mirrors/passt
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
passt/test/demo/passt
Stefano Brivio 0515adceaa passt, pasta: Namespace-based sandboxing, defer seccomp policy application 
To reach (at least) a conceptually equivalent security level as
implemented by --enable-sandbox in slirp4netns, we need to create a
new mount namespace and pivot_root() into a new (empty) mountpoint, so
that passt and pasta can't access any filesystem resource after
initialisation.

While at it, also detach IPC, PID (only for passt, to prevent
vulnerabilities based on the knowledge of a target PID), and UTS
namespaces.

With this approach, if we apply the seccomp filters right after the
configuration step, the number of allowed syscalls grows further. To
prevent this, defer the application of seccomp policies after the
initialisation phase, before the main loop, that's where we expect bad
things to happen, potentially. This way, we get back to 22 allowed
syscalls for passt and 34 for pasta, on x86_64.

While at it, move #syscalls notes to specific code paths wherever it
conceptually makes sense.

We have to open all the file handles we'll ever need before
sandboxing:

- the packet capture file can only be opened once, drop instance
  numbers from the default path and use the (pre-sandbox) PID instead

- /proc/net/tcp{,v6} and /proc/net/udp{,v6}, for automatic detection
  of bound ports in pasta mode, are now opened only once, before
  sandboxing, and their handles are stored in the execution context

- the UNIX domain socket for passt is also bound only once, before
  sandboxing: to reject clients after the first one, instead of
  closing the listening socket, keep it open, accept and immediately
  discard new connection if we already have a valid one

Clarify the (unchanged) behaviour for --netns-only in the man page.

To actually make passt and pasta processes run in a separate PID
namespace, we need to unshare(CLONE_NEWPID) before forking to
background (if configured to do so). Introduce a small daemon()
implementation, __daemon(), that additionally saves the PID file
before forking. While running in foreground, the process itself can't
move to a new PID namespace (a process can't change the notion of its
own PID): mention that in the man page.

For some reason, fork() in a detached PID namespace causes SIGTERM
and SIGQUIT to be ignored, even if the handler is still reported as
SIG_DFL: add a signal handler that just exits.

We can now drop most of the pasta_child_handler() implementation,
that took care of terminating all processes running in the same
namespace, if pasta started a shell: the shell itself is now the
init process in that namespace, and all children will terminate
once the init process exits.

Issuing 'echo $$' in a detached PID namespace won't return the
actual namespace PID as seen from the init namespace: adapt
demo and test setup scripts to reflect that.

Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
2022-02-21 13:41:13 +01:00

249 lines
5.7 KiB
Text
Raw Blame History

# SPDX-License-Identifier: AGPL-3.0-or-later
#
# PASST - Plug A Simple Socket Transport
# for qemu/UNIX domain socket mode
#
# PASTA - Pack A Subtle Tap Abstraction
# for network namespace/tap device mode
#
# test/demo/passt - Quick introduction to passt
#
# Copyright (c) 2021 Red Hat GmbH
# Author: Stefano Brivio <sbrivio@redhat.com>
onlyfor passt
say This is a short introduction to
em passt
say .
nl
nl
sleep 3
say Let's fetch the source
sleep 1
tempdir TEMPDIR
host cd __TEMPDIR__
host git clone https://passt.top/passt
sleep 1
say and build it.
sleep 1
host cd passt
host make avx2
sleep 1
nl
nl
say A quick look at the man page...
sleep 1
hostb man ./passt.1
sleep 5
hostb /ports
sleep 2
hostb n
sleep 2
hostb n
sleep 10
nl
say '-t' to forward TCP ports.
sleep 3
host q
nl
nl
say Let's create a small initramfs image for the guest.
guest cd __TEMPDIR__
guest git clone https://mbuto.lameexcu.se/mbuto
guest cd mbuto
guest ./mbuto -f passt.img -p passt -c lz4
sleep 2
nl
nl
say We want to isolate passt and guest in a
nl
say network namespace. For convenience, we'll
nl
say create it with 'pasta', see also the
nl
say 'pasta' demo above.
sleep 3
passt cd __TEMPDIR__/passt
passtb ./pasta
sleep 3
passt /sbin/dhclient
sleep 2
passt /sbin/dhclient -6
sleep 2
nl
nl
say Now let's run 'passt' in the new namespace, and
nl
say enter this namespace from the guest terminal too.
sleep 3
guest pstree -p | grep pasta
gout TARGET_PID pstree -p | grep pasta | sed -n 's/.*(\([0-9].*\))$/\1/p'
sleep 1
passtb ./passt -f -t 5201,5203
sleep 2
guest nsenter -t __TARGET_PID__ -U -n --preserve-credentials
sleep 5
nl
nl
say We're ready to start qemu with the qrap wrapper,
nl
say that we currently need to connect the netdev
nl
say back-end to passt's UNIX domain socket.
sleep 2
hout VMLINUZ echo "/boot/vmlinuz-$(uname -r)"
guest ../passt/qrap 5 kvm -m 4096 -cpu host -smp 4 -kernel __VMLINUZ__ -initrd passt.img -nographic -serial stdio -nodefaults -append "console=ttyS0 virtio-net.napi_tx=1" -device virtio-net-pci,netdev=hostnet0,x-txburst=16384 -netdev socket,fd=5,id=hostnet0
sleep 10
nl
nl
guest ip li sh
sleep 3
say Guest is up. Let's configure IPv4 first...
sleep 2
guest dhclient
sleep 2
guest ip ad sh
sleep 5
nl
say SLAAC is already done, but we can also
nl
say get another address via DHCPv6.
sleep 3
guest dhclient -6
sleep 3
nl
nl
say Let's try to communicate between host and guest.
sleep 2
guestb nc -6 -l -p 5201
sleep 2
host echo "Hello from the host" | nc -N ::1 5201
sleep 5
nl
nl
say Now the other way around... using
nl
say the address of the default gateway.
sleep 2
gout GW ip -j -4 ro sh|jq -rM '.[] | select(.dst == "default").gateway'
sleep 5
hostb nc -l -p 31337
sleep 2
guest echo "Hello from the guest" | nc -N __GW__ 31337
sleep 3
nl
nl
say Let's have a (quick!) look at performance
nl
say more in the "Performance" section below.
sleep 3
host nsenter -t __TARGET_PID__ -U -n --preserve-credentials
guest /sbin/sysctl -w net.core.rmem_max=536870912
guest /sbin/sysctl -w net.core.wmem_max=536870912
guest /sbin/sysctl -w net.core.rmem_default=33554432
guest /sbin/sysctl -w net.core.wmem_default=33554432
guest /sbin/sysctl -w net.ipv4.tcp_rmem="4096 131072 268435456"
guest /sbin/sysctl -w net.ipv4.tcp_wmem="4096 131072 268435456"
guest /sbin/sysctl -w net.ipv4.tcp_timestamps=0
host sysctl -w net.ipv4.tcp_rmem="4096 524288 134217728"
host sysctl -w net.ipv4.tcp_wmem="4096 524288 134217728"
host sysctl -w net.ipv4.tcp_timestamps=0
gout GW6 ip -j -6 ro sh|jq -rM '.[] | select(.dst == "default").gateway'
gout IFNAME ip -j li sh | jq -rM '.[] | select(.link_type == "ether").ifname'
nl
nl
info Throughput in Gbps, latency in µs
th flow host>guest guest>host
set OPTS -P4 -w 64M -l 1M -i1 --pacing-timer 100000
tr TCP/IPv6 throughput
hostb sleep 10; iperf3 -c ::1 __OPTS__
gout BW iperf3 -s1J | jq -rM ".end.sum_received.bits_per_second"
bw __BW__ 2.0 3.0
sleep 5
guestb sleep 10; iperf3 -c __GW6__%__IFNAME__ -p 5202 __OPTS__ -O3
hout BW iperf3 -s1J -p 5202 | jq -rM ".end.sum_received.bits_per_second"
bw __BW__ 2.0 3.0
tl TCP/IPv6 RR latency
guestb tcp_rr -C 5201 -P 5203 -6 --nolog
sleep 2
hout LAT tcp_rr -C 5201 -P 5203 --nolog -c -H ::1 | sed -n 's/^throughput=\(.*\)/\1/p'
lat __LAT__ 1000 800
sleep 2
hostb tcp_rr -6 --nolog
sleep 2
gout LAT tcp_rr --nolog -c -H __GW6__%__IFNAME__ | sed -n 's/^throughput=\(.*\)/\1/p'
lat __LAT__ 1000 800
sleep 2
tl TCP/IPv6 CRR latency
guestb tcp_crr -C 5201 -P 5203 -6 --nolog
sleep 2
hout LAT tcp_crr -C 5201 -P 5203 --nolog -c -H ::1 | sed -n 's/^throughput=\(.*\)/\1/p'
lat __LAT__ 1000 800
sleep 2
hostb tcp_crr -6 --nolog
sleep 2
gout LAT tcp_crr --nolog -c -H __GW6__%__IFNAME__ | sed -n 's/^throughput=\(.*\)/\1/p'
lat __LAT__ 1000 800
sleep 2
tr TCP/IPv4 throughput
hostb sleep 10; iperf3 -c 127.0.0.1 __OPTS__
gout BW iperf3 -s1J | jq -rM ".end.sum_received.bits_per_second"
bw __BW__ 2.0 3.0
sleep 5
guestb sleep 10; iperf3 -c __GW__ -p 5202 __OPTS__ -O3
hout BW iperf3 -s1J -p 5202 | jq -rM ".end.sum_received.bits_per_second"
bw __BW__ 2.0 3.0
tl TCP/IPv4 RR latency
guestb tcp_rr -C 5201 -P 5203 -4 --nolog
sleep 2
hout LAT tcp_rr -C 5201 -P 5203 --nolog -c -H 127.0.0.1 | sed -n 's/^throughput=\(.*\)/\1/p'
lat __LAT__ 1000 800
sleep 2
hostb tcp_rr -4 --nolog
sleep 2
gout LAT tcp_rr --nolog -c -H __GW__ | sed -n 's/^throughput=\(.*\)/\1/p'
lat __LAT__ 1000 800
sleep 2
tl TCP/IPv4 CRR latency
guestb tcp_crr -C 5201 -P 5203 -4 --nolog
sleep 2
hout LAT tcp_crr -C 5201 -P 5203 --nolog -c -H 127.0.0.1 | sed -n 's/^throughput=\(.*\)/\1/p'
lat __LAT__ 1000 800
sleep 2
hostb tcp_crr -4 --nolog
sleep 2
gout LAT tcp_crr --nolog -c -H __GW__ | sed -n 's/^throughput=\(.*\)/\1/p'
lat __LAT__ 1000 800
sleep 2
nl
nl
say Thanks for watching!
sleep 5
Reference in a new issue View git blame Copy permalink