xref: /linux/tools/testing/selftests/net/openvswitch/openvswitch.sh (revision 6443f4f20bdae726fe01cf5946fba9742a0ffda6)
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$KSFT_MACHINE_SLOW" == "Xyes"; then
17	WAIT_TIMEOUT=10
18fi
19
20tests="
21	arp_ping				eth-arp: Basic arp ping between two NS
22	ct_connect_v4				ip4-ct-xon: Basic ipv4 tcp connection using ct
23	connect_v4				ip4-xon: Basic ipv4 ping between two NS
24	nat_connect_v4				ip4-nat-xon: Basic ipv4 tcp connection via NAT
25	nat_related_v4				ip4-nat-related: ICMP related matches work with SNAT
26	netlink_checks				ovsnl: validate netlink attrs and settings
27	upcall_interfaces			ovs: test the upcall interfaces
28	tunnel_metadata				ovs: test extraction of tunnel metadata
29	tunnel_refcount				ovs: test tunnel vport reference cleanup
30	drop_reason				drop: test drop reasons are emitted
31	pop_vlan				vlan: POP_VLAN action strips tag
32	dec_ttl					ttl: dec_ttl decrements IP TTL
33	psample					psample: Sampling packets with psample"
34
35info() {
36	[ "${ovs_dir}" != "" ] &&
37		echo "`date +"[%m-%d %H:%M:%S]"` $*" >> ${ovs_dir}/debug.log
38	[ $VERBOSE = 0 ] || echo $*
39}
40
41ovs_wait() {
42	info "waiting $WAIT_TIMEOUT s for: $@"
43
44	if "$@" ; then
45		info "wait succeeded immediately"
46		return 0
47	fi
48
49	# A quick re-check helps speed up small races in fast systems.
50	# However, fractional sleeps might not necessarily work.
51	local start=0
52	sleep 0.1 || { sleep 1; start=1; }
53
54	for (( i=start; i<WAIT_TIMEOUT; i++ )); do
55		if "$@" ; then
56			info "wait succeeded after $i seconds"
57			return 0
58		fi
59		sleep 1
60	done
61	info "wait failed after $i seconds"
62	return 1
63}
64
65ovs_base=`pwd`
66sbxs=
67sbx_add () {
68	info "adding sandbox '$1'"
69
70	sbxs="$sbxs $1"
71
72	NO_BIN=0
73
74	# Create sandbox.
75	local d="$ovs_base"/$1
76	if [ -e $d ]; then
77		info "removing $d"
78		rm -rf "$d"
79	fi
80	mkdir "$d" || return 1
81	ovs_setenv $1
82}
83
84ovs_exit_sig() {
85	[ -e ${ovs_dir}/cleanup ] && . "$ovs_dir/cleanup"
86}
87
88on_exit() {
89	echo "$1" > ${ovs_dir}/cleanup.tmp
90	cat ${ovs_dir}/cleanup >> ${ovs_dir}/cleanup.tmp
91	mv ${ovs_dir}/cleanup.tmp ${ovs_dir}/cleanup
92}
93
94ovs_setenv() {
95	sandbox=$1
96
97	ovs_dir=$ovs_base${1:+/$1}; export ovs_dir
98
99	test -e ${ovs_dir}/cleanup || : > ${ovs_dir}/cleanup
100}
101
102ovs_sbx() {
103	if test "X$2" != X; then
104		(ovs_setenv $1; shift;
105		 info "run cmd: $@"; "$@" >> ${ovs_dir}/debug.log)
106	else
107		ovs_setenv $1
108	fi
109}
110
111ovs_add_dp () {
112	info "Adding DP/Bridge IF: sbx:$1 dp:$2 {$3, $4, $5}"
113	sbxname="$1"
114	shift
115	ovs_sbx "$sbxname" python3 $ovs_base/ovs-dpctl.py add-dp $*
116	on_exit "ovs_sbx $sbxname python3 $ovs_base/ovs-dpctl.py del-dp $1;"
117}
118
119ovs_add_if () {
120	info "Adding IF to DP: br:$3 if:$4 ($2)"
121	if [ "$5" != "-u" ]; then
122		ovs_sbx "$1" python3 $ovs_base/ovs-dpctl.py add-if \
123		    -t "$2" "$3" "$4" || return 1
124	else
125		python3 $ovs_base/ovs-dpctl.py add-if \
126		    -u -t "$2" "$3" "$4" >$ovs_dir/$4.out 2>$ovs_dir/$4.err &
127		pid=$!
128		on_exit "ovs_sbx $1 kill -TERM $pid 2>/dev/null"
129	fi
130}
131
132ovs_del_if () {
133	info "Deleting IF from DP: br:$2 if:$3"
134	ovs_sbx "$1" python3 $ovs_base/ovs-dpctl.py del-if "$2" "$3" || return 1
135}
136
137ovs_netns_spawn_daemon() {
138	sbx=$1
139	shift
140	netns=$1
141	shift
142	if [ "$netns" == "_default" ]; then
143		$*  >> $ovs_dir/stdout  2>> $ovs_dir/stderr &
144	else
145		ip netns exec $netns $*  >> $ovs_dir/stdout  2>> $ovs_dir/stderr &
146	fi
147	pid=$!
