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 1.20 2005/06/14 18:37:26 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/thread2.h>
56 #include <sys/ttycom.h>
58 #include <sys/vnode.h>
61 #include <net/ethernet.h>
63 #include <net/ifq_var.h>
64 #include <net/if_arp.h>
65 #include <net/route.h>
67 #include <netinet/in.h>
69 #include "if_tapvar.h"
73 #define CDEV_NAME "tap"
74 #define CDEV_MAJOR 149
75 #define TAPDEBUG if (tapdebug) if_printf
79 #define VMNET_DEV_MASK 0x00010000
82 static int tapmodevent (module_t, int, void *);
85 static void tapcreate (dev_t);
87 /* network interface */
88 static void tapifstart (struct ifnet *);
89 static int tapifioctl (struct ifnet *, u_long, caddr_t,
91 static void tapifinit (void *);
93 /* character device */
94 static d_open_t tapopen;
95 static d_close_t tapclose;
96 static d_read_t tapread;
97 static d_write_t tapwrite;
98 static d_ioctl_t tapioctl;
99 static d_poll_t tappoll;
101 static struct cdevsw tap_cdevsw = {
102 /* dev name */ CDEV_NAME,
103 /* dev major */ CDEV_MAJOR,
109 /* close */ tapclose,
111 /* write */ tapwrite,
112 /* ioctl */ tapioctl,
115 /* startegy */ nostrategy,
120 static int taprefcnt = 0; /* module ref. counter */
121 static int taplastunit = -1; /* max. open unit number */
122 static int tapdebug = 0; /* debug flag */
124 MALLOC_DECLARE(M_TAP);
125 MALLOC_DEFINE(M_TAP, CDEV_NAME, "Ethernet tunnel interface");
126 SYSCTL_INT(_debug, OID_AUTO, if_tap_debug, CTLFLAG_RW, &tapdebug, 0, "");
127 DEV_MODULE(if_tap, tapmodevent, NULL);
132 * module event handler
135 tapmodevent(mod, type, data)
140 static int attached = 0;
141 struct ifnet *ifp = NULL;
149 cdevsw_add(&tap_cdevsw, 0, 0);
157 cdevsw_remove(&tap_cdevsw, 0, 0);
159 /* XXX: maintain tap ifs in a local list */
161 while (unit <= taplastunit) {
163 TAILQ_FOREACH(ifp, &ifnet, if_link)
164 if ((strcmp(ifp->if_dname, TAP) == 0) ||
165 (strcmp(ifp->if_dname, VMNET) == 0))
166 if (ifp->if_dunit == unit)
171 struct tap_softc *tp = ifp->if_softc;
173 TAPDEBUG(ifp, "detached. minor = %#x, " \
174 "taplastunit = %d\n",
175 minor(tp->tap_dev), taplastunit);
180 destroy_dev(tp->tap_dev);
201 * to create interface
207 struct ifnet *ifp = NULL;
208 struct tap_softc *tp = NULL;
209 uint8_t ether_addr[ETHER_ADDR_LEN];
213 /* allocate driver storage and create device */
214 MALLOC(tp, struct tap_softc *, sizeof(*tp), M_TAP, M_WAITOK);
215 bzero(tp, sizeof(*tp));
217 /* select device: tap or vmnet */
218 if (minor(dev) & VMNET_DEV_MASK) {
220 unit = lminor(dev) & 0xff;
221 tp->tap_flags |= TAP_VMNET;
228 tp->tap_dev = make_dev(&tap_cdevsw, minor(dev), UID_ROOT, GID_WHEEL,
229 0600, "%s%d", name, unit);
230 tp->tap_dev->si_drv1 = dev->si_drv1 = tp;
231 reference_dev(tp->tap_dev); /* so we can destroy it later */
233 /* generate fake MAC address: 00 bd xx xx xx unit_no */
234 ether_addr[0] = 0x00;
235 ether_addr[1] = 0xbd;
236 bcopy(&ticks, ether_addr, 4);
237 ether_addr[5] = (u_char)unit;
239 /* fill the rest and attach interface */
243 if_initname(ifp, name, unit);
244 if (unit > taplastunit)
