oracular (1) pasta.1.gz

Provided by: passt_0.0~git20240814.61c0b0d-1_amd64 bug

NAME

       passt - Unprivileged user-mode network connectivity for virtual machines
       pasta - Unprivileged user-mode network connectivity for network namespaces

SYNOPSIS

       passt [OPTION]...
       pasta [OPTION]... [COMMAND [ARG]...]
       pasta [OPTION]... PID
       pasta [OPTION]... --netns [PATH|NAME]

DESCRIPTION

   passt
       passt  (Plug  A  Simple  Socket  Transport)  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  slirp4netns(1),  passt  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  qemu(1),  connecting to passt by means of a UNIX domain
       socket. This is supported starting from qemu 7.2. For older qemu versions, see the qrap(1) wrapper.

   pasta
       pasta (Pack A Subtle Tap Abstraction) provides equivalent functionality to network namespaces, as the one
       offered by passt for virtual machines.

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

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

OPTIONS

       Unless otherwise noted below, if conflicting or multiple options are given, the last one takes effect.

       -d, --debug
              Be verbose, don't log to the system logger.

       --trace
              Be extra verbose, show single packets. Implies --debug.

       -q, --quiet
              Don't print informational messages.

       -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.

       -e, --stderr
              This option has no effect, and is maintained for compatibility purposes only.

              Note that this configuration option is deprecated and will be removed in a future version.

       -l, --log-file PATH
              Log to file PATH, and not to the system logger.

              Specifying  this  option multiple times does not lead to multiple log files: the last given option
              takes effect.

       --log-size SIZE
              Limit log file size to SIZE 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).

       --runas UID|UID:GID|LOGIN|LOGIN:GROUP
              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 nobody if started as root.

       -h, --help
              Display a help message and exit.

       --version
              Show version and exit.

       -p, --pcap file
              Capture tap-facing (that is, guest-side or namespace-side) network packets to file in pcap format.

              Specifying this option multiple times does not lead to multiple  capture  files:  the  last  given
              option takes effect.

       -P, --pid file
              Write  own PID to file once initialisation is done, before forking to background (if configured to
              do so).

       -m, --mtu mtu
              Assign mtu 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).

       -a, --address addr
              Assign IPv4 addr via DHCP (yiaddr), or addr via DHCPv6 (option 5) and an addr-based prefix via NDP
              Router  Advertisement  (option  type  3) for an IPv6 addr.  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, if any, for the
              corresponding IP version. If no default routes are available and there is any interface  with  any
              route for a given IP version, the first of these interfaces will be chosen instead.

       -n, --netmask mask
              Assign  IPv4  netmask  mask,  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).

       -M, --mac-addr addr
              Use source MAC address addr 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.

       -g, --gateway addr
              Assign  IPv4  addr  as  default gateway via DHCP (option 3), or IPv6 addr 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, if any, 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. If no default routes are available and there is  just  one
              interface with any route, that interface will be chosen instead.

              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 NOTES below for more details about this mechanism.

       -i, --interface name
              Use  host  interface  name  to  derive  addresses  and  routes.   Default is to use the interfaces
              specified  by  --outbound-if4  and  --outbound-if6,  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. If no default routes are available and there is just one interface with any route,  that
              interface will be chosen instead.

       -o, --outbound addr
              Use an IPv4 addr as source address for IPv4 outbound TCP connections, UDP flows, ICMP requests, or
              an IPv6 addr 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.

       --outbound-if4 name
              Bind IPv4 outbound sockets to host interface name, and, unless another interface is specified  via
              -i, --interface, use this interface to derive IPv4 addresses and routes.

              By  default,  the  interface  given  by  the  default  route is selected. If no default routes are
              available and there is just one interface with any route, that interface will be chosen instead.

       --outbound-if6 name
              Bind IPv6 outbound sockets to host interface name, and, unless another interface is specified  via
              -i, --interface, use this interface to derive IPv6 addresses and routes.

