Index: head/sys/net/if_tap.c =================================================================== --- head/sys/net/if_tap.c (revision 71920) +++ head/sys/net/if_tap.c (revision 71921) @@ -1,865 +1,865 @@ /* * Copyright (C) 1999-2000 by Maksim Yevmenkin * 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. * * BASED ON: * ------------------------------------------------------------------------- * * Copyright (c) 1988, Julian Onions * Nottingham University 1987. */ /* * $FreeBSD$ * $Id: if_tap.c,v 0.21 2000/07/23 21:46:02 max Exp $ */ #include "opt_inet.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #define CDEV_NAME "tap" #define CDEV_MAJOR 149 #define TAPDEBUG if (tapdebug) printf #define TAP "tap" #define VMNET "vmnet" #define VMNET_DEV_MASK 0x00010000 /* module */ static int tapmodevent __P((module_t, int, void *)); /* device */ static void tapclone __P((void *, char *, int, dev_t *)); static void tapcreate __P((dev_t)); /* network interface */ static void tapifstart __P((struct ifnet *)); static int tapifioctl __P((struct ifnet *, u_long, caddr_t)); static void tapifinit __P((void *)); /* character device */ static d_open_t tapopen; static d_close_t tapclose; static d_read_t tapread; static d_write_t tapwrite; static d_ioctl_t tapioctl; static d_poll_t tappoll; static struct cdevsw tap_cdevsw = { /* open */ tapopen, /* close */ tapclose, /* read */ tapread, /* write */ tapwrite, /* ioctl */ tapioctl, /* poll */ tappoll, /* mmap */ nommap, /* startegy */ nostrategy, /* dev name */ CDEV_NAME, /* dev major */ CDEV_MAJOR, /* dump */ nodump, /* psize */ nopsize, /* flags */ 0, /* bmaj */ -1 }; static int taprefcnt = 0; /* module ref. counter */ static int taplastunit = -1; /* max. open unit number */ static int tapdebug = 0; /* debug flag */ MALLOC_DECLARE(M_TAP); MALLOC_DEFINE(M_TAP, CDEV_NAME, "Ethernet tunnel interface"); SYSCTL_INT(_debug, OID_AUTO, if_tap_debug, CTLFLAG_RW, &tapdebug, 0, ""); DEV_MODULE(if_tap, tapmodevent, NULL); /* * tapmodevent * * module event handler */ static int tapmodevent(mod, type, data) module_t mod; int type; void *data; { static int attached = 0; static eventhandler_tag eh_tag = NULL; switch (type) { case MOD_LOAD: if (attached) return (EEXIST); eh_tag = EVENTHANDLER_REGISTER(dev_clone, tapclone, 0, 1000); cdevsw_add(&tap_cdevsw); attached = 1; break; case MOD_UNLOAD: { int unit; if (taprefcnt > 0) return (EBUSY); EVENTHANDLER_DEREGISTER(dev_clone, eh_tag); cdevsw_remove(&tap_cdevsw); unit = 0; while (unit <= taplastunit) { int s; struct ifnet *ifp = NULL; s = splimp(); TAILQ_FOREACH(ifp, &ifnet, if_link) if ((strcmp(ifp->if_name, TAP) == 0) || (strcmp(ifp->if_name, VMNET) == 0)) if (ifp->if_unit == unit) break; splx(s); if (ifp != NULL) { struct tap_softc *tp = ifp->if_softc; TAPDEBUG("detaching %s%d. minor = %#x, " \ "taplastunit = %d\n", ifp->if_name, unit, minor(tp->tap_dev), taplastunit); s = splimp(); ether_ifdetach(ifp, 1); splx(s); destroy_dev(tp->tap_dev); FREE(tp, M_TAP); } else unit ++; } attached = 0; } break; default: return (EOPNOTSUPP); } return (0); } /* tapmodevent */ /* * DEVFS handler * * We need to support two kind of devices - tap and vmnet */ static void tapclone(arg, name, namelen, dev) void *arg; char *name; int namelen; dev_t *dev; { int unit, minor; char *device_name = NULL; if (*dev != NODEV) return; device_name = TAP; if (dev_stdclone(name, NULL, device_name, &unit) != 1) { device_name = VMNET; if (dev_stdclone(name, NULL, device_name, &unit) != 1) return; - minor = (unit | VMNET_DEV_MASK); + minor = unit2minor(unit | VMNET_DEV_MASK); } else - minor = unit; + minor = unit2minor(unit); *dev = make_dev(&tap_cdevsw, minor, UID_ROOT, GID_WHEEL, 0600, "%s%d", device_name, unit); } /* tapclone */ /* * tapcreate * * to create interface */ static void tapcreate(dev) dev_t dev; { struct