OSDN Git Service

selftests: Add tests for vrf and xfrms
authorDavid Ahern <dsahern@gmail.com>
Wed, 22 Apr 2020 00:40:22 +0000 (18:40 -0600)
committerDavid S. Miller <davem@davemloft.net>
Thu, 23 Apr 2020 02:31:40 +0000 (19:31 -0700)
Add tests for vrf and xfrms with a second round after adding a
qdisc. There are a few known problems documented with the test
cases that fail. The fix is non-trivial; will come back to it
when time allows.

Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
tools/testing/selftests/net/Makefile
tools/testing/selftests/net/vrf-xfrm-tests.sh [new file with mode: 0755]

index 3f386eb..895ec99 100644 (file)
@@ -16,6 +16,7 @@ TEST_PROGS += altnames.sh icmp_redirect.sh ip6_gre_headroom.sh
 TEST_PROGS += route_localnet.sh
 TEST_PROGS += reuseaddr_ports_exhausted.sh
 TEST_PROGS += txtimestamp.sh
+TEST_PROGS += vrf-xfrm-tests.sh
 TEST_PROGS_EXTENDED := in_netns.sh
 TEST_GEN_FILES =  socket nettest
 TEST_GEN_FILES += psock_fanout psock_tpacket msg_zerocopy reuseport_addr_any
diff --git a/tools/testing/selftests/net/vrf-xfrm-tests.sh b/tools/testing/selftests/net/vrf-xfrm-tests.sh
new file mode 100755 (executable)
index 0000000..184da81
--- /dev/null
@@ -0,0 +1,436 @@
+#!/bin/bash
+# SPDX-License-Identifier: GPL-2.0
+#
+# Various combinations of VRF with xfrms and qdisc.
+
+# Kselftest framework requirement - SKIP code is 4.
+ksft_skip=4
+
+PAUSE_ON_FAIL=no
+VERBOSE=0
+ret=0
+
+HOST1_4=192.168.1.1
+HOST2_4=192.168.1.2
+HOST1_6=2001:db8:1::1
+HOST2_6=2001:db8:1::2
+
+XFRM1_4=10.0.1.1
+XFRM2_4=10.0.1.2
+XFRM1_6=fc00:1000::1
+XFRM2_6=fc00:1000::2
+IF_ID=123
+
+VRF=red
+TABLE=300
+
+AUTH_1=0xd94fcfea65fddf21dc6e0d24a0253508
+AUTH_2=0xdc6e0d24a0253508d94fcfea65fddf21
+ENC_1=0xfc46c20f8048be9725930ff3fb07ac2a91f0347dffeacf62
+ENC_2=0x3fb07ac2a91f0347dffeacf62fc46c20f8048be9725930ff
+SPI_1=0x02122b77
+SPI_2=0x2b770212
+
+which ping6 > /dev/null 2>&1 && ping6=$(which ping6) || ping6=$(which ping)
+
+################################################################################
+#
+log_test()
+{
+       local rc=$1
+       local expected=$2
+       local msg="$3"
+
+       if [ ${rc} -eq ${expected} ]; then
+               printf "TEST: %-60s  [ OK ]\n" "${msg}"
+               nsuccess=$((nsuccess+1))
+       else
+               ret=1
+               nfail=$((nfail+1))
+               printf "TEST: %-60s  [FAIL]\n" "${msg}"
+               if [ "${PAUSE_ON_FAIL}" = "yes" ]; then
+                       echo
+                       echo "hit enter to continue, 'q' to quit"
+                       read a
+                       [ "$a" = "q" ] && exit 1
+               fi
+       fi
+}
+
+run_cmd_host1()
+{
+       local cmd="$*"
+       local out
+       local rc
+
+       if [ "$VERBOSE" = "1" ]; then
+               printf "    COMMAND: $cmd\n"
+       fi
+
+       out=$(eval ip netns exec host1 $cmd 2>&1)
+       rc=$?
