#ifndef INET
#error "IPDIVERT requires INET."
#endif
#include <sys/param.h>
#include <sys/kernel.h>
#include <sys/malloc.h>
#include <sys/mbuf.h>
#include <sys/socket.h>
#include <sys/domain.h>
#include <sys/protosw.h>
#include <sys/socketvar.h>
#include <sys/sysctl.h>
#include <sys/systm.h>
#include <sys/proc.h>
#include <net/if.h>
#include <net/route.h>
#include <netinet/in.h>
#include <netinet/in_systm.h>
#include <netinet/ip.h>
#include <netinet/in_pcb.h>
#include <netinet/in_var.h>
#include <netinet/ip_var.h>
#include <netinet/ip_fw.h>
#include <netinet/ip_divert.h>
#include <kern/zalloc.h>
#define DIVSNDQ (65536 + 100)
#define DIVRCVQ (65536 + 100)
static struct inpcbhead divcb;
static struct inpcbinfo divcbinfo;
static u_long div_sendspace = DIVSNDQ;
static u_long div_recvspace = DIVRCVQ;
static struct sockaddr_in divsrc = { sizeof(divsrc), AF_INET, };
static int div_output(struct socket *so,
struct mbuf *m, struct sockaddr *addr, struct mbuf *control);
extern int load_ipfw(void);
void
div_init(void)
{
struct inpcbinfo *pcbinfo;
LIST_INIT(&divcb);
divcbinfo.listhead = &divcb;
divcbinfo.hashbase = hashinit(1, M_PCB, &divcbinfo.hashmask);
divcbinfo.porthashbase = hashinit(1, M_PCB, &divcbinfo.porthashmask);
divcbinfo.ipi_zone = (void *) zinit(sizeof(struct inpcb),(maxsockets * sizeof(struct inpcb)),
4096, "divzone");
pcbinfo = &divcbinfo;
pcbinfo->mtx_grp_attr = lck_grp_attr_alloc_init();
pcbinfo->mtx_grp = lck_grp_alloc_init("divcb", pcbinfo->mtx_grp_attr);
pcbinfo->mtx_attr = lck_attr_alloc_init();
lck_attr_setdefault(pcbinfo->mtx_attr);
if ((pcbinfo->mtx = lck_rw_alloc_init(pcbinfo->mtx_grp, pcbinfo->mtx_attr)) == NULL)
return;
if (!IPFW_LOADED) {
load_ipfw();
}
}
void
div_input(struct mbuf *m, __unused int off)
{
ipstat.ips_noproto++;
m_freem(m);
}
void
divert_packet(struct mbuf *m, int incoming, int port, int rule)
{
struct ip *ip;
struct inpcb *inp;
struct socket *sa;
u_int16_t nport;
KASSERT(port != 0, ("%s: port=0", __FUNCTION__));
divsrc.sin_port = rule;
if (m->m_len < sizeof(struct ip) &&
(m = m_pullup(m, sizeof(struct ip))) == 0) {
return;
}
ip = mtod(m, struct ip *);
divsrc.sin_addr.s_addr = 0;
if (incoming) {
struct ifaddr *ifa;
KASSERT((m->m_flags & M_PKTHDR), ("%s: !PKTHDR", __FUNCTION__));
ifnet_lock_shared(m->m_pkthdr.rcvif);
TAILQ_FOREACH(ifa, &m->m_pkthdr.rcvif->if_addrhead, ifa_link) {
if (ifa->ifa_addr == NULL)
continue;
if (ifa->ifa_addr->sa_family != AF_INET)
continue;
divsrc.sin_addr =
((struct sockaddr_in *) ifa->ifa_addr)->sin_addr;
break;
}
ifnet_lock_done(m->m_pkthdr.rcvif);
}
bzero(&divsrc.sin_zero, sizeof(divsrc.sin_zero));
if (m->m_pkthdr.rcvif) {
snprintf(divsrc.sin_zero, sizeof(divsrc.sin_zero),
"%s%d", m->m_pkthdr.rcvif->if_name,
m->m_pkthdr.rcvif->if_unit);
}
sa = NULL;
nport = htons((u_int16_t)port);
lck_rw_lock_shared(divcbinfo.mtx);
LIST_FOREACH(inp, &divcb, inp_list) {
if (inp->inp_lport == nport)
sa = inp->inp_socket;
