2014-11-07 19:13:19 +00:00
|
|
|
/*-
|
2018-06-14 14:53:24 +00:00
|
|
|
* Copyright (c) 2014, 2018 Andrey V. Elsukov <ae@FreeBSD.org>
|
2014-11-07 19:13:19 +00:00
|
|
|
* 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 AUTHOR ``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 AUTHOR 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>
|
|
|
|
__FBSDID("$FreeBSD$");
|
|
|
|
|
|
|
|
#include <sys/param.h>
|
2018-06-14 14:53:24 +00:00
|
|
|
#include <sys/systm.h>
|
2014-11-07 19:13:19 +00:00
|
|
|
#include <sys/jail.h>
|
|
|
|
#include <sys/kernel.h>
|
|
|
|
#include <sys/lock.h>
|
|
|
|
#include <sys/malloc.h>
|
|
|
|
#include <sys/module.h>
|
|
|
|
#include <sys/mbuf.h>
|
|
|
|
#include <sys/priv.h>
|
|
|
|
#include <sys/proc.h>
|
|
|
|
#include <sys/socket.h>
|
|
|
|
#include <sys/sockio.h>
|
|
|
|
#include <sys/sx.h>
|
|
|
|
#include <sys/sysctl.h>
|
|
|
|
#include <sys/syslog.h>
|
|
|
|
|
|
|
|
#include <net/bpf.h>
|
|
|
|
#include <net/ethernet.h>
|
|
|
|
#include <net/if.h>
|
|
|
|
#include <net/if_var.h>
|
|
|
|
#include <net/if_clone.h>
|
|
|
|
#include <net/if_types.h>
|
|
|
|
#include <net/netisr.h>
|
|
|
|
#include <net/vnet.h>
|
2015-05-12 07:37:27 +00:00
|
|
|
#include <net/route.h>
|
2014-11-07 19:13:19 +00:00
|
|
|
|
|
|
|
#include <netinet/in.h>
|
|
|
|
#include <netinet/in_systm.h>
|
|
|
|
#include <netinet/in_var.h>
|
|
|
|
#include <netinet/ip.h>
|
|
|
|
#include <netinet/ip_var.h>
|
|
|
|
#include <netinet/ip_encap.h>
|
|
|
|
|
|
|
|
#include <machine/in_cksum.h>
|
|
|
|
#include <security/mac/mac_framework.h>
|
|
|
|
|
2017-04-11 08:56:18 +00:00
|
|
|
#define MEMTU (1500 - sizeof(struct mobhdr))
|
2014-11-07 19:13:19 +00:00
|
|
|
static const char mename[] = "me";
|
|
|
|
static MALLOC_DEFINE(M_IFME, mename, "Minimal Encapsulation for IP");
|
|
|
|
/* Minimal forwarding header RFC 2004 */
|
|
|
|
struct mobhdr {
|
|
|
|
uint8_t mob_proto; /* protocol */
|
|
|
|
uint8_t mob_flags; /* flags */
|
|
|
|
#define MOB_FLAGS_SP 0x80 /* source present */
|
|
|
|
uint16_t mob_csum; /* header checksum */
|
|
|
|
struct in_addr mob_dst; /* original destination address */
|
|
|
|
struct in_addr mob_src; /* original source addr (optional) */
|
|
|
|
} __packed;
|
|
|
|
|
|
|
|
struct me_softc {
|
|
|
|
struct ifnet *me_ifp;
|
|
|
|
u_int me_fibnum;
|
|
|
|
struct in_addr me_src;
|
|
|
|
struct in_addr me_dst;
|
2018-06-14 14:53:24 +00:00
|
|
|
|
|
|
|
CK_LIST_ENTRY(me_softc) chain;
|
2014-11-07 19:13:19 +00:00
|
|
|
};
|
2018-06-14 14:53:24 +00:00
|
|
|
CK_LIST_HEAD(me_list, me_softc);
|
2014-11-07 19:13:19 +00:00
|
|
|
#define ME2IFP(sc) ((sc)->me_ifp)
|
|
|
|
#define ME_READY(sc) ((sc)->me_src.s_addr != 0)
|
2018-06-14 14:53:24 +00:00
|
|
|
#define ME_RLOCK() epoch_enter_preempt(net_epoch_preempt)
|
|
|
|
#define ME_RUNLOCK() epoch_exit_preempt(net_epoch_preempt)
|
|
|
|
#define ME_WAIT() epoch_wait_preempt(net_epoch_preempt)
|
|
|
|
|
|
|
|
#ifndef ME_HASH_SIZE
|
|
|
|
#define ME_HASH_SIZE (1 << 4)
|
|
|
|
#endif
|
|
|
|
static VNET_DEFINE(struct me_list *, me_hashtbl) = NULL;
|
|
|
|
#define V_me_hashtbl VNET(me_hashtbl)
|
|
|
|
#define ME_HASH(src, dst) (V_me_hashtbl[\
|
|
|
|
me_hashval((src), (dst)) & (ME_HASH_SIZE - 1)])
|
|
|
|
|
2014-11-07 19:13:19 +00:00
|
|
|
static struct sx me_ioctl_sx;
|
|
|
|
SX_SYSINIT(me_ioctl_sx, &me_ioctl_sx, "me_ioctl");
|
|
|
|
|
|
|
|
static int me_clone_create(struct if_clone *, int, caddr_t);
|
|
|
|
static void me_clone_destroy(struct ifnet *);
|
|
|
|
static VNET_DEFINE(struct if_clone *, me_cloner);
|
|
|
|
#define V_me_cloner VNET(me_cloner)
|
|
|
|
|
|
|
|
static void me_qflush(struct ifnet *);
|
|
|
|
static int me_transmit(struct ifnet *, struct mbuf *);
|
|
|
|
static int me_ioctl(struct ifnet *, u_long, caddr_t);
|
|
|
|
static int me_output(struct ifnet *, struct mbuf *,
|
|
|
|
const struct sockaddr *, struct route *);
|
Rework IP encapsulation handling code.
