
681	run_cmd ${ns_a} ip -${proto} xfrm state add src ${veth_a_addr} dst ${veth_b_addr} spi 0x1000 proto esp aead 'rfc4106(gcm(aes))' 0x0f0f0f0f0f0f0f0f0f0f0f0f0f0f0f0f0f0f0f0f 128 mode tunnel ${encap} || return 1
682	run_cmd ${ns_a} ip -${proto} xfrm state add src ${veth_b_addr} dst ${veth_a_addr} spi 0x1001 proto esp aead 'rfc4106(gcm(aes))' 0x0f0f0f0f0f0f0f0f0f0f0f0f0f0f0f0f0f0f0f0f 128 mode tunnel ${encap}
683	run_cmd ${ns_a} ip -${proto} xfrm policy add dir out mark 10 tmpl src ${veth_a_addr} dst ${veth_b_addr} proto esp mode tunnel
684	run_cmd ${ns_a} ip -${proto} xfrm policy add dir in mark 10 tmpl src ${veth_b_addr} dst ${veth_a_addr} proto esp mode tunnel
685
686	run_cmd ${ns_b} ip -${proto} xfrm state add src ${veth_a_addr} dst ${veth_b_addr} spi 0x1000 proto esp aead 'rfc4106(gcm(aes))' 0x0f0f0f0f0f0f0f0f0f0f0f0f0f0f0f0f0f0f0f0f 128 mode tunnel ${encap}
687	run_cmd ${ns_b} ip -${proto} xfrm state add src ${veth_b_addr} dst ${veth_a_addr} spi 0x1001 proto esp aead 'rfc4106(gcm(aes))' 0x0f0f0f0f0f0f0f0f0f0f0f0f0f0f0f0f0f0f0f0f 128 mode tunnel ${encap}
688	run_cmd ${ns_b} ip -${proto} xfrm policy add dir out mark 10 tmpl src ${veth_b_addr} dst ${veth_a_addr} proto esp mode tunnel
689	run_cmd ${ns_b} ip -${proto} xfrm policy add dir in mark 10 tmpl src ${veth_a_addr} dst ${veth_b_addr} proto esp mode tunnel
690}
691
692setup_nettest_xfrm() {
693	check_gen_prog "nettest"
694
695	[ ${1} -eq 6 ] && proto="-6" || proto=""
696	port=${2}
697
698	run_cmd_bg "${ns_a}" nettest "${proto}" -q -D -s -x -p "${port}" -t 5
699	nettest_pids="${nettest_pids} $!"
700
701	run_cmd_bg "${ns_b}" nettest "${proto}" -q -D -s -x -p "${port}" -t 5
702	nettest_pids="${nettest_pids} $!"
703}
704
705setup_xfrm4() {
706	setup_xfrm 4 ${veth4_a_addr} ${veth4_b_addr}
707}
708
709setup_xfrm6() {
710	setup_xfrm 6 ${veth6_a_addr} ${veth6_b_addr}
711}
712
713setup_xfrm4udp() {
714	setup_xfrm 4 ${veth4_a_addr} ${veth4_b_addr} "encap espinudp 4500 4500 0.0.0.0" && \
715		setup_nettest_xfrm 4 4500
716}
717
718setup_xfrm6udp() {
719	setup_xfrm 6 ${veth6_a_addr} ${veth6_b_addr} "encap espinudp 4500 4500 0.0.0.0" && \
720		setup_nettest_xfrm 6 4500
721}
722
723setup_xfrm4udprouted() {
724	setup_xfrm 4 ${prefix4}.${a_r1}.1 ${prefix4}.${b_r1}.1 "encap espinudp 4500 4500 0.0.0.0" && \
725		setup_nettest_xfrm 4 4500
726}
727
728setup_xfrm6udprouted() {
729	setup_xfrm 6 ${prefix6}:${a_r1}::1 ${prefix6}:${b_r1}::1 "encap espinudp 4500 4500 0.0.0.0" && \
730		setup_nettest_xfrm 6 4500
731}
732
733setup_routing_old() {
734	for i in ${routes}; do
735		[ "${ns}" = "" ]	&& ns="${i}"		&& continue
736		[ "${addr}" = "" ]	&& addr="${i}"		&& continue
737		[ "${gw}" = "" ]	&& gw="${i}"
738
739		ns_name="$(nsname ${ns})"
740
741		ip -n "${ns_name}" route add "${addr}" table "${rt_table}" via "${gw}"
742
743		ns=""; addr=""; gw=""
744	done
745}
746
747setup_routing_new() {
748	for i in ${nexthops}; do
749		[ "${ns}" = "" ]	&& ns="${i}"		&& continue
750		[ "${fam}" = "" ]	&& fam="${i}"		&& continue
751		[ "${nhid}" = "" ]	&& nhid="${i}"		&& continue
752		[ "${gw}" = "" ]	&& gw="${i}"		&& continue
753		[ "${dev}" = "" ]	&& dev="${i}"
754
755		ns_name="$(nsname ${ns})"
756
757		ip -n ${ns_name} -${fam} nexthop add id ${nhid} via ${gw} dev ${dev}
758
759		ns=""; fam=""; nhid=""; gw=""; dev=""
760
761	done
762
763	for i in ${routes_nh}; do
764		[ "${ns}" = "" ]	&& ns="${i}"		&& continue
765		[ "${fam}" = "" ]	&& fam="${i}"		&& continue
766		[ "${addr}" = "" ]	&& addr="${i}"		&& continue
767		[ "${nhid}" = "" ]	&& nhid="${i}"
768
769		ns_name="$(nsname ${ns})"
770
771		ip -n "${ns_name}" -"${fam}" route add "${addr}" table "${rt_table}" nhid "${nhid}"
772
773		ns=""; fam=""; addr=""; nhid=""
774	done
775}
776
777setup_routing() {
778	for i in ${NS_R1} ${NS_R2}; do
779		ip netns exec ${i} sysctl -q net/ipv4/ip_forward=1
780		ip netns exec ${i} sysctl -q net/ipv6/conf/all/forwarding=1
781	done
782
783	for i in ${routing_addrs}; do
784		[ "${ns}" = "" ]	&& ns="${i}"		&& continue
785		[ "${peer}" = "" ]	&& peer="${i}"		&& continue
786		[ "${segment}" = "" ]	&& segment="${i}"
787
788		ns_name="$(nsname ${ns})"
789		peer_name="$(nsname ${peer})"
790		if="veth_${ns}-${peer}"
791		ifpeer="veth_${peer}-${ns}"
792
793		# Create veth links
794		ip link add ${if} up netns ${ns_name} type veth peer name ${ifpeer} netns ${peer_name} || return 1
795		ip -n ${peer_name} link set dev ${ifpeer} up
796
797		# Add addresses
798		ip -n ${ns_name}   addr add ${prefix4}.${segment}.1/24  dev ${if}
799		ip -n ${ns_name}   addr add ${prefix6}:${segment}::1/64 dev ${if}
800
801		ip -n ${peer_name} addr add ${prefix4}.${segment}.2/24  dev ${ifpeer}
802		ip -n ${peer_name} addr add ${prefix6}:${segment}::2/64 dev ${ifpeer}
803
804		ns=""; peer=""; segment=""
805	done
806
807	if [ "$USE_NH" = "yes" ]; then
808		setup_routing_new
809	else
810		setup_routing_old
811	fi
812
813	return 0
814}
815
816setup_policy_routing() {
817	setup_routing
818
819	ip -netns "${NS_A}" -4 rule add dsfield "${policy_mark}" \
820		table "${rt_table}"
821
822	# Set the IPv4 Don't Fragment bit with tc, since socat doesn't seem to
823	# have an option do to it.
824	tc -netns "${NS_A}" qdisc replace dev veth_A-R1 root prio
825	tc -netns "${NS_A}" qdisc replace dev veth_A-R2 root prio
826	tc -netns "${NS_A}" filter add dev veth_A-R1                      \
827		protocol ipv4 flower ip_proto udp                         \
828		action pedit ex munge ip df set 0x40 pipe csum ip and udp
829	tc -netns "${NS_A}" filter add dev veth_A-R2                      \
830		protocol ipv4 flower ip_proto udp                         \
831		action pedit ex munge ip df set 0x40 pipe csum ip and udp
832}
833
834setup_bridge() {
835	run_cmd ${ns_a} ip link add br0 type bridge || return $ksft_skip
836	run_cmd ${ns_a} ip link set br0 up
837
838	run_cmd ${ns_c} ip link add veth_C-A type veth peer name veth_A-C
839	run_cmd ${ns_c} ip link set veth_A-C netns ${NS_A}
840
841	run_cmd ${ns_a} ip link set veth_A-C up
842	run_cmd ${ns_c} ip link set veth_C-A up
843	run_cmd ${ns_c} ip addr add ${veth4_c_addr}/${veth4_mask} dev veth_C-A
844	run_cmd ${ns_c} ip addr add ${veth6_c_addr}/${veth6_mask} dev veth_C-A
845	run_cmd ${ns_a} ip link set veth_A-C master br0
846}
847
848setup_ovs_via_internal_utility() {
849	type="${1}"
850	a_addr="${2}"
851	b_addr="${3}"
852	dport="${4}"
853
854	run_cmd python3 ./openvswitch/ovs-dpctl.py add-if ovs_br0 ${type}_a -t ${type} || return 1
855
856	ports=$(python3 ./openvswitch/ovs-dpctl.py show)
857	br0_port=$(echo "$ports" | grep -E "\sovs_br0" | sed -e 's@port @@' | cut -d: -f1 | xargs)
858	type_a_port=$(echo "$ports" | grep ${type}_a | sed -e 's@port @@' | cut -d: -f1 | xargs)
859	veth_a_port=$(echo "$ports" | grep veth_A | sed -e 's@port @@' | cut -d: -f1 | xargs)
860
861	v4_a_tun="${prefix4}.${a_r1}.1"
862	v4_b_tun="${prefix4}.${b_r1}.1"
863
864	v6_a_tun="${prefix6}:${a_r1}::1"
865	v6_b_tun="${prefix6}:${b_r1}::1"
866
867	if [ "${v4_a_tun}" = "${a_addr}" ]; then
868		run_cmd python3 ./openvswitch/ovs-dpctl.py add-flow ovs_br0 \
869		    "recirc_id(0),in_port(${veth_a_port}),eth(),eth_type(0x0800),ipv4()" \
870		    "set(tunnel(tun_id=1,dst=${v4_b_tun},ttl=64,tp_dst=${dport},flags(df|csum))),${type_a_port}"
871		run_cmd python3 ./openvswitch/ovs-dpctl.py add-flow ovs_br0 \
872		    "recirc_id(0),in_port(${veth_a_port}),eth(),eth_type(0x86dd),ipv6()" \
873		    "set(tunnel(tun_id=1,dst=${v4_b_tun},ttl=64,tp_dst=${dport},flags(df|csum))),${type_a_port}"
874		run_cmd python3 ./openvswitch/ovs-dpctl.py add-flow ovs_br0 \
875		    "recirc_id(0),tunnel(tun_id=1,src=${v4_b_tun},dst=${v4_a_tun}),in_port(${type_a_port}),eth(),eth_type(0x0800),ipv4()" \
