freebsd-dev/sys/net/if_ef.c
sobomax 3ce7730799 When neither of supported frame type is enabled via kernel options enable
them all, otherwise the driver will be useless and will only confuse user
as manual page says nothing about the need to enable one of those frame
types explicitly in the kernel config.

PR:		kern/47152
Submitted by:	Andriy Gapon <avg@icyb.net.ua>
MFC after:	3 days
2005-03-06 23:03:58 +00:00

607 lines
13 KiB
C

/*-
* Copyright (c) 1999, 2000 Boris Popov
* 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 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 AUTHOR 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.
*
* $FreeBSD$
*/
#include "opt_inet.h"
#include "opt_ipx.h"
#include "opt_ef.h"
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/sockio.h>
#include <sys/malloc.h>
#include <sys/mbuf.h>
#include <sys/socket.h>
#include <sys/syslog.h>
#include <sys/kernel.h>
#include <sys/module.h>
#include <net/ethernet.h>
#include <net/if_llc.h>
#include <net/if.h>
#include <net/if_arp.h>
#include <net/if_dl.h>
#include <net/if_types.h>
#include <net/netisr.h>
#include <net/route.h>
#include <net/bpf.h>
#ifdef INET
#include <netinet/in.h>
#include <netinet/in_var.h>
#include <netinet/if_ether.h>
#endif
#ifdef IPX
#include <netipx/ipx.h>
#include <netipx/ipx_if.h>
#endif
/* If none of the supported layers is enabled explicitly enable them all */
#if !defined(ETHER_II) && !defined(ETHER_8023) && !defined(ETHER_8022) && \
!defined(ETHER_SNAP)
#define ETHER_II 1
#define ETHER_8023 1
#define ETHER_8022 1
#define ETHER_SNAP 1
#endif
/* internal frame types */
#define ETHER_FT_EII 0 /* Ethernet_II - default */
#define ETHER_FT_8023 1 /* 802.3 (Novell) */
#define ETHER_FT_8022 2 /* 802.2 */
#define ETHER_FT_SNAP 3 /* SNAP */
#define EF_NFT 4 /* total number of frame types */
#ifdef EF_DEBUG
#define EFDEBUG(format, args...) printf("%s: "format, __func__ ,## args)
#else
#define EFDEBUG(format, args...)
#endif
#define EFERROR(format, args...) printf("%s: "format, __func__ ,## args)
struct efnet {
struct arpcom ef_ac;
struct ifnet * ef_ifp;
int ef_frametype;
};
struct ef_link {
SLIST_ENTRY(ef_link) el_next;
struct ifnet *el_ifp; /* raw device for this clones */
struct efnet *el_units[EF_NFT]; /* our clones */
};
static SLIST_HEAD(ef_link_head, ef_link) efdev = {NULL};
static int efcount;
extern int (*ef_inputp)(struct ifnet*, struct ether_header *eh, struct mbuf *m);
extern int (*ef_outputp)(struct ifnet *ifp, struct mbuf **mp,
struct sockaddr *dst, short *tp, int *hlen);
/*
static void ef_reset (struct ifnet *);
*/
static int ef_attach(struct efnet *sc);
static int ef_detach(struct efnet *sc);
static void ef_init(void *);
static int ef_ioctl(struct ifnet *, u_long, caddr_t);
static void ef_start(struct ifnet *);
static int ef_input(struct ifnet*, struct ether_header *, struct mbuf *);
static int ef_output(struct ifnet *ifp, struct mbuf **mp,
struct sockaddr *dst, short *tp, int *hlen);
