2 * Copyright (C) 1999-2000 by Maksim Yevmenkin <m_evmenkin@yahoo.com>
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
8 * 1. Redistributions of source code must retain the above copyright
9 * notice, this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright
11 * notice, this list of conditions and the following disclaimer in the
12 * documentation and/or other materials provided with the distribution.
14 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
15 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
16 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
17 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
18 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
19 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
20 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
21 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
22 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
23 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
27 * -------------------------------------------------------------------------
29 * Copyright (c) 1988, Julian Onions <jpo@cs.nott.ac.uk>
30 * Nottingham University 1987.
34 * $FreeBSD: src/sys/net/if_tap.c,v 1.3.2.3 2002/04/14 21:41:48 luigi Exp $
35 * $DragonFly: src/sys/net/tap/if_tap.c,v 1.5 2003/07/19 21:14:43 dillon Exp $
36 * $Id: if_tap.c,v 0.21 2000/07/23 21:46:02 max Exp $
41 #include <sys/param.h>
43 #include <sys/filedesc.h>
44 #include <sys/filio.h>
45 #include <sys/kernel.h>
46 #include <sys/malloc.h>
50 #include <sys/signalvar.h>
51 #include <sys/socket.h>
52 #include <sys/sockio.h>
53 #include <sys/sysctl.h>
54 #include <sys/systm.h>
55 #include <sys/ttycom.h>
57 #include <sys/vnode.h>
60 #include <net/ethernet.h>
62 #include <net/if_arp.h>
63 #include <net/route.h>
65 #include <netinet/in.h>
67 #include <net/if_tapvar.h>
68 #include <net/if_tap.h>
71 #define CDEV_NAME "tap"
72 #define CDEV_MAJOR 149
73 #define TAPDEBUG if (tapdebug) printf
77 #define VMNET_DEV_MASK 0x00010000
80 static int tapmodevent __P((module_t, int, void *));
83 static void tapcreate __P((dev_t));
85 /* network interface */
86 static void tapifstart __P((struct ifnet *));
87 static int tapifioctl __P((struct ifnet *, u_long, caddr_t));
88 static void tapifinit __P((void *));
90 /* character device */
91 static d_open_t tapopen;
92 static d_close_t tapclose;
93 static d_read_t tapread;
94 static d_write_t tapwrite;
95 static d_ioctl_t tapioctl;
96 static d_poll_t tappoll;
98 static struct cdevsw tap_cdevsw = {
100 /* close */ tapclose,
102 /* write */ tapwrite,
103 /* ioctl */ tapioctl,
106 /* startegy */ nostrategy,
107 /* dev name */ CDEV_NAME,
108 /* dev major */ CDEV_MAJOR,
115 static int taprefcnt = 0; /* module ref. counter */
116 static int taplastunit = -1; /* max. open unit number */
117 static int tapdebug = 0; /* debug flag */
119 MALLOC_DECLARE(M_TAP);
120 MALLOC_DEFINE(M_TAP, CDEV_NAME, "Ethernet tunnel interface");
121 SYSCTL_INT(_debug, OID_AUTO, if_tap_debug, CTLFLAG_RW, &tapdebug, 0, "");
122 DEV_MODULE(if_tap, tapmodevent, NULL);
127 * module event handler
130 tapmodevent(mod, type, data)
135 static int attached = 0;
136 struct ifnet *ifp = NULL;
144 cdevsw_add(&tap_cdevsw);
152 cdevsw_remove(&tap_cdevsw);
155 while (unit <= taplastunit) {
157 TAILQ_FOREACH(ifp, &ifnet, if_link)
158 if ((strcmp(ifp->if_name, TAP) == 0) ||
159 (strcmp(ifp->if_name, VMNET) == 0))
160 if (ifp->if_unit == unit)
165 struct tap_softc *tp = ifp->if_softc;
167 TAPDEBUG("detaching %s%d. minor = %#x, " \
168 "taplastunit = %d\n",
169 ifp->if_name, unit, minor(tp->tap_dev),
173 ether_ifdetach(ifp, 1);
175 destroy_dev(tp->tap_dev);
196 * to create interface
202 struct ifnet *ifp = NULL;
203 struct tap_softc *tp = NULL;
204 unsigned short macaddr_hi;
208 /* allocate driver storage and create device */
209 MALLOC(tp, struct tap_softc *, sizeof(*tp), M_TAP, M_WAITOK);
210 bzero(tp, sizeof(*tp));
212 /* select device: tap or vmnet */
213 if (minor(dev) & VMNET_DEV_MASK) {
215 unit = lminor(dev) & 0xff;
216 tp->tap_flags |= TAP_VMNET;
