freebsd-dev/tests/sys/netpfil/pf/ether.sh
Kristof Provost 07ffa50ba0 pf tests: test short packets
Test sending very short packets (i.e. too short for an IP header)
packets in the Ethernet filtering code.

Sponsored by:	Rubicon Communications, LLC ("Netgate")
2022-06-28 10:31:23 +02:00

696 lines
17 KiB
Bash

# $FreeBSD$
#
# SPDX-License-Identifier: BSD-2-Clause-FreeBSD
#
# Copyright © 2021. Rubicon Communications, LLC (Netgate). All Rights Reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# 1. Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# 2. Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
#
# THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
# ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
# ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
# FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
# DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
# OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
# HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
# LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
# OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
# SUCH DAMAGE.
. $(atf_get_srcdir)/utils.subr
atf_test_case "mac" "cleanup"
mac_head()
{
atf_set descr 'Test MAC address filtering'
atf_set require.user root
}
mac_body()
{
pft_init
epair=$(vnet_mkepair)
epair_a_mac=$(ifconfig ${epair}a ether | awk '/ether/ { print $2; }')
ifconfig ${epair}a 192.0.2.1/24 up
vnet_mkjail alcatraz ${epair}b
jexec alcatraz ifconfig ${epair}b 192.0.2.2/24 up
pft_set_rules alcatraz \
"ether block from ${epair_a_mac}"
atf_check -s exit:0 -o ignore ping -c 1 -t 1 192.0.2.2
# Now enable. Ping should fail.
jexec alcatraz pfctl -e
atf_check -s exit:2 -o ignore ping -c 1 -t 1 192.0.2.2
# Should still fail for 'to'
pft_set_rules alcatraz \
"ether block to ${epair_a_mac}"
atf_check -s exit:2 -o ignore ping -c 1 -t 1 192.0.2.2
# Succeeds if we block a different MAC address
pft_set_rules alcatraz \
"ether block to 00:01:02:03:04:05"
atf_check -s exit:0 -o ignore ping -c 1 -t 1 192.0.2.2
# Should still fail for 'to', even if it's in a list
pft_set_rules alcatraz \
"ether block to { ${epair_a_mac}, 00:01:02:0:04:05 }"
atf_check -s exit:2 -o ignore ping -c 1 -t 1 192.0.2.2
# Now try this with an interface specified
pft_set_rules alcatraz \
"ether block on ${epair}b from ${epair_a_mac}"
atf_check -s exit:2 -o ignore ping -c 1 -t 1 192.0.2.2
# Wrong interface should not match
pft_set_rules alcatraz \
"ether block on ${epair}a from ${epair_a_mac}"
atf_check -s exit:0 -o ignore ping -c 1 -t 1 192.0.2.2
# Test negation
pft_set_rules alcatraz \
"ether block in on ${epair}b from ! ${epair_a_mac}"
atf_check -s exit:0 -o ignore ping -c 1 -t 1 192.0.2.2
pft_set_rules alcatraz \
"ether block out on ${epair}b to ! ${epair_a_mac}"
atf_check -s exit:0 -o ignore ping -c 1 -t 1 192.0.2.2
# Block everything not us
pft_set_rules alcatraz \
"ether block out on ${epair}b to { ! ${epair_a_mac} }"
atf_check -s exit:0 -o ignore ping -c 1 -t 1 192.0.2.2
# Block us now
pft_set_rules alcatraz \
"ether block out on ${epair}b to { ! 00:01:02:03:04:05 }"
atf_check -s exit:2 -o ignore ping -c 1 -t 1 192.0.2.2
# Block with a masked address
pft_set_rules alcatraz \
"ether block out on ${epair}b to { ! 00:01:02:03:00:00/32 }"
