xref: /linux/tools/testing/selftests/net/openvswitch/openvswitch.sh (revision d603517771d8e08a2d8fc9e1f7682ce393d3973a)
1#!/bin/bash
2# SPDX-License-Identifier: GPL-2.0
3#
4# OVS kernel module self tests
5
6trap ovs_exit_sig EXIT TERM INT ERR
7
8# Kselftest framework requirement - SKIP code is 4.
9ksft_skip=4
10
11PAUSE_ON_FAIL=no
12VERBOSE=0
13TRACING=0
14WAIT_TIMEOUT=5
15
16if test "X$KSFT_MACHINE_SLOW" == "Xyes"; then
17	WAIT_TIMEOUT=10
18fi
19
20tests="
21	arp_ping				eth-arp: Basic arp ping between two NS
22	ct_connect_v4				ip4-ct-xon: Basic ipv4 tcp connection using ct
23	connect_v4				ip4-xon: Basic ipv4 ping between two NS
24	nat_connect_v4				ip4-nat-xon: Basic ipv4 tcp connection via NAT
25	nat_related_v4				ip4-nat-related: ICMP related matches work with SNAT
26	netlink_checks				ovsnl: validate netlink attrs and settings
27	upcall_interfaces			ovs: test the upcall interfaces
28	tunnel_metadata				ovs: test extraction of tunnel metadata
29	tunnel_refcount				ovs: test tunnel vport reference cleanup
30	drop_reason				drop: test drop reasons are emitted
31	pop_vlan				vlan: POP_VLAN action strips tag
32	psample					psample: Sampling packets with psample"
33
34info() {
35	[ "${ovs_dir}" != "" ] &&
36		echo "`date +"[%m-%d %H:%M:%S]"` $*" >> ${ovs_dir}/debug.log
37	[ $VERBOSE = 0 ] || echo $*
38}
39
40ovs_wait() {
41	info "waiting $WAIT_TIMEOUT s for: $@"
42
43	if "$@" ; then
44		info "wait succeeded immediately"
45		return 0
46	fi
47
48	# A quick re-check helps speed up small races in fast systems.
49	# However, fractional sleeps might not necessarily work.
50	local start=0
51	sleep 0.1 || { sleep 1; start=1; }
52
53	for (( i=start; i<WAIT_TIMEOUT; i++ )); do
54		if "$@" ; then
55			info "wait succeeded after $i seconds"
56			return 0
57		fi
58		sleep 1
59	done
60	info "wait failed after $i seconds"
61	return 1
62}
63
64ovs_base=`pwd`
65sbxs=
66sbx_add () {
67	info "adding sandbox '$1'"
68
69	sbxs="$sbxs $1"
70
71	NO_BIN=0
72
73	# Create sandbox.
74	local d="$ovs_base"/$1
75	if [ -e $d ]; then
76		info "removing $d"
77		rm -rf "$d"
78	fi
79	mkdir "$d" || return 1
80	ovs_setenv $1
81}
82
83ovs_exit_sig() {
84	[ -e ${ovs_dir}/cleanup ] && . "$ovs_dir/cleanup"
85}
86
87on_exit() {
88	echo "$1" > ${ovs_dir}/cleanup.tmp
89	cat ${ovs_dir}/cleanup >> ${ovs_dir}/cleanup.tmp
90	mv ${ovs_dir}/cleanup.tmp ${ovs_dir}/cleanup
91}
92
93ovs_setenv() {
94	sandbox=$1
95
96	ovs_dir=$ovs_base${1:+/$1}; export ovs_dir
97
98	test -e ${ovs_dir}/cleanup || : > ${ovs_dir}/cleanup
99}
100
101ovs_sbx() {
102	if test "X$2" != X; then
103		(ovs_setenv $1; shift;
104		 info "run cmd: $@"; "$@" >> ${ovs_dir}/debug.log)
105	else
106		ovs_setenv $1
107	fi
108}
109
110ovs_add_dp () {
111	info "Adding DP/Bridge IF: sbx:$1 dp:$2 {$3, $4, $5}"
112	sbxname="$1"
113	shift
114	ovs_sbx "$sbxname" python3 $ovs_base/ovs-dpctl.py add-dp $*
115	on_exit "ovs_sbx $sbxname python3 $ovs_base/ovs-dpctl.py del-dp $1;"
116}
117
118ovs_add_if () {
119	info "Adding IF to DP: br:$3 if:$4 ($2)"
120	if [ "$5" != "-u" ]; then
121		ovs_sbx "$1" python3 $ovs_base/ovs-dpctl.py add-if \
122		    -t "$2" "$3" "$4" || return 1
123	else
124		python3 $ovs_base/ovs-dpctl.py add-if \
125		    -u -t "$2" "$3" "$4" >$ovs_dir/$4.out 2>$ovs_dir/$4.err &
126		pid=$!
127		on_exit "ovs_sbx $1 kill -TERM $pid 2>/dev/null"
128	fi
129}
130
131ovs_del_if () {
132	info "Deleting IF from DP: br:$2 if:$3"
133	ovs_sbx "$1" python3 $ovs_base/ovs-dpctl.py del-if "$2" "$3" || return 1
134}
135
136ovs_netns_spawn_daemon() {
137	sbx=$1
138	shift
139	netns=$1
140	shift
141	if [ "$netns" == "_default" ]; then
142		$*  >> $ovs_dir/stdout  2>> $ovs_dir/stderr &
143	else
144		ip netns exec $netns $*  >> $ovs_dir/stdout  2>> $ovs_dir/stderr &
145	fi
146	pid=$!
