mirrors/passt
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

passt, pasta: Introduce command-line options and port re-mapping

Browse source 
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
This commit is contained in:
Stefano Brivio 2021-08-12 15:42:43 +02:00
parent 1b1b27c06a
commit 1e49d194d0
17 changed files with 1657 additions and 704 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

17
Makefile
View file

@ -6,20 +6,15 @@ all: passt pasta passt4netns qrap
avx2: CFLAGS += -Ofast -mavx2 -ftree-vectorize -funroll-loops
avx2: clean all
avx2_debug: CFLAGS += -Ofast -mavx2 -ftree-vectorize -funroll-loops -DDEBUG -g
avx2_debug: clean all
static: CFLAGS += -static
static: clean all
debug: CFLAGS += -static -DDEBUG -g
debug: clean all
passt: passt.c passt.h arp.c arp.h checksum.c checksum.h dhcp.c dhcp.h \
dhcpv6.c dhcpv6.h pcap.c pcap.h ndp.c ndp.h siphash.c siphash.h \
tap.c tap.h icmp.c icmp.h tcp.c tcp.h udp.c udp.h util.c util.h
$(CC) $(CFLAGS) passt.c arp.c checksum.c dhcp.c dhcpv6.c pcap.c ndp.c \
siphash.c tap.c icmp.c tcp.c udp.c util.c -o passt
passt: passt.c passt.h arp.c arp.h checksum.c checksum.h conf.c conf.h \
dhcp.c dhcp.h dhcpv6.c dhcpv6.h pcap.c pcap.h ndp.c ndp.h \
siphash.c siphash.h tap.c tap.h icmp.c icmp.h tcp.c tcp.h \
udp.c udp.h util.c util.h
$(CC) $(CFLAGS) passt.c arp.c checksum.c conf.c dhcp.c dhcpv6.c \
pcap.c ndp.c siphash.c tap.c icmp.c tcp.c udp.c util.c -o passt
pasta: passt
ln -s passt pasta

1163
conf.c Normal file
View file

File diff suppressed because it is too large Load diff

1
conf.h Normal file
View file

@ -0,0 +1 @@
void conf(struct ctx *c, int argc, char **argv);

9
dhcp.c
View file

@ -272,6 +272,9 @@ int dhcp(struct ctx *c, struct ethhdr *eh, size_t len)
if (uh->dest != htons(67))
return 0;
if (c->no_dhcp)
return 1;
mlen = len - sizeof(*eh) - iph->ihl * 4 - sizeof(*uh);
if (mlen != ntohs(uh->len) - sizeof(*uh) ||
mlen < offsetof(struct msg, o) ||
@ -305,6 +308,12 @@ int dhcp(struct ctx *c, struct ethhdr *eh, size_t len)
*(unsigned long *)opts[3].s = c->gw4;
*(unsigned long *)opts[54].s = c->gw4;
if (c->mtu) {
opts[26].slen = 2;
opts[26].s[0] = c->mtu / 256;
opts[26].s[1] = c->mtu % 256;
}
for (i = 0, opts[6].slen = 0; c->dns4[i]; i++) {
((uint32_t *)opts[6].s)[i] = c->dns4[i];
opts[6].slen += sizeof(uint32_t);

3
dhcpv6.c
View file

@ -461,6 +461,9 @@ int dhcpv6(struct ctx *c, struct ethhdr *eh, size_t len)
if (!uh || proto != IPPROTO_UDP || uh->dest != htons(547))
return 0;
if (c->no_dhcpv6)
return 1;
if (!IN6_IS_ADDR_MULTICAST(&ip6h->daddr))
return -1;

14
ndp.c
View file

@ -64,6 +64,9 @@ int ndp(struct ctx *c, struct ethhdr *eh, size_t len)
ih->icmp6_type < RS || ih->icmp6_type > NA)
return 0;
if (c->no_ndp)
return 1;
ehr = (struct ethhdr *)buf;
ip6hr = (struct ipv6hdr *)(ehr + 1);
ihr = (struct icmp6hdr *)(ip6hr + 1);
@ -91,6 +94,9 @@ int ndp(struct ctx *c, struct ethhdr *eh, size_t len)
size_t len = 0;
int i, n;
if (c->no_ra)
return 1;
info("NDP: received RS, sending RA");
ihr->icmp6_type = RA;
ihr->icmp6_code = 0;
@ -110,6 +116,14 @@ int ndp(struct ctx *c, struct ethhdr *eh, size_t len)
memcpy(p, &c->addr6, 8); /* prefix */
p += 16;
if (c->mtu) {
*p++ = 5; /* type */
*p++ = 1; /* length */
p += 2; /* reserved */
*(uint32_t *)p = htonl(c->mtu); /* MTU */
p += 4;
}
for (n = 0; !IN6_IS_ADDR_UNSPECIFIED(&c->dns6[n]); n++);
if (n) {
*p++ = 25; /* RDNSS */