148	ovs_sbx "$sbx" on_exit "kill -TERM $pid 2>/dev/null"
149}
150
151ovs_spawn_daemon() {
152	sbx=$1
153	shift
154	ovs_netns_spawn_daemon $sbx "_default" $*
155}
156
157ovs_add_netns_and_veths () {
158	info "Adding netns attached: sbx:$1 dp:$2 {$3, $4, $5}"
159	ovs_sbx "$1" ip netns add "$3" || return 1
160	on_exit "ovs_sbx $1 ip netns del $3"
161	ovs_sbx "$1" ip link add "$4" type veth peer name "$5" || return 1
162	on_exit "ovs_sbx $1 ip link del $4 >/dev/null 2>&1"
163	ovs_sbx "$1" ip link set "$4" up || return 1
164	ovs_sbx "$1" ip link set "$5" netns "$3" || return 1
165	ovs_sbx "$1" ip netns exec "$3" ip link set "$5" up || return 1
166
167	if [ "$6" != "" ]; then
168		ovs_sbx "$1" ip netns exec "$3" ip addr add "$6" dev "$5" \
169		    || return 1
170	fi
171
172	if [ "$7" != "-u" ]; then
173		ovs_add_if "$1" "netdev" "$2" "$4" || return 1
174	else
175		ovs_add_if "$1" "netdev" "$2" "$4" -u || return 1
176	fi
177
178	if [ $TRACING -eq 1 ]; then
179		ovs_netns_spawn_daemon "$1" "$3" tcpdump -l -i any -s 6553
180		ovs_wait grep -q "listening on any" ${ovs_dir}/stderr
181	fi
182
183	return 0
184}
185
186ovs_add_flow () {
187	info "Adding flow to DP: sbx:$1 br:$2 flow:$3 act:$4"
188	ovs_sbx "$1" python3 $ovs_base/ovs-dpctl.py add-flow "$2" "$3" "$4"
189	if [ $? -ne 0 ]; then
190		info "Flow [ $3 : $4 ] failed"
191		return 1
192	fi
193	return 0
194}
195
196ovs_del_flows () {
197	info "Deleting all flows from DP: sbx:$1 br:$2"
198	ovs_sbx "$1" python3 $ovs_base/ovs-dpctl.py del-flows "$2"
199	return 0
200}
201
202ovs_drop_record_and_run () {
203	local sbx=$1
204	shift
205
206	perf record -a -q -e skb:kfree_skb -o ${ovs_dir}/perf.data $* \
207		>> ${ovs_dir}/stdout 2>> ${ovs_dir}/stderr
208	return $?
209}
210
211ovs_drop_reason_count()
212{
213	local reason=$1
214
215	local perf_output=`perf script -i ${ovs_dir}/perf.data -F trace:event,trace`
216	local pattern="skb:kfree_skb:.*reason: $reason"
217
218	return `echo "$perf_output" | grep "$pattern" | wc -l`
219}
220
221ovs_test_flow_fails () {
222	ERR_MSG="Flow actions may not be safe on all matching packets"
223
224	PRE_TEST=$(dmesg | grep -c "${ERR_MSG}")
225	ovs_add_flow $@ &> /dev/null $@ && return 1
226	POST_TEST=$(dmesg | grep -c "${ERR_MSG}")
227
228	if [ "$PRE_TEST" == "$POST_TEST" ]; then
229		return 1
230	fi
231	return 0
232}
233
234usage() {
235	echo
236	echo "$0 [OPTIONS] [TEST]..."
237	echo "If no TEST argument is given, all tests will be run."
238	echo
239	echo "Options"
240	echo "  -t: capture traffic via tcpdump"
241	echo "  -v: verbose"
242	echo "  -p: pause on failure"
243	echo
244	echo "Available tests${tests}"
245	exit 1
246}
247
248
249test_dec_ttl() {
250	sbx_add "test_dec_ttl" || return $?
251	ovs_add_dp "test_dec_ttl" decttl || return 1
252
253	info "create namespaces"
254	for ns in client server; do
255		ovs_add_netns_and_veths "test_dec_ttl" "decttl" "$ns" \
256			"${ns:0:1}0" "${ns:0:1}1" || return 1
257	done
258
259	ip netns exec client ip addr add 10.0.0.1/24 dev c1
260	ip netns exec client ip link set c1 up
261	ip netns exec server ip addr add 10.0.0.2/24 dev s1
262	ip netns exec server ip link set s1 up
263
264	# Probe: check if kernel supports dec_ttl action.
265	ovs_add_flow "test_dec_ttl" decttl \
266		'in_port(1),eth(),eth_type(0x0800),ipv4()' \
267		'dec_ttl(le_1())' &>/dev/null
268	if [ $? -ne 0 ]; then
269		info "no support for dec_ttl - skipping"
270		ovs_exit_sig
271		return $ksft_skip
272	fi
273
274	ovs_del_flows "test_dec_ttl" decttl
275
276	# ARP flows (bidirectional)
277	ovs_add_flow "test_dec_ttl" decttl \
278		'in_port(1),eth(),eth_type(0x0806),arp()' '2' || return 1
279	ovs_add_flow "test_dec_ttl" decttl \
280		'in_port(2),eth(),eth_type(0x0806),arp()' '1' || return 1
281
282	# IP flows with dec_ttl action
283	ovs_add_flow "test_dec_ttl" decttl \
284		'in_port(1),eth(),eth_type(0x0800),ipv4()' \
285		'dec_ttl(le_1()),2' || return 1
286	ovs_add_flow "test_dec_ttl" decttl \
287		'in_port(2),eth(),eth_type(0x0800),ipv4()' \
288		'dec_ttl(le_1()),1' || return 1
289
290	info "verify connectivity with dec_ttl"
291	ovs_sbx "test_dec_ttl" ip netns exec client ping -c 1 -W 2 \
292		10.0.0.2 || return 1
293
294	info "verify TTL=1 is dropped by dec_ttl"
295	ovs_sbx "test_dec_ttl" ip netns exec client ping -c 1 -W 2 \
296		-t 1 10.0.0.2 >/dev/null 2>&1 \
297		&& { info "FAIL: ping should fail with TTL=1 and dec_ttl"
298		     return 1; }
299
300	return 0
301}
302
303# psample test
304# - use psample to observe packets
305test_psample() {
306	sbx_add "test_psample" || return $?
307
308	# Add a datapath with per-vport dispatching.
309	ovs_add_dp "test_psample" psample -V 2:1 || return 1
310
311	info "create namespaces"
312	ovs_add_netns_and_veths "test_psample" "psample" \
313		client c0 c1 172.31.110.10/24 -u || return 1
314	ovs_add_netns_and_veths "test_psample" "psample" \
315		server s0 s1 172.31.110.20/24 -u || return 1
316
317	# Check if psample actions can be configured.
318	ovs_add_flow "test_psample" psample \
319	'in_port(1),eth(),eth_type(0x0806),arp()' 'psample(group=1)' &> /dev/null
320	if [ $? == 1 ]; then
321		info "no support for psample - skipping"
322		ovs_exit_sig
323		return $ksft_skip
324	fi
325
326	ovs_del_flows "test_psample" psample
327
328	# Test action verification.