Currently it has several disadvantages:
- it uses single mutex to protect internal structures. It is used by
data- and control- path, thus there are no parallelism at all.
- it uses single list to keep encap handlers for both INET and INET6
families.
- struct encaptab keeps unneeded information (src, dst, masks, protosw),
that isn't used by code in the source tree.
- matches are prioritized and when many tunneling interfaces are
registered, encapcheck handler of each interface is invoked for each
packet. The search takes O(n) for n interfaces. All this work is done
with exclusive lock held.
What this patch includes:
- the datapath is converted to be lockless using epoch(9) KPI.
- struct encaptab now linked using CK_LIST.
- all unused fields removed from struct encaptab. Several new fields
addedr: min_length is the minimum packet length, that encapsulation
handler expects to see; exact_match is maximum number of bits, that
can return an encapsulation handler, when it wants to consume a packet.
- IPv6 and IPv4 handlers are stored in separate lists;
- added new "encap_lookup_t" method, that will be used later. It is
targeted to speedup lookup of needed interface, when gif(4)/gre(4) have
many interfaces.
- the need to use protosw structure is eliminated. The only pr_input
method was used from this structure, so I don't see the need to keep
using it.
- encap_input_t method changed to avoid using mbuf tags to store softc
pointer. Now it is passed directly trough encap_input_t method.
encap_getarg() funtions is removed.
- all sockaddr structures and code that uses them removed. We don't have
any code in the tree that uses them. All consumers use encap_attach_func()
method, that relies on invoking of encapcheck() to determine the needed
handler.
- introduced struct encap_config, it contains parameters of encap handler
that is going to be registered by encap_attach() function.
- encap handlers are stored in lists ordered by exact_match value, thus
handlers that need more bits to match will be checked first, and if
encapcheck method returns exact_match value, the search will be stopped.
- all current consumers changed to use new KPI.
Reviewed by: mmacy
Sponsored by: Yandex LLC
Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
|
|
|
static int me_input(struct mbuf *, int, int, void *);
|
2014-11-07 19:13:19 +00:00
|
|
|
|
2018-06-14 14:53:24 +00:00
|
|
|
static int me_set_tunnel(struct me_softc *, in_addr_t, in_addr_t);
|
|
|
|
static void me_delete_tunnel(struct me_softc *);
|
2014-11-07 19:13:19 +00:00
|
|
|
|
|
|
|
SYSCTL_DECL(_net_link);
|
|
|
|
static SYSCTL_NODE(_net_link, IFT_TUNNEL, me, CTLFLAG_RW, 0,
|
|
|
|
"Minimal Encapsulation for IP (RFC 2004)");
|
|
|
|
#ifndef MAX_ME_NEST
|
|
|
|
#define MAX_ME_NEST 1
|
|
|
|
#endif
|
|
|
|
|
|
|
|
static VNET_DEFINE(int, max_me_nesting) = MAX_ME_NEST;
|
|
|
|
#define V_max_me_nesting VNET(max_me_nesting)
|
|
|
|
SYSCTL_INT(_net_link_me, OID_AUTO, max_nesting, CTLFLAG_RW | CTLFLAG_VNET,
|
|
|
|
&VNET_NAME(max_me_nesting), 0, "Max nested tunnels");
|
|
|
|
|
2018-06-14 14:53:24 +00:00
|
|
|
static uint32_t
|
|
|
|
me_hashval(in_addr_t src, in_addr_t dst)
|
|
|
|
{
|
|
|
|
uint32_t ret;
|
|
|
|
|
|
|
|
ret = fnv_32_buf(&src, sizeof(src), FNV1_32_INIT);
|
|
|
|
return (fnv_32_buf(&dst, sizeof(dst), ret));
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct me_list *
|
|
|
|
me_hashinit(void)
|
|
|
|
{
|
|
|
|
struct me_list *hash;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
hash = malloc(sizeof(struct me_list) * ME_HASH_SIZE,
|
|
|
|
M_IFME, M_WAITOK);
|
|
|