876		    "${veth_a_port}"
877		run_cmd python3 ./openvswitch/ovs-dpctl.py add-flow ovs_br0 \
878		    "recirc_id(0),tunnel(tun_id=1,src=${v4_b_tun},dst=${v4_a_tun}),in_port(${type_a_port}),eth(),eth_type(0x86dd),ipv6()" \
879		    "${veth_a_port}"
880		run_cmd python3 ./openvswitch/ovs-dpctl.py add-flow ovs_br0 \
881		    "recirc_id(0),tunnel(tun_id=1,src=${v4_b_tun},dst=${v4_a_tun}),in_port(${type_a_port}),eth(),eth_type(0x0806),arp()" \
882		    "${veth_a_port}"
883		run_cmd python3 ./openvswitch/ovs-dpctl.py add-flow ovs_br0 \
884		    "recirc_id(0),in_port(${veth_a_port}),eth(),eth_type(0x0806),arp(sip=${veth4_c_addr},tip=${tunnel4_b_addr})" \
885		    "set(tunnel(tun_id=1,dst=${v4_b_tun},ttl=64,tp_dst=${dport},flags(df|csum))),${type_a_port}"
886	else
887		run_cmd python3 ./openvswitch/ovs-dpctl.py add-flow ovs_br0 \
888		    "recirc_id(0),in_port(${veth_a_port}),eth(),eth_type(0x0800),ipv4()" \
889		    "set(tunnel(tun_id=1,ipv6_dst=${v6_b_tun},ttl=64,tp_dst=${dport},flags(df|csum))),${type_a_port}"
890		run_cmd python3 ./openvswitch/ovs-dpctl.py add-flow ovs_br0 \
891		    "recirc_id(0),in_port(${veth_a_port}),eth(),eth_type(0x86dd),ipv6()" \
892		    "set(tunnel(tun_id=1,ipv6_dst=${v6_b_tun},ttl=64,tp_dst=${dport},flags(df|csum))),${type_a_port}"
893		run_cmd python3 ./openvswitch/ovs-dpctl.py add-flow ovs_br0 \
894		    "recirc_id(0),tunnel(tun_id=1,ipv6_src=${v6_b_tun},ipv6_dst=${v6_a_tun}),in_port(${type_a_port}),eth(),eth_type(0x0800),ipv4()" \
895		    "${veth_a_port}"
896		run_cmd python3 ./openvswitch/ovs-dpctl.py add-flow ovs_br0 \
897		    "recirc_id(0),tunnel(tun_id=1,ipv6_src=${v6_b_tun},ipv6_dst=${v6_a_tun}),in_port(${type_a_port}),eth(),eth_type(0x86dd),ipv6()" \
898		    "${veth_a_port}"
899		run_cmd python3 ./openvswitch/ovs-dpctl.py add-flow ovs_br0 \
900		    "recirc_id(0),tunnel(tun_id=1,ipv6_src=${v6_b_tun},ipv6_dst=${v6_a_tun}),in_port(${type_a_port}),eth(),eth_type(0x0806),arp()" \
901		    "${veth_a_port}"
902		run_cmd python3 ./openvswitch/ovs-dpctl.py add-flow ovs_br0 \
903		    "recirc_id(0),in_port(${veth_a_port}),eth(),eth_type(0x0806),arp(sip=${veth4_c_addr},tip=${tunnel4_b_addr})" \
904		    "set(tunnel(tun_id=1,ipv6_dst=${v6_b_tun},ttl=64,tp_dst=${dport},flags(df|csum))),${type_a_port}"
905	fi
906}
907
908setup_ovs_via_vswitchd() {
909	type="${1}"
910	b_addr="${2}"
911
912	run_cmd ovs-vsctl add-port ovs_br0 ${type}_a -- \
913		set interface ${type}_a type=${type} \
914		options:remote_ip=${b_addr} options:key=1 options:csum=true || return 1
915}
916
917setup_ovs_vxlan_or_geneve() {
918	type="${1}"
919	a_addr="${2}"
920	b_addr="${3}"
921	dport="6081"
922
923	if [ "${type}" = "vxlan" ]; then
924		dport="4789"
925		opts="${opts} ttl 64 dstport 4789"
926		opts_b="local ${b_addr}"
927	fi
928
929	setup_ovs_via_internal_utility "${type}" "${a_addr}" "${b_addr}" \
930				       "${dport}" || \
931	    setup_ovs_via_vswitchd "${type}" "${b_addr}" || return 1
932
933	run_cmd ${ns_b} ip link add ${type}_b type ${type} id 1 ${opts_b} remote ${a_addr} ${opts} || return 1
934
935	run_cmd ${ns_b} ip addr add ${tunnel4_b_addr}/${tunnel4_mask} dev ${type}_b
936	run_cmd ${ns_b} ip addr add ${tunnel6_b_addr}/${tunnel6_mask} dev ${type}_b
937
938	run_cmd ip link set ${type}_a up
939	run_cmd ${ns_b} ip link set ${type}_b up
940}
941
942setup_ovs_geneve4() {
943	setup_ovs_vxlan_or_geneve geneve ${prefix4}.${a_r1}.1  ${prefix4}.${b_r1}.1
944}
945
946setup_ovs_vxlan4() {
947	setup_ovs_vxlan_or_geneve vxlan  ${prefix4}.${a_r1}.1  ${prefix4}.${b_r1}.1
948}
949
950setup_ovs_geneve6() {
951	setup_ovs_vxlan_or_geneve geneve ${prefix6}:${a_r1}::1 ${prefix6}:${b_r1}::1
952}
953
954setup_ovs_vxlan6() {
955	setup_ovs_vxlan_or_geneve vxlan  ${prefix6}:${a_r1}::1 ${prefix6}:${b_r1}::1
956}
957
958setup_ovs_br_internal() {
959	run_cmd python3 ./openvswitch/ovs-dpctl.py add-dp ovs_br0 || \
960		return 1
961}
962
963setup_ovs_br_vswitchd() {
964	run_cmd ovs-vsctl add-br ovs_br0 || return 1
965}
966
967setup_ovs_add_if() {
968	ifname="${1}"
969	run_cmd python3 ./openvswitch/ovs-dpctl.py add-if ovs_br0 \
970		"${ifname}" || \
971		run_cmd ovs-vsctl add-port ovs_br0 "${ifname}"
972}
973
974setup_ovs_bridge() {
975	setup_ovs_br_internal || setup_ovs_br_vswitchd || return $ksft_skip
976	run_cmd ip link set ovs_br0 up
977
978	run_cmd ${ns_c} ip link add veth_C-A type veth peer name veth_A-C
979	run_cmd ${ns_c} ip link set veth_A-C netns 1
980
981	run_cmd         ip link set veth_A-C up
982	run_cmd ${ns_c} ip link set veth_C-A up
983	run_cmd ${ns_c} ip addr add ${veth4_c_addr}/${veth4_mask} dev veth_C-A
984	run_cmd ${ns_c} ip addr add ${veth6_c_addr}/${veth6_mask} dev veth_C-A
985	setup_ovs_add_if veth_A-C
986
987	# Move veth_A-R1 to init
988	run_cmd ${ns_a} ip link set veth_A-R1 netns 1
989	run_cmd ip addr add ${prefix4}.${a_r1}.1/${veth4_mask} dev veth_A-R1
990	run_cmd ip addr add ${prefix6}:${a_r1}::1/${veth6_mask} dev veth_A-R1
991	run_cmd ip link set veth_A-R1 up
992	run_cmd ip route add ${prefix4}.${b_r1}.1 via ${prefix4}.${a_r1}.2
993	run_cmd ip route add ${prefix6}:${b_r1}::1 via ${prefix6}:${a_r1}::2
994}
995
996setup_multipath_new() {
997	# Set up host A with multipath routes to host B host4_b_addr
998	run_cmd ${ns_a} ip addr add ${host4_a_addr} dev lo
999	run_cmd ${ns_a} ip nexthop add id 401 via ${prefix4}.${a_r1}.2 dev veth_A-R1
1000	run_cmd ${ns_a} ip nexthop add id 402 via ${prefix4}.${a_r2}.2 dev veth_A-R2
1001	run_cmd ${ns_a} ip nexthop add id 403 group 401/402
1002	run_cmd ${ns_a} ip route add ${host4_b_addr} src ${host4_a_addr} nhid 403
1003
1004	# Set up host B with multipath routes to host A host4_a_addr
1005	run_cmd ${ns_b} ip addr add ${host4_b_addr} dev lo
1006	run_cmd ${ns_b} ip nexthop add id 401 via ${prefix4}.${b_r1}.2 dev veth_B-R1
1007	run_cmd ${ns_b} ip nexthop add id 402 via ${prefix4}.${b_r2}.2 dev veth_B-R2
1008	run_cmd ${ns_b} ip nexthop add id 403 group 401/402
1009	run_cmd ${ns_b} ip route add ${host4_a_addr} src ${host4_b_addr} nhid 403
1010}
1011
1012setup_multipath_old() {
1013	# Set up host A with multipath routes to host B host4_b_addr
1014	run_cmd ${ns_a} ip addr add ${host4_a_addr} dev lo
1015	run_cmd ${ns_a} ip route add ${host4_b_addr} \
1016			src ${host4_a_addr} \
1017			nexthop via ${prefix4}.${a_r1}.2 weight 1 \
1018			nexthop via ${prefix4}.${a_r2}.2 weight 1
1019
1020	# Set up host B with multipath routes to host A host4_a_addr
1021	run_cmd ${ns_b} ip addr add ${host4_b_addr} dev lo
1022	run_cmd ${ns_b} ip route add ${host4_a_addr} \
1023			src ${host4_b_addr} \
1024			nexthop via ${prefix4}.${b_r1}.2 weight 1 \
1025			nexthop via ${prefix4}.${b_r2}.2 weight 1
1026}
1027
1028setup_multipath() {
1029	if [ "$USE_NH" = "yes" ]; then
1030		setup_multipath_new
1031	else
1032		setup_multipath_old
1033	fi
1034
1035	# Set up routers with routes to dummies
1036	run_cmd ${ns_r1} ip route add ${host4_a_addr} via ${prefix4}.${a_r1}.1
1037	run_cmd ${ns_r2} ip route add ${host4_a_addr} via ${prefix4}.${a_r2}.1
1038	run_cmd ${ns_r1} ip route add ${host4_b_addr} via ${prefix4}.${b_r1}.1
1039	run_cmd ${ns_r2} ip route add ${host4_b_addr} via ${prefix4}.${b_r2}.1
1040}
1041
1042setup() {
1043	[ "$(id -u)" -ne 0 ] && echo "  need to run as root" && return $ksft_skip
1044
1045	for arg do
1046		eval setup_${arg} || { echo "  ${arg} not supported"; return 1; }
1047	done
1048}
1049
1050trace() {
1051	[ $TRACING -eq 0 ] && return
1052
1053	for arg do
1054		[ "${ns_cmd}" = "" ] && ns_cmd="${arg}" && continue
1055		${ns_cmd} tcpdump --immediate-mode -s 0 -i "${arg}" -w "${name}_${arg}.pcap" 2> /dev/null &
1056		tcpdump_pids="${tcpdump_pids} $!"