329	OLDIFS=$IFS
330	IFS='*'
331	min_key='in_port(1),eth(),eth_type(0x0800),ipv4()'
332	for testcase in \
333		"cookie to large"*"psample(group=1,cookie=1615141312111009080706050403020100)" \
334		"no group with cookie"*"psample(cookie=abcd)" \
335		"no group"*"psample()";
336	do
337		set -- $testcase;
338		ovs_test_flow_fails "test_psample" psample $min_key $2
339		if [ $? == 1 ]; then
340			info "failed - $1"
341			return 1
342		fi
343	done
344	IFS=$OLDIFS
345
346	ovs_del_flows "test_psample" psample
347	# Allow ARP
348	ovs_add_flow "test_psample" psample \
349		'in_port(1),eth(),eth_type(0x0806),arp()' '2' || return 1
350	ovs_add_flow "test_psample" psample \
351		'in_port(2),eth(),eth_type(0x0806),arp()' '1' || return 1
352
353	# Sample first 14 bytes of all traffic.
354	ovs_add_flow "test_psample" psample \
355	    "in_port(1),eth(),eth_type(0x0800),ipv4()" \
356            "trunc(14),psample(group=1,cookie=c0ffee),2"
357
358	# Sample all traffic. In this case, use a sample() action with both
359	# psample and an upcall emulating simultaneous local sampling and
360	# sFlow / IPFIX.
361	nlpid=$(grep -E "listening on upcall packet handler" \
362            $ovs_dir/s0.out | cut -d ":" -f 2 | tr -d ' ')
363
364	ovs_add_flow "test_psample" psample \
365            "in_port(2),eth(),eth_type(0x0800),ipv4()" \
366            "sample(sample=100%,actions(psample(group=2,cookie=eeff0c),userspace(pid=${nlpid},userdata=eeff0c))),1"
367
368	# Record psample data.
369	ovs_spawn_daemon "test_psample" python3 $ovs_base/ovs-dpctl.py psample-events
370	ovs_wait grep -q "listening for psample events" ${ovs_dir}/stdout
371
372	# Send a single ping.
373	ovs_sbx "test_psample" ip netns exec client ping -I c1 172.31.110.20 -c 1 || return 1
374
375	# We should have received one userspace action upcall and 2 psample packets.
376	ovs_wait grep -q "userspace action command" $ovs_dir/s0.out || return 1
377
378	# client -> server samples should only contain the first 14 bytes of the packet.
379	ovs_wait grep -qE "rate:4294967295,group:1,cookie:c0ffee data:[0-9a-f]{28}$" \
380		$ovs_dir/stdout || return 1
381
382	ovs_wait grep -q "rate:4294967295,group:2,cookie:eeff0c" $ovs_dir/stdout || return 1
383
384	return 0
385}
386
387# drop_reason test
388# - drop packets and verify the right drop reason is reported
389test_drop_reason() {
390	which perf >/dev/null 2>&1 || return $ksft_skip
391	which pahole >/dev/null 2>&1 || return $ksft_skip
392
393	ovs_drop_subsys=$(pahole -C skb_drop_reason_subsys |
394			      awk '/OPENVSWITCH/ { print $3; }' |
395			      tr -d ,)
396
397	sbx_add "test_drop_reason" || return $?
398
399	ovs_add_dp "test_drop_reason" dropreason || return 1
400
401	info "create namespaces"
402	for ns in client server; do
403		ovs_add_netns_and_veths "test_drop_reason" "dropreason" "$ns" \
404			"${ns:0:1}0" "${ns:0:1}1" || return 1
405	done
406
407	# Setup client namespace
408	ip netns exec client ip addr add 172.31.110.10/24 dev c1
409	ip netns exec client ip link set c1 up
410
411	# Setup server namespace
412	ip netns exec server ip addr add 172.31.110.20/24 dev s1
413	ip netns exec server ip link set s1 up
414
415	# Check if drop reasons can be sent
416	ovs_add_flow "test_drop_reason" dropreason \
417		'in_port(1),eth(),eth_type(0x0806),arp()' 'drop(10)' 2>/dev/null
418	if [ $? == 1 ]; then
419		info "no support for drop reasons - skipping"
420		ovs_exit_sig
421		return $ksft_skip
422	fi
423
424	ovs_del_flows "test_drop_reason" dropreason
425
426	# Allow ARP
427	ovs_add_flow "test_drop_reason" dropreason \
428		'in_port(1),eth(),eth_type(0x0806),arp()' '2' || return 1
429	ovs_add_flow "test_drop_reason" dropreason \
430		'in_port(2),eth(),eth_type(0x0806),arp()' '1' || return 1
431
432	# Allow client ICMP traffic but drop return path
433	ovs_add_flow "test_drop_reason" dropreason \
434		"in_port(1),eth(),eth_type(0x0800),ipv4(src=172.31.110.10,proto=1),icmp()" '2'
435	ovs_add_flow "test_drop_reason" dropreason \
436		"in_port(2),eth(),eth_type(0x0800),ipv4(src=172.31.110.20,proto=1),icmp()" 'drop'
437
438	ovs_drop_record_and_run "test_drop_reason" ip netns exec client ping -c 2 172.31.110.20
439	ovs_drop_reason_count 0x${ovs_drop_subsys}0001 # OVS_DROP_FLOW_ACTION
440	if [[ "$?" -ne "2" ]]; then
441		info "Did not detect expected drops: $?"
442		return 1
443	fi
444
445	# Drop UDP 6000 traffic with an explicit action and an error code.
446	ovs_add_flow "test_drop_reason" dropreason \
447		"in_port(1),eth(),eth_type(0x0800),ipv4(src=172.31.110.10,proto=17),udp(dst=6000)" \
448                'drop(42)'
449	# Drop UDP 7000 traffic with an explicit action with no error code.
450	ovs_add_flow "test_drop_reason" dropreason \
451		"in_port(1),eth(),eth_type(0x0800),ipv4(src=172.31.110.10,proto=17),udp(dst=7000)" \
452                'drop(0)'
453
454	ovs_drop_record_and_run \
455            "test_drop_reason" ip netns exec client nc -i 1 -zuv 172.31.110.20 6000
456	ovs_drop_reason_count 0x${ovs_drop_subsys}0004 # OVS_DROP_EXPLICIT_ACTION_ERROR
457	if [[ "$?" -ne "1" ]]; then
458		info "Did not detect expected explicit error drops: $?"