              By  default,  the  interface  given  by  the  default  route is selected. If no default routes are
              available and there is just one interface with any route, that interface will be chosen instead.

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

       --dns-forward addr
              Map addr (IPv4 or IPv6) as seen from guest or namespace to the first configured DNS resolver (with
              corresponding IP version). Maps only UDP and TCP traffic to port 53  or  port  853.   Replies  are
              translated  back with a reverse mapping.  This option can be specified zero to two times (once for
              IPv4, once for IPv6).

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

       --no-dhcp-dns
              In  passt 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.

       --dhcp-dns
              In  pasta  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.

       --no-dhcp-search
              In  passt  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.

       --dhcp-search
              In  pasta  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.

       --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.

       --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.

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

       --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.

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

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

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

       --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, or if there is no  default
              route, for any of the enabled address families.

       -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 route and  an  interface  address  are  configured  on  a  given  host
              interface.

       -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 route and  an  interface  address  are  configured  on  a  given  host
              interface.

   passt-only options
       -s, --socket path
              Path for UNIX domain socket used by qemu(1) or qrap(1) to connect to passt.  Default is to probe a
              free socket, not accepting connections, starting from /tmp/passt_1.socket to /tmp/passt_64.socket.

       -F, --fd FD
              Pass a pre-opened, connected socket to passt. Usually the socket is opened in the  parent  process
              and  passt  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.

       -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.

       -t, --tcp-ports spec
              Configure TCP port forwarding to guest. spec can be one of:

              none   Don't forward any ports

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

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

                     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:

                     -t 22  Forward local port 22 to port 22 on the guest

                     -t 22:23
                            Forward local port 22 to port 23 on the guest

                     -t 22,25
                            Forward local ports 22 and 25 to ports 22 and 25 on the guest

                     -t 22-80
                            Forward local ports between 22 and 80 to corresponding ports on the guest

                     -t 22-80:32-90
                            Forward local ports between 22 and 80 to ports between 32 and 90 on the guest

                     -t 192.0.2.1/22
                            Forward local port 22, bound to 192.0.2.1, to port 22 on the guest

                     -t 192.0.2.1%eth0/22
                            Forward local port 22, bound to 192.0.2.1 and interface eth0, to port 22

                     -t %eth0/22
                            Forward local port 22, bound to any address on interface eth0, to port 22

                     -t 2000-5000,~3000-3010
                            Forward local ports between 2000 and 5000, except for those between 3000 and 3010

                     -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

                     -t ~20000-20010
                            Forward all ports to the guest, except for the range from 20000 to 20010

              Default is none.

       -u, --udp-ports spec
              Configure UDP port forwarding to guest. spec 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 none.

   pasta-only options
       -I, --ns-ifname name
              Name of tap interface to be created in target namespace.  By default, the same interface  name  as
              the external, routable interface is used.

       -t, --tcp-ports spec
              Configure TCP port forwarding to namespace. spec can be one of:

              none   Don't forward any ports

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

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

                     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:

                     -t 22  Forward local port 22 to 22 in the target namespace

                     -t 22:23
                            Forward local port 22 to port 23 in the target namespace

                     -t 22,25
                            Forward local ports 22 and 25 to ports 22 and 25 in the target namespace

                     -t 22-80
                            Forward local ports between 22 and 80 to corresponding ports in the target namespace

                     -t 22-80:32-90
                            Forward  local  ports  between  22  and  80 to ports between 32 and 90 in the target
                            namespace

                     -t 192.0.2.1/22
                            Forward local port 22, bound to 192.0.2.1, to port 22 in the target namespace

                     -t 192.0.2.1%eth0/22
                            Forward local port 22, bound to 192.0.2.1 and interface eth0, to port 22

                     -t %eth0/22
                            Forward local port 22, bound to any address on interface eth0, to port 22

                     -t 2000-5000,~3000-3010
                            Forward local ports between 2000 and 5000, except for those between 3000 and 3010

                     -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

                     -t ~20000-20010
                            Forward all ports to the namespace, except for those between 20000 and 20010

              IPv6 bound ports are also forwarded for IPv4.