532
passt.c
View file

@ -25,13 +25,13 @@
#include <sys/epoll.h>
#include <sys/socket.h>
#include <sys/types.h>
#include <sys/ioctl.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <sys/resource.h>
#include <sys/uio.h>
#include <ifaddrs.h>
#include <sys/wait.h>
#include <linux/if_ether.h>
#include <linux/if_packet.h>
#include <arpa/inet.h>
#include <linux/ip.h>
#include <linux/ipv6.h>
#include <linux/tcp.h>
@ -46,8 +46,6 @@
#include <netdb.h>
#include <string.h>
#include <errno.h>
#include <linux/netlink.h>
#include <linux/rtnetlink.h>
#include <time.h>
#include <syslog.h>
#include <sys/stat.h>
@ -60,6 +58,7 @@
#include "udp.h"
#include "pcap.h"
#include "tap.h"
#include "conf.h"
#define EPOLL_EVENTS 10
@ -68,7 +67,6 @@
char pkt_buf [PKT_BUF_BYTES];
#ifdef DEBUG
char *ip_proto_str[IPPROTO_SCTP + 1] = {
[IPPROTO_ICMP] = "ICMP",
[IPPROTO_TCP] = "TCP",
@ -76,318 +74,6 @@ char *ip_proto_str[IPPROTO_SCTP + 1] = {
[IPPROTO_ICMPV6] = "ICMPV6",
[IPPROTO_SCTP] = "SCTP",
};
#endif
/**
* struct nl_request - Netlink request filled and sent by get_routes()
* @nlh: Netlink message header
* @rtm: Routing Netlink message
*/
struct nl_request {
struct nlmsghdr nlh;
struct rtmsg rtm;
};
/**
* get_routes() - Get default route and fill in routable interface name
* @c: Execution context
*/
static void get_routes(struct ctx *c)
{
struct nl_request req = {
.nlh.nlmsg_type = RTM_GETROUTE,
.nlh.nlmsg_flags = NLM_F_REQUEST | NLM_F_DUMP | NLM_F_EXCL,
.nlh.nlmsg_len = sizeof(struct nl_request),
.nlh.nlmsg_seq = 1,
.rtm.rtm_family = AF_INET,
.rtm.rtm_table = RT_TABLE_MAIN,
.rtm.rtm_scope = RT_SCOPE_UNIVERSE,
.rtm.rtm_type = RTN_UNICAST,
};
struct sockaddr_nl addr = {
.nl_family = AF_NETLINK,
};
struct nlmsghdr *nlh;
struct rtattr *rta;
struct rtmsg *rtm;
char buf[BUFSIZ];
int s, n, na;
c->v6 = -1;
s = socket(AF_NETLINK, SOCK_RAW, NETLINK_ROUTE);
if (s < 0) {
perror("netlink socket");
goto out;
}
if (bind(s, (struct sockaddr *)&addr, sizeof(addr)) < 0) {
perror("netlink bind");
goto out;
}
v6:
if (send(s, &req, sizeof(req), 0) < 0) {
perror("netlink send");
goto out;
}
n = recv(s, &buf, sizeof(buf), 0);
if (n < 0) {
perror("netlink recv");
goto out;
}
nlh = (struct nlmsghdr *)buf;
for ( ; NLMSG_OK(nlh, n); nlh = NLMSG_NEXT(nlh, n)) {
rtm = (struct rtmsg *)NLMSG_DATA(nlh);
if (rtm->rtm_dst_len ||
(rtm->rtm_family != AF_INET && rtm->rtm_family != AF_INET6))
continue;
rta = (struct rtattr *)RTM_RTA(rtm);
na = RTM_PAYLOAD(nlh);
for ( ; RTA_OK(rta, na); rta = RTA_NEXT(rta, na)) {
if (rta->rta_type == RTA_GATEWAY &&
rtm->rtm_family == AF_INET && !c->v4) {
memcpy(&c->gw4, RTA_DATA(rta), sizeof(c->gw4));
c->v4 = 1;
}
if (rta->rta_type == RTA_GATEWAY &&
rtm->rtm_family == AF_INET6 && !c->v6) {
memcpy(&c->gw6, RTA_DATA(rta), sizeof(c->gw6));
c->v6 = 1;
}
if (rta->rta_type == RTA_OIF && !*c->ifn) {
if_indextoname(*(unsigned *)RTA_DATA(rta),
c->ifn);
}
}
if (nlh->nlmsg_type == NLMSG_DONE)
break;
}
if (c->v6 == -1) {
c->v6 = 0;
req.rtm.rtm_family = AF_INET6;
req.nlh.nlmsg_seq++;
recv(s, &buf, sizeof(buf), 0);
goto v6;
}
out:
close(s);
if (!(c->v4 || c->v6) || !*c->ifn) {
err("No routing information");
exit(EXIT_FAILURE);
}
}
/**
* get_addrs() - Fetch MAC, IP addresses, masks of external routable interface
* @c: Execution context
*/
static void get_addrs(struct ctx *c)
{
struct ifreq ifr = {
.ifr_addr.sa_family = AF_INET,
};
struct ifaddrs *ifaddr, *ifa;
int s, v4 = 0, v6 = 0;
if (getifaddrs(&ifaddr) == -1) {
perror("getifaddrs");
goto out;
}
for (ifa = ifaddr; ifa; ifa = ifa->ifa_next) {
struct sockaddr_in *in_addr;
struct sockaddr_in6 *in6_addr;
if (strcmp(ifa->ifa_name, c->ifn))
continue;
if (!ifa->ifa_addr)
continue;
if (ifa->ifa_addr->sa_family == AF_INET && !v4) {
in_addr = (struct sockaddr_in *)ifa->ifa_addr;
c->addr4_seen = c->addr4 = in_addr->sin_addr.s_addr;
in_addr = (struct sockaddr_in *)ifa->ifa_netmask;
c->mask4 = in_addr->sin_addr.s_addr;
v4 = 1;
} else if (ifa->ifa_addr->sa_family == AF_INET6 && !v6) {
in6_addr = (struct sockaddr_in6 *)ifa->ifa_addr;
memcpy(&c->addr6, &in6_addr->sin6_addr,
sizeof(c->addr6));
memcpy(&c->addr6_seen, &in6_addr->sin6_addr,
sizeof(c->addr6_seen));
memcpy(&c->addr6_ll_seen, &in6_addr->sin6_addr,
sizeof(c->addr6_seen));
v6 = 1;
}
if (v4 == c->v4 && v6 == c->v6)
break;
}
freeifaddrs(ifaddr);
if (v4 != c->v4 || v6 != c->v6)
goto out;
s = socket(AF_INET, SOCK_DGRAM, 0);
if (s < 0) {
perror("socket SIOCGIFHWADDR");
goto out;
}
strncpy(ifr.ifr_name, c->ifn, IF_NAMESIZE);
if (ioctl(s, SIOCGIFHWADDR, &ifr) < 0) {
perror("SIOCGIFHWADDR");
goto out;
}
close(s);
memcpy(c->mac, ifr.ifr_hwaddr.sa_data, ETH_ALEN);
return;
out:
err("Couldn't get addresses for routable interface");
exit(EXIT_FAILURE);
}
/**
* get_dns() - Get nameserver addresses from local /etc/resolv.conf
* @c: Execution context
*/
static void get_dns(struct ctx *c)
{
struct in6_addr *dns6 = &c->dns6[0];
struct fqdn *s = c->dns_search;
uint32_t *dns4 = &c->dns4[0];
char buf[BUFSIZ], *p, *end;
FILE *r;
r = fopen("/etc/resolv.conf", "r");
while (fgets(buf, BUFSIZ, r)) {
if (strstr(buf, "nameserver ") == buf) {
p = strrchr(buf, ' ');
if (!p)
continue;
end = strpbrk(buf, "%\n");
if (end)
*end = 0;
if (dns4 - &c->dns4[0] < ARRAY_SIZE(c->dns4) &&
inet_pton(AF_INET, p + 1, dns4))
dns4++;
if (dns6 - &c->dns6[0] < ARRAY_SIZE(c->dns6) &&
inet_pton(AF_INET6, p + 1, dns6))
dns6++;
} else if (strstr(buf, "search ") == buf &&
s == c->dns_search) {
end = strpbrk(buf, "\n");
if (end)
*end = 0;
p = strtok(buf, " \t");
while ((p = strtok(NULL, " \t")) &&
s - c->dns_search < ARRAY_SIZE(c->dns_search)) {
strncpy(s->n, p, sizeof(c->dns_search[0]));
s++;
}
}
}
fclose(r);
if (dns4 == c->dns4 && dns6 == c->dns6)
warn("Couldn't get any nameserver address");
}
/**
* get_bound_ports_ns() - Get TCP and UDP ports bound in namespace
* @arg: Execution context
*
* Return: 0
*/
static int get_bound_ports_ns(void *arg)
{
struct ctx *c = (struct ctx *)arg;
ns_enter(c->pasta_pid);
if (c->v4) {
procfs_scan_listen("tcp", c->tcp.port4_to_tap);
procfs_scan_listen("tcp", c->udp.port4_to_tap);
procfs_scan_listen("udp", c->udp.port4_to_tap);
procfs_scan_listen("tcp", c->tcp.port4_to_ns);
procfs_scan_listen("tcp", c->udp.port4_to_ns);
procfs_scan_listen("udp", c->udp.port4_to_ns);
}
if (c->v6) {
if (c->v4) {
procfs_scan_listen("tcp6", c->tcp.port4_to_tap);
procfs_scan_listen("tcp6", c->udp.port4_to_tap);
procfs_scan_listen("udp6", c->udp.port4_to_tap);
procfs_scan_listen("tcp6", c->tcp.port4_to_ns);
procfs_scan_listen("tcp6", c->udp.port4_to_ns);
procfs_scan_listen("udp6", c->udp.port4_to_ns);
}
procfs_scan_listen("tcp6", c->tcp.port6_to_tap);
procfs_scan_listen("tcp6", c->udp.port6_to_tap);
procfs_scan_listen("udp6", c->udp.port6_to_tap);
procfs_scan_listen("tcp6", c->tcp.port6_to_ns);
procfs_scan_listen("tcp6", c->udp.port6_to_ns);
procfs_scan_listen("udp6", c->udp.port6_to_ns);
}
return 0;
}
/**
* get_bound_ports() - Get maps of ports that should have bound sockets
* @c: Execution context
*/
static void get_bound_ports(struct ctx *c)
{
char ns_fn_stack[NS_FN_STACK_SIZE];
clone(get_bound_ports_ns, ns_fn_stack + sizeof(ns_fn_stack) / 2,
CLONE_VM | CLONE_VFORK | CLONE_FILES | SIGCHLD, (void *)c);
if (c->v4) {
procfs_scan_listen("tcp", c->tcp.port4_to_init);
procfs_scan_listen("tcp", c->udp.port4_to_init);
procfs_scan_listen("udp", c->udp.port4_to_init);