459		return 1
460	fi
461
462	ovs_drop_record_and_run \
463            "test_drop_reason" ip netns exec client nc -i 1 -zuv 172.31.110.20 7000
464	ovs_drop_reason_count 0x${ovs_drop_subsys}0003 # OVS_DROP_EXPLICIT_ACTION
465	if [[ "$?" -ne "1" ]]; then
466		info "Did not detect expected explicit drops: $?"
467		return 1
468	fi
469
470	return 0
471}
472
473# arp_ping test
474# - client has 1500 byte MTU
475# - server has 1500 byte MTU
476# - send ARP ping between two ns
477test_arp_ping () {
478
479	which arping >/dev/null 2>&1 || return $ksft_skip
480
481	sbx_add "test_arp_ping" || return $?
482
483	ovs_add_dp "test_arp_ping" arpping || return 1
484
485	info "create namespaces"
486	for ns in client server; do
487		ovs_add_netns_and_veths "test_arp_ping" "arpping" "$ns" \
488		    "${ns:0:1}0" "${ns:0:1}1" || return 1
489	done
490
491	# Setup client namespace
492	ip netns exec client ip addr add 172.31.110.10/24 dev c1
493	ip netns exec client ip link set c1 up
494	HW_CLIENT=`ip netns exec client ip link show dev c1 | grep -E 'link/ether [0-9a-f:]+' | awk '{print $2;}'`
495	info "Client hwaddr: $HW_CLIENT"
496
497	# Setup server namespace
498	ip netns exec server ip addr add 172.31.110.20/24 dev s1
499	ip netns exec server ip link set s1 up
500	HW_SERVER=`ip netns exec server ip link show dev s1 | grep -E 'link/ether [0-9a-f:]+' | awk '{print $2;}'`
501	info "Server hwaddr: $HW_SERVER"
502
503	ovs_add_flow "test_arp_ping" arpping \
504		"in_port(1),eth(),eth_type(0x0806),arp(sip=172.31.110.10,tip=172.31.110.20,sha=$HW_CLIENT,tha=ff:ff:ff:ff:ff:ff)" '2' || return 1
505	ovs_add_flow "test_arp_ping" arpping \
506		"in_port(2),eth(),eth_type(0x0806),arp()" '1' || return 1
507
508	ovs_sbx "test_arp_ping" ip netns exec client arping -I c1 172.31.110.20 -c 1 || return 1
509
510	return 0
511}
512
513# ct_connect_v4 test
514#  - client has 1500 byte MTU
515#  - server has 1500 byte MTU
516#  - use ICMP to ping in each direction
517#  - only allow CT state stuff to pass through new in c -> s
518test_ct_connect_v4 () {
519
520	which nc >/dev/null 2>/dev/null || return $ksft_skip
521
522	sbx_add "test_ct_connect_v4" || return $?
523
524	ovs_add_dp "test_ct_connect_v4" ct4 || return 1
525	info "create namespaces"
526	for ns in client server; do
527		ovs_add_netns_and_veths "test_ct_connect_v4" "ct4" "$ns" \
528		    "${ns:0:1}0" "${ns:0:1}1" || return 1
529	done
530
531	ip netns exec client ip addr add 172.31.110.10/24 dev c1
532	ip netns exec client ip link set c1 up
533	ip netns exec server ip addr add 172.31.110.20/24 dev s1
534	ip netns exec server ip link set s1 up
535
536	# Add forwarding for ARP and ip packets - completely wildcarded
537	ovs_add_flow "test_ct_connect_v4" ct4 \
538		'in_port(1),eth(),eth_type(0x0806),arp()' '2' || return 1
539	ovs_add_flow "test_ct_connect_v4" ct4 \
540		'in_port(2),eth(),eth_type(0x0806),arp()' '1' || return 1
541	ovs_add_flow "test_ct_connect_v4" ct4 \
542		     'ct_state(-trk),eth(),eth_type(0x0800),ipv4()' \
543		     'ct(commit),recirc(0x1)' || return 1
544	ovs_add_flow "test_ct_connect_v4" ct4 \
545		     'recirc_id(0x1),ct_state(+trk+new),in_port(1),eth(),eth_type(0x0800),ipv4(src=172.31.110.10)' \
546		     '2' || return 1
547	ovs_add_flow "test_ct_connect_v4" ct4 \
548		     'recirc_id(0x1),ct_state(+trk+est),in_port(1),eth(),eth_type(0x0800),ipv4(src=172.31.110.10)' \
549		     '2' || return 1
550	ovs_add_flow "test_ct_connect_v4" ct4 \
551		     'recirc_id(0x1),ct_state(+trk+est),in_port(2),eth(),eth_type(0x0800),ipv4(dst=172.31.110.10)' \
552		     '1' || return 1
553	ovs_add_flow "test_ct_connect_v4" ct4 \
554		     'recirc_id(0x1),ct_state(+trk+inv),eth(),eth_type(0x0800),ipv4()' 'drop' || \
555		     return 1
556
557	# do a ping
558	ovs_sbx "test_ct_connect_v4" ip netns exec client ping 172.31.110.20 -c 3 || return 1
559
560	# create an echo server in 'server'
561	echo "server" | \
562		ovs_netns_spawn_daemon "test_ct_connect_v4" "server" \
563				nc -lvnp 4443
564	ovs_sbx "test_ct_connect_v4" ip netns exec client nc -i 1 -zv 172.31.110.20 4443 || return 1
565
566	# Now test in the other direction (should fail)
567	echo "client" | \
568		ovs_netns_spawn_daemon "test_ct_connect_v4" "client" \
569				nc -lvnp 4443
570	ovs_sbx "test_ct_connect_v4" ip netns exec client nc -i 1 -zv 172.31.110.10 4443
571	if [ $? == 0 ]; then
572	   info "ct connect to client was successful"
573	   return 1
574	fi
575
576	info "done..."
577	return 0
578}
579
580# connect_v4 test
581#  - client has 1500 byte MTU
582#  - server has 1500 byte MTU
583#  - use ICMP to ping in each direction
584test_connect_v4 () {
585
586	sbx_add "test_connect_v4" || return $?