147	ovs_sbx "$sbx" on_exit "kill -TERM $pid 2>/dev/null"
148}
149
150ovs_spawn_daemon() {
151	sbx=$1
152	shift
153	ovs_netns_spawn_daemon $sbx "_default" $*
154}
155
156ovs_add_netns_and_veths () {
157	info "Adding netns attached: sbx:$1 dp:$2 {$3, $4, $5}"
158	ovs_sbx "$1" ip netns add "$3" || return 1
159	on_exit "ovs_sbx $1 ip netns del $3"
160	ovs_sbx "$1" ip link add "$4" type veth peer name "$5" || return 1
161	on_exit "ovs_sbx $1 ip link del $4 >/dev/null 2>&1"
162	ovs_sbx "$1" ip link set "$4" up || return 1
163	ovs_sbx "$1" ip link set "$5" netns "$3" || return 1
164	ovs_sbx "$1" ip netns exec "$3" ip link set "$5" up || return 1
165
166	if [ "$6" != "" ]; then
167		ovs_sbx "$1" ip netns exec "$3" ip addr add "$6" dev "$5" \
168		    || return 1
169	fi
170
171	if [ "$7" != "-u" ]; then
172		ovs_add_if "$1" "netdev" "$2" "$4" || return 1
173	else
174		ovs_add_if "$1" "netdev" "$2" "$4" -u || return 1
175	fi
176
177	if [ $TRACING -eq 1 ]; then
178		ovs_netns_spawn_daemon "$1" "$3" tcpdump -l -i any -s 6553
179		ovs_wait grep -q "listening on any" ${ovs_dir}/stderr
180	fi
181
182	return 0
183}
184
185ovs_add_flow () {
186	info "Adding flow to DP: sbx:$1 br:$2 flow:$3 act:$4"
187	ovs_sbx "$1" python3 $ovs_base/ovs-dpctl.py add-flow "$2" "$3" "$4"
188	if [ $? -ne 0 ]; then
189		info "Flow [ $3 : $4 ] failed"
190		return 1
191	fi
192	return 0
193}
194
195ovs_del_flows () {
196	info "Deleting all flows from DP: sbx:$1 br:$2"
197	ovs_sbx "$1" python3 $ovs_base/ovs-dpctl.py del-flows "$2"
198	return 0
199}
200
201ovs_drop_record_and_run () {
202	local sbx=$1
203	shift
204
205	perf record -a -q -e skb:kfree_skb -o ${ovs_dir}/perf.data $* \
206		>> ${ovs_dir}/stdout 2>> ${ovs_dir}/stderr
207	return $?
208}
209
210ovs_drop_reason_count()
211{
212	local reason=$1
213
214	local perf_output=`perf script -i ${ovs_dir}/perf.data -F trace:event,trace`
215	local pattern="skb:kfree_skb:.*reason: $reason"
216
217	return `echo "$perf_output" | grep "$pattern" | wc -l`
218}
219
220ovs_test_flow_fails () {
221	ERR_MSG="Flow actions may not be safe on all matching packets"
222
223	PRE_TEST=$(dmesg | grep -c "${ERR_MSG}")
224	ovs_add_flow $@ &> /dev/null $@ && return 1
225	POST_TEST=$(dmesg | grep -c "${ERR_MSG}")
226
227	if [ "$PRE_TEST" == "$POST_TEST" ]; then
228		return 1
229	fi
230	return 0
231}
232
233usage() {
234	echo
235	echo "$0 [OPTIONS] [TEST]..."
236	echo "If no TEST argument is given, all tests will be run."
237	echo
238	echo "Options"
239	echo "  -t: capture traffic via tcpdump"
240	echo "  -v: verbose"
241	echo "  -p: pause on failure"
242	echo
243	echo "Available tests${tests}"
244	exit 1
245}
246
247
248# psample test
249# - use psample to observe packets
250test_psample() {
251	sbx_add "test_psample" || return $?
252
253	# Add a datapath with per-vport dispatching.
254	ovs_add_dp "test_psample" psample -V 2:1 || return 1
255
256	info "create namespaces"
257	ovs_add_netns_and_veths "test_psample" "psample" \
258		client c0 c1 172.31.110.10/24 -u || return 1
259	ovs_add_netns_and_veths "test_psample" "psample" \
260		server s0 s1 172.31.110.20/24 -u || return 1
261
262	# Check if psample actions can be configured.
263	ovs_add_flow "test_psample" psample \
264	'in_port(1),eth(),eth_type(0x0806),arp()' 'psample(group=1)' &> /dev/null
265	if [ $? == 1 ]; then
266		info "no support for psample - skipping"
267		ovs_exit_sig
268		return $ksft_skip
269	fi
270
271	ovs_del_flows "test_psample" psample
272
273	# Test action verification.
274	OLDIFS=$IFS
275	IFS='*'
276	min_key='in_port(1),eth(),eth_type(0x0800),ipv4()'
277	for testcase in \
278		"cookie to large"*"psample(group=1,cookie=1615141312111009080706050403020100)" \
279		"no group with cookie"*"psample(cookie=abcd)" \
280		"no group"*"psample()";
281	do
282		set -- $testcase;
283		ovs_test_flow_fails "test_psample" psample $min_key $2
284		if [ $? == 1 ]; then
285			info "failed - $1"
286			return 1
287		fi
288	done
289	IFS=$OLDIFS
290
291	ovs_del_flows "test_psample" psample
292	# Allow ARP
293	ovs_add_flow "test_psample" psample \
294		'in_port(1),eth(),eth_type(0x0806),arp()' '2' || return 1
295	ovs_add_flow "test_psample" psample \
296		'in_port(2),eth(),eth_type(0x0806),arp()' '1' || return 1
297
298	# Sample first 14 bytes of all traffic.
299	ovs_add_flow "test_psample" psample \
300	    "in_port(1),eth(),eth_type(0x0800),ipv4()" \
301            "trunc(14),psample(group=1,cookie=c0ffee),2"
302
303	# Sample all traffic. In this case, use a sample() action with both
304	# psample and an upcall emulating simultaneous local sampling and
305	# sFlow / IPFIX.
306	nlpid=$(grep -E "listening on upcall packet handler" \
307            $ovs_dir/s0.out | cut -d ":" -f 2 | tr -d ' ')
308
309	ovs_add_flow "test_psample" psample \
310            "in_port(2),eth(),eth_type(0x0800),ipv4()" \
311            "sample(sample=100%,actions(psample(group=2,cookie=eeff0c),userspace(pid=${nlpid},userdata=eeff0c))),1"
312
313	# Record psample data.