1057		ns_cmd=
1058	done
1059	sleep 1
1060}
1061
1062cleanup_del_ovs_internal() {
1063	# squelch the output of the del-if commands since it can be wordy
1064	python3 ./openvswitch/ovs-dpctl.py del-if ovs_br0 -d true vxlan_a	>/dev/null	2>&1
1065	python3 ./openvswitch/ovs-dpctl.py del-if ovs_br0 -d true geneve_a	>/dev/null	2>&1
1066	python3 ./openvswitch/ovs-dpctl.py del-dp ovs_br0			>/dev/null	2>&1
1067}
1068
1069cleanup_del_ovs_vswitchd() {
1070	ovs-vsctl --if-exists del-port vxlan_a	2>/dev/null
1071	ovs-vsctl --if-exists del-br ovs_br0	2>/dev/null
1072}
1073
1074cleanup() {
1075	for pid in ${tcpdump_pids}; do
1076		kill ${pid}
1077	done
1078	tcpdump_pids=
1079
1080	for pid in ${nettest_pids}; do
1081		kill ${pid}
1082	done
1083	nettest_pids=
1084
1085	for pid in ${socat_pids}; do
1086		kill "${pid}"
1087	done
1088	socat_pids=
1089
1090	cleanup_all_ns
1091
1092	ip link del veth_A-C		2>/dev/null
1093	ip link del veth_A-R1		2>/dev/null
1094	cleanup_del_ovs_internal
1095	cleanup_del_ovs_vswitchd
1096	rm -f "$tmpoutfile"
1097}
1098
1099mtu() {
1100	ns_cmd="${1}"
1101	dev="${2}"
1102	mtu="${3}"
1103
1104	${ns_cmd} ip link set dev ${dev} mtu ${mtu}
1105}
1106
1107mtu_parse() {
1108	input="${1}"
1109
1110	next=0
1111	for i in ${input}; do
1112		[ ${next} -eq 1 -a "${i}" = "lock" ] && next=2 && continue
1113		[ ${next} -eq 1 ] && echo "${i}" && return
1114		[ ${next} -eq 2 ] && echo "lock ${i}" && return
1115		[ "${i}" = "mtu" ] && next=1
1116	done
1117}
1118
1119link_get() {
1120	ns_cmd="${1}"
1121	name="${2}"
1122
1123	${ns_cmd} ip link show dev "${name}"
1124}
1125
1126link_get_mtu() {
1127	ns_cmd="${1}"
1128	name="${2}"
1129
1130	mtu_parse "$(link_get "${ns_cmd}" ${name})"
1131}
1132
1133route_get_dst_exception() {
1134	ns_cmd="${1}"; shift
1135
1136	${ns_cmd} ip route get "$@"
1137}
1138
1139route_get_dst_pmtu_from_exception() {
1140	ns_cmd="${1}"; shift
1141
1142	mtu_parse "$(route_get_dst_exception "${ns_cmd}" "$@")"
1143}
1144
1145check_pmtu_value() {
1146	expected="${1}"
1147	value="${2}"
1148	event="${3}"
1149
1150	[ "${expected}" = "any" ] && [ -n "${value}" ] && return 0
1151	[ "${value}" = "${expected}" ] && return 0
1152	[ -z "${value}" ] &&    err "  PMTU exception wasn't created after ${event}" && return 1
1153	[ -z "${expected}" ] && err "  PMTU exception shouldn't exist after ${event}" && return 1
1154	err "  found PMTU exception with incorrect MTU ${value}, expected ${expected}, after ${event}"
1155	return 1
1156}
1157
1158test_pmtu_ipvX() {
1159	family=${1}
1160
1161	setup namespaces routing || return $ksft_skip
1162	trace "${ns_a}"  veth_A-R1    "${ns_r1}" veth_R1-A \
1163	      "${ns_r1}" veth_R1-B    "${ns_b}"  veth_B-R1 \
1164	      "${ns_a}"  veth_A-R2    "${ns_r2}" veth_R2-A \
1165	      "${ns_r2}" veth_R2-B    "${ns_b}"  veth_B-R2
1166
1167	if [ ${family} -eq 4 ]; then
1168		ping=ping
1169		dst1="${prefix4}.${b_r1}.1"
1170		dst2="${prefix4}.${b_r2}.1"
1171	else
1172		ping=${ping6}
1173		dst1="${prefix6}:${b_r1}::1"
1174		dst2="${prefix6}:${b_r2}::1"
1175	fi
1176
1177	# Set up initial MTU values
1178	mtu "${ns_a}"  veth_A-R1 2000
1179	mtu "${ns_r1}" veth_R1-A 2000
1180	mtu "${ns_r1}" veth_R1-B 1400
1181	mtu "${ns_b}"  veth_B-R1 1400
1182
1183	mtu "${ns_a}"  veth_A-R2 2000
1184	mtu "${ns_r2}" veth_R2-A 2000
1185	mtu "${ns_r2}" veth_R2-B 1500
1186	mtu "${ns_b}"  veth_B-R2 1500
1187
1188	# Create route exceptions
1189	run_cmd ${ns_a} ${ping} -q -M want -i 0.1 -w 1 -s 1800 ${dst1}
1190	run_cmd ${ns_a} ${ping} -q -M want -i 0.1 -w 1 -s 1800 ${dst2}
1191
1192	# Check that exceptions have been created with the correct PMTU
1193	pmtu_1="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst1})"
1194	check_pmtu_value "1400" "${pmtu_1}" "exceeding MTU" || return 1
1195	pmtu_2="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst2})"
1196	check_pmtu_value "1500" "${pmtu_2}" "exceeding MTU" || return 1
1197
1198	# Decrease local MTU below PMTU, check for PMTU decrease in route exception
1199	mtu "${ns_a}"  veth_A-R1 1300
1200	mtu "${ns_r1}" veth_R1-A 1300
1201	pmtu_1="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst1})"
1202	check_pmtu_value "1300" "${pmtu_1}" "decreasing local MTU" || return 1
1203	# Second exception shouldn't be modified
1204	pmtu_2="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst2})"
1205	check_pmtu_value "1500" "${pmtu_2}" "changing local MTU on a link not on this path" || return 1
1206
1207	# Increase MTU, check for PMTU increase in route exception
1208	mtu "${ns_a}"  veth_A-R1 1700
1209	mtu "${ns_r1}" veth_R1-A 1700
1210	pmtu_1="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst1})"
1211	check_pmtu_value "1700" "${pmtu_1}" "increasing local MTU" || return 1
1212	# Second exception shouldn't be modified
1213	pmtu_2="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst2})"
1214	check_pmtu_value "1500" "${pmtu_2}" "changing local MTU on a link not on this path" || return 1
1215
1216	# Skip PMTU locking tests for IPv6
1217	[ $family -eq 6 ] && return 0
1218
1219	# Decrease remote MTU on path via R2, get new exception
1220	mtu "${ns_r2}" veth_R2-B 400
1221	mtu "${ns_b}"  veth_B-R2 400
1222	run_cmd ${ns_a} ${ping} -q -M want -i 0.1 -w 1 -s 1400 ${dst2}
1223	pmtu_2="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst2})"
1224	check_pmtu_value "lock 552" "${pmtu_2}" "exceeding MTU, with MTU < min_pmtu" || return 1
1225
1226	# Decrease local MTU below PMTU
1227	mtu "${ns_a}"  veth_A-R2 500
1228	mtu "${ns_r2}" veth_R2-A 500
1229	pmtu_2="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst2})"
1230	check_pmtu_value "500" "${pmtu_2}" "decreasing local MTU" || return 1
1231
1232	# Increase local MTU
1233	mtu "${ns_a}"  veth_A-R2 1500
1234	mtu "${ns_r2}" veth_R2-A 1500
1235	pmtu_2="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst2})"
1236	check_pmtu_value "1500" "${pmtu_2}" "increasing local MTU" || return 1
1237
1238	# Get new exception
1239	run_cmd ${ns_a} ${ping} -q -M want -i 0.1 -w 1 -s 1400 ${dst2}
1240	pmtu_2="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst2})"
1241	check_pmtu_value "lock 552" "${pmtu_2}" "exceeding MTU, with MTU < min_pmtu" || return 1
1242}
1243
1244test_pmtu_ipv4_exception() {
1245	test_pmtu_ipvX 4
1246}
1247
1248test_pmtu_ipv6_exception() {
1249	test_pmtu_ipvX 6
1250}
1251
1252test_pmtu_ipv4_dscp_icmp_exception() {
1253	rt_table=100
1254
1255	setup namespaces policy_routing || return $ksft_skip
1256	trace "${ns_a}"  veth_A-R1    "${ns_r1}" veth_R1-A \
1257	      "${ns_r1}" veth_R1-B    "${ns_b}"  veth_B-R1 \
1258	      "${ns_a}"  veth_A-R2    "${ns_r2}" veth_R2-A \
1259	      "${ns_r2}" veth_R2-B    "${ns_b}"  veth_B-R2
1260
1261	# Set up initial MTU values
1262	mtu "${ns_a}"  veth_A-R1 2000
1263	mtu "${ns_r1}" veth_R1-A 2000
1264	mtu "${ns_r1}" veth_R1-B 1400
1265	mtu "${ns_b}"  veth_B-R1 1400
1266
1267	mtu "${ns_a}"  veth_A-R2 2000
1268	mtu "${ns_r2}" veth_R2-A 2000
1269	mtu "${ns_r2}" veth_R2-B 1500
1270	mtu "${ns_b}"  veth_B-R2 1500
1271
1272	len=$((2000 - 20 - 8)) # Fills MTU of veth_A-R1
1273
1274	dst1="${prefix4}.${b_r1}.1"
1275	dst2="${prefix4}.${b_r2}.1"
1276
1277	# Create route exceptions
1278	dsfield=${policy_mark} # No ECN bit set (Not-ECT)
1279	run_cmd "${ns_a}" ping -q -M want -Q "${dsfield}" -c 1 -w 1 -s "${len}" "${dst1}"
1280
1281	dsfield=$(printf "%#x" $((policy_mark + 0x02))) # ECN=2 (ECT(0))
1282	run_cmd "${ns_a}" ping -q -M want -Q "${dsfield}" -c 1 -w 1 -s "${len}" "${dst2}"
1283
1284	# Check that exceptions have been created with the correct PMTU
1285	pmtu_1="$(route_get_dst_pmtu_from_exception "${ns_a}" "${dst1}" dsfield "${policy_mark}")"
1286	check_pmtu_value "1400" "${pmtu_1}" "exceeding MTU" || return 1
1287
1288	pmtu_2="$(route_get_dst_pmtu_from_exception "${ns_a}" "${dst2}" dsfield "${policy_mark}")"
1289	check_pmtu_value "1500" "${pmtu_2}" "exceeding MTU" || return 1
1290}
1291
1292test_pmtu_ipv4_dscp_udp_exception() {
1293	rt_table=100
1294
1295	if ! which socat > /dev/null 2>&1; then
1296		echo "'socat' command not found; skipping tests"
1297		return $ksft_skip
1298	fi
1299
1300	setup namespaces policy_routing || return $ksft_skip
1301	trace "${ns_a}"  veth_A-R1    "${ns_r1}" veth_R1-A \
1302	      "${ns_r1}" veth_R1-B    "${ns_b}"  veth_B-R1 \
1303	      "${ns_a}"  veth_A-R2    "${ns_r2}" veth_R2-A \
1304	      "${ns_r2}" veth_R2-B    "${ns_b}"  veth_B-R2
1305
1306	# Set up initial MTU values
1307	mtu "${ns_a}"  veth_A-R1 2000
1308	mtu "${ns_r1}" veth_R1-A 2000
1309	mtu "${ns_r1}" veth_R1-B 1400
1310	mtu "${ns_b}"  veth_B-R1 1400
1311
1312	mtu "${ns_a}"  veth_A-R2 2000
1313	mtu "${ns_r2}" veth_R2-A 2000
1314	mtu "${ns_r2}" veth_R2-B 1500
1315	mtu "${ns_b}"  veth_B-R2 1500
1316
1317	len=$((2000 - 20 - 8)) # Fills MTU of veth_A-R1
1318
1319	dst1="${prefix4}.${b_r1}.1"
1320	dst2="${prefix4}.${b_r2}.1"
1321
1322	# Create route exceptions
1323	run_cmd_bg "${ns_b}" socat UDP-LISTEN:50000 OPEN:/dev/null,wronly=1
1324	socat_pids="${socat_pids} $!"
