• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1#!/bin/sh
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_vxlan4_exception
30#	Set up the same network topology as pmtu_ipv4, create a VXLAN tunnel
31#	over IPv4 between A and B, routed via R1. On the link between R1 and B,
32#	set a MTU lower than the VXLAN MTU and the MTU on the link between A and
33#	R1. Send IPv4 packets, exceeding the MTU between R1 and B, over VXLAN
34#	from A to B and check that the PMTU exception is created with the right
35#	value on A
36#
37# - pmtu_ipv6_vxlan4_exception
38#	Same as pmtu_ipv4_vxlan4_exception, but send IPv6 packets from A to B
39#
40# - pmtu_ipv4_vxlan6_exception
41#	Same as pmtu_ipv4_vxlan4_exception, but use IPv6 transport from A to B
42#
43# - pmtu_ipv6_vxlan6_exception
44#	Same as pmtu_ipv4_vxlan6_exception, but send IPv6 packets from A to B
45#
46# - pmtu_ipv4_geneve4_exception
47#	Same as pmtu_ipv4_vxlan4_exception, but using a GENEVE tunnel instead of
48#	VXLAN
49#
50# - pmtu_ipv6_geneve4_exception
51#	Same as pmtu_ipv6_vxlan4_exception, but using a GENEVE tunnel instead of
52#	VXLAN
53#
54# - pmtu_ipv4_geneve6_exception
55#	Same as pmtu_ipv4_vxlan6_exception, but using a GENEVE tunnel instead of
56#	VXLAN
57#
58# - pmtu_ipv6_geneve6_exception
59#	Same as pmtu_ipv6_vxlan6_exception, but using a GENEVE tunnel instead of
60#	VXLAN
61#
62# - pmtu_ipv{4,6}_br_vxlan{4,6}_exception
63#	Set up three namespaces, A, B, and C, with routing between A and B over
64#	R1. R2 is unused in these tests. A has a veth connection to C, and is
65#	connected to B via a VXLAN endpoint, which is directly bridged to C.
66#	MTU on the B-R1 link is lower than other MTUs.
67#
68#	Check that both C and A are able to communicate with B over the VXLAN
69#	tunnel, and that PMTU exceptions with the correct values are created.
70#
71#	                  segment a_r1    segment b_r1            b_r1: 4000
72#	                .--------------R1--------------.    everything
73#	   C---veth     A                               B         else: 5000
74#	        ' bridge                                |
75#	            '---- - - - - - VXLAN - - - - - - - '
76#
77# - pmtu_ipv{4,6}_br_geneve{4,6}_exception
78#	Same as pmtu_ipv{4,6}_br_vxlan{4,6}_exception, with a GENEVE tunnel
79#	instead.
80#
81# - pmtu_ipv{4,6}_ovs_vxlan{4,6}_exception
82#	Set up two namespaces, B, and C, with routing between the init namespace
83#	and B over R1. A and R2 are unused in these tests. The init namespace
84#	has a veth connection to C, and is connected to B via a VXLAN endpoint,
85#	which is handled by Open vSwitch and bridged to C. MTU on the B-R1 link
86#	is lower than other MTUs.
87#
88#	Check that C is able to communicate with B over the VXLAN tunnel, and
89#	that PMTU exceptions with the correct values are created.
90#
91#	                  segment a_r1    segment b_r1            b_r1: 4000
92#	                .--------------R1--------------.    everything
93#	   C---veth    init                             B         else: 5000
94#	        '- ovs                                  |
95#	            '---- - - - - - VXLAN - - - - - - - '
96#
97# - pmtu_ipv{4,6}_ovs_geneve{4,6}_exception
98#	Same as pmtu_ipv{4,6}_ovs_vxlan{4,6}_exception, with a GENEVE tunnel
99#	instead.
100#
101# - pmtu_ipv{4,6}_fou{4,6}_exception
102#	Same as pmtu_ipv4_vxlan4, but using a direct IPv4/IPv6 encapsulation
103#	(FoU) over IPv4/IPv6, instead of VXLAN
104#
105# - pmtu_ipv{4,6}_fou{4,6}_exception
106#	Same as pmtu_ipv4_vxlan4, but using a generic UDP IPv4/IPv6
107#	encapsulation (GUE) over IPv4/IPv6, instead of VXLAN
108#
109# - pmtu_ipv{4,6}_ipv{4,6}_exception
110#	Same as pmtu_ipv4_vxlan4, but using a IPv4/IPv6 tunnel over IPv4/IPv6,
111#	instead of VXLAN
112#
113# - pmtu_vti4_exception
114#	Set up vti tunnel on top of veth, with xfrm states and policies, in two
115#	namespaces with matching endpoints. Check that route exception is not
116#	created if link layer MTU is not exceeded, then exceed it and check that
117#	exception is created with the expected PMTU. The approach described
118#	below for IPv6 doesn't apply here, because, on IPv4, administrative MTU
119#	changes alone won't affect PMTU
120#
121# - pmtu_vti4_udp_exception
122#       Same as pmtu_vti4_exception, but using ESP-in-UDP
123#
124# - pmtu_vti4_udp_routed_exception
125#       Set up vti tunnel on top of veth connected through routing namespace and
126#	add xfrm states and policies with ESP-in-UDP encapsulation. Check that
127#	route exception is not created if link layer MTU is not exceeded, then
128#	lower MTU on second part of routed environment and check that exception
129#	is created with the expected PMTU.
130#
131# - pmtu_vti6_exception
132#	Set up vti6 tunnel on top of veth, with xfrm states and policies, in two
133#	namespaces with matching endpoints. Check that route exception is
134#	created by exceeding link layer MTU with ping to other endpoint. Then
135#	decrease and increase MTU of tunnel, checking that route exception PMTU
136#	changes accordingly
137#
138# - pmtu_vti6_udp_exception
139#       Same as pmtu_vti6_exception, but using ESP-in-UDP
140#
141# - pmtu_vti6_udp_routed_exception
142#	Same as pmtu_vti6_udp_routed_exception but with routing between vti
143#	endpoints
144#
145# - pmtu_vti4_default_mtu
146#	Set up vti4 tunnel on top of veth, in two namespaces with matching
147#	endpoints. Check that MTU assigned to vti interface is the MTU of the
148#	lower layer (veth) minus additional lower layer headers (zero, for veth)
149#	minus IPv4 header length
150#
151# - pmtu_vti6_default_mtu
152#	Same as above, for IPv6
153#
154# - pmtu_vti4_link_add_mtu
155#	Set up vti4 interface passing MTU value at link creation, check MTU is
156#	configured, and that link is not created with invalid MTU values
157#
158# - pmtu_vti6_link_add_mtu
159#	Same as above, for IPv6
160#
161# - pmtu_vti6_link_change_mtu
162#	Set up two dummy interfaces with different MTUs, create a vti6 tunnel
163#	and check that configured MTU is used on link creation and changes, and
164#	that MTU is properly calculated instead when MTU is not configured from
165#	userspace
166#
167# - cleanup_ipv4_exception
168#	Similar to pmtu_ipv4_vxlan4_exception, but explicitly generate PMTU
169#	exceptions on multiple CPUs and check that the veth device tear-down
170# 	happens in a timely manner
171#
172# - cleanup_ipv6_exception
173#	Same as above, but use IPv6 transport from A to B
174#
175# - list_flush_ipv4_exception
176#	Using the same topology as in pmtu_ipv4, create exceptions, and check
177#	they are shown when listing exception caches, gone after flushing them
178#
179# - list_flush_ipv6_exception
180#	Using the same topology as in pmtu_ipv6, create exceptions, and check
181#	they are shown when listing exception caches, gone after flushing them
182#
183# - pmtu_ipv4_route_change
184#	Use the same topology as in pmtu_ipv4, but issue a route replacement
185#	command and delete the corresponding device afterward. This tests for
186#	proper cleanup of the PMTU exceptions by the route replacement path.
187#	Device unregistration should complete successfully
188#
189# - pmtu_ipv6_route_change
190#	Same as above but with IPv6
191
192# Kselftest framework requirement - SKIP code is 4.
193ksft_skip=4
194
195PAUSE_ON_FAIL=no
196VERBOSE=0
197TRACING=0
198
199# Some systems don't have a ping6 binary anymore
200which ping6 > /dev/null 2>&1 && ping6=$(which ping6) || ping6=$(which ping)
201
202#               Name                          Description                  re-run with nh
203tests="
204	pmtu_ipv4_exception		ipv4: PMTU exceptions			1
205	pmtu_ipv6_exception		ipv6: PMTU exceptions			1
206	pmtu_ipv4_vxlan4_exception	IPv4 over vxlan4: PMTU exceptions	1
207	pmtu_ipv6_vxlan4_exception	IPv6 over vxlan4: PMTU exceptions	1
208	pmtu_ipv4_vxlan6_exception	IPv4 over vxlan6: PMTU exceptions	1
209	pmtu_ipv6_vxlan6_exception	IPv6 over vxlan6: PMTU exceptions	1
210	pmtu_ipv4_geneve4_exception	IPv4 over geneve4: PMTU exceptions	1
211	pmtu_ipv6_geneve4_exception	IPv6 over geneve4: PMTU exceptions	1
212	pmtu_ipv4_geneve6_exception	IPv4 over geneve6: PMTU exceptions	1
213	pmtu_ipv6_geneve6_exception	IPv6 over geneve6: PMTU exceptions	1
214	pmtu_ipv4_br_vxlan4_exception	IPv4, bridged vxlan4: PMTU exceptions	1
215	pmtu_ipv6_br_vxlan4_exception	IPv6, bridged vxlan4: PMTU exceptions	1
216	pmtu_ipv4_br_vxlan6_exception	IPv4, bridged vxlan6: PMTU exceptions	1
217	pmtu_ipv6_br_vxlan6_exception	IPv6, bridged vxlan6: PMTU exceptions	1
218	pmtu_ipv4_br_geneve4_exception	IPv4, bridged geneve4: PMTU exceptions	1
219	pmtu_ipv6_br_geneve4_exception	IPv6, bridged geneve4: PMTU exceptions	1
220	pmtu_ipv4_br_geneve6_exception	IPv4, bridged geneve6: PMTU exceptions	1
221	pmtu_ipv6_br_geneve6_exception	IPv6, bridged geneve6: PMTU exceptions	1
222	pmtu_ipv4_ovs_vxlan4_exception	IPv4, OVS vxlan4: PMTU exceptions	1
223	pmtu_ipv6_ovs_vxlan4_exception	IPv6, OVS vxlan4: PMTU exceptions	1
224	pmtu_ipv4_ovs_vxlan6_exception	IPv4, OVS vxlan6: PMTU exceptions	1
225	pmtu_ipv6_ovs_vxlan6_exception	IPv6, OVS vxlan6: PMTU exceptions	1
226	pmtu_ipv4_ovs_geneve4_exception	IPv4, OVS geneve4: PMTU exceptions	1
227	pmtu_ipv6_ovs_geneve4_exception	IPv6, OVS geneve4: PMTU exceptions	1
228	pmtu_ipv4_ovs_geneve6_exception	IPv4, OVS geneve6: PMTU exceptions	1
229	pmtu_ipv6_ovs_geneve6_exception	IPv6, OVS geneve6: PMTU exceptions	1
230	pmtu_ipv4_fou4_exception	IPv4 over fou4: PMTU exceptions		1
231	pmtu_ipv6_fou4_exception	IPv6 over fou4: PMTU exceptions		1
232	pmtu_ipv4_fou6_exception	IPv4 over fou6: PMTU exceptions		1
233	pmtu_ipv6_fou6_exception	IPv6 over fou6: PMTU exceptions		1
234	pmtu_ipv4_gue4_exception	IPv4 over gue4: PMTU exceptions		1
235	pmtu_ipv6_gue4_exception	IPv6 over gue4: PMTU exceptions		1
236	pmtu_ipv4_gue6_exception	IPv4 over gue6: PMTU exceptions		1
237	pmtu_ipv6_gue6_exception	IPv6 over gue6: PMTU exceptions		1
238	pmtu_ipv4_ipv4_exception	IPv4 over IPv4: PMTU exceptions		1
239	pmtu_ipv6_ipv4_exception	IPv6 over IPv4: PMTU exceptions		1
240	pmtu_ipv4_ipv6_exception	IPv4 over IPv6: PMTU exceptions		1
241	pmtu_ipv6_ipv6_exception	IPv6 over IPv6: PMTU exceptions		1
242	pmtu_vti6_exception		vti6: PMTU exceptions			0
243	pmtu_vti4_exception		vti4: PMTU exceptions			0
244	pmtu_vti6_udp_exception		vti6: PMTU exceptions (ESP-in-UDP)	0
245	pmtu_vti4_udp_exception		vti4: PMTU exceptions (ESP-in-UDP)	0
246	pmtu_vti6_udp_routed_exception	vti6: PMTU exceptions, routed (ESP-in-UDP)	0
247	pmtu_vti4_udp_routed_exception	vti4: PMTU exceptions, routed (ESP-in-UDP)	0
248	pmtu_vti4_default_mtu		vti4: default MTU assignment		0
249	pmtu_vti6_default_mtu		vti6: default MTU assignment		0
250	pmtu_vti4_link_add_mtu		vti4: MTU setting on link creation	0
251	pmtu_vti6_link_add_mtu		vti6: MTU setting on link creation	0
252	pmtu_vti6_link_change_mtu	vti6: MTU changes on link changes	0
253	cleanup_ipv4_exception		ipv4: cleanup of cached exceptions	1
254	cleanup_ipv6_exception		ipv6: cleanup of cached exceptions	1
255	list_flush_ipv4_exception	ipv4: list and flush cached exceptions	1
256	list_flush_ipv6_exception	ipv6: list and flush cached exceptions	1
257	pmtu_ipv4_route_change		ipv4: PMTU exception w/route replace	1
258	pmtu_ipv6_route_change		ipv6: PMTU exception w/route replace	1"
259
260NS_A="ns-A"
261NS_B="ns-B"
262NS_C="ns-C"
263NS_R1="ns-R1"
264NS_R2="ns-R2"
265ns_a="ip netns exec ${NS_A}"
266ns_b="ip netns exec ${NS_B}"
267ns_c="ip netns exec ${NS_C}"
268ns_r1="ip netns exec ${NS_R1}"
269ns_r2="ip netns exec ${NS_R2}"
270# Addressing and routing for tests with routers: four network segments, with
271# index SEGMENT between 1 and 4, a common prefix (PREFIX4 or PREFIX6) and an
272# identifier ID, which is 1 for hosts (A and B), 2 for routers (R1 and R2).