static int ef_load(void);
static int ef_unload(void);
/*
* Install the interface, most of structure initialization done in ef_clone()
*/
static int
ef_attach(struct efnet *sc)
{
struct ifnet *ifp = (struct ifnet*)&sc->ef_ac.ac_if;
struct ifaddr *ifa2;
struct sockaddr_dl *sdl2;
ifp->if_start = ef_start;
ifp->if_watchdog = NULL;
ifp->if_init = ef_init;
ifp->if_snd.ifq_maxlen = IFQ_MAXLEN;
ifp->if_flags = (IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST);
/*
* Attach the interface
*/
ifa2 = ifaddr_byindex(sc->ef_ifp->if_index);
sdl2 = (struct sockaddr_dl *)ifa2->ifa_addr;
ether_ifattach(ifp, LLADDR(sdl2));
ifp->if_resolvemulti = 0;
ifp->if_type = IFT_XETHER;
ifp->if_flags |= IFF_RUNNING;
bcopy(LLADDR(sdl2), sc->ef_ac.ac_enaddr, ETHER_ADDR_LEN);
EFDEBUG("%s: attached\n", ifp->if_xname);
return 1;
}
/*
* This is for _testing_only_, just removes interface from interfaces list
*/
static int
ef_detach(struct efnet *sc)
{
struct ifnet *ifp = (struct ifnet*)&sc->ef_ac.ac_if;
int s;
s = splimp();
if (ifp->if_flags & IFF_UP) {
if_down(ifp);
if (ifp->if_flags & IFF_RUNNING) {
/* find internet addresses and delete routes */
register struct ifaddr *ifa;
TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) {
rtinit(ifa, (int)RTM_DELETE, 0);
}
}
}
IFNET_WLOCK();
TAILQ_REMOVE(&ifnet, ifp, if_link);
IFNET_WUNLOCK();
splx(s);
return 0;
}
static void
ef_init(void *foo) {
return;
}
static int
ef_ioctl(struct ifnet *ifp, u_long cmd, caddr_t data)
{
struct efnet *sc = ifp->if_softc;
struct ifaddr *ifa = (struct ifaddr*)data;
int s, error;
EFDEBUG("IOCTL %ld for %s\n", cmd, ifp->if_xname);
error = 0;
s = splimp();
switch (cmd) {
case SIOCSIFFLAGS:
error = 0;
break;
case SIOCSIFADDR:
if (sc->ef_frametype == ETHER_FT_8023 &&
ifa->ifa_addr->sa_family != AF_IPX) {
error = EAFNOSUPPORT;
break;
}
ifp->if_flags |= IFF_UP;
/* FALL THROUGH */
default:
error = ether_ioctl(ifp, cmd, data);
break;
}
splx(s);
return error;
}
/*
* Currently packet prepared in the ether_output(), but this can be a better
* place.
*/
static void
ef_start(struct ifnet *ifp)
{
struct efnet *sc = (struct efnet*)ifp->if_softc;
struct ifnet *p;
struct mbuf *m;
int error;
ifp->if_flags |= IFF_OACTIVE;
p = sc->ef_ifp;
EFDEBUG("\n");
for (;;) {
IF_DEQUEUE(&ifp->if_snd, m);
if (m == 0)
break;
BPF_MTAP(ifp, m);
IFQ_HANDOFF(p, m, error);
if (error) {
ifp->if_oerrors++;
continue;
}
ifp->if_opackets++;
}
ifp->if_flags &= ~IFF_OACTIVE;
return;
}
/*
* Inline functions do not put additional overhead to procedure call or
* parameter passing but simplify the code
*/
static int __inline
ef_inputEII(struct mbuf *m, struct ether_header *eh, u_short ether_type)
{
int isr;
switch(ether_type) {
#ifdef IPX
case ETHERTYPE_IPX:
isr = NETISR_IPX;
break;
#endif
#ifdef INET
case ETHERTYPE_IP:
if (ip_fastforward(m))
return (0);
isr = NETISR_IP;
break;
case ETHERTYPE_ARP:
isr = NETISR_ARP;
break;
#endif
default:
return (EPROTONOSUPPORT);
}
netisr_dispatch(isr, m);
return (0);
}