}
if (sa) {
int error = 0;
socket_lock(sa, 1);
if (sbappendaddr(&sa->so_rcv, (struct sockaddr *)&divsrc,
m, (struct mbuf *)0, &error) != 0)
sorwakeup(sa);
socket_unlock(sa, 1);
} else {
m_freem(m);
ipstat.ips_noproto++;
ipstat.ips_delivered--;
}
lck_rw_done(divcbinfo.mtx);
}
static int
div_output(so, m, addr, control)
struct socket *so;
register struct mbuf *m;
struct sockaddr *addr;
struct mbuf *control;
{
register struct inpcb *const inp = sotoinpcb(so);
register struct ip *const ip = mtod(m, struct ip *);
struct sockaddr_in *sin = (struct sockaddr_in *)addr;
int error = 0;
if (control)
m_freem(control);
if (sin) {
struct m_tag *mtag;
struct divert_tag *dt;
int len = 0;
char *c = sin->sin_zero;
mtag = m_tag_alloc(KERNEL_MODULE_TAG_ID, KERNEL_TAG_TYPE_DIVERT,
sizeof(struct divert_tag), M_NOWAIT);
if (mtag == NULL) {
error = ENOBUFS;
goto cantsend;
}
dt = (struct divert_tag *)(mtag+1);
dt->info = 0;
dt->cookie = sin->sin_port;
m_tag_prepend(m, mtag);
while (*c++ && (len++ < sizeof(sin->sin_zero)));
if ((len > 0) && (len < sizeof(sin->sin_zero)))
m->m_pkthdr.rcvif = ifunit(sin->sin_zero);
}
if (!sin || sin->sin_addr.s_addr == 0) {
if (((ip->ip_hl != (sizeof (*ip) >> 2)) && inp->inp_options) ||
((u_short)ntohs(ip->ip_len) > m->m_pkthdr.len)) {
error = EINVAL;
goto cantsend;
}
NTOHS(ip->ip_len);
NTOHS(ip->ip_off);
ipstat.ips_rawout++;
socket_unlock(so, 0);
error = ip_output(m,
inp->inp_options, &inp->inp_route,
(so->so_options & SO_DONTROUTE) |
IP_ALLOWBROADCAST | IP_RAWOUTPUT,
inp->inp_moptions);
socket_lock(so, 0);
} else {
struct ifaddr *ifa;
if (m->m_pkthdr.rcvif == NULL) {
bzero(sin->sin_zero, sizeof(sin->sin_zero));
sin->sin_port = 0;
if (!(ifa = ifa_ifwithaddr((struct sockaddr *) sin))) {
error = EADDRNOTAVAIL;
goto cantsend;
}
m->m_pkthdr.rcvif = ifa->ifa_ifp;
ifafree(ifa);
}
if ((~IF_HWASSIST_CSUM_FLAGS(m->m_pkthdr.rcvif->if_hwassist) &
m->m_pkthdr.csum_flags) == 0) {
if (m->m_pkthdr.csum_flags & CSUM_DELAY_DATA) {
m->m_pkthdr.csum_flags &= ~CSUM_DELAY_DATA;
}
m->m_pkthdr.csum_flags |=
CSUM_DATA_VALID | CSUM_PSEUDO_HDR |
CSUM_IP_CHECKED | CSUM_IP_VALID;
m->m_pkthdr.csum_data = 0xffff;
}
else if (m->m_pkthdr.csum_flags & CSUM_DELAY_DATA) {
int hlen;
#ifdef _IP_VHL
hlen = IP_VHL_HL(ip->ip_vhl) << 2;
#else
hlen = ip->ip_hl << 2;
#endif
in_delayed_cksum(m);
m->m_pkthdr.csum_flags &= ~CSUM_DELAY_DATA;
ip->ip_sum = in_cksum(m, hlen);
}
proto_inject(PF_INET, m);
}
return error;
cantsend:
m_freem(m);
return error;
}
static int
div_attach(struct socket *so, int proto, struct proc *p)
{
struct inpcb *inp;
int error;
inp = sotoinpcb(so);
if (inp)
panic("div_attach");
if (p && (error = proc_suser(p)) != 0)
return error;
error = soreserve(so, div_sendspace, div_recvspace);
if (error)
return error;
error = in_pcballoc(so, &divcbinfo, p);
if (error)
return error;
inp = (struct inpcb *)so->so_pcb;
inp->inp_ip_p = proto;
inp->inp_vflag |= INP_IPV4;
inp->inp_flags |= INP_HDRINCL;