jexec alcatraz pfctl -se
atf_check -s exit:2 -o ignore ping -c 1 -t 1 192.0.2.2
epair_prefix=$(echo $epair_a_mac | cut -c-8)
pft_set_rules alcatraz \
"ether block out on ${epair}b to { ${epair_prefix}:00:00:00/24 }"
atf_check -s exit:2 -o ignore ping -c 1 -t 1 192.0.2.2
pft_set_rules alcatraz \
"ether block out on ${epair}b to { ${epair_prefix}:00:00:00&ff:ff:ff:00:00:00 }"
atf_check -s exit:2 -o ignore ping -c 1 -t 1 192.0.2.2
# Check '-F ethernet' works
jexec alcatraz pfctl -F ethernet
atf_check -s exit:0 -o ignore ping -c 1 -t 1 192.0.2.2
}
mac_cleanup()
{
pft_cleanup
}
atf_test_case "proto" "cleanup"
proto_head()
{
atf_set descr 'Test EtherType filtering'
atf_set require.user root
}
proto_body()
{
pft_init
epair=$(vnet_mkepair)
epair_a_mac=$(ifconfig ${epair}a ether | awk '/ether/ { print $2; }')
ifconfig ${epair}a 192.0.2.1/24 up
vnet_mkjail alcatraz ${epair}b
jexec alcatraz ifconfig ${epair}b 192.0.2.2/24 up
pft_set_rules alcatraz \
"ether block proto 0x0810"
jexec alcatraz pfctl -e
atf_check -s exit:0 -o ignore ping -c 1 -t 1 192.0.2.2
# Block IP
pft_set_rules alcatraz \
"ether block proto 0x0800"
atf_check -s exit:2 -o ignore ping -c 1 -t 1 192.0.2.2
# Block ARP
pft_set_rules alcatraz \
"ether block proto 0x0806"
arp -d 192.0.2.2
atf_check -s exit:2 -o ignore ping -c 1 -t 1 192.0.2.2
}
proto_cleanup()
{
pft_cleanup
}
atf_test_case "direction" "cleanup"
direction_head()
{
atf_set descr 'Test directionality of ether rules'
atf_set require.user root
atf_set require.progs jq
}
direction_body()
{
pft_init
epair=$(vnet_mkepair)
epair_a_mac=$(ifconfig ${epair}a ether | awk '/ether/ { print $2; }')
epair_b_mac=$(ifconfig ${epair}b ether | awk '/ether/ { print $2; }')
ifconfig ${epair}a 192.0.2.1/24 up
vnet_mkjail alcatraz ${epair}b
jexec alcatraz ifconfig ${epair}b 192.0.2.2/24 up
pft_set_rules alcatraz \
"ether block in proto 0x0806"
jexec alcatraz pfctl -e
arp -d 192.0.2.2
jexec alcatraz arp -d 192.0.2.1
# We don't allow the jail to receive ARP requests, so if we try to ping
# from host to jail the host can't resolve the MAC address
ping -c 1 -t 1 192.0.2.2
mac=$(arp -an --libxo json \
| jq '."arp"."arp-cache"[] |
select(."ip-address"=="192.0.2.2")."mac-address"')
atf_check_not_equal "$mac" "$epair_b_mac"
# Clear ARP table again
arp -d 192.0.2.2
jexec alcatraz arp -d 192.0.2.1
# However, we allow outbound ARP, so the host will learn our MAC if the
# jail tries to ping
jexec alcatraz ping -c 1 -t 1 192.0.2.1
mac=$(arp -an --libxo json \
| jq '."arp"."arp-cache"[] |
select(."ip-address"=="192.0.2.2")."mac-address"')
atf_check_equal "$mac" "$epair_b_mac"
# Now do the same, but with outbound ARP blocking
pft_set_rules alcatraz \
"ether block out proto 0x0806"
# Clear ARP table again
arp -d 192.0.2.2
jexec alcatraz arp -d 192.0.2.1
# The jail can't send ARP requests to us, so we'll never learn our MAC
# address
jexec alcatraz ping -c 1 -t 1 192.0.2.1
mac=$(jexec alcatraz arp -an --libxo json \
| jq '."arp"."arp-cache"[] |
select(."ip-address"=="192.0.2.1")."mac-address"')
atf_check_not_equal "$mac" "$epair_a_mac"
}
direction_cleanup()
{
pft_cleanup
}
atf_test_case "captive" "cleanup"
captive_head()
{
atf_set descr 'Test a basic captive portal-like setup'
atf_set require.user root
}
captive_body()
