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 1.18 2005/06/14 18:15:37 joerg 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 "if_tapvar.h"
71 #define CDEV_NAME "tap"
72 #define CDEV_MAJOR 149
73 #define TAPDEBUG if (tapdebug) if_printf
77 #define VMNET_DEV_MASK 0x00010000
80 static int tapmodevent (module_t, int, void *);
83 static void tapcreate (dev_t);
85 /* network interface */
86 static void tapifstart (struct ifnet *);
87 static int tapifioctl (struct ifnet *, u_long, caddr_t,
89 static void tapifinit (void *);
91 /* character device */
92 static d_open_t tapopen;
93 static d_close_t tapclose;
94 static d_read_t tapread;
95 static d_write_t tapwrite;
96 static d_ioctl_t tapioctl;
97 static d_poll_t tappoll;
99 static struct cdevsw tap_cdevsw = {
100 /* dev name */ CDEV_NAME,
101 /* dev major */ CDEV_MAJOR,
107 /* close */ tapclose,
109 /* write */ tapwrite,
110 /* ioctl */ tapioctl,
113 /* startegy */ nostrategy,
118 static int taprefcnt = 0; /* module ref. counter */
119 static int taplastunit = -1; /* max. open unit number */
120 static int tapdebug = 0; /* debug flag */
122 MALLOC_DECLARE(M_TAP);
123 MALLOC_DEFINE(M_TAP, CDEV_NAME, "Ethernet tunnel interface");
124 SYSCTL_INT(_debug, OID_AUTO, if_tap_debug, CTLFLAG_RW, &tapdebug, 0, "");
125 DEV_MODULE(if_tap, tapmodevent, NULL);
130 * module event handler
133 tapmodevent(mod, type, data)
138 static int attached = 0;
139 struct ifnet *ifp = NULL;
147 cdevsw_add(&tap_cdevsw, 0, 0);
155 cdevsw_remove(&tap_cdevsw, 0, 0);
157 /* XXX: maintain tap ifs in a local list */
159 while (unit <= taplastunit) {
161 TAILQ_FOREACH(ifp, &ifnet, if_link)
162 if ((strcmp(ifp->if_dname, TAP) == 0) ||
163 (strcmp(ifp->if_dname, VMNET) == 0))
164 if (ifp->if_dunit == unit)
169 struct tap_softc *tp = ifp->if_softc;
171 TAPDEBUG(ifp, "detached. minor = %#x, " \
172 "taplastunit = %d\n",
173 minor(tp->tap_dev), taplastunit);
178 destroy_dev(tp->tap_dev);
199 * to create interface
205 struct ifnet *ifp = NULL;
206 struct tap_softc *tp = NULL;
207 uint8_t ether_addr[ETHER_ADDR_LEN];
211 /* allocate driver storage and create device */
212 MALLOC(tp, struct tap_softc *, sizeof(*tp), M_TAP, M_WAITOK);
213 bzero(tp, sizeof(*tp));
215 /* select device: tap or vmnet */
216 if (minor(dev) & VMNET_DEV_MASK) {
218 unit = lminor(dev) & 0xff;
219 tp->tap_flags |= TAP_VMNET;
226 tp->tap_dev = make_dev(&tap_cdevsw, minor(dev), UID_ROOT, GID_WHEEL,
227 0600, "%s%d", name, unit);
228 tp->tap_dev->si_drv1 = dev->si_drv1 = tp;
229 reference_dev(tp->tap_dev); /* so we can destroy it later */
231 /* generate fake MAC address: 00 bd xx xx xx unit_no */
232 ether_addr[0] = 0x00;
233 ether_addr[1] = 0xbd;
234 bcopy(&ticks, ether_addr, 4);
235 ether_addr[5] = (u_char)unit;
237 /* fill the rest and attach interface */
241 if_initname(ifp, name, unit);
242 if (unit > taplastunit)
245 ifp->if_init = tapifinit;
246 ifp->if_start = tapifstart;
247 ifp->if_ioctl = tapifioctl;
248 ifp->if_mtu = ETHERMTU;
249 ifp->if_flags = (IFF_BROADCAST|IFF_SIMPLEX|IFF_MULTICAST);
250 ifp->if_snd.ifq_maxlen = ifqmaxlen;