1325
1326	dsfield=${policy_mark} # No ECN bit set (Not-ECT)
1327	run_cmd "${ns_a}" socat OPEN:/dev/zero,rdonly=1,readbytes="${len}" \
1328		UDP:"${dst1}":50000,tos="${dsfield}"
1329
1330	dsfield=$(printf "%#x" $((policy_mark + 0x02))) # ECN=2 (ECT(0))
1331	run_cmd "${ns_a}" socat OPEN:/dev/zero,rdonly=1,readbytes="${len}" \
1332		UDP:"${dst2}":50000,tos="${dsfield}"
1333
1334	# Check that exceptions have been created with the correct PMTU
1335	pmtu_1="$(route_get_dst_pmtu_from_exception "${ns_a}" "${dst1}" dsfield "${policy_mark}")"
1336	check_pmtu_value "1400" "${pmtu_1}" "exceeding MTU" || return 1
1337	pmtu_2="$(route_get_dst_pmtu_from_exception "${ns_a}" "${dst2}" dsfield "${policy_mark}")"
1338	check_pmtu_value "1500" "${pmtu_2}" "exceeding MTU" || return 1
1339}
1340
1341test_pmtu_ipvX_over_vxlanY_or_geneveY_exception() {
1342	type=${1}
1343	family=${2}
1344	outer_family=${3}
1345	ll_mtu=4000
1346
1347	if [ ${outer_family} -eq 4 ]; then
1348		setup namespaces routing ${type}4 || return $ksft_skip
1349		#                      IPv4 header   UDP header   VXLAN/GENEVE header   Ethernet header
1350		exp_mtu=$((${ll_mtu} - 20          - 8          - 8                   - 14))
1351	else
1352		setup namespaces routing ${type}6 || return $ksft_skip
1353		#                      IPv6 header   UDP header   VXLAN/GENEVE header   Ethernet header
1354		exp_mtu=$((${ll_mtu} - 40          - 8          - 8                   - 14))
1355	fi
1356
1357	trace "${ns_a}" ${type}_a    "${ns_b}"  ${type}_b \
1358	      "${ns_a}" veth_A-R1    "${ns_r1}" veth_R1-A \
1359	      "${ns_b}" veth_B-R1    "${ns_r1}" veth_R1-B
1360
1361	if [ ${family} -eq 4 ]; then
1362		ping=ping
1363		dst=${tunnel4_b_addr}
1364	else
1365		ping=${ping6}
1366		dst=${tunnel6_b_addr}
1367	fi
1368
1369	# Create route exception by exceeding link layer MTU
1370	mtu "${ns_a}"  veth_A-R1 $((${ll_mtu} + 1000))
1371	mtu "${ns_r1}" veth_R1-A $((${ll_mtu} + 1000))
1372	mtu "${ns_b}"  veth_B-R1 ${ll_mtu}
1373	mtu "${ns_r1}" veth_R1-B ${ll_mtu}
1374
1375	mtu "${ns_a}" ${type}_a $((${ll_mtu} + 1000))
1376	mtu "${ns_b}" ${type}_b $((${ll_mtu} + 1000))
1377	run_cmd ${ns_a} ${ping} -q -M want -i 0.1 -w 1 -s $((${ll_mtu} + 500)) ${dst}
1378
1379	# Check that exception was created
1380	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst})"
1381	check_pmtu_value ${exp_mtu} "${pmtu}" "exceeding link layer MTU on ${type} interface"
1382}
1383
1384test_pmtu_ipv4_vxlan4_exception() {
1385	test_pmtu_ipvX_over_vxlanY_or_geneveY_exception vxlan  4 4
1386}
1387
1388test_pmtu_ipv6_vxlan4_exception() {
1389	test_pmtu_ipvX_over_vxlanY_or_geneveY_exception vxlan  6 4
1390}
1391
1392test_pmtu_ipv4_geneve4_exception() {
1393	test_pmtu_ipvX_over_vxlanY_or_geneveY_exception geneve 4 4
1394}
1395
1396test_pmtu_ipv6_geneve4_exception() {
1397	test_pmtu_ipvX_over_vxlanY_or_geneveY_exception geneve 6 4
1398}
1399
1400test_pmtu_ipv4_vxlan6_exception() {
1401	test_pmtu_ipvX_over_vxlanY_or_geneveY_exception vxlan  4 6
1402}
1403
1404test_pmtu_ipv6_vxlan6_exception() {
1405	test_pmtu_ipvX_over_vxlanY_or_geneveY_exception vxlan  6 6
1406}
1407
1408test_pmtu_ipv4_geneve6_exception() {
1409	test_pmtu_ipvX_over_vxlanY_or_geneveY_exception geneve 4 6
1410}
1411
1412test_pmtu_ipv6_geneve6_exception() {
1413	test_pmtu_ipvX_over_vxlanY_or_geneveY_exception geneve 6 6
1414}
1415
1416test_pmtu_ipvX_over_bridged_vxlanY_or_geneveY_exception() {
1417	type=${1}
1418	family=${2}
1419	outer_family=${3}
1420	ll_mtu=4000
1421
1422	if [ ${outer_family} -eq 4 ]; then
1423		setup namespaces routing bridge bridged_${type}4 || return $ksft_skip
1424		#                      IPv4 header   UDP header   VXLAN/GENEVE header   Ethernet header
1425		exp_mtu=$((${ll_mtu} - 20          - 8          - 8                   - 14))
1426	else
1427		setup namespaces routing bridge bridged_${type}6 || return $ksft_skip
1428		#                      IPv6 header   UDP header   VXLAN/GENEVE header   Ethernet header
1429		exp_mtu=$((${ll_mtu} - 40          - 8          - 8                   - 14))
1430	fi
1431
1432	trace "${ns_a}" ${type}_a    "${ns_b}"  ${type}_b \
1433	      "${ns_a}" veth_A-R1    "${ns_r1}" veth_R1-A \
1434	      "${ns_b}" veth_B-R1    "${ns_r1}" veth_R1-B \
1435	      "${ns_a}" br0          "${ns_a}"  veth-A-C  \
1436	      "${ns_c}" veth_C-A
1437
1438	if [ ${family} -eq 4 ]; then
1439		ping=ping
1440		dst=${tunnel4_b_addr}
1441	else
1442		ping=${ping6}
1443		dst=${tunnel6_b_addr}
1444	fi
1445
1446	# Create route exception by exceeding link layer MTU
1447	mtu "${ns_a}"  veth_A-R1 $((${ll_mtu} + 1000))
1448	mtu "${ns_a}"  br0       $((${ll_mtu} + 1000))
1449	mtu "${ns_a}"  veth_A-C  $((${ll_mtu} + 1000))
1450	mtu "${ns_c}"  veth_C-A  $((${ll_mtu} + 1000))
1451	mtu "${ns_r1}" veth_R1-A $((${ll_mtu} + 1000))
1452	mtu "${ns_b}"  veth_B-R1 ${ll_mtu}
1453	mtu "${ns_r1}" veth_R1-B ${ll_mtu}
1454
1455	mtu "${ns_a}" ${type}_a $((${ll_mtu} + 1000))
1456	mtu "${ns_b}" ${type}_b $((${ll_mtu} + 1000))
1457
1458	run_cmd ${ns_c} ${ping} -q -M want -i 0.1 -c 10 -s $((${ll_mtu} + 500)) ${dst} || return 1
1459	run_cmd ${ns_a} ${ping} -q -M want -i 0.1 -w 1  -s $((${ll_mtu} + 500)) ${dst} || return 1
1460
1461	# Check that exceptions were created
1462	pmtu="$(route_get_dst_pmtu_from_exception "${ns_c}" ${dst})"
1463	check_pmtu_value ${exp_mtu} "${pmtu}" "exceeding link layer MTU on bridged ${type} interface"
1464	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst})"
1465	check_pmtu_value ${exp_mtu} "${pmtu}" "exceeding link layer MTU on locally bridged ${type} interface"
1466
1467	tmpoutfile=$(mktemp)
1468
1469	# Flush Exceptions, retry with TCP
1470	run_cmd ${ns_a} ip route flush cached ${dst}
1471	run_cmd ${ns_b} ip route flush cached ${dst}
1472	run_cmd ${ns_c} ip route flush cached ${dst}
1473
1474	for target in "${ns_a}" "${ns_c}" ; do
1475		if [ ${family} -eq 4 ]; then
1476			TCPDST=TCP:${dst}:50000
1477		else
1478			TCPDST="TCP:[${dst}]:50000"
1479		fi
1480		${ns_b} socat -T 3 -u -6 TCP-LISTEN:50000,reuseaddr STDOUT > $tmpoutfile &
1481		local socat_pid=$!