223 tp->tap_dev = make_dev(&tap_cdevsw, minor(dev), UID_ROOT, GID_WHEEL,
224 0600, "%s%d", name, unit);
225 tp->tap_dev->si_drv1 = dev->si_drv1 = tp;
227 /* generate fake MAC address: 00 bd xx xx xx unit_no */
228 macaddr_hi = htons(0x00bd);
229 bcopy(&macaddr_hi, &tp->arpcom.ac_enaddr[0], sizeof(short));
230 bcopy(&ticks, &tp->arpcom.ac_enaddr[2], sizeof(long));
231 tp->arpcom.ac_enaddr[5] = (u_char)unit;
233 /* fill the rest and attach interface */
238 if (unit > taplastunit)
242 ifp->if_init = tapifinit;
243 ifp->if_output = ether_output;
244 ifp->if_start = tapifstart;
245 ifp->if_ioctl = tapifioctl;
246 ifp->if_mtu = ETHERMTU;
247 ifp->if_flags = (IFF_BROADCAST|IFF_SIMPLEX|IFF_MULTICAST);
248 ifp->if_snd.ifq_maxlen = ifqmaxlen;
251 ether_ifattach(ifp, 1);
254 tp->tap_flags |= TAP_INITED;
256 TAPDEBUG("interface %s%d created. minor = %#x\n",
257 ifp->if_name, ifp->if_unit, minor(tp->tap_dev));
264 * to open tunnel. must be superuser
267 tapopen(dev_t dev, int flag, int mode, d_thread_t *td)
269 struct tap_softc *tp = NULL;
274 if ((error = suser(td)) != 0)
283 if (tp->tap_flags & TAP_OPEN)
286 bcopy(tp->arpcom.ac_enaddr, tp->ether_addr, sizeof(tp->ether_addr));
289 tp->tap_flags |= TAP_OPEN;
292 TAPDEBUG("%s%d is open. minor = %#x, refcnt = %d, taplastunit = %d\n",
293 tp->tap_if.if_name, tp->tap_if.if_unit,
294 minor(tp->tap_dev), taprefcnt, taplastunit);
303 * close the device - mark i/f down & delete routing info
306 tapclose(dev_t dev, int foo, int bar, d_thread_t *td)
309 struct tap_softc *tp = dev->si_drv1;
310 struct ifnet *ifp = &tp->tap_if;
311 struct mbuf *m = NULL;
313 /* junk all pending output */
317 IF_DEQUEUE(&ifp->if_snd, m);
324 * do not bring the interface down, and do not anything with
325 * interface, if we are in VMnet mode. just close the device.
328 if (((tp->tap_flags & TAP_VMNET) == 0) && (ifp->if_flags & IFF_UP)) {
331 if (ifp->if_flags & IFF_RUNNING) {
332 /* find internet addresses and delete routes */
333 struct ifaddr *ifa = NULL;
335 TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) {
336 if (ifa->ifa_addr->sa_family == AF_INET) {
337 rtinit(ifa, (int)RTM_DELETE, 0);
339 /* remove address from interface */
341 sizeof(*(ifa->ifa_addr)));
342 bzero(ifa->ifa_dstaddr,
343 sizeof(*(ifa->ifa_dstaddr)));
344 bzero(ifa->ifa_netmask,
345 sizeof(*(ifa->ifa_netmask)));
349 ifp->if_flags &= ~IFF_RUNNING;
354 funsetown(tp->tap_sigio);
355 selwakeup(&tp->tap_rsel);
357 tp->tap_flags &= ~TAP_OPEN;
363 printf("%s%d minor = %#x, refcnt = %d is out of sync. " \
364 "set refcnt to 0\n", ifp->if_name, ifp->if_unit,
365 minor(tp->tap_dev), taprefcnt);
368 TAPDEBUG("%s%d is closed. minor = %#x, refcnt = %d, taplastunit = %d\n",
369 ifp->if_name, ifp->if_unit, minor(tp->tap_dev),
370 taprefcnt, taplastunit);
379 * network interface initialization function
385 struct tap_softc *tp = (struct tap_softc *)xtp;
386 struct ifnet *ifp = &tp->tap_if;
388 TAPDEBUG("initializing %s%d, minor = %#x\n",
389 ifp->if_name, ifp->if_unit, minor(tp->tap_dev));
391 ifp->if_flags |= IFF_RUNNING;
392 ifp->if_flags &= ~IFF_OACTIVE;
394 /* attempt to start output */
402 * Process an ioctl request on network interface
405 tapifioctl(ifp, cmd, data)
410 struct tap_softc *tp = (struct tap_softc *)(ifp->if_softc);
411 struct ifstat *ifs = NULL;
419 dummy = ether_ioctl(ifp, cmd, data);
423 case SIOCSIFFLAGS: /* XXX -- just like vmnet does */
430 ifs = (struct ifstat *)data;
431 dummy = strlen(ifs->ascii);
432 if (tp->tap_td != NULL && dummy < sizeof(ifs->ascii)) {
433 if (tp->tap_td->td_proc) {
434 snprintf(ifs->ascii + dummy,
435 sizeof(ifs->ascii) - dummy,
436 "\tOpened by pid %d\n",
437 (int)tp->tap_td->td_proc->p_pid);
439 snprintf(ifs->ascii + dummy,
440 sizeof(ifs->ascii) - dummy,
441 "\tOpened by td %p\n", tp->tap_td);
458 * queue packets from higher level ready to put out
464 struct tap_softc *tp = ifp->if_softc;
467 TAPDEBUG("%s%d starting, minor = %#x\n",
468 ifp->if_name, ifp->if_unit, minor(tp->tap_dev));
471 * do not junk pending output if we are in VMnet mode.
