diff --git a/UPDATING b/UPDATING index 0f6df8dc3e81..44438de07cb7 100644 --- a/UPDATING +++ b/UPDATING @@ -31,6 +31,14 @@ NOTE TO PEOPLE WHO THINK THAT FreeBSD 13.x IS SLOW: disable the most expensive debugging functionality run "ln -s 'abort:false,junk:false' /etc/malloc.conf".) +20190507: + The tap(4) driver has been folded into tun(4), and the module has been + renamed to tuntap. You should update any kld_load="if_tap" or + kld_load="if_tun" entries in /etc/rc.conf, if_tap_load="YES" or + if_tun_load="YES" entries in /boot/loader.conf to load the if_tuntap + module instead, and "device tap" or "device tun" entries in kernel + config files to select the tuntap device instead. + 20190418: The following knobs have been added related to tradeoffs between safe use of the random device and availability in the absence of diff --git a/sbin/ifconfig/ifconfig.c b/sbin/ifconfig/ifconfig.c index e330cc915667..1c46247b36a7 100644 --- a/sbin/ifconfig/ifconfig.c +++ b/sbin/ifconfig/ifconfig.c @@ -130,6 +130,25 @@ struct ifa_order_elt { TAILQ_HEAD(ifa_queue, ifa_order_elt); +static struct module_map_entry { + const char *ifname; + const char *kldname; +} module_map[] = { + { + .ifname = "tun", + .kldname = "if_tuntap", + }, + { + .ifname = "tap", + .kldname = "if_tuntap", + }, + { + .ifname = "vmnet", + .kldname = "if_tuntap", + }, +}; + + void opt_register(struct option *p) { @@ -1413,9 +1432,10 @@ ifmaybeload(const char *name) { #define MOD_PREFIX_LEN 3 /* "if_" */ struct module_stat mstat; - int fileid, modid; - char ifkind[IFNAMSIZ + MOD_PREFIX_LEN], ifname[IFNAMSIZ], *dp; + int i, fileid, modid; + char ifname[IFNAMSIZ], *ifkind, *dp; const char *cp; + struct module_map_entry *mme; /* loading suppressed by the user */ if (noload) @@ -1429,9 +1449,26 @@ ifmaybeload(const char *name) break; } - /* turn interface and unit into module name */ - strlcpy(ifkind, "if_", sizeof(ifkind)); - strlcat(ifkind, ifname, sizeof(ifkind)); + /* Either derive it from the map or guess otherwise */ + ifkind = NULL; + for (i = 0; i < nitems(module_map); ++i) { + mme = &module_map[i]; + if (strcmp(mme->ifname, ifname) == 0) { + ifkind = strdup(mme->kldname); + if (ifkind == NULL) + err(EXIT_FAILURE, "ifmaybeload"); + break; + } + } + + /* We didn't have an alias for it... we'll guess. */ + if (ifkind == NULL) { + ifkind = malloc(IFNAMSIZ + MOD_PREFIX_LEN); + + /* turn interface and unit into module name */ + strlcpy(ifkind, "if_", sizeof(ifkind)); + strlcat(ifkind, ifname, sizeof(ifkind)); + } /* scan files in kernel */ mstat.version = sizeof(struct module_stat); @@ -1450,7 +1487,7 @@ ifmaybeload(const char *name) /* already loaded? */ if (strcmp(ifname, cp) == 0 || strcmp(ifkind, cp) == 0) - return; + goto out; } } @@ -1459,6 +1496,8 @@ ifmaybeload(const char *name) * infer the names of all drivers (eg mlx4en(4)). */ (void) kldload(ifkind); +out: + free(ifkind); } static struct cmd basic_cmds[] = { diff --git a/share/man/man4/tap.4 b/share/man/man4/tap.4 index cee97074682c..b0d0d729a62e 100644 --- a/share/man/man4/tap.4 +++ b/share/man/man4/tap.4 @@ -1,14 +1,14 @@ .\" $FreeBSD$ .\" Based on PR#2411 .\" -.Dd November 29, 2017 +.Dd April 29, 2019 .Dt TAP 4 .Os .Sh NAME .Nm tap .Nd Ethernet tunnel software network interface .Sh SYNOPSIS -.Cd device tap +.Cd device tuntap .Sh DESCRIPTION The .Nm @@ -51,7 +51,7 @@ The network interfaces are named .Dq Li tap1 , etc., one for each control device that has been opened. These Ethernet network interfaces persist until -.Pa if_tap.ko +.Pa if_tuntap.ko module is unloaded, or until removed with "ifconfig destroy" (see below). .Pp .Nm @@ -96,7 +96,7 @@ It therefore defaults to being enabled until further