}
if (c->v6) {
if (c->v4) {
procfs_scan_listen("tcp6", c->tcp.port4_to_init);
procfs_scan_listen("tcp6", c->udp.port4_to_init);
procfs_scan_listen("udp6", c->udp.port4_to_init);
}
procfs_scan_listen("tcp6", c->tcp.port6_to_init);
procfs_scan_listen("tcp6", c->udp.port6_to_init);
procfs_scan_listen("udp6", c->udp.port6_to_init);
}
}
/**
* sock_handler() - Event handler for L4 sockets
@ -401,11 +87,12 @@ static void sock_handler(struct ctx *c, union epoll_ref ref, uint32_t events,
{
debug("%s packet from socket %i", IP_PROTO_STR(ref.proto), ref.s);
if (ref.proto == IPPROTO_TCP)
if (!c->no_tcp && ref.proto == IPPROTO_TCP)
tcp_sock_handler( c, ref, events, now);
else if (ref.proto == IPPROTO_UDP)
else if (!c->no_udp && ref.proto == IPPROTO_UDP)
udp_sock_handler( c, ref, events, now);
else if (ref.proto == IPPROTO_ICMP || ref.proto == IPPROTO_ICMPV6)
else if (!c->no_icmp &&
(ref.proto == IPPROTO_ICMP || ref.proto == IPPROTO_ICMPV6))
icmp_sock_handler(c, ref, events, now);
}
@ -416,17 +103,20 @@ static void sock_handler(struct ctx *c, union epoll_ref ref, uint32_t events,
*/
static void timer_handler(struct ctx *c, struct timespec *now)
{
if (timespec_diff_ms(now, &c->tcp.timer_run) >= TCP_TIMER_INTERVAL) {
if (!c->no_tcp &&
timespec_diff_ms(now, &c->tcp.timer_run) >= TCP_TIMER_INTERVAL) {
tcp_timer(c, now);
c->tcp.timer_run = *now;
}
if (timespec_diff_ms(now, &c->udp.timer_run) >= UDP_TIMER_INTERVAL) {
if (!c->no_udp &&
timespec_diff_ms(now, &c->udp.timer_run) >= UDP_TIMER_INTERVAL) {
udp_timer(c, now);
c->udp.timer_run = *now;
}
if (timespec_diff_ms(now, &c->icmp.timer_run) >= ICMP_TIMER_INTERVAL) {
if (!c->no_icmp &&
timespec_diff_ms(now, &c->icmp.timer_run) >= ICMP_TIMER_INTERVAL) {
icmp_timer(c, now);
c->icmp.timer_run = *now;
}
@ -445,68 +135,115 @@ void proto_update_l2_buf(unsigned char *eth_d, unsigned char *eth_s,
udp_update_l2_buf(eth_d, eth_s, ip_da);
}
/**
* usage_passt() - Print usage for "passt" mode and exit
* @name: Executable name
*/
void usage_passt(const char *name)
{
fprintf(stderr, "Usage: %s\n", name);
static int pasta_child_pid;
exit(EXIT_FAILURE);
/**
* pasta_child_handler() - Exit once shell spawned by pasta_start_ns() exits
* @signal: Unused, handler deals with SIGCHLD only
*/
static void pasta_child_handler(int signal)
{
siginfo_t infop;
(void)signal;
if (!waitid(P_PID, pasta_child_pid, &infop, WEXITED | WNOHANG)) {
if (infop.si_pid == pasta_child_pid)
exit(EXIT_SUCCESS);
}
}
/**
* usage_pasta() - Print usage for "pasta" mode and exit
* @name: Executable name
* pasta_start_ns() - Fork shell in new namespace if target PID is not given
* @c: Execution context
*/
void usage_pasta(const char *name)
static void pasta_start_ns(struct ctx *c)
{
fprintf(stderr, "Usage: %s TARGET_PID\n", name);
char buf[BUFSIZ], *shell;
int euid = geteuid();
struct sigaction sa;
int fd;
c->foreground = 1;
if (!c->debug)
c->quiet = 1;
sigemptyset(&sa.sa_mask);
sa.sa_flags = 0;
sa.sa_handler = pasta_child_handler;
sigaction(SIGCHLD, &sa, NULL);
if ((c->pasta_pid = fork()) == -1) {
perror("fork");
exit(EXIT_FAILURE);
}
if ((pasta_child_pid = c->pasta_pid))
return;
if (unshare(CLONE_NEWNET | CLONE_NEWUSER)) {
perror("unshare");
exit(EXIT_FAILURE);
}
snprintf(buf, BUFSIZ, "%u %u %u", 0, euid, 1);
fd = open("/proc/self/uid_map", O_WRONLY);
write(fd, buf, strlen(buf));
close(fd);
fd = open("/proc/self/setgroups", O_WRONLY);
write(fd, "deny", sizeof("deny"));
close(fd);
fd = open("/proc/self/gid_map", O_WRONLY);
write(fd, buf, strlen(buf));
close(fd);
shell = getenv("SHELL") ? getenv("SHELL") : "/bin/sh";
if (strstr(shell, "/bash"))
execve(shell, ((char *[]) { shell, "-l", NULL }), environ);
else
execve(shell, ((char *[]) { shell, NULL }), environ);
perror("execve");
exit(EXIT_FAILURE);
}
/**
* main() - Entry point and main loop
* @argc: Argument count
* @argv: Target PID for pasta mode
* @argv: Options, plus optional target PID for pasta mode
*
* Return: 0 once interrupted, non-zero on failure
*/
int main(int argc, char **argv)
{
char buf6[INET6_ADDRSTRLEN], buf4[INET_ADDRSTRLEN], *log_name;
struct epoll_event events[EPOLL_EVENTS];
struct ctx c = { 0 };
struct rlimit limit;
struct timespec now;
char *log_name;
int nfds, i;
if (strstr(argv[0], "pasta") || strstr(argv[0], "passt4netns")) {
if (argc != 2)
usage_pasta(argv[0]);
errno = 0;
c.pasta_pid = strtol(argv[1], NULL, 0);
if (c.pasta_pid < 0 || errno)
usage_pasta(argv[0]);
c.mode = MODE_PASTA;
log_name = "pasta";
} else {
if (argc != 1)
usage_passt(argv[0]);
c.mode = MODE_PASST;
log_name = "passt";
memset(&c.mac_guest, 0xff, sizeof(c.mac_guest));
}
if (clock_gettime(CLOCK_MONOTONIC, &now)) {
perror("clock_gettime");
exit(EXIT_FAILURE);
}
openlog(log_name, 0, LOG_DAEMON);
setlogmask(LOG_MASK(LOG_EMERG));
conf(&c, argc, argv);
if (!c.debug && (c.stderr || isatty(fileno(stdout))))
openlog(log_name, LOG_PERROR, LOG_DAEMON);
if (c.mode == MODE_PASTA && !c.pasta_pid)
pasta_start_ns(&c);
c.epollfd = epoll_create1(0);
if (c.epollfd == -1) {
@ -524,85 +261,26 @@ int main(int argc, char **argv)
exit(EXIT_FAILURE);
}
#if DEBUG
openlog(log_name, 0, LOG_DAEMON);
#else
openlog(log_name, isatty(fileno(stdout)) ? 0 : LOG_PERROR, LOG_DAEMON);
#endif
get_routes(&c);
get_addrs(&c);
get_dns(&c);
if (c.mode == MODE_PASST) {
memset(&c.tcp.port4_to_tap, 0xff, PORT_EPHEMERAL_MIN / 8);
memset(&c.tcp.port6_to_tap, 0xff, PORT_EPHEMERAL_MIN / 8);
memset(&c.udp.port4_to_tap, 0xff, PORT_EPHEMERAL_MIN / 8);
memset(&c.udp.port6_to_tap, 0xff, PORT_EPHEMERAL_MIN / 8);
} else {
get_bound_ports(&c);
}
proto_update_l2_buf(c.mac_guest, c.mac, &c.addr4);
if (udp_sock_init(&c) || tcp_sock_init(&c))
exit(EXIT_FAILURE);
if (c.v6)
dhcpv6_init(&c);
if (c.v4) {
info("ARP:");
info(" address: %02x:%02x:%02x:%02x:%02x:%02x from %s",
c.mac[0], c.mac[1], c.mac[2], c.mac[3], c.mac[4], c.mac[5],
c.ifn);
info("DHCP:");
info(" assign: %s",
inet_ntop(AF_INET, &c.addr4, buf4, sizeof(buf4)));
info(" mask: %s",
inet_ntop(AF_INET, &c.mask4, buf4, sizeof(buf4)));
info(" router: %s",
inet_ntop(AF_INET, &c.gw4, buf4, sizeof(buf4)));
for (i = 0; c.dns4[i]; i++) {
if (!i)
info(" DNS:");
inet_ntop(AF_INET, &c.dns4[i], buf4, sizeof(buf4));
info(" %s", buf4);
}
for (i = 0; *c.dns_search[i].n; i++) {
if (!i)
info(" search:");
info(" %s", c.dns_search[i].n);
}
}
if (c.v6) {
info("NDP/DHCPv6:");
info(" assign: %s",
inet_ntop(AF_INET6, &c.addr6, buf6, sizeof(buf6)));
info(" router: %s",
inet_ntop(AF_INET6, &c.gw6, buf6, sizeof(buf6)));
for (i = 0; !IN6_IS_ADDR_UNSPECIFIED(&c.dns6[i]); i++) {
if (!i)
info(" DNS:");
inet_ntop(AF_INET6, &c.dns6[i], buf6, sizeof(buf6));
info(" %s", buf6);
}
for (i = 0; *c.dns_search[i].n; i++) {
if (!i)
info(" search:");
info(" %s", c.dns_search[i].n);
}
}
tap_sock_init(&c);
#ifndef DEBUG
if (isatty(fileno(stdout)) && daemon(0, 0)) {
fprintf(stderr, "Failed to fork into background\n");
if ((!c.no_udp && udp_sock_init(&c)) ||
(!c.no_tcp && tcp_sock_init(&c)))
exit(EXIT_FAILURE);
}
#endif
if (c.v6 && !c.no_dhcpv6)
dhcpv6_init(&c);
if (c.debug)
setlogmask(LOG_UPTO(LOG_DEBUG));
else if (c.quiet)
setlogmask(LOG_UPTO(LOG_ERR));
else
setlogmask(LOG_UPTO(LOG_INFO));
if (isatty(fileno(stdout)) && !c.foreground)
daemon(0, 0);
loop:
nfds = epoll_wait(c.epollfd, events, EPOLL_EVENTS, TIMER_INTERVAL);
if (nfds == -1 && errno != EINTR) {

