freebsd-src/tests/sys/netpfil/pf/fragmentation_pass.sh
Kajetan Staszkiewicz 65074f6f31 pf: fix double ruleset evaluation for fragments sent to dummynet
The function `pf_setup_pdesc()` handles ruleset evaluation for non-reassembled
packets. Having it called before `pf_mtag` is checked for flags
`PF_MTAG_FLAG_ROUTE_TO` and `PF_MTAG_FLAG_DUMMYNET` will cause loops for
fragmented packets if reassembly is disabled.

Move `pd` zeroing and `pf_mtag` extraction from `pf_setup_pdesc()` to a separate
function `pf_init_pdesc()` and change the order of function calls: first
call `pf_init_pdesc()`, then check if the currently processed packet has been
reinjected from dummynet, finally call `pf_setup_pdesc()`.

Add functionality of sending UDP packets to `pft_ping.py` with fragmentation
support and fix broken IPv6 reassembly.

Reviewed by:	kp
Differential Revision:	https://reviews.freebsd.org/D46880
2024-10-03 13:49:57 +02:00

613 lines
15 KiB
Bash

#
# SPDX-License-Identifier: BSD-2-Clause
#
# Copyright (c) 2017 Kristof Provost <kp@FreeBSD.org>
#
# 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.
. $(atf_get_srcdir)/utils.subr
common_dir=$(atf_get_srcdir)/../common
atf_test_case "too_many_fragments" "cleanup"
too_many_fragments_head()
{
atf_set descr 'IPv4 fragment limitation test'
atf_set require.user root
}
too_many_fragments_body()
{
pft_init
epair=$(vnet_mkepair)
vnet_mkjail alcatraz ${epair}a
ifconfig ${epair}b inet 192.0.2.1/24 up
jexec alcatraz ifconfig ${epair}a 192.0.2.2/24 up
ifconfig ${epair}b mtu 200
jexec alcatraz ifconfig ${epair}a mtu 200
jexec alcatraz pfctl -e
pft_set_rules alcatraz \
"set reassemble yes" \
"pass keep state"
# So we know pf is limiting things
jexec alcatraz sysctl net.inet.ip.maxfragsperpacket=1024
# Sanity check
atf_check -s exit:0 -o ignore ping -c 1 192.0.2.2
# We can ping with < 64 fragments
atf_check -s exit:0 -o ignore ping -c 1 -s 800 192.0.2.2
# Too many fragments should fail
atf_check -s exit:2 -o ignore ping -c 1 -s 20000 192.0.2.2
}
too_many_fragments_cleanup()
{
pft_cleanup
}
atf_test_case "v6" "cleanup"
v6_head()
{
atf_set descr 'IPv6 fragmentation test'
atf_set require.user root
atf_set require.progs scapy
}
v6_body()
{
pft_init
epair_send=$(vnet_mkepair)
epair_link=$(vnet_mkepair)
vnet_mkjail alcatraz ${epair_send}b ${epair_link}a
vnet_mkjail singsing ${epair_link}b
ifconfig ${epair_send}a inet6 2001:db8:42::1/64 no_dad up
jexec alcatraz ifconfig ${epair_send}b inet6 2001:db8:42::2/64 no_dad up
jexec alcatraz ifconfig ${epair_link}a inet6 2001:db8:43::2/64 no_dad up
jexec alcatraz sysctl net.inet6.ip6.forwarding=1
jexec singsing ifconfig ${epair_link}b inet6 2001:db8:43::3/64 no_dad up
jexec singsing route add -6 2001:db8:42::/64 2001:db8:43::2
route add -6 2001:db8:43::/64 2001:db8:42::2
jexec alcatraz ifconfig ${epair_send}b inet6 -ifdisabled
jexec alcatraz ifconfig ${epair_link}a inet6 -ifdisabled
jexec singsing ifconfig ${epair_link}b inet6 -ifdisabled
ifconfig ${epair_send}a inet6 -ifdisabled
ifconfig ${epair_send}a
jexec alcatraz ifconfig ${epair_send}b
lladdr=$(jexec alcatraz ifconfig ${epair_send}b | awk '/ scopeid / { print($2); }' | cut -f 1 -d %)
jexec alcatraz pfctl -e
pft_set_rules alcatraz \
"set reassemble yes" \
"pass keep state" \
"block in" \
"pass in inet6 proto icmp6 icmp6-type { neighbrsol, neighbradv }" \
"pass in inet6 proto icmp6 icmp6-type { echoreq, echorep }" \
"set skip on lo"
# Host test
atf_check -s exit:0 -o ignore \
ping -6 -c 1 2001:db8:42::2
atf_check -s exit:0 -o ignore \
ping -6 -c 1 -s 4500 2001:db8:42::2
atf_check -s exit:0 -o ignore\
ping -6 -c 1 -b 70000 -s 65000 2001:db8:42::2
# Force an NDP lookup
ping -6 -c 1 ${lladdr}%${epair_send}a
atf_check -s exit:0 -o ignore\
ping -6 -c 1 -b 70000 -s 65000 ${lladdr}%${epair_send}a