notice. .Ef .Pp Control devices (once successfully opened) persist until -.Pa if_tap.ko +.Pa if_tuntap.ko is unloaded or the interface is destroyed. .Pp Each interface supports the usual Ethernet network interface @@ -296,27 +296,6 @@ device can also be used with the VMware port as a replacement for the old .Em VMnet device driver. -The driver uses the minor number -to select between -.Nm -and -.Nm vmnet -devices. -.Em VMnet -minor numbers begin at -.Va 0x800000 -+ -.Va N ; -where -.Va N -is a -.Em VMnet -unit number. -In this case the control device is expected to be -.Pa /dev/vmnet Ns Sy N , -and the network interface will be -.Sy vmnet Ns Ar N . -Additionally, .Em VMnet devices do not .Xr ifconfig 8 diff --git a/share/man/man4/tun.4 b/share/man/man4/tun.4 index 82e5f4510aa8..a5d15f8901d9 100644 --- a/share/man/man4/tun.4 +++ b/share/man/man4/tun.4 @@ -2,14 +2,14 @@ .\" $FreeBSD$ .\" Based on PR#2411 .\" -.Dd November 29, 2017 +.Dd April 29, 2019 .Dt TUN 4 .Os .Sh NAME .Nm tun .Nd tunnel software network interface .Sh SYNOPSIS -.Cd device tun +.Cd device tuntap .Sh DESCRIPTION The .Nm @@ -52,7 +52,7 @@ The network interfaces are named .Dq Li tun1 , etc., one for each control device that has been opened. These network interfaces persist until the -.Pa if_tun.ko +.Pa if_tuntap.ko module is unloaded, or until removed with the .Xr ifconfig 8 command. @@ -99,7 +99,7 @@ It therefore defaults to being enabled until further notice. .Ef .Pp Control devices (once successfully opened) persist until -.Pa if_tun.ko +.Pa if_tuntap.ko is unloaded in the same way that network interfaces persist (see above). .Pp Each interface supports the usual network-interface diff --git a/sys/amd64/conf/GENERIC b/sys/amd64/conf/GENERIC index 7969031f60b5..9dc14700a16c 100644 --- a/sys/amd64/conf/GENERIC +++ b/sys/amd64/conf/GENERIC @@ -323,7 +323,7 @@ device padlock_rng # VIA Padlock RNG device rdrand_rng # Intel Bull Mountain RNG device ether # Ethernet support device vlan # 802.1Q VLAN support -device tun # Packet tunnel. +device tuntap # Packet tunnel. device md # Memory "disks" device gif # IPv6 and IPv4 tunneling device firmware # firmware assist module diff --git a/sys/amd64/conf/MINIMAL b/sys/amd64/conf/MINIMAL index 648dbbee0fa5..481a7ce96e44 100644 --- a/sys/amd64/conf/MINIMAL +++ b/sys/amd64/conf/MINIMAL @@ -136,7 +136,7 @@ device padlock_rng # VIA Padlock RNG device rdrand_rng # Intel Bull Mountain RNG device ether # Ethernet support device vlan # 802.1Q VLAN support -device tun # Packet tunnel. +device tuntap # Packet tunnel. device gif # IPv6 and IPv4 tunneling # The `bpf' device enables the Berkeley Packet Filter. diff --git a/sys/arm/conf/DOCKSTAR b/sys/arm/conf/DOCKSTAR index b9d1ad81b022..5684d29ef87b 100644 --- a/sys/arm/conf/DOCKSTAR +++ b/sys/arm/conf/DOCKSTAR @@ -69,7 +69,7 @@ device loop # Network loopback device md # Memory/malloc disk device pty # BSD-style compatibility pseudo ttys device random # Entropy device -device tun # Packet tunnel. +device tuntap # Packet tunnel. device ether # Required for all ethernet devices device vlan # 802.1Q VLAN support device wlan # 802.11 WLAN support diff --git a/sys/arm/conf/DREAMPLUG-1001 b/sys/arm/conf/DREAMPLUG-1001 index 8f309b8e1f61..5d6928470e5d 100644 --- a/sys/arm/conf/DREAMPLUG-1001 +++ b/sys/arm/conf/DREAMPLUG-1001 @@ -72,7 +72,7 @@ device loop # Network loopback device md # Memory/malloc disk device pty # BSD-style compatibility pseudo ttys device random # Entropy device -device tun # Packet tunnel. +device tuntap # Packet tunnel. device ether # Required for all ethernet devices device vlan # 802.1Q VLAN support device wlan # 802.11 WLAN support diff --git