|
for (i = 0; i < ME_HASH_SIZE; i++)
|
|
|
|
CK_LIST_INIT(&hash[i]);
|
|
|
|
|
|
|
|
return (hash);
|
|
|
|
}
|
|
|
|
|
2014-11-07 19:13:19 +00:00
|
|
|
static void
|
|
|
|
vnet_me_init(const void *unused __unused)
|
|
|
|
{
|
|
|
|
V_me_cloner = if_clone_simple(mename, me_clone_create,
|
|
|
|
me_clone_destroy, 0);
|
|
|
|
}
|
|
|
|
VNET_SYSINIT(vnet_me_init, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY,
|
|
|
|
vnet_me_init, NULL);
|
|
|
|
|
|
|
|
static void
|
|
|
|
vnet_me_uninit(const void *unused __unused)
|
|
|
|
{
|
|
|
|
|
2018-06-14 14:53:24 +00:00
|
|
|
if (V_me_hashtbl != NULL)
|
|
|
|
free(V_me_hashtbl, M_IFME);
|
2014-11-07 19:13:19 +00:00
|
|
|
if_clone_detach(V_me_cloner);
|
|
|
|
}
|
|
|
|
VNET_SYSUNINIT(vnet_me_uninit, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY,
|
|
|
|
vnet_me_uninit, NULL);
|
|
|
|
|
|
|
|
static int
|
|
|
|
me_clone_create(struct if_clone *ifc, int unit, caddr_t params)
|
|
|
|
{
|
|
|
|
struct me_softc *sc;
|
|
|
|
|
|
|
|
sc = malloc(sizeof(struct me_softc), M_IFME, M_WAITOK | M_ZERO);
|
|
|
|
sc->me_fibnum = curthread->td_proc->p_fibnum;
|
|
|
|
ME2IFP(sc) = if_alloc(IFT_TUNNEL);
|
|
|
|
ME2IFP(sc)->if_softc = sc;
|
|
|
|
if_initname(ME2IFP(sc), mename, unit);
|
|
|
|
|
2017-04-11 08:56:18 +00:00
|
|
|
ME2IFP(sc)->if_mtu = MEMTU;;
|
2014-11-07 19:13:19 +00:00
|
|
|
ME2IFP(sc)->if_flags = IFF_POINTOPOINT|IFF_MULTICAST;
|
|
|
|
ME2IFP(sc)->if_output = me_output;
|
|
|
|
ME2IFP(sc)->if_ioctl = me_ioctl;
|
|
|
|
ME2IFP(sc)->if_transmit = me_transmit;
|
|
|
|
ME2IFP(sc)->if_qflush = me_qflush;
|
2015-10-03 09:15:23 +00:00
|
|
|
ME2IFP(sc)->if_capabilities |= IFCAP_LINKSTATE;
|
|
|
|
ME2IFP(sc)->if_capenable |= IFCAP_LINKSTATE;
|
2014-11-07 19:13:19 +00:00
|
|
|
if_attach(ME2IFP(sc));
|
|
|
|
bpfattach(ME2IFP(sc), DLT_NULL, sizeof(u_int32_t));
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
|
|
|
me_clone_destroy(struct ifnet *ifp)
|
|
|
|
{
|
|
|
|
struct me_softc *sc;
|
|
|
|
|
|
|
|
sx_xlock(&me_ioctl_sx);
|
|
|
|
sc = ifp->if_softc;
|
2018-06-14 14:53:24 +00:00
|
|
|
me_delete_tunnel(sc);
|
2014-11-07 19:13:19 +00:00
|
|
|
bpfdetach(ifp);
|
|
|
|
if_detach(ifp);
|
|
|
|
ifp->if_softc = NULL;
|
|
|
|
sx_xunlock(&me_ioctl_sx);
|
|
|
|
|
2018-06-14 14:53:24 +00:00
|
|
|
ME_WAIT();
|
2014-11-07 19:13:19 +00:00
|
|
|
if_free(ifp);
|
|
|
|
free(sc, M_IFME);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int
|
|
|
|
me_ioctl(struct ifnet *ifp, u_long cmd, caddr_t data)
|
|
|
|
{
|
|
|
|
struct ifreq *ifr = (struct ifreq *)data;
|
|
|
|
struct sockaddr_in *src, *dst;
|
|
|
|
struct me_softc *sc;
|
|
|
|
int error;
|
|
|
|
|
|
|
|
switch (cmd) {
|
|
|
|
case SIOCSIFMTU:
|
|
|
|
if (ifr->ifr_mtu < 576)
|
|
|
|
return (EINVAL);
|
2017-04-11 08:56:18 +00:00
|
|
|
ifp->if_mtu = ifr->ifr_mtu;
|
2014-11-07 19:13:19 +00:00
|
|
|
return (0);
|
|
|
|
case SIOCSIFADDR:
|
|
|
|
ifp->if_flags |= IFF_UP;
|
|
|
|
case SIOCSIFFLAGS:
|
|
|
|
case SIOCADDMULTI:
|
|
|
|
case SIOCDELMULTI:
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
sx_xlock(&me_ioctl_sx);
|
|
|
|
sc = ifp->if_softc;
|
|
|
|
if (sc == NULL) {
|
|
|
|
error = ENXIO;
|
|
|
|
goto end;
|
|
|
|
}
|
|
|
|
error = 0;
|
|
|
|
switch (cmd) {
|
|
|
|
case SIOCSIFPHYADDR:
|
2018-06-14 14:53:24 +00:00
|
|
|
src = &((struct in_aliasreq *)data)->ifra_addr;
|
|
|
|
dst = &((struct in_aliasreq *)data)->ifra_dstaddr;
|
2014-11-07 19:13:19 +00:00
|
|
|
if (src->sin_family != dst->sin_family ||
|
|
|
|
src->sin_family != AF_INET ||
|
|
|
|
src->sin_len != dst->sin_len ||
|
|
|
|
src->sin_len != sizeof(struct sockaddr_in)) {