273# Addresses are:
274# - IPv4: PREFIX4.SEGMENT.ID (/24)
275# - IPv6: PREFIX6:SEGMENT::ID (/64)
276prefix4="10.0"
277prefix6="fc00"
278a_r1=1
279a_r2=2
280b_r1=3
281b_r2=4
282#	ns	peer	segment
283routing_addrs="
284	A	R1	${a_r1}
285	A	R2	${a_r2}
286	B	R1	${b_r1}
287	B	R2	${b_r2}
288"
289# Traffic from A to B goes through R1 by default, and through R2, if destined to
290# B's address on the b_r2 segment.
291# Traffic from B to A goes through R1.
292#	ns	destination		gateway
293routes="
294	A	default			${prefix4}.${a_r1}.2
295	A	${prefix4}.${b_r2}.1	${prefix4}.${a_r2}.2
296	B	default			${prefix4}.${b_r1}.2
297
298	A	default			${prefix6}:${a_r1}::2
299	A	${prefix6}:${b_r2}::1	${prefix6}:${a_r2}::2
300	B	default			${prefix6}:${b_r1}::2
301"
302USE_NH="no"
303#	ns	family	nh id	   destination		gateway
304nexthops="
305	A	4	41	${prefix4}.${a_r1}.2	veth_A-R1
306	A	4	42	${prefix4}.${a_r2}.2	veth_A-R2
307	B	4	41	${prefix4}.${b_r1}.2	veth_B-R1
308
309	A	6	61	${prefix6}:${a_r1}::2	veth_A-R1
310	A	6	62	${prefix6}:${a_r2}::2	veth_A-R2
311	B	6	61	${prefix6}:${b_r1}::2	veth_B-R1
312"
313
314# nexthop id correlates to id in nexthops config above
315#	ns    family	prefix			nh id
316routes_nh="
317	A	4	default			41
318	A	4	${prefix4}.${b_r2}.1	42
319	B	4	default			41
320
321	A	6	default			61
322	A	6	${prefix6}:${b_r2}::1	62
323	B	6	default			61
324"
325
326veth4_a_addr="192.168.1.1"
327veth4_b_addr="192.168.1.2"
328veth4_c_addr="192.168.2.10"
329veth4_mask="24"
330veth6_a_addr="fd00:1::a"
331veth6_b_addr="fd00:1::b"
332veth6_c_addr="fd00:2::c"
333veth6_mask="64"
334
335tunnel4_a_addr="192.168.2.1"
336tunnel4_b_addr="192.168.2.2"
337tunnel4_mask="24"
338tunnel6_a_addr="fd00:2::a"
339tunnel6_b_addr="fd00:2::b"
340tunnel6_mask="64"
341
342dummy6_0_prefix="fc00:1000::"
343dummy6_1_prefix="fc00:1001::"
344dummy6_mask="64"
345
346err_buf=
347tcpdump_pids=
348nettest_pids=
349
350err() {
351	err_buf="${err_buf}${1}
352"
353}
354
355err_flush() {
356	echo -n "${err_buf}"
357	err_buf=
358}
359
360run_cmd() {
361	cmd="$*"
362
363	if [ "$VERBOSE" = "1" ]; then
364		printf "    COMMAND: $cmd\n"
365	fi
366
367	out="$($cmd 2>&1)"
368	rc=$?
369	if [ "$VERBOSE" = "1" -a -n "$out" ]; then
370		echo "    $out"
371		echo
372	fi
373
374	return $rc
375}
376
377run_cmd_bg() {
378	cmd="$*"
379
380	if [ "$VERBOSE" = "1" ]; then
381		printf "    COMMAND: %s &\n" "${cmd}"
382	fi
383
384	$cmd 2>&1 &
385}
386
387# Find the auto-generated name for this namespace
388nsname() {
389	eval echo \$NS_$1
390}
391
392setup_fou_or_gue() {
393	outer="${1}"
394	inner="${2}"
395	encap="${3}"
396
397	if [ "${outer}" = "4" ]; then
398		modprobe fou || return $ksft_skip
399		a_addr="${prefix4}.${a_r1}.1"
400		b_addr="${prefix4}.${b_r1}.1"
401		if [ "${inner}" = "4" ]; then
402			type="ipip"
403			ipproto="4"
404		else
405			type="sit"
406			ipproto="41"
407		fi
408	else
409		modprobe fou6 || return $ksft_skip
410		a_addr="${prefix6}:${a_r1}::1"
411		b_addr="${prefix6}:${b_r1}::1"
412		if [ "${inner}" = "4" ]; then
413			type="ip6tnl"
414			mode="mode ipip6"
415			ipproto="4 -6"
416		else
417			type="ip6tnl"
418			mode="mode ip6ip6"
419			ipproto="41 -6"
420		fi
421	fi
422
423	run_cmd ${ns_a} ip fou add port 5555 ipproto ${ipproto} || return $ksft_skip
424	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
425
426	run_cmd ${ns_b} ip fou add port 5556 ipproto ${ipproto}
427	run_cmd ${ns_b} ip link add ${encap}_b type ${type} ${mode} local ${b_addr} remote ${a_addr} encap ${encap} encap-sport auto encap-dport 5555
428
429	if [ "${inner}" = "4" ]; then
430		run_cmd ${ns_a} ip addr add ${tunnel4_a_addr}/${tunnel4_mask} dev ${encap}_a
431		run_cmd ${ns_b} ip addr add ${tunnel4_b_addr}/${tunnel4_mask} dev ${encap}_b
432	else
433		run_cmd ${ns_a} ip addr add ${tunnel6_a_addr}/${tunnel6_mask} dev ${encap}_a
434		run_cmd ${ns_b} ip addr add ${tunnel6_b_addr}/${tunnel6_mask} dev ${encap}_b
435	fi
436
437	run_cmd ${ns_a} ip link set ${encap}_a up
438	run_cmd ${ns_b} ip link set ${encap}_b up
439}
440
441setup_fou44() {
442	setup_fou_or_gue 4 4 fou
443}
444
445setup_fou46() {
446	setup_fou_or_gue 4 6 fou
447}
448
449setup_fou64() {
450	setup_fou_or_gue 6 4 fou
451}
452
453setup_fou66() {
454	setup_fou_or_gue 6 6 fou
455}
456
457setup_gue44() {
458	setup_fou_or_gue 4 4 gue
459}
460
461setup_gue46() {
462	setup_fou_or_gue 4 6 gue
463}
464
465setup_gue64() {
466	setup_fou_or_gue 6 4 gue
467}
468
469setup_gue66() {
470	setup_fou_or_gue 6 6 gue
471}
472
473setup_ipvX_over_ipvY() {
474	inner=${1}
475	outer=${2}
476
477	if [ "${outer}" -eq 4 ]; then
478		a_addr="${prefix4}.${a_r1}.1"
479		b_addr="${prefix4}.${b_r1}.1"
480		if [ "${inner}" -eq 4 ]; then
481			type="ipip"
482			mode="ipip"
483		else
484			type="sit"
485			mode="ip6ip"
486		fi
487	else
488		a_addr="${prefix6}:${a_r1}::1"
489		b_addr="${prefix6}:${b_r1}::1"
490		type="ip6tnl"
491		if [ "${inner}" -eq 4 ]; then
492			mode="ipip6"
493		else
494			mode="ip6ip6"
495		fi
496	fi
497
498	run_cmd ${ns_a} ip link add ip_a type ${type} local ${a_addr} remote ${b_addr} mode ${mode} || return $ksft_skip
499	run_cmd ${ns_b} ip link add ip_b type ${type} local ${b_addr} remote ${a_addr} mode ${mode}
500
501	run_cmd ${ns_a} ip link set ip_a up
502	run_cmd ${ns_b} ip link set ip_b up
503
504	if [ "${inner}" = "4" ]; then
505		run_cmd ${ns_a} ip addr add ${tunnel4_a_addr}/${tunnel4_mask} dev ip_a
506		run_cmd ${ns_b} ip addr add ${tunnel4_b_addr}/${tunnel4_mask} dev ip_b
507	else
508		run_cmd ${ns_a} ip addr add ${tunnel6_a_addr}/${tunnel6_mask} dev ip_a
509		run_cmd ${ns_b} ip addr add ${tunnel6_b_addr}/${tunnel6_mask} dev ip_b
510	fi
511}
512
513setup_ip4ip4() {
514	setup_ipvX_over_ipvY 4 4
515}
516
517setup_ip6ip4() {
518	setup_ipvX_over_ipvY 6 4
519}
520
521setup_ip4ip6() {
522	setup_ipvX_over_ipvY 4 6
523}
524
525setup_ip6ip6() {
526	setup_ipvX_over_ipvY 6 6
527}
528
529setup_namespaces() {
530	for n in ${NS_A} ${NS_B} ${NS_C} ${NS_R1} ${NS_R2}; do
531		ip netns add ${n} || return 1
532
533		# Disable DAD, so that we don't have to wait to use the
534		# configured IPv6 addresses
535		ip netns exec ${n} sysctl -q net/ipv6/conf/default/accept_dad=0
536	done
537}
538
539setup_veth() {
540	run_cmd ${ns_a} ip link add veth_a type veth peer name veth_b || return 1
541	run_cmd ${ns_a} ip link set veth_b netns ${NS_B}
542
543	run_cmd ${ns_a} ip addr add ${veth4_a_addr}/${veth4_mask} dev veth_a
544	run_cmd ${ns_b} ip addr add ${veth4_b_addr}/${veth4_mask} dev veth_b
545
546	run_cmd ${ns_a} ip addr add ${veth6_a_addr}/${veth6_mask} dev veth_a
547	run_cmd ${ns_b} ip addr add ${veth6_b_addr}/${veth6_mask} dev veth_b
548
549	run_cmd ${ns_a} ip link set veth_a up
550	run_cmd ${ns_b} ip link set veth_b up
551}
552
553setup_vti() {
554	proto=${1}
555	veth_a_addr="${2}"
556	veth_b_addr="${3}"
557	vti_a_addr="${4}"
558	vti_b_addr="${5}"
559	vti_mask=${6}
560
561	[ ${proto} -eq 6 ] && vti_type="vti6" || vti_type="vti"
562
563	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
564	run_cmd ${ns_b} ip link add vti${proto}_b type ${vti_type} local ${veth_b_addr} remote ${veth_a_addr} key 10
565
566	run_cmd ${ns_a} ip addr add ${vti_a_addr}/${vti_mask} dev vti${proto}_a
567	run_cmd ${ns_b} ip addr add ${vti_b_addr}/${vti_mask} dev vti${proto}_b
568
569	run_cmd ${ns_a} ip link set vti${proto}_a up
570	run_cmd ${ns_b} ip link set vti${proto}_b up
571}
572
573setup_vti4() {
574	setup_vti 4 ${veth4_a_addr} ${veth4_b_addr} ${tunnel4_a_addr} ${tunnel4_b_addr} ${tunnel4_mask}
575}
576
577setup_vti6() {
578	setup_vti 6 ${veth6_a_addr} ${veth6_b_addr} ${tunnel6_a_addr} ${tunnel6_b_addr} ${tunnel6_mask}
579}
580
581setup_vti4routed() {
582	setup_vti 4 ${prefix4}.${a_r1}.1 ${prefix4}.${b_r1}.1 ${tunnel4_a_addr} ${tunnel4_b_addr} ${tunnel4_mask}
583}
584
585setup_vti6routed() {
586	setup_vti 6 ${prefix6}:${a_r1}::1 ${prefix6}:${b_r1}::1 ${tunnel6_a_addr} ${tunnel6_b_addr} ${tunnel6_mask}
587}
588
589setup_vxlan_or_geneve() {
590	type="${1}"
591	a_addr="${2}"
592	b_addr="${3}"
593	opts="${4}"
594	br_if_a="${5}"
595
596	if [ "${type}" = "vxlan" ]; then
597		opts="${opts} ttl 64 dstport 4789"
598		opts_a="local ${a_addr}"
599		opts_b="local ${b_addr}"
600	else
601		opts_a=""
602		opts_b=""
603	fi
604
605	run_cmd ${ns_a} ip link add ${type}_a type ${type} id 1 ${opts_a} remote ${b_addr} ${opts} || return 1
606	run_cmd ${ns_b} ip link add ${type}_b type ${type} id 1 ${opts_b} remote ${a_addr} ${opts}
607
608	if [ -n "${br_if_a}" ]; then
609		run_cmd ${ns_a} ip addr add ${tunnel4_a_addr}/${tunnel4_mask} dev ${br_if_a}
610		run_cmd ${ns_a} ip addr add ${tunnel6_a_addr}/${tunnel6_mask} dev ${br_if_a}
611		run_cmd ${ns_a} ip link set ${type}_a master ${br_if_a}
612	else
613		run_cmd ${ns_a} ip addr add ${tunnel4_a_addr}/${tunnel4_mask} dev ${type}_a