a/sys/arm/conf/EFIKA_MX b/sys/arm/conf/EFIKA_MX index 4a5a6a3ea133..248060b6cf04 100644 --- a/sys/arm/conf/EFIKA_MX +++ b/sys/arm/conf/EFIKA_MX @@ -60,7 +60,7 @@ device loop # Network loopback device random # Entropy device device ether # Ethernet support #device vlan # 802.1Q VLAN support -#device tun # Packet tunnel. +#device tuntap # Packet tunnel. #device md # Memory "disks" #device gif # IPv6 and IPv4 tunneling #device firmware # firmware assist module diff --git a/sys/arm/conf/IMX53 b/sys/arm/conf/IMX53 index baaa71e72d86..7b7d67bbca9c 100644 --- a/sys/arm/conf/IMX53 +++ b/sys/arm/conf/IMX53 @@ -47,7 +47,7 @@ device loop # Network loopback device random # Entropy device device ether # Ethernet support #device vlan # 802.1Q VLAN support -#device tun # Packet tunnel. +#device tuntap # Packet tunnel. device md # Memory "disks" #device gif # IPv6 and IPv4 tunneling #device firmware # firmware assist module diff --git a/sys/arm/conf/IMX6 b/sys/arm/conf/IMX6 index 81eb1e0df313..3baf9f06f039 100644 --- a/sys/arm/conf/IMX6 +++ b/sys/arm/conf/IMX6 @@ -51,7 +51,7 @@ device mpcore_timer device loop # Network loopback device random # Entropy device device vlan # 802.1Q VLAN support -device tun # Packet tunnel. +device tuntap # Packet tunnel. device md # Memory "disks" #device gif # IPv6 and IPv4 tunneling #device firmware # firmware assist module diff --git a/sys/arm/conf/TEGRA124 b/sys/arm/conf/TEGRA124 index aa84ed86e91e..a830519d909e 100644 --- a/sys/arm/conf/TEGRA124 +++ b/sys/arm/conf/TEGRA124 @@ -45,7 +45,7 @@ device regulator device loop # Network loopback device random # Entropy device device vlan # 802.1Q VLAN support -#device tun # Packet tunnel. +#device tuntap # Packet tunnel. device md # Memory "disks" #device gif # IPv6 and IPv4 tunneling device firmware # firmware assist module diff --git a/sys/arm64/conf/GENERIC b/sys/arm64/conf/GENERIC index 0097a08f589d..b3f27def13c3 100644 --- a/sys/arm64/conf/GENERIC +++ b/sys/arm64/conf/GENERIC @@ -296,7 +296,7 @@ device loop # Network loopback device random # Entropy device device ether # Ethernet support device vlan # 802.1Q VLAN support -device tun # Packet tunnel. +device tuntap # Packet tunnel. device md # Memory "disks" device gif # IPv6 and IPv4 tunneling device firmware # firmware assist module diff --git a/sys/conf/NOTES b/sys/conf/NOTES index 319e4f2d6bc1..aa5838b74f04 100644 --- a/sys/conf/NOTES +++ b/sys/conf/NOTES @@ -898,11 +898,9 @@ device epair # which discards all packets sent and receives none. device edsc -# The `tap' device is a pty-like virtual Ethernet interface -device tap - -# The `tun' device implements (user-)ppp and nos-tun(8) -device tun +# The `tuntap' device implements (user-)ppp, nos-tun(8) and a pty-like virtual +# Ethernet interface +device tuntap # The `gif' device implements IPv6 over IP4 tunneling, # IPv4 over IPv6 tunneling, IPv4 over IPv4 tunneling and diff --git a/sys/conf/files b/sys/conf/files index 90057c557cb5..7327e650589c 100644 --- a/sys/conf/files +++ b/sys/conf/files @@ -4079,8 +4079,7 @@ net/if_mib.c standard net/if_spppfr.c optional sppp | netgraph_sppp net/if_spppsubr.c optional sppp | netgraph_sppp net/if_stf.c optional stf inet inet6 -net/if_tun.c optional tun -net/if_tap.c optional tap +net/if_tuntap.c optional tuntap net/if_vlan.c optional vlan net/if_vxlan.c optional vxlan inet | vxlan inet6 net/ifdi_if.m optional ether pci iflib diff --git a/sys/i386/conf/GENERIC b/sys/i386/conf/GENERIC index da3627857cb5..be1b5e45bdc1 100644 --- a/sys/i386/conf/GENERIC +++ b/sys/i386/conf/GENERIC @@ -316,7 +316,7 @@ device padlock_rng # VIA Padlock RNG device rdrand_rng # Intel Bull Mountain RNG device ether # Ethernet support device