passt/tap.c

140 lines
3.3 KiB
C
Raw Normal View History

passt: New design and implementation with native Layer 4 sockets This is a reimplementation, partially building on the earlier draft, that uses L4 sockets (SOCK_DGRAM, SOCK_STREAM) instead of SOCK_RAW, providing L4-L2 translation functionality without requiring any security capability. Conceptually, this follows the design presented at: https://gitlab.com/abologna/kubevirt-and-kvm/-/blob/master/Networking.md The most significant novelty here comes from TCP and UDP translation layers. In particular, the TCP state and translation logic follows the intent of being minimalistic, without reimplementing a full TCP stack in either direction, and synchronising as much as possible the TCP dynamic and flows between guest and host kernel. Another important introduction concerns addressing, port translation and forwarding. The Layer 4 implementations now attempt to bind on all unbound ports, in order to forward connections in a transparent way. While at it: - the qemu 'tap' back-end can't be used as-is by qrap anymore, because of explicit checks now introduced in qemu to ensure that the corresponding file descriptor is actually a tap device. For this reason, qrap now operates on a 'socket' back-end type, accounting for and building the additional header reporting frame length - provide a demo script that sets up namespaces, addresses and routes, and starts the daemon. A virtual machine started in the network namespace, wrapped by qrap, will now directly interface with passt and communicate using Layer 4 sockets provided by the host kernel. Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
2021-02-16 07:25:09 +01:00
// SPDX-License-Identifier: AGPL-3.0-or-later
/* PASST - Plug A Simple Socket Transport
*
* tap.c - Functions to communicate with guest-facing tap interface
*
* Copyright (c) 2020-2021 Red Hat GmbH
* Author: Stefano Brivio <sbrivio@redhat.com>
*
*/
#include <stdio.h>
#include <limits.h>
#include <string.h>
#include <net/ethernet.h>
#include <net/if.h>
#include <netinet/in.h>
#include <stdint.h>
#include <linux/ip.h>
#include <linux/ipv6.h>
#include <linux/tcp.h>
#include <linux/udp.h>
#include <linux/icmp.h>
#include <linux/icmpv6.h>
#include "passt.h"
#include "util.h"
/**
* tap_send() - Send frame and qemu socket header with indication of length
* @fd: tap file descriptor
* @len: Total L2 packet length
* @flags: Flags for send(), if any
*
* Return: return code from send()
*/
int tap_send(int fd, void *data, size_t len, int flags)
{
uint32_t vnet_len = htonl(len);
send(fd, &vnet_len, 4, MSG_DONTWAIT | MSG_NOSIGNAL);
passt: New design and implementation with native Layer 4 sockets This is a reimplementation, partially building on the earlier draft, that uses L4 sockets (SOCK_DGRAM, SOCK_STREAM) instead of SOCK_RAW, providing L4-L2 translation functionality without requiring any security capability. Conceptually, this follows the design presented at: https://gitlab.com/abologna/kubevirt-and-kvm/-/blob/master/Networking.md The most significant novelty here comes from TCP and UDP translation layers. In particular, the TCP state and translation logic follows the intent of being minimalistic, without reimplementing a full TCP stack in either direction, and synchronising as much as possible the TCP dynamic and flows between guest and host kernel. Another important introduction concerns addressing, port translation and forwarding. The Layer 4 implementations now attempt to bind on all unbound ports, in order to forward connections in a transparent way. While at it: - the qemu 'tap' back-end can't be used as-is by qrap anymore, because of explicit checks now introduced in qemu to ensure that the corresponding file descriptor is actually a tap device. For this reason, qrap now operates on a 'socket' back-end type, accounting for and building the additional header reporting frame length - provide a demo script that sets up namespaces, addresses and routes, and starts the daemon. A virtual machine started in the network namespace, wrapped by qrap, will now directly interface with passt and communicate using Layer 4 sockets provided by the host kernel. Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
2021-02-16 07:25:09 +01:00
return send(fd, data, len, flags | MSG_DONTWAIT | MSG_NOSIGNAL);