614		run_cmd ${ns_a} ip addr add ${tunnel6_a_addr}/${tunnel6_mask} dev ${type}_a
615	fi
616
617	run_cmd ${ns_b} ip addr add ${tunnel4_b_addr}/${tunnel4_mask} dev ${type}_b
618	run_cmd ${ns_b} ip addr add ${tunnel6_b_addr}/${tunnel6_mask} dev ${type}_b
619
620	run_cmd ${ns_a} ip link set ${type}_a up
621	run_cmd ${ns_b} ip link set ${type}_b up
622}
623
624setup_geneve4() {
625	setup_vxlan_or_geneve geneve ${prefix4}.${a_r1}.1  ${prefix4}.${b_r1}.1  "df set"
626}
627
628setup_vxlan4() {
629	setup_vxlan_or_geneve vxlan  ${prefix4}.${a_r1}.1  ${prefix4}.${b_r1}.1  "df set"
630}
631
632setup_geneve6() {
633	setup_vxlan_or_geneve geneve ${prefix6}:${a_r1}::1 ${prefix6}:${b_r1}::1 ""
634}
635
636setup_vxlan6() {
637	setup_vxlan_or_geneve vxlan  ${prefix6}:${a_r1}::1 ${prefix6}:${b_r1}::1 ""
638}
639
640setup_bridged_geneve4() {
641	setup_vxlan_or_geneve geneve ${prefix4}.${a_r1}.1  ${prefix4}.${b_r1}.1  "df set" "br0"
642}
643
644setup_bridged_vxlan4() {
645	setup_vxlan_or_geneve vxlan  ${prefix4}.${a_r1}.1  ${prefix4}.${b_r1}.1  "df set" "br0"
646}
647
648setup_bridged_geneve6() {
649	setup_vxlan_or_geneve geneve ${prefix6}:${a_r1}::1 ${prefix6}:${b_r1}::1 "" "br0"
650}
651
652setup_bridged_vxlan6() {
653	setup_vxlan_or_geneve vxlan  ${prefix6}:${a_r1}::1 ${prefix6}:${b_r1}::1 "" "br0"
654}
655
656setup_xfrm() {
657	proto=${1}
658	veth_a_addr="${2}"
659	veth_b_addr="${3}"
660	encap=${4}
661
662	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
663	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}
664	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
665	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
666
667	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}
668	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}
669	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
670	run_cmd ${ns_b} ip -${proto} xfrm policy add dir in mark 10 tmpl src ${veth_a_addr} dst ${veth_b_addr} proto esp mode tunnel
671}
672
673setup_nettest_xfrm() {
674	if ! which nettest >/dev/null; then
675		PATH=$PWD:$PATH
676		if ! which nettest >/dev/null; then
677			echo "'nettest' command not found; skipping tests"
678			return 1
679		fi
680	fi
681
682	[ ${1} -eq 6 ] && proto="-6" || proto=""
683	port=${2}
684
685	run_cmd_bg "${ns_a}" nettest "${proto}" -q -D -s -x -p "${port}" -t 5
686	nettest_pids="${nettest_pids} $!"
687
688	run_cmd_bg "${ns_b}" nettest "${proto}" -q -D -s -x -p "${port}" -t 5
689	nettest_pids="${nettest_pids} $!"
690}
691
692setup_xfrm4() {
693	setup_xfrm 4 ${veth4_a_addr} ${veth4_b_addr}
694}
695
696setup_xfrm6() {
697	setup_xfrm 6 ${veth6_a_addr} ${veth6_b_addr}
698}
699
700setup_xfrm4udp() {
701	setup_xfrm 4 ${veth4_a_addr} ${veth4_b_addr} "encap espinudp 4500 4500 0.0.0.0" && \
702		setup_nettest_xfrm 4 4500
703}
704
705setup_xfrm6udp() {
706	setup_xfrm 6 ${veth6_a_addr} ${veth6_b_addr} "encap espinudp 4500 4500 0.0.0.0" && \
707		setup_nettest_xfrm 6 4500
708}
709
710setup_xfrm4udprouted() {
711	setup_xfrm 4 ${prefix4}.${a_r1}.1 ${prefix4}.${b_r1}.1 "encap espinudp 4500 4500 0.0.0.0" && \
712		setup_nettest_xfrm 4 4500
713}
714
715setup_xfrm6udprouted() {
716	setup_xfrm 6 ${prefix6}:${a_r1}::1 ${prefix6}:${b_r1}::1 "encap espinudp 4500 4500 0.0.0.0" && \
717		setup_nettest_xfrm 6 4500
718}
719
720setup_routing_old() {
721	for i in ${routes}; do
722		[ "${ns}" = "" ]	&& ns="${i}"		&& continue
723		[ "${addr}" = "" ]	&& addr="${i}"		&& continue
724		[ "${gw}" = "" ]	&& gw="${i}"
725
726		ns_name="$(nsname ${ns})"
727
728		ip -n ${ns_name} route add ${addr} via ${gw}
729
730		ns=""; addr=""; gw=""
731	done
732}
733
734setup_routing_new() {
735	for i in ${nexthops}; do
736		[ "${ns}" = "" ]	&& ns="${i}"		&& continue
737		[ "${fam}" = "" ]	&& fam="${i}"		&& continue
738		[ "${nhid}" = "" ]	&& nhid="${i}"		&& continue
739		[ "${gw}" = "" ]	&& gw="${i}"		&& continue
740		[ "${dev}" = "" ]	&& dev="${i}"
741
742		ns_name="$(nsname ${ns})"
743
744		ip -n ${ns_name} -${fam} nexthop add id ${nhid} via ${gw} dev ${dev}
745
746		ns=""; fam=""; nhid=""; gw=""; dev=""
747
748	done
749
750	for i in ${routes_nh}; do
751		[ "${ns}" = "" ]	&& ns="${i}"		&& continue
752		[ "${fam}" = "" ]	&& fam="${i}"		&& continue
753		[ "${addr}" = "" ]	&& addr="${i}"		&& continue
754		[ "${nhid}" = "" ]	&& nhid="${i}"
755
756		ns_name="$(nsname ${ns})"
757
758		ip -n ${ns_name} -${fam} route add ${addr} nhid ${nhid}
759
760		ns=""; fam=""; addr=""; nhid=""
761	done
762}
763
764setup_routing() {
765	for i in ${NS_R1} ${NS_R2}; do
766		ip netns exec ${i} sysctl -q net/ipv4/ip_forward=1
767		ip netns exec ${i} sysctl -q net/ipv6/conf/all/forwarding=1
768	done
769
770	for i in ${routing_addrs}; do
771		[ "${ns}" = "" ]	&& ns="${i}"		&& continue
772		[ "${peer}" = "" ]	&& peer="${i}"		&& continue
773		[ "${segment}" = "" ]	&& segment="${i}"
774
775		ns_name="$(nsname ${ns})"
776		peer_name="$(nsname ${peer})"
777		if="veth_${ns}-${peer}"
778		ifpeer="veth_${peer}-${ns}"
779
780		# Create veth links
781		ip link add ${if} up netns ${ns_name} type veth peer name ${ifpeer} netns ${peer_name} || return 1
782		ip -n ${peer_name} link set dev ${ifpeer} up
783
784		# Add addresses
785		ip -n ${ns_name}   addr add ${prefix4}.${segment}.1/24  dev ${if}
786		ip -n ${ns_name}   addr add ${prefix6}:${segment}::1/64 dev ${if}
787
788		ip -n ${peer_name} addr add ${prefix4}.${segment}.2/24  dev ${ifpeer}
789		ip -n ${peer_name} addr add ${prefix6}:${segment}::2/64 dev ${ifpeer}
790
791		ns=""; peer=""; segment=""
792	done
793
794	if [ "$USE_NH" = "yes" ]; then
795		setup_routing_new
796	else
797		setup_routing_old
798	fi
799
800	return 0
801}
802
803setup_bridge() {
804	run_cmd ${ns_a} ip link add br0 type bridge || return $ksft_skip
805	run_cmd ${ns_a} ip link set br0 up
806
807	run_cmd ${ns_c} ip link add veth_C-A type veth peer name veth_A-C
808	run_cmd ${ns_c} ip link set veth_A-C netns ns-A
809
810	run_cmd ${ns_a} ip link set veth_A-C up
811	run_cmd ${ns_c} ip link set veth_C-A up
812	run_cmd ${ns_c} ip addr add ${veth4_c_addr}/${veth4_mask} dev veth_C-A
813	run_cmd ${ns_c} ip addr add ${veth6_c_addr}/${veth6_mask} dev veth_C-A
814	run_cmd ${ns_a} ip link set veth_A-C master br0
815}
816
817setup_ovs_vxlan_or_geneve() {
818	type="${1}"
819	a_addr="${2}"
820	b_addr="${3}"
821
822	if [ "${type}" = "vxlan" ]; then
823		opts="${opts} ttl 64 dstport 4789"
824		opts_b="local ${b_addr}"
825	fi
826
827	run_cmd ovs-vsctl add-port ovs_br0 ${type}_a -- \
828		set interface ${type}_a type=${type} \
829		options:remote_ip=${b_addr} options:key=1 options:csum=true || return 1
830
831	run_cmd ${ns_b} ip link add ${type}_b type ${type} id 1 ${opts_b} remote ${a_addr} ${opts} || return 1
832
833	run_cmd ${ns_b} ip addr add ${tunnel4_b_addr}/${tunnel4_mask} dev ${type}_b
834	run_cmd ${ns_b} ip addr add ${tunnel6_b_addr}/${tunnel6_mask} dev ${type}_b
835
836	run_cmd ${ns_b} ip link set ${type}_b up
837}
838
839setup_ovs_geneve4() {
840	setup_ovs_vxlan_or_geneve geneve ${prefix4}.${a_r1}.1  ${prefix4}.${b_r1}.1
841}
842
843setup_ovs_vxlan4() {
844	setup_ovs_vxlan_or_geneve vxlan  ${prefix4}.${a_r1}.1  ${prefix4}.${b_r1}.1
845}
846
847setup_ovs_geneve6() {
848	setup_ovs_vxlan_or_geneve geneve ${prefix6}:${a_r1}::1 ${prefix6}:${b_r1}::1
849}
850
851setup_ovs_vxlan6() {
852	setup_ovs_vxlan_or_geneve vxlan  ${prefix6}:${a_r1}::1 ${prefix6}:${b_r1}::1
853}
854
855setup_ovs_bridge() {
856	run_cmd ovs-vsctl add-br ovs_br0 || return $ksft_skip
857	run_cmd ip link set ovs_br0 up
858
859	run_cmd ${ns_c} ip link add veth_C-A type veth peer name veth_A-C
860	run_cmd ${ns_c} ip link set veth_A-C netns 1
861
862	run_cmd         ip link set veth_A-C up
863	run_cmd ${ns_c} ip link set veth_C-A up
864	run_cmd ${ns_c} ip addr add ${veth4_c_addr}/${veth4_mask} dev veth_C-A
865	run_cmd ${ns_c} ip addr add ${veth6_c_addr}/${veth6_mask} dev veth_C-A
866	run_cmd ovs-vsctl add-port ovs_br0 veth_A-C
867
868	# Move veth_A-R1 to init
869	run_cmd ${ns_a} ip link set veth_A-R1 netns 1
870	run_cmd ip addr add ${prefix4}.${a_r1}.1/${veth4_mask} dev veth_A-R1
871	run_cmd ip addr add ${prefix6}:${a_r1}::1/${veth6_mask} dev veth_A-R1
872	run_cmd ip link set veth_A-R1 up
873	run_cmd ip route add ${prefix4}.${b_r1}.1 via ${prefix4}.${a_r1}.2
874	run_cmd ip route add ${prefix6}:${b_r1}::1 via ${prefix6}:${a_r1}::2
875}
876
877setup() {
878	[ "$(id -u)" -ne 0 ] && echo "  need to run as root" && return $ksft_skip
879
880	for arg do
881		eval setup_${arg} || { echo "  ${arg} not supported"; return 1; }
882	done
883}
884
885trace() {
886	[ $TRACING -eq 0 ] && return
887
888	for arg do
889		[ "${ns_cmd}" = "" ] && ns_cmd="${arg}" && continue
890		${ns_cmd} tcpdump --immediate-mode -s 0 -i "${arg}" -w "${name}_${arg}.pcap" 2> /dev/null &
891		tcpdump_pids="${tcpdump_pids} $!"