vlan # 802.1Q VLAN support -device tun # Packet tunnel. +device tuntap # Packet tunnel. device md # Memory "disks" device gif # IPv6 and IPv4 tunneling device firmware # firmware assist module diff --git a/sys/i386/conf/MINIMAL b/sys/i386/conf/MINIMAL index a2b8549196d3..0b0bd23f40b4 100644 --- a/sys/i386/conf/MINIMAL +++ b/sys/i386/conf/MINIMAL @@ -137,7 +137,7 @@ device padlock_rng # VIA Padlock RNG device rdrand_rng # Intel Bull Mountain RNG device ether # Ethernet support device vlan # 802.1Q VLAN support -device tun # Packet tunnel. +device tuntap # Packet tunnel. device gif # IPv6 and IPv4 tunneling # The `bpf' device enables the Berkeley Packet Filter. diff --git a/sys/mips/conf/ERL b/sys/mips/conf/ERL index d23beca827d6..f12f6228bd95 100644 --- a/sys/mips/conf/ERL +++ b/sys/mips/conf/ERL @@ -152,7 +152,7 @@ device loop # Network loopback device random # Entropy device device ether # Ethernet support device vlan # 802.1Q VLAN support -device tun # Packet tunnel. +device tuntap # Packet tunnel. device md # Memory "disks" device gif # IPv6 and IPv4 tunneling device firmware # firmware assist module diff --git a/sys/mips/conf/OCTEON1 b/sys/mips/conf/OCTEON1 index 89bcf2cc7ff1..6c536f0d6ed4 100644 --- a/sys/mips/conf/OCTEON1 +++ b/sys/mips/conf/OCTEON1 @@ -187,7 +187,7 @@ device loop # Network loopback device random # Entropy device device ether # Ethernet support device vlan # 802.1Q VLAN support -device tun # Packet tunnel. +device tuntap # Packet tunnel. device md # Memory "disks" device gif # IPv6 and IPv4 tunneling device firmware # firmware assist module diff --git a/sys/modules/Makefile b/sys/modules/Makefile index 2c4a8047f3e4..83bbdad7c615 100644 --- a/sys/modules/Makefile +++ b/sys/modules/Makefile @@ -163,8 +163,7 @@ SUBDIR= \ if_lagg \ ${_if_ndis} \ ${_if_stf} \ - if_tap \ - if_tun \ + if_tuntap \ if_vlan \ if_vxlan \ iflib \ diff --git a/sys/modules/if_tap/Makefile b/sys/modules/if_tap/Makefile deleted file mode 100644 index 0df56f3b105b..000000000000 --- a/sys/modules/if_tap/Makefile +++ /dev/null @@ -1,10 +0,0 @@ -# Copyright (C) 1999-2000 by Maksim Yevmenkin -# -# $FreeBSD$ - -.PATH: ${SRCTOP}/sys/net - -KMOD= if_tap -SRCS= if_tap.c opt_inet.h vnode_if.h - -.include diff --git a/sys/modules/if_tun/Makefile b/sys/modules/if_tun/Makefile deleted file mode 100644 index ddfc8d32709d..000000000000 --- a/sys/modules/if_tun/Makefile +++ /dev/null @@ -1,8 +0,0 @@ -# $FreeBSD$ - -.PATH: ${SRCTOP}/sys/net - -KMOD= if_tun -SRCS= if_tun.c opt_inet.h opt_inet6.h - -.include diff --git a/sys/modules/if_tuntap/Makefile b/sys/modules/if_tuntap/Makefile new file mode 100644 index 000000000000..c473fa07df40 --- /dev/null +++ b/sys/modules/if_tuntap/Makefile @@ -0,0 +1,35 @@ +# $FreeBSD$ + +.PATH: ${SRCTOP}/sys/net + +KMOD= if_tuntap +SRCS= if_tuntap.c opt_inet.h opt_inet6.h + +# Symlink for backwards compatibility with systems installed at 12.0 or older +.if ${MACHINE_CPUARCH} != "powerpc" +SYMLINKS= ${KMOD}.ko ${KMODDIR}/if_tun.ko \ + ${KMOD}.ko ${KMODDIR}/if_tap.ko +.else +# Some PPC systems use msdosfs for /boot, which can't handle links or symlinks +afterinstall: alias alias_debug +alias: .PHONY + ${INSTALL} -T release -o ${KMODOWN} -g ${KMODGRP} -m ${KMODMODE} \ + ${_INSTALLFLAGS} ${PROG} ${DESTDIR}${KMODDIR}/if_tun.ko + ${INSTALL} -T release -o ${KMODOWN} -g ${KMODGRP} -m ${KMODMODE} \ + ${_INSTALLFLAGS} ${PROG} ${DESTDIR}${KMODDIR}/if_tap.ko +.if defined(DEBUG_FLAGS) && !defined(INSTALL_NODEBUG) && \ + "${MK_KERNEL_SYMBOLS}" != "no" +alias_debug: .PHONY + ${INSTALL} -T debug -o ${KMODOWN} -g ${KMODGRP} -m ${KMODMODE} \ + ${_INSTALLFLAGS} ${PROG}.debug \ + ${DESTDIR}${KERN_DEBUGDIR}${KMODDIR}/if_tun.ko + ${INSTALL} -T debug -o ${KMODOWN} -g ${KMODGRP} -m ${KMODMODE} \ + ${_INSTALLFLAGS} ${PROG}.debug \ + ${DESTDIR}${KERN_DEBUGDIR}${KMODDIR}/if_tap.ko +.else +alias_debug: .PHONY +.endif +.endif + + +.include diff --git a/sys/net/if_tap.c b/sys/net/if_tap.c deleted file mode 100644 index ff681b9bee7f..000000000000 --- a/sys/net/if_tap.c +++ /dev/null @@ -1,1145 +0,0 @@ -/*- - * SPDX-License-Identifier: BSD-2-Clause-FreeBSD - * - * 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 -#include -#include -#include -#include -#include -#include -#include -#include - -#include - -#include -#include - -#define CDEV_NAME "tap" -#define TAPDEBUG if (tapdebug) printf - -static const char tapname[] = "tap"; -static const char vmnetname[] = "vmnet"; -#define TAPMAXUNIT 0x7fff -#define VMNET_DEV_MASK CLONE_FLAG0 - -/* module */ -static int tapmodevent(module_t, int, void *); - -/* device */ -static void tapclone(void *, struct ucred *, char *, int, - struct cdev **); -static void tapcreate(struct cdev *); - -/* network interface */ -static void tapifstart(struct ifnet *); -static int tapifioctl(struct ifnet *, u_long, caddr_t); -static void tapifinit(void *); - -static int tap_clone_create(struct if_clone *, int, caddr_t); -static void tap_clone_destroy(struct ifnet *); -static struct if_clone *tap_cloner; -static int vmnet_clone_create(struct if_clone *, int, caddr_t); -static void vmnet_clone_destroy(struct ifnet *); -static struct if_clone *vmnet_cloner; - -/* 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 d_kqfilter_t tapkqfilter; - -/* kqueue(2) */ -static int tapkqread(struct knote *, long); -static int tapkqwrite(struct knote *, long); -static void tapkqdetach(struct knote *); - -static struct filterops tap_read_filterops = { - .f_isfd = 1, - .f_attach = NULL, - .f_detach = tapkqdetach, - .f_event = tapkqread, -}; - -static struct filterops tap_write_filterops = { - .f_isfd = 1, - .f_attach = NULL, - .f_detach = tapkqdetach, - .f_event = tapkqwrite, -}; - -static struct cdevsw tap_cdevsw = { - .d_version = D_VERSION, - .d_flags = D_NEEDMINOR, - .d_open = tapopen, - .d_close = tapclose, - .d_read = tapread, - .d_write = tapwrite, - .d_ioctl = tapioctl, - .d_poll = tappoll, - .d_name = CDEV_NAME, - .d_kqfilter = tapkqfilter, -}; - -/* - * All global variables in if_tap.c are locked with tapmtx, with the - * exception of tapdebug, which is accessed unlocked; tapclones is - * static at runtime. - */ -static struct mtx tapmtx; -static int tapdebug = 0; /* debug flag */ -static int tapuopen = 0; /* allow user open() */ -static int tapuponopen = 0; /* IFF_UP on open() */ -static int tapdclone = 1; /* enable devfs cloning */ -static SLIST_HEAD(, tap_softc) taphead; /* first device */ -static struct clonedevs *tapclones; - -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, ""); - -static struct sx tap_ioctl_sx; -SX_SYSINIT(tap_ioctl_sx, &tap_ioctl_sx, "tap_ioctl"); - -SYSCTL_DECL(_net_link); -static SYSCTL_NODE(_net_link, OID_AUTO, tap, CTLFLAG_RW, 0, - "Ethernet tunnel software network interface"); -SYSCTL_INT(_net_link_tap, OID_AUTO, user_open, CTLFLAG_RW, &tapuopen, 0, - "Allow user to open /dev/tap (based on node permissions)"); -SYSCTL_INT(_net_link_tap, OID_AUTO, up_on_open, CTLFLAG_RW, &tapuponopen, 0, - "Bring interface up when /dev/tap is opened"); -SYSCTL_INT(_net_link_tap, OID_AUTO, devfs_cloning, CTLFLAG_RWTUN, &tapdclone, 0, - "Enable legacy devfs interface creation"); -SYSCTL_INT(_net_link_tap, OID_AUTO, debug, CTLFLAG_RW, &tapdebug, 0, ""); - -DEV_MODULE(if_tap, tapmodevent, NULL); -MODULE_VERSION(if_tap, 1); - -static int -tap_clone_create(struct if_clone *ifc, int unit, caddr_t params) -{ - struct