1062 lines
24 KiB
C
1062 lines
24 KiB
C
/* $NetBSD: ipsecif.c,v 1.22 2023/09/01 11:23:39 andvar Exp $ */
|
|
|
|
/*
|
|
* Copyright (c) 2017 Internet Initiative Japan Inc.
|
|
* All rights reserved.
|
|
*
|
|
* Redistribution and use in source and binary forms, with or without
|
|
* modification, are permitted provided that the following conditions
|
|
* are met:
|
|
* 1. Redistributions of source code must retain the above copyright
|
|
* notice, this list of conditions and the following disclaimer.
|
|
* 2. Redistributions in binary form must reproduce the above copyright
|
|
* notice, this list of conditions and the following disclaimer in the
|
|
* documentation and/or other materials provided with the distribution.
|
|
*
|
|
* THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
|
|
* ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
|
|
* TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
|
|
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
|
|
* BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
|
|
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
|
|
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
|
|
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
|
|
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
|
|
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
|
|
* POSSIBILITY OF SUCH DAMAGE.
|
|
*/
|
|
|
|
#include <sys/cdefs.h>
|
|
__KERNEL_RCSID(0, "$NetBSD: ipsecif.c,v 1.22 2023/09/01 11:23:39 andvar Exp $");
|
|
|
|
#ifdef _KERNEL_OPT
|
|
#include "opt_inet.h"
|
|
#include "opt_ipsec.h"
|
|
#endif
|
|
|
|
#include <sys/param.h>
|
|
#include <sys/systm.h>
|
|
#include <sys/socket.h>
|
|
#include <sys/sockio.h>
|
|
#include <sys/mbuf.h>
|
|
#include <sys/errno.h>
|
|
#include <sys/ioctl.h>
|
|
#include <sys/syslog.h>
|
|
#include <sys/kernel.h>
|
|
|
|
#include <net/if.h>
|
|
#include <net/route.h>
|
|
|
|
#include <netinet/in.h>
|
|
#include <netinet/in_systm.h>
|
|
#include <netinet/ip.h>
|
|
#include <netinet/ip_var.h>
|
|
#include <netinet/in_var.h>
|
|
#include <netinet/ip_encap.h>
|
|
#include <netinet/ip_ecn.h>
|
|
#include <netinet/ip_private.h>
|
|
#include <netinet/udp.h>
|
|
|
|
#ifdef INET6
|
|
#include <netinet/ip6.h>
|
|
#include <netinet6/ip6_var.h>
|
|
#include <netinet6/ip6_private.h>
|
|
#include <netinet6/in6_var.h>
|
|
#include <netinet6/ip6protosw.h> /* for struct ip6ctlparam */
|
|
#include <netinet/ip_ecn.h>
|
|
#endif
|
|
|
|
#include <netipsec/key.h>
|
|
#include <netipsec/ipsecif.h>
|
|
|
|
#include <net/if_ipsec.h>
|
|
|
|
static int ipsecif_set_natt_ports(struct ipsec_variant *, struct mbuf *);
|
|
static void ipsecif4_input(struct mbuf *, int, int, void *);
|
|
static int ipsecif4_output(struct ipsec_variant *, int, struct mbuf *);
|
|
static int ipsecif4_filter4(const struct ip *, struct ipsec_variant *,
|
|
struct ifnet *);
|
|
|
|
#ifdef INET6
|
|
static int ipsecif6_input(struct mbuf **, int *, int, void *);
|
|
static int ipsecif6_output(struct ipsec_variant *, int, struct mbuf *);
|
|
static int ipsecif6_filter6(const struct ip6_hdr *, struct ipsec_variant *,
|
|
struct ifnet *);
|
|
#endif
|
|
|
|
static int ip_ipsec_ttl = IPSEC_TTL;
|
|
static int ip_ipsec_copy_tos = 0;
|
|
#ifdef INET6
|
|
int ip6_ipsec_hlim = IPSEC_HLIM;
|
|
int ip6_ipsec_pmtu = 0;
|
|
static int ip6_ipsec_copy_tos = 0;
|
|
#endif
|
|
|
|
static const struct encapsw ipsecif4_encapsw = {
|
|
.encapsw4 = {
|
|
.pr_input = ipsecif4_input,
|
|
.pr_ctlinput = NULL,
|
|
}
|
