path: root/passt.1

.\" SPDX-License-Identifier: GPL-2.0-or-later
.\" Copyright (c) 2020-2022 Red Hat GmbH
.\" Author: Stefano Brivio <sbrivio@redhat.com>
.TH passt 1

.SH NAME
.B passt
\- Unprivileged user-mode network connectivity for virtual machines
.br
.B pasta
\- Unprivileged user-mode network connectivity for network namespaces

.SH SYNOPSIS
.B passt
[\fIOPTION\fR]...
.br
.B pasta
[\fIOPTION\fR]... [\fICOMMAND\fR [\fIARG\fR]...]
.br
.B pasta
[\fIOPTION\fR]... \fIPID\fR
.br
.B pasta
[\fIOPTION\fR]... \fB--netns\fR [\fIPATH\fR|\fINAME\fR]

.SH DESCRIPTION

.SS passt

.B passt
(\fIP\fRlug \fIA\fR \fIS\fRimple \fIS\fRocket \fIT\fRransport) provides full,
quasi-native network connectivity to virtual machines in user-mode without
requiring any capabilities or privileges.

The data plane implements a translation layer between a Layer-2 virtual network
interface and native Layer-4 (TCP, UDP, ping) sockets on the host, giving the
illusion that application processes residing on the guest are running on the
local host, from a networking perspective.

Built-in ARP, DHCP, NDP, and DHCPv6 implementations are designed to provide the
guest with a network configuration that tightly resembles the host native
configuration. With the default options, guest and host share IP addresses,
routes, and port bindings.

Port forwarding and translation allow networking services running in the guest
to be reachable from both local and remote hosts.

Unlike \fBslirp4netns\fR(1), \fBpasst\fR doesn't implement a full TCP stack: the
TCP translation layer has no stateful data buffering and operates by reflecting
one peer's observed parameters (congestion window size, acknowledged data, etc.)
to the corresponding peer.

Currently, the only supported hypervisor is \fBqemu\fR(1), connecting to
\fBpasst\fR by means of a UNIX domain socket. This is supported starting from
qemu 7.2. For older qemu versions, see the \fBqrap\fR(1) wrapper.

.SS pasta

.B pasta
(\fIP\fRack \fIA\fR \fIS\fRubtle \fIT\fRap \fIA\fRbstraction) provides
equivalent functionality to network namespaces, as the one offered by
\fBpasst\fR for virtual machines.

If PID or --netns are given, \fBpasta\fR associates to an existing
user and network namespace. Otherwise, \fBpasta\fR creates a new user
and network namespace, and spawns the given command or a default shell
within this context. A \fItap\fR device within the network namespace
is created to provide network connectivity.

For local TCP and UDP traffic only, \fBpasta\fR also implements a bypass path
directly mapping Layer-4 sockets between \fIinit\fR and target namespaces,
for performance reasons.

.SH OPTIONS

.TP
.BR \-d ", " \-\-debug
Be verbose, don't log to the system logger.

.TP
.BR \-\-trace
Be extra verbose, show single packets. Implies \fB--debug\fR.

.TP
.BR \-q ", " \-\-quiet
Don't print informational messages.

.TP
.BR \-f ", " \-\-foreground
Don't run in background. This implies that the process is not moved to a
detached PID namespace after starting, because the PID itself cannot change.
Default is to fork into background.

.TP
.BR \-e ", " \-\-stderr
Log to standard error too.
Default is to log to the system logger only, if started from an interactive
terminal, and to both system logger and standard error otherwise.

.TP
.BR \-l ", " \-\-log-file " " \fIPATH\fR
Log to file \fIPATH\fR, not to standard error, and not to the system logger.

.TP
.BR \-\-log-size " " \fISIZE\fR
Limit log file size to \fISIZE\fR bytes. When the log file is full, make room
for new entries by removing old ones at the beginning. This limit is mandatory.
Default is 1048576 (1 MiB).

.TP
.BR \-\-runas " " \fIUID\fR|\fIUID:GID\fR|\fILOGIN\fR|\fILOGIN:GROUP\fR
Attempt to change to given UID and corresponding group if UID is given,
or to given UID and given GID if both are given. Alternatively, login name, or
login name and group name can be passed. This requires privileges (either
initial effective UID 0 or CAP_SETUID capability) to work.
Default is to change to user \fInobody\fR if started as root.

.TP
.BR \-h ", " \-\-help
Display a help message and exit.

.TP
.BR \-\-version
Show version and exit.

.TP
.BR \-p ", " \-\-pcap " " \fIfile
Capture tap-facing (that is, guest-side or namespace-side) network packets to
\fIfile\fR in \fBpcap\fR format.

.TP
.BR \-P ", " \-\-pid " " \fIfile
Write own PID to \fIfile\fR once initialisation is done, before forking to
background (if configured to do so).