cdev *dev; - int i; - - /* Find any existing device, or allocate new unit number. */ - i = clone_create(&tapclones, &tap_cdevsw, &unit, &dev, 0); - if (i) { - dev = make_dev(&tap_cdevsw, unit, UID_ROOT, GID_WHEEL, 0600, - "%s%d", tapname, unit); - } - - tapcreate(dev); - return (0); -} - -/* vmnet devices are tap devices in disguise */ -static int -vmnet_clone_create(struct if_clone *ifc, int unit, caddr_t params) -{ - struct cdev *dev; - int i; - - /* Find any existing device, or allocate new unit number. */ - i = clone_create(&tapclones, &tap_cdevsw, &unit, &dev, VMNET_DEV_MASK); - if (i) { - dev = make_dev(&tap_cdevsw, unit | VMNET_DEV_MASK, UID_ROOT, - GID_WHEEL, 0600, "%s%d", vmnetname, unit); - } - - tapcreate(dev); - return (0); -} - -static void -tap_destroy(struct tap_softc *tp) -{ - struct ifnet *ifp = tp->tap_ifp; - - CURVNET_SET(ifp->if_vnet); - sx_xlock(&tap_ioctl_sx); - ifp->if_softc = NULL; - sx_xunlock(&tap_ioctl_sx); - - destroy_dev(tp->tap_dev); - seldrain(&tp->tap_rsel); - knlist_clear(&tp->tap_rsel.si_note, 0); - knlist_destroy(&tp->tap_rsel.si_note); - ether_ifdetach(ifp); - if_free(ifp); - - mtx_destroy(&tp->tap_mtx); - free(tp, M_TAP); - CURVNET_RESTORE(); -} - -static void -tap_clone_destroy(struct ifnet *ifp) -{ - struct tap_softc *tp = ifp->if_softc; - - mtx_lock(&tapmtx); - SLIST_REMOVE(&taphead, tp, tap_softc, tap_next); - mtx_unlock(&tapmtx); - tap_destroy(tp); -} - -/* vmnet devices are tap devices in disguise */ -static void -vmnet_clone_destroy(struct ifnet *ifp) -{ - tap_clone_destroy(ifp); -} - -/* - * tapmodevent - * - * module event handler - */ -static int -tapmodevent(module_t mod, int type, void *data) -{ - static eventhandler_tag eh_tag = NULL; - struct tap_softc *tp = NULL; - struct ifnet *ifp = NULL; - - switch (type) { - case MOD_LOAD: - - /* intitialize device */ - - mtx_init(&tapmtx, "tapmtx", NULL, MTX_DEF); - SLIST_INIT(&taphead); - - clone_setup(&tapclones); - eh_tag = EVENTHANDLER_REGISTER(dev_clone, tapclone, 0, 1000); - if (eh_tag == NULL) { - clone_cleanup(&tapclones); - mtx_destroy(&tapmtx); - return (ENOMEM); - } - tap_cloner = if_clone_simple(tapname, tap_clone_create, - tap_clone_destroy, 0); - vmnet_cloner = if_clone_simple(vmnetname, vmnet_clone_create, - vmnet_clone_destroy, 0); - return (0); - - case MOD_UNLOAD: - /* - * The EBUSY algorithm here can't quite atomically - * guarantee that this is race-free since we have to - * release the tap mtx to deregister the clone handler. - */ - mtx_lock(&tapmtx); - SLIST_FOREACH(tp, &taphead, tap_next) { - mtx_lock(&tp->tap_mtx); - if (tp->tap_flags & TAP_OPEN) { - mtx_unlock(&tp->tap_mtx); - mtx_unlock(&tapmtx); - return (EBUSY); - } - mtx_unlock(&tp->tap_mtx); - } - mtx_unlock(&tapmtx); - - EVENTHANDLER_DEREGISTER(dev_clone, eh_tag); - if_clone_detach(tap_cloner); - if_clone_detach(vmnet_cloner); - drain_dev_clone_events(); - - mtx_lock(&tapmtx); - while ((tp = SLIST_FIRST(&taphead)) != NULL) { - SLIST_REMOVE_HEAD(&taphead, tap_next); - mtx_unlock(&tapmtx); - - ifp = tp->tap_ifp; - - TAPDEBUG("detaching %s\n", ifp->if_xname); - - tap_destroy(tp); - mtx_lock(&tapmtx); - } - mtx_unlock(&tapmtx); - clone_cleanup(&tapclones); - - mtx_destroy(&tapmtx); - - break; - - default: - return (EOPNOTSUPP); - } - - return (0); -} /* tapmodevent */ - - -/* - * DEVFS handler - * - * We need to support two kind of devices - tap and vmnet - */ -static void -tapclone(void *arg, struct ucred *cred, char *name, int namelen, struct cdev **dev) -{ - char devname[SPECNAMELEN + 1]; - int i, unit, append_unit; - int extra; - - if (*dev != NULL) - return; - - if (!tapdclone || - (!tapuopen && priv_check_cred(cred, PRIV_NET_IFCREATE) != 