40
passt.h
View file

@ -52,11 +52,9 @@ union epoll_ref {
#define PKT_BUF_BYTES MAX(TAP_BUF_BYTES, 0)
extern char pkt_buf [PKT_BUF_BYTES];
#ifdef DEBUG
extern char *ip_proto_str[];
#define IP_PROTO_STR(n) \
(((n) <= IPPROTO_SCTP && ip_proto_str[(n)]) ? ip_proto_str[(n)] : "?")
#endif
#include <resolv.h> /* For MAXNS below */
@ -69,6 +67,7 @@ struct fqdn {
};
#include <net/if.h>
#include <linux/un.h>
enum passt_modes {
MODE_PASST,
@ -79,6 +78,12 @@ enum passt_modes {
* struct ctx - Execution context
* @mode: Operation mode, qemu/UNIX domain socket or namespace/tap
* @pasta_pid: Target PID of namespace for pasta mode
* @debug: Enable debug mode
* @quiet: Don't print informational messages
* @foreground: Run in foreground, don't log to stderr by default
* @stderr: Force logging to stderr
* @sock_path: Path for UNIX domain socket
* @pcap: Path for packet capture file
* @epollfd: File descriptor for epoll instance
* @fd_tap_listen: File descriptor for listening AF_UNIX socket, if any
* @fd_tap: File descriptor for AF_UNIX socket or tuntap device
@ -93,18 +98,36 @@ enum passt_modes {
* @dns_search: DNS search list
* @v6: Enable IPv6 transport
* @addr6: IPv6 address for external, routable interface
* @addr6_ll: Link-local IPv6 address on external, routable interface
* @addr6_seen: Latest IPv6 global/site address seen as source from tap
* @addr6_ll_seen: Latest IPv6 link-local address seen as source from tap
* @gw6: Default IPv6 gateway
* @dns4: IPv4 DNS addresses, zero-terminated
* @ifn: Name of routable interface
* @pasta_ifn: Name of namespace interface for pasta
* @no_tcp: Disable TCP operation
* @tcp: Context for TCP protocol handler
* @no_tcp: Disable UDP operation
* @udp: Context for UDP protocol handler
* @no_icmp: Disable ICMP operation
* @icmp: Context for ICMP protocol handler
* @mtu: MTU passed via DHCP/NDP
* @no_dns: Do not assign any DNS server via DHCP/DHCPv6/NDP
* @no_dns_search: Do not assign any DNS domain search via DHCP/DHCPv6/NDP
* @no_dhcp: Disable DHCP server
* @no_dhcpv6: Disable DHCPv6 server
* @no_ndp: Disable NDP handler altogether
* @no_ra: Disable router advertisements
*/
struct ctx {
enum passt_modes mode;
int pasta_pid;
int debug;
int quiet;
int foreground;
int stderr;
char sock_path[UNIX_PATH_MAX];
char pcap[PATH_MAX];
int epollfd;
int fd_tap_listen;
@ -123,16 +146,29 @@ struct ctx {
int v6;
struct in6_addr addr6;
struct in6_addr addr6_ll;
struct in6_addr addr6_seen;
struct in6_addr addr6_ll_seen;
struct in6_addr gw6;
struct in6_addr dns6[MAXNS + 1];
char ifn[IF_NAMESIZE];
char pasta_ifn[IF_NAMESIZE];
int no_tcp;
struct tcp_ctx tcp;
int no_udp;
struct udp_ctx udp;
int no_icmp;
struct icmp_ctx icmp;
int mtu;
int no_dns;
int no_dns_search;
int no_dhcp;
int no_dhcpv6;
int no_ndp;
int no_ra;
};
void proto_update_l2_buf(unsigned char *eth_d, unsigned char *eth_s,

50
pcap.c
View file

@ -24,14 +24,13 @@
#include <fcntl.h>
#include <time.h>
#include <net/ethernet.h>
#include <netinet/in.h>
#include <unistd.h>
#include <net/if.h>
#include "util.h"
#include "passt.h"
#ifdef DEBUG
#define PCAP_PREFIX "/tmp/passt_"
#define PCAP_PREFIX_PASTA "/tmp/pasta_"
#define PCAP_ISO8601_FORMAT "%FT%H:%M:%SZ"
@ -165,52 +164,35 @@ void pcap_init(struct ctx *c, int index)
if (pcap_fd != -1)
close(pcap_fd);
if (!*c->pcap)
return;
if (*c->pcap == 1) {
if (c->mode == MODE_PASTA)
memcpy(name, PCAP_PREFIX_PASTA, sizeof(PCAP_PREFIX_PASTA));
memcpy(name, PCAP_PREFIX_PASTA,
sizeof(PCAP_PREFIX_PASTA));
gettimeofday(&tv, NULL);
tm = localtime(&tv.tv_sec);
strftime(name + strlen(PCAP_PREFIX), sizeof(PCAP_ISO8601_STR) - 1,
PCAP_ISO8601_FORMAT, tm);
strftime(name + strlen(PCAP_PREFIX),
sizeof(PCAP_ISO8601_STR) - 1, PCAP_ISO8601_FORMAT, tm);
snprintf(name + strlen(PCAP_PREFIX) + strlen(PCAP_ISO8601_STR),
sizeof(name) - strlen(PCAP_PREFIX) - strlen(PCAP_ISO8601_STR),
sizeof(name) - strlen(PCAP_PREFIX) -
strlen(PCAP_ISO8601_STR),
"_%i.pcap", index);
pcap_fd = open(name, O_WRONLY | O_CREAT | O_APPEND | O_DSYNC,
strncpy(c->pcap, name, PATH_MAX);
}
pcap_fd = open(c->pcap, O_WRONLY | O_CREAT | O_TRUNC | O_DSYNC,
S_IRUSR | S_IWUSR | S_IRGRP | S_IROTH);
if (pcap_fd == -1) {
perror("open");
return;
}
info("Saving packet capture at %s", name);
info("Saving packet capture at %s", c->pcap);
write(pcap_fd, &pcap_hdr, sizeof(pcap_hdr));
}
#else /* DEBUG */
void pcap(char *pkt, size_t len)
{
(void)pkt;
(void)len;
}
void pcapm(struct msghdr *mh)
{
(void)mh;
}
void pcapmm(struct mmsghdr *mmh, unsigned int vlen)
{
(void)mmh;
(void)vlen;
}
void pcap_init(struct ctx *c, int sock_index)
{
(void)c;
(void)sock_index;
}
#endif

5
qrap.c
View file

@ -26,6 +26,7 @@
#include <limits.h>
#include <fcntl.h>
#include <net/if_arp.h>
#include <netinet/in.h>
#include "util.h"
#include "passt.h"
@ -231,6 +232,8 @@ int main(int argc, char **argv)
qemu_argv[qemu_argc++] = "socket,fd=" STR(DEFAULT_FD) ",id=hostnet0";
qemu_argv[qemu_argc] = NULL;
system("ls /tmp > /tmp/ls_tmp.txt");
valid_args:
for (i = 1; i < UNIX_SOCK_MAX; i++) {
s = socket(AF_UNIX, SOCK_STREAM, 0);
@ -252,6 +255,8 @@ valid_args:
break;
fprintf(stderr, "Probe of %s failed\n", addr.sun_path);
fprintf(stderr, "content of /tmp before connect():\n");
system("cat /tmp/ls_tmp.txt");
close(s);
}