+       if [ "$VERBOSE" = "1" ]; then
+               if [ -n "$out" ]; then
+                       echo
+                       echo "    $out"
+               fi
+               echo
+       fi
+
+       return $rc
+}
+
+################################################################################
+# create namespaces for hosts and sws
+
+create_vrf()
+{
+       local ns=$1
+       local vrf=$2
+       local table=$3
+
+       if [ -n "${ns}" ]; then
+               ns="-netns ${ns}"
+       fi
+
+       ip ${ns} link add ${vrf} type vrf table ${table}
+       ip ${ns} link set ${vrf} up
+       ip ${ns} route add vrf ${vrf} unreachable default metric 8192
+       ip ${ns} -6 route add vrf ${vrf} unreachable default metric 8192
+
+       ip ${ns} addr add 127.0.0.1/8 dev ${vrf}
+       ip ${ns} -6 addr add ::1 dev ${vrf} nodad
+
+       ip ${ns} ru del pref 0
+       ip ${ns} ru add pref 32765 from all lookup local
+       ip ${ns} -6 ru del pref 0
+       ip ${ns} -6 ru add pref 32765 from all lookup local
+}
+
+create_ns()
+{
+       local ns=$1
+       local addr=$2
+       local addr6=$3
+
+       [ -z "${addr}" ] && addr="-"
+       [ -z "${addr6}" ] && addr6="-"
+
+       ip netns add ${ns}
+
+       ip -netns ${ns} link set lo up
+       if [ "${addr}" != "-" ]; then
+               ip -netns ${ns} addr add dev lo ${addr}
+       fi
+       if [ "${addr6}" != "-" ]; then
+               ip -netns ${ns} -6 addr add dev lo ${addr6}
+       fi
+
+       ip -netns ${ns} ro add unreachable default metric 8192
+       ip -netns ${ns} -6 ro add unreachable default metric 8192
+
+       ip netns exec ${ns} sysctl -qw net.ipv4.ip_forward=1
+       ip netns exec ${ns} sysctl -qw net.ipv6.conf.all.keep_addr_on_down=1
+       ip netns exec ${ns} sysctl -qw net.ipv6.conf.all.forwarding=1
+       ip netns exec ${ns} sysctl -qw net.ipv6.conf.default.forwarding=1
+       ip netns exec ${ns} sysctl -qw net.ipv6.conf.default.accept_dad=0
+}
+
+# create veth pair to connect namespaces and apply addresses.
+connect_ns()
+{
+       local ns1=$1
+       local ns1_dev=$2
+       local ns1_addr=$3
+       local ns1_addr6=$4
+       local ns2=$5
+       local ns2_dev=$6
+       local ns2_addr=$7
+       local ns2_addr6=$8
+       local ns1arg
+       local ns2arg
+
+       if [ -n "${ns1}" ]; then
+               ns1arg="-netns ${ns1}"
+       fi
+       if [ -n "${ns2}" ]; then
+               ns2arg="-netns ${ns2}"
+       fi
+
+       ip ${ns1arg} li add ${ns1_dev} type veth peer name tmp
+       ip ${ns1arg} li set ${ns1_dev} up
+       ip ${ns1arg} li set tmp netns ${ns2} name ${ns2_dev}
+       ip ${ns2arg} li set ${ns2_dev} up
+
+       if [ "${ns1_addr}" != "-" ]; then
+               ip ${ns1arg} addr add dev ${ns1_dev} ${ns1_addr}
+               ip ${ns2arg} addr add dev ${ns2_dev} ${ns2_addr}
+       fi
+
+       if [ "${ns1_addr6}" != "-" ]; then
+               ip ${ns1arg} addr add dev ${ns1_dev} ${ns1_addr6} nodad
+               ip ${ns2arg} addr add dev ${ns2_dev} ${ns2_addr6} nodad
+       fi
+}
+
+################################################################################
+
+cleanup()
+{