.TP
.BR \-m ", " \-\-mtu " " \fImtu
Assign \fImtu\fR via DHCP (option 26) and NDP (option type 5). A zero value
disables assignment.
By default, the advertised MTU is 65520 bytes, that is, the maximum 802.3 MTU
minus the length of a 802.3 header, rounded to 32 bits (IPv4 words).

.TP
.BR \-a ", " \-\-address " " \fIaddr
Assign IPv4 \fIaddr\fR via DHCP (\fByiaddr\fR), or \fIaddr\fR via DHCPv6 (option
5) and an \fIaddr\fR-based prefix via NDP Router Advertisement (option type 3)
for an IPv6 \fIaddr\fR.
This option can be specified zero (for defaults) to two times (once for IPv4,
once for IPv6).
By default, assigned IPv4 and IPv6 addresses are taken from the host interfaces
with the first default route for the corresponding IP version.

.TP
.BR \-n ", " \-\-netmask " " \fImask
Assign IPv4 netmask \fImask\fR, expressed as dot-decimal or number of bits, via
DHCP (option 1).
By default, the netmask associated to the host address matching the assigned one
is used. If there's no matching address on the host, the netmask is determined
according to the CIDR block of the assigned address (RFC 4632).

.TP
.BR \-M ", " \-\-mac-addr " " \fIaddr
Use source MAC address \fIaddr\fR when communicating to the guest or to the
target namespace.
Default is to use the MAC address of the interface with the first IPv4 default
route on the host.

.TP
.BR \-g ", " \-\-gateway " " \fIaddr
Assign IPv4 \fIaddr\fR as default gateway via DHCP (option 3), or IPv6
\fIaddr\fR as source for NDP Router Advertisement and DHCPv6 messages.
This option can be specified zero (for defaults) to two times (once for IPv4,
once for IPv6).
By default, IPv4 and IPv6 gateways are taken from the host interface with the
first default route for the corresponding IP version. If the default route is a
multipath one, the gateway is the first nexthop router returned by the kernel
which has the highest weight in the set of paths.

Note: these addresses are also used as source address for packets directed to
the guest or to the target namespace having a loopback or local source address,
to allow mapping of local traffic to guest and target namespace. See the
\fBNOTES\fR below for more details about this mechanism.

.TP
.BR \-i ", " \-\-interface " " \fIname
Use host interface \fIname\fR to derive addresses and routes.
Default is to use the interfaces specified by \fB--outbound-if4\fR and
\fB--outbound-if6\fR, for IPv4 and IPv6 addresses and routes, respectively. If
no interfaces are given, the interface with the first default routes for each IP
version is selected.

.TP
.BR \-o ", " \-\-outbound " " \fIaddr
Use an IPv4 \fIaddr\fR as source address for IPv4 outbound TCP connections, UDP
flows, ICMP requests, or an IPv6 \fIaddr\fR for IPv6 ones, by binding outbound
sockets to it.
This option can be specified zero (for defaults) to two times (once for IPv4,
once for IPv6).
By default, the source address is selected by the routing tables.

.TP
.BR \-\-outbound-if4 " " \fIname
Bind IPv4 outbound sockets to host interface \fIname\fR, and, unless another
interface is specified via \fB-i\fR, \fB--interface\fR, use this interface to
derive IPv4 addresses and routes.
By default, the interface given by the default route is selected.

.TP
.BR \-\-outbound-if6 " " \fIname
Bind IPv6 outbound sockets to host interface \fIname\fR, and, unless another
interface is specified via \fB-i\fR, \fB--interface\fR, use this interface to
derive IPv6 addresses and routes.
By default, the interface given by the default route is selected.

.TP
.BR \-D ", " \-\-dns " " \fIaddr
Use \fIaddr\fR (IPv4 or IPv6) for DHCP, DHCPv6, NDP or DNS forwarding, as
configured (see options \fB--no-dhcp-dns\fR, \fB--dhcp-dns\fR,
\fB--dns-forward\fR) instead of reading addresses from \fI/etc/resolv.conf\fR.
This option can be specified multiple times.  Specifying \fB-D none\fR disables
usage of DNS addresses altogether.

.TP
.BR \-\-dns-forward " " \fIaddr
Map \fIaddr\fR (IPv4 or IPv6) as seen from guest or namespace to the first
configured DNS resolver (with corresponding IP version). Mapping is limited to
UDP traffic directed to port 53, and DNS answers are translated back with a
reverse mapping.
This option can be specified zero to two times (once for IPv4, once for IPv6).