so->so_state |= SS_ISCONNECTED;
#ifdef MORE_DICVLOCK_DEBUG
printf("div_attach: so=%x sopcb=%x lock=%x ref=%x\n",
so, so->so_pcb, ((struct inpcb *)so->so_pcb)->inpcb_mtx, so->so_usecount);
#endif
return 0;
}
static int
div_detach(struct socket *so)
{
struct inpcb *inp;
#ifdef MORE_DICVLOCK_DEBUG
printf("div_detach: so=%x sopcb=%x lock=%x ref=%x\n",
so, so->so_pcb, ((struct inpcb *)so->so_pcb)->inpcb_mtx, so->so_usecount);
#endif
inp = sotoinpcb(so);
if (inp == 0)
panic("div_detach: so=%x null inp\n", so);
in_pcbdetach(inp);
inp->inp_state = INPCB_STATE_DEAD;
return 0;
}
static int
div_abort(struct socket *so)
{
soisdisconnected(so);
return div_detach(so);
}
static int
div_disconnect(struct socket *so)
{
if ((so->so_state & SS_ISCONNECTED) == 0)
return ENOTCONN;
return div_abort(so);
}
static int
div_bind(struct socket *so, struct sockaddr *nam, struct proc *p)
{
struct inpcb *inp;
int error;
inp = sotoinpcb(so);
if (nam->sa_family != AF_INET) {
error = EAFNOSUPPORT;
} else {
((struct sockaddr_in *)nam)->sin_addr.s_addr = INADDR_ANY;
error = in_pcbbind(inp, nam, p);
}
return error;
}
static int
div_shutdown(struct socket *so)
{
socantsendmore(so);
return 0;
}
static int
div_send(struct socket *so, __unused int flags, struct mbuf *m, struct sockaddr *nam,
struct mbuf *control, __unused struct proc *p)
{
if (m->m_len < sizeof (struct ip) &&
(m = m_pullup(m, sizeof (struct ip))) == 0) {
ipstat.ips_toosmall++;
m_freem(m);
return EINVAL;
}
return div_output(so, m, nam, control);
}
static int
div_pcblist SYSCTL_HANDLER_ARGS
{
int error, i, n;
struct inpcb *inp, **inp_list;
inp_gen_t gencnt;
struct xinpgen xig;
lck_rw_lock_exclusive(divcbinfo.mtx);
if (req->oldptr == USER_ADDR_NULL) {
n = divcbinfo.ipi_count;
req->oldidx = 2 * (sizeof xig)
+ (n + n/8) * sizeof(struct xinpcb);
lck_rw_done(divcbinfo.mtx);
return 0;
}
if (req->newptr != USER_ADDR_NULL) {
lck_rw_done(divcbinfo.mtx);
return EPERM;
}
gencnt = divcbinfo.ipi_gencnt;
n = divcbinfo.ipi_count;
bzero(&xig, sizeof(xig));
xig.xig_len = sizeof xig;
xig.xig_count = n;
xig.xig_gen = gencnt;
xig.xig_sogen = so_gencnt;
error = SYSCTL_OUT(req, &xig, sizeof xig);
if (error) {
lck_rw_done(divcbinfo.mtx);
return error;
}
inp_list = _MALLOC(n * sizeof *inp_list, M_TEMP, M_WAITOK);
if (inp_list == 0) {
lck_rw_done(divcbinfo.mtx);
return ENOMEM;
}
for (inp = LIST_FIRST(divcbinfo.listhead), i = 0; inp && i < n;
inp = LIST_NEXT(inp, inp_list)) {
#ifdef __APPLE__
if (inp->inp_gencnt <= gencnt && inp->inp_state != INPCB_STATE_DEAD)
#else
if (inp->inp_gencnt <= gencnt && !prison_xinpcb(req->p, inp))
#endif
inp_list[i++] = inp;
}
n = i;
error = 0;
for (i = 0; i < n; i++) {
inp = inp_list[i];
if (inp->inp_gencnt <= gencnt && inp->inp_state != INPCB_STATE_DEAD) {
struct xinpcb xi;
bzero(&xi, sizeof(xi));
xi.xi_len = sizeof xi;
inpcb_to_compat(inp, &xi.xi_inp);
if (inp->inp_socket)
sotoxsocket(inp->inp_socket, &xi.xi_socket);