static int __inline
ef_inputSNAP(struct mbuf *m, struct ether_header *eh, struct llc* l,
u_short ether_type)
{
int isr;
switch(ether_type) {
#ifdef IPX
case ETHERTYPE_IPX:
m_adj(m, 8);
isr = NETISR_IPX;
break;
#endif
default:
return (EPROTONOSUPPORT);
}
netisr_dispatch(isr, m);
return (0);
}
static int __inline
ef_input8022(struct mbuf *m, struct ether_header *eh, struct llc* l,
u_short ether_type)
{
int isr;
switch(ether_type) {
#ifdef IPX
case 0xe0:
m_adj(m, 3);
isr = NETISR_IPX;
break;
#endif
default:
return (EPROTONOSUPPORT);
}
netisr_dispatch(isr, m);
return (0);
}
/*
* Called from ether_input()
*/
static int
ef_input(struct ifnet *ifp, struct ether_header *eh, struct mbuf *m)
{
u_short ether_type;
int ft = -1;
struct efnet *efp;
struct ifnet *eifp;
struct llc *l;
struct ef_link *efl;
int isr;
ether_type = ntohs(eh->ether_type);
l = NULL;
if (ether_type < ETHERMTU) {
l = mtod(m, struct llc*);
if (l->llc_dsap == 0xff && l->llc_ssap == 0xff) {
/*
* Novell's "802.3" frame
*/
ft = ETHER_FT_8023;
} else if (l->llc_dsap == 0xaa && l->llc_ssap == 0xaa) {
/*
* 802.2/SNAP
*/
ft = ETHER_FT_SNAP;
ether_type = ntohs(l->llc_un.type_snap.ether_type);
} else if (l->llc_dsap == l->llc_ssap) {
/*
* 802.3/802.2
*/
ft = ETHER_FT_8022;
ether_type = l->llc_ssap;
}
} else
ft = ETHER_FT_EII;
if (ft == -1) {
EFDEBUG("Unrecognised ether_type %x\n", ether_type);
return EPROTONOSUPPORT;
}
/*
* Check if interface configured for the given frame
*/
efp = NULL;
SLIST_FOREACH(efl, &efdev, el_next) {
if (efl->el_ifp == ifp) {
efp = efl->el_units[ft];
break;
}
}
if (efp == NULL) {
EFDEBUG("Can't find if for %d\n", ft);
return EPROTONOSUPPORT;
}
eifp = &efp->ef_ac.ac_if;
if ((eifp->if_flags & IFF_UP) == 0)
return EPROTONOSUPPORT;
eifp->if_ibytes += m->m_pkthdr.len + sizeof (*eh);
m->m_pkthdr.rcvif = eifp;
BPF_MTAP2(eifp, eh, ETHER_HDR_LEN, m);
/*
* Now we ready to adjust mbufs and pass them to protocol intr's
*/
switch(ft) {
case ETHER_FT_EII:
return (ef_inputEII(m, eh, ether_type));
#ifdef IPX
case ETHER_FT_8023: /* only IPX can be here */
isr = NETISR_IPX;
break;
#endif
case ETHER_FT_SNAP:
return (ef_inputSNAP(m, eh, l, ether_type));
case ETHER_FT_8022:
return (ef_input8022(m, eh, l, ether_type));
default:
EFDEBUG("No support for frame %d and proto %04x\n",
ft, ether_type);
return (EPROTONOSUPPORT);
}
netisr_dispatch(isr, m);
return (0);
}
static int
ef_output(struct ifnet *ifp, struct mbuf **mp, struct sockaddr *dst, short *tp,
int *hlen)
{
struct efnet *sc = (struct efnet*)ifp->if_softc;
struct mbuf *m = *mp;
u_char *cp;
short type;
if (ifp->if_type != IFT_XETHER)
return ENETDOWN;
switch (sc->ef_frametype) {
case ETHER_FT_EII:
#ifdef IPX
type = htons(ETHERTYPE_IPX);
#else
return EPFNOSUPPORT;
#endif
break;
case ETHER_FT_8023:
type = htons(m->m_pkthdr.len);
break;
case ETHER_FT_8022:
M_PREPEND(m, ETHER_HDR_LEN + 3, M_TRYWAIT);
if (m == NULL) {
*mp = NULL;
return ENOBUFS;
}
/*
* Ensure that ethernet header and next three bytes