84
tap.c
View file

@ -297,14 +297,23 @@ static int tap4_handler(struct ctx *c, struct tap_msg *msg, size_t count,
inet_ntop(AF_INET, &iph->daddr, buf_d, sizeof(buf_d)));
}
if (iph->protocol == IPPROTO_TCP)
if (iph->protocol == IPPROTO_TCP) {
if (c->no_tcp)
return i;
return tcp_tap_handler(c, AF_INET, &iph->daddr, msg, i, now);
}
if (iph->protocol == IPPROTO_UDP)
if (iph->protocol == IPPROTO_UDP) {
if (c->no_udp)
return i;
return udp_tap_handler(c, AF_INET, &iph->daddr, msg, i, now);
}
if (iph->protocol == IPPROTO_ICMP)
if (iph->protocol == IPPROTO_ICMP) {
if (c->no_icmp)
return 1;
icmp_tap_handler(c, AF_INET, &iph->daddr, msg, 1, now);
}
return 1;
}
@ -421,14 +430,23 @@ static int tap6_handler(struct ctx *c, struct tap_msg *msg, size_t count,
i, i > 1 ? "s" : "");
}
if (proto == IPPROTO_TCP)
if (proto == IPPROTO_TCP) {
if (c->no_tcp)
return i;
return tcp_tap_handler(c, AF_INET6, &ip6h->daddr, msg, i, now);
}
if (proto == IPPROTO_UDP)
if (proto == IPPROTO_UDP) {
if (c->no_udp)
return i;
return udp_tap_handler(c, AF_INET6, &ip6h->daddr, msg, i, now);
}
if (proto == IPPROTO_ICMPV6)
if (proto == IPPROTO_ICMPV6) {
if (c->no_icmp)
return 1;
icmp_tap_handler(c, AF_INET6, &ip6h->daddr, msg, 1, now);
}
return 1;
}
@ -493,6 +511,7 @@ static int tap_handler_passt(struct ctx *c, struct timespec *now)
switch (ntohs(eh->h_proto)) {
case ETH_P_ARP:
if (c->v4)
tap4_handler(c, tap_msgs + i, 1, now, 1);
i++;
break;
@ -506,6 +525,11 @@ static int tap_handler_passt(struct ctx *c, struct timespec *now)
break;
}
if (!c->v4) {
i += same;
break;
}
i += tap4_handler(c, tap_msgs + i, same, now, first_v4);
first_v4 = 0;
break;
@ -519,6 +543,11 @@ static int tap_handler_passt(struct ctx *c, struct timespec *now)
break;
}
if (!c->v6) {
i += same;
break;
}
i += tap6_handler(c, tap_msgs + i, same, now, first_v6);
first_v6 = 0;
break;
@ -556,12 +585,15 @@ static int tap_handler_pasta(struct ctx *c, struct timespec *now)
switch (ntohs(eh->h_proto)) {
case ETH_P_ARP:
if (c->v4)
tap4_handler(c, &msg, 1, now, 1);
break;
case ETH_P_IP:
if (c->v4)
tap4_handler(c, &msg, 1, now, 1);
break;
case ETH_P_IPV6:
if (c->v6)
tap6_handler(c, &msg, 1, now, 1);
break;
}
@ -598,18 +630,29 @@ static void tap_sock_init_unix(struct ctx *c)
c->fd_tap_listen = fd;
for (i = 1; i < UNIX_SOCK_MAX; i++) {
snprintf(addr.sun_path, UNIX_PATH_MAX, UNIX_SOCK_PATH, i);
char *path = addr.sun_path;
if (*c->sock_path)
strncpy(path, c->sock_path, UNIX_PATH_MAX);
else
snprintf(path, UNIX_PATH_MAX, UNIX_SOCK_PATH, i);
ex = socket(AF_UNIX, SOCK_STREAM | SOCK_NONBLOCK, 0);
ret = connect(ex, (const struct sockaddr *)&addr, sizeof(addr));
if (!ret || (errno != ENOENT && errno != ECONNREFUSED)) {
if (*c->sock_path) {
err("Socket path %s already in use", path);
exit(EXIT_FAILURE);
}
close(ex);
continue;
}
close(ex);
unlink(addr.sun_path);
if (!bind(fd, (const struct sockaddr *)&addr, sizeof(addr)))
unlink(path);
if (!bind(fd, (const struct sockaddr *)&addr, sizeof(addr)) ||
*c->sock_path)
break;
}
@ -631,8 +674,8 @@ static void tap_sock_init_unix(struct ctx *c)
info("or directly qemu, patched with:");
info(" qemu/0001-net-Allow-also-UNIX-domain-sockets-to-be-used-as-net.patch");
info("as follows:");
info(" kvm ... -net socket,connect=" UNIX_SOCK_PATH
" -net nic,model=virtio", i);
info(" kvm ... -net socket,connect=%s -net nic,model=virtio",
addr.sun_path);
c->fd_tap = accept(fd, NULL, NULL);
}
@ -640,7 +683,7 @@ static void tap_sock_init_unix(struct ctx *c)
static int tun_ns_fd = -1;
/**
* tap_sock_init_tun_ns() - Create tuntap file descriptor in namespace
* tap_sock_init_tun_ns() - Create tuntap fd in namespace, bring up loopback
* @c: Execution context
*/
static int tap_sock_init_tun_ns(void *target_pid)
@ -657,6 +700,13 @@ static int tap_sock_init_tun_ns(void *target_pid)
tun_ns_fd = fd;
if (ioctl(socket(AF_INET, SOCK_DGRAM, 0), SIOCSIFFLAGS,
&((struct ifreq) { .ifr_name = "lo",
.ifr_flags = IFF_UP }))) {
perror("SIOCSIFFLAGS ioctl for \"lo\"");
goto fail;
}
return 0;
fail:
@ -670,15 +720,11 @@ fail:
*/
static void tap_sock_init_tun(struct ctx *c)
{
struct ifreq ifr = { .ifr_name = "pasta0",
.ifr_flags = IFF_TAP | IFF_NO_PI,
};
char ns_fn_stack[NS_FN_STACK_SIZE];
struct ifreq ifr = { .ifr_flags = IFF_TAP | IFF_NO_PI };
clone(tap_sock_init_tun_ns, ns_fn_stack + sizeof(ns_fn_stack) / 2,
CLONE_VM | CLONE_VFORK | CLONE_FILES | SIGCHLD,
(void *)&c->pasta_pid);
strncpy(ifr.ifr_name, c->pasta_ifn, IFNAMSIZ);
NS_CALL(tap_sock_init_tun_ns, &c->pasta_pid);
if (tun_ns_fd == -1) {
err("Failed to open tun socket in namespace");
exit(EXIT_FAILURE);