ifnet *ifp = NULL; struct tap_softc *tp = NULL; unsigned short macaddr_hi; int unit, s; char *name = NULL; /* allocate driver storage and create device */ MALLOC(tp, struct tap_softc *, sizeof(*tp), M_TAP, M_WAITOK | M_ZERO); /* select device: tap or vmnet */ if (minor(dev) & VMNET_DEV_MASK) { name = VMNET; unit = dev2unit(dev) & 0xff; tp->tap_flags |= TAP_VMNET; } else { name = TAP; unit = dev2unit(dev); } tp->tap_dev = make_dev(&tap_cdevsw, minor(dev), UID_ROOT, GID_WHEEL, 0600, "%s%d", name, unit); tp->tap_dev->si_drv1 = dev->si_drv1 = tp; /* generate fake MAC address: 00 bd xx xx xx unit_no */ macaddr_hi = htons(0x00bd); bcopy(&macaddr_hi, &tp->arpcom.ac_enaddr[0], sizeof(short)); bcopy(&ticks, &tp->arpcom.ac_enaddr[2], sizeof(long)); tp->arpcom.ac_enaddr[5] = (u_char)unit; /* fill the rest and attach interface */ ifp = &tp->tap_if; ifp->if_softc = tp; ifp->if_unit = unit; if (unit > taplastunit) taplastunit = unit; ifp->if_name = name; ifp->if_init = tapifinit; ifp->if_output = ether_output; ifp->if_start = tapifstart; ifp->if_ioctl = tapifioctl; ifp->if_mtu = ETHERMTU; ifp->if_flags = (IFF_BROADCAST|IFF_SIMPLEX|IFF_MULTICAST); ifp->if_snd.ifq_maxlen = ifqmaxlen; s = splimp(); ether_ifattach(ifp, 1); splx(s); tp->tap_flags |= TAP_INITED; TAPDEBUG("interface %s%d created. minor = %#x\n", ifp->if_name, ifp->if_unit, minor(tp->tap_dev)); } /* tapcreate */ /* * tapopen * * to open tunnel. must be superuser */ static int tapopen(dev, flag, mode, p) dev_t dev; int flag; int mode; struct proc *p; { struct tap_softc *tp = NULL; int error; if ((error = suser(p)) != 0) return (error); tp = dev->si_drv1; if (tp == NULL) { tapcreate(dev); tp = dev->si_drv1; } if (tp->tap_flags & TAP_OPEN) return (EBUSY); bcopy(tp->arpcom.ac_enaddr, tp->ether_addr, sizeof(tp->ether_addr)); tp->tap_pid = p->p_pid; tp->tap_flags |= TAP_OPEN; taprefcnt ++; TAPDEBUG("%s%d is open. minor = %#x, refcnt = %d, taplastunit = %d\n", tp->tap_if.if_name, tp->tap_if.if_unit, minor(tp->tap_dev), taprefcnt, taplastunit); return (0); } /* tapopen */ /* * tapclose * * close the device - mark i/f down & delete routing info */ static int tapclose(dev, foo, bar, p) dev_t dev; int foo; int bar; struct proc *p; { int s; struct tap_softc *tp = dev->si_drv1; struct ifnet *ifp = &tp->tap_if; struct mbuf *m = NULL; /* junk all pending output */ s = splimp(); do { IF_DEQUEUE(&ifp->if_snd, m); if (m != NULL) m_freem(m); } while (m != NULL); splx(s); /* * do not bring the interface down, and do not anything with * interface, if we are in VMnet mode. just close the device. */ if (((tp->tap_flags & TAP_VMNET) == 0) && (ifp->if_flags & IFF_UP)) { s = splimp(); if_down(ifp); if (ifp->if_flags & IFF_RUNNING) { /* find internet addresses and delete routes */ struct ifaddr *ifa = NULL; TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family == AF_INET) { rtinit(ifa, (int)RTM_DELETE, 0); /* remove address from interface */ bzero(ifa->ifa_addr, sizeof(*(ifa->ifa_addr))); bzero(ifa->ifa_dstaddr, sizeof(*(ifa->ifa_dstaddr))); bzero(ifa->ifa_netmask, sizeof(*(ifa->ifa_netmask))); } } ifp->if_flags &= ~IFF_RUNNING; } splx(s); } funsetown(tp->tap_sigio); selwakeup(&tp->tap_rsel); tp->tap_flags &= ~TAP_OPEN; tp->tap_pid = 0; taprefcnt --; if (taprefcnt < 0) { taprefcnt = 0; printf("%s%d minor = %#x, refcnt = %d is out of sync. " \ "set refcnt to 0\n", ifp->if_name, ifp->if_unit, minor(tp->tap_dev), taprefcnt); } TAPDEBUG("%s%d is closed. minor = %#x, refcnt = %d, taplastunit = %d\n", ifp->if_name, ifp->if_unit, minor(tp->tap_dev), taprefcnt, taplastunit); return (0); } /* tapclose */ /* * tapifinit * * network interface initialization function */ static void tapifinit(xtp) void *xtp; { struct tap_softc *tp = (struct tap_softc *)xtp; struct ifnet *ifp = &tp->tap_if; TAPDEBUG("initializing %s%d, minor = %#x\n", ifp->if_name, ifp->if_unit, minor(tp->tap_dev)); ifp->if_flags |= IFF_RUNNING; ifp->if_flags &= ~IFF_OACTIVE; /* attempt to start output */ tapifstart(ifp); } /* tapifinit */ /* * tapifioctl * * Process an ioctl request on network interface */ int tapifioctl(ifp, cmd, data) struct ifnet *ifp; u_long cmd; caddr_t data; { struct tap_softc *tp = (struct tap_softc *)(ifp->if_softc); struct ifstat *ifs = NULL; int s, dummy; switch (cmd) { case SIOCSIFADDR: case SIOCGIFADDR: case SIOCSIFMTU: s = splimp(); dummy = ether_ioctl(ifp, cmd, data); splx(s); return (dummy); case SIOCSIFFLAGS: /* XXX -- just like vmnet does */ case SIOCADDMULTI: case SIOCDELMULTI: break; case SIOCGIFSTATUS: s = splimp(); ifs = (struct ifstat *)data; dummy = strlen(ifs->ascii); if (tp->tap_pid != 0 && dummy < sizeof(ifs->ascii)) snprintf(ifs->ascii + dummy, sizeof(ifs->ascii) - dummy, "\tOpened by PID %d\n", tp->tap_pid); splx(s); break; default: return (EINVAL); } return (0); } /* tapifioctl */ /* * tapifstart * * queue packets from higher level ready to put out */ static void tapifstart(ifp) struct ifnet *ifp; { struct tap_softc *tp = ifp->if_softc; int s; TAPDEBUG("%s%d starting, minor = %#x\n", ifp->if_name, ifp->if_unit, minor(tp->tap_dev)); /* * do not junk pending output if we are in VMnet mode. * XXX: can this do any harm because of queue overflow? */ if (((tp->tap_flags & TAP_VMNET) == 0) && ((tp->tap_flags & TAP_READY) != TAP_READY)) { struct mbuf *m = NULL; TAPDEBUG("%s%d not ready. minor = %#x, tap_flags = 0x%x\n", ifp->if_name, ifp->if_unit, minor(tp->tap_dev), tp->tap_flags); s = splimp(); do { IF_DEQUEUE(&ifp->if_snd, m); if (m != NULL) m_freem(m); ifp->if_oerrors ++; } while (m != NULL); splx(s); return; } s = splimp(); ifp->if_flags |= IFF_OACTIVE; if (ifp->if_snd.ifq_len != 0) { if (tp->tap_flags & TAP_RWAIT) { tp->tap_flags &= ~TAP_RWAIT; wakeup((caddr_t)tp); } if ((tp->tap_flags & TAP_ASYNC) && (tp->tap_sigio != NULL)) pgsigio(tp->tap_sigio, SIGIO, 0); selwakeup(&tp->tap_rsel); ifp->if_opackets ++; /* obytes are counted in ether_output */ } ifp->if_flags &= ~IFF_OACTIVE; splx(s); } /* tapifstart */ /* * tapioctl * * the cdevsw interface is now pretty minimal */ static int tapioctl(dev, cmd, data, flag, p) dev_t dev; u_long cmd; caddr_t data; int flag; struct proc *p; { struct tap_softc *tp = dev->si_drv1; struct ifnet *ifp = &tp->tap_if; struct tapinfo *tapp = NULL; int s; switch (cmd) { case TAPSIFINFO: s = splimp(); tapp = (struct tapinfo *)data; ifp->if_mtu = tapp->mtu; ifp->if_type = tapp->type; ifp->if_baudrate = tapp->baudrate; splx(s); break; case TAPGIFINFO: tapp = (struct tapinfo *)data; tapp->mtu = ifp->if_mtu; tapp->type = ifp->if_type; tapp->baudrate = ifp->if_baudrate; break; case TAPSDEBUG: tapdebug = *(int *)data; break; case TAPGDEBUG: *(int *)data = tapdebug; break; case FIONBIO: break; case FIOASYNC: s = splimp(); if (*(int *)data) tp->tap_flags |= TAP_ASYNC; else tp->tap_flags &= ~TAP_ASYNC; splx(s); break; case FIONREAD: s = splimp(); if (ifp->if_snd.ifq_head) { struct mbuf *mb = ifp->if_snd.ifq_head; for(*(int *)data = 0;mb != NULL;mb = mb->m_next) *(int *)data += mb->m_len; } else *(int *)data = 0; splx(s); break; case FIOSETOWN: return (fsetown(*(int *)data, &tp->tap_sigio)); case FIOGETOWN: *(int *)data = fgetown(tp->tap_sigio); return (0); /* this is deprecated, FIOSETOWN should be used instead */ case TIOCSPGRP: return (fsetown(-(*(int *)data), &tp->tap_sigio)); /* this is deprecated, FIOGETOWN should be used instead */ case TIOCGPGRP: *(int *)data = -fgetown(tp->tap_sigio); return (0); /* VMware/VMnet port ioctl's */ case SIOCGIFFLAGS: /* get ifnet flags */ bcopy(&ifp->if_flags, data, sizeof(ifp->if_flags)); break; case VMIO_SIOCSIFFLAGS: { /* VMware/VMnet SIOCSIFFLAGS */ short f = *(short *)data; f &= 0x0fff; f &= ~IFF_CANTCHANGE; f |= IFF_UP; s = splimp(); ifp->if_flags = f | (ifp->if_flags & IFF_CANTCHANGE); splx(s); } break; case OSIOCGIFADDR: /* get MAC address of the remote side */ case SIOCGIFADDR: bcopy(tp->ether_addr, data, sizeof(tp->ether_addr)); break; case SIOCSIFADDR: /* set MAC address of the remote side */ bcopy(data, tp->ether_addr, sizeof(tp->ether_addr)); break; default: return (ENOTTY); } return (0); } /* tapioctl */ /* * tapread * * the cdevsw read interface - reads a packet at a time, or at * least as much of a packet as can be read */ static int tapread(dev, uio, flag) dev_t dev; struct uio *uio; int flag; { struct tap_softc *tp = dev->si_drv1; struct ifnet *ifp = &tp->tap_if; struct mbuf *m = NULL, *m0 = NULL; int error = 0, len, s; TAPDEBUG("%s%d reading, minor = %#x\n", ifp->if_name, ifp->if_unit, minor(tp->tap_dev)); if ((tp->tap_flags & TAP_READY) != TAP_READY) { TAPDEBUG("%s%d not ready. minor = %#x, tap_flags = 0x%x\n", ifp->if_name, ifp->if_unit, minor(tp->tap_dev), tp->tap_flags); return (EHOSTDOWN); } tp->tap_flags &= ~TAP_RWAIT; /* sleep until we get a packet */ do { s = splimp(); IF_DEQUEUE(&ifp->if_snd, m0); splx(s); if (m0 == NULL) { if (flag & IO_NDELAY) return (EWOULDBLOCK); tp->tap_flags |= TAP_RWAIT; error = tsleep((caddr_t)tp,PCATCH|(PZERO+1),"taprd",0); if (error) return (error); } } while (m0 == NULL); /* feed packet to bpf */ if (ifp->if_bpf != NULL) bpf_mtap(ifp, m0); /* xfer packet to user space */ while ((m0 != NULL) && (uio->uio_resid > 0) && (error == 0)) { len = min(uio->uio_resid, m0->m_len); if (len == 0) break; error = uiomove(mtod(m0, caddr_t), len, uio); MFREE(m0, m); m0 = m; } if (m0 != NULL) { TAPDEBUG("%s%d dropping mbuf, minor = %#x\n", ifp->if_name, ifp->if_unit, minor(tp->tap_dev)); m_freem(m0); } return (error); } /* tapread */ /* * tapwrite * * the cdevsw write interface - an atomic write is a packet - or else! */ static int tapwrite(dev, uio, flag) dev_t dev; struct uio *uio; int flag; { struct tap_softc *tp = dev->si_drv1; struct ifnet *ifp = &tp->tap_if; struct mbuf *top = NULL, **mp = NULL, *m = NULL; struct ether_header *eh = NULL; int error = 0, tlen, mlen; TAPDEBUG("%s%d writting, minor = %#x\n", ifp->if_name, ifp->if_unit, minor(tp->tap_dev)); if (uio->uio_resid == 0) return (0); if ((uio->uio_resid < 0) || (uio->uio_resid > TAPMRU)) { TAPDEBUG("%s%d invalid packet len = %d, minor = %#x\n", ifp->if_name, ifp->if_unit, uio->uio_resid, minor(tp->tap_dev)); return (EIO); } tlen = uio->uio_resid; /* get a header mbuf */ MGETHDR(m, M_DONTWAIT, MT_DATA); if (m == NULL) return (ENOBUFS); mlen = MHLEN; top = 0; mp = ⊤ while ((error == 0) && (uio->uio_resid > 0)) { m->m_len = min(mlen, uio->uio_resid); error = uiomove(mtod(m, caddr_t), m->m_len, uio); *mp = m; mp = &m->m_next; if (uio->uio_resid > 0) { MGET(m, M_DONTWAIT, MT_DATA); if (m == NULL) { error = ENOBUFS; break; } mlen = MLEN; } } if (error) { ifp->if_ierrors ++; if (top) m_freem(top); return (error); } top->m_pkthdr.len = tlen; top->m_pkthdr.rcvif = ifp; /* * Ethernet bridge and bpf are handled in ether_input * * adjust mbuf and give packet to the ether_input */ eh = mtod(top, struct ether_header *); m_adj(top, sizeof(struct ether_header)); ether_input(ifp, eh, top); ifp->if_ipackets ++; /* ibytes are counted in ether_input */ return (0); } /* tapwrite */ /* * tappoll * * the poll interface, this is only useful on reads * really. the write detect always returns true, write never blocks * anyway, it either accepts the packet or drops it */ static int tappoll(dev, events, p) dev_t dev; int events; struct proc *p; { struct tap_softc *tp = dev->si_drv1; struct ifnet *ifp = &tp->tap_if; int s, revents = 0; TAPDEBUG("%s%d polling, minor = %#x\n", ifp->if_name, ifp->if_unit, minor(tp->tap_dev)); s = splimp(); if (events & (POLLIN | POLLRDNORM)) { if (ifp->if_snd.ifq_len > 0) { TAPDEBUG("%s%d have data in queue. len = %d, " \ "minor = %#x\n", ifp->if_name, ifp->if_unit, ifp->if_snd.ifq_len, minor(tp->tap_dev)); revents |= (events & (POLLIN | POLLRDNORM)); } else { TAPDEBUG("%s%d waiting for data, minor = %#x\n", ifp->if_name, ifp->if_unit, minor(tp->tap_dev)); selrecord(p, &tp->tap_rsel); } } if (events & (POLLOUT | POLLWRNORM)) revents |= (events & (POLLOUT | POLLWRNORM)); splx(s); return (revents); } /* tappoll */ Index: head/sys/net/if_tun.c =================================================================== --- head/sys/net/if_tun.c (revision 71920) +++ head/sys/net/if_tun.c (revision 71921) @@ -1,755 +1,754 @@ /* $NetBSD: if_tun.c,v 1.14 1994/06/29 06:36:25 cgd Exp $ */ /* * Copyright (c) 1988, Julian Onions * Nottingham University 1987. * * This source may be freely distributed, however I would be interested * in any changes that are made. * * This driver takes packets off the IP i/f and hands them up to a * user process to have its wicked way with. This driver has it's * roots in a similar driver written by Phil Cockcroft (formerly) at * UCL. This driver is based much more on read/write/poll mode of * operation though. * * $FreeBSD$ */ #include "opt_inet.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef INET #include #endif #include #include #include static MALLOC_DEFINE(M_TUN, "tun", "Tunnel Interface"); static void tuncreate __P((dev_t dev)); #define TUNDEBUG if (tundebug) printf static int tundebug = 0; SYSCTL_INT(_debug, OID_AUTO, if_tun_debug, CTLFLAG_RW, &tundebug, 0, ""); static int tunoutput __P((struct ifnet *, struct mbuf *, struct sockaddr *, struct rtentry *rt)); static int tunifioctl __P((struct ifnet *, u_long, caddr_t)); static int tuninit __P((struct ifnet *)); static void tunstart __P((struct ifnet *)); static d_open_t tunopen; static d_close_t tunclose; static d_read_t tunread; static d_write_t tunwrite; static d_ioctl_t tunioctl; static d_poll_t tunpoll; #define CDEV_MAJOR 52 static struct cdevsw tun_cdevsw = { /* open */ tunopen, /* close */ tunclose, /* read */ tunread, /* write */ tunwrite, /* ioctl */ tunioctl, /* poll */ tunpoll, /* mmap */ nommap, /* strategy */ nostrategy, /* name */ "tun", /* maj */ CDEV_MAJOR, /* dump */ nodump, /* psize */ nopsize, /* flags */ 0, /* bmaj */ -1 }; static void tun_clone __P((void *arg, char *name, int namelen, dev_t *dev)); static void tun_clone(arg, name, namelen, dev) void *arg; char *name; int namelen; dev_t *dev; { int u; if (*dev != NODEV) return; if (dev_stdclone(name, NULL, "tun", &u) != 1) return; - /* XXX: minor encoding if u > 255 */ - *dev = make_dev(&tun_cdevsw, u, + *dev = make_dev(&tun_cdevsw, unit2minor(u), UID_UUCP, GID_DIALER, 0600, "tun%d", u); } static int tun_modevent(module_t mod, int type, void *data) { switch (type) { case MOD_LOAD: EVENTHANDLER_REGISTER(dev_clone, tun_clone, 0, 1000); cdevsw_add(&tun_cdevsw); break; case MOD_UNLOAD: printf("if_tun module unload - not possible for this module type\n"); return EINVAL; } return 0; } static moduledata_t tun_mod = { "if_tun", tun_modevent, 0 }; DECLARE_MODULE(if_tun, tun_mod, SI_SUB_PSEUDO, SI_ORDER_ANY); static void tunstart(ifp) struct ifnet *ifp; { struct tun_softc *tp = ifp->if_softc; if (tp->tun_flags & TUN_RWAIT) { tp->tun_flags &= ~TUN_RWAIT; wakeup((caddr_t)tp); } if (tp->tun_flags & TUN_ASYNC && tp->tun_sigio) pgsigio(tp->tun_sigio, SIGIO, 0); selwakeup(&tp->tun_rsel); } static void tuncreate(dev) dev_t dev; { struct tun_softc *sc; struct ifnet *ifp; dev = make_dev(&tun_cdevsw, minor(dev), UID_UUCP, GID_DIALER, 0600, "tun%d", dev2unit(dev)); MALLOC(sc, struct tun_softc *, sizeof(*sc), M_TUN, M_WAITOK | M_ZERO); sc->tun_flags = TUN_INITED; ifp = &sc->tun_if; ifp->if_unit = dev2unit(dev); ifp->if_name = "tun"; ifp->if_mtu = TUNMTU; ifp->if_ioctl = tunifioctl; ifp->if_output = tunoutput; ifp->if_start = tunstart; ifp->if_flags = IFF_POINTOPOINT | IFF_MULTICAST; ifp->if_type = IFT_PPP; ifp->if_snd.ifq_maxlen = ifqmaxlen; ifp->if_softc = sc; if_attach(ifp); bpfattach(ifp, DLT_NULL, sizeof(u_int)); dev->si_drv1 = sc; } /* * tunnel open - must be superuser & the device must be * configured in */ static int tunopen(dev, flag, mode, p) dev_t dev; int flag, mode; struct proc *p; { struct ifnet *ifp; struct tun_softc *tp; register int error; error = suser(p); if (error) return (error); tp = dev->si_drv1; if (!tp) { tuncreate(dev); tp = dev->si_drv1; } if (tp->tun_flags & TUN_OPEN) return EBUSY; tp->tun_pid = p->p_pid; ifp = &tp->tun_if; tp->tun_flags |= TUN_OPEN; TUNDEBUG("%s%d: open\n", ifp->if_name, ifp->if_unit); return (0); } /* * tunclose - close the device - mark i/f down & delete * routing info */ static int tunclose(dev, foo, bar, p) dev_t dev; int foo; int bar; struct proc *p; { register int s; struct tun_softc *tp; struct ifnet *ifp; tp = dev->si_drv1; ifp = &tp->tun_if; tp->tun_flags &= ~TUN_OPEN; tp->tun_pid = 0; /* * junk all pending output */ IF_DRAIN(&ifp->if_snd); if (ifp->if_flags & IFF_UP) { s = splimp(); if_down(ifp); splx(s); } if (ifp->if_flags & IFF_RUNNING) { register struct ifaddr *ifa; s = splimp(); /* find internet addresses and delete routes */ for (ifa = ifp->if_addrhead.tqh_first; ifa; ifa = ifa->ifa_link.tqe_next) if (ifa->ifa_addr->sa_family == AF_INET) rtinit(ifa, (int)RTM_DELETE, tp->tun_flags & TUN_DSTADDR ? RTF_HOST : 0); ifp->if_flags &= ~IFF_RUNNING; splx(s); } funsetown(tp->tun_sigio); selwakeup(&tp->tun_rsel); TUNDEBUG ("%s%d: closed\n", ifp->if_name, ifp->if_unit); return (0); } static int tuninit(ifp) struct ifnet *ifp; { struct tun_softc *tp = ifp->if_softc; register struct ifaddr *ifa; int error = 0; TUNDEBUG("%s%d: tuninit\n", ifp->if_name, ifp->if_unit); ifp->if_flags |= IFF_UP | IFF_RUNNING; getmicrotime(&ifp->if_lastchange); for (ifa = ifp->if_addrhead.tqh_first; ifa; ifa = ifa->ifa_link.tqe_next) { if (ifa->ifa_addr == NULL) error = EFAULT; /* XXX: Should maybe return straight off? */ else { #ifdef INET if (ifa->ifa_addr->sa_family == AF_INET) { struct sockaddr_in *si; si = (struct sockaddr_in *)ifa->ifa_addr; if (si->sin_addr.s_addr) tp->tun_flags |= TUN_IASET; si = (struct sockaddr_in *)ifa->ifa_dstaddr; if (si && si->sin_addr.s_addr) tp->tun_flags |= TUN_DSTADDR; } #endif } } return (error); } /* * Process an ioctl request. */ int tunifioctl(ifp, cmd, data) struct ifnet *ifp; u_long cmd; caddr_t data; { struct ifreq *ifr = (struct ifreq *)data; struct tun_softc *tp = ifp->if_softc; struct ifstat *ifs; int error = 0, s; s = splimp(); switch(cmd) { case SIOCGIFSTATUS: ifs = (struct ifstat *)data; if (tp->tun_pid) sprintf(ifs->ascii + strlen(ifs->ascii), "\tOpened by PID %d\n", tp->tun_pid); break; case SIOCSIFADDR: error = tuninit(ifp); TUNDEBUG("%s%d: address set, error=%d\n", ifp->if_name, ifp->if_unit, error); break; case SIOCSIFDSTADDR: error = tuninit(ifp); TUNDEBUG("%s%d: destination address set, error=%d\n", ifp->if_name, ifp->if_unit, error); break; case SIOCSIFMTU: ifp->if_mtu = ifr->ifr_mtu; TUNDEBUG("%s%d: mtu set\n", ifp->if_name, ifp->if_unit); break; case SIOCADDMULTI: case SIOCDELMULTI: break; default: error = EINVAL; } splx(s); return (error); } /* * tunoutput - queue packets from higher level ready to put out. */ int tunoutput(ifp, m0, dst, rt) struct ifnet *ifp; struct mbuf *m0; struct sockaddr *dst; struct rtentry *rt; { struct tun_softc *tp = ifp->if_softc; TUNDEBUG ("%s%d: tunoutput\n", ifp->if_name, ifp->if_unit); if ((tp->tun_flags & TUN_READY) != TUN_READY) { TUNDEBUG ("%s%d: not ready 0%o\n", ifp->if_name, ifp->if_unit, tp->tun_flags); m_freem (m0); return EHOSTDOWN; } /* BPF write needs to be