314	ovs_spawn_daemon "test_psample" python3 $ovs_base/ovs-dpctl.py psample-events
315	ovs_wait grep -q "listening for psample events" ${ovs_dir}/stdout
316
317	# Send a single ping.
318	ovs_sbx "test_psample" ip netns exec client ping -I c1 172.31.110.20 -c 1 || return 1
319
320	# We should have received one userspace action upcall and 2 psample packets.
321	ovs_wait grep -q "userspace action command" $ovs_dir/s0.out || return 1
322
323	# client -> server samples should only contain the first 14 bytes of the packet.
324	ovs_wait grep -qE "rate:4294967295,group:1,cookie:c0ffee data:[0-9a-f]{28}$" \
325		$ovs_dir/stdout || return 1
326
327	ovs_wait grep -q "rate:4294967295,group:2,cookie:eeff0c" $ovs_dir/stdout || return 1
328
329	return 0
330}
331
332# drop_reason test
333# - drop packets and verify the right drop reason is reported
334test_drop_reason() {
335	which perf >/dev/null 2>&1 || return $ksft_skip
336	which pahole >/dev/null 2>&1 || return $ksft_skip
337
338	ovs_drop_subsys=$(pahole -C skb_drop_reason_subsys |
339			      awk '/OPENVSWITCH/ { print $3; }' |
340			      tr -d ,)
341
342	sbx_add "test_drop_reason" || return $?
343
344	ovs_add_dp "test_drop_reason" dropreason || return 1
345
346	info "create namespaces"
347	for ns in client server; do
348		ovs_add_netns_and_veths "test_drop_reason" "dropreason" "$ns" \
349			"${ns:0:1}0" "${ns:0:1}1" || return 1
350	done
351
352	# Setup client namespace
353	ip netns exec client ip addr add 172.31.110.10/24 dev c1
354	ip netns exec client ip link set c1 up
355
356	# Setup server namespace
357	ip netns exec server ip addr add 172.31.110.20/24 dev s1
358	ip netns exec server ip link set s1 up
359
360	# Check if drop reasons can be sent
361	ovs_add_flow "test_drop_reason" dropreason \
362		'in_port(1),eth(),eth_type(0x0806),arp()' 'drop(10)' 2>/dev/null
363	if [ $? == 1 ]; then
364		info "no support for drop reasons - skipping"
365		ovs_exit_sig
366		return $ksft_skip
367	fi
368
369	ovs_del_flows "test_drop_reason" dropreason
370
371	# Allow ARP
372	ovs_add_flow "test_drop_reason" dropreason \
373		'in_port(1),eth(),eth_type(0x0806),arp()' '2' || return 1
374	ovs_add_flow "test_drop_reason" dropreason \
375		'in_port(2),eth(),eth_type(0x0806),arp()' '1' || return 1
376
377	# Allow client ICMP traffic but drop return path
378	ovs_add_flow "test_drop_reason" dropreason \
379		"in_port(1),eth(),eth_type(0x0800),ipv4(src=172.31.110.10,proto=1),icmp()" '2'
380	ovs_add_flow "test_drop_reason" dropreason \
381		"in_port(2),eth(),eth_type(0x0800),ipv4(src=172.31.110.20,proto=1),icmp()" 'drop'
382
383	ovs_drop_record_and_run "test_drop_reason" ip netns exec client ping -c 2 172.31.110.20
384	ovs_drop_reason_count 0x${ovs_drop_subsys}0001 # OVS_DROP_FLOW_ACTION
385	if [[ "$?" -ne "2" ]]; then
386		info "Did not detect expected drops: $?"
387		return 1
388	fi
389
390	# Drop UDP 6000 traffic with an explicit action and an error code.
391	ovs_add_flow "test_drop_reason" dropreason \
392		"in_port(1),eth(),eth_type(0x0800),ipv4(src=172.31.110.10,proto=17),udp(dst=6000)" \
393                'drop(42)'
394	# Drop UDP 7000 traffic with an explicit action with no error code.
395	ovs_add_flow "test_drop_reason" dropreason \
396		"in_port(1),eth(),eth_type(0x0800),ipv4(src=172.31.110.10,proto=17),udp(dst=7000)" \
397                'drop(0)'
398
399	ovs_drop_record_and_run \
400            "test_drop_reason" ip netns exec client nc -i 1 -zuv 172.31.110.20 6000
401	ovs_drop_reason_count 0x${ovs_drop_subsys}0004 # OVS_DROP_EXPLICIT_ACTION_ERROR
402	if [[ "$?" -ne "1" ]]; then
403		info "Did not detect expected explicit error drops: $?"
404		return 1
405	fi
406
407	ovs_drop_record_and_run \
408            "test_drop_reason" ip netns exec client nc -i 1 -zuv 172.31.110.20 7000
409	ovs_drop_reason_count 0x${ovs_drop_subsys}0003 # OVS_DROP_EXPLICIT_ACTION
410	if [[ "$?" -ne "1" ]]; then
411		info "Did not detect expected explicit drops: $?"
412		return 1
413	fi
414
415	return 0
416}
417
418# arp_ping test
419# - client has 1500 byte MTU
420# - server has 1500 byte MTU
421# - send ARP ping between two ns
422test_arp_ping () {
423
424	which arping >/dev/null 2>&1 || return $ksft_skip
425
426	sbx_add "test_arp_ping" || return $?