{
# Host is client, jail 'gw' is the captive portal gateway, jail 'srv'
# is a random (web)server. We use the echo protocol rather than http
# for the test, because that's easier.
pft_init
epair_gw=$(vnet_mkepair)
epair_srv=$(vnet_mkepair)
epair_gw_a_mac=$(ifconfig ${epair_gw}a ether | awk '/ether/ { print $2; }')
vnet_mkjail gw ${epair_gw}b ${epair_srv}a
vnet_mkjail srv ${epair_srv}b
ifconfig ${epair_gw}a 192.0.2.2/24 up
route add -net 198.51.100.0/24 192.0.2.1
jexec gw ifconfig ${epair_gw}b 192.0.2.1/24 up
jexec gw ifconfig lo0 127.0.0.1/8 up
jexec gw sysctl net.inet.ip.forwarding=1
jexec gw ifconfig ${epair_srv}a 198.51.100.1/24 up
jexec srv ifconfig ${epair_srv}b 198.51.100.2/24 up
jexec srv route add -net 192.0.2.0/24 198.51.100.1
# Sanity check
atf_check -s exit:0 -o ignore ping -c 1 -t 1 198.51.100.2
pft_set_rules gw \
"ether pass quick proto 0x0806" \
"ether pass tag captive" \
"rdr on ${epair_gw}b proto tcp to port echo tagged captive -> 127.0.0.1 port echo"
jexec gw pfctl -e
# ICMP should still work, because we don't redirect it.
atf_check -s exit:0 -o ignore ping -c 1 -t 1 198.51.100.2
# Run the echo server only on the gw, so we know we've redirectly
# correctly if we get an echo message.
jexec gw /usr/sbin/inetd $(atf_get_srcdir)/echo_inetd.conf
# Confirm that we're getting redirected
atf_check -s exit:0 -o match:"^foo$" -x "echo foo | nc -N 198.51.100.2 7"
jexec gw killall inetd
# Now pretend we've authenticated, so add the client's MAC address
pft_set_rules gw \
"ether pass quick proto 0x0806" \
"ether pass quick from ${epair_gw_a_mac}" \
"ether pass tag captive" \
"rdr on ${epair_gw}b proto tcp to port echo tagged captive -> 127.0.0.1 port echo"
# No redirect, so failure.
atf_check -s exit:1 -x "echo foo | nc -N 198.51.100.2 7"
# Start a server in srv
jexec srv /usr/sbin/inetd $(atf_get_srcdir)/echo_inetd.conf
# And now we can talk to that one.
atf_check -s exit:0 -o match:"^foo$" -x "echo foo | nc -N 198.51.100.2 7"
}
captive_cleanup()
{
pft_cleanup
}
atf_test_case "captive_long" "cleanup"
captive_long_head()
{
atf_set descr 'More complex captive portal setup'
atf_set require.user root
}
captive_long_body()
{
# Host is client, jail 'gw' is the captive portal gateway, jail 'srv'
# is a random (web)server. We use the echo protocol rather than http
# for the test, because that's easier.
dummynet_init
epair_gw=$(vnet_mkepair)
epair_srv=$(vnet_mkepair)
epair_gw_a_mac=$(ifconfig ${epair_gw}a ether | awk '/ether/ { print $2; }')
vnet_mkjail gw ${epair_gw}b ${epair_srv}a
vnet_mkjail srv ${epair_srv}b
ifconfig ${epair_gw}a 192.0.2.2/24 up
route add -net 198.51.100.0/24 192.0.2.1
jexec gw ifconfig ${epair_gw}b 192.0.2.1/24 up
jexec gw ifconfig lo0 127.0.0.1/8 up
jexec gw sysctl net.inet.ip.forwarding=1
jexec gw ifconfig ${epair_srv}a 198.51.100.1/24 up
jexec srv ifconfig ${epair_srv}b 198.51.100.2/24 up
jexec srv route add -net 192.0.2.0/24 198.51.100.1
jexec gw dnctl pipe 1 config bw 300KByte/s
# Sanity check
atf_check -s exit:0 -o ignore ping -c 1 -t 1 198.51.100.2
pft_set_rules gw \
"ether anchor \"captiveportal\" on { ${epair_gw}b } {" \
"ether pass quick proto { 0x0806, 0x8035, 0x888e, 0x88c7, 0x8863, 0x8864 }" \
"ether pass tag \"captive\"" \
"}" \
"rdr on ${epair_gw}b proto tcp to port daytime tagged captive -> 127.0.0.1 port echo"