587
588	ovs_add_dp "test_connect_v4" cv4 || return 1
589
590	info "create namespaces"
591	for ns in client server; do
592		ovs_add_netns_and_veths "test_connect_v4" "cv4" "$ns" \
593		    "${ns:0:1}0" "${ns:0:1}1" || return 1
594	done
595
596
597	ip netns exec client ip addr add 172.31.110.10/24 dev c1
598	ip netns exec client ip link set c1 up
599	ip netns exec server ip addr add 172.31.110.20/24 dev s1
600	ip netns exec server ip link set s1 up
601
602	# Add forwarding for ARP and ip packets - completely wildcarded
603	ovs_add_flow "test_connect_v4" cv4 \
604		'in_port(1),eth(),eth_type(0x0806),arp()' '2' || return 1
605	ovs_add_flow "test_connect_v4" cv4 \
606		'in_port(2),eth(),eth_type(0x0806),arp()' '1' || return 1
607	ovs_add_flow "test_connect_v4" cv4 \
608		'in_port(1),eth(),eth_type(0x0800),ipv4(src=172.31.110.10)' '2' || return 1
609	ovs_add_flow "test_connect_v4" cv4 \
610		'in_port(2),eth(),eth_type(0x0800),ipv4(src=172.31.110.20)' '1' || return 1
611
612	# do a ping
613	ovs_sbx "test_connect_v4" ip netns exec client ping 172.31.110.20 -c 3 || return 1
614
615	info "done..."
616	return 0
617}
618
619# nat_connect_v4 test
620#  - client has 1500 byte MTU
621#  - server has 1500 byte MTU
622#  - use ICMP to ping in each direction
623#  - only allow CT state stuff to pass through new in c -> s
624test_nat_connect_v4 () {
625	which nc >/dev/null 2>/dev/null || return $ksft_skip
626
627	sbx_add "test_nat_connect_v4" || return $?
628
629	ovs_add_dp "test_nat_connect_v4" nat4 || return 1
630	info "create namespaces"
631	for ns in client server; do
632		ovs_add_netns_and_veths "test_nat_connect_v4" "nat4" "$ns" \
633		    "${ns:0:1}0" "${ns:0:1}1" || return 1
634	done
635
636	ip netns exec client ip addr add 172.31.110.10/24 dev c1
637	ip netns exec client ip link set c1 up
638	ip netns exec server ip addr add 172.31.110.20/24 dev s1
639	ip netns exec server ip link set s1 up
640
641	ip netns exec client ip route add default via 172.31.110.20
642
643	ovs_add_flow "test_nat_connect_v4" nat4 \
644		'in_port(1),eth(),eth_type(0x0806),arp()' '2' || return 1
645	ovs_add_flow "test_nat_connect_v4" nat4 \
646		'in_port(2),eth(),eth_type(0x0806),arp()' '1' || return 1
647	ovs_add_flow "test_nat_connect_v4" nat4 \
648		"ct_state(-trk),in_port(1),eth(),eth_type(0x0800),ipv4(dst=192.168.0.20)" \
649		"ct(commit,nat(dst=172.31.110.20)),recirc(0x1)"
650	ovs_add_flow "test_nat_connect_v4" nat4 \
651		"ct_state(-trk),in_port(2),eth(),eth_type(0x0800),ipv4()" \
652		"ct(commit,nat),recirc(0x2)"
653
654	ovs_add_flow "test_nat_connect_v4" nat4 \
655		"recirc_id(0x1),ct_state(+trk-inv),in_port(1),eth(),eth_type(0x0800),ipv4()" "2"
656	ovs_add_flow "test_nat_connect_v4" nat4 \
657		"recirc_id(0x2),ct_state(+trk-inv),in_port(2),eth(),eth_type(0x0800),ipv4()" "1"
658
659	# do a ping
660	ovs_sbx "test_nat_connect_v4" ip netns exec client ping 192.168.0.20 -c 3 || return 1
661
662	# create an echo server in 'server'
663	echo "server" | \
664		ovs_netns_spawn_daemon "test_nat_connect_v4" "server" \
665				nc -lvnp 4443
666	ovs_sbx "test_nat_connect_v4" ip netns exec client nc -i 1 -zv 192.168.0.20 4443 || return 1
667
668	# Now test in the other direction (should fail)
669	echo "client" | \
670		ovs_netns_spawn_daemon "test_nat_connect_v4" "client" \
671				nc -lvnp 4443
672	ovs_sbx "test_nat_connect_v4" ip netns exec client nc -i 1 -zv 172.31.110.10 4443
673	if [ $? == 0 ]; then
674	   info "connect to client was successful"
675	   return 1
676	fi
677
678	info "done..."
679	return 0
680}
681
682# nat_related_v4 test
683#  - client->server ip packets go via SNAT
684#  - client solicits ICMP destination unreachable packet from server
685#  - undo NAT for ICMP reply and test dst ip has been updated
686test_nat_related_v4 () {
687	which nc >/dev/null 2>/dev/null || return $ksft_skip
688
689	sbx_add "test_nat_related_v4" || return $?