892		ns_cmd=
893	done
894	sleep 1
895}
896
897cleanup() {
898	for pid in ${tcpdump_pids}; do
899		kill ${pid}
900	done
901	tcpdump_pids=
902
903	for pid in ${nettest_pids}; do
904		kill ${pid}
905	done
906	nettest_pids=
907
908	for n in ${NS_A} ${NS_B} ${NS_C} ${NS_R1} ${NS_R2}; do
909		ip netns del ${n} 2> /dev/null
910	done
911
912	ip link del veth_A-C			2>/dev/null
913	ip link del veth_A-R1			2>/dev/null
914	ovs-vsctl --if-exists del-port vxlan_a	2>/dev/null
915	ovs-vsctl --if-exists del-br ovs_br0	2>/dev/null
916}
917
918mtu() {
919	ns_cmd="${1}"
920	dev="${2}"
921	mtu="${3}"
922
923	${ns_cmd} ip link set dev ${dev} mtu ${mtu}
924}
925
926mtu_parse() {
927	input="${1}"
928
929	next=0
930	for i in ${input}; do
931		[ ${next} -eq 1 -a "${i}" = "lock" ] && next=2 && continue
932		[ ${next} -eq 1 ] && echo "${i}" && return
933		[ ${next} -eq 2 ] && echo "lock ${i}" && return
934		[ "${i}" = "mtu" ] && next=1
935	done
936}
937
938link_get() {
939	ns_cmd="${1}"
940	name="${2}"
941
942	${ns_cmd} ip link show dev "${name}"
943}
944
945link_get_mtu() {
946	ns_cmd="${1}"
947	name="${2}"
948
949	mtu_parse "$(link_get "${ns_cmd}" ${name})"
950}
951
952route_get_dst_exception() {
953	ns_cmd="${1}"
954	dst="${2}"
955
956	${ns_cmd} ip route get "${dst}"
957}
958
959route_get_dst_pmtu_from_exception() {
960	ns_cmd="${1}"
961	dst="${2}"
962
963	mtu_parse "$(route_get_dst_exception "${ns_cmd}" ${dst})"
964}
965
966check_pmtu_value() {
967	expected="${1}"
968	value="${2}"
969	event="${3}"
970
971	[ "${expected}" = "any" ] && [ -n "${value}" ] && return 0
972	[ "${value}" = "${expected}" ] && return 0
973	[ -z "${value}" ] &&    err "  PMTU exception wasn't created after ${event}" && return 1
974	[ -z "${expected}" ] && err "  PMTU exception shouldn't exist after ${event}" && return 1
975	err "  found PMTU exception with incorrect MTU ${value}, expected ${expected}, after ${event}"
976	return 1
977}
978
979test_pmtu_ipvX() {
980	family=${1}
981
982	setup namespaces routing || return $ksft_skip
983	trace "${ns_a}"  veth_A-R1    "${ns_r1}" veth_R1-A \
984	      "${ns_r1}" veth_R1-B    "${ns_b}"  veth_B-R1 \
985	      "${ns_a}"  veth_A-R2    "${ns_r2}" veth_R2-A \
986	      "${ns_r2}" veth_R2-B    "${ns_b}"  veth_B-R2
987
988	if [ ${family} -eq 4 ]; then
989		ping=ping
990		dst1="${prefix4}.${b_r1}.1"
991		dst2="${prefix4}.${b_r2}.1"
992	else
993		ping=${ping6}
994		dst1="${prefix6}:${b_r1}::1"
995		dst2="${prefix6}:${b_r2}::1"
996	fi
997
998	# Set up initial MTU values
999	mtu "${ns_a}"  veth_A-R1 2000
1000	mtu "${ns_r1}" veth_R1-A 2000
1001	mtu "${ns_r1}" veth_R1-B 1400
1002	mtu "${ns_b}"  veth_B-R1 1400
1003
1004	mtu "${ns_a}"  veth_A-R2 2000
1005	mtu "${ns_r2}" veth_R2-A 2000
1006	mtu "${ns_r2}" veth_R2-B 1500
1007	mtu "${ns_b}"  veth_B-R2 1500
1008
1009	# Create route exceptions
1010	run_cmd ${ns_a} ${ping} -q -M want -i 0.1 -w 1 -s 1800 ${dst1}
1011	run_cmd ${ns_a} ${ping} -q -M want -i 0.1 -w 1 -s 1800 ${dst2}
1012
1013	# Check that exceptions have been created with the correct PMTU
1014	pmtu_1="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst1})"
1015	check_pmtu_value "1400" "${pmtu_1}" "exceeding MTU" || return 1
1016	pmtu_2="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst2})"
1017	check_pmtu_value "1500" "${pmtu_2}" "exceeding MTU" || return 1
1018
1019	# Decrease local MTU below PMTU, check for PMTU decrease in route exception
1020	mtu "${ns_a}"  veth_A-R1 1300
1021	mtu "${ns_r1}" veth_R1-A 1300
1022	pmtu_1="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst1})"
1023	check_pmtu_value "1300" "${pmtu_1}" "decreasing local MTU" || return 1
1024	# Second exception shouldn't be modified
1025	pmtu_2="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst2})"
1026	check_pmtu_value "1500" "${pmtu_2}" "changing local MTU on a link not on this path" || return 1
1027
1028	# Increase MTU, check for PMTU increase in route exception
1029	mtu "${ns_a}"  veth_A-R1 1700
1030	mtu "${ns_r1}" veth_R1-A 1700
1031	pmtu_1="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst1})"
1032	check_pmtu_value "1700" "${pmtu_1}" "increasing local MTU" || return 1
1033	# Second exception shouldn't be modified
1034	pmtu_2="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst2})"
1035	check_pmtu_value "1500" "${pmtu_2}" "changing local MTU on a link not on this path" || return 1
1036
1037	# Skip PMTU locking tests for IPv6
1038	[ $family -eq 6 ] && return 0
1039
1040	# Decrease remote MTU on path via R2, get new exception
1041	mtu "${ns_r2}" veth_R2-B 400
1042	mtu "${ns_b}"  veth_B-R2 400
1043	run_cmd ${ns_a} ${ping} -q -M want -i 0.1 -w 1 -s 1400 ${dst2}
1044	pmtu_2="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst2})"
1045	check_pmtu_value "lock 552" "${pmtu_2}" "exceeding MTU, with MTU < min_pmtu" || return 1
1046
1047	# Decrease local MTU below PMTU
1048	mtu "${ns_a}"  veth_A-R2 500
1049	mtu "${ns_r2}" veth_R2-A 500
1050	pmtu_2="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst2})"
1051	check_pmtu_value "500" "${pmtu_2}" "decreasing local MTU" || return 1
1052
1053	# Increase local MTU
1054	mtu "${ns_a}"  veth_A-R2 1500
1055	mtu "${ns_r2}" veth_R2-A 1500
1056	pmtu_2="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst2})"
1057	check_pmtu_value "1500" "${pmtu_2}" "increasing local MTU" || return 1
1058
1059	# Get new exception
1060	run_cmd ${ns_a} ${ping} -q -M want -i 0.1 -w 1 -s 1400 ${dst2}
1061	pmtu_2="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst2})"
1062	check_pmtu_value "lock 552" "${pmtu_2}" "exceeding MTU, with MTU < min_pmtu" || return 1
1063}
1064
1065test_pmtu_ipv4_exception() {
1066	test_pmtu_ipvX 4
1067}
1068
1069test_pmtu_ipv6_exception() {
1070	test_pmtu_ipvX 6
1071}
1072
1073test_pmtu_ipvX_over_vxlanY_or_geneveY_exception() {
1074	type=${1}
1075	family=${2}
1076	outer_family=${3}
1077	ll_mtu=4000
1078
1079	if [ ${outer_family} -eq 4 ]; then
1080		setup namespaces routing ${type}4 || return $ksft_skip
1081		#                      IPv4 header   UDP header   VXLAN/GENEVE header   Ethernet header
1082		exp_mtu=$((${ll_mtu} - 20          - 8          - 8                   - 14))
1083	else
1084		setup namespaces routing ${type}6 || return $ksft_skip
1085		#                      IPv6 header   UDP header   VXLAN/GENEVE header   Ethernet header
1086		exp_mtu=$((${ll_mtu} - 40          - 8          - 8                   - 14))
1087	fi
1088
1089	trace "${ns_a}" ${type}_a    "${ns_b}"  ${type}_b \
1090	      "${ns_a}" veth_A-R1    "${ns_r1}" veth_R1-A \
1091	      "${ns_b}" veth_B-R1    "${ns_r1}" veth_R1-B
1092
1093	if [ ${family} -eq 4 ]; then
1094		ping=ping
1095		dst=${tunnel4_b_addr}
1096	else
1097		ping=${ping6}
1098		dst=${tunnel6_b_addr}
1099	fi
1100
1101	# Create route exception by exceeding link layer MTU
1102	mtu "${ns_a}"  veth_A-R1 $((${ll_mtu} + 1000))
1103	mtu "${ns_r1}" veth_R1-A $((${ll_mtu} + 1000))
1104	mtu "${ns_b}"  veth_B-R1 ${ll_mtu}
1105	mtu "${ns_r1}" veth_R1-B ${ll_mtu}
1106
1107	mtu "${ns_a}" ${type}_a $((${ll_mtu} + 1000))
1108	mtu "${ns_b}" ${type}_b $((${ll_mtu} + 1000))
1109	run_cmd ${ns_a} ${ping} -q -M want -i 0.1 -w 1 -s $((${ll_mtu} + 500)) ${dst}
1110
1111	# Check that exception was created
1112	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst})"
1113	check_pmtu_value ${exp_mtu} "${pmtu}" "exceeding link layer MTU on ${type} interface"
1114}
1115
1116test_pmtu_ipv4_vxlan4_exception() {
1117	test_pmtu_ipvX_over_vxlanY_or_geneveY_exception vxlan  4 4
1118}
1119
1120test_pmtu_ipv6_vxlan4_exception() {
1121	test_pmtu_ipvX_over_vxlanY_or_geneveY_exception vxlan  6 4
1122}
1123
1124test_pmtu_ipv4_geneve4_exception() {
1125	test_pmtu_ipvX_over_vxlanY_or_geneveY_exception geneve 4 4
1126}
1127
1128test_pmtu_ipv6_geneve4_exception() {
1129	test_pmtu_ipvX_over_vxlanY_or_geneveY_exception geneve 6 4
1130}
1131
1132test_pmtu_ipv4_vxlan6_exception() {
1133	test_pmtu_ipvX_over_vxlanY_or_geneveY_exception vxlan  4 6
1134}
1135
1136test_pmtu_ipv6_vxlan6_exception() {
1137	test_pmtu_ipvX_over_vxlanY_or_geneveY_exception vxlan  6 6
1138}
1139
1140test_pmtu_ipv4_geneve6_exception() {
1141	test_pmtu_ipvX_over_vxlanY_or_geneveY_exception geneve 4 6
1142}
1143
1144test_pmtu_ipv6_geneve6_exception() {
1145	test_pmtu_ipvX_over_vxlanY_or_geneveY_exception geneve 6 6
1146}
1147
1148test_pmtu_ipvX_over_bridged_vxlanY_or_geneveY_exception() {
1149	type=${1}
1150	family=${2}
1151	outer_family=${3}
1152	ll_mtu=4000
1153
1154	if [ ${outer_family} -eq 4 ]; then
1155		setup namespaces routing bridge bridged_${type}4 || return $ksft_skip