253 ether_ifattach(ifp, ether_addr);
256 tp->tap_flags |= TAP_INITED;
258 TAPDEBUG(ifp, "created. minor = %#x\n", minor(tp->tap_dev));
265 * to open tunnel. must be superuser
268 tapopen(dev_t dev, int flag, int mode, d_thread_t *td)
270 struct tap_softc *tp = NULL;
273 if ((error = suser(td)) != 0)
282 if (tp->tap_flags & TAP_OPEN)
285 bcopy(tp->arpcom.ac_enaddr, tp->ether_addr, sizeof(tp->ether_addr));
288 tp->tap_flags |= TAP_OPEN;
291 TAPDEBUG(&tp->arpcom.ac_if,
292 "opened. minor = %#x, refcnt = %d, taplastunit = %d\n",
293 minor(tp->tap_dev), taprefcnt, taplastunit);
302 * close the device - mark i/f down & delete routing info
305 tapclose(dev_t dev, int foo, int bar, d_thread_t *td)
308 struct tap_softc *tp = dev->si_drv1;
309 struct ifnet *ifp = &tp->tap_if;
310 struct mbuf *m = NULL;
312 /* junk all pending output */
316 IF_DEQUEUE(&ifp->if_snd, m);
323 * do not bring the interface down, and do not anything with
324 * interface, if we are in VMnet mode. just close the device.
327 if (((tp->tap_flags & TAP_VMNET) == 0) && (ifp->if_flags & IFF_UP)) {
330 if (ifp->if_flags & IFF_RUNNING) {
331 /* find internet addresses and delete routes */
332 struct ifaddr *ifa = NULL;
334 TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) {
335 if (ifa->ifa_addr->sa_family == AF_INET) {
336 rtinit(ifa, (int)RTM_DELETE, 0);
338 /* remove address from interface */
340 sizeof(*(ifa->ifa_addr)));
341 bzero(ifa->ifa_dstaddr,
342 sizeof(*(ifa->ifa_dstaddr)));
343 bzero(ifa->ifa_netmask,
344 sizeof(*(ifa->ifa_netmask)));
348 ifp->if_flags &= ~IFF_RUNNING;
353 funsetown(tp->tap_sigio);
354 selwakeup(&tp->tap_rsel);
356 tp->tap_flags &= ~TAP_OPEN;
362 if_printf(ifp, "minor = %#x, refcnt = %d is out of sync. "
363 "set refcnt to 0\n", minor(tp->tap_dev), taprefcnt);
366 TAPDEBUG(ifp, "closed. minor = %#x, refcnt = %d, taplastunit = %d\n",
367 minor(tp->tap_dev), taprefcnt, taplastunit);
376 * network interface initialization function
382 struct tap_softc *tp = (struct tap_softc *)xtp;
383 struct ifnet *ifp = &tp->tap_if;
385 TAPDEBUG(ifp, "initializing, minor = %#x\n", minor(tp->tap_dev));
387 ifp->if_flags |= IFF_RUNNING;
388 ifp->if_flags &= ~IFF_OACTIVE;
390 /* attempt to start output */
398 * Process an ioctl request on network interface
401 tapifioctl(ifp, cmd, data, cr)
407 struct tap_softc *tp = (struct tap_softc *)(ifp->if_softc);
408 struct ifstat *ifs = NULL;
416 dummy = ether_ioctl(ifp, cmd, data);
420 case SIOCSIFFLAGS: /* XXX -- just like vmnet does */
427 ifs = (struct ifstat *)data;
428 dummy = strlen(ifs->ascii);
429 if (tp->tap_td != NULL && dummy < sizeof(ifs->ascii)) {
430 if (tp->tap_td->td_proc) {
431 snprintf(ifs->ascii + dummy,
432 sizeof(ifs->ascii) - dummy,
433 "\tOpened by pid %d\n",
434 (int)tp->tap_td->td_proc->p_pid);
436 snprintf(ifs->ascii + dummy,
437 sizeof(ifs->ascii) - dummy,
438 "\tOpened by td %p\n", tp->tap_td);
455 * queue packets from higher level ready to put out
461 struct tap_softc *tp = ifp->if_softc;
464 TAPDEBUG(ifp, "starting, minor = %#x\n", minor(tp->tap_dev));
467 * do not junk pending output if we are in VMnet mode.