690
691	ovs_add_dp "test_nat_related_v4" natrelated4 || return 1
692	info "create namespaces"
693	for ns in client server; do
694		ovs_add_netns_and_veths "test_nat_related_v4" "natrelated4" "$ns" \
695			"${ns:0:1}0" "${ns:0:1}1" || return 1
696	done
697
698	ip netns exec client ip addr add 172.31.110.10/24 dev c1
699	ip netns exec client ip link set c1 up
700	ip netns exec server ip addr add 172.31.110.20/24 dev s1
701	ip netns exec server ip link set s1 up
702
703	ip netns exec server ip route add 192.168.0.20/32 via 172.31.110.10
704
705	# Allow ARP
706	ovs_add_flow "test_nat_related_v4" natrelated4 \
707		"in_port(1),eth(),eth_type(0x0806),arp()" "2" || return 1
708	ovs_add_flow "test_nat_related_v4" natrelated4 \
709		"in_port(2),eth(),eth_type(0x0806),arp()" "1" || return 1
710
711	# Allow IP traffic from client->server, rewrite source IP with SNAT to 192.168.0.20
712	ovs_add_flow "test_nat_related_v4" natrelated4 \
713		"ct_state(-trk),in_port(1),eth(),eth_type(0x0800),ipv4(dst=172.31.110.20)" \
714		"ct(commit,nat(src=192.168.0.20)),recirc(0x1)" || return 1
715	ovs_add_flow "test_nat_related_v4" natrelated4 \
716		"recirc_id(0x1),ct_state(+trk-inv),in_port(1),eth(),eth_type(0x0800),ipv4()" \
717		"2" || return 1
718
719	# Allow related ICMP responses back from server and undo NAT to restore original IP
720	# Drop any ICMP related packets where dst ip hasn't been restored back to original IP
721	ovs_add_flow "test_nat_related_v4" natrelated4 \
722		"ct_state(-trk),in_port(2),eth(),eth_type(0x0800),ipv4()" \
723		"ct(commit,nat),recirc(0x2)" || return 1
724	ovs_add_flow "test_nat_related_v4" natrelated4 \
725		"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()" \
726		"1" || return 1
727	ovs_add_flow "test_nat_related_v4" natrelated4 \
728		"recirc_id(0x2),ct_state(+rel+trk),in_port(2),eth(),eth_type(0x0800),ipv4(dst=192.168.0.20,proto=1),icmp()" \
729		"drop" || return 1
730
731	# Solicit destination unreachable response from server
732	ovs_sbx "test_nat_related_v4" ip netns exec client \
733		bash -c "echo a | nc -u -w 1 172.31.110.20 10000"
734
735	# Check to make sure no packets matched the drop rule with incorrect dst ip
736	python3 "$ovs_base/ovs-dpctl.py" dump-flows natrelated4 \
737		| grep "drop" | grep "packets:0" >/dev/null || return 1
738
739	info "done..."
740	return 0
741}
742
743# netlink_validation
744# - Create a dp
745# - check no warning with "old version" simulation
746test_netlink_checks () {
747	sbx_add "test_netlink_checks" || return 1
748
749	info "setting up new DP"
750	ovs_add_dp "test_netlink_checks" nv0 || return 1
751	# now try again
752	PRE_TEST=$(dmesg | grep -E "RIP: [0-9a-fA-Fx]+:ovs_dp_cmd_new\+")
753	ovs_add_dp "test_netlink_checks" nv0 -V 0 || return 1
754	POST_TEST=$(dmesg | grep -E "RIP: [0-9a-fA-Fx]+:ovs_dp_cmd_new\+")
755	if [ "$PRE_TEST" != "$POST_TEST" ]; then
756		info "failed - gen warning"
757		return 1
758	fi
759
760	ovs_add_netns_and_veths "test_netlink_checks" nv0 left left0 l0 || \
761	    return 1
762	ovs_add_netns_and_veths "test_netlink_checks" nv0 right right0 r0 || \
763	    return 1
764	[ $(python3 $ovs_base/ovs-dpctl.py show nv0 | grep port | \
765	    wc -l) == 3 ] || \
766	      return 1
767	ovs_del_if "test_netlink_checks" nv0 right0 || return 1
768	[ $(python3 $ovs_base/ovs-dpctl.py show nv0 | grep port | \
769	    wc -l) == 2 ] || \
770	      return 1
771
772	info "Checking clone depth"
773	ERR_MSG="Flow actions may not be safe on all matching packets"
774	PRE_TEST=$(dmesg | grep -c "${ERR_MSG}")
775	ovs_add_flow "test_netlink_checks" nv0 \
776		'in_port(1),eth(),eth_type(0x800),ipv4()' \
777		'clone(clone(clone(clone(clone(clone(clone(clone(clone(clone(clone(clone(clone(clone(clone(clone(clone(drop)))))))))))))))))' \
778		>/dev/null 2>&1 && return 1
779	POST_TEST=$(dmesg | grep -c "${ERR_MSG}")
780
781	if [ "$PRE_TEST" == "$POST_TEST" ]; then
782		info "failed - clone depth too large"
783		return 1
784	fi
785
786	PRE_TEST=$(dmesg | grep -c "${ERR_MSG}")
787	ovs_add_flow "test_netlink_checks" nv0 \
788		'in_port(1),eth(),eth_type(0x0806),arp()' 'drop(0),2' \
789		&> /dev/null && return 1
790	POST_TEST=$(dmesg | grep -c "${ERR_MSG}")
791	if [ "$PRE_TEST" == "$POST_TEST" ]; then
792		info "failed - error not generated"
793		return 1
794	fi
795	return 0
796}
797
798test_upcall_interfaces() {
799	sbx_add "test_upcall_interfaces" || return 1
800
801	info "setting up new DP"
802	ovs_add_dp "test_upcall_interfaces" ui0 -V 2:1 || return 1
803