|
};
|
|
|
|
#ifdef INET6
|
|
static const struct encapsw ipsecif6_encapsw;
|
|
#endif
|
|
|
|
static int
|
|
ipsecif_set_natt_ports(struct ipsec_variant *var, struct mbuf *m)
|
|
{
|
|
|
|
KASSERT(if_ipsec_heldref_variant(var));
|
|
|
|
if (var->iv_sport || var->iv_dport) {
|
|
struct m_tag *mtag;
|
|
|
|
mtag = m_tag_get(PACKET_TAG_IPSEC_NAT_T_PORTS,
|
|
sizeof(uint16_t) + sizeof(uint16_t), M_DONTWAIT);
|
|
if (mtag) {
|
|
uint16_t *natt_port;
|
|
|
|
natt_port = (uint16_t *)(mtag + 1);
|
|
natt_port[0] = var->iv_dport;
|
|
natt_port[1] = var->iv_sport;
|
|
m_tag_prepend(m, mtag);
|
|
} else {
|
|
return ENOBUFS;
|
|
}
|
|
}
|
|
|
|
return 0;
|
|
}
|
|
|
|
static struct mbuf *
|
|
ipsecif4_prepend_hdr(struct ipsec_variant *var, struct mbuf *m,
|
|
uint8_t proto, uint8_t tos)
|
|
{
|
|
struct ip *ip;
|
|
struct sockaddr_in *src, *dst;
|
|
|
|
src = satosin(var->iv_psrc);
|
|
dst = satosin(var->iv_pdst);
|
|
|
|
if (in_nullhost(src->sin_addr) || in_nullhost(src->sin_addr) ||
|
|
src->sin_addr.s_addr == INADDR_BROADCAST ||
|
|
dst->sin_addr.s_addr == INADDR_BROADCAST) {
|
|
m_freem(m);
|
|
return NULL;
|
|
}
|
|
m->m_flags &= ~M_BCAST;
|
|
|
|
if (IN_MULTICAST(src->sin_addr.s_addr) ||
|
|
IN_MULTICAST(dst->sin_addr.s_addr)) {
|
|
m_freem(m);
|
|
return NULL;
|
|
}
|
|
|
|
M_PREPEND(m, sizeof(struct ip), M_DONTWAIT);
|
|
if (m && M_UNWRITABLE(m, sizeof(struct ip)))
|
|
m = m_pullup(m, sizeof(struct ip));
|
|
if (m == NULL)
|
|
return NULL;
|
|
|
|
ip = mtod(m, struct ip *);
|
|
ip->ip_v = IPVERSION;
|
|
ip->ip_off = htons(0);
|
|
if (m->m_pkthdr.len < IP_MINFRAGSIZE)
|
|
ip->ip_id = 0;
|
|
else
|
|
ip->ip_id = ip_newid(NULL);
|
|
ip->ip_hl = sizeof(*ip) >> 2;
|
|
if (ip_ipsec_copy_tos)
|
|
ip->ip_tos = tos;
|
|
else
|
|
ip->ip_tos = 0;
|
|
ip->ip_sum = 0;
|
|
ip->ip_src = src->sin_addr;
|
|
ip->ip_dst = dst->sin_addr;
|
|
ip->ip_p = proto;
|
|
ip->ip_ttl = ip_ipsec_ttl;
|
|
ip->ip_len = htons(m->m_pkthdr.len);
|
|
#ifndef IPSEC_TX_TOS_CLEAR
|
|
struct ifnet *ifp = &var->iv_softc->ipsec_if;
|
|
if (ifp->if_flags & IFF_ECN)
|
|
ip_ecn_ingress(ECN_ALLOWED, &ip->ip_tos, &tos);
|
|
else
|
|
ip_ecn_ingress(ECN_NOCARE, &ip->ip_tos, &tos);
|
|
#endif
|
|
|
|
return m;
|
|
}
|
|
|
|
static int
|
|
ipsecif4_needfrag(struct mbuf *m, struct ipsecrequest *isr)
|
|
{
|
|
struct ip ip0;
|
|
struct ip *ip;
|
|
int mtu;
|
|
struct secasvar *sav;
|
|
|
|
sav = key_lookup_sa_bysaidx(&isr->saidx);
|
|
if (sav == NULL)
|
|
return 0;
|
|
|
|
if (!(sav->natt_type & UDP_ENCAP_ESPINUDP)) {
|
|
mtu = 0;
|
|
goto out;
|
|
}
|
|
|
|
if (m->m_len < sizeof(struct ip)) {
|
|
m_copydata(m, 0, sizeof(ip0), &ip0);
|
|
ip = &ip0;
|
|
} else {
|
|
ip = mtod(m, struct ip *);
|
|
}
|
|
mtu = sav->esp_frag;
|
|
if (ntohs(ip->ip_len) <= mtu)
|
|
mtu = 0;
|
|
|
|
out:
|
|
KEY_SA_UNREF(&sav);
|
|
return mtu;
|
|
}
|
|
|
|
static struct mbuf *
|
|
ipsecif4_flowinfo(struct mbuf *m, int family, int *proto0, u_int8_t *tos0)
|
|
{
|
|
const struct ip *ip;
|
|
int proto;
|
|
int tos;
|
|
|
|
KASSERT(proto0 != NULL);
|
|
KASSERT(tos0 != NULL);
|
|
|
|
switch (family) {
|
|
case AF_INET:
|
|
proto = IPPROTO_IPV4;
|
|
if (m->m_len < sizeof(*ip)) {
|
|
m = m_pullup(m, sizeof(*ip));
|
|
if (m == NULL) {
|
|
*tos0 = 0;
|
|
*proto0 = 0;
|
|
return NULL;
|
|
}
|
|
}
|
|
ip = mtod(m, const struct ip *);
|
|
tos = ip->ip_tos;
|
|
/* TODO: support ALTQ for inner packet */
|
|
break;
|
|
#ifdef INET6
|
|
case AF_INET6: {
|
|
const struct ip6_hdr *ip6;
|
|
proto = IPPROTO_IPV6;
|
|