427
428	ovs_add_dp "test_arp_ping" arpping || return 1
429
430	info "create namespaces"
431	for ns in client server; do
432		ovs_add_netns_and_veths "test_arp_ping" "arpping" "$ns" \
433		    "${ns:0:1}0" "${ns:0:1}1" || return 1
434	done
435
436	# Setup client namespace
437	ip netns exec client ip addr add 172.31.110.10/24 dev c1
438	ip netns exec client ip link set c1 up
439	HW_CLIENT=`ip netns exec client ip link show dev c1 | grep -E 'link/ether [0-9a-f:]+' | awk '{print $2;}'`
440	info "Client hwaddr: $HW_CLIENT"
441
442	# Setup server namespace
443	ip netns exec server ip addr add 172.31.110.20/24 dev s1
444	ip netns exec server ip link set s1 up
445	HW_SERVER=`ip netns exec server ip link show dev s1 | grep -E 'link/ether [0-9a-f:]+' | awk '{print $2;}'`
446	info "Server hwaddr: $HW_SERVER"
447
448	ovs_add_flow "test_arp_ping" arpping \
449		"in_port(1),eth(),eth_type(0x0806),arp(sip=172.31.110.10,tip=172.31.110.20,sha=$HW_CLIENT,tha=ff:ff:ff:ff:ff:ff)" '2' || return 1
450	ovs_add_flow "test_arp_ping" arpping \
451		"in_port(2),eth(),eth_type(0x0806),arp()" '1' || return 1
452
453	ovs_sbx "test_arp_ping" ip netns exec client arping -I c1 172.31.110.20 -c 1 || return 1
454
455	return 0
456}
457
458# ct_connect_v4 test
459#  - client has 1500 byte MTU
460#  - server has 1500 byte MTU
461#  - use ICMP to ping in each direction
462#  - only allow CT state stuff to pass through new in c -> s
463test_ct_connect_v4 () {
464
465	which nc >/dev/null 2>/dev/null || return $ksft_skip
466
467	sbx_add "test_ct_connect_v4" || return $?
468
469	ovs_add_dp "test_ct_connect_v4" ct4 || return 1
470	info "create namespaces"
471	for ns in client server; do
472		ovs_add_netns_and_veths "test_ct_connect_v4" "ct4" "$ns" \
473		    "${ns:0:1}0" "${ns:0:1}1" || return 1
474	done
475
476	ip netns exec client ip addr add 172.31.110.10/24 dev c1
477	ip netns exec client ip link set c1 up
478	ip netns exec server ip addr add 172.31.110.20/24 dev s1
479	ip netns exec server ip link set s1 up
480
481	# Add forwarding for ARP and ip packets - completely wildcarded
482	ovs_add_flow "test_ct_connect_v4" ct4 \
483		'in_port(1),eth(),eth_type(0x0806),arp()' '2' || return 1
484	ovs_add_flow "test_ct_connect_v4" ct4 \
485		'in_port(2),eth(),eth_type(0x0806),arp()' '1' || return 1
486	ovs_add_flow "test_ct_connect_v4" ct4 \
487		     'ct_state(-trk),eth(),eth_type(0x0800),ipv4()' \
488		     'ct(commit),recirc(0x1)' || return 1
489	ovs_add_flow "test_ct_connect_v4" ct4 \
490		     'recirc_id(0x1),ct_state(+trk+new),in_port(1),eth(),eth_type(0x0800),ipv4(src=172.31.110.10)' \
491		     '2' || return 1
492	ovs_add_flow "test_ct_connect_v4" ct4 \
493		     'recirc_id(0x1),ct_state(+trk+est),in_port(1),eth(),eth_type(0x0800),ipv4(src=172.31.110.10)' \
494		     '2' || return 1
495	ovs_add_flow "test_ct_connect_v4" ct4 \
496		     'recirc_id(0x1),ct_state(+trk+est),in_port(2),eth(),eth_type(0x0800),ipv4(dst=172.31.110.10)' \
497		     '1' || return 1
498	ovs_add_flow "test_ct_connect_v4" ct4 \
499		     'recirc_id(0x1),ct_state(+trk+inv),eth(),eth_type(0x0800),ipv4()' 'drop' || \
500		     return 1
501
502	# do a ping
503	ovs_sbx "test_ct_connect_v4" ip netns exec client ping 172.31.110.20 -c 3 || return 1
504
505	# create an echo server in 'server'
506	echo "server" | \
507		ovs_netns_spawn_daemon "test_ct_connect_v4" "server" \
508				nc -lvnp 4443
509	ovs_sbx "test_ct_connect_v4" ip netns exec client nc -i 1 -zv 172.31.110.20 4443 || return 1
510
511	# Now test in the other direction (should fail)
512	echo "client" | \
513		ovs_netns_spawn_daemon "test_ct_connect_v4" "client" \
514				nc -lvnp 4443
515	ovs_sbx "test_ct_connect_v4" ip netns exec client nc -i 1 -zv 172.31.110.10 4443
516	if [ $? == 0 ]; then
517	   info "ct connect to client was successful"
518	   return 1
519	fi
520
521	info "done..."
522	return 0
523}
524
525# connect_v4 test
526#  - client has 1500 byte MTU
527#  - server has 1500 byte MTU
528#  - use ICMP to ping in each direction
529test_connect_v4 () {
530
531	sbx_add "test_connect_v4" || return $?
532
533	ovs_add_dp "test_connect_v4" cv4 || return 1
534
535	info "create namespaces"
536	for ns in client server; do
537		ovs_add_netns_and_veths "test_connect_v4" "cv4" "$ns" \
538		    "${ns:0:1}0" "${ns:0:1}1" || return 1
539	done
540
541
542	ip netns exec client ip addr add 172.31.110.10/24 dev c1
543	ip netns exec client ip link set c1 up
544	ip netns exec server ip addr add 172.31.110.20/24 dev s1
545	ip netns exec server ip link set s1 up
546
547	# Add forwarding for ARP and ip packets - completely wildcarded
548	ovs_add_flow "test_connect_v4" cv4 \
549		'in_port(1),eth(),eth_type(0x0806),arp()' '2' || return 1
550	ovs_add_flow "test_connect_v4" cv4 \
551		'in_port(2),eth(),eth_type(0x0806),arp()' '1' || return 1
552	ovs_add_flow "test_connect_v4" cv4 \
553		'in_port(1),eth(),eth_type(0x0800),ipv4(src=172.31.110.10)' '2' || return 1
554	ovs_add_flow "test_connect_v4" cv4 \
555		'in_port(2),eth(),eth_type(0x0800),ipv4(src=172.31.110.20)' '1' || return 1
556
557	# do a ping
558	ovs_sbx "test_connect_v4" ip netns exec client ping 172.31.110.20 -c 3 || return 1
559
560	info "done..."