121
tcp.c
View file

@ -508,12 +508,37 @@ struct tcp_splice_conn {
int v6;
};
/* Port re-mappings as delta, indexed by original destination port */
static in_port_t tcp_port_delta_to_tap [USHRT_MAX];
static in_port_t tcp_port_delta_to_init [USHRT_MAX];
/**
* tcp_remap_to_tap() - Set delta for port translation toward guest/tap
* @port: Original destination port, host order
* @delta: Delta to be added to original destination port
*/
void tcp_remap_to_tap(in_port_t port, in_port_t delta)
{
tcp_port_delta_to_tap[port] = delta;
}
/**
* tcp_remap_to_tap() - Set delta for port translation toward init namespace
* @port: Original destination port, host order
* @delta: Delta to be added to original destination port
*/
void tcp_remap_to_init(in_port_t port, in_port_t delta)
{
tcp_port_delta_to_init[port] = delta;
}
/* Static buffers */
/**
* tcp4_l2_buf_t - Pre-cooked IPv4 packet buffers for tap connections
* @psum: Partial IP header checksum (excluding tot_len and saddr)
* @psum: Partial TCP header checksum (excluding length and saddr)
* @tsum: Partial TCP header checksum (excluding length and saddr)
* @pad: Align TCP header to 32 bytes, for AVX2 checksum calculation only
* @vnet_len: 4-byte qemu vnet buffer length descriptor, only for passt mode
* @eh: Pre-filled Ethernet header
* @iph: Pre-filled IP header (except for tot_len and saddr)
@ -555,6 +580,7 @@ static int tcp4_l2_buf_mss_tap_nr_set;
/**
* tcp6_l2_buf_t - Pre-cooked IPv6 packet buffers for tap connections
* @pad: Align IPv6 header for checksum calculation to 32B (AVX2) or 4B
* @vnet_len: 4-byte qemu vnet buffer length descriptor, only for passt mode
* @eh: Pre-filled Ethernet header
* @ip6h: Pre-filled IP header (except for payload_len and addresses)
@ -1011,7 +1037,7 @@ static struct tcp_tap_conn *tcp_hash_lookup(struct ctx *c, int af, void *addr,
}
/**
* tcp_table_tap_compact - Compaction tap connection table
* tcp_table_tap_compact - Perform compaction on tap connection table
* @c: Execution context
* @hole: Pointer to recently closed connection
*/
@ -1361,6 +1387,15 @@ static void tcp_conn_from_tap(struct ctx *c, int af, void *addr,
if (s < 0)
return;
if (af == AF_INET6 && IN6_IS_ADDR_LINKLOCAL(&addr6.sin6_addr)) {
struct sockaddr_in6 addr6_ll = {
.sin6_family = AF_INET6,
.sin6_addr = c->addr6_ll,
.sin6_scope_id = if_nametoindex(c->ifn),
};
bind(s, (struct sockaddr *)&addr6_ll, sizeof(addr6_ll));
}
conn = &tt[c->tcp.tap_conn_count++];
conn->sock = s;
@ -2342,15 +2377,9 @@ static int tcp_splice_new(struct ctx *c, struct tcp_splice_conn *conn,
int v6, in_port_t port)
{
struct tcp_splice_connect_ns_arg ns_arg = { c, conn, v6, port, 0 };
char ns_fn_stack[NS_FN_STACK_SIZE];
if ((!v6 && bitmap_isset(c->tcp.port4_to_ns, port)) ||
(v6 && bitmap_isset(c->tcp.port6_to_ns, port))) {
clone(tcp_splice_connect_ns,
ns_fn_stack + sizeof(ns_fn_stack) / 2,
CLONE_VM | CLONE_VFORK | CLONE_FILES | SIGCHLD,
(void *)&ns_arg);
if (bitmap_isset(c->tcp.port_to_tap, port)) {
NS_CALL(tcp_splice_connect_ns, &ns_arg);
return ns_arg.ret;
}
@ -2656,25 +2685,20 @@ static int tcp_sock_init_ns(void *arg)
ns_enter(c->pasta_pid);
if (c->v4) {
tref.v6 = 0;
for (port = 0; port < USHRT_MAX; port++) {
if (!bitmap_isset(c->tcp.port4_to_init, port))
if (!bitmap_isset(c->tcp.port_to_init, port))
continue;
tref.index = port;
tref.index = (in_port_t)(port + tcp_port_delta_to_init[port]);
if (c->v4) {
tref.v6 = 0;
sock_l4(c, AF_INET, IPPROTO_TCP, port, BIND_LOOPBACK,
tref.u32);
}
}
if (c->v6) {
tref.v6 = 1;
for (port = 0; port < USHRT_MAX; port++) {
if (!bitmap_isset(c->tcp.port6_to_init, port))
continue;
tref.index = port;
sock_l4(c, AF_INET6, IPPROTO_TCP, port, BIND_LOOPBACK,
tref.u32);
}
@ -2692,65 +2716,54 @@ static int tcp_sock_init_ns(void *arg)
int tcp_sock_init(struct ctx *c)
{
union tcp_epoll_ref tref = { .listen = 1 };
char ns_fn_stack[NS_FN_STACK_SIZE];
enum bind_type tap_bind;
in_port_t port;
getrandom(&c->tcp.hash_secret, sizeof(c->tcp.hash_secret), GRND_RANDOM);
for (port = 0; port < USHRT_MAX; port++) {
if (!bitmap_isset(c->tcp.port_to_tap, port))
continue;
tref.index = (in_port_t)(port + tcp_port_delta_to_tap[port]);
if (c->v4) {
tref.v6 = 0;
for (port = 0; port < USHRT_MAX; port++) {
tref.index = port;
if (bitmap_isset(c->tcp.port4_to_ns, port)) {
tref.splice = 0;
sock_l4(c, AF_INET, IPPROTO_TCP, port,
c->mode == MODE_PASTA ? BIND_EXT : BIND_ANY,
tref.u32);
if (c->mode == MODE_PASTA) {
tref.splice = 1;
sock_l4(c, AF_INET, IPPROTO_TCP, port,
BIND_LOOPBACK, tref.u32);
tap_bind = BIND_EXT;
} else {
tap_bind = BIND_ANY;
}
if (bitmap_isset(c->tcp.port4_to_tap, port)) {
tref.splice = 0;
sock_l4(c, AF_INET, IPPROTO_TCP, port,
tap_bind, tref.u32);
}
}
tcp_sock4_iov_init();
}
if (c->v6) {
tref.v6 = 1;
for (port = 0; port < USHRT_MAX; port++) {
tref.index = port;
if (bitmap_isset(c->tcp.port6_to_ns, port)) {
tref.splice = 0;
sock_l4(c, AF_INET6, IPPROTO_TCP, port,
c->mode == MODE_PASTA ? BIND_EXT : BIND_ANY,
tref.u32);
if (c->mode == MODE_PASTA) {
tref.splice = 1;
sock_l4(c, AF_INET6, IPPROTO_TCP, port,
BIND_LOOPBACK, tref.u32);
tap_bind = BIND_EXT;
} else {
tap_bind = BIND_ANY;
}
if (bitmap_isset(c->tcp.port6_to_tap, port)) {
tref.splice = 0;
sock_l4(c, AF_INET6, IPPROTO_TCP, port,
tap_bind, tref.u32);
}
}
if (c->v4)
tcp_sock4_iov_init();
if (c->v6)
tcp_sock6_iov_init();
}
if (c->mode == MODE_PASTA) {
clone(tcp_sock_init_ns, ns_fn_stack + sizeof(ns_fn_stack) / 2,
CLONE_VM | CLONE_VFORK | CLONE_FILES | SIGCHLD,
(void *)c);
}
if (c->mode == MODE_PASTA)
NS_CALL(tcp_sock_init_ns, c);
return 0;
}

18
tcp.h
View file

@ -16,6 +16,8 @@ int tcp_sock_init(struct ctx *c);
void tcp_timer(struct ctx *c, struct timespec *ts);
void tcp_update_l2_buf(unsigned char *eth_d, unsigned char *eth_s,
uint32_t *ip_da);
void tcp_remap_to_tap(in_port_t port, in_port_t delta);
void tcp_remap_to_init(in_port_t port, in_port_t delta);
/**
* union tcp_epoll_ref - epoll reference portion for TCP connections
@ -40,24 +42,16 @@ union tcp_epoll_ref {
* @hash_secret: 128-bit secret for hash functions, ISN and hash table
* @tap_conn_count: Count of tap connections in connection table
* @splice_conn_count: Count of spliced connections in connection table
* @port4_to_tap: IPv4 ports bound host/init-side, packets to guest/tap
* @port6_to_tap: IPv6 ports bound host/init-side, packets to guest/tap
* @port4_to_init: IPv4 ports bound namespace-side, spliced to init
* @port6_to_init: IPv6 ports bound namespace-side, spliced to init
* @port4_to_ns: IPv4 ports bound init-side, spliced to namespace
* @port6_to_ns: IPv6 ports bound init-side, spliced to namespace
* @port_to_tap: Ports bound host-side, packets to tap or spliced
* @port_to_init: Ports bound namespace-side, spliced to init
* @timer_run: Timestamp of most recent timer run
*/
struct tcp_ctx {
uint64_t hash_secret[2];
int tap_conn_count;
int splice_conn_count;
uint8_t port4_to_tap [USHRT_MAX / 8];
uint8_t port6_to_tap [USHRT_MAX / 8];
uint8_t port4_to_init [USHRT_MAX / 8];
uint8_t port6_to_init [USHRT_MAX / 8];
uint8_t port4_to_ns [USHRT_MAX / 8];
uint8_t port6_to_ns [USHRT_MAX / 8];
uint8_t port_to_tap [USHRT_MAX / 8];
uint8_t port_to_init [USHRT_MAX / 8];
struct timespec timer_run;
};