if (m->m_len < sizeof(*ip6)) {
|
|
m = m_pullup(m, sizeof(*ip6));
|
|
if (m == NULL) {
|
|
*tos0 = 0;
|
|
*proto0 = 0;
|
|
return NULL;
|
|
}
|
|
}
|
|
ip6 = mtod(m, const struct ip6_hdr *);
|
|
tos = (ntohl(ip6->ip6_flow) >> 20) & 0xff;
|
|
/* TODO: support ALTQ for inner packet */
|
|
break;
|
|
}
|
|
#endif /* INET6 */
|
|
default:
|
|
*tos0 = 0;
|
|
*proto0 = 0;
|
|
return NULL;
|
|
}
|
|
|
|
*proto0 = proto;
|
|
*tos0 = tos;
|
|
return m;
|
|
}
|
|
|
|
static int
|
|
ipsecif4_fragout(struct ipsec_variant *var, int family, struct mbuf *m, int mtu)
|
|
{
|
|
struct ifnet *ifp = &var->iv_softc->ipsec_if;
|
|
struct mbuf *next;
|
|
struct m_tag *mtag;
|
|
int error;
|
|
|
|
KASSERT(if_ipsec_heldref_variant(var));
|
|
|
|
mtag = m_tag_find(m, PACKET_TAG_IPSEC_NAT_T_PORTS);
|
|
if (mtag)
|
|
m_tag_delete(m, mtag);
|
|
|
|
/* consider new IP header prepended in ipsecif4_output() */
|
|
if (mtu <= sizeof(struct ip)) {
|
|
m_freem(m);
|
|
return ENETUNREACH;
|
|
}
|
|
m->m_pkthdr.csum_flags |= M_CSUM_IPv4;
|
|
error = ip_fragment(m, ifp, mtu - sizeof(struct ip));
|
|
if (error)
|
|
return error;
|
|
|
|
for (error = 0; m; m = next) {
|
|
next = m->m_nextpkt;
|
|
m->m_nextpkt = NULL;
|
|
if (error) {
|
|
m_freem(m);
|
|
continue;
|
|
}
|
|
|
|
error = ipsecif4_output(var, family, m);
|
|
}
|
|
if (error == 0)
|
|
IP_STATINC(IP_STAT_FRAGMENTED);
|
|
|
|
return error;
|
|
}
|
|
|
|
int
|
|
ipsecif4_encap_func(struct mbuf *m, struct ip *ip, struct ipsec_variant *var)
|
|
{
|
|
struct m_tag *mtag;
|
|
struct sockaddr_in *src, *dst;
|
|
u_int16_t src_port = 0;
|
|
u_int16_t dst_port = 0;
|
|
|
|
KASSERT(var != NULL);
|
|
|
|
src = satosin(var->iv_psrc);
|
|
dst = satosin(var->iv_pdst);
|
|
mtag = m_tag_find(m, PACKET_TAG_IPSEC_NAT_T_PORTS);
|
|
if (mtag) {
|
|
u_int16_t *ports;
|
|
|
|
ports = (u_int16_t *)(mtag + 1);
|
|
src_port = ports[0];
|
|
dst_port = ports[1];
|
|
}
|
|
|
|
/* address match */
|
|
if (src->sin_addr.s_addr != ip->ip_dst.s_addr ||
|
|
dst->sin_addr.s_addr != ip->ip_src.s_addr)
|
|
return 0;
|
|
|
|
/* UDP encap? */
|
|
if (mtag == NULL && var->iv_sport == 0 && var->iv_dport == 0)
|
|
goto match;
|
|
|
|
/* port match */
|
|
if (src_port != var->iv_dport ||
|
|
dst_port != var->iv_sport) {
|
|
#ifdef DEBUG
|
|
printf("%s: port mismatch: pkt(%u, %u), if(%u, %u)\n",
|
|
__func__, ntohs(src_port), ntohs(dst_port),
|
|
ntohs(var->iv_sport), ntohs(var->iv_dport));
|
|
#endif
|
|
return 0;
|
|
}
|
|
|
|
match:
|
|
/*
|
|
* hide NAT-T information from encapsulated traffics.
|
|
* they don't know about IPsec.
|
|
*/
|
|
if (mtag)
|
|
m_tag_delete(m, mtag);
|
|
return sizeof(src->sin_addr) + sizeof(dst->sin_addr);
|
|
}
|
|
|
|
static int
|
|
ipsecif4_output(struct ipsec_variant *var, int family, struct mbuf *m)
|
|
{
|
|
struct secpolicy *sp = NULL;
|
|
u_int8_t tos;
|
|
int proto;
|
|
int error;
|
|
int mtu;
|
|
u_long sa_mtu = 0;
|
|
|
|
KASSERT(if_ipsec_heldref_variant(var));
|
|
KASSERT(if_ipsec_variant_is_configured(var));
|
|
KASSERT(var->iv_psrc->sa_family == AF_INET);
|
|
KASSERT(var->iv_pdst->sa_family == AF_INET);
|
|
|
|
switch (family) {
|
|
case AF_INET:
|
|
sp = IV_SP_OUT(var);
|
|
break;
|
|
case AF_INET6:
|
|
sp = IV_SP_OUT6(var);
|
|
break;
|
|
default:
|
|
m_freem(m);
|
|
return EAFNOSUPPORT;
|
|
}
|
|
KASSERT(sp != NULL);
|
|
/*
|
|
* The SPs in ipsec_variant are prevented from freed by
|
|
* ipsec_variant->iv_psref. So, KEY_SP_REF() is unnecessary here.
|
|
*
|
|
* However, lastused should be updated.