804	ovs_add_netns_and_veths "test_upcall_interfaces" ui0 upc left0 l0 \
805	    172.31.110.1/24 -u || return 1
806
807	ovs_wait grep -q "listening on upcall packet handler" ${ovs_dir}/left0.out
808
809	info "sending arping"
810	ip netns exec upc arping -I l0 172.31.110.20 -c 1 \
811	    >$ovs_dir/arping.stdout 2>$ovs_dir/arping.stderr
812
813	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
814	return 0
815}
816
817ovs_add_kernel_tunnel() {
818	local sbxname=$1; shift
819	local ns=$1; shift
820	local tnl_type=$1; shift
821	local name=$1; shift
822	local addr=$1; shift
823
824	info "setting up kernel ${tnl_type} tunnel ${name}"
825	ovs_sbx "${sbxname}" ip -netns ${ns} link add dev ${name} type ${tnl_type} $* || return 1
826	on_exit "ovs_sbx ${sbxname} ip -netns ${ns} link del ${name} >/dev/null 2>&1"
827	ovs_sbx "${sbxname}" ip -netns ${ns} addr add dev ${name} ${addr} || return 1
828	ovs_sbx "${sbxname}" ip -netns ${ns} link set dev ${name} mtu 1450 up || return 1
829}
830
831test_tunnel_metadata() {
832	which arping >/dev/null 2>&1 || return $ksft_skip
833
834	sbxname="test_tunnel_metadata"
835	sbx_add "${sbxname}" || return 1
836
837	info "setting up new DP"
838	ovs_add_dp "${sbxname}" tdp0 -V 2:1 || return 1
839
840	ovs_add_netns_and_veths "${sbxname}" tdp0 tns left0 l0 \
841		172.31.110.1/24 || return 1
842
843	info "removing veth interface from openvswitch and setting IP"
844	ovs_del_if "${sbxname}" tdp0 left0 || return 1
845	ovs_sbx "${sbxname}" ip addr add 172.31.110.2/24 dev left0 || return 1
846	ovs_sbx "${sbxname}" ip link set left0 up || return 1
847
848	info "setting up tunnel port in openvswitch"
849	ovs_add_if "${sbxname}" "vxlan" tdp0 ovs-vxlan0 -u || return 1
850	on_exit "ovs_sbx ${sbxname} ip link del ovs-vxlan0"
851	ovs_wait ip link show ovs-vxlan0 &>/dev/null || return 1
852	ovs_sbx "${sbxname}" ip link set ovs-vxlan0 up || return 1
853
854	configs=$(echo '
855	    1 172.31.221.1/24 1155332 32   set   udpcsum flags\(df\|csum\)
856	    2 172.31.222.1/24 1234567 45   set noudpcsum flags\(df\)
857	    3 172.31.223.1/24 1020304 23 unset   udpcsum flags\(csum\)
858	    4 172.31.224.1/24 1357986 15 unset noudpcsum' | sed '/^$/d')
859
860	while read -r i addr id ttl df csum flags; do
861		ovs_add_kernel_tunnel "${sbxname}" tns vxlan vxlan${i} ${addr} \
862			remote 172.31.110.2 id ${id} dstport 4789 \
863			ttl ${ttl} df ${df} ${csum} || return 1
864	done <<< "${configs}"
865
866	ovs_wait grep -q 'listening on upcall packet handler' \
867		${ovs_dir}/ovs-vxlan0.out || return 1
868
869	info "sending arping"
870	for i in 1 2 3 4; do
871		ovs_sbx "${sbxname}" ip netns exec tns \
872			arping -I vxlan${i} 172.31.22${i}.2 -c 1 \
873			>${ovs_dir}/arping.stdout 2>${ovs_dir}/arping.stderr
874	done
875
876	info "checking that received decapsulated packets carry correct metadata"
877	while read -r i addr id ttl df csum flags; do
878		arp_hdr="arp\\(sip=172.31.22${i}.1,tip=172.31.22${i}.2,op=1,sha="
879		addrs="src=172.31.110.1,dst=172.31.110.2"
880		ports="tp_src=[0-9]*,tp_dst=4789"
881		tnl_md="tunnel\\(tun_id=${id},${addrs},ttl=${ttl},${ports},${flags}\\)"
882
883		ovs_sbx "${sbxname}" grep -qE "MISS upcall.*${tnl_md}.*${arp_hdr}" \
884			${ovs_dir}/ovs-vxlan0.out || return 1
885	done <<< "${configs}"
886
887	return 0
888}
889
890test_tunnel_refcount() {
891	sbxname="test_tunnel_refcount"
892	sbx_add "${sbxname}" || return 1
893
894	ovs_sbx "${sbxname}" ip netns add trefns || return 1
895	on_exit "ovs_sbx ${sbxname} ip netns del trefns"
896
897	for tun_type in gre vxlan geneve; do
898		info "testing ${tun_type} tunnel vport refcount"
899
900		ovs_sbx "${sbxname}" ip netns exec trefns \
901			python3 $ovs_base/ovs-dpctl.py \
902			add-dp dp-${tun_type} || return 1
903
904		ovs_sbx "${sbxname}" ip netns exec trefns \
905			python3 $ovs_base/ovs-dpctl.py \
906			add-if --no-lwt -t ${tun_type} \
907			dp-${tun_type} ovs-${tun_type}0 || return 1
908
909		ovs_wait ip -netns trefns link show \
910			ovs-${tun_type}0 >/dev/null 2>&1 || return 1
911
912		info "deleting dp - may hang if reference counting is broken"
913		ovs_sbx "${sbxname}" ip netns exec trefns \
914			python3 $ovs_base/ovs-dpctl.py \
915			del-dp dp-${tun_type} &
916
917		dev_removed() {
918			! ip -netns trefns link show "$1" >/dev/null 2>&1
919		}
920		ovs_wait dev_removed dp-${tun_type} || return 1
921		ovs_wait dev_removed ovs-${tun_type}0 || return 1
922	done
923
924	return 0
925}
926
927test_pop_vlan() {
928	local sbx="test_pop_vlan"
929	sbx_add "$sbx" || return $?