1156		#                      IPv4 header   UDP header   VXLAN/GENEVE header   Ethernet header
1157		exp_mtu=$((${ll_mtu} - 20          - 8          - 8                   - 14))
1158	else
1159		setup namespaces routing bridge bridged_${type}6 || return $ksft_skip
1160		#                      IPv6 header   UDP header   VXLAN/GENEVE header   Ethernet header
1161		exp_mtu=$((${ll_mtu} - 40          - 8          - 8                   - 14))
1162	fi
1163
1164	trace "${ns_a}" ${type}_a    "${ns_b}"  ${type}_b \
1165	      "${ns_a}" veth_A-R1    "${ns_r1}" veth_R1-A \
1166	      "${ns_b}" veth_B-R1    "${ns_r1}" veth_R1-B \
1167	      "${ns_a}" br0          "${ns_a}"  veth-A-C  \
1168	      "${ns_c}" veth_C-A
1169
1170	if [ ${family} -eq 4 ]; then
1171		ping=ping
1172		dst=${tunnel4_b_addr}
1173	else
1174		ping=${ping6}
1175		dst=${tunnel6_b_addr}
1176	fi
1177
1178	# Create route exception by exceeding link layer MTU
1179	mtu "${ns_a}"  veth_A-R1 $((${ll_mtu} + 1000))
1180	mtu "${ns_a}"  br0       $((${ll_mtu} + 1000))
1181	mtu "${ns_a}"  veth_A-C  $((${ll_mtu} + 1000))
1182	mtu "${ns_c}"  veth_C-A  $((${ll_mtu} + 1000))
1183	mtu "${ns_r1}" veth_R1-A $((${ll_mtu} + 1000))
1184	mtu "${ns_b}"  veth_B-R1 ${ll_mtu}
1185	mtu "${ns_r1}" veth_R1-B ${ll_mtu}
1186
1187	mtu "${ns_a}" ${type}_a $((${ll_mtu} + 1000))
1188	mtu "${ns_b}" ${type}_b $((${ll_mtu} + 1000))
1189
1190	run_cmd ${ns_c} ${ping} -q -M want -i 0.1 -c 10 -s $((${ll_mtu} + 500)) ${dst} || return 1
1191	run_cmd ${ns_a} ${ping} -q -M want -i 0.1 -w 1  -s $((${ll_mtu} + 500)) ${dst} || return 1
1192
1193	# Check that exceptions were created
1194	pmtu="$(route_get_dst_pmtu_from_exception "${ns_c}" ${dst})"
1195	check_pmtu_value ${exp_mtu} "${pmtu}" "exceeding link layer MTU on bridged ${type} interface"
1196	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst})"
1197	check_pmtu_value ${exp_mtu} "${pmtu}" "exceeding link layer MTU on locally bridged ${type} interface"
1198}
1199
1200test_pmtu_ipv4_br_vxlan4_exception() {
1201	test_pmtu_ipvX_over_bridged_vxlanY_or_geneveY_exception vxlan  4 4
1202}
1203
1204test_pmtu_ipv6_br_vxlan4_exception() {
1205	test_pmtu_ipvX_over_bridged_vxlanY_or_geneveY_exception vxlan  6 4
1206}
1207
1208test_pmtu_ipv4_br_geneve4_exception() {
1209	test_pmtu_ipvX_over_bridged_vxlanY_or_geneveY_exception geneve 4 4
1210}
1211
1212test_pmtu_ipv6_br_geneve4_exception() {
1213	test_pmtu_ipvX_over_bridged_vxlanY_or_geneveY_exception geneve 6 4
1214}
1215
1216test_pmtu_ipv4_br_vxlan6_exception() {
1217	test_pmtu_ipvX_over_bridged_vxlanY_or_geneveY_exception vxlan  4 6
1218}
1219
1220test_pmtu_ipv6_br_vxlan6_exception() {
1221	test_pmtu_ipvX_over_bridged_vxlanY_or_geneveY_exception vxlan  6 6
1222}
1223
1224test_pmtu_ipv4_br_geneve6_exception() {
1225	test_pmtu_ipvX_over_bridged_vxlanY_or_geneveY_exception geneve 4 6
1226}
1227
1228test_pmtu_ipv6_br_geneve6_exception() {
1229	test_pmtu_ipvX_over_bridged_vxlanY_or_geneveY_exception geneve 6 6
1230}
1231
1232test_pmtu_ipvX_over_ovs_vxlanY_or_geneveY_exception() {
1233	type=${1}
1234	family=${2}
1235	outer_family=${3}
1236	ll_mtu=4000
1237
1238	if [ ${outer_family} -eq 4 ]; then
1239		setup namespaces routing ovs_bridge ovs_${type}4 || return $ksft_skip
1240		#                      IPv4 header   UDP header   VXLAN/GENEVE header   Ethernet header
1241		exp_mtu=$((${ll_mtu} - 20          - 8          - 8                   - 14))
1242	else
1243		setup namespaces routing ovs_bridge ovs_${type}6 || return $ksft_skip
1244		#                      IPv6 header   UDP header   VXLAN/GENEVE header   Ethernet header
1245		exp_mtu=$((${ll_mtu} - 40          - 8          - 8                   - 14))
1246	fi
1247
1248	if [ "${type}" = "vxlan" ]; then
1249		tun_a="vxlan_sys_4789"
1250	elif [ "${type}" = "geneve" ]; then
1251		tun_a="genev_sys_6081"
1252	fi
1253
1254	trace ""        "${tun_a}"  "${ns_b}"  ${type}_b \
1255	      ""        veth_A-R1   "${ns_r1}" veth_R1-A \
1256	      "${ns_b}" veth_B-R1   "${ns_r1}" veth_R1-B \
1257	      ""        ovs_br0     ""         veth-A-C  \
1258	      "${ns_c}" veth_C-A
1259
1260	if [ ${family} -eq 4 ]; then
1261		ping=ping
1262		dst=${tunnel4_b_addr}
1263	else
1264		ping=${ping6}
1265		dst=${tunnel6_b_addr}
1266	fi
1267
1268	# Create route exception by exceeding link layer MTU
1269	mtu ""         veth_A-R1 $((${ll_mtu} + 1000))
1270	mtu ""         ovs_br0   $((${ll_mtu} + 1000))
1271	mtu ""         veth_A-C  $((${ll_mtu} + 1000))
1272	mtu "${ns_c}"  veth_C-A  $((${ll_mtu} + 1000))
1273	mtu "${ns_r1}" veth_R1-A $((${ll_mtu} + 1000))
1274	mtu "${ns_b}"  veth_B-R1 ${ll_mtu}
1275	mtu "${ns_r1}" veth_R1-B ${ll_mtu}
1276
1277	mtu ""        ${tun_a}  $((${ll_mtu} + 1000))
1278	mtu "${ns_b}" ${type}_b $((${ll_mtu} + 1000))
1279
1280	run_cmd ${ns_c} ${ping} -q -M want -i 0.1 -c 20 -s $((${ll_mtu} + 500)) ${dst} || return 1
1281
1282	# Check that exceptions were created
1283	pmtu="$(route_get_dst_pmtu_from_exception "${ns_c}" ${dst})"
1284	check_pmtu_value ${exp_mtu} "${pmtu}" "exceeding link layer MTU on Open vSwitch ${type} interface"
1285}
1286
1287test_pmtu_ipv4_ovs_vxlan4_exception() {
1288	test_pmtu_ipvX_over_ovs_vxlanY_or_geneveY_exception vxlan  4 4
1289}
1290
1291test_pmtu_ipv6_ovs_vxlan4_exception() {
1292	test_pmtu_ipvX_over_ovs_vxlanY_or_geneveY_exception vxlan  6 4
1293}
1294
1295test_pmtu_ipv4_ovs_geneve4_exception() {
1296	test_pmtu_ipvX_over_ovs_vxlanY_or_geneveY_exception geneve 4 4
1297}
1298
1299test_pmtu_ipv6_ovs_geneve4_exception() {
1300	test_pmtu_ipvX_over_ovs_vxlanY_or_geneveY_exception geneve 6 4
1301}
1302
1303test_pmtu_ipv4_ovs_vxlan6_exception() {
1304	test_pmtu_ipvX_over_ovs_vxlanY_or_geneveY_exception vxlan  4 6
1305}
1306
1307test_pmtu_ipv6_ovs_vxlan6_exception() {
1308	test_pmtu_ipvX_over_ovs_vxlanY_or_geneveY_exception vxlan  6 6
1309}
1310
1311test_pmtu_ipv4_ovs_geneve6_exception() {
1312	test_pmtu_ipvX_over_ovs_vxlanY_or_geneveY_exception geneve 4 6
1313}
1314
1315test_pmtu_ipv6_ovs_geneve6_exception() {
1316	test_pmtu_ipvX_over_ovs_vxlanY_or_geneveY_exception geneve 6 6
1317}
1318
1319test_pmtu_ipvX_over_fouY_or_gueY() {
1320	inner_family=${1}
1321	outer_family=${2}
1322	encap=${3}
1323	ll_mtu=4000
1324
1325	setup namespaces routing ${encap}${outer_family}${inner_family} || return $ksft_skip
1326	trace "${ns_a}" ${encap}_a   "${ns_b}"  ${encap}_b \
1327	      "${ns_a}" veth_A-R1    "${ns_r1}" veth_R1-A \
1328	      "${ns_b}" veth_B-R1    "${ns_r1}" veth_R1-B
1329
1330	if [ ${inner_family} -eq 4 ]; then
1331		ping=ping
1332		dst=${tunnel4_b_addr}
1333	else
1334		ping=${ping6}
1335		dst=${tunnel6_b_addr}
1336	fi
1337
1338	if [ "${encap}" = "gue" ]; then
1339		encap_overhead=4
1340	else
1341		encap_overhead=0
1342	fi
1343
1344	if [ ${outer_family} -eq 4 ]; then
1345		#                      IPv4 header   UDP header
1346		exp_mtu=$((${ll_mtu} - 20          - 8         - ${encap_overhead}))
1347	else
1348		#                      IPv6 header   Option 4   UDP header
1349		exp_mtu=$((${ll_mtu} - 40          - 8        - 8       - ${encap_overhead}))
1350	fi
1351
1352	# Create route exception by exceeding link layer MTU
1353	mtu "${ns_a}"  veth_A-R1 $((${ll_mtu} + 1000))
1354	mtu "${ns_r1}" veth_R1-A $((${ll_mtu} + 1000))
1355	mtu "${ns_b}"  veth_B-R1 ${ll_mtu}
1356	mtu "${ns_r1}" veth_R1-B ${ll_mtu}
1357
1358	mtu "${ns_a}" ${encap}_a $((${ll_mtu} + 1000))
1359	mtu "${ns_b}" ${encap}_b $((${ll_mtu} + 1000))
1360	run_cmd ${ns_a} ${ping} -q -M want -i 0.1 -w 1 -s $((${ll_mtu} + 500)) ${dst}
1361
1362	# Check that exception was created
1363	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst})"
1364	check_pmtu_value ${exp_mtu} "${pmtu}" "exceeding link layer MTU on ${encap} interface"
1365}
1366
1367test_pmtu_ipv4_fou4_exception() {
1368	test_pmtu_ipvX_over_fouY_or_gueY 4 4 fou
1369}
1370
1371test_pmtu_ipv6_fou4_exception() {
1372	test_pmtu_ipvX_over_fouY_or_gueY 6 4 fou
1373}
1374
1375test_pmtu_ipv4_fou6_exception() {
1376	test_pmtu_ipvX_over_fouY_or_gueY 4 6 fou
1377}
1378
1379test_pmtu_ipv6_fou6_exception() {
1380	test_pmtu_ipvX_over_fouY_or_gueY 6 6 fou
1381}
1382
1383test_pmtu_ipv4_gue4_exception() {
1384	test_pmtu_ipvX_over_fouY_or_gueY 4 4 gue
1385}
1386
1387test_pmtu_ipv6_gue4_exception() {
1388	test_pmtu_ipvX_over_fouY_or_gueY 6 4 gue
1389}
1390
1391test_pmtu_ipv4_gue6_exception() {
1392	test_pmtu_ipvX_over_fouY_or_gueY 4 6 gue
1393}
1394
1395test_pmtu_ipv6_gue6_exception() {
1396	test_pmtu_ipvX_over_fouY_or_gueY 6 6 gue
1397}
1398