1482
1483		wait_local_port_listen ${NS_B} 50000 tcp
1484
1485		dd if=/dev/zero status=none bs=1M count=1 | ${target} socat -T 3 -u STDIN $TCPDST,connect-timeout=3
1486
1487		wait ${socat_pid}
1488		size=$(du -sb $tmpoutfile)
1489		size=${size%%/tmp/*}
1490
1491		[ $size -ne 1048576 ] && err "File size $size mismatches expected value in locally bridged vxlan test" && return 1
1492	done
1493
1494	rm -f "$tmpoutfile"
1495
1496	# Check that exceptions were created
1497	pmtu="$(route_get_dst_pmtu_from_exception "${ns_c}" ${dst})"
1498	check_pmtu_value ${exp_mtu} "${pmtu}" "tcp: exceeding link layer MTU on bridged ${type} interface"
1499	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst})"
1500	check_pmtu_value ${exp_mtu} "${pmtu}" "tcp exceeding link layer MTU on locally bridged ${type} interface"
1501}
1502
1503test_pmtu_ipv4_br_vxlan4_exception() {
1504	test_pmtu_ipvX_over_bridged_vxlanY_or_geneveY_exception vxlan  4 4
1505}
1506
1507test_pmtu_ipv6_br_vxlan4_exception() {
1508	test_pmtu_ipvX_over_bridged_vxlanY_or_geneveY_exception vxlan  6 4
1509}
1510
1511test_pmtu_ipv4_br_geneve4_exception() {
1512	test_pmtu_ipvX_over_bridged_vxlanY_or_geneveY_exception geneve 4 4
1513}
1514
1515test_pmtu_ipv6_br_geneve4_exception() {
1516	test_pmtu_ipvX_over_bridged_vxlanY_or_geneveY_exception geneve 6 4
1517}
1518
1519test_pmtu_ipv4_br_vxlan6_exception() {
1520	test_pmtu_ipvX_over_bridged_vxlanY_or_geneveY_exception vxlan  4 6
1521}
1522
1523test_pmtu_ipv6_br_vxlan6_exception() {
1524	test_pmtu_ipvX_over_bridged_vxlanY_or_geneveY_exception vxlan  6 6
1525}
1526
1527test_pmtu_ipv4_br_geneve6_exception() {
1528	test_pmtu_ipvX_over_bridged_vxlanY_or_geneveY_exception geneve 4 6
1529}
1530
1531test_pmtu_ipv6_br_geneve6_exception() {
1532	test_pmtu_ipvX_over_bridged_vxlanY_or_geneveY_exception geneve 6 6
1533}
1534
1535test_pmtu_ipvX_over_ovs_vxlanY_or_geneveY_exception() {
1536	type=${1}
1537	family=${2}
1538	outer_family=${3}
1539	ll_mtu=4000
1540
1541	if [ "${type}" = "vxlan" ]; then
1542		tun_a="vxlan_sys_4789"
1543	elif [ "${type}" = "geneve" ]; then
1544		tun_a="genev_sys_6081"
1545	fi
1546
1547	if [ ${outer_family} -eq 4 ]; then
1548		setup namespaces routing ovs_bridge ovs_${type}4 || return $ksft_skip
1549		#                      IPv4 header   UDP header   VXLAN/GENEVE header   Ethernet header
1550		exp_mtu=$((${ll_mtu} - 20          - 8          - 8                   - 14))
1551	else
1552		setup namespaces routing ovs_bridge ovs_${type}6 || return $ksft_skip
1553		#                      IPv6 header   UDP header   VXLAN/GENEVE header   Ethernet header
1554		exp_mtu=$((${ll_mtu} - 40          - 8          - 8                   - 14))
1555	fi
1556
1557	trace ""        ${type}_a    "${ns_b}"  ${type}_b \
1558	      ""        veth_A-R1    "${ns_r1}" veth_R1-A \
1559	      "${ns_b}" veth_B-R1    "${ns_r1}" veth_R1-B \
1560	      ""        ovs_br0      ""         veth-A_C  \
1561	      "${ns_c}" veth_C-A     ""         "${tun_a}"
1562
1563	if [ ${family} -eq 4 ]; then
1564		ping=ping
1565		dst=${tunnel4_b_addr}
1566	else
1567		ping=${ping6}
1568		dst=${tunnel6_b_addr}
1569	fi
1570
1571	# Create route exception by exceeding link layer MTU
1572	mtu ""         veth_A-R1 $((${ll_mtu} + 1000))
1573	mtu ""         ovs_br0   $((${ll_mtu} + 1000))
1574	mtu ""         veth_A-C  $((${ll_mtu} + 1000))
1575	mtu "${ns_c}"  veth_C-A  $((${ll_mtu} + 1000))
1576	mtu "${ns_r1}" veth_R1-A $((${ll_mtu} + 1000))
1577	mtu "${ns_b}"  veth_B-R1 ${ll_mtu}
1578	mtu "${ns_r1}" veth_R1-B ${ll_mtu}
1579
1580	mtu ""        ${tun_a}  $((${ll_mtu} + 1000)) 2>/dev/null || \
1581		mtu ""        ${type}_a  $((${ll_mtu} + 1000)) 2>/dev/null
1582	mtu "${ns_b}" ${type}_b  $((${ll_mtu} + 1000))
1583
1584	run_cmd ${ns_c} ${ping} -q -M want -i 0.1 -c 20 -s $((${ll_mtu} + 500)) ${dst} || return 1
1585
1586	# Check that exceptions were created
1587	pmtu="$(route_get_dst_pmtu_from_exception "${ns_c}" ${dst})"
1588	check_pmtu_value ${exp_mtu} "${pmtu}" "exceeding link layer MTU on Open vSwitch ${type} interface"
1589}
1590
1591test_pmtu_ipv4_ovs_vxlan4_exception() {
1592	test_pmtu_ipvX_over_ovs_vxlanY_or_geneveY_exception vxlan  4 4
1593}
1594
1595test_pmtu_ipv6_ovs_vxlan4_exception() {
1596	test_pmtu_ipvX_over_ovs_vxlanY_or_geneveY_exception vxlan  6 4
1597}
1598
1599test_pmtu_ipv4_ovs_geneve4_exception() {
1600	test_pmtu_ipvX_over_ovs_vxlanY_or_geneveY_exception geneve 4 4
1601}
1602
1603test_pmtu_ipv6_ovs_geneve4_exception() {
1604	test_pmtu_ipvX_over_ovs_vxlanY_or_geneveY_exception geneve 6 4
1605}
1606
1607test_pmtu_ipv4_ovs_vxlan6_exception() {
1608	test_pmtu_ipvX_over_ovs_vxlanY_or_geneveY_exception vxlan  4 6
1609}
1610
1611test_pmtu_ipv6_ovs_vxlan6_exception() {
1612	test_pmtu_ipvX_over_ovs_vxlanY_or_geneveY_exception vxlan  6 6
1613}
1614
1615test_pmtu_ipv4_ovs_geneve6_exception() {
1616	test_pmtu_ipvX_over_ovs_vxlanY_or_geneveY_exception geneve 4 6
1617}
1618
1619test_pmtu_ipv6_ovs_geneve6_exception() {
1620	test_pmtu_ipvX_over_ovs_vxlanY_or_geneveY_exception geneve 6 6
1621}
1622
1623test_pmtu_ipvX_over_fouY_or_gueY() {
1624	inner_family=${1}
1625	outer_family=${2}
1626	encap=${3}
1627	ll_mtu=4000
1628
1629	setup namespaces routing ${encap}${outer_family}${inner_family} || return $ksft_skip
1630	trace "${ns_a}" ${encap}_a   "${ns_b}"  ${encap}_b \
1631	      "${ns_a}" veth_A-R1    "${ns_r1}" veth_R1-A \
1632	      "${ns_b}" veth_B-R1    "${ns_r1}" veth_R1-B
1633
1634	if [ ${inner_family} -eq 4 ]; then
1635		ping=ping
1636		dst=${tunnel4_b_addr}
1637	else
1638		ping=${ping6}
1639		dst=${tunnel6_b_addr}
1640	fi
1641
1642	if [ "${encap}" = "gue" ]; then
1643		encap_overhead=4
1644	else
1645		encap_overhead=0
1646	fi
1647
1648	if [ ${outer_family} -eq 4 ]; then
1649		#                      IPv4 header   UDP header
1650		exp_mtu=$((${ll_mtu} - 20          - 8         - ${encap_overhead}))
1651	else
1652		#                      IPv6 header   Option 4   UDP header
1653		exp_mtu=$((${ll_mtu} - 40          - 8        - 8       - ${encap_overhead}))
1654	fi
1655
1656	# Create route exception by exceeding link layer MTU
1657	mtu "${ns_a}"  veth_A-R1 $((${ll_mtu} + 1000))
1658	mtu "${ns_r1}" veth_R1-A $((${ll_mtu} + 1000))
1659	mtu "${ns_b}"  veth_B-R1 ${ll_mtu}
1660	mtu "${ns_r1}" veth_R1-B ${ll_mtu}
1661
1662	mtu "${ns_a}" ${encap}_a $((${ll_mtu} + 1000))
1663	mtu "${ns_b}" ${encap}_b $((${ll_mtu} + 1000))
1664	run_cmd ${ns_a} ${ping} -q -M want -i 0.1 -w 1 -s $((${ll_mtu} + 500)) ${dst}
1665
1666	# Check that exception was created
1667	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst})"
1668	check_pmtu_value ${exp_mtu} "${pmtu}" "exceeding link layer MTU on ${encap} interface"
1669}
1670
1671test_pmtu_ipv4_fou4_exception() {
1672	test_pmtu_ipvX_over_fouY_or_gueY 4 4 fou
1673}
1674
1675test_pmtu_ipv6_fou4_exception() {
1676	test_pmtu_ipvX_over_fouY_or_gueY 6 4 fou
1677}
1678
1679test_pmtu_ipv4_fou6_exception() {
1680	test_pmtu_ipvX_over_fouY_or_gueY 4 6 fou
1681}
1682
1683test_pmtu_ipv6_fou6_exception() {
1684	test_pmtu_ipvX_over_fouY_or_gueY 6 6 fou
1685}
1686
1687test_pmtu_ipv4_gue4_exception() {
1688	test_pmtu_ipvX_over_fouY_or_gueY 4 4 gue
1689}
1690
1691test_pmtu_ipv6_gue4_exception() {
1692	test_pmtu_ipvX_over_fouY_or_gueY 6 4 gue
1693}
1694
1695test_pmtu_ipv4_gue6_exception() {
1696	test_pmtu_ipvX_over_fouY_or_gueY 4 6 gue
1697}
1698
1699test_pmtu_ipv6_gue6_exception() {
1700	test_pmtu_ipvX_over_fouY_or_gueY 6 6 gue
1701}
1702
1703test_pmtu_ipvX_over_ipvY_exception() {
1704	inner=${1}
1705	outer=${2}
1706	ll_mtu=4000
1707
1708	setup namespaces routing ip${inner}ip${outer} || return $ksft_skip
1709
1710	trace "${ns_a}" ip_a         "${ns_b}"  ip_b  \
1711	      "${ns_a}" veth_A-R1    "${ns_r1}" veth_R1-A \
1712	      "${ns_b}" veth_B-R1    "${ns_r1}" veth_R1-B
1713
1714	if [ ${inner} -eq 4 ]; then
1715		ping=ping
1716		dst=${tunnel4_b_addr}
1717	else
1718		ping=${ping6}
1719		dst=${tunnel6_b_addr}
1720	fi
1721
1722	if [ ${outer} -eq 4 ]; then
1723		#                      IPv4 header
1724		exp_mtu=$((${ll_mtu} - 20))
1725	else
1726		#                      IPv6 header   Option 4
1727		exp_mtu=$((${ll_mtu} - 40          - 8))
1728	fi
1729
1730	# Create route exception by exceeding link layer MTU
1731	mtu "${ns_a}"  veth_A-R1 $((${ll_mtu} + 1000))
1732	mtu "${ns_r1}" veth_R1-A $((${ll_mtu} + 1000))
1733	mtu "${ns_b}"  veth_B-R1 ${ll_mtu}
1734	mtu "${ns_r1}" veth_R1-B ${ll_mtu}
1735
1736	mtu "${ns_a}" ip_a $((${ll_mtu} + 1000)) || return
1737	mtu "${ns_b}" ip_b $((${ll_mtu} + 1000)) || return
1738	run_cmd ${ns_a} ${ping} -q -M want -i 0.1 -w 1 -s $((${ll_mtu} + 500)) ${dst}
1739
1740	# Check that exception was created
1741	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst})"
1742	check_pmtu_value ${exp_mtu} "${pmtu}" "exceeding link layer MTU on ip${inner}ip${outer} interface"
1743}
1744