|
|
|
|
error = EINVAL;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (src->sin_addr.s_addr == INADDR_ANY ||
|
|
|
|
dst->sin_addr.s_addr == INADDR_ANY) {
|
|
|
|
error = EADDRNOTAVAIL;
|
|
|
|
break;
|
|
|
|
}
|
2018-06-14 14:53:24 +00:00
|
|
|
error = me_set_tunnel(sc, src->sin_addr.s_addr,
|
|
|
|
dst->sin_addr.s_addr);
|
2014-11-07 19:13:19 +00:00
|
|
|
break;
|
|
|
|
case SIOCDIFPHYADDR:
|
2018-06-14 14:53:24 +00:00
|
|
|
me_delete_tunnel(sc);
|
2014-11-07 19:13:19 +00:00
|
|
|
break;
|
|
|
|
case SIOCGIFPSRCADDR:
|
|
|
|
case SIOCGIFPDSTADDR:
|
|
|
|
if (!ME_READY(sc)) {
|
|
|
|
error = EADDRNOTAVAIL;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
src = (struct sockaddr_in *)&ifr->ifr_addr;
|
|
|
|
memset(src, 0, sizeof(*src));
|
|
|
|
src->sin_family = AF_INET;
|
|
|
|
src->sin_len = sizeof(*src);
|
|
|
|
switch (cmd) {
|
|
|
|
case SIOCGIFPSRCADDR:
|
|
|
|
src->sin_addr = sc->me_src;
|
|
|
|
break;
|
|
|
|
case SIOCGIFPDSTADDR:
|
|
|
|
src->sin_addr = sc->me_dst;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
error = prison_if(curthread->td_ucred, sintosa(src));
|
|
|
|
if (error != 0)
|
|
|
|
memset(src, 0, sizeof(*src));
|
|
|
|
break;
|
2015-05-12 07:37:27 +00:00
|
|
|
case SIOCGTUNFIB:
|
|
|
|
ifr->ifr_fib = sc->me_fibnum;
|
|
|
|
break;
|
|
|
|
case SIOCSTUNFIB:
|
|
|
|
if ((error = priv_check(curthread, PRIV_NET_GRE)) != 0)
|
|
|
|
break;
|
|
|
|
if (ifr->ifr_fib >= rt_numfibs)
|
|
|
|
error = EINVAL;
|
|
|
|
else
|
|
|
|
sc->me_fibnum = ifr->ifr_fib;
|
|
|
|
break;
|
2014-11-07 19:13:19 +00:00
|
|
|
default:
|
|
|
|
error = EINVAL;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
end:
|
|
|
|
sx_xunlock(&me_ioctl_sx);
|
|
|
|
return (error);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int
|
2018-06-14 14:53:24 +00:00
|
|
|
me_lookup(const struct mbuf *m, int off, int proto, void **arg)
|
2014-11-07 19:13:19 +00:00
|
|
|
{
|
2018-06-14 14:53:24 +00:00
|
|
|
const struct ip *ip;
|
2014-11-07 19:13:19 +00:00
|
|
|
struct me_softc *sc;
|
|
|
|
|
2018-06-14 14:53:24 +00:00
|
|
|
MPASS(in_epoch());
|
|
|
|
ip = mtod(m, const struct ip *);
|
|
|
|
CK_LIST_FOREACH(sc, &ME_HASH(ip->ip_dst.s_addr,
|
|
|
|
ip->ip_src.s_addr), chain) {
|
2014-11-07 19:13:19 +00:00
|
|
|
if (sc->me_src.s_addr == ip->ip_dst.s_addr &&
|
2018-06-14 14:53:24 +00:00
|
|
|
sc->me_dst.s_addr == ip->ip_src.s_addr) {
|
|
|
|
if ((ME2IFP(sc)->if_flags & IFF_UP) == 0)
|
|
|
|
return (0);
|
|
|
|
*arg = sc;
|
|
|
|
return (ENCAP_DRV_LOOKUP);
|
|
|
|
}
|
2014-11-07 19:13:19 +00:00
|
|
|
}
|
2018-06-14 14:53:24 +00:00
|
|
|
return (0);
|
2014-11-07 19:13:19 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static int
|
2018-06-14 14:53:24 +00:00
|
|
|
me_set_tunnel(struct me_softc *sc, in_addr_t src, in_addr_t dst)
|
2014-11-07 19:13:19 +00:00
|
|
|
{
|
2018-06-14 14:53:24 +00:00
|
|
|
struct me_softc *tmp;
|
2014-11-07 19:13:19 +00:00
|
|
|
|
|
|
|
sx_assert(&me_ioctl_sx, SA_XLOCKED);
|
2018-06-14 14:53:24 +00:00
|
|
|
|
|
|
|
if (V_me_hashtbl == NULL)
|
|
|
|
V_me_hashtbl = me_hashinit();
|
|
|
|
|
|
|
|
if (sc->me_src.s_addr == src && sc->me_dst.s_addr == dst)
|
|
|
|
return (0);
|
|
|
|
|
|
|
|
CK_LIST_FOREACH(tmp, &ME_HASH(src, dst), chain) {
|
|
|
|
if (tmp == sc)
|
2014-11-07 19:13:19 +00:00
|
|
|
continue;
|
2018-06-14 14:53:24 +00:00
|
|
|
if (tmp->me_src.s_addr == src &&
|
|
|
|
tmp->me_dst.s_addr == dst)
|
2014-11-07 19:13:19 +00:00
|
|
|
return (EADDRNOTAVAIL);
|
2015-10-03 09:15:23 +00:00
|
|
|
}
|
2018-06-14 14:53:24 +00:00
|
|
|
|
|
|
|
me_delete_tunnel(sc);
|
|
|
|