              Default is auto.

       -u, --udp-ports spec
              Configure  UDP  port  forwarding to namespace. spec is as described for TCP above, and the list of
              ports is derived from listening sockets reported by /proc/net/udp and /proc/net/udp6, see proc(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 auto.

       -T, --tcp-ns spec
              Configure TCP port forwarding from target namespace to init namespace.  spec is as described above
              for TCP.

              Default is auto.

       -U, --udp-ns spec
              Configure UDP port forwarding from target namespace to init namespace.  spec is as described above
              for UDP.

              Default is auto.

       --userns spec
              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.

       --netns spec
              Target network namespace to join, as a path or a name.  A name is treated as with  ip-netns(8)  as
              equivalent to a path in /run/netns.

              This option can't be specified with a PID.

       --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.

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

              Without this option, pasta 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.

       --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.

       --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, for the corresponding IP
              version only.

              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 deprecated 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 REPORTING BUGS below.

       --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, for the corresponding  IP  version
              only.

              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 deprecated 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 REPORTING BUGS below.

       --ns-mac-addr addr
              Configure MAC address addr on the tap interface in the namespace.

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

EXAMPLES

   pasta
       Create and use a new, connected, user and network namespace
              $ 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
              $

       Connect an existing user and network namespace
              $ unshare -rUn
              # echo $$
              2446678

                   [From another terminal]
              $ ./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

                   [Back to the original terminal]
              # 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

   passt
       Start and connect a guest with basic port forwarding
              $ ./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

                   [From another terminal]
              $ ./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

                   [Back to the original terminal]
              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

                   [From yet another terminal]
              $ ssh -p 2222 root@localhost
              root@localhost's password:
                   [...]
              # 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

NOTES

   Handling of traffic with local destination and source addresses
       Both passt and pasta 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 (if any) -- 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 passt or pasta 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.

   Handling of local traffic in pasta
       Depending  on  the  configuration,  pasta  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:  pasta  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 splice(2) system call, and for UDP, a pair of recvmmsg(2) and sendmmsg(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.

   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, passt and
       pasta will try to bind to them, but will fail, unless, either:

       • the sys.net.ipv4.ip_unprivileged_port_start sysctl is set to the number of the lowest  port  passt  and
         pasta need. For example, as root:

              sysctl -w net.ipv4.ip_unprivileged_port_start=443

         Note: this is the recommended way of enabling passt and pasta to bind to ports with numbers below 1024.

       • or the CAP_NET_BIND_SERVICE Linux capability is granted, see services(5) and capabilities(7).

         This  is,  in  general,  not  the  recommended  way, because passt and pasta 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
         AppArmor,  SELinux,  Smack  or  TOMOYO,  and no other processes can interact in such a way in virtue of
         this, granting this capability to passt and pasta only can effectively  prevent  other  processes  from
         utilising it.

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

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

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

   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), passt and pasta try to bind  these  sockets  using  the  observed
       source identifier as "port" -- that corresponds to Echo identifiers for "ping" sockets.

       As  bind(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 bind(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 ping_group_range parameter needs to  include  the  PID  of
       passt or pasta, see icmp(7).

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

   TCP sending window and TCP_INFO 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, passt and pasta routinely query the size of the sending  window
       seen by the kernel on the corresponding socket using the TCP_INFO socket option, see tcp(7). Before Linux
       5.3, i.e. before Linux kernel commit 8f7baad7f035 ("tcp: Add snd_wnd to TCP_INFO"),  the  sending  window
       (snd_wnd 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.

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.

AUTHORS

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

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.

       Copyright (c) 2020-2022 Red Hat GmbH.

       passt and pasta are free software: you can redistribute them and/or modify them under the  terms  of  the
       GNU General Public License as published by the Free Software Foundation, either version 2 of the License,
       or (at your option) any later version.

SEE ALSO

       namespaces(7), qemu(1), qrap(1), slirp4netns(1).

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

                                                                                                        passt(1)