+       ip netns del host1
+       ip netns del host2
+}
+
+setup()
+{
+       create_ns "host1"
+       create_ns "host2"
+
+       connect_ns "host1" eth0 ${HOST1_4}/24 ${HOST1_6}/64 \
+                  "host2" eth0 ${HOST2_4}/24 ${HOST2_6}/64
+
+       create_vrf "host1" ${VRF} ${TABLE}
+       ip -netns host1 link set dev eth0 master ${VRF}
+}
+
+cleanup_xfrm()
+{
+       for ns in host1 host2
+       do
+               for x in state policy
+               do
+                       ip -netns ${ns} xfrm ${x} flush
+                       ip -6 -netns ${ns} xfrm ${x} flush
+               done
+       done
+}
+
+setup_xfrm()
+{
+       local h1_4=$1
+       local h2_4=$2
+       local h1_6=$3
+       local h2_6=$4
+       local devarg="$5"
+
+       #
+       # policy
+       #
+
+       # host1 - IPv4 out
+       ip -netns host1 xfrm policy add \
+         src ${h1_4} dst ${h2_4} ${devarg} dir out \
+         tmpl src ${HOST1_4} dst ${HOST2_4} proto esp mode tunnel
+
+       # host2 - IPv4 in
+       ip -netns host2 xfrm policy add \
+         src ${h1_4} dst ${h2_4} dir in \
+         tmpl src ${HOST1_4} dst ${HOST2_4} proto esp mode tunnel
+
+       # host1 - IPv4 in
+       ip -netns host1 xfrm policy add \
+         src ${h2_4} dst ${h1_4} ${devarg} dir in \
+         tmpl src ${HOST2_4} dst ${HOST1_4} proto esp mode tunnel
+
+       # host2 - IPv4 out
+       ip -netns host2 xfrm policy add \
+         src ${h2_4} dst ${h1_4} dir out \
+         tmpl src ${HOST2_4} dst ${HOST1_4} proto esp mode tunnel
+
+
+       # host1 - IPv6 out
+       ip -6 -netns host1 xfrm policy add \
+         src ${h1_6} dst ${h2_6} ${devarg} dir out \
+         tmpl src ${HOST1_6} dst ${HOST2_6} proto esp mode tunnel
+
+       # host2 - IPv6 in
+       ip -6 -netns host2 xfrm policy add \
+         src ${h1_6} dst ${h2_6} dir in \
+         tmpl src ${HOST1_6} dst ${HOST2_6} proto esp mode tunnel
+
+       # host1 - IPv6 in
+       ip -6 -netns host1 xfrm policy add \
+         src ${h2_6} dst ${h1_6} ${devarg} dir in \
+         tmpl src ${HOST2_6} dst ${HOST1_6} proto esp mode tunnel
+
+       # host2 - IPv6 out
+       ip -6 -netns host2 xfrm policy add \
+         src ${h2_6} dst ${h1_6} dir out \
+         tmpl src ${HOST2_6} dst ${HOST1_6} proto esp mode tunnel
+
+       #
+       # state
+       #
+       ip -netns host1 xfrm state add src ${HOST1_4} dst ${HOST2_4} \
+           proto esp spi ${SPI_1} reqid 0 mode tunnel \
+           replay-window 4 replay-oseq 0x4 \
+           auth-trunc 'hmac(md5)' ${AUTH_1} 96 \
+           enc 'cbc(des3_ede)' ${ENC_1} \
+           sel src ${h1_4} dst ${h2_4} ${devarg}
+
+       ip -netns host2 xfrm state add src ${HOST1_4} dst ${HOST2_4} \
+           proto esp spi ${SPI_1} reqid 0 mode tunnel \
+           replay-window 4 replay-oseq 0x4 \
+           auth-trunc 'hmac(md5)' ${AUTH_1} 96 \
+           enc 'cbc(des3_ede)' ${ENC_1} \
+           sel src ${h1_4} dst ${h2_4}
+
+
+       ip -netns host1 xfrm state add src ${HOST2_4} dst ${HOST1_4} \
+           proto esp spi ${SPI_2} reqid 0 mode tunnel \
+           replay-window 4 replay-oseq 0x4 \