1745test_pmtu_ipv4_ipv4_exception() {
1746	test_pmtu_ipvX_over_ipvY_exception 4 4
1747}
1748
1749test_pmtu_ipv6_ipv4_exception() {
1750	test_pmtu_ipvX_over_ipvY_exception 6 4
1751}
1752
1753test_pmtu_ipv4_ipv6_exception() {
1754	test_pmtu_ipvX_over_ipvY_exception 4 6
1755}
1756
1757test_pmtu_ipv6_ipv6_exception() {
1758	test_pmtu_ipvX_over_ipvY_exception 6 6
1759}
1760
1761test_pmtu_vti4_exception() {
1762	setup namespaces veth vti4 xfrm4 || return $ksft_skip
1763	trace "${ns_a}" veth_a    "${ns_b}" veth_b \
1764	      "${ns_a}" vti4_a    "${ns_b}" vti4_b
1765
1766	veth_mtu=1500
1767	vti_mtu=$((veth_mtu - 20))
1768
1769	#                                SPI   SN   IV  ICV   pad length   next header
1770	esp_payload_rfc4106=$((vti_mtu - 4   - 4  - 8 - 16  - 1          - 1))
1771	ping_payload=$((esp_payload_rfc4106 - 28))
1772
1773	mtu "${ns_a}" veth_a ${veth_mtu}
1774	mtu "${ns_b}" veth_b ${veth_mtu}
1775	mtu "${ns_a}" vti4_a ${vti_mtu}
1776	mtu "${ns_b}" vti4_b ${vti_mtu}
1777
1778	# Send DF packet without exceeding link layer MTU, check that no
1779	# exception is created
1780	run_cmd ${ns_a} ping -q -M want -i 0.1 -w 1 -s ${ping_payload} ${tunnel4_b_addr}
1781	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel4_b_addr})"
1782	check_pmtu_value "" "${pmtu}" "sending packet smaller than PMTU (IP payload length ${esp_payload_rfc4106})" || return 1
1783
1784	# Now exceed link layer MTU by one byte, check that exception is created
1785	# with the right PMTU value
1786	run_cmd ${ns_a} ping -q -M want -i 0.1 -w 1 -s $((ping_payload + 1)) ${tunnel4_b_addr}
1787	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel4_b_addr})"
1788	check_pmtu_value "${esp_payload_rfc4106}" "${pmtu}" "exceeding PMTU (IP payload length $((esp_payload_rfc4106 + 1)))"
1789}
1790
1791test_pmtu_vti6_exception() {
1792	setup namespaces veth vti6 xfrm6 || return $ksft_skip
1793	trace "${ns_a}" veth_a    "${ns_b}" veth_b \
1794	      "${ns_a}" vti6_a    "${ns_b}" vti6_b
1795	fail=0
1796
1797	# Create route exception by exceeding link layer MTU
1798	mtu "${ns_a}" veth_a 4000
1799	mtu "${ns_b}" veth_b 4000
1800	mtu "${ns_a}" vti6_a 5000
1801	mtu "${ns_b}" vti6_b 5000
1802	run_cmd ${ns_a} ${ping6} -q -i 0.1 -w 1 -s 60000 ${tunnel6_b_addr}
1803
1804	# Check that exception was created
1805	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel6_b_addr})"
1806	check_pmtu_value any "${pmtu}" "creating tunnel exceeding link layer MTU" || return 1
1807
1808	# Decrease tunnel MTU, check for PMTU decrease in route exception
1809	mtu "${ns_a}" vti6_a 3000
1810	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel6_b_addr})"
1811	check_pmtu_value "3000" "${pmtu}" "decreasing tunnel MTU" || fail=1
1812
1813	# Increase tunnel MTU, check for PMTU increase in route exception
1814	mtu "${ns_a}" vti6_a 9000
1815	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel6_b_addr})"
1816	check_pmtu_value "9000" "${pmtu}" "increasing tunnel MTU" || fail=1
1817
1818	return ${fail}
1819}
1820
1821test_pmtu_vti4_udp_exception() {
1822	setup namespaces veth vti4 xfrm4udp || return $ksft_skip
1823	trace "${ns_a}" veth_a    "${ns_b}" veth_b \
1824	      "${ns_a}" vti4_a    "${ns_b}" vti4_b
1825
1826	veth_mtu=1500
1827	vti_mtu=$((veth_mtu - 20))
1828
1829	#                                UDP   SPI   SN   IV  ICV   pad length   next header
1830	esp_payload_rfc4106=$((vti_mtu - 8   - 4   - 4  - 8 - 16  - 1          - 1))
1831	ping_payload=$((esp_payload_rfc4106 - 28))
1832
1833	mtu "${ns_a}" veth_a ${veth_mtu}
1834	mtu "${ns_b}" veth_b ${veth_mtu}
1835	mtu "${ns_a}" vti4_a ${vti_mtu}
1836	mtu "${ns_b}" vti4_b ${vti_mtu}
1837
1838	# Send DF packet without exceeding link layer MTU, check that no
1839	# exception is created
1840	run_cmd ${ns_a} ping -q -M want -i 0.1 -w 1 -s ${ping_payload} ${tunnel4_b_addr}
1841	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel4_b_addr})"
1842	check_pmtu_value "" "${pmtu}" "sending packet smaller than PMTU (IP payload length ${esp_payload_rfc4106})" || return 1
1843
1844	# Now exceed link layer MTU by one byte, check that exception is created
1845	# with the right PMTU value
1846	run_cmd ${ns_a} ping -q -M want -i 0.1 -w 1 -s $((ping_payload + 1)) ${tunnel4_b_addr}
1847	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel4_b_addr})"
1848	check_pmtu_value "${esp_payload_rfc4106}" "${pmtu}" "exceeding PMTU (IP payload length $((esp_payload_rfc4106 + 1)))"
1849}
1850
1851test_pmtu_vti6_udp_exception() {
1852	setup namespaces veth vti6 xfrm6udp || return $ksft_skip
1853	trace "${ns_a}" veth_a    "${ns_b}" veth_b \
1854	      "${ns_a}" vti6_a    "${ns_b}" vti6_b
1855	fail=0
1856
1857	# Create route exception by exceeding link layer MTU
1858	mtu "${ns_a}" veth_a 4000
1859	mtu "${ns_b}" veth_b 4000
1860	mtu "${ns_a}" vti6_a 5000
1861	mtu "${ns_b}" vti6_b 5000
1862	run_cmd ${ns_a} ${ping6} -q -i 0.1 -w 1 -s 60000 ${tunnel6_b_addr}
1863
1864	# Check that exception was created
1865	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel6_b_addr})"
1866	check_pmtu_value any "${pmtu}" "creating tunnel exceeding link layer MTU" || return 1
1867
1868	# Decrease tunnel MTU, check for PMTU decrease in route exception
1869	mtu "${ns_a}" vti6_a 3000
1870	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel6_b_addr})"
1871	check_pmtu_value "3000" "${pmtu}" "decreasing tunnel MTU" || fail=1
1872
1873	# Increase tunnel MTU, check for PMTU increase in route exception
1874	mtu "${ns_a}" vti6_a 9000
1875	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel6_b_addr})"
1876	check_pmtu_value "9000" "${pmtu}" "increasing tunnel MTU" || fail=1
1877
1878	return ${fail}
1879}
1880
1881test_pmtu_vti4_udp_routed_exception() {
1882	setup namespaces routing vti4routed xfrm4udprouted || return $ksft_skip
1883	trace "${ns_a}" veth_A-R1    "${ns_b}" veth_B-R1 \
1884	      "${ns_a}" vti4_a       "${ns_b}" vti4_b
1885
1886	veth_mtu=1500
1887	vti_mtu=$((veth_mtu - 20))
1888
1889	#                                UDP   SPI   SN   IV  ICV   pad length   next header
1890	esp_payload_rfc4106=$((vti_mtu - 8   - 4   - 4  - 8 - 16  - 1          - 1))
1891	ping_payload=$((esp_payload_rfc4106 - 28))
1892
1893        mtu "${ns_a}"  veth_A-R1 ${veth_mtu}
1894        mtu "${ns_r1}" veth_R1-A ${veth_mtu}
1895        mtu "${ns_b}"  veth_B-R1 ${veth_mtu}
1896        mtu "${ns_r1}" veth_R1-B ${veth_mtu}
1897
1898	mtu "${ns_a}" vti4_a ${vti_mtu}
1899	mtu "${ns_b}" vti4_b ${vti_mtu}
1900
1901	# Send DF packet without exceeding link layer MTU, check that no
1902	# exception is created
1903	run_cmd ${ns_a} ping -q -M want -i 0.1 -w 1 -s ${ping_payload} ${tunnel4_b_addr}
1904	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel4_b_addr})"
1905	check_pmtu_value "" "${pmtu}" "sending packet smaller than PMTU (IP payload length ${esp_payload_rfc4106})" || return 1
1906
1907	# Now decrease link layer MTU by 8 bytes on R1, check that exception is created
1908	# with the right PMTU value
1909        mtu "${ns_r1}" veth_R1-B $((veth_mtu - 8))
1910	run_cmd ${ns_a} ping -q -M want -i 0.1 -w 1 -s $((ping_payload)) ${tunnel4_b_addr}
1911	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel4_b_addr})"
1912	check_pmtu_value "$((esp_payload_rfc4106 - 8))" "${pmtu}" "exceeding PMTU (IP payload length $((esp_payload_rfc4106)))"
1913}
1914
1915test_pmtu_vti6_udp_routed_exception() {
1916	setup namespaces routing vti6routed xfrm6udprouted || return $ksft_skip
1917	trace "${ns_a}" veth_A-R1    "${ns_b}" veth_B-R1 \
1918	      "${ns_a}" vti6_a       "${ns_b}" vti6_b
1919
1920	veth_mtu=1500
1921	vti_mtu=$((veth_mtu - 40))
1922
1923	#                                UDP   SPI   SN   IV  ICV   pad length   next header
1924	esp_payload_rfc4106=$((vti_mtu - 8   - 4   - 4  - 8 - 16  - 1          - 1))
1925	ping_payload=$((esp_payload_rfc4106 - 48))
1926
1927        mtu "${ns_a}"  veth_A-R1 ${veth_mtu}
1928        mtu "${ns_r1}" veth_R1-A ${veth_mtu}
1929        mtu "${ns_b}"  veth_B-R1 ${veth_mtu}
1930        mtu "${ns_r1}" veth_R1-B ${veth_mtu}
1931
1932	# mtu "${ns_a}" vti6_a ${vti_mtu}
1933	# mtu "${ns_b}" vti6_b ${vti_mtu}
1934
1935	run_cmd ${ns_a} ${ping6} -q -M want -i 0.1 -w 1 -s ${ping_payload} ${tunnel6_b_addr}
1936
1937	# Check that exception was not created
1938	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel6_b_addr})"
1939	check_pmtu_value "" "${pmtu}" "sending packet smaller than PMTU (IP payload length ${esp_payload_rfc4106})" || return 1
1940
1941	# Now decrease link layer MTU by 8 bytes on R1, check that exception is created
1942	# with the right PMTU value
1943        mtu "${ns_r1}" veth_R1-B $((veth_mtu - 8))
1944	run_cmd ${ns_a} ${ping6} -q -M want -i 0.1 -w 1 -s $((ping_payload)) ${tunnel6_b_addr}
1945	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel6_b_addr})"
1946	check_pmtu_value "$((esp_payload_rfc4106 - 8))" "${pmtu}" "exceeding PMTU (IP payload length $((esp_payload_rfc4106)))"