|
|
*/
|
|
key_sp_touch(sp);
|
|
|
|
KASSERT(sp->policy != IPSEC_POLICY_NONE);
|
|
KASSERT(sp->policy != IPSEC_POLICY_ENTRUST);
|
|
KASSERT(sp->policy != IPSEC_POLICY_BYPASS);
|
|
if (sp->policy != IPSEC_POLICY_IPSEC) {
|
|
m_freem(m);
|
|
error = ENETUNREACH;
|
|
goto done;
|
|
}
|
|
|
|
/* get flowinfo */
|
|
m = ipsecif4_flowinfo(m, family, &proto, &tos);
|
|
if (m == NULL) {
|
|
error = ENETUNREACH;
|
|
goto done;
|
|
}
|
|
|
|
/* prepend new IP header */
|
|
m = ipsecif4_prepend_hdr(var, m, proto, tos);
|
|
if (m == NULL) {
|
|
error = ENETUNREACH;
|
|
goto done;
|
|
}
|
|
|
|
/*
|
|
* Normal netipsec's NAT-T fragmentation is done in ip_output().
|
|
* See "natt_frag" processing.
|
|
* However, ipsec(4) interface's one is not done in the same way,
|
|
* so we must do NAT-T fragmentation by own code.
|
|
*/
|
|
/* NAT-T ESP fragmentation */
|
|
mtu = ipsecif4_needfrag(m, sp->req);
|
|
if (mtu > 0)
|
|
return ipsecif4_fragout(var, family, m, mtu);
|
|
|
|
/* set NAT-T ports */
|
|
error = ipsecif_set_natt_ports(var, m);
|
|
if (error) {
|
|
m_freem(m);
|
|
goto done;
|
|
}
|
|
|
|
/* IPsec output */
|
|
IP_STATINC(IP_STAT_LOCALOUT);
|
|
error = ipsec4_process_packet(m, sp->req, &sa_mtu);
|
|
if (error == ENOENT)
|
|
error = 0;
|
|
/*
|
|
* frangmentation is already done in ipsecif4_fragout(),
|
|
* so ipsec4_process_packet() must not do fragmentation here.
|
|
*/
|
|
KASSERT(sa_mtu == 0);
|
|
|
|
done:
|
|
return error;
|
|
}
|
|
|
|
#ifdef INET6
|
|
int
|
|
ipsecif6_encap_func(struct mbuf *m, struct ip6_hdr *ip6, struct ipsec_variant *var)
|
|
{
|
|
struct m_tag *mtag;
|
|
struct sockaddr_in6 *src, *dst;
|
|
u_int16_t src_port = 0;
|
|
u_int16_t dst_port = 0;
|
|
|
|
KASSERT(var != NULL);
|
|
|
|
src = satosin6(var->iv_psrc);
|
|
dst = satosin6(var->iv_pdst);
|
|
mtag = m_tag_find(m, PACKET_TAG_IPSEC_NAT_T_PORTS);
|
|
if (mtag) {
|
|
u_int16_t *ports;
|
|
|
|
ports = (u_int16_t *)(mtag + 1);
|
|
src_port = ports[0];
|
|
dst_port = ports[1];
|
|
}
|
|
|
|
/* address match */
|
|
if (!IN6_ARE_ADDR_EQUAL(&src->sin6_addr, &ip6->ip6_dst) ||
|
|
!IN6_ARE_ADDR_EQUAL(&dst->sin6_addr, &ip6->ip6_src))
|
|
return 0;
|
|
|
|
/* UDP encap? */
|
|
if (mtag == NULL && var->iv_sport == 0 && var->iv_dport == 0)
|
|
goto match;
|
|
|
|
/* port match */
|
|
if (src_port != var->iv_dport ||
|
|
dst_port != var->iv_sport) {
|
|
#ifdef DEBUG
|
|
printf("%s: port mismatch: pkt(%u, %u), if(%u, %u)\n",
|
|
__func__, ntohs(src_port), ntohs(dst_port),
|
|
ntohs(var->iv_sport), ntohs(var->iv_dport));
|
|
#endif
|
|
return 0;
|
|
}
|
|
|
|
match:
|
|
/*
|
|
* hide NAT-T information from encapsulated traffics.
|
|
* they don't know about IPsec.