jexec gw pfctl -e
# ICMP should still work, because we don't redirect it.
atf_check -s exit:0 -o ignore ping -c 1 -t 1 198.51.100.2
jexec gw /usr/sbin/inetd -p gw.pid $(atf_get_srcdir)/echo_inetd.conf
jexec srv /usr/sbin/inetd -p srv.pid $(atf_get_srcdir)/daytime_inetd.conf
echo foo | nc -N 198.51.100.2 13
# Confirm that we're getting redirected
atf_check -s exit:0 -o match:"^foo$" -x "echo foo | nc -N 198.51.100.2 13"
# Now update the rules to allow our client to pass without redirect
pft_set_rules gw \
"ether anchor \"captiveportal\" on { ${epair_gw}b } {" \
"ether pass quick proto { 0x0806, 0x8035, 0x888e, 0x88c7, 0x8863, 0x8864 }" \
"ether pass quick from { ${epair_gw_a_mac} } dnpipe 1" \
"ether pass tag \"captive\"" \
"}" \
"rdr on ${epair_gw}b proto tcp to port daytime tagged captive -> 127.0.0.1 port echo"
# We're not being redirected and get datime information now
atf_check -s exit:0 -o match:"^(Mon|Tue|Wed|Thu|Fri|Sat|Sun)" -x "echo foo | nc -N 198.51.100.2 13"
jexec gw killall inetd
jexec srv killall inetd
}
captive_long_cleanup()
{
pft_cleanup
}
atf_test_case "dummynet" "cleanup"
dummynet_head()
{
atf_set descr 'Test dummynet for L2 traffic'
atf_set require.user root
}
dummynet_body()
{
pft_init
if ! kldstat -q -m dummynet; then
atf_skip "This test requires dummynet"
fi
epair=$(vnet_mkepair)
vnet_mkjail alcatraz ${epair}b
ifconfig ${epair}a 192.0.2.1/24 up
jexec alcatraz ifconfig ${epair}b 192.0.2.2/24 up
# Sanity check
atf_check -s exit:0 -o ignore ping -i .1 -c 3 -s 1200 192.0.2.2
jexec alcatraz dnctl pipe 1 config bw 30Byte/s
jexec alcatraz pfctl -e
pft_set_rules alcatraz \
"ether pass in dnpipe 1"
# single ping succeeds just fine
atf_check -s exit:0 -o ignore ping -c 1 192.0.2.2
# Saturate the link
ping -i .1 -c 5 -s 1200 192.0.2.2
# We should now be hitting the limits and get this packet dropped.
atf_check -s exit:2 -o ignore ping -c 1 -s 1200 192.0.2.2
# We can now also dummynet outbound traffic!
pft_set_rules alcatraz \
"ether pass out dnpipe 1"
# We should still be hitting the limits and get this packet dropped.
atf_check -s exit:2 -o ignore ping -c 1 -s 1200 192.0.2.2
}
dummynet_cleanup()
{
pft_cleanup
}
atf_test_case "anchor" "cleanup"
anchor_head()
{
atf_set descr 'Test ether anchors'
atf_set require.user root
}
anchor_body()
{
pft_init
epair=$(vnet_mkepair)
epair_a_mac=$(ifconfig ${epair}a ether | awk '/ether/ { print $2; }')
vnet_mkjail alcatraz ${epair}b
ifconfig ${epair}a 192.0.2.1/24 up
jexec alcatraz ifconfig ${epair}b 192.0.2.2/24 up
# Sanity check
atf_check -s exit:0 -o ignore ping -c 1 192.0.2.2
jexec alcatraz pfctl -e
pft_set_rules alcatraz \
"ether anchor \"foo\" in on lo0 {" \
"ether block" \
"}"
# That only filters on lo0, so we should still be able to pass traffic
atf_check -s exit:0 -o ignore ping -c 1 192.0.2.2
pft_set_rules alcatraz \
"ether block in" \
"ether anchor \"foo\" in on ${epair}b {" \
"ether pass" \
"}"
atf_check -s exit:0 -o ignore ping -c 1 192.0.2.2
pft_set_rules alcatraz \
"ether pass" \
"ether anchor \"bar\" in on ${epair}b {" \
"ether block" \
"}"
atf_check -s exit:2 -o ignore ping -c 1 -t 2 192.0.2.2
pft_set_rules alcatraz \
"ether block in" \
"ether anchor \"baz\" on ${epair}b {" \
"ether pass in from 01:02:03:04:05:06" \
"}" \
"ether pass in from ${epair_a_mac}"
atf_check -s exit:0 -o ignore ping -c 1 192.0.2.2
atf_check -s exit:0 -o match:'baz' jexec alcatraz pfctl -sA
}
anchor_cleanup()
{
pft_cleanup
}
atf_test_case "ip" "cleanup"
ip_head()
