.\" SPDX-License-Identifier: GPL-2.0-or-later .\" Copyright (c) 2020-2022 Red Hat GmbH .\" Author: Stefano Brivio .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 Unless otherwise noted below, \fBif conflicting or multiple options are given, the last one takes effect.\fR .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 This option has no effect, and is maintained for compatibility purposes only. Note that this configuration option is \fBdeprecated\fR and will be removed in a future version. .TP .BR \-l ", " \-\-log-file " " \fIPATH\fR Log to file \fIPATH\fR, and not to the system logger. Specifying this option multiple times does \fInot\fR lead to multiple log files: the last given option takes effect. .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. Specifying this option multiple times does \fInot\fR lead to multiple capture files: the last given option takes effect. .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, 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. .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, 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 \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. If no default routes are available and there is just one interface with any route, that interface will be chosen instead. .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. If no default routes are available and there is just one interface with any route, that interface will be chosen instead. .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. If no default routes are available and there is just one interface with any route, that interface will be chosen instead. .TP .BR \-D ", " \-\-dns " " \fIaddr Instruct the guest (via DHCP, DHVPv6 or NDP) to use \fIaddr\fR (IPv4 or IPv6) as a nameserver, as configured (see options \fB--no-dhcp-dns\fR, \fB--dhcp-dns\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. Unlike addresses from \fI/etc/resolv.conf\fR, \fIaddr\fR is given to the guest without remapping. For example \fB--dns 127.0.0.1\fR will instruct the guest to use itself as nameserver, not the host. .TP .BR \-\-dns-forward " " \fIaddr Map \fIaddr\fR (IPv4 or IPv6) as seen from guest or namespace to the nameserver (with corresponding IP version) specified by the \fB\-\-dns-host\fR option. 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). .TP .BR \-\-dns-host " " \fIaddr Configure the host nameserver which guest or namespace queries to the \fB\-\-dns-forward\fR address will be redirected to. This option can be specified zero to two times (once for IPv4, once for IPv6). By default, the first nameserver from the host's \fI/etc/resolv.conf\fR. .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 \-\-freebind Allow any binding address to be specified for \fB-t\fR and \fB-u\fR options. Usually binding addresses must be addresses currently configured on the host. With \fB\-\-freebind\fR, the \fBIP_FREEBIND\fR or \fBIPV6_FREEBIND\fR socket option is enabled allowing any address to be used. This is typically used to bind addresses which might be configured on the host in future, at which point the forwarding will immediately start operating. .TP .BR \-\-map-host-loopback " " \fIaddr Translate \fIaddr\fR to refer to the host. Packets from the guest to \fIaddr\fR will be redirected to the host. On the host such packets will appear to have both source and destination of 127.0.0.1 or ::1. If \fIaddr\fR is 'none', no address is mapped (this implies \fB--no-map-gw\fR). Only one IPv4 and one IPv6 address can be translated, if the option is specified multiple times, the last one takes effect. Default is to translate the guest's default gateway address, unless \fB--no-map-gw\fR is given, in which case no address is mapped. .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, or if there is no default route, for any of the enabled address families. .TP .BR \-\-map-guest-addr " " \fIaddr Translate \fIaddr\fR in the guest to be equal to the guest's assigned address on the host. That is, packets from the guest to \fIaddr\fR will be redirected to the address assigned to the guest with \fB-a\fR, or by default the host's global address. This allows the guest to access services availble on the host's global address, even though its own address shadows that of the host. If \fIaddr\fR is 'none', no address is mapped. Only one IPv4 and one IPv6 address can be translated, and if the option is specified multiple times, the last one for each address type takes effect. Default is no mapping. .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 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 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, 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 \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, 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 \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: 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: 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: 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: 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: 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: 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 (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 \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 , David Gibson . .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 General Public License as published by the Free Software Foundation, either version 2 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/.