error = SYSCTL_OUT(req, &xi, sizeof xi);
}
}
if (!error) {
bzero(&xig, sizeof(xig));
xig.xig_len = sizeof xig;
xig.xig_gen = divcbinfo.ipi_gencnt;
xig.xig_sogen = so_gencnt;
xig.xig_count = divcbinfo.ipi_count;
error = SYSCTL_OUT(req, &xig, sizeof xig);
}
FREE(inp_list, M_TEMP);
lck_rw_done(divcbinfo.mtx);
return error;
}
__private_extern__ int
div_lock(struct socket *so, int refcount, int lr)
{
int lr_saved;
#ifdef __ppc__
if (lr == 0) {
__asm__ volatile("mflr %0" : "=r" (lr_saved));
}
else lr_saved = lr;
#endif
#ifdef MORE_DICVLOCK_DEBUG
printf("div_lock: so=%x sopcb=%x lock=%x ref=%x lr=%x\n",
so,
so->so_pcb,
so->so_pcb ? ((struct inpcb *)so->so_pcb)->inpcb_mtx : 0,
so->so_usecount,
lr_saved);
#endif
if (so->so_pcb) {
lck_mtx_lock(((struct inpcb *)so->so_pcb)->inpcb_mtx);
} else {
panic("div_lock: so=%x NO PCB! lr=%x\n", so, lr_saved);
lck_mtx_lock(so->so_proto->pr_domain->dom_mtx);
}
if (so->so_usecount < 0)
panic("div_lock: so=%x so_pcb=%x lr=%x ref=%x\n",
so, so->so_pcb, lr_saved, so->so_usecount);
if (refcount)
so->so_usecount++;
so->reserved3 = (void *)lr_saved;
return (0);
}
__private_extern__ int
div_unlock(struct socket *so, int refcount, int lr)
{
int lr_saved;
lck_mtx_t * mutex_held;
struct inpcb *inp = sotoinpcb(so);
#ifdef __ppc__
if (lr == 0) {
__asm__ volatile("mflr %0" : "=r" (lr_saved));
}
else lr_saved = lr;
#endif
#ifdef MORE_DICVLOCK_DEBUG
printf("div_unlock: so=%x sopcb=%x lock=%x ref=%x lr=%x\n",
so,
so->so_pcb,
so->so_pcb ? ((struct inpcb *)so->so_pcb)->inpcb_mtx : 0,
so->so_usecount,
lr_saved);
#endif
if (refcount)
so->so_usecount--;
if (so->so_usecount < 0)
panic("div_unlock: so=%x usecount=%x\n", so, so->so_usecount);
if (so->so_pcb == NULL) {
panic("div_unlock: so=%x NO PCB usecount=%x lr=%x\n", so, so->so_usecount, lr_saved);
mutex_held = so->so_proto->pr_domain->dom_mtx;
} else {
mutex_held = ((struct inpcb *)so->so_pcb)->inpcb_mtx;
}
if (so->so_usecount == 0 && (inp->inp_wantcnt == WNT_STOPUSING)) {
lck_rw_lock_exclusive(divcbinfo.mtx);
in_pcbdispose(inp);
lck_rw_done(divcbinfo.mtx);
return (0);
}
lck_mtx_assert(mutex_held, LCK_MTX_ASSERT_OWNED);
lck_mtx_unlock(mutex_held);
so->reserved4 = (void *)lr_saved;
return (0);
}
__private_extern__ lck_mtx_t *
div_getlock(struct socket *so, __unused int locktype)
{
struct inpcb *inpcb = (struct inpcb *)so->so_pcb;
if (so->so_pcb) {
if (so->so_usecount < 0)
panic("div_getlock: so=%x usecount=%x\n", so, so->so_usecount);
return(inpcb->inpcb_mtx);
} else {
panic("div_getlock: so=%x NULL so_pcb\n", so);
return (so->so_proto->pr_domain->dom_mtx);
}
}
struct pr_usrreqs div_usrreqs = {
div_abort, pru_accept_notsupp, div_attach, div_bind,
pru_connect_notsupp, pru_connect2_notsupp, in_control, div_detach,
div_disconnect, pru_listen_notsupp, in_setpeeraddr, pru_rcvd_notsupp,
pru_rcvoob_notsupp, div_send, pru_sense_null, div_shutdown,
in_setsockaddr, sosend, soreceive, pru_sopoll_notsupp
};