1399test_pmtu_ipvX_over_ipvY_exception() {
1400	inner=${1}
1401	outer=${2}
1402	ll_mtu=4000
1403
1404	setup namespaces routing ip${inner}ip${outer} || return $ksft_skip
1405
1406	trace "${ns_a}" ip_a         "${ns_b}"  ip_b  \
1407	      "${ns_a}" veth_A-R1    "${ns_r1}" veth_R1-A \
1408	      "${ns_b}" veth_B-R1    "${ns_r1}" veth_R1-B
1409
1410	if [ ${inner} -eq 4 ]; then
1411		ping=ping
1412		dst=${tunnel4_b_addr}
1413	else
1414		ping=${ping6}
1415		dst=${tunnel6_b_addr}
1416	fi
1417
1418	if [ ${outer} -eq 4 ]; then
1419		#                      IPv4 header
1420		exp_mtu=$((${ll_mtu} - 20))
1421	else
1422		#                      IPv6 header   Option 4
1423		exp_mtu=$((${ll_mtu} - 40          - 8))
1424	fi
1425
1426	# Create route exception by exceeding link layer MTU
1427	mtu "${ns_a}"  veth_A-R1 $((${ll_mtu} + 1000))
1428	mtu "${ns_r1}" veth_R1-A $((${ll_mtu} + 1000))
1429	mtu "${ns_b}"  veth_B-R1 ${ll_mtu}
1430	mtu "${ns_r1}" veth_R1-B ${ll_mtu}
1431
1432	mtu "${ns_a}" ip_a $((${ll_mtu} + 1000)) || return
1433	mtu "${ns_b}" ip_b $((${ll_mtu} + 1000)) || return
1434	run_cmd ${ns_a} ${ping} -q -M want -i 0.1 -w 1 -s $((${ll_mtu} + 500)) ${dst}
1435
1436	# Check that exception was created
1437	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst})"
1438	check_pmtu_value ${exp_mtu} "${pmtu}" "exceeding link layer MTU on ip${inner}ip${outer} interface"
1439}
1440
1441test_pmtu_ipv4_ipv4_exception() {
1442	test_pmtu_ipvX_over_ipvY_exception 4 4
1443}
1444
1445test_pmtu_ipv6_ipv4_exception() {
1446	test_pmtu_ipvX_over_ipvY_exception 6 4
1447}
1448
1449test_pmtu_ipv4_ipv6_exception() {
1450	test_pmtu_ipvX_over_ipvY_exception 4 6
1451}
1452
1453test_pmtu_ipv6_ipv6_exception() {
1454	test_pmtu_ipvX_over_ipvY_exception 6 6
1455}
1456
1457test_pmtu_vti4_exception() {
1458	setup namespaces veth vti4 xfrm4 || return $ksft_skip
1459	trace "${ns_a}" veth_a    "${ns_b}" veth_b \
1460	      "${ns_a}" vti4_a    "${ns_b}" vti4_b
1461
1462	veth_mtu=1500
1463	vti_mtu=$((veth_mtu - 20))
1464
1465	#                                SPI   SN   IV  ICV   pad length   next header
1466	esp_payload_rfc4106=$((vti_mtu - 4   - 4  - 8 - 16  - 1          - 1))
1467	ping_payload=$((esp_payload_rfc4106 - 28))
1468
1469	mtu "${ns_a}" veth_a ${veth_mtu}
1470	mtu "${ns_b}" veth_b ${veth_mtu}
1471	mtu "${ns_a}" vti4_a ${vti_mtu}
1472	mtu "${ns_b}" vti4_b ${vti_mtu}
1473
1474	# Send DF packet without exceeding link layer MTU, check that no
1475	# exception is created
1476	run_cmd ${ns_a} ping -q -M want -i 0.1 -w 1 -s ${ping_payload} ${tunnel4_b_addr}
1477	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel4_b_addr})"
1478	check_pmtu_value "" "${pmtu}" "sending packet smaller than PMTU (IP payload length ${esp_payload_rfc4106})" || return 1
1479
1480	# Now exceed link layer MTU by one byte, check that exception is created
1481	# with the right PMTU value
1482	run_cmd ${ns_a} ping -q -M want -i 0.1 -w 1 -s $((ping_payload + 1)) ${tunnel4_b_addr}
1483	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel4_b_addr})"
1484	check_pmtu_value "${esp_payload_rfc4106}" "${pmtu}" "exceeding PMTU (IP payload length $((esp_payload_rfc4106 + 1)))"
1485}
1486
1487test_pmtu_vti6_exception() {
1488	setup namespaces veth vti6 xfrm6 || return $ksft_skip
1489	trace "${ns_a}" veth_a    "${ns_b}" veth_b \
1490	      "${ns_a}" vti6_a    "${ns_b}" vti6_b
1491	fail=0
1492
1493	# Create route exception by exceeding link layer MTU
1494	mtu "${ns_a}" veth_a 4000
1495	mtu "${ns_b}" veth_b 4000
1496	mtu "${ns_a}" vti6_a 5000
1497	mtu "${ns_b}" vti6_b 5000
1498	run_cmd ${ns_a} ${ping6} -q -i 0.1 -w 1 -s 60000 ${tunnel6_b_addr}
1499
1500	# Check that exception was created
1501	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel6_b_addr})"
1502	check_pmtu_value any "${pmtu}" "creating tunnel exceeding link layer MTU" || return 1
1503
1504	# Decrease tunnel MTU, check for PMTU decrease in route exception
1505	mtu "${ns_a}" vti6_a 3000
1506	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel6_b_addr})"
1507	check_pmtu_value "3000" "${pmtu}" "decreasing tunnel MTU" || fail=1
1508
1509	# Increase tunnel MTU, check for PMTU increase in route exception
1510	mtu "${ns_a}" vti6_a 9000
1511	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel6_b_addr})"
1512	check_pmtu_value "9000" "${pmtu}" "increasing tunnel MTU" || fail=1
1513
1514	return ${fail}
1515}
1516
1517test_pmtu_vti4_udp_exception() {
1518	setup namespaces veth vti4 xfrm4udp || return $ksft_skip
1519	trace "${ns_a}" veth_a    "${ns_b}" veth_b \
1520	      "${ns_a}" vti4_a    "${ns_b}" vti4_b
1521
1522	veth_mtu=1500
1523	vti_mtu=$((veth_mtu - 20))
1524
1525	#                                UDP   SPI   SN   IV  ICV   pad length   next header
1526	esp_payload_rfc4106=$((vti_mtu - 8   - 4   - 4  - 8 - 16  - 1          - 1))
1527	ping_payload=$((esp_payload_rfc4106 - 28))
1528
1529	mtu "${ns_a}" veth_a ${veth_mtu}
1530	mtu "${ns_b}" veth_b ${veth_mtu}
1531	mtu "${ns_a}" vti4_a ${vti_mtu}
1532	mtu "${ns_b}" vti4_b ${vti_mtu}
1533
1534	# Send DF packet without exceeding link layer MTU, check that no
1535	# exception is created
1536	run_cmd ${ns_a} ping -q -M want -i 0.1 -w 1 -s ${ping_payload} ${tunnel4_b_addr}
1537	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel4_b_addr})"
1538	check_pmtu_value "" "${pmtu}" "sending packet smaller than PMTU (IP payload length ${esp_payload_rfc4106})" || return 1
1539
1540	# Now exceed link layer MTU by one byte, check that exception is created
1541	# with the right PMTU value
1542	run_cmd ${ns_a} ping -q -M want -i 0.1 -w 1 -s $((ping_payload + 1)) ${tunnel4_b_addr}
1543	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel4_b_addr})"
1544	check_pmtu_value "${esp_payload_rfc4106}" "${pmtu}" "exceeding PMTU (IP payload length $((esp_payload_rfc4106 + 1)))"
1545}
1546
1547test_pmtu_vti6_udp_exception() {
1548	setup namespaces veth vti6 xfrm6udp || return $ksft_skip
1549	trace "${ns_a}" veth_a    "${ns_b}" veth_b \
1550	      "${ns_a}" vti6_a    "${ns_b}" vti6_b
1551	fail=0
1552
1553	# Create route exception by exceeding link layer MTU
1554	mtu "${ns_a}" veth_a 4000
1555	mtu "${ns_b}" veth_b 4000
1556	mtu "${ns_a}" vti6_a 5000
1557	mtu "${ns_b}" vti6_b 5000
1558	run_cmd ${ns_a} ${ping6} -q -i 0.1 -w 1 -s 60000 ${tunnel6_b_addr}
1559
1560	# Check that exception was created
1561	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel6_b_addr})"
1562	check_pmtu_value any "${pmtu}" "creating tunnel exceeding link layer MTU" || return 1
1563
1564	# Decrease tunnel MTU, check for PMTU decrease in route exception
1565	mtu "${ns_a}" vti6_a 3000
1566	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel6_b_addr})"
1567	check_pmtu_value "3000" "${pmtu}" "decreasing tunnel MTU" || fail=1
1568
1569	# Increase tunnel MTU, check for PMTU increase in route exception
1570	mtu "${ns_a}" vti6_a 9000
1571	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel6_b_addr})"
1572	check_pmtu_value "9000" "${pmtu}" "increasing tunnel MTU" || fail=1
1573
1574	return ${fail}
1575}
1576
1577test_pmtu_vti4_udp_routed_exception() {
1578	setup namespaces routing vti4routed xfrm4udprouted || return $ksft_skip
1579	trace "${ns_a}" veth_A-R1    "${ns_b}" veth_B-R1 \
1580	      "${ns_a}" vti4_a       "${ns_b}" vti4_b
1581
1582	veth_mtu=1500
1583	vti_mtu=$((veth_mtu - 20))
1584
1585	#                                UDP   SPI   SN   IV  ICV   pad length   next header
1586	esp_payload_rfc4106=$((vti_mtu - 8   - 4   - 4  - 8 - 16  - 1          - 1))
1587	ping_payload=$((esp_payload_rfc4106 - 28))
1588
1589        mtu "${ns_a}"  veth_A-R1 ${veth_mtu}
1590        mtu "${ns_r1}" veth_R1-A ${veth_mtu}
1591        mtu "${ns_b}"  veth_B-R1 ${veth_mtu}
1592        mtu "${ns_r1}" veth_R1-B ${veth_mtu}
1593
1594	mtu "${ns_a}" vti4_a ${vti_mtu}
1595	mtu "${ns_b}" vti4_b ${vti_mtu}
1596
1597	# Send DF packet without exceeding link layer MTU, check that no
1598	# exception is created
1599	run_cmd ${ns_a} ping -q -M want -i 0.1 -w 1 -s ${ping_payload} ${tunnel4_b_addr}
1600	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel4_b_addr})"
1601	check_pmtu_value "" "${pmtu}" "sending packet smaller than PMTU (IP payload length ${esp_payload_rfc4106})" || return 1
1602
1603	# Now decrease link layer MTU by 8 bytes on R1, check that exception is created
1604	# with the right PMTU value
1605        mtu "${ns_r1}" veth_R1-B $((veth_mtu - 8))
1606	run_cmd ${ns_a} ping -q -M want -i 0.1 -w 1 -s $((ping_payload)) ${tunnel4_b_addr}