# Forwarding test
atf_check -s exit:0 -o ignore \
ping -6 -c 1 2001:db8:43::3
atf_check -s exit:0 -o ignore \
ping -6 -c 1 -s 4500 2001:db8:43::3
atf_check -s exit:0 -o ignore\
ping -6 -c 1 -b 70000 -s 65000 2001:db8:43::3
$(atf_get_srcdir)/CVE-2019-5597.py \
${epair_send}a \
2001:db8:42::1 \
2001:db8:43::3
}
v6_cleanup()
{
pft_cleanup
}
atf_test_case "mtu_diff" "cleanup"
mtu_diff_head()
{
atf_set descr 'Test reassembly across different MTUs, PR #255432'
atf_set require.user root
}
mtu_diff_body()
{
pft_init
epair_small=$(vnet_mkepair)
epair_large=$(vnet_mkepair)
vnet_mkjail first ${epair_small}b ${epair_large}a
vnet_mkjail second ${epair_large}b
ifconfig ${epair_small}a 192.0.2.1/25 up
jexec first ifconfig ${epair_small}b 192.0.2.2/25 up
jexec first sysctl net.inet.ip.forwarding=1
jexec first ifconfig ${epair_large}a 192.0.2.130/25 up
jexec first ifconfig ${epair_large}a mtu 9000
jexec second ifconfig ${epair_large}b 192.0.2.131/25 up
jexec second ifconfig ${epair_large}b mtu 9000
jexec second route add default 192.0.2.130
route add 192.0.2.128/25 192.0.2.2
jexec first pfctl -e
pft_set_rules first \
"set reassemble yes" \
"pass keep state"
# Sanity checks
atf_check -s exit:0 -o ignore ping -c 1 192.0.2.2
atf_check -s exit:0 -o ignore ping -c 1 192.0.2.130
atf_check -s exit:0 -o ignore ping -c 1 192.0.2.131
# Large packet that'll get reassembled and sent out in one on the large
# epair
atf_check -s exit:0 -o ignore ping -c 1 -s 8000 192.0.2.131
}
mtu_diff_cleanup()
{
pft_cleanup
}
frag_common()
{
name=$1
pft_init
epair=$(vnet_mkepair)
vnet_mkjail alcatraz ${epair}a
ifconfig ${epair}b inet 192.0.2.1/24 up
jexec alcatraz ifconfig ${epair}a 192.0.2.2/24 up
jexec alcatraz pfctl -e
pft_set_rules alcatraz \
"set reassemble yes" \
"pass keep state"
# Sanity check
atf_check -s exit:0 -o ignore ping -c 1 192.0.2.2
atf_check -s exit:0 -o ignore $(atf_get_srcdir)/frag-${1}.py \
--to 192.0.2.2 \
--fromaddr 192.0.2.1 \
--sendif ${epair}b \
--recvif ${epair}b
}
atf_test_case "overreplace" "cleanup"
overreplace_head()
{
atf_set descr 'ping fragment that overlaps fragment at index boundary and replace it'
atf_set require.user root
atf_set require.progs scapy
}
overreplace_body()
{
frag_common overreplace
}
overreplace_cleanup()
{
pft_cleanup
}
atf_test_case "overindex" "cleanup"
overindex_head()
{
atf_set descr 'ping fragment that overlaps the first fragment at index boundary'
atf_set require.user root
atf_set require.progs scapy
}
overindex_body()
{
frag_common overindex
}
overindex_cleanup()
{
pft_cleanup
}
atf_test_case "overlimit" "cleanup"
overlimit_head()
{
atf_set descr 'ping fragment at index boundary that cannot be requeued'
atf_set require.user root
atf_set require.progs scapy
}
overlimit_body()
{
frag_common overlimit
}
overlimit_cleanup()
{
pft_cleanup
}
atf_test_case "reassemble" "cleanup"
reassemble_head()
{
atf_set descr 'Test reassembly'
atf_set require.user root
}
reassemble_body()
{
pft_init
epair=$(vnet_mkepair)
vnet_mkjail alcatraz ${epair}a
ifconfig ${epair}b inet 192.0.2.1/24 up
jexec alcatraz ifconfig ${epair}a 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 \
"pass out" \
"block in" \
"pass in inet proto icmp all icmp-type echoreq"
# Single fragment passes
atf_check -s exit:0 -o ignore ping -c 1 192.0.2.2
# But a fragmented ping does not
atf_check -s exit:2 -o ignore ping -c 1 -s 2000 192.0.2.2
pft_set_rules alcatraz \
"set reassemble yes" \
"pass out" \
"block in" \
"pass in inet proto icmp all icmp-type echoreq"
# Both single packet & fragmented pass when we scrub
atf_check -s exit:0 -o ignore ping -c 1 192.0.2.2
atf_check -s exit:0 -o ignore ping -c 1 -s 2000 192.0.2.2
}
reassemble_cleanup()
{
pft_cleanup
}
atf_test_case "no_df" "cleanup"
no_df_head()
{
atf_set descr 'Test removing of DF flag'
atf_set require.user root
}
no_df_body()