.TP
.BR \-S ", " \-\-search " " \fIlist
Use space-separated \fIlist\fR for DHCP, DHCPv6, and NDP purposes, instead of
reading entries from \fI/etc/resolv.conf\fR. See options \fB--no-dhcp-search\fR
and \fB--dhcp-search\fR. \fB--search none\fR disables the DNS domain search
list altogether (if you need to search a domain called "none" you can use
\fB--search none.\fR).

.TP
.BR \-\-no-dhcp-dns
In \fIpasst\fR mode, do not assign IPv4 addresses via DHCP (option 23) or IPv6
addresses via NDP Router Advertisement (option type 25) and DHCPv6 (option 23)
as DNS resolvers.
By default, all the configured addresses are passed.

.TP
.BR \-\-dhcp-dns
In \fIpasta\fR mode, assign IPv4 addresses via DHCP (option 23) or IPv6
addresses via NDP Router Advertisement (option type 25) and DHCPv6 (option 23)
as DNS resolvers.
By default, configured addresses, if any, are not passed.

.TP
.BR \-\-no-dhcp-search
In \fIpasst\fR mode, do not send the DNS domain search list addresses via DHCP
(option 119), via NDP Router Advertisement (option type 31) and DHCPv6 (option
24).
By default, the DNS domain search list resulting from configuration is passed.

.TP
.BR \-\-dhcp-search
In \fIpasta\fR mode, send the DNS domain search list addresses via DHCP (option
119), via NDP Router Advertisement (option type 31) and DHCPv6 (option 24).
By default, the DNS domain search list resulting from configuration is not
passed.

.TP
.BR \-\-no-tcp
Disable the TCP protocol handler. No TCP connections will be accepted host-side,
and TCP packets coming from guest or target namespace will be silently dropped.

.TP
.BR \-\-no-udp
Disable the UDP protocol handler. No UDP traffic coming from the host side will
be forwarded, and UDP packets coming from guest or target namespace will be
silently dropped.

.TP
.BR \-\-no-icmp
Disable the ICMP/ICMPv6 echo handler. ICMP and ICMPv6 echo requests coming from
guest or target namespace will be silently dropped.

.TP
.BR \-\-no-dhcp
Disable the DHCP server. DHCP client requests coming from guest or target
namespace will be silently dropped. Implied if there is no gateway on the
selected IPv4 default route.

.TP
.BR \-\-no-ndp
Disable NDP responses. NDP messages coming from guest or target namespace will
be ignored.

.TP
.BR \-\-no-dhcpv6
Disable the DHCPv6 server. DHCPv6 client requests coming from guest or target
namespace will be silently dropped.

.TP
.BR \-\-no-ra
Disable Router Advertisements. Router Solicitations coming from guest or target
namespace will be ignored.

.TP
.BR \-\-no-map-gw
Don't remap TCP connections and untracked UDP traffic, with the gateway address
as destination, to the host. Implied if there is no gateway on the selected
default route for any of the enabled address families.

.TP
.BR \-4 ", " \-\-ipv4-only
Enable IPv4-only operation. IPv6 traffic will be ignored.
By default, IPv6 operation is enabled as long as at least an IPv6 default route
and an interface address are configured on a given host interface.

.TP
.BR \-6 ", " \-\-ipv6-only
Enable IPv6-only operation. IPv4 traffic will be ignored.
By default, IPv4 operation is enabled as long as at least an IPv4 default route
and an interface address are configured on a given host interface.

.SS \fBpasst\fR-only options

.TP
.BR \-s ", " \-\-socket " " \fIpath
Path for UNIX domain socket used by \fBqemu\fR(1) or \fBqrap\fR(1) to connect to
\fBpasst\fR.
Default is to probe a free socket, not accepting connections, starting from
\fI/tmp/passt_1.socket\fR to \fI/tmp/passt_64.socket\fR.

.TP
.BR \-F ", " \-\-fd " " \fIFD
Pass a pre-opened, connected socket to \fBpasst\fR. Usually the socket is opened
in the parent process and \fBpasst\fR inherits it when run as a child. This
allows the parent process to open sockets using another address family or
requiring special privileges.

This option implies the behaviour described for \-\-one-off, once this socket
is closed.

.TP
.BR \-1 ", " \-\-one-off
Quit after handling a single client connection, that is, once the client closes
the socket, or once we get a socket error.

.TP
.BR \-t ", " \-\-tcp-ports " " \fIspec
Configure TCP port forwarding to guest. \fIspec\fR can be one of:
.RS

.TP
.BR none
Don't forward any ports

.TP
.BR all
Forward all unbound, non-ephemeral ports, as permitted by current capabilities.
For low (< 1024) ports, see \fBNOTES\fR. No failures are reported for
unavailable ports, unless no ports could be forwarded at all.