* will fit into single mbuf
*/
m = m_pullup(m, ETHER_HDR_LEN + 3);
if (m == NULL) {
*mp = NULL;
return ENOBUFS;
}
m_adj(m, ETHER_HDR_LEN);
type = htons(m->m_pkthdr.len);
cp = mtod(m, u_char *);
*cp++ = 0xE0;
*cp++ = 0xE0;
*cp++ = 0x03;
*hlen += 3;
break;
case ETHER_FT_SNAP:
M_PREPEND(m, 8, M_TRYWAIT);
if (m == NULL) {
*mp = NULL;
return ENOBUFS;
}
type = htons(m->m_pkthdr.len);
cp = mtod(m, u_char *);
bcopy("\xAA\xAA\x03\x00\x00\x00\x81\x37", cp, 8);
*hlen += 8;
break;
default:
return EPFNOSUPPORT;
}
*mp = m;
*tp = type;
return 0;
}
/*
* Create clone from the given interface
*/
static int
ef_clone(struct ef_link *efl, int ft)
{
struct efnet *efp;
struct ifnet *eifp;
struct ifnet *ifp = efl->el_ifp;
efp = (struct efnet*)malloc(sizeof(struct efnet), M_IFADDR,
M_WAITOK | M_ZERO);
if (efp == NULL)
return ENOMEM;
efp->ef_ifp = ifp;
efp->ef_frametype = ft;
eifp = &efp->ef_ac.ac_if;
snprintf(eifp->if_xname, IFNAMSIZ,
"%sf%d", ifp->if_xname, efp->ef_frametype);
eifp->if_dname = "ef";
eifp->if_dunit = IF_DUNIT_NONE;
eifp->if_softc = efp;
if (ifp->if_ioctl)
eifp->if_ioctl = ef_ioctl;
efl->el_units[ft] = efp;
return 0;
}
static int
ef_load(void)
{
struct ifnet *ifp;
struct efnet *efp;
struct ef_link *efl = NULL;
int error = 0, d;
IFNET_RLOCK();
TAILQ_FOREACH(ifp, &ifnet, if_link) {
if (ifp->if_type != IFT_ETHER) continue;
EFDEBUG("Found interface %s\n", ifp->if_xname);
efl = (struct ef_link*)malloc(sizeof(struct ef_link),
M_IFADDR, M_WAITOK | M_ZERO);
if (efl == NULL) {
error = ENOMEM;
break;
}
efl->el_ifp = ifp;
#ifdef ETHER_II
error = ef_clone(efl, ETHER_FT_EII);
if (error) break;
#endif
#ifdef ETHER_8023
error = ef_clone(efl, ETHER_FT_8023);
if (error) break;
#endif
#ifdef ETHER_8022
error = ef_clone(efl, ETHER_FT_8022);
if (error) break;
#endif
#ifdef ETHER_SNAP
error = ef_clone(efl, ETHER_FT_SNAP);
if (error) break;
#endif
efcount++;
SLIST_INSERT_HEAD(&efdev, efl, el_next);
}
IFNET_RUNLOCK();
if (error) {
if (efl)
SLIST_INSERT_HEAD(&efdev, efl, el_next);
SLIST_FOREACH(efl, &efdev, el_next) {
for (d = 0; d < EF_NFT; d++)
if (efl->el_units[d])
free(efl->el_units[d], M_IFADDR);
free(efl, M_IFADDR);
}
return error;
}
SLIST_FOREACH(efl, &efdev, el_next) {
for (d = 0; d < EF_NFT; d++) {
efp = efl->el_units[d];
if (efp)
ef_attach(efp);
}
}
ef_inputp = ef_input;
ef_outputp = ef_output;
EFDEBUG("Loaded\n");
return 0;
}
static int
ef_unload(void)
{
struct efnet *efp;
struct ef_link *efl;
int d;
ef_inputp = NULL;
ef_outputp = NULL;
SLIST_FOREACH(efl, &efdev, el_next) {
for (d = 0; d < EF_NFT; d++) {
efp = efl->el_units[d];
if (efp) {
ef_detach(efp);
}
}
}
EFDEBUG("Unloaded\n");
return 0;
}
static int
if_ef_modevent(module_t mod, int type, void *data)
{
switch ((modeventtype_t)type) {
case MOD_LOAD:
return ef_load();
case MOD_UNLOAD:
return ef_unload();
default:
return EOPNOTSUPP;
}
return 0;
}
static moduledata_t if_ef_mod = {
"if_ef", if_ef_modevent, NULL
};
DECLARE_MODULE(if_ef, if_ef_mod, SI_SUB_PSEUDO, SI_ORDER_MIDDLE);