|
|
*/
|
|
if (mtag)
|
|
m_tag_delete(m, mtag);
|
|
return sizeof(src->sin6_addr) + sizeof(dst->sin6_addr);
|
|
}
|
|
|
|
static int
|
|
ipsecif6_output(struct ipsec_variant *var, int family, struct mbuf *m)
|
|
{
|
|
struct ifnet *ifp = &var->iv_softc->ipsec_if;
|
|
struct ipsec_softc *sc = ifp->if_softc;
|
|
struct route *ro_pc;
|
|
kmutex_t *lock_pc;
|
|
struct rtentry *rt;
|
|
struct sockaddr_in6 *sin6_src;
|
|
struct sockaddr_in6 *sin6_dst;
|
|
struct ip6_hdr *ip6;
|
|
int proto, error, flags;
|
|
u_int8_t itos, otos;
|
|
union {
|
|
struct sockaddr dst;
|
|
struct sockaddr_in6 dst6;
|
|
} u;
|
|
|
|
KASSERT(if_ipsec_heldref_variant(var));
|
|
KASSERT(if_ipsec_variant_is_configured(var));
|
|
|
|
sin6_src = satosin6(var->iv_psrc);
|
|
sin6_dst = satosin6(var->iv_pdst);
|
|
|
|
KASSERT(sin6_src->sin6_family == AF_INET6);
|
|
KASSERT(sin6_dst->sin6_family == AF_INET6);
|
|
|
|
switch (family) {
|
|
#ifdef INET
|
|
case AF_INET:
|
|
{
|
|
struct ip *ip;
|
|
|
|
proto = IPPROTO_IPV4;
|
|
if (m->m_len < sizeof(*ip)) {
|
|
m = m_pullup(m, sizeof(*ip));
|
|
if (m == NULL)
|
|
return ENOBUFS;
|
|
}
|
|
ip = mtod(m, struct ip *);
|
|
itos = ip->ip_tos;
|
|
/* TODO: support ALTQ for inner packet */
|
|
break;
|
|
}
|
|
#endif /* INET */
|
|
case AF_INET6:
|
|
{
|
|
struct ip6_hdr *xip6;
|
|
proto = IPPROTO_IPV6;
|
|
if (m->m_len < sizeof(*xip6)) {
|
|
m = m_pullup(m, sizeof(*xip6));
|
|
if (m == NULL)
|
|
return ENOBUFS;
|
|
}
|
|
xip6 = mtod(m, struct ip6_hdr *);
|
|
itos = (ntohl(xip6->ip6_flow) >> 20) & 0xff;
|
|
/* TODO: support ALTQ for inner packet */
|
|
break;
|
|
}
|
|
default:
|
|
m_freem(m);
|
|
return EAFNOSUPPORT;
|
|
}
|
|
|
|
/* prepend new IP header */
|
|
M_PREPEND(m, sizeof(struct ip6_hdr), M_DONTWAIT);
|
|
if (m && M_UNWRITABLE(m, sizeof(struct ip6_hdr)))
|
|
m = m_pullup(m, sizeof(struct ip6_hdr));
|
|
if (m == NULL)
|
|
return ENOBUFS;
|
|
|
|
ip6 = mtod(m, struct ip6_hdr *);
|
|
ip6->ip6_flow = 0;
|
|
ip6->ip6_vfc &= ~IPV6_VERSION_MASK;
|
|
ip6->ip6_vfc |= IPV6_VERSION;
|
|
#if 0 /* ip6->ip6_plen will be filled by ip6_output */
|
|
ip6->ip6_plen = htons((u_short)m->m_pkthdr.len - sizeof(*ip6));
|
|
#endif
|
|
ip6->ip6_nxt = proto;
|
|
ip6->ip6_hlim = ip6_ipsec_hlim;
|
|
ip6->ip6_src = sin6_src->sin6_addr;
|
|
/* bidirectional configured tunnel mode */
|
|
if (!IN6_IS_ADDR_UNSPECIFIED(&sin6_dst->sin6_addr)) {
|
|
ip6->ip6_dst = sin6_dst->sin6_addr;
|
|
} else {
|
|
m_freem(m);
|
|
return ENETUNREACH;
|
|
}
|
|
#ifndef IPSEC_TX_TOS_CLEAR
|
|
if (!ip6_ipsec_copy_tos)
|
|
otos = 0;
|
|
|
|
if (ifp->if_flags & IFF_ECN)
|
|
ip_ecn_ingress(ECN_ALLOWED, &otos, &itos);
|
|
else
|
|
ip_ecn_ingress(ECN_NOCARE, &otos, &itos);
|
|
#else
|
|
if (ip6_ipsec_copy_tos)
|
|
otos = itos;
|
|
else
|
|
otos = 0;
|
|
#endif
|
|
ip6->ip6_flow &= ~ntohl(0xff00000);
|
|
ip6->ip6_flow |= htonl((u_int32_t)otos << 20);
|
|
|
|
sockaddr_in6_init(&u.dst6, &sin6_dst->sin6_addr, 0, 0, 0);
|
|
|
|
if_tunnel_get_ro(sc->ipsec_ro_percpu, &ro_pc, &lock_pc);
|
|
if ((rt = rtcache_lookup(ro_pc, &u.dst)) == NULL) {
|
|
if_tunnel_put_ro(sc->ipsec_ro_percpu, lock_pc);
|
|
m_freem(m);
|
|
return ENETUNREACH;
|
|
}
|
|
|
|
if (rt->rt_ifp == ifp) {
|
|
rtcache_unref(rt, ro_pc);
|
|
rtcache_free(ro_pc);
|
|
if_tunnel_put_ro(sc->ipsec_ro_percpu, lock_pc);
|
|
m_freem(m);
|
|
return ENETUNREACH;
|
|
}
|
|
rtcache_unref(rt, ro_pc);
|
|
|
|
/* set NAT-T ports */
|
|
error = ipsecif_set_natt_ports(var, m);
|
|
if (error) {
|
|
m_freem(m);
|
|
goto out;
|
|
}
|
|
|
|
/*
|
|
* - IPSEC_PMTU_MINMTU
|
|
* Force fragmentation to minimum MTU to avoid path MTU discovery
|
|
* - IPSEC_PMTU_OUTERMTU
|
|
* Trust outer MTU is large enough to send all packets
|
|
*
|
|
* It is too painful to ask for resend of inner packet, to achieve
|
|
* path MTU discovery for encapsulated packets.
|
|
*
|
|
* See RFC4459.