.TP
.BR ports
A comma-separated list of ports, optionally ranged with \fI-\fR, and,
optionally, with target ports after \fI:\fR, if they differ. Specific addresses
can be bound as well, separated by \fI/\fR, and also, since Linux 5.7, limited
to specific interfaces, prefixed by \fI%\fR. Within given ranges, selected ports
and ranges can be excluded by an additional specification prefixed by \fI~\fR.

Specifying excluded ranges only implies that all other ports are forwarded. In
this case, no failures are reported for unavailable ports, unless no ports could
be forwarded at all.

Examples:
.RS
.TP
-t 22
Forward local port 22 to port 22 on the guest
.TP
-t 22:23
Forward local port 22 to port 23 on the guest
.TP
-t 22,25
Forward local ports 22 and 25 to ports 22 and 25 on the guest
.TP
-t 22-80
Forward local ports between 22 and 80 to corresponding ports on the guest
.TP
-t 22-80:32-90
Forward local ports between 22 and 80 to ports between 32 and 90 on the guest
.TP
-t 192.0.2.1/22
Forward local port 22, bound to 192.0.2.1, to port 22 on the guest
.TP
-t 192.0.2.1%eth0/22
Forward local port 22, bound to 192.0.2.1 and interface eth0, to port 22
.TP
-t %eth0/22
Forward local port 22, bound to any address on interface eth0, to port 22
.TP
-t 2000-5000,~3000-3010
Forward local ports between 2000 and 5000, except for those between 3000 and
3010
.TP
-t 192.0.2.1/20-30,~25
For the local address 192.0.2.1, forward ports between 20 and 24 and between 26
and 30
.TP
-t ~20000-20010
Forward all ports to the guest, except for the range from 20000 to 20010
.RE

Default is \fBnone\fR.
.RE

.TP
.BR \-u ", " \-\-udp-ports " " \fIspec
Configure UDP port forwarding to guest. \fIspec\fR is as described for TCP
above.

Note: unless overridden, UDP ports with numbers corresponding to forwarded TCP
port numbers are forwarded too, without, however, any port translation. IPv6
bound ports are also forwarded for IPv4.

Default is \fBnone\fR.

.SS \fBpasta\fR-only options

.TP
.BR \-I ", " \-\-ns-ifname " " \fIname
Name of tap interface to be created in target namespace.
By default, the same interface name as the external, routable interface is used.

.TP
.BR \-t ", " \-\-tcp-ports " " \fIspec
Configure TCP port forwarding to namespace. \fIspec\fR can be one of:
.RS

.TP
.BR none
Don't forward any ports

.TP
.BR auto
Dynamically forward ports bound in the namespace. The list of ports is
periodically derived (every second) from listening sockets reported by
\fI/proc/net/tcp\fR and \fI/proc/net/tcp6\fR, see \fBproc\fR(5).

.TP
.BR ports
A comma-separated list of ports, optionally ranged with \fI-\fR, and,
optionally, with target ports after \fI:\fR, if they differ. Specific addresses
can be bound as well, separated by \fI/\fR, and also, since Linux 5.7, limited
to specific interfaces, prefixed by \fI%\fR. Within given ranges, selected ports
and ranges can be excluded by an additional specification prefixed by \fI~\fR.

Specifying excluded ranges only implies that all other ports are forwarded. In
this case, no failures are reported for unavailable ports, unless no ports could
be forwarded at all.

Examples:
.RS
.TP
-t 22
Forward local port 22 to 22 in the target namespace
.TP
-t 22:23
Forward local port 22 to port 23 in the target namespace
.TP
-t 22,25
Forward local ports 22 and 25 to ports 22 and 25 in the target namespace
.TP
-t 22-80
Forward local ports between 22 and 80 to corresponding ports in the target
namespace
.TP
-t 22-80:32-90
Forward local ports between 22 and 80 to ports between 32 and 90 in the target
namespace
.TP
-t 192.0.2.1/22
Forward local port 22, bound to 192.0.2.1, to port 22 in the target namespace
.TP
-t 192.0.2.1%eth0/22
Forward local port 22, bound to 192.0.2.1 and interface eth0, to port 22
.TP
-t %eth0/22
Forward local port 22, bound to any address on interface eth0, to port 22
.TP
-t 2000-5000,~3000-3010
Forward local ports between 2000 and 5000, except for those between 3000 and
3010
.TP
-t 192.0.2.1/20-30,~25
For the local address 192.0.2.1, forward ports between 20 and 24 and between 26
and 30
.TP
-t ~20000-20010
Forward all ports to the namespace, except for those between 20000 and 20010
.RE

IPv6 bound ports are also forwarded for IPv4.

Default is \fBauto\fR.
.RE

.TP
.BR \-u ", " \-\-udp-ports " " \fIspec
Configure UDP port forwarding to namespace. \fIspec\fR is as described for TCP
above, and the list of ports is derived from listening sockets reported by
\fI/proc/net/udp\fR and \fI/proc/net/udp6\fR, see \fBproc\fR(5).

