ZeroTierOne/osdep/MacEthernetTapAgent.c
2019-01-14 10:25:53 -08:00

435 lines
12 KiB
C

/*
* ZeroTier One - Network Virtualization Everywhere
* Copyright (C) 2011-2019 ZeroTier, Inc. https://www.zerotier.com/
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*
* --
*
* You can be released from the requirements of the license by purchasing
* a commercial license. Buying such a license is mandatory as soon as you
* develop commercial closed-source software that incorporates or links
* directly against ZeroTier software without disclosing the source code
* of your own application.
*/
/*
* This creates a pair of feth devices with the lower numbered device
* being the ZeroTier virtual interface and the other being the device
* used to actually read and write packets. The latter gets no IP config
* and is only used for I/O. The behavior of feth is similar to the
* veth pairs that exist on Linux.
*
* The feth device has only existed since MacOS Sierra, but that's fairly
* long ago in Mac terms.
*
* I/O with feth must be done using two different sockets. The BPF socket
* is used to receive packets, while an AF_NDRV (low-level network driver
* access) socket must be used to inject. AF_NDRV can't read IP frames
* since BSD doesn't forward packets out the NDRV tap if they've already
* been handled, and while BPF can inject its MTU for injected packets
* is limited to 2048. AF_NDRV packet injection is required to inject
* ZeroTier's large MTU frames.
*
* Benchmarks show that this performs similarly to the old tap.kext driver,
* and a kext is no longer required. Splitting it off into an agent will
* also make it easier to have zerotier-one itself drop permissions.
*
* All this stuff is basically undocumented. A lot of tracing through
* the Darwin/XNU kernel source was required to figure out how to make
* this actually work.
*
* See also:
*
* https://apple.stackexchange.com/questions/337715/fake-ethernet-interfaces-feth-if-fake-anyone-ever-seen-this
* https://opensource.apple.com/source/xnu/xnu-4570.41.2/bsd/net/if_fake.c.auto.html
*
*/
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <stdarg.h>
#include <unistd.h>
#include <signal.h>
#include <fcntl.h>
#include <errno.h>
#include <sys/signal.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <sys/ioctl.h>
#include <sys/wait.h>
#include <sys/select.h>
#include <sys/cdefs.h>
#include <sys/uio.h>
#include <sys/param.h>
#include <sys/ioctl.h>
#include <sys/socket.h>
#include <sys/sysctl.h>
#include <netinet/in.h>
#include <arpa/inet.h>
#include <net/bpf.h>
#include <net/route.h>
#include <net/if.h>
#include <net/if_arp.h>
#include <net/if_dl.h>
#include <net/if_media.h>
#include <net/ndrv.h>
#include <netinet/in_var.h>
#include <netinet/icmp6.h>
#include <netinet6/in6_var.h>
#include <netinet6/nd6.h>
#include <ifaddrs.h>
#include "../version.h"
#include "MacEthernetTapAgent.h"
#ifndef SIOCAUTOCONF_START
#define SIOCAUTOCONF_START _IOWR('i', 132, struct in6_ifreq) /* accept rtadvd on this interface */
#endif
#ifndef SIOCAUTOCONF_STOP
#define SIOCAUTOCONF_STOP _IOWR('i', 133, struct in6_ifreq) /* stop accepting rtadv for this interface */
#endif
#define P_IFCONFIG "/sbin/ifconfig"
static unsigned char s_pktReadBuf[262144] __attribute__ ((__aligned__(16)));
static unsigned char s_stdinReadBuf[262144] __attribute__ ((__aligned__(16)));
static char s_deviceName[IFNAMSIZ];
static char s_peerDeviceName[IFNAMSIZ];
static int s_bpffd = -1;
static int s_ndrvfd = -1;
static pid_t s_parentPid;
static void configureIpv6Parameters(const char *ifname,int performNUD,int acceptRouterAdverts)
{
struct in6_ndireq nd;
struct in6_ifreq ifr;
int s = socket(AF_INET6,SOCK_DGRAM,0);
if (s <= 0)
return;
memset(&nd,0,sizeof(nd));
strncpy(nd.ifname,ifname,sizeof(nd.ifname));
if (ioctl(s,SIOCGIFINFO_IN6,&nd)) {
close(s);
return;
}
unsigned long oldFlags = (unsigned long)nd.ndi.flags;
if (performNUD)
nd.ndi.flags |= ND6_IFF_PERFORMNUD;
else nd.ndi.flags &= ~ND6_IFF_PERFORMNUD;
if (oldFlags != (unsigned long)nd.ndi.flags) {
if (ioctl(s,SIOCSIFINFO_FLAGS,&nd)) {
close(s);
return;
}
}
memset(&ifr,0,sizeof(ifr));
strncpy(ifr.ifr_name,ifname,sizeof(ifr.ifr_name));
if (ioctl(s,acceptRouterAdverts ? SIOCAUTOCONF_START : SIOCAUTOCONF_STOP,&ifr)) {
close(s);
return;
}
close(s);
}
static int run(const char *path,...)