sc->me_dst.s_addr = dst;
|
|
|
|
sc->me_src.s_addr = src;
|
|
|
|
CK_LIST_INSERT_HEAD(&ME_HASH(src, dst), sc, chain);
|
|
|
|
|
|
|
|
ME2IFP(sc)->if_drv_flags |= IFF_DRV_RUNNING;
|
|
|
|
if_link_state_change(ME2IFP(sc), LINK_STATE_UP);
|
2014-11-07 19:13:19 +00:00
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
2018-06-14 14:53:24 +00:00
|
|
|
me_delete_tunnel(struct me_softc *sc)
|
2014-11-07 19:13:19 +00:00
|
|
|
{
|
|
|
|
|
|
|
|
sx_assert(&me_ioctl_sx, SA_XLOCKED);
|
2018-06-14 14:53:24 +00:00
|
|
|
if (ME_READY(sc)) {
|
|
|
|
CK_LIST_REMOVE(sc, chain);
|
|
|
|
ME_WAIT();
|
|
|
|
|
|
|
|
sc->me_src.s_addr = 0;
|
|
|
|
sc->me_dst.s_addr = 0;
|
|
|
|
ME2IFP(sc)->if_drv_flags &= ~IFF_DRV_RUNNING;
|
|
|
|
if_link_state_change(ME2IFP(sc), LINK_STATE_DOWN);
|
|
|
|
}
|
2014-11-07 19:13:19 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static uint16_t
|
|
|
|
me_in_cksum(uint16_t *p, int nwords)
|
|
|
|
{
|
|
|
|
uint32_t sum = 0;
|
|
|
|
|
|
|
|
while (nwords-- > 0)
|
|
|
|
sum += *p++;
|
|
|
|
sum = (sum >> 16) + (sum & 0xffff);
|
|
|
|
sum += (sum >> 16);
|
|
|
|
return (~sum);
|
|
|
|
}
|
|
|
|
|
Rework IP encapsulation handling code.
Currently it has several disadvantages:
- it uses single mutex to protect internal structures. It is used by
data- and control- path, thus there are no parallelism at all.
- it uses single list to keep encap handlers for both INET and INET6
families.
- struct encaptab keeps unneeded information (src, dst, masks, protosw),
that isn't used by code in the source tree.
- matches are prioritized and when many tunneling interfaces are
registered, encapcheck handler of each interface is invoked for each
packet. The search takes O(n) for n interfaces. All this work is done
with exclusive lock held.
What this patch includes:
- the datapath is converted to be lockless using epoch(9) KPI.
- struct encaptab now linked using CK_LIST.
- all unused fields removed from struct encaptab. Several new fields
addedr: min_length is the minimum packet length, that encapsulation
handler expects to see; exact_match is maximum number of bits, that
can return an encapsulation handler, when it wants to consume a packet.
- IPv6 and IPv4 handlers are stored in separate lists;
- added new "encap_lookup_t" method, that will be used later. It is
targeted to speedup lookup of needed interface, when gif(4)/gre(4) have
many interfaces.
- the need to use protosw structure is eliminated. The only pr_input
method was used from this structure, so I don't see the need to keep
using it.
- encap_input_t method changed to avoid using mbuf tags to store softc
pointer. Now it is passed directly trough encap_input_t method.
encap_getarg() funtions is removed.
- all sockaddr structures and code that uses them removed. We don't have
any code in the tree that uses them. All consumers use encap_attach_func()
method, that relies on invoking of encapcheck() to determine the needed
handler.
- introduced struct encap_config, it contains parameters of encap handler
that is going to be registered by encap_attach() function.
- encap handlers are stored in lists ordered by exact_match value, thus
handlers that need more bits to match will be checked first, and if
encapcheck method returns exact_match value, the search will be stopped.
- all current consumers changed to use new KPI.
Reviewed by: mmacy
Sponsored by: Yandex LLC
Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
|
|
|
static int
|
|
|
|
me_input(struct mbuf *m, int off, int proto, void *arg)
|
2014-11-07 19:13:19 +00:00
|
|
|
{
|
Rework IP encapsulation handling code.