468 * XXX: can this do any harm because of queue overflow?
471 if (((tp->tap_flags & TAP_VMNET) == 0) &&
472 ((tp->tap_flags & TAP_READY) != TAP_READY)) {
473 struct mbuf *m = NULL;
475 TAPDEBUG(ifp, "not ready. minor = %#x, tap_flags = 0x%x\n",
476 minor(tp->tap_dev), tp->tap_flags);
480 IF_DEQUEUE(&ifp->if_snd, m);
491 ifp->if_flags |= IFF_OACTIVE;
493 if (ifp->if_snd.ifq_len != 0) {
494 if (tp->tap_flags & TAP_RWAIT) {
495 tp->tap_flags &= ~TAP_RWAIT;
499 if ((tp->tap_flags & TAP_ASYNC) && (tp->tap_sigio != NULL))
500 pgsigio(tp->tap_sigio, SIGIO, 0);
502 selwakeup(&tp->tap_rsel);
503 ifp->if_opackets ++; /* obytes are counted in ether_output */
506 ifp->if_flags &= ~IFF_OACTIVE;
514 * the cdevsw interface is now pretty minimal
517 tapioctl(dev_t dev, u_long cmd, caddr_t data, int flag, d_thread_t *td)
519 struct tap_softc *tp = dev->si_drv1;
520 struct ifnet *ifp = &tp->tap_if;
521 struct tapinfo *tapp = NULL;
527 tapp = (struct tapinfo *)data;
528 ifp->if_mtu = tapp->mtu;
529 ifp->if_type = tapp->type;
530 ifp->if_baudrate = tapp->baudrate;
535 tapp = (struct tapinfo *)data;
536 tapp->mtu = ifp->if_mtu;
537 tapp->type = ifp->if_type;
538 tapp->baudrate = ifp->if_baudrate;
542 tapdebug = *(int *)data;
546 *(int *)data = tapdebug;
555 tp->tap_flags |= TAP_ASYNC;
557 tp->tap_flags &= ~TAP_ASYNC;
563 if (ifp->if_snd.ifq_head) {
564 struct mbuf *mb = ifp->if_snd.ifq_head;
566 for(*(int *)data = 0;mb != NULL;mb = mb->m_next)
567 *(int *)data += mb->m_len;
575 return (fsetown(*(int *)data, &tp->tap_sigio));
578 *(int *)data = fgetown(tp->tap_sigio);
581 /* this is deprecated, FIOSETOWN should be used instead */
583 return (fsetown(-(*(int *)data), &tp->tap_sigio));
585 /* this is deprecated, FIOGETOWN should be used instead */
587 *(int *)data = -fgetown(tp->tap_sigio);
590 /* VMware/VMnet port ioctl's */
592 case SIOCGIFFLAGS: /* get ifnet flags */
593 bcopy(&ifp->if_flags, data, sizeof(ifp->if_flags));
596 case VMIO_SIOCSIFFLAGS: { /* VMware/VMnet SIOCSIFFLAGS */
597 short f = *(short *)data;
600 f &= ~IFF_CANTCHANGE;
604 ifp->if_flags = f | (ifp->if_flags & IFF_CANTCHANGE);
608 case OSIOCGIFADDR: /* get MAC address of the remote side */
610 bcopy(tp->ether_addr, data, sizeof(tp->ether_addr));
613 case SIOCSIFADDR: /* set MAC address of the remote side */
614 bcopy(data, tp->ether_addr, sizeof(tp->ether_addr));
627 * the cdevsw read interface - reads a packet at a time, or at
628 * least as much of a packet as can be read
631 tapread(dev, uio, flag)
636 struct tap_softc *tp = dev->si_drv1;
637 struct ifnet *ifp = &tp->tap_if;
638 struct mbuf *m0 = NULL;
639 int error = 0, len, s;