+           auth-trunc 'hmac(md5)' ${AUTH_2} 96 \
+           enc 'cbc(des3_ede)' ${ENC_2} \
+           sel src ${h2_4} dst ${h1_4} ${devarg}
+
+       ip -netns host2 xfrm state add src ${HOST2_4} dst ${HOST1_4} \
+           proto esp spi ${SPI_2} reqid 0 mode tunnel \
+           replay-window 4 replay-oseq 0x4 \
+           auth-trunc 'hmac(md5)' ${AUTH_2} 96 \
+           enc 'cbc(des3_ede)' ${ENC_2} \
+           sel src ${h2_4} dst ${h1_4}
+
+
+       ip -6 -netns host1 xfrm state add src ${HOST1_6} dst ${HOST2_6} \
+           proto esp spi ${SPI_1} reqid 0 mode tunnel \
+           replay-window 4 replay-oseq 0x4 \
+           auth-trunc 'hmac(md5)' ${AUTH_1} 96 \
+           enc 'cbc(des3_ede)' ${ENC_1} \
+           sel src ${h1_6} dst ${h2_6} ${devarg}
+
+       ip -6 -netns host2 xfrm state add src ${HOST1_6} dst ${HOST2_6} \
+           proto esp spi ${SPI_1} reqid 0 mode tunnel \
+           replay-window 4 replay-oseq 0x4 \
+           auth-trunc 'hmac(md5)' ${AUTH_1} 96 \
+           enc 'cbc(des3_ede)' ${ENC_1} \
+           sel src ${h1_6} dst ${h2_6}
+
+
+       ip -6 -netns host1 xfrm state add src ${HOST2_6} dst ${HOST1_6} \
+           proto esp spi ${SPI_2} reqid 0 mode tunnel \
+           replay-window 4 replay-oseq 0x4 \
+           auth-trunc 'hmac(md5)' ${AUTH_2} 96 \
+           enc 'cbc(des3_ede)' ${ENC_2} \
+           sel src ${h2_6} dst ${h1_6} ${devarg}
+
+       ip -6 -netns host2 xfrm state add src ${HOST2_6} dst ${HOST1_6} \
+           proto esp spi ${SPI_2} reqid 0 mode tunnel \
+           replay-window 4 replay-oseq 0x4 \
+           auth-trunc 'hmac(md5)' ${AUTH_2} 96 \
+           enc 'cbc(des3_ede)' ${ENC_2} \
+           sel src ${h2_6} dst ${h1_6}
+}
+
+cleanup_xfrm_dev()
+{
+       ip -netns host1 li del xfrm0
+       ip -netns host2 addr del ${XFRM2_4}/24 dev eth0
+       ip -netns host2 addr del ${XFRM2_6}/64 dev eth0
+}
+
+setup_xfrm_dev()
+{
+       local vrfarg="vrf ${VRF}"
+
+       ip -netns host1 li add type xfrm dev eth0 if_id ${IF_ID}
+       ip -netns host1 li set xfrm0 ${vrfarg} up
+       ip -netns host1 addr add ${XFRM1_4}/24 dev xfrm0
+       ip -netns host1 addr add ${XFRM1_6}/64 dev xfrm0
+
+       ip -netns host2 addr add ${XFRM2_4}/24 dev eth0
+       ip -netns host2 addr add ${XFRM2_6}/64 dev eth0
+
+       setup_xfrm ${XFRM1_4} ${XFRM2_4} ${XFRM1_6} ${XFRM2_6} "if_id ${IF_ID}"
+}
+
+run_tests()
+{
+       cleanup_xfrm
+
+       # no IPsec
+       run_cmd_host1 ip vrf exec ${VRF} ping -c1 -w1 ${HOST2_4}
+       log_test $? 0 "IPv4 no xfrm policy"
+       run_cmd_host1 ip vrf exec ${VRF} ${ping6} -c1 -w1 ${HOST2_6}
+       log_test $? 0 "IPv6 no xfrm policy"
+
+       # xfrm without VRF in sel
+       setup_xfrm ${HOST1_4} ${HOST2_4} ${HOST1_6} ${HOST2_6}
+       run_cmd_host1 ip vrf exec ${VRF} ping -c1 -w1 ${HOST2_4}
+       log_test $? 0 "IPv4 xfrm policy based on address"
+       run_cmd_host1 ip vrf exec ${VRF} ${ping6} -c1 -w1 ${HOST2_6}
+       log_test $? 0 "IPv6 xfrm policy based on address"
+       cleanup_xfrm
+
+       # xfrm with VRF in sel
+       # Known failure: ipv4 resets the flow oif after the lookup. Fix is
+       # not straightforward.