561	return 0
562}
563
564# nat_connect_v4 test
565#  - client has 1500 byte MTU
566#  - server has 1500 byte MTU
567#  - use ICMP to ping in each direction
568#  - only allow CT state stuff to pass through new in c -> s
569test_nat_connect_v4 () {
570	which nc >/dev/null 2>/dev/null || return $ksft_skip
571
572	sbx_add "test_nat_connect_v4" || return $?
573
574	ovs_add_dp "test_nat_connect_v4" nat4 || return 1
575	info "create namespaces"
576	for ns in client server; do
577		ovs_add_netns_and_veths "test_nat_connect_v4" "nat4" "$ns" \
578		    "${ns:0:1}0" "${ns:0:1}1" || return 1
579	done
580
581	ip netns exec client ip addr add 172.31.110.10/24 dev c1
582	ip netns exec client ip link set c1 up
583	ip netns exec server ip addr add 172.31.110.20/24 dev s1
584	ip netns exec server ip link set s1 up
585
586	ip netns exec client ip route add default via 172.31.110.20
587
588	ovs_add_flow "test_nat_connect_v4" nat4 \
589		'in_port(1),eth(),eth_type(0x0806),arp()' '2' || return 1
590	ovs_add_flow "test_nat_connect_v4" nat4 \
591		'in_port(2),eth(),eth_type(0x0806),arp()' '1' || return 1
592	ovs_add_flow "test_nat_connect_v4" nat4 \
593		"ct_state(-trk),in_port(1),eth(),eth_type(0x0800),ipv4(dst=192.168.0.20)" \
594		"ct(commit,nat(dst=172.31.110.20)),recirc(0x1)"
595	ovs_add_flow "test_nat_connect_v4" nat4 \
596		"ct_state(-trk),in_port(2),eth(),eth_type(0x0800),ipv4()" \
597		"ct(commit,nat),recirc(0x2)"
598
599	ovs_add_flow "test_nat_connect_v4" nat4 \
600		"recirc_id(0x1),ct_state(+trk-inv),in_port(1),eth(),eth_type(0x0800),ipv4()" "2"
601	ovs_add_flow "test_nat_connect_v4" nat4 \
602		"recirc_id(0x2),ct_state(+trk-inv),in_port(2),eth(),eth_type(0x0800),ipv4()" "1"
603
604	# do a ping
605	ovs_sbx "test_nat_connect_v4" ip netns exec client ping 192.168.0.20 -c 3 || return 1
606
607	# create an echo server in 'server'
608	echo "server" | \
609		ovs_netns_spawn_daemon "test_nat_connect_v4" "server" \
610				nc -lvnp 4443
611	ovs_sbx "test_nat_connect_v4" ip netns exec client nc -i 1 -zv 192.168.0.20 4443 || return 1
612
613	# Now test in the other direction (should fail)
614	echo "client" | \
615		ovs_netns_spawn_daemon "test_nat_connect_v4" "client" \
616				nc -lvnp 4443
617	ovs_sbx "test_nat_connect_v4" ip netns exec client nc -i 1 -zv 172.31.110.10 4443
618	if [ $? == 0 ]; then
619	   info "connect to client was successful"
620	   return 1
621	fi
622
623	info "done..."
624	return 0
625}
626
627# nat_related_v4 test
628#  - client->server ip packets go via SNAT
629#  - client solicits ICMP destination unreachable packet from server
630#  - undo NAT for ICMP reply and test dst ip has been updated
631test_nat_related_v4 () {
632	which nc >/dev/null 2>/dev/null || return $ksft_skip
633
634	sbx_add "test_nat_related_v4" || return $?
635
636	ovs_add_dp "test_nat_related_v4" natrelated4 || return 1
637	info "create namespaces"
638	for ns in client server; do
639		ovs_add_netns_and_veths "test_nat_related_v4" "natrelated4" "$ns" \
640			"${ns:0:1}0" "${ns:0:1}1" || return 1
641	done
642
643	ip netns exec client ip addr add 172.31.110.10/24 dev c1
644	ip netns exec client ip link set c1 up
645	ip netns exec server ip addr add 172.31.110.20/24 dev s1
646	ip netns exec server ip link set s1 up
647
648	ip netns exec server ip route add 192.168.0.20/32 via 172.31.110.10
649
650	# Allow ARP
651	ovs_add_flow "test_nat_related_v4" natrelated4 \
652		"in_port(1),eth(),eth_type(0x0806),arp()" "2" || return 1
653	ovs_add_flow "test_nat_related_v4" natrelated4 \
654		"in_port(2),eth(),eth_type(0x0806),arp()" "1" || return 1
655
656	# Allow IP traffic from client->server, rewrite source IP with SNAT to 192.168.0.20
657	ovs_add_flow "test_nat_related_v4" natrelated4 \
658		"ct_state(-trk),in_port(1),eth(),eth_type(0x0800),ipv4(dst=172.31.110.20)" \
659		"ct(commit,nat(src=192.168.0.20)),recirc(0x1)" || return 1
660	ovs_add_flow "test_nat_related_v4" natrelated4 \
661		"recirc_id(0x1),ct_state(+trk-inv),in_port(1),eth(),eth_type(0x0800),ipv4()" \
662		"2" || return 1
663
664	# Allow related ICMP responses back from server and undo NAT to restore original IP
665	# Drop any ICMP related packets where dst ip hasn't been restored back to original IP
666	ovs_add_flow "test_nat_related_v4" natrelated4 \
667		"ct_state(-trk),in_port(2),eth(),eth_type(0x0800),ipv4()" \
668		"ct(commit,nat),recirc(0x2)" || return 1
669	ovs_add_flow "test_nat_related_v4" natrelated4 \
670		"recirc_id(0x2),ct_state(+rel+trk),in_port(2),eth(),eth_type(0x0800),ipv4(src=172.31.110.20,dst=172.31.110.10,proto=1),icmp()" \
671		"1" || return 1
672	ovs_add_flow "test_nat_related_v4" natrelated4 \
673		"recirc_id(0x2),ct_state(+rel+trk),in_port(2),eth(),eth_type(0x0800),ipv4(dst=192.168.0.20,proto=1),icmp()" \
674		"drop" || return 1
675
676	# Solicit destination unreachable response from server
677	ovs_sbx "test_nat_related_v4" ip netns exec client \
678		bash -c "echo a | nc -u -w 1 172.31.110.20 10000"
679
680	# Check to make sure no packets matched the drop rule with incorrect dst ip
681	python3 "$ovs_base/ovs-dpctl.py" dump-flows natrelated4 \
682		| grep "drop" | grep "packets:0" >/dev/null || return 1
683
684	info "done..."