Note: unless overridden, UDP ports with numbers corresponding to forwarded TCP
port numbers are forwarded too, without, however, any port translation. 

IPv6 bound ports are also forwarded for IPv4.

Default is \fBauto\fR.

.TP
.BR \-T ", " \-\-tcp-ns " " \fIspec
Configure TCP port forwarding from target namespace to init namespace.
\fIspec\fR is as described above for TCP.

Default is \fBauto\fR.

.TP
.BR \-U ", " \-\-udp-ns " " \fIspec
Configure UDP port forwarding from target namespace to init namespace.
\fIspec\fR is as described above for UDP.

Default is \fBauto\fR.

.TP
.BR \-\-userns " " \fIspec
Target user namespace to join, as a path. If PID is given, without this option,
the user namespace will be the one of the corresponding process.

.TP
.BR \-\-netns " " \fIspec
Target network namespace to join, as a path or a name.  A name is treated as
with \fBip-netns(8)\fR as equivalent to a path in \fI/run/netns\fR.

This option can't be specified with a PID.

.TP
.BR \-\-netns-only
Join only a target network namespace, not a user namespace, and don't create one
for sandboxing purposes either. This is implied if PATH or NAME are given
without \-\-userns.

.TP
.BR \-\-no-netns-quit
Do not exit once the target namespace reference is removed.

Without this option, \fBpasta\fR will terminate if the target network namespace
is bound to the filesystem, and the given path is deleted, or if the target
network namespace is represented by a procfs entry, and that entry is deleted,
representing the fact that a process with the given PID terminated.

.TP
.BR \-\-config-net
Configure networking in the namespace: set up addresses and routes as configured
or sourced from the host, and bring up the tap interface.

.TP
.BR \-\-no-copy-routes " " (DEPRECATED)
With \-\-config-net, do not copy all the routes associated to the interface we
derive addresses and routes from: set up only the default gateway. Implied by
-g, \-\-gateway.

Default is to copy all the routing entries from the interface in the outer
namespace to the target namespace, translating the output interface attribute to
the outbound interface in the namespace.

Note that this configuration option is \fBdeprecated\fR and will be removed in a
future version. It is not expected to be of any use, and it simply reflects a
legacy behaviour. If you have any use for this, refer to \fBREPORTING BUGS\fR
below.

.TP
.BR \-\-no-copy-addrs " " (DEPRECATED)
With \-\-config-net, do not copy all the addresses associated to the interface
we derive addresses and routes from: set up a single one. Implied by \-a,
\-\-address.

Default is to copy all the addresses, except for link-local ones, from the
interface from the outer namespace to the target namespace.

Note that this configuration option is \fBdeprecated\fR and will be removed in a
future version. It is not expected to be of any use, and it simply reflects a
legacy behaviour. If you have any use for this, refer to \fBREPORTING BUGS\fR
below.

.TP
.BR \-\-ns-mac-addr " " \fIaddr
Configure MAC address \fIaddr\fR on the tap interface in the namespace.

Default is to let the tap driver build a pseudorandom hardware address.

.SH EXAMPLES

.SS \fBpasta
.BR "Create and use a new, connected, user and network namespace"
.RS
.nf
$ iperf3 -s -D
$ ./pasta
Outbound interface: eth0, namespace interface: eth0
ARP:
    address: 28:16:ad:39:a9:ea
DHCP:
    assign: 192.168.1.118
    mask: 255.255.255.0
    router: 192.168.1.1
NDP/DHCPv6:
    assign: 2a02:6d40:3ca5:2001:b81d:fa4a:8cdd:cf17
    router: fe80::62e3:27ff:fe33:2b01
#
# dhclient -4 --no-pid
# dhclient -6 --no-pid
# ip address show
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 65520 qdisc pfifo_fast state UNKNOWN group default qlen 1000
    link/ether 5e:90:02:eb:b0:2a brd ff:ff:ff:ff:ff:ff
    inet 192.168.1.118/24 brd 192.168.1.255 scope global eth0
       valid_lft forever preferred_lft forever
    inet6 2a02:6d40:3ca5:2001:b81d:fa4a:8cdd:cf17/128 scope global 
       valid_lft forever preferred_lft forever
    inet6 2a02:6d40:3ca5:2001:5c90:2ff:feeb:b02a/64 scope global dynamic mngtmpaddr 
       valid_lft 3591sec preferred_lft 3591sec
    inet6 fe80::5c90:2ff:feeb:b02a/64 scope link 
       valid_lft forever preferred_lft forever
# ip route show
default via 192.168.1.1 dev eth0 
192.168.1.0/24 dev eth0 proto kernel scope link src 192.168.1.118 
# ip -6 route show
2a02:6d40:3ca5:2001:b81d:fa4a:8cdd:cf17 dev eth0 proto kernel metric 256 pref medium
2a02:6d40:3ca5:2001::/64 dev eth0 proto kernel metric 256 expires 3584sec pref medium
fe80::/64 dev eth0 proto kernel metric 256 pref medium
default via fe80::62e3:27ff:fe33:2b01 dev eth0 proto ra metric 1024 expires 3584sec pref medium
# iperf3 -c 127.0.0.1 -t1
Connecting to host 127.0.0.1, port 5201
[  5] local 127.0.0.1 port 51938 connected to 127.0.0.1 port 5201
[ ID] Interval           Transfer     Bitrate         Retr  Cwnd
[  5]   0.00-1.00   sec  4.46 GBytes  38.3 Gbits/sec    0   3.93 MBytes       
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bitrate         Retr
[  5]   0.00-1.00   sec  4.46 GBytes  38.3 Gbits/sec    0             sender
[  5]   0.00-1.41   sec  4.45 GBytes  27.1 Gbits/sec                  receiver