472 * XXX: can this do any harm because of queue overflow?
475 if (((tp->tap_flags & TAP_VMNET) == 0) &&
476 ((tp->tap_flags & TAP_READY) != TAP_READY)) {
477 struct mbuf *m = NULL;
479 TAPDEBUG("%s%d not ready. minor = %#x, tap_flags = 0x%x\n",
480 ifp->if_name, ifp->if_unit,
481 minor(tp->tap_dev), tp->tap_flags);
485 IF_DEQUEUE(&ifp->if_snd, m);
496 ifp->if_flags |= IFF_OACTIVE;
498 if (ifp->if_snd.ifq_len != 0) {
499 if (tp->tap_flags & TAP_RWAIT) {
500 tp->tap_flags &= ~TAP_RWAIT;
504 if ((tp->tap_flags & TAP_ASYNC) && (tp->tap_sigio != NULL))
505 pgsigio(tp->tap_sigio, SIGIO, 0);
507 selwakeup(&tp->tap_rsel);
508 ifp->if_opackets ++; /* obytes are counted in ether_output */
511 ifp->if_flags &= ~IFF_OACTIVE;
519 * the cdevsw interface is now pretty minimal
522 tapioctl(dev_t dev, u_long cmd, caddr_t data, int flag, d_thread_t *td)
524 struct tap_softc *tp = dev->si_drv1;
525 struct ifnet *ifp = &tp->tap_if;
526 struct tapinfo *tapp = NULL;
532 tapp = (struct tapinfo *)data;
533 ifp->if_mtu = tapp->mtu;
534 ifp->if_type = tapp->type;
535 ifp->if_baudrate = tapp->baudrate;
540 tapp = (struct tapinfo *)data;
541 tapp->mtu = ifp->if_mtu;
542 tapp->type = ifp->if_type;
543 tapp->baudrate = ifp->if_baudrate;
547 tapdebug = *(int *)data;
551 *(int *)data = tapdebug;
560 tp->tap_flags |= TAP_ASYNC;
562 tp->tap_flags &= ~TAP_ASYNC;
568 if (ifp->if_snd.ifq_head) {
569 struct mbuf *mb = ifp->if_snd.ifq_head;
571 for(*(int *)data = 0;mb != NULL;mb = mb->m_next)
572 *(int *)data += mb->m_len;
580 return (fsetown(*(int *)data, &tp->tap_sigio));
583 *(int *)data = fgetown(tp->tap_sigio);
586 /* this is deprecated, FIOSETOWN should be used instead */
588 return (fsetown(-(*(int *)data), &tp->tap_sigio));
590 /* this is deprecated, FIOGETOWN should be used instead */
592 *(int *)data = -fgetown(tp->tap_sigio);
595 /* VMware/VMnet port ioctl's */
597 case SIOCGIFFLAGS: /* get ifnet flags */
598 bcopy(&ifp->if_flags, data, sizeof(ifp->if_flags));
601 case VMIO_SIOCSIFFLAGS: { /* VMware/VMnet SIOCSIFFLAGS */
602 short f = *(short *)data;
605 f &= ~IFF_CANTCHANGE;
609 ifp->if_flags = f | (ifp->if_flags & IFF_CANTCHANGE);
613 case OSIOCGIFADDR: /* get MAC address of the remote side */
615 bcopy(tp->ether_addr, data, sizeof(tp->ether_addr));
618 case SIOCSIFADDR: /* set MAC address of the remote side */
619 bcopy(data, tp->ether_addr, sizeof(tp->ether_addr));
632 * the cdevsw read interface - reads a packet at a time, or at
633 * least as much of a packet as can be read
636 tapread(dev, uio, flag)
641 struct tap_softc *tp = dev->si_drv1;
642 struct ifnet *ifp = &tp->tap_if;
643 struct mbuf *m0 = NULL;
644 int error = 0, len, s;
646 TAPDEBUG("%s%d reading, minor = %#x\n",
647 ifp->if_name, ifp->if_unit, minor(tp->tap_dev));
649 if ((tp->tap_flags & TAP_READY) != TAP_READY) {
650 TAPDEBUG("%s%d not ready. minor = %#x, tap_flags = 0x%x\n",