247 ifp->if_init = tapifinit;
248 ifp->if_start = tapifstart;
249 ifp->if_ioctl = tapifioctl;
250 ifp->if_mtu = ETHERMTU;
251 ifp->if_flags = (IFF_BROADCAST|IFF_SIMPLEX|IFF_MULTICAST);
252 ifq_set_maxlen(&ifp->if_snd, ifqmaxlen);
253 ifq_set_ready(&ifp->if_snd);
255 ether_ifattach(ifp, ether_addr);
257 tp->tap_flags |= TAP_INITED;
259 TAPDEBUG(ifp, "created. minor = %#x\n", minor(tp->tap_dev));
266 * to open tunnel. must be superuser
269 tapopen(dev_t dev, int flag, int mode, d_thread_t *td)
271 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(&tp->arpcom.ac_if,
293 "opened. minor = %#x, refcnt = %d, taplastunit = %d\n",
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)
308 struct tap_softc *tp = dev->si_drv1;
309 struct ifnet *ifp = &tp->tap_if;
311 /* junk all pending output */
314 ifq_purge(&ifp->if_snd);
318 * do not bring the interface down, and do not anything with
319 * interface, if we are in VMnet mode. just close the device.
322 if (((tp->tap_flags & TAP_VMNET) == 0) && (ifp->if_flags & IFF_UP)) {
325 if (ifp->if_flags & IFF_RUNNING) {
326 /* find internet addresses and delete routes */
327 struct ifaddr *ifa = NULL;
329 TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) {
330 if (ifa->ifa_addr->sa_family == AF_INET) {
331 rtinit(ifa, (int)RTM_DELETE, 0);
333 /* remove address from interface */
335 sizeof(*(ifa->ifa_addr)));
336 bzero(ifa->ifa_dstaddr,
337 sizeof(*(ifa->ifa_dstaddr)));
338 bzero(ifa->ifa_netmask,
339 sizeof(*(ifa->ifa_netmask)));
343 ifp->if_flags &= ~IFF_RUNNING;
348 funsetown(tp->tap_sigio);
349 selwakeup(&tp->tap_rsel);
351 tp->tap_flags &= ~TAP_OPEN;
357 if_printf(ifp, "minor = %#x, refcnt = %d is out of sync. "
358 "set refcnt to 0\n", minor(tp->tap_dev), taprefcnt);
361 TAPDEBUG(ifp, "closed. minor = %#x, refcnt = %d, taplastunit = %d\n",
362 minor(tp->tap_dev), taprefcnt, taplastunit);
371 * network interface initialization function
377 struct tap_softc *tp = (struct tap_softc *)xtp;
378 struct ifnet *ifp = &tp->tap_if;
380 TAPDEBUG(ifp, "initializing, minor = %#x\n", minor(tp->tap_dev));
382 ifp->if_flags |= IFF_RUNNING;
383 ifp->if_flags &= ~IFF_OACTIVE;
385 /* attempt to start output */
393 * Process an ioctl request on network interface
396 tapifioctl(ifp, cmd, data, cr)
402 struct tap_softc *tp = (struct tap_softc *)(ifp->if_softc);
403 struct ifstat *ifs = NULL;
411 dummy = ether_ioctl(ifp, cmd, data);
415 case SIOCSIFFLAGS: /* XXX -- just like vmnet does */
422 ifs = (struct ifstat *)data;
423 dummy = strlen(ifs->ascii);
424 if (tp->tap_td != NULL && dummy < sizeof(ifs->ascii)) {
425 if (tp->tap_td->td_proc) {
426 snprintf(ifs->ascii + dummy,
427 sizeof(ifs->ascii) - dummy,
428 "\tOpened by pid %d\n",
429 (int)tp->tap_td->td_proc->p_pid);
431 snprintf(ifs->ascii + dummy,
432 sizeof(ifs->ascii) - dummy,
433 "\tOpened by td %p\n", tp->tap_td);
450 * queue packets from higher level ready to put out
456 struct tap_softc *tp = ifp->if_softc;
458 TAPDEBUG(ifp, "starting, minor = %#x\n", minor(tp->tap_dev));
461 * do not junk pending output if we are in VMnet mode.