|
|
*/
|
|
if (sc->ipsec_pmtu == IPSEC_PMTU_SYSDEFAULT) {
|
|
switch (ip6_ipsec_pmtu) {
|
|
case IPSEC_PMTU_MINMTU:
|
|
flags = IPV6_MINMTU;
|
|
break;
|
|
case IPSEC_PMTU_OUTERMTU:
|
|
flags = 0;
|
|
break;
|
|
default:
|
|
#ifdef DEBUG
|
|
log(LOG_DEBUG, "%s: ignore unexpected ip6_ipsec_pmtu %d\n",
|
|
__func__, ip6_ipsec_pmtu);
|
|
#endif
|
|
flags = IPV6_MINMTU;
|
|
break;
|
|
}
|
|
} else {
|
|
switch (sc->ipsec_pmtu) {
|
|
case IPSEC_PMTU_MINMTU:
|
|
flags = IPV6_MINMTU;
|
|
break;
|
|
case IPSEC_PMTU_OUTERMTU:
|
|
flags = 0;
|
|
break;
|
|
default:
|
|
#ifdef DEBUG
|
|
log(LOG_DEBUG, "%s: ignore unexpected ipsec_pmtu of %s %d\n",
|
|
__func__, ifp->if_xname, sc->ipsec_pmtu);
|
|
#endif
|
|
flags = IPV6_MINMTU;
|
|
break;
|
|
}
|
|
}
|
|
error = ip6_output(m, 0, ro_pc, flags, 0, NULL, NULL);
|
|
|
|
out:
|
|
if (error)
|
|
rtcache_free(ro_pc);
|
|
if_tunnel_put_ro(sc->ipsec_ro_percpu, lock_pc);
|
|
|
|
return error;
|
|
}
|
|
#endif /* INET6 */
|
|
|
|
static void
|
|
ipsecif4_input(struct mbuf *m, int off, int proto, void *eparg)
|
|
{
|
|
struct ifnet *ipsecp;
|
|
struct ipsec_softc *sc = eparg;
|
|
struct ipsec_variant *var;
|
|
const struct ip *ip;
|
|
int af;
|
|
#ifndef IPSEC_TX_TOS_CLEAR
|
|
u_int8_t otos;
|
|
#endif
|
|
struct psref psref_rcvif;
|
|
struct psref psref_var;
|
|
struct ifnet *rcvif;
|
|
|
|
KASSERT(sc != NULL);
|
|
|
|
ipsecp = &sc->ipsec_if;
|
|
if ((ipsecp->if_flags & IFF_UP) == 0) {
|
|
m_freem(m);
|
|
ip_statinc(IP_STAT_NOIPSEC);
|
|
return;
|
|
}
|
|
|
|
var = if_ipsec_getref_variant(sc, &psref_var);
|
|
if (if_ipsec_variant_is_unconfigured(var)) {
|
|
if_ipsec_putref_variant(var, &psref_var);
|
|
m_freem(m);
|
|
ip_statinc(IP_STAT_NOIPSEC);
|
|
return;
|
|
}
|
|
|
|
ip = mtod(m, const struct ip *);
|
|
|
|
rcvif = m_get_rcvif_psref(m, &psref_rcvif);
|
|
if (rcvif == NULL || !ipsecif4_filter4(ip, var, rcvif)) {
|
|
m_put_rcvif_psref(rcvif, &psref_rcvif);
|
|
if_ipsec_putref_variant(var, &psref_var);
|
|
m_freem(m);
|
|
ip_statinc(IP_STAT_NOIPSEC);
|
|
return;
|
|
}
|
|
m_put_rcvif_psref(rcvif, &psref_rcvif);
|
|
if_ipsec_putref_variant(var, &psref_var);
|
|
#ifndef IPSEC_TX_TOS_CLEAR
|
|
otos = ip->ip_tos;
|
|
#endif
|
|
m_adj(m, off);
|
|
|
|
switch (proto) {
|
|
case IPPROTO_IPV4:
|
|
{
|
|
struct ip *xip;
|
|
af = AF_INET;
|
|
if (M_UNWRITABLE(m, sizeof(*xip))) {
|
|
m = m_pullup(m, sizeof(*xip));
|
|
if (m == NULL)
|
|
return;
|
|
}
|
|
xip = mtod(m, struct ip *);
|
|
#ifndef IPSEC_TX_TOS_CLEAR
|
|
if (ipsecp->if_flags & IFF_ECN)
|
|
ip_ecn_egress(ECN_ALLOWED, &otos, &xip->ip_tos);
|
|
else
|
|
ip_ecn_egress(ECN_NOCARE, &otos, &xip->ip_tos);
|
|
#endif
|
|
break;
|
|
}
|
|
#ifdef INET6
|
|
case IPPROTO_IPV6:
|
|
{
|
|
struct ip6_hdr *ip6;
|
|
u_int8_t itos;
|
|
af = AF_INET6;
|
|
if (M_UNWRITABLE(m, sizeof(*ip6))) {
|
|
m = m_pullup(m, sizeof(*ip6));
|
|
if (m == NULL)
|
|
return;
|
|
}
|
|
ip6 = mtod(m, struct ip6_hdr *);
|
|
itos = (ntohl(ip6->ip6_flow) >> 20) & 0xff;
|
|
#ifndef IPSEC_TX_TOS_CLEAR
|
|
if (ipsecp->if_flags & IFF_ECN)
|
|
ip_ecn_egress(ECN_ALLOWED, &otos, &itos);
|
|
else
|
|
ip_ecn_egress(ECN_NOCARE, &otos, &itos);
|
|
#endif
|
|
ip6->ip6_flow &= ~htonl(0xff << 20);