{
va_list ap;
char *args[16];
int argNo = 1;
va_start(ap,path);
args[0] = (char *)path;
for(;argNo<15;++argNo) {
args[argNo] = va_arg(ap,char *);
if (!args[argNo]) {
break;
}
}
args[argNo++] = (char *)0;
va_end(ap);
pid_t pid = vfork();
if (pid < 0) {
return -1;
} else if (pid == 0) {
dup2(STDERR_FILENO,STDOUT_FILENO);
execv(args[0],args);
_exit(-1);
}
int rv = 0;
waitpid(pid,&rv,0);
return rv;
}
static void die()
{
if (s_ndrvfd >= 0)
close(s_ndrvfd);
if (s_bpffd >= 0)
close(s_bpffd);
if (s_deviceName[0])
run("/sbin/ifconfig",s_deviceName,"destroy",(char *)0);
if (s_peerDeviceName[0])
run("/sbin/ifconfig",s_peerDeviceName,"destroy",(char *)0);
}
int main(int argc,char **argv)
{
char buf[128];
struct ifreq ifr;
u_int fl;
fd_set rfds,wfds,efds;
struct iovec iov[2];
s_deviceName[0] = 0;
s_peerDeviceName[0] = 0;
s_parentPid = getppid();
atexit(&die);
signal(SIGIO,SIG_IGN);
signal(SIGCHLD,SIG_IGN);
signal(SIGPIPE,SIG_IGN);
signal(SIGUSR1,SIG_IGN);
signal(SIGUSR2,SIG_IGN);
signal(SIGALRM,SIG_IGN);
signal(SIGQUIT,&exit);
signal(SIGTERM,&exit);
signal(SIGKILL,&exit);
signal(SIGINT,&exit);
signal(SIGPIPE,&exit);
if (getuid() != 0) {
if (setuid(0) != 0) {
fprintf(stderr,"E must be run as root or with root setuid bit on executable\n");
return ZT_MACETHERNETTAPAGENT_EXIT_CODE_INVALID_REQUEST;
}
}
if (argc < 5) {
fprintf(stderr,"E invalid or missing argument(s) (usage: MacEthernetTapAgent <0-4999> <mac> <mtu> <metric>)\n");
return ZT_MACETHERNETTAPAGENT_EXIT_CODE_INVALID_REQUEST;
}
const int deviceNo = atoi(argv[1]);
if ((deviceNo < 0)||(deviceNo > 4999)) {
fprintf(stderr,"E invalid or missing argument(s) (usage: MacEthernetTapAgent <0-4999> <mac> <mtu> <metric>)\n");
return ZT_MACETHERNETTAPAGENT_EXIT_CODE_INVALID_REQUEST;
}
const char *mac = argv[2];
const char *mtu = argv[3];
const char *metric = argv[4];
s_ndrvfd = socket(AF_NDRV,SOCK_RAW,0);
if (s_ndrvfd < 0) {
fprintf(stderr,"E unable to open AF_NDRV socket\n");
return ZT_MACETHERNETTAPAGENT_EXIT_CODE_UNABLE_TO_CREATE;
}
snprintf(s_peerDeviceName,sizeof(s_peerDeviceName),"feth%d",deviceNo+5000);
if (run(P_IFCONFIG,s_peerDeviceName,"create",(char *)0) != 0) {
fprintf(stderr,"E unable to create %s\n",s_deviceName);
return ZT_MACETHERNETTAPAGENT_EXIT_CODE_UNABLE_TO_CREATE;
}
usleep(10);
snprintf(s_deviceName,sizeof(s_deviceName),"feth%d",deviceNo);