passt: New design and implementation with native Layer 4 sockets This is a reimplementation, partially building on the earlier draft, that uses L4 sockets (SOCK_DGRAM, SOCK_STREAM) instead of SOCK_RAW, providing L4-L2 translation functionality without requiring any security capability. Conceptually, this follows the design presented at: https://gitlab.com/abologna/kubevirt-and-kvm/-/blob/master/Networking.md The most significant novelty here comes from TCP and UDP translation layers. In particular, the TCP state and translation logic follows the intent of being minimalistic, without reimplementing a full TCP stack in either direction, and synchronising as much as possible the TCP dynamic and flows between guest and host kernel. Another important introduction concerns addressing, port translation and forwarding. The Layer 4 implementations now attempt to bind on all unbound ports, in order to forward connections in a transparent way. While at it: - the qemu 'tap' back-end can't be used as-is by qrap anymore, because of explicit checks now introduced in qemu to ensure that the corresponding file descriptor is actually a tap device. For this reason, qrap now operates on a 'socket' back-end type, accounting for and building the additional header reporting frame length - provide a demo script that sets up namespaces, addresses and routes, and starts the daemon. A virtual machine started in the network namespace, wrapped by qrap, will now directly interface with passt and communicate using Layer 4 sockets provided by the host kernel. Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
2021-02-16 07:25:09 +01:00
}
/**
* tap_ip_send() - Send IP packet, with L2 headers, calculating L3/L4 checksums
* @c: Execution context
* @src: IPv6 source address, IPv4-mapped for IPv4 sources
* @proto: L4 protocol number
* @in: Payload
* @len: L4 payload length
*/
void tap_ip_send(struct ctx *c, struct in6_addr *src, uint8_t proto,
char *in, size_t len)
{
char pkt[USHRT_MAX];
struct ethhdr *eh;
eh = (struct ethhdr *)pkt;
/* TODO: ARP table lookup */
memcpy(eh->h_dest, c->mac_guest, ETH_ALEN);
memcpy(eh->h_source, c->mac, ETH_ALEN);
if (IN6_IS_ADDR_V4MAPPED(src)) {
struct iphdr *iph = (struct iphdr *)(eh + 1);
char *data = (char *)(iph + 1);
eh->h_proto = ntohs(ETH_P_IP);
iph->version = 4;
iph->ihl = 5;
iph->tos = 0;
iph->tot_len = htons(len + 20);
iph->id = 0;
iph->frag_off = 0;
iph->ttl = 255;
iph->protocol = proto;
udp: Connection tracking for ephemeral, local ports, and related fixes As we support UDP forwarding for packets that are sent to local ports, we actually need some kind of connection tracking for UDP. While at it, this commit introduces a number of vaguely related fixes for issues observed while trying this out. In detail: - implement an explicit, albeit minimalistic, connection tracking for UDP, to allow usage of ephemeral ports by the guest and by the host at the same time, by binding them dynamically as needed, and to allow mapping address changes for packets with a loopback address as destination - set the guest MAC address whenever we receive a packet from tap instead of waiting for an ARP request, and set it to broadcast on start, otherwise DHCPv6 might not work if all DHCPv6 requests time out before the guest starts talking IPv4 - split context IPv6 address into address we assign, global or site address seen on tap, and link-local address seen on tap, and make sure we use the addresses we've seen as destination (link-local choice depends on source address). Similarly, for IPv4, split into address we assign and address we observe, and use the address we observe as destination - introduce a clock_gettime() syscall right after epoll_wait() wakes up, so that we can remove all the other ones and pass the current timestamp to tap and socket handlers -- this is additionally needed by UDP to time out bindings to ephemeral ports and mappings between loopback address and a local address - rename sock_l4_add() to sock_l4(), no semantic changes intended - include <arpa/inet.h> in passt.c before kernel headers so that we can use <netinet/in.h> macros to check IPv6 address types, and remove a duplicate <linux/ip.h> inclusion Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
2021-04-29 16:59:20 +02:00
iph->daddr = c->addr4_seen;
passt: New design and implementation with native Layer 4 sockets This is a reimplementation, partially building on the earlier draft, that uses L4 sockets (SOCK_DGRAM, SOCK_STREAM) instead of SOCK_RAW, providing L4-L2 translation functionality without requiring any security capability. Conceptually, this follows the design presented at: https://gitlab.com/abologna/kubevirt-and-kvm/-/blob/master/Networking.md The most significant novelty here comes from TCP and UDP translation layers. In particular, the TCP state and translation logic follows the intent of being minimalistic, without reimplementing a full TCP stack in either direction, and synchronising as much as possible the TCP dynamic and flows between guest and host kernel. Another important introduction concerns addressing, port translation and forwarding. The Layer 4 implementations now attempt to bind on all unbound ports, in order to forward connections in a transparent way. While at it: - the qemu 'tap' back-end can't be used as-is by qrap anymore, because of explicit checks now introduced in qemu to ensure that the corresponding file descriptor is actually a tap device. For this reason, qrap now operates on a 'socket' back-end type, accounting for and building the additional header reporting frame length - provide a demo script that sets up namespaces, addresses and routes, and starts the daemon. A virtual machine started in the network namespace, wrapped by qrap, will now directly interface with passt and communicate using Layer 4 sockets provided by the host kernel. Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