685	return 0
686}
687
688# netlink_validation
689# - Create a dp
690# - check no warning with "old version" simulation
691test_netlink_checks () {
692	sbx_add "test_netlink_checks" || return 1
693
694	info "setting up new DP"
695	ovs_add_dp "test_netlink_checks" nv0 || return 1
696	# now try again
697	PRE_TEST=$(dmesg | grep -E "RIP: [0-9a-fA-Fx]+:ovs_dp_cmd_new\+")
698	ovs_add_dp "test_netlink_checks" nv0 -V 0 || return 1
699	POST_TEST=$(dmesg | grep -E "RIP: [0-9a-fA-Fx]+:ovs_dp_cmd_new\+")
700	if [ "$PRE_TEST" != "$POST_TEST" ]; then
701		info "failed - gen warning"
702		return 1
703	fi
704
705	ovs_add_netns_and_veths "test_netlink_checks" nv0 left left0 l0 || \
706	    return 1
707	ovs_add_netns_and_veths "test_netlink_checks" nv0 right right0 r0 || \
708	    return 1
709	[ $(python3 $ovs_base/ovs-dpctl.py show nv0 | grep port | \
710	    wc -l) == 3 ] || \
711	      return 1
712	ovs_del_if "test_netlink_checks" nv0 right0 || return 1
713	[ $(python3 $ovs_base/ovs-dpctl.py show nv0 | grep port | \
714	    wc -l) == 2 ] || \
715	      return 1
716
717	info "Checking clone depth"
718	ERR_MSG="Flow actions may not be safe on all matching packets"
719	PRE_TEST=$(dmesg | grep -c "${ERR_MSG}")
720	ovs_add_flow "test_netlink_checks" nv0 \
721		'in_port(1),eth(),eth_type(0x800),ipv4()' \
722		'clone(clone(clone(clone(clone(clone(clone(clone(clone(clone(clone(clone(clone(clone(clone(clone(clone(drop)))))))))))))))))' \
723		>/dev/null 2>&1 && return 1
724	POST_TEST=$(dmesg | grep -c "${ERR_MSG}")
725
726	if [ "$PRE_TEST" == "$POST_TEST" ]; then
727		info "failed - clone depth too large"
728		return 1
729	fi
730
731	PRE_TEST=$(dmesg | grep -c "${ERR_MSG}")
732	ovs_add_flow "test_netlink_checks" nv0 \
733		'in_port(1),eth(),eth_type(0x0806),arp()' 'drop(0),2' \
734		&> /dev/null && return 1
735	POST_TEST=$(dmesg | grep -c "${ERR_MSG}")
736	if [ "$PRE_TEST" == "$POST_TEST" ]; then
737		info "failed - error not generated"
738		return 1
739	fi
740	return 0
741}
742
743test_upcall_interfaces() {
744	sbx_add "test_upcall_interfaces" || return 1
745
746	info "setting up new DP"
747	ovs_add_dp "test_upcall_interfaces" ui0 -V 2:1 || return 1
748
749	ovs_add_netns_and_veths "test_upcall_interfaces" ui0 upc left0 l0 \
750	    172.31.110.1/24 -u || return 1
751
752	ovs_wait grep -q "listening on upcall packet handler" ${ovs_dir}/left0.out
753
754	info "sending arping"
755	ip netns exec upc arping -I l0 172.31.110.20 -c 1 \
756	    >$ovs_dir/arping.stdout 2>$ovs_dir/arping.stderr
757
758	grep -E "MISS upcall\[0/yes\]: .*arp\(sip=172.31.110.1,tip=172.31.110.20,op=1,sha=" $ovs_dir/left0.out >/dev/null 2>&1 || return 1
759	return 0
760}
761
762ovs_add_kernel_tunnel() {
763	local sbxname=$1; shift
764	local ns=$1; shift
765	local tnl_type=$1; shift
766	local name=$1; shift
767	local addr=$1; shift
768
769	info "setting up kernel ${tnl_type} tunnel ${name}"
770	ovs_sbx "${sbxname}" ip -netns ${ns} link add dev ${name} type ${tnl_type} $* || return 1
771	on_exit "ovs_sbx ${sbxname} ip -netns ${ns} link del ${name} >/dev/null 2>&1"
772	ovs_sbx "${sbxname}" ip -netns ${ns} addr add dev ${name} ${addr} || return 1
773	ovs_sbx "${sbxname}" ip -netns ${ns} link set dev ${name} mtu 1450 up || return 1
774}
775
776test_tunnel_metadata() {
777	which arping >/dev/null 2>&1 || return $ksft_skip
778
779	sbxname="test_tunnel_metadata"
780	sbx_add "${sbxname}" || return 1
781
782	info "setting up new DP"
783	ovs_add_dp "${sbxname}" tdp0 -V 2:1 || return 1
784
785	ovs_add_netns_and_veths "${sbxname}" tdp0 tns left0 l0 \
786		172.31.110.1/24 || return 1
787
788	info "removing veth interface from openvswitch and setting IP"
789	ovs_del_if "${sbxname}" tdp0 left0 || return 1
790	ovs_sbx "${sbxname}" ip addr add 172.31.110.2/24 dev left0 || return 1
791	ovs_sbx "${sbxname}" ip link set left0 up || return 1
792
793	info "setting up tunnel port in openvswitch"
794	ovs_add_if "${sbxname}" "vxlan" tdp0 ovs-vxlan0 -u || return 1
795	on_exit "ovs_sbx ${sbxname} ip link del ovs-vxlan0"
796	ovs_wait ip link show ovs-vxlan0 &>/dev/null || return 1