{
setup_router_server_ipv4
ifconfig ${epair_tester}a mtu 9000
jexec router ifconfig ${epair_tester}b mtu 9000
jexec router ifconfig ${epair_server}a mtu 1500
jexec server ifconfig ${epair_server}b mtu 1500
# Sanity check.
ping_server_check_reply exit:0 --ping-type=icmp
pft_set_rules router \
"set reassemble no" \
"pass out" \
"block in" \
"pass in inet proto icmp all icmp-type echoreq"
# Ping with normal, fragmentable packets.
ping_server_check_reply exit:1 --ping-type=icmp --send-length=2000
pft_set_rules router \
"set reassemble yes" \
"pass out" \
"block in" \
"pass in inet proto icmp all icmp-type echoreq"
# Ping with normal, fragmentable packets.
ping_server_check_reply exit:0 --ping-type=icmp --send-length=2000
# Ping with non-fragmentable packets.
ping_server_check_reply exit:1 --ping-type=icmp --send-length=2000 --send-flags DF
pft_set_rules router \
"set reassemble yes no-df" \
"pass out" \
"block in" \
"pass in inet proto icmp all icmp-type echoreq"
# Ping with non-fragmentable packets again.
# This time pf will strip the DF flag.
ping_server_check_reply exit:0 --ping-type=icmp --send-length=2000 --send-flags DF
}
no_df_cleanup()
{
pft_cleanup
}
atf_test_case "no_df" "cleanup"
no_df_head()
{
atf_set descr 'Test removing of DF flag'
atf_set require.user root
}
no_df_body()
{
setup_router_server_ipv4
# Tester can send long packets which will get fragmented by the router.
# Replies from server will come in fragments which might get
# reassembled resulting in a long reply packet sent back to tester.
ifconfig ${epair_tester}a mtu 9000
jexec router ifconfig ${epair_tester}b mtu 9000
jexec router ifconfig ${epair_server}a mtu 1500
jexec server ifconfig ${epair_server}b mtu 1500
# Sanity check.
ping_server_check_reply exit:0 --ping-type=icmp
# Enable packet reassembly with clearing of the no-df flag.
pft_set_rules router \
"scrub all fragment reassemble no-df" \
"block" \
"pass inet proto icmp all icmp-type echoreq"
# Ping with non-fragmentable packets.
# pf will strip the DF flag resulting in fragmentation and packets
# getting properly forwarded.
ping_server_check_reply exit:0 --ping-type=icmp --send-length=2000 --send-flags DF
}
no_df_cleanup()
{
pft_cleanup
}
atf_test_case "reassemble_slowpath" "cleanup"
reassemble_slowpath_head()
{
atf_set descr 'Test reassembly on the slow path'
atf_set require.user root
}
reassemble_slowpath_body()
{
if ! sysctl -q kern.features.ipsec >/dev/null ; then
atf_skip "This test requires ipsec"
fi
setup_router_server_ipv4
# Now define an ipsec policy so we end up taking the slow path.
# We don't actually need the traffic to go through ipsec, we just don't
# want to go through ip_tryforward().
echo "flush;
spdflush;
spdadd 203.0.113.1/32 203.0.113.2/32 any -P out ipsec esp/transport//require;
add 203.0.113.1 203.0.113.2 esp 0x1001 -E aes-gcm-16 \"12345678901234567890\";" \