2021-02-16 07:25:09 +01:00
memcpy(&iph->saddr, &src->s6_addr[12], 4);
iph->check = 0;
iph->check = csum_ip4(iph, iph->ihl * 4);
memcpy(data, in, len);
if (iph->protocol == IPPROTO_TCP) {
csum_tcp4(iph);
} else if (iph->protocol == IPPROTO_UDP) {
struct udphdr *uh = (struct udphdr *)(iph + 1);
uh->check = 0;
}
tap_send(c->fd_unix, pkt, len + sizeof(*iph) + sizeof(*eh), 0);
} else {
struct ipv6hdr *ip6h = (struct ipv6hdr *)(eh + 1);
char *data = (char *)(ip6h + 1);
eh->h_proto = ntohs(ETH_P_IPV6);
memset(ip6h->flow_lbl, 0, 3);
ip6h->payload_len = htons(len);
ip6h->priority = 0;
ip6h->saddr = *src;
udp: Connection tracking for ephemeral, local ports, and related fixes As we support UDP forwarding for packets that are sent to local ports, we actually need some kind of connection tracking for UDP. While at it, this commit introduces a number of vaguely related fixes for issues observed while trying this out. In detail: - implement an explicit, albeit minimalistic, connection tracking for UDP, to allow usage of ephemeral ports by the guest and by the host at the same time, by binding them dynamically as needed, and to allow mapping address changes for packets with a loopback address as destination - set the guest MAC address whenever we receive a packet from tap instead of waiting for an ARP request, and set it to broadcast on start, otherwise DHCPv6 might not work if all DHCPv6 requests time out before the guest starts talking IPv4 - split context IPv6 address into address we assign, global or site address seen on tap, and link-local address seen on tap, and make sure we use the addresses we've seen as destination (link-local choice depends on source address). Similarly, for IPv4, split into address we assign and address we observe, and use the address we observe as destination - introduce a clock_gettime() syscall right after epoll_wait() wakes up, so that we can remove all the other ones and pass the current timestamp to tap and socket handlers -- this is additionally needed by UDP to time out bindings to ephemeral ports and mappings between loopback address and a local address - rename sock_l4_add() to sock_l4(), no semantic changes intended - include <arpa/inet.h> in passt.c before kernel headers so that we can use <netinet/in.h> macros to check IPv6 address types, and remove a duplicate <linux/ip.h> inclusion Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
2021-04-29 16:59:20 +02:00
if (IN6_IS_ADDR_LINKLOCAL(src))
ip6h->daddr = c->addr6_ll_seen;
else
ip6h->daddr = c->addr6_seen;
passt: New design and implementation with native Layer 4 sockets This is a reimplementation, partially building on the earlier draft, that uses L4 sockets (SOCK_DGRAM, SOCK_STREAM) instead of SOCK_RAW, providing L4-L2 translation functionality without requiring any security capability. Conceptually, this follows the design presented at: https://gitlab.com/abologna/kubevirt-and-kvm/-/blob/master/Networking.md The most significant novelty here comes from TCP and UDP translation layers. In particular, the TCP state and translation logic follows the intent of being minimalistic, without reimplementing a full TCP stack in either direction, and synchronising as much as possible the TCP dynamic and flows between guest and host kernel. Another important introduction concerns addressing, port translation and forwarding. The Layer 4 implementations now attempt to bind on all unbound ports, in order to forward connections in a transparent way. While at it: - the qemu 'tap' back-end can't be used as-is by qrap anymore, because of explicit checks now introduced in qemu to ensure that the corresponding file descriptor is actually a tap device. For this reason, qrap now operates on a 'socket' back-end type, accounting for and building the additional header reporting frame length - provide a demo script that sets up namespaces, addresses and routes, and starts the daemon. A virtual machine started in the network namespace, wrapped by qrap, will now directly interface with passt and communicate using Layer 4 sockets provided by the host kernel. Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
2021-02-16 07:25:09 +01:00
memcpy(data, in, len);
ip6h->hop_limit = proto;
ip6h->version = 0;
ip6h->nexthdr = 0;
if (proto == IPPROTO_TCP) {
struct tcphdr *th = (struct tcphdr *)(ip6h + 1);
th->check = 0;
th->check = csum_ip4(ip6h, len + sizeof(*ip6h));
} else if (proto == IPPROTO_UDP) {
struct udphdr *uh = (struct udphdr *)(ip6h + 1);
uh->check = 0;
uh->check = csum_ip4(ip6h, len + sizeof(*ip6h));
} else if (proto == IPPROTO_ICMPV6) {
struct icmp6hdr *ih = (struct icmp6hdr *)(ip6h + 1);
ih->icmp6_cksum = 0;
ih->icmp6_cksum = csum_ip4(ip6h, len + sizeof(*ip6h));
}
ip6h->version = 6;
ip6h->nexthdr = proto;
ip6h->hop_limit = 255;
tap_send(c->fd_unix, pkt, len + sizeof(*ip6h) + sizeof(*eh), 0);
}
}