797	ovs_sbx "${sbxname}" ip link set ovs-vxlan0 up || return 1
798
799	configs=$(echo '
800	    1 172.31.221.1/24 1155332 32   set   udpcsum flags\(df\|csum\)
801	    2 172.31.222.1/24 1234567 45   set noudpcsum flags\(df\)
802	    3 172.31.223.1/24 1020304 23 unset   udpcsum flags\(csum\)
803	    4 172.31.224.1/24 1357986 15 unset noudpcsum' | sed '/^$/d')
804
805	while read -r i addr id ttl df csum flags; do
806		ovs_add_kernel_tunnel "${sbxname}" tns vxlan vxlan${i} ${addr} \
807			remote 172.31.110.2 id ${id} dstport 4789 \
808			ttl ${ttl} df ${df} ${csum} || return 1
809	done <<< "${configs}"
810
811	ovs_wait grep -q 'listening on upcall packet handler' \
812		${ovs_dir}/ovs-vxlan0.out || return 1
813
814	info "sending arping"
815	for i in 1 2 3 4; do
816		ovs_sbx "${sbxname}" ip netns exec tns \
817			arping -I vxlan${i} 172.31.22${i}.2 -c 1 \
818			>${ovs_dir}/arping.stdout 2>${ovs_dir}/arping.stderr
819	done
820
821	info "checking that received decapsulated packets carry correct metadata"
822	while read -r i addr id ttl df csum flags; do
823		arp_hdr="arp\\(sip=172.31.22${i}.1,tip=172.31.22${i}.2,op=1,sha="
824		addrs="src=172.31.110.1,dst=172.31.110.2"
825		ports="tp_src=[0-9]*,tp_dst=4789"
826		tnl_md="tunnel\\(tun_id=${id},${addrs},ttl=${ttl},${ports},${flags}\\)"
827
828		ovs_sbx "${sbxname}" grep -qE "MISS upcall.*${tnl_md}.*${arp_hdr}" \
829			${ovs_dir}/ovs-vxlan0.out || return 1
830	done <<< "${configs}"
831
832	return 0
833}
834
835test_tunnel_refcount() {
836	sbxname="test_tunnel_refcount"
837	sbx_add "${sbxname}" || return 1
838
839	ovs_sbx "${sbxname}" ip netns add trefns || return 1
840	on_exit "ovs_sbx ${sbxname} ip netns del trefns"
841
842	for tun_type in gre vxlan geneve; do
843		info "testing ${tun_type} tunnel vport refcount"
844
845		ovs_sbx "${sbxname}" ip netns exec trefns \
846			python3 $ovs_base/ovs-dpctl.py \
847			add-dp dp-${tun_type} || return 1
848
849		ovs_sbx "${sbxname}" ip netns exec trefns \
850			python3 $ovs_base/ovs-dpctl.py \
851			add-if --no-lwt -t ${tun_type} \
852			dp-${tun_type} ovs-${tun_type}0 || return 1
853
854		ovs_wait ip -netns trefns link show \
855			ovs-${tun_type}0 >/dev/null 2>&1 || return 1
856
857		info "deleting dp - may hang if reference counting is broken"
858		ovs_sbx "${sbxname}" ip netns exec trefns \
859			python3 $ovs_base/ovs-dpctl.py \
860			del-dp dp-${tun_type} &
861
862		dev_removed() {
863			! ip -netns trefns link show "$1" >/dev/null 2>&1
864		}
865		ovs_wait dev_removed dp-${tun_type} || return 1
866		ovs_wait dev_removed ovs-${tun_type}0 || return 1
867	done
868
869	return 0
870}
871
872test_pop_vlan() {
873	local sbx="test_pop_vlan"
874	sbx_add "$sbx" || return $?
875	ovs_add_dp "$sbx" vlandp || return 1
876
877	ovs_add_netns_and_veths "$sbx" vlandp \
878		ns1 veth1 ns1veth 192.0.2.1/24 || return 1
879	ovs_add_netns_and_veths "$sbx" vlandp \
880		ns2 veth2 ns2veth 192.0.2.2/24 || return 1
881
882	# Baseline: untagged bidirectional forwarding
883	ovs_add_flow "$sbx" vlandp \
884		'in_port(1),eth(),eth_type(0x0806),arp()' '2' || return 1
885	ovs_add_flow "$sbx" vlandp \
886		'in_port(2),eth(),eth_type(0x0806),arp()' '1' || return 1
887	ovs_add_flow "$sbx" vlandp \
888		'in_port(1),eth(),eth_type(0x0800),ipv4()' '2' || return 1
889	ovs_add_flow "$sbx" vlandp \
890		'in_port(2),eth(),eth_type(0x0800),ipv4()' '1' || return 1
891	ovs_sbx "$sbx" ip netns exec ns1 ping -c 3 -W 2 \
892		192.0.2.2 || return 1
893
894	# VLAN topology: ns1 uses VLAN sub-interface, ns2 is plain
895	ip -n ns1 link add link ns1veth name ns1veth.10 \
896		type vlan id 10 || return 1
897	on_exit "ip -n ns1 link del ns1veth.10 2>/dev/null"
898	ip -n ns1 addr add 198.51.100.1/24 dev ns1veth.10 || return 1
899	ip -n ns1 link set ns1veth.10 up || return 1
900	ip -n ns2 addr add 198.51.100.2/24 dev ns2veth || return 1
901
902	ovs_del_flows "$sbx" vlandp
903
904	# Static ARP: avoids VLAN-tagged ARP complexity
905	local ns1veth10mac ns2mac
906	ns1veth10mac=$(ip -n ns1 link show ns1veth.10 \
907		| awk '/link\/ether/ {print $2}')
908	[ -z "$ns1veth10mac" ] && \