|
|
ip6->ip6_flow |= htonl((u_int32_t)itos << 20);
|
|
break;
|
|
}
|
|
#endif /* INET6 */
|
|
default:
|
|
ip_statinc(IP_STAT_NOIPSEC);
|
|
m_freem(m);
|
|
return;
|
|
}
|
|
if_ipsec_input(m, af, ipsecp);
|
|
|
|
return;
|
|
}
|
|
|
|
/*
|
|
* validate and filter the packet
|
|
*/
|
|
static int
|
|
ipsecif4_filter4(const struct ip *ip, struct ipsec_variant *var,
|
|
struct ifnet *ifp)
|
|
{
|
|
struct sockaddr_in *src, *dst;
|
|
|
|
src = satosin(var->iv_psrc);
|
|
dst = satosin(var->iv_pdst);
|
|
|
|
return in_tunnel_validate(ip, src->sin_addr, dst->sin_addr);
|
|
}
|
|
|
|
#ifdef INET6
|
|
static int
|
|
ipsecif6_input(struct mbuf **mp, int *offp, int proto, void *eparg)
|
|
{
|
|
struct mbuf *m = *mp;
|
|
struct ifnet *ipsecp;
|
|
struct ipsec_softc *sc = eparg;
|
|
struct ipsec_variant *var;
|
|
struct ip6_hdr *ip6;
|
|
int af = 0;
|
|
#ifndef IPSEC_TX_TOS_CLEAR
|
|
u_int32_t otos;
|
|
#endif
|
|
struct psref psref_rcvif;
|
|
struct psref psref_var;
|
|
struct ifnet *rcvif;
|
|
|
|
KASSERT(eparg != NULL);
|
|
|
|
ipsecp = &sc->ipsec_if;
|
|
if ((ipsecp->if_flags & IFF_UP) == 0) {
|
|
m_freem(m);
|
|
IP6_STATINC(IP6_STAT_NOIPSEC);
|
|
return IPPROTO_DONE;
|
|
}
|
|
|
|
var = if_ipsec_getref_variant(sc, &psref_var);
|
|
if (if_ipsec_variant_is_unconfigured(var)) {
|
|
if_ipsec_putref_variant(var, &psref_var);
|
|
m_freem(m);
|
|
IP6_STATINC(IP6_STAT_NOIPSEC);
|
|
return IPPROTO_DONE;
|
|
}
|
|
|
|
ip6 = mtod(m, struct ip6_hdr *);
|
|
|
|
rcvif = m_get_rcvif_psref(m, &psref_rcvif);
|
|
if (rcvif == NULL || !ipsecif6_filter6(ip6, var, rcvif)) {
|
|
m_put_rcvif_psref(rcvif, &psref_rcvif);
|
|
if_ipsec_putref_variant(var, &psref_var);
|
|
m_freem(m);
|
|
IP6_STATINC(IP6_STAT_NOIPSEC);
|
|
return IPPROTO_DONE;
|
|
}
|
|
m_put_rcvif_psref(rcvif, &psref_rcvif);
|
|
if_ipsec_putref_variant(var, &psref_var);
|
|
|
|
#ifndef IPSEC_TX_TOS_CLEAR
|
|
otos = ip6->ip6_flow;
|
|
#endif
|
|
m_adj(m, *offp);
|
|
|
|
switch (proto) {
|
|
#ifdef INET
|
|
case IPPROTO_IPV4:
|
|
{
|
|
af = AF_INET;
|
|
#ifndef IPSEC_TX_TOS_CLEAR
|
|
struct ip *ip;
|
|
u_int8_t otos8;
|
|
otos8 = (ntohl(otos) >> 20) & 0xff;
|
|
|
|
if (M_UNWRITABLE(m, sizeof(*ip))) {
|
|
m = m_pullup(m, sizeof(*ip));
|
|
if (m == NULL)
|
|
return IPPROTO_DONE;
|
|
}
|
|
ip = mtod(m, struct ip *);
|
|
if (ipsecp->if_flags & IFF_ECN)
|
|
ip_ecn_egress(ECN_ALLOWED, &otos8, &ip->ip_tos);
|
|
else
|
|
ip_ecn_egress(ECN_NOCARE, &otos8, &ip->ip_tos);
|
|
#endif
|
|
break;
|
|
}
|
|
#endif /* INET */
|
|
case IPPROTO_IPV6:
|
|
{
|
|
af = AF_INET6;
|
|
#ifndef IPSEC_TX_TOS_CLEAR
|
|
struct ip6_hdr *xip6;
|
|
|
|
if (M_UNWRITABLE(m, sizeof(*xip6))) {
|
|
m = m_pullup(m, sizeof(*xip6));
|
|
if (m == NULL)
|
|
return IPPROTO_DONE;
|
|
}
|
|
xip6 = mtod(m, struct ip6_hdr *);
|
|
if (ipsecp->if_flags & IFF_ECN)
|
|
ip6_ecn_egress(ECN_ALLOWED, &otos, &xip6->ip6_flow);
|
|
else
|
|
ip6_ecn_egress(ECN_NOCARE, &otos, &xip6->ip6_flow);
|
|
break;
|
|
#endif
|
|
}
|
|
default:
|
|
IP6_STATINC(IP6_STAT_NOIPSEC);
|
|
m_freem(m);
|
|
return IPPROTO_DONE;
|
|
}
|
|
|
|
if_ipsec_input(m, af, ipsecp);
|
|
return IPPROTO_DONE;
|
|
}
|
|
|
|
/*
|
|
* validate and filter the packet.