462 * XXX: can this do any harm because of queue overflow?
465 if (((tp->tap_flags & TAP_VMNET) == 0) &&
466 ((tp->tap_flags & TAP_READY) != TAP_READY)) {
467 TAPDEBUG(ifp, "not ready. minor = %#x, tap_flags = 0x%x\n",
468 minor(tp->tap_dev), tp->tap_flags);
471 ifq_purge(&ifp->if_snd);
478 ifp->if_flags |= IFF_OACTIVE;
480 if (!ifq_is_empty(&ifp->if_snd)) {
481 if (tp->tap_flags & TAP_RWAIT) {
482 tp->tap_flags &= ~TAP_RWAIT;
486 if ((tp->tap_flags & TAP_ASYNC) && (tp->tap_sigio != NULL))
487 pgsigio(tp->tap_sigio, SIGIO, 0);
489 selwakeup(&tp->tap_rsel);
490 ifp->if_opackets ++; /* obytes are counted in ether_output */
493 ifp->if_flags &= ~IFF_OACTIVE;
502 * the cdevsw interface is now pretty minimal
505 tapioctl(dev_t dev, u_long cmd, caddr_t data, int flag, d_thread_t *td)
507 struct tap_softc *tp = dev->si_drv1;
508 struct ifnet *ifp = &tp->tap_if;
509 struct tapinfo *tapp = NULL;
515 tapp = (struct tapinfo *)data;
516 ifp->if_mtu = tapp->mtu;
517 ifp->if_type = tapp->type;
518 ifp->if_baudrate = tapp->baudrate;
523 tapp = (struct tapinfo *)data;
524 tapp->mtu = ifp->if_mtu;
525 tapp->type = ifp->if_type;
526 tapp->baudrate = ifp->if_baudrate;
530 tapdebug = *(int *)data;
534 *(int *)data = tapdebug;
543 tp->tap_flags |= TAP_ASYNC;
545 tp->tap_flags &= ~TAP_ASYNC;
552 if ((mb = ifq_poll(&ifp->if_snd)) != NULL) {
553 for(; mb != NULL; mb = mb->m_next)
554 *(int *)data += mb->m_len;
560 return (fsetown(*(int *)data, &tp->tap_sigio));
563 *(int *)data = fgetown(tp->tap_sigio);
566 /* this is deprecated, FIOSETOWN should be used instead */
568 return (fsetown(-(*(int *)data), &tp->tap_sigio));
570 /* this is deprecated, FIOGETOWN should be used instead */
572 *(int *)data = -fgetown(tp->tap_sigio);
575 /* VMware/VMnet port ioctl's */
577 case SIOCGIFFLAGS: /* get ifnet flags */
578 bcopy(&ifp->if_flags, data, sizeof(ifp->if_flags));
581 case VMIO_SIOCSIFFLAGS: { /* VMware/VMnet SIOCSIFFLAGS */
582 short f = *(short *)data;
585 f &= ~IFF_CANTCHANGE;
589 ifp->if_flags = f | (ifp->if_flags & IFF_CANTCHANGE);
593 case OSIOCGIFADDR: /* get MAC address of the remote side */
595 bcopy(tp->ether_addr, data, sizeof(tp->ether_addr));
598 case SIOCSIFADDR: /* set MAC address of the remote side */
599 bcopy(data, tp->ether_addr, sizeof(tp->ether_addr));
612 * the cdevsw read interface - reads a packet at a time, or at
613 * least as much of a packet as can be read
616 tapread(dev, uio, flag)
621 struct tap_softc *tp = dev->si_drv1;
622 struct ifnet *ifp = &tp->tap_if;
623 struct mbuf *m0 = NULL;