909		{ info "failed to get ns1veth10mac"; return 1; }
910	ns2mac=$(ip -n ns2 link show ns2veth \
911		| awk '/link\/ether/ {print $2}')
912	[ -z "$ns2mac" ] && \
913		{ info "failed to get ns2mac"; return 1; }
914	ip -n ns1 neigh replace 198.51.100.2 lladdr "$ns2mac" \
915		dev ns1veth.10 nud permanent || return 1
916	ip -n ns2 neigh replace 198.51.100.1 \
917		lladdr "$ns1veth10mac" \
918		dev ns2veth nud permanent || return 1
919
920	local vlan_match='in_port(1),eth(),eth_type(0x8100),'
921	vlan_match+='vlan(vid=10),'
922	vlan_match+='encap(eth_type(0x0800),'
923	vlan_match+='ipv4(src=198.51.100.1,proto=1),icmp())'
924
925	# Negative: forward without pop_vlan -- tagged frame
926	# is invisible to ns2 (no VLAN sub-interface), ping fails
927	ovs_add_flow "$sbx" vlandp "$vlan_match" '2' || return 1
928	ovs_sbx "$sbx" ip netns exec ns1 ping -I ns1veth.10 \
929		-c 3 -W 1 198.51.100.2 >/dev/null 2>&1 \
930		&& { info "FAIL: ping should fail without pop_vlan"
931		     return 1; }
932
933	ovs_del_flows "$sbx" vlandp
934
935	# Positive: pop_vlan strips tag on forward path,
936	# push_vlan restores tag on return path -- ping succeeds
937	ovs_add_flow "$sbx" vlandp \
938		"$vlan_match" 'pop_vlan,2' || return 1
939	ovs_add_flow "$sbx" vlandp \
940		'in_port(2),eth(),eth_type(0x0800),ipv4()' \
941		'push_vlan(vid=10,pcp=0,tpid=0x8100),1' || return 1
942	ovs_sbx "$sbx" ip netns exec ns1 ping -I ns1veth.10 \
943		-c 3 -W 2 198.51.100.2 || return 1
944
945	return 0
946}
947
948run_test() {
949	(
950	tname="$1"
951	tdesc="$2"
952
953	if python3 ovs-dpctl.py -h 2>&1 | \
954	     grep -E "Need to (install|upgrade) the python" >/dev/null 2>&1; then
955		stdbuf -o0 printf "TEST: %-60s  [PYLIB]\n" "${tdesc}"
956		return $ksft_skip
957	fi
958
959	python3 ovs-dpctl.py show >/dev/null 2>&1 || \
960		echo "[DPCTL] show exception."
961
962	if ! lsmod | grep openvswitch >/dev/null 2>&1; then
963		stdbuf -o0 printf "TEST: %-60s  [NOMOD]\n" "${tdesc}"
964		return $ksft_skip
965	fi
966
967	printf "TEST: %-60s  [START]\n" "${tname}"
968
969	unset IFS
970
971	eval test_${tname}
972	ret=$?
973
974	if [ $ret -eq 0 ]; then
975		printf "TEST: %-60s  [ OK ]\n" "${tdesc}"
976		ovs_exit_sig
977		rm -rf "$ovs_dir"
978	elif [ $ret -eq 1 ]; then
979		printf "TEST: %-60s  [FAIL]\n" "${tdesc}"
980		if [ "${PAUSE_ON_FAIL}" = "yes" ]; then
981			echo
982			echo "Pausing. Logs in $ovs_dir/. Hit enter to continue"
983			read a
984		fi
985		ovs_exit_sig
986		[ "${PAUSE_ON_FAIL}" = "yes" ] || rm -rf "$ovs_dir"
987		exit 1
988	elif [ $ret -eq $ksft_skip ]; then
989		printf "TEST: %-60s  [SKIP]\n" "${tdesc}"
990	elif [ $ret -eq 2 ]; then
991		rm -rf test_${tname}
992		run_test "$1" "$2"
993	fi
994
995	return $ret
996	)
997	ret=$?
998	case $ret in
999		0)
1000			[ $all_skipped = true ] && [ $exitcode=$ksft_skip ] && exitcode=0
1001			all_skipped=false
1002		;;
1003		$ksft_skip)
1004			[ $all_skipped = true ] && exitcode=$ksft_skip
1005		;;
1006		*)
1007			all_skipped=false
1008			exitcode=1
1009		;;
1010	esac
1011
1012	return $ret
1013}
1014
1015
1016exitcode=0
1017desc=0
1018all_skipped=true
1019
1020while getopts :pvt o
1021do
1022	case $o in
1023	p) PAUSE_ON_FAIL=yes;;
1024	v) VERBOSE=1;;
1025	t) if which tcpdump > /dev/null 2>&1; then
1026		TRACING=1
1027	   else
1028		echo "=== tcpdump not available, tracing disabled"
1029	   fi
1030	   ;;
1031	*) usage;;
1032	esac
1033done
1034shift $(($OPTIND-1))
1035
1036IFS="
1037"
1038
1039for arg do
1040	# Check first that all requested tests are available before running any
1041	command -v > /dev/null "test_${arg}" || { echo "=== Test ${arg} not found"; usage; }
1042done
1043
1044name=""
1045desc=""
1046for t in ${tests}; do
1047	[ "${name}" = "" ]	&& name="${t}"	&& continue
1048	[ "${desc}" = "" ]	&& desc="${t}"
1049
1050	run_this=1
1051	for arg do
1052		[ "${arg}" != "${arg#--*}" ] && continue
1053		[ "${arg}" = "${name}" ] && run_this=1 && break
1054		run_this=0
1055	done
1056	if [ $run_this -eq 1 ]; then
1057		run_test "${name}" "${desc}"
1058	fi
1059	name=""
1060	desc=""
1061done
1062
1063exit ${exitcode}
1064