{
atf_set descr 'Test filtering based on IP source/destination'
atf_set require.user root
}
ip_body()
{
pft_init
epair=$(vnet_mkepair)
vnet_mkjail alcatraz ${epair}b
ifconfig ${epair}a 192.0.2.1/24 up
jexec alcatraz ifconfig ${epair}b 192.0.2.2/24 up
# Sanity check
atf_check -s exit:0 -o ignore ping -c 1 192.0.2.2
jexec alcatraz pfctl -e
pft_set_rules alcatraz \
"ether pass" \
"ether block in l3 from 192.0.2.1"
atf_check -s exit:2 -o ignore ping -c 1 192.0.2.2
# Change IP address and we can ping again
ifconfig ${epair}a 192.0.2.3/24 up
atf_check -s exit:0 -o ignore ping -c 1 192.0.2.2
# Test the 'to' keyword too
pft_set_rules alcatraz \
"ether pass" \
"ether block out l3 to 192.0.2.3"
atf_check -s exit:2 -o ignore ping -c 1 192.0.2.2
# Test table
pft_set_rules alcatraz \
"table <tbl> { 192.0.2.3 }" \
"ether pass" \
"ether block out l3 to <tbl>"
atf_check -s exit:2 -o ignore ping -c 1 192.0.2.2
}
ip_cleanup()
{
pft_cleanup
}
atf_test_case "tag" "cleanup"
tag_head()
{
atf_set descr 'Test setting tags'
atf_set require.user root
}
tag_body()
{
pft_init
epair=$(vnet_mkepair)
vnet_mkjail alcatraz ${epair}b
ifconfig ${epair}a 192.0.2.1/24 up
jexec alcatraz ifconfig ${epair}b 192.0.2.2/24 up
# Sanity check
atf_check -s exit:0 -o ignore ping -c 1 192.0.2.2
jexec alcatraz pfctl -e
pft_set_rules alcatraz \
"ether pass in tag foo" \
"block in tagged foo"
atf_check -s exit:2 -o ignore ping -c 1 192.0.2.2
pft_set_rules alcatraz \
"ether pass in tag bar" \
"block in tagged foo"
# Still passes when tagged differently
atf_check -s exit:0 -o ignore ping -c 1 192.0.2.2
}
tag_cleanup()
{
pft_cleanup
}
atf_test_case "match_tag" "cleanup"
match_tag_head()
{
atf_set descr 'Test matching tags'
atf_set require.user root
}
match_tag_body()
{
pft_init
epair=$(vnet_mkepair)
vnet_mkjail alcatraz ${epair}b
ifconfig ${epair}a 192.0.2.1/24 up
jexec alcatraz ifconfig ${epair}b 192.0.2.2/24 up
# Sanity check
atf_check -s exit:0 -o ignore ping -c 1 192.0.2.2
jexec alcatraz pfctl -e
pft_set_rules alcatraz \
"ether block out tagged foo" \
"pass in proto icmp tag foo"
atf_check -s exit:2 -o ignore ping -c 1 192.0.2.2
pft_set_rules alcatraz \
"ether block out tagged bar" \
"pass in proto icmp tag foo"
# Still passes when tagged differently
atf_check -s exit:0 -o ignore ping -c 1 192.0.2.2
}
match_tag_cleanup()
{
pft_cleanup
}
atf_test_case "short_pkt" "cleanup"
short_pkt_head()
{
atf_set descr 'Test overly short Ethernet packets'
atf_set require.user root
}
short_pkt_body()
{
pft_init
epair=$(vnet_mkepair)
ifconfig ${epair}a 192.0.2.1/24 up
vnet_mkjail alcatraz ${epair}b
jexec alcatraz ifconfig ${epair}b 192.0.2.2/24 up
jexec alcatraz pfctl -e
pft_set_rules alcatraz \
"ether pass in" \
"ether pass out" \
"ether pass in l3 from 192.0.2.1"
# Sanity check
atf_check -s exit:0 -o ignore ping -c 1 192.0.2.2
jexec alcatraz pfctl -se -v
# Try sending ever shorter ping requests
# BPF won't let us send anything shorter than an Ethernet header, but
# that's good enough for this test
for i in `seq 46 14`
do
$(atf_get_srcdir)/pft_ether.py \
--sendif ${epair}a \
--to 192.0.2.2 \
--len ${i}
done
}
short_pkt_cleanup()
{
pft_cleanup
}
atf_init_test_cases()
{
atf_add_test_case "mac"
atf_add_test_case "proto"
atf_add_test_case "direction"
atf_add_test_case "captive"
atf_add_test_case "captive_long"
atf_add_test_case "dummynet"
atf_add_test_case "anchor"
atf_add_test_case "ip"
atf_add_test_case "tag"
atf_add_test_case "match_tag"
atf_add_test_case "short_pkt"
}