1947
1948}
1949
1950test_pmtu_vti4_default_mtu() {
1951	setup namespaces veth vti4 || return $ksft_skip
1952
1953	# Check that MTU of vti device is MTU of veth minus IPv4 header length
1954	veth_mtu="$(link_get_mtu "${ns_a}" veth_a)"
1955	vti4_mtu="$(link_get_mtu "${ns_a}" vti4_a)"
1956	if [ $((veth_mtu - vti4_mtu)) -ne 20 ]; then
1957		err "  vti MTU ${vti4_mtu} is not veth MTU ${veth_mtu} minus IPv4 header length"
1958		return 1
1959	fi
1960}
1961
1962test_pmtu_vti6_default_mtu() {
1963	setup namespaces veth vti6 || return $ksft_skip
1964
1965	# Check that MTU of vti device is MTU of veth minus IPv6 header length
1966	veth_mtu="$(link_get_mtu "${ns_a}" veth_a)"
1967	vti6_mtu="$(link_get_mtu "${ns_a}" vti6_a)"
1968	if [ $((veth_mtu - vti6_mtu)) -ne 40 ]; then
1969		err "  vti MTU ${vti6_mtu} is not veth MTU ${veth_mtu} minus IPv6 header length"
1970		return 1
1971	fi
1972}
1973
1974test_pmtu_vti4_link_add_mtu() {
1975	setup namespaces || return $ksft_skip
1976
1977	run_cmd ${ns_a} ip link add vti4_a type vti local ${veth4_a_addr} remote ${veth4_b_addr} key 10
1978	[ $? -ne 0 ] && err "  vti not supported" && return $ksft_skip
1979	run_cmd ${ns_a} ip link del vti4_a
1980
1981	fail=0
1982
1983	min=68
1984	max=$((65535 - 20))
1985	# Check invalid values first
1986	for v in $((min - 1)) $((max + 1)); do
1987		run_cmd ${ns_a} ip link add vti4_a mtu ${v} type vti local ${veth4_a_addr} remote ${veth4_b_addr} key 10
1988		# This can fail, or MTU can be adjusted to a proper value
1989		[ $? -ne 0 ] && continue
1990		mtu="$(link_get_mtu "${ns_a}" vti4_a)"
1991		if [ ${mtu} -lt ${min} -o ${mtu} -gt ${max} ]; then
1992			err "  vti tunnel created with invalid MTU ${mtu}"
1993			fail=1
1994		fi
1995		run_cmd ${ns_a} ip link del vti4_a
1996	done
1997
1998	# Now check valid values
1999	for v in ${min} 1300 ${max}; do
2000		run_cmd ${ns_a} ip link add vti4_a mtu ${v} type vti local ${veth4_a_addr} remote ${veth4_b_addr} key 10
2001		mtu="$(link_get_mtu "${ns_a}" vti4_a)"
2002		run_cmd ${ns_a} ip link del vti4_a
2003		if [ "${mtu}" != "${v}" ]; then
2004			err "  vti MTU ${mtu} doesn't match configured value ${v}"
2005			fail=1
2006		fi
2007	done
2008
2009	return ${fail}
2010}
2011
2012test_pmtu_vti6_link_add_mtu() {
2013	setup namespaces || return $ksft_skip
2014
2015	run_cmd ${ns_a} ip link add vti6_a type vti6 local ${veth6_a_addr} remote ${veth6_b_addr} key 10
2016	[ $? -ne 0 ] && err "  vti6 not supported" && return $ksft_skip
2017	run_cmd ${ns_a} ip link del vti6_a
2018
2019	fail=0
2020
2021	min=68			# vti6 can carry IPv4 packets too
2022	max=$((65535 - 40))
2023	# Check invalid values first
2024	for v in $((min - 1)) $((max + 1)); do
2025		run_cmd ${ns_a} ip link add vti6_a mtu ${v} type vti6 local ${veth6_a_addr} remote ${veth6_b_addr} key 10
2026		# This can fail, or MTU can be adjusted to a proper value
2027		[ $? -ne 0 ] && continue
2028		mtu="$(link_get_mtu "${ns_a}" vti6_a)"
2029		if [ ${mtu} -lt ${min} -o ${mtu} -gt ${max} ]; then
2030			err "  vti6 tunnel created with invalid MTU ${v}"
2031			fail=1
2032		fi
2033		run_cmd ${ns_a} ip link del vti6_a
2034	done
2035
2036	# Now check valid values
2037	for v in 68 1280 1300 $((65535 - 40)); do
2038		run_cmd ${ns_a} ip link add vti6_a mtu ${v} type vti6 local ${veth6_a_addr} remote ${veth6_b_addr} key 10
2039		mtu="$(link_get_mtu "${ns_a}" vti6_a)"
2040		run_cmd ${ns_a} ip link del vti6_a
2041		if [ "${mtu}" != "${v}" ]; then
2042			err "  vti6 MTU ${mtu} doesn't match configured value ${v}"
2043			fail=1
2044		fi
2045	done
2046
2047	return ${fail}
2048}
2049
2050test_pmtu_vti6_link_change_mtu() {
2051	setup namespaces || return $ksft_skip
2052
2053	run_cmd ${ns_a} ip link add dummy0 mtu 1500 type dummy
2054	[ $? -ne 0 ] && err "  dummy not supported" && return $ksft_skip
2055	run_cmd ${ns_a} ip link add dummy1 mtu 3000 type dummy
2056	run_cmd ${ns_a} ip link set dummy0 up
2057	run_cmd ${ns_a} ip link set dummy1 up
2058
2059	run_cmd ${ns_a} ip addr add ${dummy6_0_prefix}1/${dummy6_mask} dev dummy0
2060	run_cmd ${ns_a} ip addr add ${dummy6_1_prefix}1/${dummy6_mask} dev dummy1
2061
2062	fail=0
2063
2064	# Create vti6 interface bound to device, passing MTU, check it
2065	run_cmd ${ns_a} ip link add vti6_a mtu 1300 type vti6 remote ${dummy6_0_prefix}2 local ${dummy6_0_prefix}1
2066	mtu="$(link_get_mtu "${ns_a}" vti6_a)"
2067	if [ ${mtu} -ne 1300 ]; then
2068		err "  vti6 MTU ${mtu} doesn't match configured value 1300"
2069		fail=1
2070	fi
2071
2072	# Move to another device with different MTU, without passing MTU, check
2073	# MTU is adjusted
2074	run_cmd ${ns_a} ip link set vti6_a type vti6 remote ${dummy6_1_prefix}2 local ${dummy6_1_prefix}1
2075	mtu="$(link_get_mtu "${ns_a}" vti6_a)"
2076	if [ ${mtu} -ne $((3000 - 40)) ]; then
2077		err "  vti MTU ${mtu} is not dummy MTU 3000 minus IPv6 header length"
2078		fail=1
2079	fi
2080
2081	# Move it back, passing MTU, check MTU is not overridden
2082	run_cmd ${ns_a} ip link set vti6_a mtu 1280 type vti6 remote ${dummy6_0_prefix}2 local ${dummy6_0_prefix}1
2083	mtu="$(link_get_mtu "${ns_a}" vti6_a)"
2084	if [ ${mtu} -ne 1280 ]; then
2085		err "  vti6 MTU ${mtu} doesn't match configured value 1280"
2086		fail=1
2087	fi
2088
2089	return ${fail}
2090}
2091
2092check_command() {
2093	cmd=${1}
2094
2095	if ! which ${cmd} > /dev/null 2>&1; then
2096		err "  missing required command: '${cmd}'"
2097		return 1
2098	fi
2099	return 0
2100}
2101
2102check_running() {
2103	pid=${1}
2104	cmd=${2}
2105
2106	[ "$(cat /proc/${pid}/cmdline 2>/dev/null | tr -d '\0')" = "${cmd}" ]
2107}
2108
2109test_cleanup_vxlanX_exception() {
2110	outer="${1}"
2111	encap="vxlan"
2112	ll_mtu=4000
2113
2114	check_command taskset || return $ksft_skip
2115	cpu_list=$(grep -m 2 processor /proc/cpuinfo | cut -d ' ' -f 2)
2116
2117	setup namespaces routing ${encap}${outer} || return $ksft_skip
2118	trace "${ns_a}" ${encap}_a   "${ns_b}"  ${encap}_b \
2119	      "${ns_a}" veth_A-R1    "${ns_r1}" veth_R1-A \
2120	      "${ns_b}" veth_B-R1    "${ns_r1}" veth_R1-B
2121
2122	# Create route exception by exceeding link layer MTU
2123	mtu "${ns_a}"  veth_A-R1 $((${ll_mtu} + 1000))
2124	mtu "${ns_r1}" veth_R1-A $((${ll_mtu} + 1000))
2125	mtu "${ns_b}"  veth_B-R1 ${ll_mtu}
2126	mtu "${ns_r1}" veth_R1-B ${ll_mtu}
2127
2128	mtu "${ns_a}" ${encap}_a $((${ll_mtu} + 1000))
2129	mtu "${ns_b}" ${encap}_b $((${ll_mtu} + 1000))
2130
2131	# Fill exception cache for multiple CPUs (2)
2132	# we can always use inner IPv4 for that
2133	for cpu in ${cpu_list}; do
2134		run_cmd taskset --cpu-list ${cpu} ${ns_a} ping -q -M want -i 0.1 -w 1 -s $((${ll_mtu} + 500)) ${tunnel4_b_addr}
2135	done
2136
2137	${ns_a} ip link del dev veth_A-R1 &
2138	iplink_pid=$!
2139	for i in $(seq 1 20); do
2140		check_running ${iplink_pid} "iplinkdeldevveth_A-R1" || return 0
2141		sleep 0.1
2142	done
2143	err "  can't delete veth device in a timely manner, PMTU dst likely leaked"
2144	return 1
2145}
2146
2147test_cleanup_ipv6_exception() {
2148	test_cleanup_vxlanX_exception 6
2149}
2150
2151test_cleanup_ipv4_exception() {
2152	test_cleanup_vxlanX_exception 4
2153}
2154
2155run_test() {
2156	(
2157	tname="$1"
2158	tdesc="$2"
2159
2160	unset IFS
2161
2162	# Since cleanup() relies on variables modified by this subshell, it
2163	# has to run in this context.
2164	trap cleanup EXIT
2165
2166	if [ "$VERBOSE" = "1" ]; then
2167		printf "\n##########################################################################\n\n"
2168	fi
2169
2170	eval test_${tname}
2171	ret=$?
2172
2173	if [ $ret -eq 0 ]; then
2174		printf "TEST: %-60s  [ OK ]\n" "${tdesc}"
2175	elif [ $ret -eq 1 ]; then
2176		printf "TEST: %-60s  [FAIL]\n" "${tdesc}"
2177		if [ "${PAUSE_ON_FAIL}" = "yes" ]; then
2178			echo
2179			echo "Pausing. Hit enter to continue"
2180			read a
2181		fi
2182		err_flush
2183		exit 1
2184	elif [ $ret -eq $ksft_skip ]; then
2185		printf "TEST: %-60s  [SKIP]\n" "${tdesc}"
2186		err_flush
2187	fi
2188
2189	return $ret
2190	)
2191	ret=$?
2192	case $ret in
2193		0)
2194			all_skipped=false
2195			[ $exitcode -eq $ksft_skip ] && exitcode=0
2196		;;
2197		$ksft_skip)
2198			[ $all_skipped = true ] && exitcode=$ksft_skip
2199		;;
2200		*)
2201			all_skipped=false
2202			exitcode=1
2203		;;
2204	esac
2205
2206	return $ret
2207}
2208
2209run_test_nh() {
2210	tname="$1"
2211	tdesc="$2"
2212
2213	USE_NH=yes
2214	run_test "${tname}" "${tdesc} - nexthop objects"
2215	USE_NH=no
2216}
2217
2218test_list_flush_ipv4_exception() {
2219	setup namespaces routing || return $ksft_skip
2220	trace "${ns_a}"  veth_A-R1    "${ns_r1}" veth_R1-A \
2221	      "${ns_r1}" veth_R1-B    "${ns_b}"  veth_B-R1 \
2222	      "${ns_a}"  veth_A-R2    "${ns_r2}" veth_R2-A \
2223	      "${ns_r2}" veth_R2-B    "${ns_b}"  veth_B-R2
2224
2225	dst_prefix1="${prefix4}.${b_r1}."