iperf Done.
# iperf3 -c ::1 -t1
Connecting to host ::1, port 5201
[  5] local ::1 port 50108 connected to ::1 port 5201
[ ID] Interval           Transfer     Bitrate         Retr  Cwnd
[  5]   0.00-1.00   sec  4.35 GBytes  37.4 Gbits/sec    0   4.99 MBytes       
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bitrate         Retr
[  5]   0.00-1.00   sec  4.35 GBytes  37.4 Gbits/sec    0             sender
[  5]   0.00-1.41   sec  4.35 GBytes  26.4 Gbits/sec                  receiver

iperf Done.
# ping -c1 -4 spaghetti.pizza
PING spaghetti.pizza (172.67.192.217) 56(84) bytes of data.
64 bytes from 172.67.192.217: icmp_seq=1 ttl=255 time=37.3 ms

--- spaghetti.pizza ping statistics ---
1 packets transmitted, 1 received, 0% packet loss, time 0ms
# ping -c1 -6 spaghetti.pizza
PING spaghetti.pizza(2606:4700:3034::6815:147a (2606:4700:3034::6815:147a)) 56 data bytes
64 bytes from 2606:4700:3034::6815:147a: icmp_seq=1 ttl=255 time=35.6 ms

--- spaghetti.pizza ping statistics ---
1 packets transmitted, 1 received, 0% packet loss, time 0ms
rtt min/avg/max/mdev = 35.605/35.605/35.605/0.000 ms
# logout
$

.RE
.fi

.BR "Connect an existing user and network namespace"
.RS
.nf
$ unshare -rUn
# echo $$
2446678

.fi
.BR "	[From another terminal]"
.nf
$ ./pasta 2446678
Outbound interface: eth0, namespace interface: eth0
ARP:
    address: 28:16:ad:39:a9:ea
DHCP:
    assign: 192.168.1.118
    mask: 255.255.255.0
    router: 192.168.1.1
NDP/DHCPv6:
    assign: 2a02:6d40:3ca5:2001:b81d:fa4a:8cdd:cf17
    router: fe80::62e3:27ff:fe33:2b01

.fi
.BR "	[Back to the original terminal]"
.nf
# dhclient -4 --no-pid
# dhclient -6 --no-pid
# ip address show
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 65520 qdisc pfifo_fast state UNKNOWN group default qlen 1000
    link/ether fa:c1:2a:27:92:a9 brd ff:ff:ff:ff:ff:ff
    inet 192.168.1.118/24 brd 192.168.1.255 scope global eth0
       valid_lft forever preferred_lft forever
    inet6 2a02:6d40:3ca5:2001:b81d:fa4a:8cdd:cf17/128 scope global 
       valid_lft forever preferred_lft forever
    inet6 2a02:6d40:3ca5:2001:f8c1:2aff:fe27:92a9/64 scope global dynamic mngtmpaddr 
       valid_lft 3594sec preferred_lft 3594sec
    inet6 fe80::f8c1:2aff:fe27:92a9/64 scope link 
       valid_lft forever preferred_lft forever
.fi
.RE

.SS \fBpasst
.BR "Start and connect a guest with basic port forwarding"
.RS
.nf
$ ./passt -f -t 2222:22
Outbound interface: eth0
ARP:
    address: 28:16:ad:39:a9:ea
DHCP:
    assign: 192.168.1.118
    mask: 255.255.255.0
    router: 192.168.1.1
        search:
            redhat.com
NDP/DHCPv6:
    assign: 2a02:6d40:3ca5:2001:b81d:fa4a:8cdd:cf17
    router: fe80::62e3:27ff:fe33:2b01
        search:
            redhat.com
UNIX domain socket bound at /tmp/passt_1.socket