Currently it has several disadvantages:
- it uses single mutex to protect internal structures. It is used by
data- and control- path, thus there are no parallelism at all.
- it uses single list to keep encap handlers for both INET and INET6
families.
- struct encaptab keeps unneeded information (src, dst, masks, protosw),
that isn't used by code in the source tree.
- matches are prioritized and when many tunneling interfaces are
registered, encapcheck handler of each interface is invoked for each
packet. The search takes O(n) for n interfaces. All this work is done
with exclusive lock held.
What this patch includes:
- the datapath is converted to be lockless using epoch(9) KPI.
- struct encaptab now linked using CK_LIST.
- all unused fields removed from struct encaptab. Several new fields
addedr: min_length is the minimum packet length, that encapsulation
handler expects to see; exact_match is maximum number of bits, that
can return an encapsulation handler, when it wants to consume a packet.
- IPv6 and IPv4 handlers are stored in separate lists;
- added new "encap_lookup_t" method, that will be used later. It is
targeted to speedup lookup of needed interface, when gif(4)/gre(4) have
many interfaces.
- the need to use protosw structure is eliminated. The only pr_input
method was used from this structure, so I don't see the need to keep
using it.
- encap_input_t method changed to avoid using mbuf tags to store softc
pointer. Now it is passed directly trough encap_input_t method.
encap_getarg() funtions is removed.
- all sockaddr structures and code that uses them removed. We don't have
any code in the tree that uses them. All consumers use encap_attach_func()
method, that relies on invoking of encapcheck() to determine the needed
handler.
- introduced struct encap_config, it contains parameters of encap handler
that is going to be registered by encap_attach() function.
- encap handlers are stored in lists ordered by exact_match value, thus
handlers that need more bits to match will be checked first, and if
encapcheck method returns exact_match value, the search will be stopped.
- all current consumers changed to use new KPI.
Reviewed by: mmacy
Sponsored by: Yandex LLC
Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
|
|
|
struct me_softc *sc = arg;
|
2014-11-07 19:13:19 +00:00
|
|
|
struct mobhdr *mh;
|
|
|
|
struct ifnet *ifp;
|
|
|
|
struct ip *ip;
|
|
|
|
int hlen;
|
|
|
|
|
|
|
|
ifp = ME2IFP(sc);
|
|
|
|
/* checks for short packets */
|
|
|
|
hlen = sizeof(struct mobhdr);
|
|
|
|
if (m->m_pkthdr.len < sizeof(struct ip) + hlen)
|
|
|
|
hlen -= sizeof(struct in_addr);
|
|
|
|
if (m->m_len < sizeof(struct ip) + hlen)
|
|
|
|
m = m_pullup(m, sizeof(struct ip) + hlen);
|
|
|
|
if (m == NULL)
|
|
|
|
goto drop;
|
|
|
|
mh = (struct mobhdr *)mtodo(m, sizeof(struct ip));
|
|
|
|
/* check for wrong flags */
|
|
|
|
if (mh->mob_flags & (~MOB_FLAGS_SP)) {
|
|
|
|
m_freem(m);
|
|
|
|
goto drop;
|
|
|
|
}
|
|
|
|
if (mh->mob_flags) {
|
|
|
|
if (hlen != sizeof(struct mobhdr)) {
|
|
|
|
m_freem(m);
|
|
|
|
goto drop;
|
|
|
|
}
|
|
|
|
} else
|
|
|
|
hlen = sizeof(struct mobhdr) - sizeof(struct in_addr);
|
|
|
|
/* check mobile header checksum */
|
|
|
|
if (me_in_cksum((uint16_t *)mh, hlen / sizeof(uint16_t)) != 0) {
|
|
|
|
m_freem(m);
|
|
|
|
goto drop;
|
|
|
|
}
|
|
|
|
#ifdef MAC
|
|
|
|