if (run(P_IFCONFIG,s_deviceName,"create",(char *)0) != 0) {
fprintf(stderr,"E unable to create %s\n",s_deviceName);
return ZT_MACETHERNETTAPAGENT_EXIT_CODE_UNABLE_TO_CREATE;
}
run(P_IFCONFIG,s_deviceName,"lladdr",mac,(char *)0);
usleep(10);
run(P_IFCONFIG,s_peerDeviceName,"peer",s_deviceName,(char *)0);
usleep(10);
run(P_IFCONFIG,s_peerDeviceName,"mtu","10000","up",(char *)0);
usleep(10);
run(P_IFCONFIG,s_deviceName,"mtu",mtu,"metric",metric,"up",(char *)0);
usleep(10);
configureIpv6Parameters(s_deviceName,1,0);
usleep(10);
struct sockaddr_ndrv nd;
nd.snd_len = sizeof(struct sockaddr_ndrv);
nd.snd_family = AF_NDRV;
memcpy(nd.snd_name,s_peerDeviceName,sizeof(nd.snd_name));
if (bind(s_ndrvfd,(struct sockaddr *)&nd,sizeof(nd)) != 0) {
fprintf(stderr,"E unable to bind AF_NDRV socket\n");
return ZT_MACETHERNETTAPAGENT_EXIT_CODE_UNABLE_TO_CREATE;
}
if (connect(s_ndrvfd,(struct sockaddr *)&nd,sizeof(nd)) != 0) {
fprintf(stderr,"E unable to connect AF_NDRV socket\n");
return ZT_MACETHERNETTAPAGENT_EXIT_CODE_UNABLE_TO_CREATE;
}
/* Start at /dev/bpf1 since some simple bpf-using net utilities hard-code /dev/bpf0.
* Things like libpcap are smart enough to search. */
for(int bpfno=1;bpfno<5000;++bpfno) {
char tmp[32];
snprintf(tmp,sizeof(tmp),"/dev/bpf%d",bpfno);
s_bpffd = open(tmp,O_RDWR);
if (s_bpffd >= 0) {
break;
}
}
if (s_bpffd < 0) {
fprintf(stderr,"E unable to open bpf device\n");
return ZT_MACETHERNETTAPAGENT_EXIT_CODE_UNABLE_TO_CREATE;
}
fl = sizeof(s_pktReadBuf);
if (ioctl(s_bpffd,BIOCSBLEN,&fl) != 0) {
return ZT_MACETHERNETTAPAGENT_EXIT_CODE_UNABLE_TO_CREATE;
}
const size_t readPktSize = (size_t)fl;
fl = 1;
if (ioctl(s_bpffd,BIOCIMMEDIATE,&fl) != 0) {
return ZT_MACETHERNETTAPAGENT_EXIT_CODE_UNABLE_TO_CREATE;
}
fl = 0;
if (ioctl(s_bpffd,BIOCSSEESENT,&fl) != 0) {
return ZT_MACETHERNETTAPAGENT_EXIT_CODE_UNABLE_TO_CREATE;
}
memset(&ifr,0,sizeof(ifr));
memcpy(ifr.ifr_name,s_peerDeviceName,IFNAMSIZ);
if (ioctl(s_bpffd,BIOCSETIF,&ifr) != 0) {
return ZT_MACETHERNETTAPAGENT_EXIT_CODE_UNABLE_TO_CREATE;
}
fl = 1;
if (ioctl(s_bpffd,BIOCSHDRCMPLT,&fl) != 0) {
return ZT_MACETHERNETTAPAGENT_EXIT_CODE_UNABLE_TO_CREATE;
}
fl = 1;
if (ioctl(s_bpffd,BIOCPROMISC,&fl) != 0) {
return ZT_MACETHERNETTAPAGENT_EXIT_CODE_UNABLE_TO_CREATE;
}