651 ifp->if_name, ifp->if_unit,
652 minor(tp->tap_dev), tp->tap_flags);
657 tp->tap_flags &= ~TAP_RWAIT;
659 /* sleep until we get a packet */
662 IF_DEQUEUE(&ifp->if_snd, m0);
666 if (flag & IO_NDELAY)
667 return (EWOULDBLOCK);
669 tp->tap_flags |= TAP_RWAIT;
670 error = tsleep((caddr_t)tp, PCATCH, "taprd", 0);
674 } while (m0 == NULL);
676 /* feed packet to bpf */
677 if (ifp->if_bpf != NULL)
680 /* xfer packet to user space */
681 while ((m0 != NULL) && (uio->uio_resid > 0) && (error == 0)) {
682 len = min(uio->uio_resid, m0->m_len);
686 error = uiomove(mtod(m0, caddr_t), len, uio);
691 TAPDEBUG("%s%d dropping mbuf, minor = %#x\n",
692 ifp->if_name, ifp->if_unit, minor(tp->tap_dev));
703 * the cdevsw write interface - an atomic write is a packet - or else!
706 tapwrite(dev, uio, flag)
711 struct tap_softc *tp = dev->si_drv1;
712 struct ifnet *ifp = &tp->tap_if;
713 struct mbuf *top = NULL, **mp = NULL, *m = NULL;
714 struct ether_header *eh = NULL;
715 int error = 0, tlen, mlen;
717 TAPDEBUG("%s%d writting, minor = %#x\n",
718 ifp->if_name, ifp->if_unit, minor(tp->tap_dev));
720 if (uio->uio_resid == 0)
723 if ((uio->uio_resid < 0) || (uio->uio_resid > TAPMRU)) {
724 TAPDEBUG("%s%d invalid packet len = %d, minor = %#x\n",
725 ifp->if_name, ifp->if_unit,
726 uio->uio_resid, minor(tp->tap_dev));
730 tlen = uio->uio_resid;
732 /* get a header mbuf */
733 MGETHDR(m, M_DONTWAIT, MT_DATA);
740 while ((error == 0) && (uio->uio_resid > 0)) {
741 m->m_len = min(mlen, uio->uio_resid);
742 error = uiomove(mtod(m, caddr_t), m->m_len, uio);
745 if (uio->uio_resid > 0) {
746 MGET(m, M_DONTWAIT, MT_DATA);
761 top->m_pkthdr.len = tlen;
762 top->m_pkthdr.rcvif = ifp;
765 * Ethernet bridge and bpf are handled in ether_input
767 * adjust mbuf and give packet to the ether_input
770 eh = mtod(top, struct ether_header *);
771 m_adj(top, sizeof(struct ether_header));
772 ether_input(ifp, eh, top);
773 ifp->if_ipackets ++; /* ibytes are counted in ether_input */
782 * the poll interface, this is only useful on reads
783 * really. the write detect always returns true, write never blocks
784 * anyway, it either accepts the packet or drops it
787 tappoll(dev_t dev, int events, d_thread_t *td)
789 struct tap_softc *tp = dev->si_drv1;
790 struct ifnet *ifp = &tp->tap_if;
793 TAPDEBUG("%s%d polling, minor = %#x\n",
794 ifp->if_name, ifp->if_unit, minor(tp->tap_dev));
797 if (events & (POLLIN | POLLRDNORM)) {
798 if (ifp->if_snd.ifq_len > 0) {
799 TAPDEBUG("%s%d have data in queue. len = %d, " \
800 "minor = %#x\n", ifp->if_name, ifp->if_unit,
801 ifp->if_snd.ifq_len, minor(tp->tap_dev));
803 revents |= (events & (POLLIN | POLLRDNORM));
806 TAPDEBUG("%s%d waiting for data, minor = %#x\n",
807 ifp->if_name, ifp->if_unit, minor(tp->tap_dev));
809 selrecord(td, &tp->tap_rsel);
813 if (events & (POLLOUT | POLLWRNORM))
814 revents |= (events & (POLLOUT | POLLWRNORM));