641 TAPDEBUG(ifp, "reading, minor = %#x\n", minor(tp->tap_dev));
643 if ((tp->tap_flags & TAP_READY) != TAP_READY) {
644 TAPDEBUG(ifp, "not ready. minor = %#x, tap_flags = 0x%x\n",
645 minor(tp->tap_dev), tp->tap_flags);
650 tp->tap_flags &= ~TAP_RWAIT;
652 /* sleep until we get a packet */
655 IF_DEQUEUE(&ifp->if_snd, m0);
659 if (flag & IO_NDELAY)
660 return (EWOULDBLOCK);
662 tp->tap_flags |= TAP_RWAIT;
663 error = tsleep((caddr_t)tp, PCATCH, "taprd", 0);
667 } while (m0 == NULL);
671 /* xfer packet to user space */
672 while ((m0 != NULL) && (uio->uio_resid > 0) && (error == 0)) {
673 len = min(uio->uio_resid, m0->m_len);
677 error = uiomove(mtod(m0, caddr_t), len, uio);
682 TAPDEBUG(ifp, "dropping mbuf, minor = %#x\n",
694 * the cdevsw write interface - an atomic write is a packet - or else!
697 tapwrite(dev, uio, flag)
702 struct tap_softc *tp = dev->si_drv1;
703 struct ifnet *ifp = &tp->tap_if;
704 struct mbuf *top = NULL, **mp = NULL, *m = NULL;
705 int error = 0, tlen, mlen;
707 TAPDEBUG(ifp, "writting, minor = %#x\n", minor(tp->tap_dev));
709 if (uio->uio_resid == 0)
712 if ((uio->uio_resid < 0) || (uio->uio_resid > TAPMRU)) {
713 TAPDEBUG(ifp, "invalid packet len = %d, minor = %#x\n",
714 uio->uio_resid, minor(tp->tap_dev));
718 tlen = uio->uio_resid;
720 /* get a header mbuf */
721 MGETHDR(m, MB_DONTWAIT, MT_DATA);
728 while ((error == 0) && (uio->uio_resid > 0)) {
729 m->m_len = min(mlen, uio->uio_resid);
730 error = uiomove(mtod(m, caddr_t), m->m_len, uio);
733 if (uio->uio_resid > 0) {
734 MGET(m, MB_DONTWAIT, MT_DATA);
749 top->m_pkthdr.len = tlen;
750 top->m_pkthdr.rcvif = ifp;
753 * Ethernet bridge and bpf are handled in ether_input
755 * adjust mbuf and give packet to the ether_input
758 (*ifp->if_input)(ifp, top);
759 ifp->if_ipackets ++; /* ibytes are counted in ether_input */
768 * the poll interface, this is only useful on reads
769 * really. the write detect always returns true, write never blocks
770 * anyway, it either accepts the packet or drops it
773 tappoll(dev_t dev, int events, d_thread_t *td)
775 struct tap_softc *tp = dev->si_drv1;
776 struct ifnet *ifp = &tp->tap_if;
779 TAPDEBUG(ifp, "polling, minor = %#x\n", minor(tp->tap_dev));
782 if (events & (POLLIN | POLLRDNORM)) {
783 if (ifp->if_snd.ifq_len > 0) {
785 "has data in queue. len = %d, minor = %#x\n",
786 ifp->if_snd.ifq_len, minor(tp->tap_dev));
788 revents |= (events & (POLLIN | POLLRDNORM));
791 TAPDEBUG(ifp, "waiting for data, minor = %#x\n",
794 selrecord(td, &tp->tap_rsel);
798 if (events & (POLLOUT | POLLWRNORM))
799 revents |= (events & (POLLOUT | POLLWRNORM));