mac_ifnet_create_mbuf(ifp, m);
|
|
|
|
#endif
|
|
|
|
ip = mtod(m, struct ip *);
|
|
|
|
ip->ip_dst = mh->mob_dst;
|
|
|
|
ip->ip_p = mh->mob_proto;
|
|
|
|
ip->ip_sum = 0;
|
|
|
|
ip->ip_len = htons(m->m_pkthdr.len - hlen);
|
|
|
|
if (mh->mob_flags)
|
|
|
|
ip->ip_src = mh->mob_src;
|
|
|
|
memmove(mtodo(m, hlen), ip, sizeof(struct ip));
|
|
|
|
m_adj(m, hlen);
|
|
|
|
m_clrprotoflags(m);
|
|
|
|
m->m_pkthdr.rcvif = ifp;
|
|
|
|
m->m_pkthdr.csum_flags |= (CSUM_IP_CHECKED | CSUM_IP_VALID);
|
2015-05-12 07:37:27 +00:00
|
|
|
M_SETFIB(m, ifp->if_fib);
|
2014-11-07 19:13:19 +00:00
|
|
|
hlen = AF_INET;
|
|
|
|
BPF_MTAP2(ifp, &hlen, sizeof(hlen), m);
|
|
|
|
if_inc_counter(ifp, IFCOUNTER_IPACKETS, 1);
|
|
|
|
if_inc_counter(ifp, IFCOUNTER_IBYTES, m->m_pkthdr.len);
|
|
|
|
if ((ifp->if_flags & IFF_MONITOR) != 0)
|
|
|
|
m_freem(m);
|
|
|
|
else
|
|
|
|
netisr_dispatch(NETISR_IP, m);
|
|
|
|
return (IPPROTO_DONE);
|
|
|
|
drop:
|
|
|
|
if_inc_counter(ifp, IFCOUNTER_IERRORS, 1);
|
|
|
|
return (IPPROTO_DONE);
|
|
|
|
}
|
|
|
|
|
|
|
|
#define MTAG_ME 1414491977
|
|
|
|
static int
|
|
|
|
me_check_nesting(struct ifnet *ifp, struct mbuf *m)
|
|
|
|
{
|
|
|
|
struct m_tag *mtag;
|
|
|
|
int count;
|
|
|
|
|
|
|
|
count = 1;
|
|
|
|
mtag = NULL;
|
2015-05-06 14:02:57 +00:00
|
|
|
while ((mtag = m_tag_locate(m, MTAG_ME, 0, mtag)) != NULL) {
|
2014-11-07 19:13:19 +00:00
|
|
|
if (*(struct ifnet **)(mtag + 1) == ifp) {
|
|
|
|
log(LOG_NOTICE, "%s: loop detected\n", ifp->if_xname);
|
|
|
|
return (EIO);
|
|
|
|
}
|
|
|
|
count++;
|
|
|
|
}
|
|
|
|
if (count > V_max_me_nesting) {
|
|
|
|
log(LOG_NOTICE,
|
|
|
|
"%s: if_output recursively called too many times(%d)\n",
|
|
|
|
ifp->if_xname, count);
|
|
|
|
return (EIO);
|
|
|
|
}
|
|
|
|
mtag = m_tag_alloc(MTAG_ME, 0, sizeof(struct ifnet *), M_NOWAIT);
|
|
|
|
if (mtag == NULL)
|
|
|
|
return (ENOMEM);
|
|
|
|
*(struct ifnet **)(mtag + 1) = ifp;
|
|
|
|
m_tag_prepend(m, mtag);
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int
|
|
|
|
me_output(struct ifnet *ifp, struct mbuf *m, const struct sockaddr *dst,
|
2018-06-14 14:53:24 +00:00
|
|
|
struct route *ro __unused)
|
2014-11-07 19:13:19 +00:00
|
|
|
{
|
|
|
|
uint32_t af;
|
|
|
|
|
|
|
|
if (dst->sa_family == AF_UNSPEC)
|
|
|
|
bcopy(dst->sa_data, &af, sizeof(af));
|
|
|
|
else
|
|
|
|
af = dst->sa_family;
|
2018-06-14 14:53:24 +00:00
|
|
|
m->m_pkthdr.csum_data = af;
|
2014-11-07 19:13:19 +00:00
|
|
|
return (ifp->if_transmit(ifp, m));
|
|
|
|
}
|
|
|
|
|
|
|
|
static int
|
|
|
|
me_transmit(struct ifnet *ifp, struct mbuf *m)
|
|
|
|
{
|
|
|
|
struct mobhdr mh;
|
|
|
|
struct me_softc *sc;
|
|
|
|
struct ip *ip;
|
2018-06-14 14:53:24 +00:00
|
|
|
uint32_t af;
|
2014-11-07 19:13:19 +00:00
|
|
|
int error, hlen, plen;
|
|
|
|
|
2018-06-14 14:53:24 +00:00
|
|
|
#ifdef MAC
|
|
|
|
error = mac_ifnet_check_transmit(ifp, m);
|
|
|
|
if (error != 0)
|
|
|
|
goto drop;
|
|
|
|
#endif
|
|
|
|
error = ENETDOWN;
|
|
|
|
ME_RLOCK();
|
2014-11-07 19:13:19 +00:00
|
|
|
sc = ifp->if_softc;
|
2018-06-14 14:53:24 +00:00
|
|
|
if (sc == NULL || !ME_READY(sc) ||
|
|
|
|
(ifp->if_flags & IFF_MONITOR) != 0 ||
|
|
|
|
(ifp->if_flags & IFF_UP) == 0 ||
|
|
|
|
(error = me_check_nesting(ifp, m) != 0)) {
|
|
|
|
m_freem(m);
|
|
|
|
goto drop;
|
|
|
|
}
|
|
|
|
af = m->m_pkthdr.csum_data;
|
|
|
|
if (af != AF_INET) {
|
|
|
|