You can now start qrap:
    ./qrap 5 qemu-system-x86_64 ... -net socket,fd=5 -net nic,model=virtio
or directly qemu, patched with:
    qemu/0001-net-Allow-also-UNIX-domain-sockets-to-be-used-as-net.patch
as follows:
    qemu-system-x86_64 ... -net socket,connect=/tmp/passt_1.socket -net nic,model=virtio

.fi
.BR "	[From another terminal]"
.nf
$ ./qrap 5 qemu-system-x86_64 test.qcow2 -m 1024 -display none -nodefaults -nographic -net socket,fd=5 -net nic,model=virtio
Connected to /tmp/passt_1.socket

.fi
.BR "	[Back to the original terminal]"
.nf
passt: DHCP: ack to request
passt:     from 52:54:00:12:34:56
passt: NDP: received NS, sending NA
passt: NDP: received RS, sending RA
passt: DHCPv6: received SOLICIT, sending ADVERTISE
passt: NDP: received NS, sending NA
passt: DHCPv6: received REQUEST/RENEW/CONFIRM, sending REPLY
passt: NDP: received NS, sending NA

.fi
.BR "	[From yet another terminal]"
.nf
$ ssh -p 2222 root@localhost
root@localhost's password: 
.fi
.BR "	[...]"
.nf
# ip address show
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: ens2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 65520 qdisc pfifo_fast state UP group default qlen 1000
    link/ether 52:54:00:12:34:56 brd ff:ff:ff:ff:ff:ff
    inet 192.168.1.118/24 brd 192.168.1.255 scope global noprefixroute ens2
       valid_lft forever preferred_lft forever
    inet6 2a02:6d40:3ca5:2001:b81d:fa4a:8cdd:cf17/128 scope global noprefixroute 
       valid_lft forever preferred_lft forever
    inet6 2a02:6d40:3ca5:2001:b019:9ae2:a2fe:e6b4/64 scope global dynamic noprefixroute 
       valid_lft 3588sec preferred_lft 3588sec
    inet6 fe80::1f98:d09f:9309:9e77/64 scope link noprefixroute 
       valid_lft forever preferred_lft forever
.fi
.RE

.SH NOTES

.SS Handling of traffic with local destination and source addresses

Both \fBpasst\fR and \fBpasta\fR can bind on ports with a local address,
depending on the configuration. Local destination or source addresses need to be
changed before packets are delivered to the guest or target namespace: most
operating systems would drop packets received from non-loopback interfaces with
local addresses, and it would also be impossible for guest or target namespace
to route answers back.

For convenience, and somewhat arbitrarily, the source address on these packets
is translated to the address of the default IPv4 or IPv6 gateway -- this is
known to be an existing, valid address on the same subnet.

Loopback destination addresses are instead translated to the observed external
address of the guest or target namespace. For IPv6 packets, if usage of a
link-local address by guest or namespace has ever been observed, and the
original destination address is also a link-local address, the observed
link-local address is used. Otherwise, the observed global address is used. For
both IPv4 and IPv6, if no addresses have been seen yet, the configured addresses
will be used instead.

For example, if \fBpasst\fR or \fBpasta\fR receive a connection from 127.0.0.1,
with destination 127.0.0.10, and the default IPv4 gateway is 192.0.2.1, while
the last observed source address from guest or namespace is 192.0.2.2, this will
be translated to a connection from 192.0.2.1 to 192.0.2.2.

Similarly, for traffic coming from guest or namespace, packets with destination
address corresponding to the default gateway will have their destination address
translated to a loopback address, if and only if a packet, in the opposite
direction, with a loopback destination or source address, port-wise matching for
UDP, or connection-wise for TCP, has been recently forwarded to guest or
namespace. This behaviour can be disabled with \-\-no\-map\-gw.

.SS Handling of local traffic in pasta

Depending on the configuration, \fBpasta\fR can bind to local ports in the init
namespace, in the target namespace, or both, and forward connections and packets
to corresponding ports in the other namespace.

To avoid unnecessary overhead, these connections and packets are not forwarded
through the tap device connecting the namespaces: \fBpasta\fR creates a socket
in the destination namespace, with matching Layer-4 protocol, and uses it to
forward local data. For TCP, data is forwarded between the originating socket
and the new socket using the \fBsplice\fR(2) system call, and for UDP, a pair
of \fBrecvmmsg\fR(2) and \fBsendmmsg\fR(2) system calls deals with packet
transfers.

This bypass only applies to local connections and traffic, because it's not
possible to bind sockets to foreign addresses.