2226	dst2="${prefix4}.${b_r2}.1"
2227
2228	# Set up initial MTU values
2229	mtu "${ns_a}"  veth_A-R1 2000
2230	mtu "${ns_r1}" veth_R1-A 2000
2231	mtu "${ns_r1}" veth_R1-B 1500
2232	mtu "${ns_b}"  veth_B-R1 1500
2233
2234	mtu "${ns_a}"  veth_A-R2 2000
2235	mtu "${ns_r2}" veth_R2-A 2000
2236	mtu "${ns_r2}" veth_R2-B 1500
2237	mtu "${ns_b}"  veth_B-R2 1500
2238
2239	fail=0
2240
2241	# Add 100 addresses for veth endpoint on B reached by default A route
2242	for i in $(seq 100 199); do
2243		run_cmd ${ns_b} ip addr add "${dst_prefix1}${i}" dev veth_B-R1
2244	done
2245
2246	# Create 100 cached route exceptions for path via R1, one via R2. Note
2247	# that with IPv4 we need to actually cause a route lookup that matches
2248	# the exception caused by ICMP, in order to actually have a cached
2249	# route, so we need to ping each destination twice
2250	for i in $(seq 100 199); do
2251		run_cmd ${ns_a} ping -q -M want -i 0.1 -c 2 -s 1800 "${dst_prefix1}${i}"
2252	done
2253	run_cmd ${ns_a} ping -q -M want -i 0.1 -c 2 -s 1800 "${dst2}"
2254
2255	if [ "$(${ns_a} ip -oneline route list cache | wc -l)" -ne 101 ]; then
2256		err "  can't list cached exceptions"
2257		fail=1
2258	fi
2259
2260	run_cmd ${ns_a} ip route flush cache
2261	pmtu1="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst_prefix}1)"
2262	pmtu2="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst_prefix}2)"
2263	if [ -n "${pmtu1}" ] || [ -n "${pmtu2}" ] || \
2264	   [ -n "$(${ns_a} ip route list cache)" ]; then
2265		err "  can't flush cached exceptions"
2266		fail=1
2267	fi
2268
2269	return ${fail}
2270}
2271
2272test_list_flush_ipv6_exception() {
2273	setup namespaces routing || return $ksft_skip
2274	trace "${ns_a}"  veth_A-R1    "${ns_r1}" veth_R1-A \
2275	      "${ns_r1}" veth_R1-B    "${ns_b}"  veth_B-R1 \
2276	      "${ns_a}"  veth_A-R2    "${ns_r2}" veth_R2-A \
2277	      "${ns_r2}" veth_R2-B    "${ns_b}"  veth_B-R2
2278
2279	dst_prefix1="${prefix6}:${b_r1}::"
2280	dst2="${prefix6}:${b_r2}::1"
2281
2282	# Set up initial MTU values
2283	mtu "${ns_a}"  veth_A-R1 2000
2284	mtu "${ns_r1}" veth_R1-A 2000
2285	mtu "${ns_r1}" veth_R1-B 1500
2286	mtu "${ns_b}"  veth_B-R1 1500
2287
2288	mtu "${ns_a}"  veth_A-R2 2000
2289	mtu "${ns_r2}" veth_R2-A 2000
2290	mtu "${ns_r2}" veth_R2-B 1500
2291	mtu "${ns_b}"  veth_B-R2 1500
2292
2293	fail=0
2294
2295	# Add 100 addresses for veth endpoint on B reached by default A route
2296	for i in $(seq 100 199); do
2297		run_cmd ${ns_b} ip addr add "${dst_prefix1}${i}" dev veth_B-R1
2298	done
2299
2300	# Create 100 cached route exceptions for path via R1, one via R2
2301	for i in $(seq 100 199); do
2302		run_cmd ${ns_a} ping -q -M want -i 0.1 -w 1 -s 1800 "${dst_prefix1}${i}"
2303	done
2304	run_cmd ${ns_a} ping -q -M want -i 0.1 -w 1 -s 1800 "${dst2}"
2305	if [ "$(${ns_a} ip -oneline -6 route list cache | wc -l)" -ne 101 ]; then
2306		err "  can't list cached exceptions"
2307		fail=1
2308	fi
2309
2310	run_cmd ${ns_a} ip -6 route flush cache
2311	pmtu1="$(route_get_dst_pmtu_from_exception "${ns_a}" "${dst_prefix1}100")"
2312	pmtu2="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst2})"
2313	if [ -n "${pmtu1}" ] || [ -n "${pmtu2}" ] || \
2314	   [ -n "$(${ns_a} ip -6 route list cache)" ]; then
2315		err "  can't flush cached exceptions"
2316		fail=1
2317	fi
2318
2319	return ${fail}
2320}
2321
2322test_pmtu_ipvX_route_change() {
2323	family=${1}
2324
2325	setup namespaces routing || return 2
2326	trace "${ns_a}"  veth_A-R1    "${ns_r1}" veth_R1-A \
2327	      "${ns_r1}" veth_R1-B    "${ns_b}"  veth_B-R1 \
2328	      "${ns_a}"  veth_A-R2    "${ns_r2}" veth_R2-A \
2329	      "${ns_r2}" veth_R2-B    "${ns_b}"  veth_B-R2
2330
2331	if [ ${family} -eq 4 ]; then
2332		ping=ping
2333		dst1="${prefix4}.${b_r1}.1"
2334		dst2="${prefix4}.${b_r2}.1"
2335		gw="${prefix4}.${a_r1}.2"
2336	else
2337		ping=${ping6}
2338		dst1="${prefix6}:${b_r1}::1"
2339		dst2="${prefix6}:${b_r2}::1"
2340		gw="${prefix6}:${a_r1}::2"
2341	fi
2342
2343	# Set up initial MTU values
2344	mtu "${ns_a}"  veth_A-R1 2000
2345	mtu "${ns_r1}" veth_R1-A 2000
2346	mtu "${ns_r1}" veth_R1-B 1400
2347	mtu "${ns_b}"  veth_B-R1 1400
2348
2349	mtu "${ns_a}"  veth_A-R2 2000
2350	mtu "${ns_r2}" veth_R2-A 2000
2351	mtu "${ns_r2}" veth_R2-B 1500
2352	mtu "${ns_b}"  veth_B-R2 1500
2353
2354	# Create route exceptions
2355	run_cmd ${ns_a} ${ping} -q -M want -i 0.1 -w 1 -s 1800 ${dst1}
2356	run_cmd ${ns_a} ${ping} -q -M want -i 0.1 -w 1 -s 1800 ${dst2}
2357
2358	# Check that exceptions have been created with the correct PMTU
2359	pmtu_1="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst1})"
2360	check_pmtu_value "1400" "${pmtu_1}" "exceeding MTU" || return 1
2361	pmtu_2="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst2})"
2362	check_pmtu_value "1500" "${pmtu_2}" "exceeding MTU" || return 1
2363
2364	# Replace the route from A to R1
2365	run_cmd ${ns_a} ip route change default via ${gw}
2366
2367	# Delete the device in A
2368	run_cmd ${ns_a} ip link del "veth_A-R1"
2369}
2370
2371test_pmtu_ipv4_route_change() {
2372	test_pmtu_ipvX_route_change 4
2373}
2374
2375test_pmtu_ipv6_route_change() {
2376	test_pmtu_ipvX_route_change 6
2377}
2378
2379test_pmtu_ipv4_mp_exceptions() {
2380	setup namespaces routing multipath || return $ksft_skip
2381
2382	trace "${ns_a}"  veth_A-R1    "${ns_r1}" veth_R1-A \
2383	      "${ns_r1}" veth_R1-B    "${ns_b}"  veth_B-R1 \
2384	      "${ns_a}"  veth_A-R2    "${ns_r2}" veth_R2-A \
2385	      "${ns_r2}" veth_R2-B    "${ns_b}"  veth_B-R2
2386
2387	# Set up initial MTU values
2388	mtu "${ns_a}"  veth_A-R1 2000
2389	mtu "${ns_r1}" veth_R1-A 2000
2390	mtu "${ns_r1}" veth_R1-B 1500
2391	mtu "${ns_b}"  veth_B-R1 1500
2392
2393	mtu "${ns_a}"  veth_A-R2 2000
2394	mtu "${ns_r2}" veth_R2-A 2000
2395	mtu "${ns_r2}" veth_R2-B 1500
2396	mtu "${ns_b}"  veth_B-R2 1500
2397
2398	# Ping and expect two nexthop exceptions for two routes
2399	run_cmd ${ns_a} ping -q -M want -i 0.1 -c 1 -s 1800 "${host4_b_addr}"
2400
2401	# Check that exceptions have been created with the correct PMTU
2402	pmtu_a_R1="$(route_get_dst_pmtu_from_exception "${ns_a}" "${host4_b_addr}" oif veth_A-R1)"
2403	pmtu_a_R2="$(route_get_dst_pmtu_from_exception "${ns_a}" "${host4_b_addr}" oif veth_A-R2)"
2404
2405	check_pmtu_value "1500" "${pmtu_a_R1}" "exceeding MTU (veth_A-R1)" || return 1
2406	check_pmtu_value "1500" "${pmtu_a_R2}" "exceeding MTU (veth_A-R2)" || return 1
2407}
2408
2409usage() {
2410	echo
2411	echo "$0 [OPTIONS] [TEST]..."
2412	echo "If no TEST argument is given, all tests will be run."
2413	echo
2414	echo "Options"
2415	echo "  --trace: capture traffic to TEST_INTERFACE.pcap"
2416	echo
2417	echo "Available tests${tests}"
2418	exit 1
2419}
2420
2421################################################################################
2422#
2423exitcode=0
2424desc=0
2425all_skipped=true
2426
2427while getopts :ptv o
2428do
2429	case $o in
2430	p) PAUSE_ON_FAIL=yes;;
2431	v) VERBOSE=1;;
2432	t) if which tcpdump > /dev/null 2>&1; then
2433		TRACING=1
2434	   else
2435		echo "=== tcpdump not available, tracing disabled"
2436	   fi
2437	   ;;
2438	*) usage;;
2439	esac
2440done
2441shift $(($OPTIND-1))
2442
2443IFS="
2444"
2445
2446for arg do
2447	# Check first that all requested tests are available before running any
2448	command -v > /dev/null "test_${arg}" || { echo "=== Test ${arg} not found"; usage; }
2449done
2450
2451trap cleanup EXIT
2452
2453# start clean
2454cleanup
2455
2456HAVE_NH=no
2457ip nexthop ls >/dev/null 2>&1
2458[ $? -eq 0 ] && HAVE_NH=yes
2459
2460name=""
2461desc=""
2462rerun_nh=0
2463for t in ${tests}; do
2464	[ "${name}" = "" ]	&& name="${t}"	&& continue
2465	[ "${desc}" = "" ]	&& desc="${t}"	&& continue
2466
2467	if [ "${HAVE_NH}" = "yes" ]; then
2468		rerun_nh="${t}"
2469	fi
2470
2471	run_this=1
2472	for arg do
2473		[ "${arg}" != "${arg#--*}" ] && continue
2474		[ "${arg}" = "${name}" ] && run_this=1 && break
2475		run_this=0
2476	done
2477	if [ $run_this -eq 1 ]; then
2478		run_test "${name}" "${desc}"
2479		# if test was skipped no need to retry with nexthop objects
2480		[ $? -eq $ksft_skip ] && rerun_nh=0
2481
2482		if [ "${rerun_nh}" = "1" ]; then
2483			run_test_nh "${name}" "${desc}"
2484		fi
2485	fi
2486	name=""
2487	desc=""
2488	rerun_nh=0
2489done
2490
2491exit ${exitcode}
2492