1607	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel4_b_addr})"
1608	check_pmtu_value "$((esp_payload_rfc4106 - 8))" "${pmtu}" "exceeding PMTU (IP payload length $((esp_payload_rfc4106)))"
1609}
1610
1611test_pmtu_vti6_udp_routed_exception() {
1612	setup namespaces routing vti6routed xfrm6udprouted || return $ksft_skip
1613	trace "${ns_a}" veth_A-R1    "${ns_b}" veth_B-R1 \
1614	      "${ns_a}" vti6_a       "${ns_b}" vti6_b
1615
1616	veth_mtu=1500
1617	vti_mtu=$((veth_mtu - 40))
1618
1619	#                                UDP   SPI   SN   IV  ICV   pad length   next header
1620	esp_payload_rfc4106=$((vti_mtu - 8   - 4   - 4  - 8 - 16  - 1          - 1))
1621	ping_payload=$((esp_payload_rfc4106 - 48))
1622
1623        mtu "${ns_a}"  veth_A-R1 ${veth_mtu}
1624        mtu "${ns_r1}" veth_R1-A ${veth_mtu}
1625        mtu "${ns_b}"  veth_B-R1 ${veth_mtu}
1626        mtu "${ns_r1}" veth_R1-B ${veth_mtu}
1627
1628	# mtu "${ns_a}" vti6_a ${vti_mtu}
1629	# mtu "${ns_b}" vti6_b ${vti_mtu}
1630
1631	run_cmd ${ns_a} ${ping6} -q -M want -i 0.1 -w 1 -s ${ping_payload} ${tunnel6_b_addr}
1632
1633	# Check that exception was not created
1634	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel6_b_addr})"
1635	check_pmtu_value "" "${pmtu}" "sending packet smaller than PMTU (IP payload length ${esp_payload_rfc4106})" || return 1
1636
1637	# Now decrease link layer MTU by 8 bytes on R1, check that exception is created
1638	# with the right PMTU value
1639        mtu "${ns_r1}" veth_R1-B $((veth_mtu - 8))
1640	run_cmd ${ns_a} ${ping6} -q -M want -i 0.1 -w 1 -s $((ping_payload)) ${tunnel6_b_addr}
1641	pmtu="$(route_get_dst_pmtu_from_exception "${ns_a}" ${tunnel6_b_addr})"
1642	check_pmtu_value "$((esp_payload_rfc4106 - 8))" "${pmtu}" "exceeding PMTU (IP payload length $((esp_payload_rfc4106)))"
1643
1644}
1645
1646test_pmtu_vti4_default_mtu() {
1647	setup namespaces veth vti4 || return $ksft_skip
1648
1649	# Check that MTU of vti device is MTU of veth minus IPv4 header length
1650	veth_mtu="$(link_get_mtu "${ns_a}" veth_a)"
1651	vti4_mtu="$(link_get_mtu "${ns_a}" vti4_a)"
1652	if [ $((veth_mtu - vti4_mtu)) -ne 20 ]; then
1653		err "  vti MTU ${vti4_mtu} is not veth MTU ${veth_mtu} minus IPv4 header length"
1654		return 1
1655	fi
1656}
1657
1658test_pmtu_vti6_default_mtu() {
1659	setup namespaces veth vti6 || return $ksft_skip
1660
1661	# Check that MTU of vti device is MTU of veth minus IPv6 header length
1662	veth_mtu="$(link_get_mtu "${ns_a}" veth_a)"
1663	vti6_mtu="$(link_get_mtu "${ns_a}" vti6_a)"
1664	if [ $((veth_mtu - vti6_mtu)) -ne 40 ]; then
1665		err "  vti MTU ${vti6_mtu} is not veth MTU ${veth_mtu} minus IPv6 header length"
1666		return 1
1667	fi
1668}
1669
1670test_pmtu_vti4_link_add_mtu() {
1671	setup namespaces || return $ksft_skip
1672
1673	run_cmd ${ns_a} ip link add vti4_a type vti local ${veth4_a_addr} remote ${veth4_b_addr} key 10
1674	[ $? -ne 0 ] && err "  vti not supported" && return $ksft_skip
1675	run_cmd ${ns_a} ip link del vti4_a
1676
1677	fail=0
1678
1679	min=68
1680	max=$((65535 - 20))
1681	# Check invalid values first
1682	for v in $((min - 1)) $((max + 1)); do
1683		run_cmd ${ns_a} ip link add vti4_a mtu ${v} type vti local ${veth4_a_addr} remote ${veth4_b_addr} key 10
1684		# This can fail, or MTU can be adjusted to a proper value
1685		[ $? -ne 0 ] && continue
1686		mtu="$(link_get_mtu "${ns_a}" vti4_a)"
1687		if [ ${mtu} -lt ${min} -o ${mtu} -gt ${max} ]; then
1688			err "  vti tunnel created with invalid MTU ${mtu}"
1689			fail=1
1690		fi
1691		run_cmd ${ns_a} ip link del vti4_a
1692	done
1693
1694	# Now check valid values
1695	for v in ${min} 1300 ${max}; do
1696		run_cmd ${ns_a} ip link add vti4_a mtu ${v} type vti local ${veth4_a_addr} remote ${veth4_b_addr} key 10
1697		mtu="$(link_get_mtu "${ns_a}" vti4_a)"
1698		run_cmd ${ns_a} ip link del vti4_a
1699		if [ "${mtu}" != "${v}" ]; then
1700			err "  vti MTU ${mtu} doesn't match configured value ${v}"
1701			fail=1
1702		fi
1703	done
1704
1705	return ${fail}
1706}
1707
1708test_pmtu_vti6_link_add_mtu() {
1709	setup namespaces || return $ksft_skip
1710
1711	run_cmd ${ns_a} ip link add vti6_a type vti6 local ${veth6_a_addr} remote ${veth6_b_addr} key 10
1712	[ $? -ne 0 ] && err "  vti6 not supported" && return $ksft_skip
1713	run_cmd ${ns_a} ip link del vti6_a
1714
1715	fail=0
1716
1717	min=68			# vti6 can carry IPv4 packets too
1718	max=$((65535 - 40))
1719	# Check invalid values first
1720	for v in $((min - 1)) $((max + 1)); do
1721		run_cmd ${ns_a} ip link add vti6_a mtu ${v} type vti6 local ${veth6_a_addr} remote ${veth6_b_addr} key 10
1722		# This can fail, or MTU can be adjusted to a proper value
1723		[ $? -ne 0 ] && continue
1724		mtu="$(link_get_mtu "${ns_a}" vti6_a)"
1725		if [ ${mtu} -lt ${min} -o ${mtu} -gt ${max} ]; then
1726			err "  vti6 tunnel created with invalid MTU ${v}"
1727			fail=1
1728		fi
1729		run_cmd ${ns_a} ip link del vti6_a
1730	done
1731
1732	# Now check valid values
1733	for v in 68 1280 1300 $((65535 - 40)); do
1734		run_cmd ${ns_a} ip link add vti6_a mtu ${v} type vti6 local ${veth6_a_addr} remote ${veth6_b_addr} key 10
1735		mtu="$(link_get_mtu "${ns_a}" vti6_a)"
1736		run_cmd ${ns_a} ip link del vti6_a
1737		if [ "${mtu}" != "${v}" ]; then
1738			err "  vti6 MTU ${mtu} doesn't match configured value ${v}"
1739			fail=1
1740		fi
1741	done
1742
1743	return ${fail}
1744}
1745
1746test_pmtu_vti6_link_change_mtu() {
1747	setup namespaces || return $ksft_skip
1748
1749	run_cmd ${ns_a} ip link add dummy0 mtu 1500 type dummy
1750	[ $? -ne 0 ] && err "  dummy not supported" && return $ksft_skip
1751	run_cmd ${ns_a} ip link add dummy1 mtu 3000 type dummy
1752	run_cmd ${ns_a} ip link set dummy0 up
1753	run_cmd ${ns_a} ip link set dummy1 up
1754
1755	run_cmd ${ns_a} ip addr add ${dummy6_0_prefix}1/${dummy6_mask} dev dummy0
1756	run_cmd ${ns_a} ip addr add ${dummy6_1_prefix}1/${dummy6_mask} dev dummy1
1757
1758	fail=0
1759
1760	# Create vti6 interface bound to device, passing MTU, check it
1761	run_cmd ${ns_a} ip link add vti6_a mtu 1300 type vti6 remote ${dummy6_0_prefix}2 local ${dummy6_0_prefix}1
1762	mtu="$(link_get_mtu "${ns_a}" vti6_a)"
1763	if [ ${mtu} -ne 1300 ]; then
1764		err "  vti6 MTU ${mtu} doesn't match configured value 1300"
1765		fail=1
1766	fi
1767
1768	# Move to another device with different MTU, without passing MTU, check
1769	# MTU is adjusted
1770	run_cmd ${ns_a} ip link set vti6_a type vti6 remote ${dummy6_1_prefix}2 local ${dummy6_1_prefix}1
1771	mtu="$(link_get_mtu "${ns_a}" vti6_a)"
1772	if [ ${mtu} -ne $((3000 - 40)) ]; then
1773		err "  vti MTU ${mtu} is not dummy MTU 3000 minus IPv6 header length"
1774		fail=1
1775	fi
1776
1777	# Move it back, passing MTU, check MTU is not overridden
1778	run_cmd ${ns_a} ip link set vti6_a mtu 1280 type vti6 remote ${dummy6_0_prefix}2 local ${dummy6_0_prefix}1
1779	mtu="$(link_get_mtu "${ns_a}" vti6_a)"
1780	if [ ${mtu} -ne 1280 ]; then
1781		err "  vti6 MTU ${mtu} doesn't match configured value 1280"
1782		fail=1
1783	fi
1784
1785	return ${fail}
1786}
1787
1788check_command() {
1789	cmd=${1}
1790
1791	if ! which ${cmd} > /dev/null 2>&1; then
1792		err "  missing required command: '${cmd}'"
1793		return 1
1794	fi
1795	return 0
1796}
1797
1798check_running() {
1799	pid=${1}
1800	cmd=${2}
1801
1802	[ "$(cat /proc/${pid}/cmdline 2>/dev/null | tr -d '\0')" = "{cmd}" ]
1803}
1804
1805test_cleanup_vxlanX_exception() {
1806	outer="${1}"
1807	encap="vxlan"
1808	ll_mtu=4000
1809
1810	check_command taskset || return $ksft_skip
1811	cpu_list=$(grep -m 2 processor /proc/cpuinfo | cut -d ' ' -f 2)
1812
1813	setup namespaces routing ${encap}${outer} || return $ksft_skip
1814	trace "${ns_a}" ${encap}_a   "${ns_b}"  ${encap}_b \
1815	      "${ns_a}" veth_A-R1    "${ns_r1}" veth_R1-A \
1816	      "${ns_b}" veth_B-R1    "${ns_r1}" veth_R1-B
1817
1818	# Create route exception by exceeding link layer MTU
1819	mtu "${ns_a}"  veth_A-R1 $((${ll_mtu} + 1000))
1820	mtu "${ns_r1}" veth_R1-A $((${ll_mtu} + 1000))
1821	mtu "${ns_b}"  veth_B-R1 ${ll_mtu}
1822	mtu "${ns_r1}" veth_R1-B ${ll_mtu}
1823
1824	mtu "${ns_a}" ${encap}_a $((${ll_mtu} + 1000))
1825	mtu "${ns_b}" ${encap}_b $((${ll_mtu} + 1000))
1826
1827	# Fill exception cache for multiple CPUs (2)
1828	# we can always use inner IPv4 for that
1829	for cpu in ${cpu_list}; do
1830		run_cmd taskset --cpu-list ${cpu} ${ns_a} ping -q -M want -i 0.1 -w 1 -s $((${ll_mtu} + 500)) ${tunnel4_b_addr}
1831	done
1832
1833	${ns_a} ip link del dev veth_A-R1 &
1834	iplink_pid=$!