626 TAPDEBUG(ifp, "reading, minor = %#x\n", minor(tp->tap_dev));
628 if ((tp->tap_flags & TAP_READY) != TAP_READY) {
629 TAPDEBUG(ifp, "not ready. minor = %#x, tap_flags = 0x%x\n",
630 minor(tp->tap_dev), tp->tap_flags);
635 tp->tap_flags &= ~TAP_RWAIT;
637 /* sleep until we get a packet */
640 m0 = ifq_dequeue(&ifp->if_snd);
644 if (flag & IO_NDELAY)
645 return (EWOULDBLOCK);
647 tp->tap_flags |= TAP_RWAIT;
648 error = tsleep((caddr_t)tp, PCATCH, "taprd", 0);
652 } while (m0 == NULL);
656 /* xfer packet to user space */
657 while ((m0 != NULL) && (uio->uio_resid > 0) && (error == 0)) {
658 len = min(uio->uio_resid, m0->m_len);
662 error = uiomove(mtod(m0, caddr_t), len, uio);
667 TAPDEBUG(ifp, "dropping mbuf, minor = %#x\n",
679 * the cdevsw write interface - an atomic write is a packet - or else!
682 tapwrite(dev, uio, flag)
687 struct tap_softc *tp = dev->si_drv1;
688 struct ifnet *ifp = &tp->tap_if;
689 struct mbuf *top = NULL, **mp = NULL, *m = NULL;
690 int error = 0, tlen, mlen;
692 TAPDEBUG(ifp, "writting, minor = %#x\n", minor(tp->tap_dev));
694 if (uio->uio_resid == 0)
697 if ((uio->uio_resid < 0) || (uio->uio_resid > TAPMRU)) {
698 TAPDEBUG(ifp, "invalid packet len = %d, minor = %#x\n",
699 uio->uio_resid, minor(tp->tap_dev));
703 tlen = uio->uio_resid;
705 /* get a header mbuf */
706 MGETHDR(m, MB_DONTWAIT, MT_DATA);
713 while ((error == 0) && (uio->uio_resid > 0)) {
714 m->m_len = min(mlen, uio->uio_resid);
715 error = uiomove(mtod(m, caddr_t), m->m_len, uio);
718 if (uio->uio_resid > 0) {
719 MGET(m, MB_DONTWAIT, MT_DATA);
734 top->m_pkthdr.len = tlen;
735 top->m_pkthdr.rcvif = ifp;
738 * Ethernet bridge and bpf are handled in ether_input
740 * adjust mbuf and give packet to the ether_input
743 (*ifp->if_input)(ifp, top);
744 ifp->if_ipackets ++; /* ibytes are counted in ether_input */
753 * the poll interface, this is only useful on reads
754 * really. the write detect always returns true, write never blocks
755 * anyway, it either accepts the packet or drops it
758 tappoll(dev_t dev, int events, d_thread_t *td)
760 struct tap_softc *tp = dev->si_drv1;
761 struct ifnet *ifp = &tp->tap_if;
764 TAPDEBUG(ifp, "polling, minor = %#x\n", minor(tp->tap_dev));
768 if (events & (POLLIN | POLLRDNORM)) {
769 if (!ifq_is_empty(&ifp->if_snd)) {
771 "has data in queue. minor = %#x\n",
774 revents |= (events & (POLLIN | POLLRDNORM));
777 TAPDEBUG(ifp, "waiting for data, minor = %#x\n",
780 selrecord(td, &tp->tap_rsel);
784 if (events & (POLLOUT | POLLWRNORM))
785 revents |= (events & (POLLOUT | POLLWRNORM));