| jexec router setkey -c
# Sanity check.
ping_server_check_reply exit:0 --ping-type=icmp
# Enable packet reassembly with clearing of the no-df flag.
pft_set_rules router \
"scrub in on ${epair_tester}b fragment no reassemble" \
"scrub on ${epair_server}a fragment reassemble" \
"pass"
# Ensure that the packet makes it through the slow path
atf_check -s exit:0 -o ignore \
ping -c 1 -s 2000 198.51.100.2
}
reassemble_slowpath_cleanup()
{
pft_cleanup
}
atf_test_case "dummynet" "cleanup"
dummynet_head()
{
atf_set descr 'dummynet + reassembly test'
atf_set require.user root
}
dummynet_body()
{
pft_init
dummynet_init
epair=$(vnet_mkepair)
vnet_mkjail alcatraz ${epair}a
ifconfig ${epair}b inet 192.0.2.1/24 up
jexec alcatraz ifconfig ${epair}a 192.0.2.2/24 up
# Sanity check
atf_check -s exit:0 -o ignore ping -c 1 192.0.2.2
jexec alcatraz dnctl pipe 1 config bw 600Byte/s
jexec alcatraz dnctl pipe 2 config bw 700Byte/s
jexec alcatraz pfctl -e
pft_set_rules alcatraz \
"set reassemble yes" \
"block" \
"pass inet proto icmp all icmp-type echoreq dnpipe (1, 2)"
atf_check -s exit:0 -o ignore ping -s 2000 -c 1 192.0.2.2
}
dummynet_cleanup()
{
pft_cleanup
}
atf_test_case "dummynet_nat" "cleanup"
dummynet_nat_head()
{
atf_set descr 'Test dummynet on NATed fragmented traffic'
atf_set require.user root
}
dummynet_nat_body()
{
pft_init
dummynet_init
epair_one=$(vnet_mkepair)
ifconfig ${epair_one}a 192.0.2.1/24 up
epair_two=$(vnet_mkepair)
vnet_mkjail alcatraz ${epair_one}b ${epair_two}a
jexec alcatraz ifconfig ${epair_one}b 192.0.2.2/24 up
jexec alcatraz ifconfig ${epair_two}a 198.51.100.1/24 up
jexec alcatraz sysctl net.inet.ip.forwarding=1
vnet_mkjail singsing ${epair_two}b
jexec singsing ifconfig ${epair_two}b 198.51.100.2/24 up
jexec singsing route add default 198.51.100.1
route add 198.51.100.0/24 192.0.2.2
jexec alcatraz dnctl pipe 1 config bw 1600Byte/s
jexec alcatraz dnctl pipe 2 config bw 1700Byte/s
jexec alcatraz pfctl -e
pft_set_rules alcatraz \
"set reassemble yes" \
"nat on ${epair_two}a from 192.0.2.0/24 -> (${epair_two}a)" \
"block in" \
"pass in inet proto icmp all icmp-type echoreq dnpipe (1, 2)"
atf_check -s exit:0 -o ignore ping -c 1 198.51.100.2
atf_check -s exit:0 -o ignore ping -c 1 -s 2000 198.51.100.2
}
dummynet_nat_cleanup()
{
pft_cleanup
}
atf_test_case "dummynet_fragmented" "cleanup"
dummynet_fragmented_head()
{
atf_set descr 'Test dummynet on NATed fragmented traffic'
atf_set require.user root
}
dummynet_fragmented_body()
{
pft_init
dummynet_init
# No test for IPv6. IPv6 fragment reassembly can't be disabled.
setup_router_dummy_ipv4
jexec router dnctl pipe 1 config delay 1
pft_set_rules router \
"set reassemble no" \
"block" \
"pass inet6 proto icmp6 icmp6-type { neighbrsol, neighbradv }" \
"pass in on ${epair_tester}b inet proto udp dnpipe (1, 1)" \
"pass out on ${epair_server}a inet proto udp" \
ping_dummy_check_request exit:0 --ping-type=udp --send-length=10000 --send-frag-length=1280
rules=$(mktemp) || exit 1
jexec router pfctl -qvsr > $rules
# Count that fragmented packets have hit the rule only once and that
# they have not created states. There is no stateful firewall support
# for fragmented packets.
grep -A2 'pass in on epair0b inet proto udp all keep state dnpipe(1, 1)' $rules |
grep -qE 'Packets: 8\s+Bytes: 10168\s+States: 0\s+' ||
atf_fail "Fragmented packets not counted correctly"
}
dummynet_fragmented_cleanup()
{
pft_cleanup
}
atf_init_test_cases()
{
atf_add_test_case "too_many_fragments"
atf_add_test_case "v6"
atf_add_test_case "mtu_diff"
atf_add_test_case "overreplace"
atf_add_test_case "overindex"
atf_add_test_case "overlimit"
atf_add_test_case "reassemble"
atf_add_test_case "no_df"
atf_add_test_case "reassemble_slowpath"
atf_add_test_case "dummynet"
atf_add_test_case "dummynet_nat"
atf_add_test_case "dummynet_fragmented"
}