1835	for i in $(seq 1 20); do
1836		check_running ${iplink_pid} "iplinkdeldevveth_A-R1" || return 0
1837		sleep 0.1
1838	done
1839	err "  can't delete veth device in a timely manner, PMTU dst likely leaked"
1840	return 1
1841}
1842
1843test_cleanup_ipv6_exception() {
1844	test_cleanup_vxlanX_exception 6
1845}
1846
1847test_cleanup_ipv4_exception() {
1848	test_cleanup_vxlanX_exception 4
1849}
1850
1851run_test() {
1852	(
1853	tname="$1"
1854	tdesc="$2"
1855
1856	unset IFS
1857
1858	# Since cleanup() relies on variables modified by this subshell, it
1859	# has to run in this context.
1860	trap cleanup EXIT
1861
1862	if [ "$VERBOSE" = "1" ]; then
1863		printf "\n##########################################################################\n\n"
1864	fi
1865
1866	eval test_${tname}
1867	ret=$?
1868
1869	if [ $ret -eq 0 ]; then
1870		printf "TEST: %-60s  [ OK ]\n" "${tdesc}"
1871	elif [ $ret -eq 1 ]; then
1872		printf "TEST: %-60s  [FAIL]\n" "${tdesc}"
1873		if [ "${PAUSE_ON_FAIL}" = "yes" ]; then
1874			echo
1875			echo "Pausing. Hit enter to continue"
1876			read a
1877		fi
1878		err_flush
1879		exit 1
1880	elif [ $ret -eq $ksft_skip ]; then
1881		printf "TEST: %-60s  [SKIP]\n" "${tdesc}"
1882		err_flush
1883	fi
1884
1885	return $ret
1886	)
1887	ret=$?
1888	case $ret in
1889		0)
1890			all_skipped=false
1891			[ $exitcode -eq $ksft_skip ] && exitcode=0
1892		;;
1893		$ksft_skip)
1894			[ $all_skipped = true ] && exitcode=$ksft_skip
1895		;;
1896		*)
1897			all_skipped=false
1898			exitcode=1
1899		;;
1900	esac
1901
1902	return $ret
1903}
1904
1905run_test_nh() {
1906	tname="$1"
1907	tdesc="$2"
1908
1909	USE_NH=yes
1910	run_test "${tname}" "${tdesc} - nexthop objects"
1911	USE_NH=no
1912}
1913
1914test_list_flush_ipv4_exception() {
1915	setup namespaces routing || return $ksft_skip
1916	trace "${ns_a}"  veth_A-R1    "${ns_r1}" veth_R1-A \
1917	      "${ns_r1}" veth_R1-B    "${ns_b}"  veth_B-R1 \
1918	      "${ns_a}"  veth_A-R2    "${ns_r2}" veth_R2-A \
1919	      "${ns_r2}" veth_R2-B    "${ns_b}"  veth_B-R2
1920
1921	dst_prefix1="${prefix4}.${b_r1}."
1922	dst2="${prefix4}.${b_r2}.1"
1923
1924	# Set up initial MTU values
1925	mtu "${ns_a}"  veth_A-R1 2000
1926	mtu "${ns_r1}" veth_R1-A 2000
1927	mtu "${ns_r1}" veth_R1-B 1500
1928	mtu "${ns_b}"  veth_B-R1 1500
1929
1930	mtu "${ns_a}"  veth_A-R2 2000
1931	mtu "${ns_r2}" veth_R2-A 2000
1932	mtu "${ns_r2}" veth_R2-B 1500
1933	mtu "${ns_b}"  veth_B-R2 1500
1934
1935	fail=0
1936
1937	# Add 100 addresses for veth endpoint on B reached by default A route
1938	for i in $(seq 100 199); do
1939		run_cmd ${ns_b} ip addr add "${dst_prefix1}${i}" dev veth_B-R1
1940	done
1941
1942	# Create 100 cached route exceptions for path via R1, one via R2. Note
1943	# that with IPv4 we need to actually cause a route lookup that matches
1944	# the exception caused by ICMP, in order to actually have a cached
1945	# route, so we need to ping each destination twice
1946	for i in $(seq 100 199); do
1947		run_cmd ${ns_a} ping -q -M want -i 0.1 -c 2 -s 1800 "${dst_prefix1}${i}"
1948	done
1949	run_cmd ${ns_a} ping -q -M want -i 0.1 -c 2 -s 1800 "${dst2}"
1950
1951	if [ "$(${ns_a} ip -oneline route list cache | wc -l)" -ne 101 ]; then
1952		err "  can't list cached exceptions"
1953		fail=1
1954	fi
1955
1956	run_cmd ${ns_a} ip route flush cache
1957	pmtu1="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst_prefix}1)"
1958	pmtu2="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst_prefix}2)"
1959	if [ -n "${pmtu1}" ] || [ -n "${pmtu2}" ] || \
1960	   [ -n "$(${ns_a} ip route list cache)" ]; then
1961		err "  can't flush cached exceptions"
1962		fail=1
1963	fi
1964
1965	return ${fail}
1966}
1967
1968test_list_flush_ipv6_exception() {
1969	setup namespaces routing || return $ksft_skip
1970	trace "${ns_a}"  veth_A-R1    "${ns_r1}" veth_R1-A \
1971	      "${ns_r1}" veth_R1-B    "${ns_b}"  veth_B-R1 \
1972	      "${ns_a}"  veth_A-R2    "${ns_r2}" veth_R2-A \
1973	      "${ns_r2}" veth_R2-B    "${ns_b}"  veth_B-R2
1974
1975	dst_prefix1="${prefix6}:${b_r1}::"
1976	dst2="${prefix6}:${b_r2}::1"
1977
1978	# Set up initial MTU values
1979	mtu "${ns_a}"  veth_A-R1 2000
1980	mtu "${ns_r1}" veth_R1-A 2000
1981	mtu "${ns_r1}" veth_R1-B 1500
1982	mtu "${ns_b}"  veth_B-R1 1500
1983
1984	mtu "${ns_a}"  veth_A-R2 2000
1985	mtu "${ns_r2}" veth_R2-A 2000
1986	mtu "${ns_r2}" veth_R2-B 1500
1987	mtu "${ns_b}"  veth_B-R2 1500
1988
1989	fail=0
1990
1991	# Add 100 addresses for veth endpoint on B reached by default A route
1992	for i in $(seq 100 199); do
1993		run_cmd ${ns_b} ip addr add "${dst_prefix1}${i}" dev veth_B-R1
1994	done
1995
1996	# Create 100 cached route exceptions for path via R1, one via R2
1997	for i in $(seq 100 199); do
1998		run_cmd ${ns_a} ping -q -M want -i 0.1 -w 1 -s 1800 "${dst_prefix1}${i}"
1999	done
2000	run_cmd ${ns_a} ping -q -M want -i 0.1 -w 1 -s 1800 "${dst2}"
2001	if [ "$(${ns_a} ip -oneline -6 route list cache | wc -l)" -ne 101 ]; then
2002		err "  can't list cached exceptions"
2003		fail=1
2004	fi
2005
2006	run_cmd ${ns_a} ip -6 route flush cache
2007	pmtu1="$(route_get_dst_pmtu_from_exception "${ns_a}" "${dst_prefix1}100")"
2008	pmtu2="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst2})"
2009	if [ -n "${pmtu1}" ] || [ -n "${pmtu2}" ] || \
2010	   [ -n "$(${ns_a} ip -6 route list cache)" ]; then
2011		err "  can't flush cached exceptions"
2012		fail=1
2013	fi
2014
2015	return ${fail}
2016}
2017
2018test_pmtu_ipvX_route_change() {
2019	family=${1}
2020
2021	setup namespaces routing || return 2
2022	trace "${ns_a}"  veth_A-R1    "${ns_r1}" veth_R1-A \
2023	      "${ns_r1}" veth_R1-B    "${ns_b}"  veth_B-R1 \
2024	      "${ns_a}"  veth_A-R2    "${ns_r2}" veth_R2-A \
2025	      "${ns_r2}" veth_R2-B    "${ns_b}"  veth_B-R2
2026
2027	if [ ${family} -eq 4 ]; then
2028		ping=ping
2029		dst1="${prefix4}.${b_r1}.1"
2030		dst2="${prefix4}.${b_r2}.1"
2031		gw="${prefix4}.${a_r1}.2"
2032	else
2033		ping=${ping6}
2034		dst1="${prefix6}:${b_r1}::1"
2035		dst2="${prefix6}:${b_r2}::1"
2036		gw="${prefix6}:${a_r1}::2"
2037	fi
2038
2039	# Set up initial MTU values
2040	mtu "${ns_a}"  veth_A-R1 2000
2041	mtu "${ns_r1}" veth_R1-A 2000
2042	mtu "${ns_r1}" veth_R1-B 1400
2043	mtu "${ns_b}"  veth_B-R1 1400
2044
2045	mtu "${ns_a}"  veth_A-R2 2000
2046	mtu "${ns_r2}" veth_R2-A 2000
2047	mtu "${ns_r2}" veth_R2-B 1500
2048	mtu "${ns_b}"  veth_B-R2 1500
2049
2050	# Create route exceptions
2051	run_cmd ${ns_a} ${ping} -q -M want -i 0.1 -w 1 -s 1800 ${dst1}
2052	run_cmd ${ns_a} ${ping} -q -M want -i 0.1 -w 1 -s 1800 ${dst2}
2053
2054	# Check that exceptions have been created with the correct PMTU
2055	pmtu_1="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst1})"
2056	check_pmtu_value "1400" "${pmtu_1}" "exceeding MTU" || return 1
2057	pmtu_2="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst2})"
2058	check_pmtu_value "1500" "${pmtu_2}" "exceeding MTU" || return 1
2059
2060	# Replace the route from A to R1
2061	run_cmd ${ns_a} ip route change default via ${gw}
2062
2063	# Delete the device in A
2064	run_cmd ${ns_a} ip link del "veth_A-R1"
2065}
2066
2067test_pmtu_ipv4_route_change() {
2068	test_pmtu_ipvX_route_change 4
2069}
2070
2071test_pmtu_ipv6_route_change() {
2072	test_pmtu_ipvX_route_change 6
2073}
2074
2075usage() {
2076	echo
2077	echo "$0 [OPTIONS] [TEST]..."
2078	echo "If no TEST argument is given, all tests will be run."
2079	echo
2080	echo "Options"
2081	echo "  --trace: capture traffic to TEST_INTERFACE.pcap"
2082	echo
2083	echo "Available tests${tests}"
2084	exit 1
2085}
2086
2087################################################################################
2088#
2089exitcode=0
2090desc=0
2091all_skipped=true
2092
2093while getopts :ptv o
2094do
2095	case $o in
2096	p) PAUSE_ON_FAIL=yes;;
2097	v) VERBOSE=1;;
2098	t) if which tcpdump > /dev/null 2>&1; then
2099		TRACING=1
2100	   else
2101		echo "=== tcpdump not available, tracing disabled"
2102	   fi
2103	   ;;
2104	*) usage;;
2105	esac
2106done
2107shift $(($OPTIND-1))
2108
2109IFS="
2110"
2111
2112for arg do
2113	# Check first that all requested tests are available before running any
2114	command -v > /dev/null "test_${arg}" || { echo "=== Test ${arg} not found"; usage; }
2115done
2116
2117trap cleanup EXIT
2118
2119# start clean
2120cleanup
2121
2122HAVE_NH=no
2123ip nexthop ls >/dev/null 2>&1
2124[ $? -eq 0 ] && HAVE_NH=yes
2125
2126name=""
2127desc=""
2128rerun_nh=0
2129for t in ${tests}; do
2130	[ "${name}" = "" ]	&& name="${t}"	&& continue
2131	[ "${desc}" = "" ]	&& desc="${t}"	&& continue
2132
2133	if [ "${HAVE_NH}" = "yes" ]; then
2134		rerun_nh="${t}"
2135	fi
2136
2137	run_this=1
2138	for arg do
2139		[ "${arg}" != "${arg#--*}" ] && continue
2140		[ "${arg}" = "${name}" ] && run_this=1 && break
2141		run_this=0
2142	done
2143	if [ $run_this -eq 1 ]; then
2144		run_test "${name}" "${desc}"
2145		# if test was skipped no need to retry with nexthop objects
2146		[ $? -eq $ksft_skip ] && rerun_nh=0
2147
2148		if [ "${rerun_nh}" = "1" ]; then
2149			run_test_nh "${name}" "${desc}"
2150		fi
2151	fi
2152	name=""
2153	desc=""
2154	rerun_nh=0
2155done
2156
2157exit ${exitcode}
2158