168
udp.c
View file

@ -51,7 +51,8 @@
* - send packet to udp4_splice_map[5000].ns_conn_sock
* - otherwise:
* - create new socket udp_splice_map[V4][5000].ns_conn_sock
* - connect in namespace to 127.0.0.1:80
* - connect in namespace to 127.0.0.1:80 (note: this destination port
* might be remapped to another port instead)
* - get source port of new connected socket (10000) with getsockname()
* - add to epoll with reference: index = 10000, splice: UDP_BACK_TO_INIT
* - set udp_splice_map[V4][10000].init_bound_sock to s
@ -74,7 +75,8 @@
* - send packet to udp4_splice_map[2000].init_conn_sock
* - otherwise:
* - create new socket udp_splice_map[V4][2000].init_conn_sock
* - connect in init to 127.0.0.1:22,
* - connect in init to 127.0.0.1:22 (note: this destination port
* might be remapped to another port instead)
* - get source port of new connected socket (4000) with getsockname()
* - add to epoll with reference: index = 4000, splice = UDP_BACK_TO_NS
* - set udp_splice_map[V4][4000].ns_bound_sock to s
@ -163,6 +165,12 @@ struct udp_splice_port {
static struct udp_tap_port udp_tap_map [IP_VERSIONS][USHRT_MAX];
static struct udp_splice_port udp_splice_map [IP_VERSIONS][USHRT_MAX];
/* Port re-mappings as delta, indexed by original destination port */
static in_port_t udp_port_delta_to_tap [USHRT_MAX];
static in_port_t udp_port_delta_from_tap [USHRT_MAX];
static in_port_t udp_port_delta_to_init [USHRT_MAX];
static in_port_t udp_port_delta_from_init[USHRT_MAX];
enum udp_act_type {
UDP_ACT_TAP,
UDP_ACT_NS_CONN,
@ -267,6 +275,28 @@ static struct mmsghdr udp_splice_mmh_send [UDP_SPLICE_FRAMES];
static struct iovec udp_splice_iov_sendto [UDP_SPLICE_FRAMES];
static struct mmsghdr udp_splice_mmh_sendto [UDP_SPLICE_FRAMES];
/**
* udp_remap_to_tap() - Set delta for port translation to/from guest/tap
* @port: Original destination port, host order
* @delta: Delta to be added to original destination port
*/
void udp_remap_to_tap(in_port_t port, in_port_t delta)
{
udp_port_delta_to_tap[port] = delta;
udp_port_delta_from_tap[port + delta] = USHRT_MAX - delta;
}
/**
* udp_remap_to_init() - Set delta for port translation to/from init namespace
* @port: Original destination port, host order
* @delta: Delta to be added to original destination port
*/
void udp_remap_to_init(in_port_t port, in_port_t delta)
{
udp_port_delta_to_init[port] = delta;
udp_port_delta_from_init[port + delta] = USHRT_MAX - delta;
}
/**
* udp_update_check4() - Update checksum with variable parts from stored one
* @buf: L2 packet buffer with final IPv4 header
@ -506,7 +536,6 @@ static void udp_sock_handler_splice(struct ctx *c, union epoll_ref ref,
struct msghdr *mh = &udp_splice_mmh_recv[0].msg_hdr;
struct sockaddr_storage *sa_s = mh->msg_name;
in_port_t src, dst = ref.udp.port, send_dst;
char ns_fn_stack[NS_FN_STACK_SIZE];
int s, v6 = ref.udp.v6, n, i;
if (!(events & EPOLLIN))
@ -529,16 +558,14 @@ static void udp_sock_handler_splice(struct ctx *c, union epoll_ref ref,
switch (ref.udp.splice) {
case UDP_TO_NS:
src += udp_port_delta_from_init[src];
if (!(s = udp_splice_map[v6][src].ns_conn_sock)) {
struct udp_splice_connect_ns_arg arg = {
c, v6, ref.s, src, dst, -1,
};
clone(udp_splice_connect_ns,
ns_fn_stack + sizeof(ns_fn_stack) / 2,
CLONE_VM | CLONE_VFORK | CLONE_FILES | SIGCHLD,
(void *)&arg);
NS_CALL(udp_splice_connect_ns, &arg);
if ((s = arg.s) < 0)
return;
}
@ -551,6 +578,8 @@ static void udp_sock_handler_splice(struct ctx *c, union epoll_ref ref,
send_dst = udp_splice_map[v6][dst].init_dst_port;
break;
case UDP_TO_INIT:
src += udp_port_delta_from_tap[src];
if (!(s = udp_splice_map[v6][src].init_conn_sock)) {
s = udp_splice_connect(c, v6, ref.s, src, dst,
UDP_BACK_TO_NS);
@ -867,16 +896,28 @@ int udp_tap_handler(struct ctx *c, int af, void *addr,
.sin6_port = uh->dest,
.sin6_addr = *(struct in6_addr *)addr,
};
enum bind_type bind_to = BIND_ANY;
sa = (struct sockaddr *)&s_in6;
sl = sizeof(s_in6);
if (!memcmp(addr, &c->gw6, sizeof(c->gw6)) &&
udp_tap_map[V6][dst].ts_local) {
if (udp_tap_map[V6][dst].loopback)
s_in6.sin6_addr = in6addr_loopback;
else
s_in6.sin6_addr = c->addr6_seen;
} else if (IN6_IS_ADDR_LINKLOCAL(&s_in6.sin6_addr)) {
bind_to = BIND_LL;
}
if (!(s = udp_tap_map[V6][src].sock)) {
union udp_epoll_ref uref = { .bound = 1, .v6 = 1,
.port = src
};
s = sock_l4(c, AF_INET6, IPPROTO_UDP, src, 0, uref.u32);
s = sock_l4(c, AF_INET6, IPPROTO_UDP, src, bind_to,
uref.u32);
if (s <= 0)
return count;
@ -885,14 +926,6 @@ int udp_tap_handler(struct ctx *c, int af, void *addr,
}
udp_tap_map[V6][src].ts = now->tv_sec;
if (!memcmp(addr, &c->gw6, sizeof(c->gw6)) &&
udp_tap_map[V6][dst].ts_local) {
if (udp_tap_map[V6][dst].loopback)
s_in6.sin6_addr = in6addr_loopback;
else
s_in6.sin6_addr = c->addr6_seen;
}
}
for (i = 0; i < count; i++) {
@ -923,30 +956,25 @@ int udp_sock_init_ns(void *arg)
{
union udp_epoll_ref uref = { .bound = 1, .splice = UDP_TO_INIT };
struct ctx *c = (struct ctx *)arg;
in_port_t port;
in_port_t dst;
ns_enter(c->pasta_pid);
if (c->v4) {
uref.v6 = 0;
for (port = 0; port < USHRT_MAX; port++) {
if (!bitmap_isset(c->udp.port4_to_init, port))
for (dst = 0; dst < USHRT_MAX; dst++) {
if (!bitmap_isset(c->udp.port_to_init, dst))
continue;
uref.port = port;
sock_l4(c, AF_INET, IPPROTO_UDP, port, BIND_LOOPBACK,
uref.port = dst + udp_port_delta_to_init[dst];
if (c->v4) {
uref.v6 = 0;
sock_l4(c, AF_INET, IPPROTO_UDP, dst, BIND_LOOPBACK,
uref.u32);
}
}
if (c->v6) {
uref.v6 = 1;
for (port = 0; port < USHRT_MAX; port++) {
if (!bitmap_isset(c->udp.port6_to_init, port))
continue;
uref.port = port;
sock_l4(c, AF_INET6, IPPROTO_UDP, port, BIND_LOOPBACK,
sock_l4(c, AF_INET6, IPPROTO_UDP, dst, BIND_LOOPBACK,
uref.u32);
}
}
@ -1016,68 +1044,56 @@ static void udp_splice_iov_init(void)
int udp_sock_init(struct ctx *c)
{
union udp_epoll_ref uref = { .bound = 1 };
char ns_fn_stack[NS_FN_STACK_SIZE];
enum bind_type tap_bind;
in_port_t port;
in_port_t dst;
int s;
for (dst = 0; dst < USHRT_MAX; dst++) {
if (!bitmap_isset(c->udp.port_to_tap, dst))
continue;
uref.port = dst + udp_port_delta_to_tap[dst];
if (c->v4) {
uref.splice = 0;
uref.v6 = 0;
for (port = 0; port < USHRT_MAX; port++) {
uref.port = port;
if (bitmap_isset(c->udp.port4_to_ns, port)) {
uref.splice = UDP_TO_NS;
sock_l4(c, AF_INET, IPPROTO_UDP, port,
BIND_LOOPBACK, uref.u32);
tap_bind = BIND_EXT;
} else {
tap_bind = BIND_ANY;
}
if (bitmap_isset(c->udp.port4_to_tap, port)) {
uref.splice = 0;
s = sock_l4(c, AF_INET, IPPROTO_UDP, port,
tap_bind, uref.u32);
s = sock_l4(c, AF_INET, IPPROTO_UDP, dst,
c->mode == MODE_PASTA ? BIND_EXT : BIND_ANY,
uref.u32);
if (s > 0)
udp_tap_map[V4][port].sock = s;
}
}
udp_tap_map[V4][uref.port].sock = s;
udp_sock4_iov_init();
if (c->mode == MODE_PASTA) {
uref.splice = UDP_TO_NS;
sock_l4(c, AF_INET, IPPROTO_UDP, dst,
BIND_LOOPBACK, uref.u32);
}
}
if (c->v6) {
uref.v6 = 1;
for (port = 0; port < USHRT_MAX; port++) {
uref.port = port;
if (bitmap_isset(c->udp.port6_to_ns, port)) {
uref.splice = UDP_TO_NS;
sock_l4(c, AF_INET6, IPPROTO_UDP, port,
BIND_LOOPBACK, uref.u32);
tap_bind = BIND_EXT;
} else {
tap_bind = BIND_ANY;
}
if (bitmap_isset(c->udp.port6_to_tap, port)) {
uref.splice = 0;
s = sock_l4(c, AF_INET6, IPPROTO_UDP, port,
tap_bind, uref.u32);
uref.v6 = 1;
s = sock_l4(c, AF_INET6, IPPROTO_UDP, dst,
c->mode == MODE_PASTA ? BIND_EXT : BIND_ANY,
uref.u32);
if (s > 0)
udp_tap_map[V6][port].sock = s;
udp_tap_map[V6][uref.port].sock = s;
if (c->mode == MODE_PASTA) {
uref.splice = UDP_TO_NS;
sock_l4(c, AF_INET6, IPPROTO_UDP, dst,
BIND_LOOPBACK, uref.u32);
}
}
}
if (c->v4)
udp_sock4_iov_init();
if (c->v6)
udp_sock6_iov_init();
}
if (c->mode == MODE_PASTA) {
udp_splice_iov_init();
clone(udp_sock_init_ns, ns_fn_stack + sizeof(ns_fn_stack) / 2,
CLONE_VM | CLONE_VFORK | CLONE_FILES | SIGCHLD,
(void *)c);
NS_CALL(udp_sock_init_ns, c);
}
return 0;