.SS Binding to low numbered ports (well-known or system ports, up to 1023)

If the port forwarding configuration requires binding to ports with numbers
lower than 1024, \fBpasst\fR and \fBpasta\fR will try to bind to them, but will
fail, unless, either:

.IP \(bu 2
the \fIsys.net.ipv4.ip_unprivileged_port_start\fR sysctl is set to the number
of the lowest port \fBpasst\fR and \fBpasta\fR need. For example, as root:

.nf
	sysctl -w net.ipv4.ip_unprivileged_port_start=443
.fi

\fBNote\fR: this is the recommended way of enabling \fBpasst\fR and \fBpasta\fR
to bind to ports with numbers below 1024.

.IP \(bu
or the \fICAP_NET_BIND_SERVICE\fR Linux capability is granted, see
\fBservices\fR(5) and \fBcapabilities\fR(7).

This is, in general, \fBnot the recommended way\fR, because \fBpasst\fR and
\fBpasta\fR might be used as vector to effectively use this capability from
another process.

However, if your environment is sufficiently controlled by an LSM (Linux
Security Module) such as \fIAppArmor\fR, \fISELinux\fR, \fISmack\fR or
\fITOMOYO\fR, and no other processes can interact in such a way in virtue of
this, granting this capability to \fBpasst\fR and \fBpasta\fR only can
effectively prevent other processes from utilising it.

Note that this will not work for automatic detection and forwarding of ports
with \fBpasta\fR, because \fBpasta\fR will relinquish this capability at
runtime.

To grant this capability, you can issue, as root:

.nf
	for p in $(which passt passt.avx2); do
		setcap 'cap_net_bind_service=+ep' "${p}"
	done
.fi

.RE

.SS ICMP/ICMPv6 Echo sockets

ICMP and ICMPv6 Echo requests coming from guest or target namespace are handled
using so-called "ping" sockets, introduced in Linux 2.6.30. To preserve the
original identifier (see RFC 792, page 14, for ICMP, and RFC 4443, section 4.1,
for ICMPv6), \fBpasst\fR and \fBpasta\fR try to bind these sockets using the
observed source identifier as "port" -- that corresponds to Echo identifiers
for "ping" sockets.

As \fBbind\fR(2) failures were seen with particularly restrictive SELinux
policies, a fall-back mechanism maps different identifiers to different sockets,
and identifiers in replies will be mapped back to the original identifier of the
request. However, if \fBbind\fR(2) fails and the fall-back mechanism is used,
echo requests will be forwarded with different, albeit unique, identifiers.

For ICMP and ICMPv6 Echo requests to work, the \fIping_group_range\fR parameter
needs to include the PID of \fBpasst\fR or \fBpasta\fR, see \fBicmp\fR(7).

.SS pasta and loopback interface

As \fBpasta\fR connects to an existing namespace, or once it creates a new
namespace, it will also ensure that the loopback interface, \fIlo\fR, is brought
up. This is needed to bind ports using the loopback address in the namespace.

.SS TCP sending window and \fITCP_INFO\fB before Linux 5.3

To synchronise the TCP sending window from host Layer-4 sockets to the TCP
parameters announced in TCP segments sent over the Layer-2 interface,
\fBpasst\fR and \fBpasta\fR routinely query the size of the sending window seen
by the kernel on the corresponding socket using the \fITCP_INFO\fR socket
option, see \fBtcp\fR(7). Before Linux 5.3, i.e. before Linux kernel commit 
8f7baad7f035 ("tcp: Add snd_wnd to TCP_INFO"), the sending window
(\fIsnd_wnd\fR field) is not available.

If the sending window cannot be queried, it will always be announced as the
current sending buffer size to guest or target namespace. This might affect
throughput of TCP connections.

.SH LIMITATIONS

Currently, IGMP/MLD proxying (RFC 4605) and support for SCTP (RFC 4960) are not
implemented.

TCP Selective Acknowledgment (RFC 2018), as well as Protection Against Wrapped
Sequences (PAWS) and Round-Trip Time Measurement (RTTM), both described by RFC
7232, are currently not implemented.

.SH AUTHORS

Stefano Brivio <sbrivio@redhat.com>, David Gibson <david@gibson.dropbear.id.au>.

.SH REPORTING BUGS

Please report issues on the bug tracker at https://passt.top/passt/bugs, or
send a message to the passt-user@passt.top mailing list, see
https://passt.top/passt/lists.

.SH COPYRIGHT

Copyright (c) 2020-2022 Red Hat GmbH.

\fBpasst\fR and \fBpasta\fR are free software: you can redistribute them and/or
modify them under the terms of the GNU Affero General Public License as
published by the Free Software Foundation, either version 3 of the License, or
(at your option) any later version. 

.SH SEE ALSO

\fBnamespaces\fR(7), \fBqemu\fR(1), \fBqrap\fR(1), \fBslirp4netns\fR(1).

High-level documentation is available at https://passt.top/passt/about/.