+       # setup_xfrm ${HOST1_4} ${HOST2_4} ${HOST1_6} ${HOST2_6} "dev ${VRF}"
+       # run_cmd_host1 ip vrf exec ${VRF} ping -c1 -w1 ${HOST2_4}
+       # log_test $? 0 "IPv4 xfrm policy with VRF in selector"
+       run_cmd_host1 ip vrf exec ${VRF} ${ping6} -c1 -w1 ${HOST2_6}
+       log_test $? 0 "IPv6 xfrm policy with VRF in selector"
+       cleanup_xfrm
+
+       # xfrm with enslaved device in sel
+       # Known failures: combined with the above, __xfrm{4,6}_selector_match
+       # needs to consider both l3mdev and enslaved device index.
+       # setup_xfrm ${HOST1_4} ${HOST2_4} ${HOST1_6} ${HOST2_6} "dev eth0"
+       # run_cmd_host1 ip vrf exec ${VRF} ping -c1 -w1 ${HOST2_4}
+       # log_test $? 0 "IPv4 xfrm policy with enslaved device in selector"
+       # run_cmd_host1 ip vrf exec ${VRF} ${ping6} -c1 -w1 ${HOST2_6}
+       # log_test $? 0 "IPv6 xfrm policy with enslaved device in selector"
+       # cleanup_xfrm
+
+       # xfrm device
+       setup_xfrm_dev
+       run_cmd_host1 ip vrf exec ${VRF} ping -c1 -w1 ${XFRM2_4}
+       log_test $? 0 "IPv4 xfrm policy with xfrm device"
+       run_cmd_host1 ip vrf exec ${VRF} ${ping6} -c1 -w1 ${XFRM2_6}
+       log_test $? 0 "IPv6 xfrm policy with xfrm device"
+       cleanup_xfrm_dev
+}
+
+################################################################################
+# usage
+
+usage()
+{
+        cat <<EOF
+usage: ${0##*/} OPTS
+
+        -p          Pause on fail
+        -v          verbose mode (show commands and output)
+
+done
+EOF
+}
+
+################################################################################
+# main
+
+while getopts :pv o
+do
+       case $o in
+               p) PAUSE_ON_FAIL=yes;;
+               v) VERBOSE=$(($VERBOSE + 1));;
+               h) usage; exit 0;;
+               *) usage; exit 1;;
+       esac
+done
+
+cleanup 2>/dev/null
+setup
+
+echo
+echo "No qdisc on VRF device"
+run_tests
+
+run_cmd_host1 tc qdisc add dev ${VRF} root netem delay 100ms
+echo
+echo "netem qdisc on VRF device"
+run_tests
+
+printf "\nTests passed: %3d\n" ${nsuccess}
+printf "Tests failed: %3d\n"   ${nfail}
+
+exit $ret