|
|
*/
|
|
static int
|
|
ipsecif6_filter6(const struct ip6_hdr *ip6, struct ipsec_variant *var,
|
|
struct ifnet *ifp)
|
|
{
|
|
struct sockaddr_in6 *src, *dst;
|
|
|
|
src = satosin6(var->iv_psrc);
|
|
dst = satosin6(var->iv_pdst);
|
|
|
|
return in6_tunnel_validate(ip6, &src->sin6_addr, &dst->sin6_addr);
|
|
}
|
|
#endif /* INET6 */
|
|
|
|
int
|
|
ipsecif4_attach(struct ipsec_variant *var)
|
|
{
|
|
struct ipsec_softc *sc = var->iv_softc;
|
|
|
|
KASSERT(if_ipsec_variant_is_configured(var));
|
|
|
|
if (var->iv_encap_cookie4 != NULL)
|
|
return EALREADY;
|
|
|
|
var->iv_encap_cookie4 = encap_attach_addr(AF_INET, -1,
|
|
var->iv_psrc, var->iv_pdst, if_ipsec_encap_func, &ipsecif4_encapsw,
|
|
sc);
|
|
if (var->iv_encap_cookie4 == NULL)
|
|
return EEXIST;
|
|
|
|
var->iv_output = ipsecif4_output;
|
|
return 0;
|
|
}
|
|
|
|
int
|
|
ipsecif4_detach(struct ipsec_variant *var)
|
|
{
|
|
int error;
|
|
|
|
if (var->iv_encap_cookie4 == NULL)
|
|
return 0;
|
|
|
|
var->iv_output = NULL;
|
|
error = encap_detach(var->iv_encap_cookie4);
|
|
if (error == 0)
|
|
var->iv_encap_cookie4 = NULL;
|
|
|
|
return error;
|
|
}
|
|
|
|
#ifdef INET6
|
|
int
|
|
ipsecif6_attach(struct ipsec_variant *var)
|
|
{
|
|
struct ipsec_softc *sc = var->iv_softc;
|
|
|
|
KASSERT(if_ipsec_variant_is_configured(var));
|
|
KASSERT(var->iv_encap_cookie6 == NULL);
|
|
|
|
var->iv_encap_cookie6 = encap_attach_addr(AF_INET6, -1,
|
|
var->iv_psrc, var->iv_pdst, if_ipsec_encap_func, &ipsecif6_encapsw,
|
|
sc);
|
|
if (var->iv_encap_cookie6 == NULL)
|
|
return EEXIST;
|
|
|
|
var->iv_output = ipsecif6_output;
|
|
return 0;
|
|
}
|
|
|
|
int
|
|
ipsecif6_detach(struct ipsec_variant *var)
|
|
{
|
|
struct ipsec_softc *sc = var->iv_softc;
|
|
int error;
|
|
|
|
KASSERT(var->iv_encap_cookie6 != NULL);
|
|
|
|
if_tunnel_ro_percpu_rtcache_free(sc->ipsec_ro_percpu);
|
|
|
|
var->iv_output = NULL;
|
|
error = encap_detach(var->iv_encap_cookie6);
|
|
if (error == 0)
|
|
var->iv_encap_cookie6 = NULL;
|
|
return error;
|
|
}
|
|
|
|
void *
|
|
ipsecif6_ctlinput(int cmd, const struct sockaddr *sa, void *d, void *eparg)
|
|
{
|
|
struct ipsec_softc *sc = eparg;
|
|
struct ip6ctlparam *ip6cp = NULL;
|
|
struct ip6_hdr *ip6;
|
|
const struct sockaddr_in6 *dst6;
|
|
struct route *ro_pc;
|
|
kmutex_t *lock_pc;
|
|
|
|
if (sa->sa_family != AF_INET6 ||
|
|
sa->sa_len != sizeof(struct sockaddr_in6))
|
|
return NULL;
|
|
|
|
if ((unsigned)cmd >= PRC_NCMDS)
|
|
return NULL;
|
|
if (cmd == PRC_HOSTDEAD)
|
|
d = NULL;
|
|
else if (inet6ctlerrmap[cmd] == 0)
|
|
return NULL;
|
|
|
|
/* if the parameter is from icmp6, decode it. */
|
|
if (d != NULL) {
|
|
ip6cp = (struct ip6ctlparam *)d;
|
|
ip6 = ip6cp->ip6c_ip6;
|
|
} else {
|
|
ip6 = NULL;
|
|
}
|
|
|
|
if (!ip6)
|
|
return NULL;
|
|
|
|
if_tunnel_get_ro(sc->ipsec_ro_percpu, &ro_pc, &lock_pc);
|
|
dst6 = satocsin6(rtcache_getdst(ro_pc));
|
|
/* XXX scope */
|
|
if (dst6 == NULL)
|
|
;
|
|
else if (IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst, &dst6->sin6_addr))
|
|
/* flush route cache */
|
|
rtcache_free(ro_pc);
|
|
|
|
if_tunnel_put_ro(sc->ipsec_ro_percpu, lock_pc);
|
|
|
|
return NULL;
|
|
}
|
|
|
|
ENCAP_PR_WRAP_CTLINPUT(ipsecif6_ctlinput)
|
|
#define ipsecif6_ctlinput ipsecif6_ctlinput_wrapper
|
|
|
|
static const struct encapsw ipsecif6_encapsw = {
|
|
.encapsw6 = {
|
|
.pr_input = ipsecif6_input,
|
|
.pr_ctlinput = ipsecif6_ctlinput,
|
|
}
|
|
};
|
|
#endif /* INET6 */
|