error = EAFNOSUPPORT;
|
2014-11-07 19:13:19 +00:00
|
|
|
m_freem(m);
|
|
|
|
goto drop;
|
|
|
|
}
|
|
|
|
if (m->m_len < sizeof(struct ip))
|
|
|
|
m = m_pullup(m, sizeof(struct ip));
|
|
|
|
if (m == NULL) {
|
|
|
|
error = ENOBUFS;
|
|
|
|
goto drop;
|
|
|
|
}
|
|
|
|
ip = mtod(m, struct ip *);
|
|
|
|
/* Fragmented datagramms shouldn't be encapsulated */
|
|
|
|
if (ip->ip_off & htons(IP_MF | IP_OFFMASK)) {
|
|
|
|
error = EINVAL;
|
|
|
|
m_freem(m);
|
|
|
|
goto drop;
|
|
|
|
}
|
|
|
|
mh.mob_proto = ip->ip_p;
|
|
|
|
mh.mob_src = ip->ip_src;
|
|
|
|
mh.mob_dst = ip->ip_dst;
|
|
|
|
if (in_hosteq(sc->me_src, ip->ip_src)) {
|
|
|
|
hlen = sizeof(struct mobhdr) - sizeof(struct in_addr);
|
|
|
|
mh.mob_flags = 0;
|
|
|
|
} else {
|
|
|
|
hlen = sizeof(struct mobhdr);
|
|
|
|
mh.mob_flags = MOB_FLAGS_SP;
|
|
|
|
}
|
|
|
|
plen = m->m_pkthdr.len;
|
|
|
|
ip->ip_src = sc->me_src;
|
|
|
|
ip->ip_dst = sc->me_dst;
|
2018-06-14 14:53:24 +00:00
|
|
|
m->m_flags &= ~(M_BCAST|M_MCAST);
|
2014-11-07 19:13:19 +00:00
|
|
|
M_SETFIB(m, sc->me_fibnum);
|
|
|
|
M_PREPEND(m, hlen, M_NOWAIT);
|
|
|
|
if (m == NULL) {
|
|
|
|
error = ENOBUFS;
|
|
|
|
goto drop;
|
|
|
|
}
|
|
|
|
if (m->m_len < sizeof(struct ip) + hlen)
|
|
|
|
m = m_pullup(m, sizeof(struct ip) + hlen);
|
|
|
|
if (m == NULL) {
|
|
|
|
error = ENOBUFS;
|
|
|
|
goto drop;
|
|
|
|
}
|
|
|
|
memmove(mtod(m, void *), mtodo(m, hlen), sizeof(struct ip));
|
|
|
|
ip = mtod(m, struct ip *);
|
|
|
|
ip->ip_len = htons(m->m_pkthdr.len);
|
|
|
|
ip->ip_p = IPPROTO_MOBILE;
|
|
|
|
ip->ip_sum = 0;
|
|
|
|
mh.mob_csum = 0;
|
|
|
|
mh.mob_csum = me_in_cksum((uint16_t *)&mh, hlen / sizeof(uint16_t));
|
|
|
|
bcopy(&mh, mtodo(m, sizeof(struct ip)), hlen);
|
|
|
|
error = ip_output(m, NULL, NULL, IP_FORWARDING, NULL, NULL);
|
|
|
|
drop:
|
|
|
|
if (error)
|
|
|
|
if_inc_counter(ifp, IFCOUNTER_OERRORS, 1);
|
|
|
|
else {
|
|
|
|
if_inc_counter(ifp, IFCOUNTER_OPACKETS, 1);
|
|
|
|
if_inc_counter(ifp, IFCOUNTER_OBYTES, plen);
|
|
|
|
}
|
2018-06-14 14:53:24 +00:00
|
|
|
ME_RUNLOCK();
|
2014-11-07 19:13:19 +00:00
|
|
|
return (error);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
|
|
|
me_qflush(struct ifnet *ifp __unused)
|
|
|
|
{
|
|
|
|
|
|
|
|
}
|
|
|
|
|
2018-06-14 14:53:24 +00:00
|
|
|
static const struct encaptab *ecookie = NULL;
|
|
|
|
static const struct encap_config me_encap_cfg = {
|
|
|
|
.proto = IPPROTO_MOBILE,
|
|
|
|
.min_length = sizeof(struct ip) + sizeof(struct mobhdr) -
|
|
|
|
sizeof(in_addr_t),
|
|
|
|
.exact_match = ENCAP_DRV_LOOKUP,
|
|
|
|
.lookup = me_lookup,
|
|
|
|
.input = me_input
|
|
|
|
};
|
|
|
|
|
2014-11-07 19:13:19 +00:00
|
|
|
static int
|
|
|
|
memodevent(module_t mod, int type, void *data)
|
|
|
|
{
|
|
|
|
|
|
|
|
switch (type) {
|
|
|
|
case MOD_LOAD:
|
2018-06-14 14:53:24 +00:00
|
|
|
ecookie = ip_encap_attach(&me_encap_cfg, NULL, M_WAITOK);
|
|
|
|
break;
|
2014-11-07 19:13:19 +00:00
|
|
|
case MOD_UNLOAD:
|
2018-06-14 14:53:24 +00:00
|
|
|
ip_encap_detach(ecookie);
|
2014-11-07 19:13:19 +00:00
|
|
|
break;
|
|
|
|
default:
|
|
|
|
return (EOPNOTSUPP);
|
|
|
|
}
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
|
|
|
|
static moduledata_t me_mod = {
|
|
|
|
"if_me",
|
|
|
|
memodevent,
|
|
|
|
0
|
|
|
|
};
|
|
|
|
|
|
|
|
DECLARE_MODULE(if_me, me_mod, SI_SUB_PSEUDO, SI_ORDER_ANY);
|
|
|
|
MODULE_VERSION(if_me, 1);
|