fprintf(stderr,"I %s %s %d.%d.%d.%d\n",s_deviceName,s_peerDeviceName,ZEROTIER_ONE_VERSION_MAJOR,ZEROTIER_ONE_VERSION_MINOR,ZEROTIER_ONE_VERSION_REVISION,ZEROTIER_ONE_VERSION_BUILD);
FD_ZERO(&rfds);
FD_ZERO(&wfds);
FD_ZERO(&efds);
long stdinReadPtr = 0;
for(;;) {
FD_SET(STDIN_FILENO,&rfds);
FD_SET(s_bpffd,&rfds);
if (select(s_bpffd+1,&rfds,&wfds,&efds,(struct timeval *)0) < 0) {
if ((errno == EAGAIN)||(errno == EINTR)) {
usleep(10);
continue;
}
return ZT_MACETHERNETTAPAGENT_EXIT_CODE_READ_ERROR;
}
if (FD_ISSET(s_bpffd,&rfds)) {
long n = (long)read(s_bpffd,s_pktReadBuf,readPktSize);
if (n > 0) {
for(unsigned char *p=s_pktReadBuf,*eof=p+n;p<eof;) {
struct bpf_hdr *h = (struct bpf_hdr *)p;
if ((h->bh_caplen > 0)&&((p + h->bh_hdrlen + h->bh_caplen) <= eof)) {
uint16_t len = (uint16_t)h->bh_caplen;
iov[0].iov_base = &len;
iov[0].iov_len = 2;
iov[1].iov_base = p + h->bh_hdrlen;
iov[1].iov_len = h->bh_caplen;
writev(STDOUT_FILENO,iov,2);
}
p += BPF_WORDALIGN(h->bh_hdrlen + h->bh_caplen);
}
}
}
if (FD_ISSET(STDIN_FILENO,&rfds)) {
long n = (long)read(STDIN_FILENO,s_stdinReadBuf + stdinReadPtr,sizeof(s_stdinReadBuf) - stdinReadPtr);
if (n > 0) {
stdinReadPtr += n;
while (stdinReadPtr >= 2) {
long len = *((uint16_t *)s_stdinReadBuf);
if (stdinReadPtr >= (len + 2)) {
if (len > 0) {
unsigned char *msg = s_stdinReadBuf + 2;
switch(msg[0]) {
case ZT_MACETHERNETTAPAGENT_STDIN_CMD_PACKET:
if (len > 1) {
if (write(s_ndrvfd,msg+1,len-1) < 0) {
fprintf(stderr,"E inject failed size==%ld errno==%d\n",len-1,errno);
}
}
break;
case ZT_MACETHERNETTAPAGENT_STDIN_CMD_IFCONFIG: {
char *args[16];
args[0] = P_IFCONFIG;
args[1] = s_deviceName;
int argNo = 2;
for(int argPtr=0,k=1,l=(int)len;k<l;++k) {
if (!msg[k]) {
if (argPtr > 0) {
argPtr = 0;
++argNo;
if (argNo >= 15) {
break;
}
}
} else {
if (argPtr == 0) {
args[argNo] = (char *)(msg + k);
}
argPtr++;
}
}
args[argNo] = (char *)0;
if (argNo > 2) {
pid_t pid = fork();
if (pid < 0) {
return -1;
} else if (pid == 0) {
dup2(STDERR_FILENO,STDOUT_FILENO);
execv(args[0],args);
_exit(-1);
}
int rv = 0;
waitpid(pid,&rv,0);
}
} break;
case ZT_MACETHERNETTAPAGENT_STDIN_CMD_EXIT:
return ZT_MACETHERNETTAPAGENT_EXIT_CODE_SUCCESS;
}
}
if (stdinReadPtr > (len + 2)) {
memmove(s_stdinReadBuf,s_stdinReadBuf + len + 2,stdinReadPtr -= (len + 2));
} else {
stdinReadPtr = 0;
}
} else {
break;
}
}
}
}
}
return ZT_MACETHERNETTAPAGENT_EXIT_CODE_SUCCESS;
}