930	ovs_add_dp "$sbx" vlandp || return 1
931
932	ovs_add_netns_and_veths "$sbx" vlandp \
933		ns1 veth1 ns1veth 192.0.2.1/24 || return 1
934	ovs_add_netns_and_veths "$sbx" vlandp \
935		ns2 veth2 ns2veth 192.0.2.2/24 || return 1
936
937	# Baseline: untagged bidirectional forwarding
938	ovs_add_flow "$sbx" vlandp \
939		'in_port(1),eth(),eth_type(0x0806),arp()' '2' || return 1
940	ovs_add_flow "$sbx" vlandp \
941		'in_port(2),eth(),eth_type(0x0806),arp()' '1' || return 1
942	ovs_add_flow "$sbx" vlandp \
943		'in_port(1),eth(),eth_type(0x0800),ipv4()' '2' || return 1
944	ovs_add_flow "$sbx" vlandp \
945		'in_port(2),eth(),eth_type(0x0800),ipv4()' '1' || return 1
946	ovs_sbx "$sbx" ip netns exec ns1 ping -c 3 -W 2 \
947		192.0.2.2 || return 1
948
949	# VLAN topology: ns1 uses VLAN sub-interface, ns2 is plain
950	ip -n ns1 link add link ns1veth name ns1veth.10 \
951		type vlan id 10 || return 1
952	on_exit "ip -n ns1 link del ns1veth.10 2>/dev/null"
953	ip -n ns1 addr add 198.51.100.1/24 dev ns1veth.10 || return 1
954	ip -n ns1 link set ns1veth.10 up || return 1
955	ip -n ns2 addr add 198.51.100.2/24 dev ns2veth || return 1
956
957	ovs_del_flows "$sbx" vlandp
958
959	# Static ARP: avoids VLAN-tagged ARP complexity
960	local ns1veth10mac ns2mac
961	ns1veth10mac=$(ip -n ns1 link show ns1veth.10 \
962		| awk '/link\/ether/ {print $2}')
963	[ -z "$ns1veth10mac" ] && \
964		{ info "failed to get ns1veth10mac"; return 1; }
965	ns2mac=$(ip -n ns2 link show ns2veth \
966		| awk '/link\/ether/ {print $2}')
967	[ -z "$ns2mac" ] && \
968		{ info "failed to get ns2mac"; return 1; }
969	ip -n ns1 neigh replace 198.51.100.2 lladdr "$ns2mac" \
970		dev ns1veth.10 nud permanent || return 1
971	ip -n ns2 neigh replace 198.51.100.1 \
972		lladdr "$ns1veth10mac" \
973		dev ns2veth nud permanent || return 1
974
975	local vlan_match='in_port(1),eth(),eth_type(0x8100),'
976	vlan_match+='vlan(vid=10),'
977	vlan_match+='encap(eth_type(0x0800),'
978	vlan_match+='ipv4(src=198.51.100.1,proto=1),icmp())'
979
980	# Negative: forward without pop_vlan -- tagged frame
981	# is invisible to ns2 (no VLAN sub-interface), ping fails
982	ovs_add_flow "$sbx" vlandp "$vlan_match" '2' || return 1
983	ovs_sbx "$sbx" ip netns exec ns1 ping -I ns1veth.10 \
984		-c 3 -W 1 198.51.100.2 >/dev/null 2>&1 \
985		&& { info "FAIL: ping should fail without pop_vlan"
986		     return 1; }
987
988	ovs_del_flows "$sbx" vlandp
989
990	# Positive: pop_vlan strips tag on forward path,
991	# push_vlan restores tag on return path -- ping succeeds
992	ovs_add_flow "$sbx" vlandp \
993		"$vlan_match" 'pop_vlan,2' || return 1
994	ovs_add_flow "$sbx" vlandp \
995		'in_port(2),eth(),eth_type(0x0800),ipv4()' \
996		'push_vlan(vid=10,pcp=0,tpid=0x8100),1' || return 1
997	ovs_sbx "$sbx" ip netns exec ns1 ping -I ns1veth.10 \
998		-c 3 -W 2 198.51.100.2 || return 1
999
1000	return 0
1001}
1002
1003run_test() {
1004	(
1005	tname="$1"
1006	tdesc="$2"
1007
1008	if python3 ovs-dpctl.py -h 2>&1 | \
1009	     grep -E "Need to (install|upgrade) the python" >/dev/null 2>&1; then
1010		stdbuf -o0 printf "TEST: %-60s  [PYLIB]\n" "${tdesc}"
1011		return $ksft_skip
1012	fi
1013
1014	python3 ovs-dpctl.py show >/dev/null 2>&1 || \
1015		echo "[DPCTL] show exception."
1016
1017	if ! lsmod | grep openvswitch >/dev/null 2>&1; then
1018		stdbuf -o0 printf "TEST: %-60s  [NOMOD]\n" "${tdesc}"
1019		return $ksft_skip
1020	fi
1021
1022	printf "TEST: %-60s  [START]\n" "${tname}"
1023
1024	unset IFS
1025
1026	eval test_${tname}
1027	ret=$?
1028
1029	if [ $ret -eq 0 ]; then
1030		printf "TEST: %-60s  [ OK ]\n" "${tdesc}"
1031		ovs_exit_sig
1032		rm -rf "$ovs_dir"
1033	elif [ $ret -eq 1 ]; then
1034		printf "TEST: %-60s  [FAIL]\n" "${tdesc}"
1035		if [ "${PAUSE_ON_FAIL}" = "yes" ]; then
1036			echo
1037			echo "Pausing. Logs in $ovs_dir/. Hit enter to continue"
1038			read a
1039		fi
1040		ovs_exit_sig
1041		[ "${PAUSE_ON_FAIL}" = "yes" ] || rm -rf "$ovs_dir"
1042		exit 1
1043	elif [ $ret -eq $ksft_skip ]; then
1044		printf "TEST: %-60s  [SKIP]\n" "${tdesc}"
1045	elif [ $ret -eq 2 ]; then
1046		rm -rf test_${tname}
1047		run_test "$1" "$2"
1048	fi
1049
1050	return $ret
1051	)
1052	ret=$?
1053	case $ret in
1054		0)
1055			[ $all_skipped = true ] && [ $exitcode=$ksft_skip ] && exitcode=0
1056			all_skipped=false
1057		;;
1058		$ksft_skip)
1059			[ $all_skipped = true ] && exitcode=$ksft_skip
1060		;;
1061		*)
1062			all_skipped=false
1063			exitcode=1
1064		;;
1065	esac
1066
1067	return $ret
1068}
1069
1070
1071exitcode=0
1072desc=0
1073all_skipped=true
1074
1075while getopts :pvt o
1076do
1077	case $o in
1078	p) PAUSE_ON_FAIL=yes;;
1079	v) VERBOSE=1;;
1080	t) if which tcpdump > /dev/null 2>&1; then
1081		TRACING=1
1082	   else
1083		echo "=== tcpdump not available, tracing disabled"
1084	   fi
1085	   ;;
1086	*) usage;;
1087	esac
1088done
1089shift $(($OPTIND-1))
1090
1091IFS="
1092"
1093
1094for arg do
1095	# Check first that all requested tests are available before running any
1096	command -v > /dev/null "test_${arg}" || { echo "=== Test ${arg} not found"; usage; }
1097done
1098
1099name=""
1100desc=""
1101for t in ${tests}; do
1102	[ "${name}" = "" ]	&& name="${t}"	&& continue
1103	[ "${desc}" = "" ]	&& desc="${t}"
1104
1105	run_this=1
1106	for arg do
1107		[ "${arg}" != "${arg#--*}" ] && continue
1108		[ "${arg}" = "${name}" ] && run_this=1 && break
1109		run_this=0
1110	done
1111	if [ $run_this -eq 1 ]; then
1112		run_test "${name}" "${desc}"
1113	fi
1114	name=""
1115	desc=""
1116done
1117
1118exit ${exitcode}
1119