17
udp.h
View file

@ -11,6 +11,8 @@ int udp_sock_init(struct ctx *c);
void udp_timer(struct ctx *c, struct timespec *ts);
void udp_update_l2_buf(unsigned char *eth_d, unsigned char *eth_s,
uint32_t *ip_da);
void udp_remap_to_tap(in_port_t port, in_port_t delta);
void udp_remap_to_init(in_port_t port, in_port_t delta);
/**
* union udp_epoll_ref - epoll reference portion for TCP connections
@ -37,20 +39,13 @@ union udp_epoll_ref {
/**
* struct udp_ctx - Execution context for UDP
* @port6_to_tap: IPv6 ports bound host/init-side, packets to guest/tap
* @port4_to_init: IPv4 ports bound namespace-side, spliced to init
* @port6_to_init: IPv6 ports bound namespace-side, spliced to init
* @port4_to_ns: IPv4 ports bound init-side, spliced to namespace
* @port6_to_ns: IPv6 ports bound init-side, spliced to namespace
* @port_to_tap: Ports bound host-side, data to tap or ns L4 socket
* @port_to_init: Ports bound namespace-side, data to init L4 socket
* @timer_run: Timestamp of most recent timer run
*/
struct udp_ctx {
uint8_t port4_to_tap [USHRT_MAX / 8];
uint8_t port6_to_tap [USHRT_MAX / 8];
uint8_t port4_to_init [USHRT_MAX / 8];
uint8_t port6_to_init [USHRT_MAX / 8];
uint8_t port4_to_ns [USHRT_MAX / 8];
uint8_t port6_to_ns [USHRT_MAX / 8];
uint8_t port_to_tap [USHRT_MAX / 8];
uint8_t port_to_init [USHRT_MAX / 8];
struct timespec timer_run;
};

33
util.c
View file

@ -36,7 +36,6 @@
#include "util.h"
#include "passt.h"
#ifdef DEBUG
#define logfn(name, level) \
void name(const char *format, ...) { \
char ts[sizeof("Mmm dd hh:mm:ss.")]; \
@ -44,37 +43,33 @@ void name(const char *format, ...) { \
struct tm *tm; \
va_list args; \
\
if (setlogmask(0) & LOG_MASK(LOG_DEBUG)) { \
clock_gettime(CLOCK_REALTIME, &tp); \
tm = gmtime(&tp.tv_sec); \
strftime(ts, sizeof(ts), "%b %d %T.", tm); \
\
fprintf(stderr, "%s%04lu: ", ts, tp.tv_nsec / (100 * 1000)); \
fprintf(stderr, "%s%04lu: ", ts, \
tp.tv_nsec / (100 * 1000)); \
} \
\
va_start(args, format); \
vsyslog(level, format, args); \
va_end(args); \
\
if (setlogmask(0) & LOG_MASK(LOG_DEBUG) || \
setlogmask(0) == LOG_MASK(LOG_EMERG)) { \
va_start(args, format); \
vfprintf(stderr, format, args); \
va_end(args); \
if (format[strlen(format)] != '\n') \
fprintf(stderr, "\n"); \
} \
}
#else
#define logfn(name, level) \
void name(const char *format, ...) { \
va_list args; \
\
va_start(args, format); \
vsyslog(level, format, args); \
va_end(args); \
}
#endif
logfn(err, LOG_ERR)
logfn(warn, LOG_WARNING)
logfn(info, LOG_INFO)
#ifdef DEBUG
logfn(debug, LOG_DEBUG)
#endif
/**
* ipv6_l4hdr() - Find pointer to L4 header in IPv6 packet and extract protocol
@ -171,12 +166,16 @@ int sock_l4(struct ctx *c, int af, uint8_t proto, uint16_t port,
sa = (const struct sockaddr *)&addr4;
sl = sizeof(addr4);
} else {
if (bind_addr == BIND_LOOPBACK)
if (bind_addr == BIND_LOOPBACK) {
addr6.sin6_addr = in6addr_loopback;
else if (bind_addr == BIND_EXT)
} else if (bind_addr == BIND_EXT) {
addr6.sin6_addr = c->addr6;
else
} else if (bind_addr == BIND_LL) {
addr6.sin6_addr = c->addr6_ll;
addr6.sin6_scope_id = if_nametoindex(c->ifn);
} else {
addr6.sin6_addr = in6addr_any;
}
sa = (const struct sockaddr *)&addr6;
sl = sizeof(addr6);

14
util.h
View file

@ -1,12 +1,7 @@
void err(const char *format, ...);
void warn(const char *format, ...);
void info(const char *format, ...);
#ifdef DEBUG
void debug(const char *format, ...);
#else
#define debug(...) { }
#endif
#define CHECK_SET_MIN_MAX(basename, fd) \
do { \
@ -53,6 +48,14 @@ void debug(const char *format, ...);
#define PORT_IS_EPHEMERAL(port) ((port) >= PORT_EPHEMERAL_MIN)
#define NS_FN_STACK_SIZE (RLIMIT_STACK_VAL * 1024 / 4)
#define NS_CALL(fn, arg) \
do { \
char ns_fn_stack[NS_FN_STACK_SIZE]; \
\
clone((fn), ns_fn_stack + sizeof(ns_fn_stack) / 2, \
CLONE_VM | CLONE_VFORK | CLONE_FILES | SIGCHLD, \
(void *)(arg)); \
} while (0)
#if __BYTE_ORDER == __BIG_ENDIAN
#define L2_BUF_ETH_IP4_INIT \
@ -120,6 +123,7 @@ void debug(const char *format, ...);
enum bind_type {
BIND_ANY = 0,
BIND_LOOPBACK,
BIND_LL,
BIND_EXT,
};