handled specially */ if (dst->sa_family == AF_UNSPEC) { dst->sa_family = *(mtod(m0, int *)); m0->m_len -= sizeof(int); m0->m_pkthdr.len -= sizeof(int); m0->m_data += sizeof(int); } if (ifp->if_bpf) { /* * We need to prepend the address family as * a four byte field. Cons up a dummy header * to pacify bpf. This is safe because bpf * will only read from the mbuf (i.e., it won't * try to free it or keep a pointer to it). */ struct mbuf m; uint32_t af = dst->sa_family; m.m_next = m0; m.m_len = 4; m.m_data = (char *)⁡ bpf_mtap(ifp, &m); } /* prepend sockaddr? this may abort if the mbuf allocation fails */ if (tp->tun_flags & TUN_LMODE) { /* allocate space for sockaddr */ M_PREPEND(m0, dst->sa_len, M_DONTWAIT); /* if allocation failed drop packet */ if (m0 == NULL) { ifp->if_iqdrops++; ifp->if_oerrors++; return (ENOBUFS); } else { bcopy(dst, m0->m_data, dst->sa_len); } } if (tp->tun_flags & TUN_IFHEAD) { /* Prepend the address family */ M_PREPEND(m0, 4, M_DONTWAIT); /* if allocation failed drop packet */ if (m0 == NULL) { ifp->if_iqdrops++; ifp->if_oerrors++; return ENOBUFS; } else *(u_int32_t *)m0->m_data = htonl(dst->sa_family); } else { #ifdef INET if (dst->sa_family != AF_INET) #endif { m_freem(m0); return EAFNOSUPPORT; } } if (! IF_HANDOFF(&ifp->if_snd, m0, ifp)) { ifp->if_collisions++; return ENOBUFS; } ifp->if_opackets++; return 0; } /* * the cdevsw interface is now pretty minimal. */ static int tunioctl(dev, cmd, data, flag, p) dev_t dev; u_long cmd; caddr_t data; int flag; struct proc *p; { int s; struct tun_softc *tp = dev->si_drv1; struct tuninfo *tunp; switch (cmd) { case TUNSIFINFO: tunp = (struct tuninfo *)data; if (tunp->mtu < IF_MINMTU) return (EINVAL); tp->tun_if.if_mtu = tunp->mtu; tp->tun_if.if_type = tunp->type; tp->tun_if.if_baudrate = tunp->baudrate; break; case TUNGIFINFO: tunp = (struct tuninfo *)data; tunp->mtu = tp->tun_if.if_mtu; tunp->type = tp->tun_if.if_type; tunp->baudrate = tp->tun_if.if_baudrate; break; case TUNSDEBUG: tundebug = *(int *)data; break; case TUNGDEBUG: *(int *)data = tundebug; break; case TUNSLMODE: if (*(int *)data) { tp->tun_flags |= TUN_LMODE; tp->tun_flags &= ~TUN_IFHEAD; } else tp->tun_flags &= ~TUN_LMODE; break; case TUNSIFHEAD: if (*(int *)data) { tp->tun_flags |= TUN_IFHEAD; tp->tun_flags &= ~TUN_LMODE; } else tp->tun_flags &= ~TUN_IFHEAD; break; case TUNGIFHEAD: *(int *)data = (tp->tun_flags & TUN_IFHEAD) ? 1 : 0; break; case TUNSIFMODE: /* deny this if UP */ if (tp->tun_if.if_flags & IFF_UP) return(EBUSY); switch (*(int *)data) { case IFF_POINTOPOINT: tp->tun_if.if_flags |= IFF_POINTOPOINT; tp->tun_if.if_flags &= ~IFF_BROADCAST; break; case IFF_BROADCAST: tp->tun_if.if_flags &= ~IFF_POINTOPOINT; tp->tun_if.if_flags |= IFF_BROADCAST; break; default: return(EINVAL); } break; case TUNSIFPID: tp->tun_pid = curproc->p_pid; break; case FIONBIO: break; case FIOASYNC: if (*(int *)data) tp->tun_flags |= TUN_ASYNC; else tp->tun_flags &= ~TUN_ASYNC; break; case FIONREAD: s = splimp(); if (tp->tun_if.if_snd.ifq_head) { struct mbuf *mb = tp->tun_if.if_snd.ifq_head; for( *(int *)data = 0; mb != 0; mb = mb->m_next) *(int *)data += mb->m_len; } else *(int *)data = 0; splx(s); break; case FIOSETOWN: return (fsetown(*(int *)data, &tp->tun_sigio)); case FIOGETOWN: *(int *)data = fgetown(tp->tun_sigio); return (0); /* This is deprecated, FIOSETOWN should be used instead. */ case TIOCSPGRP: return (fsetown(-(*(int *)data), &tp->tun_sigio)); /* This is deprecated, FIOGETOWN should be used instead. */ case TIOCGPGRP: *(int *)data = -fgetown(tp->tun_sigio); return (0); default: return (ENOTTY); } return (0); } /* * The cdevsw read interface - reads a packet at a time, or at * least as much of a packet as can be read. */ static int tunread(dev, uio, flag) dev_t dev; struct uio *uio; int flag; { struct tun_softc *tp = dev->si_drv1; struct ifnet *ifp = &tp->tun_if; struct mbuf *m, *m0; int error=0, len, s; TUNDEBUG ("%s%d: read\n", ifp->if_name, ifp->if_unit); if ((tp->tun_flags & TUN_READY) != TUN_READY) { TUNDEBUG ("%s%d: not ready 0%o\n", ifp->if_name, ifp->if_unit, tp->tun_flags); return EHOSTDOWN; } tp->tun_flags &= ~TUN_RWAIT; s = splimp(); do { IF_DEQUEUE(&ifp->if_snd, m0); if (m0 == 0) { if (flag & IO_NDELAY) { splx(s); return EWOULDBLOCK; } tp->tun_flags |= TUN_RWAIT; if((error = tsleep((caddr_t)tp, PCATCH | (PZERO + 1), "tunread", 0)) != 0) { splx(s); return error; } } } while (m0 == 0); splx(s); while (m0 && uio->uio_resid > 0 && error == 0) { len = min(uio->uio_resid, m0->m_len); if (len == 0) break; error = uiomove(mtod(m0, caddr_t), len, uio); MFREE(m0, m); m0 = m; } if (m0) { TUNDEBUG("Dropping mbuf\n"); m_freem(m0); } return error; } /* * the cdevsw write interface - an atomic write is a packet - or else! */ static int tunwrite(dev, uio, flag) dev_t dev; struct uio *uio; int flag; { struct tun_softc *tp = dev->si_drv1; struct ifnet *ifp = &tp->tun_if; struct mbuf *top, **mp, *m; int error=0, tlen, mlen; uint32_t family; TUNDEBUG("%s%d: tunwrite\n", ifp->if_name, ifp->if_unit); if (uio->uio_resid == 0) return 0; if (uio->uio_resid < 0 || uio->uio_resid > TUNMRU) { TUNDEBUG("%s%d: len=%d!\n", ifp->if_name, ifp->if_unit, uio->uio_resid); return EIO; } tlen = uio->uio_resid; /* get a header mbuf */ MGETHDR(m, M_DONTWAIT, MT_DATA); if (m == NULL) return ENOBUFS; mlen = MHLEN; top = 0; mp = ⊤ while (error == 0 && uio->uio_resid > 0) { m->m_len = min(mlen, uio->uio_resid); error = uiomove(mtod (m, caddr_t), m->m_len, uio); *mp = m; mp = &m->m_next; if (uio->uio_resid > 0) { MGET (m, M_DONTWAIT, MT_DATA); if (m == 0) { error = ENOBUFS; break; } mlen = MLEN; } } if (error) { if (top) m_freem (top); ifp->if_ierrors++; return error; } top->m_pkthdr.len = tlen; top->m_pkthdr.rcvif = ifp; if (ifp->if_bpf) { if (tp->tun_flags & TUN_IFHEAD) { /* * Conveniently, we already have a 4-byte address * family prepended to our packet ! * Inconveniently, it's in the wrong byte order ! */ if ((top = m_pullup(top, sizeof(family))) == NULL) return ENOBUFS; *mtod(top, u_int32_t *) = ntohl(*mtod(top, u_int32_t *)); bpf_mtap(ifp, top); *mtod(top, u_int32_t *) = htonl(*mtod(top, u_int32_t *)); } else { /* * We need to prepend the address family as * a four byte field. Cons up a dummy header * to pacify bpf. This is safe because bpf * will only read from the mbuf (i.e., it won't * try to free it or keep a pointer to it). */ struct mbuf m; uint32_t af = AF_INET; m.m_next = top; m.m_len = 4; m.m_data = (char *)⁡ bpf_mtap(ifp, &m); } } if (tp->tun_flags & TUN_IFHEAD) { if (top->m_len < sizeof(family) && (top = m_pullup(top, sizeof(family))) == NULL) return ENOBUFS; family = ntohl(*mtod(top, u_int32_t *)); m_adj(top, sizeof(family)); } else family = AF_INET; ifp->if_ibytes += top->m_pkthdr.len; ifp->if_ipackets++; return family_enqueue(family, top); } /* * tunpoll - the poll interface, this is only useful on reads * really. The write detect always returns true, write never blocks * anyway, it either accepts the packet or drops it. */ static int tunpoll(dev, events, p) dev_t dev; int events; struct proc *p; { int s; struct tun_softc *tp = dev->si_drv1; struct ifnet *ifp = &tp->tun_if; int revents = 0; s = splimp(); TUNDEBUG("%s%d: tunpoll\n", ifp->if_name, ifp->if_unit); if (events & (POLLIN | POLLRDNORM)) { if (ifp->if_snd.ifq_len > 0) { TUNDEBUG("%s%d: tunpoll q=%d\n", ifp->if_name, ifp->if_unit, ifp->if_snd.ifq_len); revents |= events & (POLLIN | POLLRDNORM); } else { TUNDEBUG("%s%d: tunpoll waiting\n", ifp->if_name, ifp->if_unit); selrecord(p, &tp->tun_rsel); } } if (events & (POLLOUT | POLLWRNORM)) revents |= events & (POLLOUT | POLLWRNORM); splx(s); return (revents); }