0)) - return; - - unit = 0; - append_unit = 0; - extra = 0; - - /* We're interested in only tap/vmnet devices. */ - if (strcmp(name, tapname) == 0) { - unit = -1; - } else if (strcmp(name, vmnetname) == 0) { - unit = -1; - extra = VMNET_DEV_MASK; - } else if (dev_stdclone(name, NULL, tapname, &unit) != 1) { - if (dev_stdclone(name, NULL, vmnetname, &unit) != 1) { - return; - } else { - extra = VMNET_DEV_MASK; - } - } - - if (unit == -1) - append_unit = 1; - - CURVNET_SET(CRED_TO_VNET(cred)); - /* find any existing device, or allocate new unit number */ - i = clone_create(&tapclones, &tap_cdevsw, &unit, dev, extra); - if (i) { - if (append_unit) { - /* - * We were passed 'tun' or 'tap', with no unit specified - * so we'll need to append it now. - */ - namelen = snprintf(devname, sizeof(devname), "%s%d", name, - unit); - name = devname; - } - - *dev = make_dev_credf(MAKEDEV_REF, &tap_cdevsw, unit | extra, - cred, UID_ROOT, GID_WHEEL, 0600, "%s", name); - } - - if_clone_create(name, namelen, NULL); - CURVNET_RESTORE(); -} /* tapclone */ - - -/* - * tapcreate - * - * to create interface - */ -static void -tapcreate(struct cdev *dev) -{ - struct ifnet *ifp = NULL; - struct tap_softc *tp = NULL; - unsigned short macaddr_hi; - uint32_t macaddr_mid; - int unit; - const char *name = NULL; - u_char eaddr[6]; - - /* allocate driver storage and create device */ - tp = malloc(sizeof(*tp), M_TAP, M_WAITOK | M_ZERO); - mtx_init(&tp->tap_mtx, "tap_mtx", NULL, MTX_DEF); - mtx_lock(&tapmtx); - SLIST_INSERT_HEAD(&taphead, tp, tap_next); - mtx_unlock(&tapmtx); - - unit = dev2unit(dev); - - /* select device: tap or vmnet */ - if (unit & VMNET_DEV_MASK) { - name = vmnetname; - tp->tap_flags |= TAP_VMNET; - } else - name = tapname; - - unit &= TAPMAXUNIT; - - TAPDEBUG("tapcreate(%s%d). minor = %#x\n", name, unit, dev2unit(dev)); - - /* generate fake MAC address: 00 bd xx xx xx unit_no */ - macaddr_hi = htons(0x00bd); - macaddr_mid = (uint32_t) ticks; - bcopy(&macaddr_hi, eaddr, sizeof(short)); - bcopy(&macaddr_mid, &eaddr[2], sizeof(uint32_t)); - eaddr[5] = (u_char)unit; - - /* fill the rest and attach interface */ - ifp = tp->tap_ifp = if_alloc(IFT_ETHER); - if (ifp == NULL) - panic("%s%d: can not if_alloc()", name, unit); - ifp->if_softc = tp; - if_initname(ifp, name, unit); - ifp->if_init = tapifinit; - ifp->if_start = tapifstart; - ifp->if_ioctl = tapifioctl; - ifp->if_mtu = ETHERMTU; - ifp->if_flags = (IFF_BROADCAST|IFF_SIMPLEX|IFF_MULTICAST); - IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen); - ifp->if_capabilities |= IFCAP_LINKSTATE; - ifp->if_capenable |= IFCAP_LINKSTATE; - - dev->si_drv1 = tp; - tp->tap_dev = dev; - - ether_ifattach(ifp, eaddr); - - mtx_lock(&tp->tap_mtx); - tp->tap_flags |= TAP_INITED; - mtx_unlock(&tp->tap_mtx); - - knlist_init_mtx(&tp->tap_rsel.si_note, &tp->tap_mtx); - - TAPDEBUG("interface %s is created. minor = %#x\n", - ifp->if_xname, dev2unit(dev)); -} /* tapcreate */ - - -/* - * tapopen - * - * to open tunnel. must be superuser - */ -static int -tapopen(struct cdev *dev, int flag, int mode, struct thread *td) -{ - struct tap_softc *tp = NULL; - struct ifnet *ifp = NULL; - int error; - - if (tapuopen == 0) { - error = priv_check(td, PRIV_NET_TAP); - if (error) - return (error); - } - - if ((dev2unit(dev) & CLONE_UNITMASK) > TAPMAXUNIT) - return (ENXIO); - - tp = dev->si_drv1; - - mtx_lock(&tp->tap_mtx); - if (tp->tap_flags & TAP_OPEN) { - mtx_unlock(&tp->tap_mtx); - return (EBUSY); - } - - bcopy(IF_LLADDR(tp->tap_ifp), tp->ether_addr, sizeof(tp->ether_addr)); - tp->tap_pid = td->td_proc->p_pid; - tp->tap_flags |= TAP_OPEN; - ifp = tp->tap_ifp; - - ifp->if_drv_flags |= IFF_DRV_RUNNING; - ifp->if_drv_flags &= ~IFF_DRV_OACTIVE; - if (tapuponopen) - ifp->if_flags |= IFF_UP; - if_link_state_change(ifp, LINK_STATE_UP); - mtx_unlock(&tp->tap_mtx); - - TAPDEBUG("%s is open. minor = %#x\n", ifp->if_xname, dev2unit(dev)); - - return (0); -} /* tapopen */ - - -/* - * tapclose - * - * close the device - mark i/f down & delete routing info - */ -static int -tapclose(struct cdev *dev, int foo, int bar, struct thread *td) -{ - struct ifaddr *ifa; - struct tap_softc *tp = dev->si_drv1; - struct ifnet *ifp = tp->tap_ifp; - - /* junk all pending output */ - mtx_lock(&tp->tap_mtx); - CURVNET_SET(ifp->if_vnet); - IF_DRAIN(&ifp->if_snd); - - /* - * 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 | IFF_LINK0)) == IFF_UP) { - mtx_unlock(&tp->tap_mtx); - if_down(ifp); - mtx_lock(&tp->tap_mtx); - if (ifp->if_drv_flags & IFF_DRV_RUNNING) { - ifp->if_drv_flags &= ~IFF_DRV_RUNNING; - mtx_unlock(&tp->tap_mtx); - CK_STAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { - rtinit(ifa, (int)RTM_DELETE, 0); - } - if_purgeaddrs(ifp); - mtx_lock(&tp->tap_mtx); - } - } - - if_link_state_change(ifp, LINK_STATE_DOWN); - CURVNET_RESTORE(); - - funsetown(&tp->tap_sigio); - selwakeuppri(&tp->tap_rsel, PZERO+1); - KNOTE_LOCKED(&tp->tap_rsel.si_note, 0); - - tp->tap_flags &= ~TAP_OPEN; - tp->tap_pid = 0; - mtx_unlock(&tp->tap_mtx); - - TAPDEBUG("%s is closed. minor = %#x\n", - ifp->if_xname, dev2unit(dev)); - - return (0); -} /* tapclose */ - - -/* - * tapifinit - * - * network interface initialization function - */ -static void -tapifinit(void *xtp) -{ - struct tap_softc *tp = (struct tap_softc *)xtp; - struct ifnet *ifp = tp->tap_ifp; - - TAPDEBUG("initializing %s\n", ifp->if_xname); - - mtx_lock(&tp->tap_mtx); - ifp->if_drv_flags |= IFF_DRV_RUNNING; - ifp->if_drv_flags &= ~IFF_DRV_OACTIVE; - mtx_unlock(&tp->tap_mtx); - - /* attempt to start output */ - tapifstart(ifp); -} /* tapifinit */ - - -/* - * tapifioctl - * - * Process an ioctl request on network interface - */ -static int -tapifioctl(struct ifnet *ifp, u_long cmd, caddr_t data) -{ - struct tap_softc *tp; - struct ifreq *ifr = (struct ifreq *)data; - struct ifstat *ifs = NULL; - struct ifmediareq *ifmr = NULL; - int dummy, error = 0; - - sx_xlock(&tap_ioctl_sx); - tp = ifp->if_softc; - if (tp == NULL) { - error = ENXIO; - goto bad; - } - switch (cmd) { - case SIOCSIFFLAGS: /* XXX -- just like vmnet does */ - case SIOCADDMULTI: - case SIOCDELMULTI: - break; - - case SIOCGIFMEDIA: - ifmr = (struct ifmediareq *)data; - dummy = ifmr->ifm_count; - ifmr->ifm_count = 1; - ifmr->ifm_status = IFM_AVALID; - ifmr->ifm_active = IFM_ETHER; - if (tp->tap_flags & TAP_OPEN) - ifmr->ifm_status |= IFM_ACTIVE; - ifmr->ifm_current = ifmr->ifm_active; - if (dummy >= 1) { - int media = IFM_ETHER; - error = copyout(&media, ifmr->ifm_ulist, - sizeof(int)); - } - break; - - case SIOCSIFMTU: - ifp->if_mtu = ifr->ifr_mtu; - break; - - case SIOCGIFSTATUS: - ifs = (struct ifstat *)data; - mtx_lock(&tp->tap_mtx); - if (tp->tap_pid != 0) - snprintf(ifs->ascii, sizeof(ifs->ascii), - "\tOpened by PID %d\n", tp->tap_pid); - else - ifs->ascii[0] = '\0'; - mtx_unlock(&tp->tap_mtx); - break; - - default: - error = ether_ioctl(ifp, cmd, data); - break; - } - -bad: - sx_xunlock(&tap_ioctl_sx); - return (error); -} /* tapifioctl */ - - -/* - * tapifstart - * - * queue packets from higher level ready to put out - */ -static void -tapifstart(struct ifnet *ifp) -{ - struct tap_softc *tp = ifp->if_softc; - - TAPDEBUG("%s starting\n", ifp->if_xname); - - /* - * do not junk pending output if we are in VMnet mode. - * XXX: can this do any harm because of queue overflow? - */ - - mtx_lock(&tp->tap_mtx); - if (((tp->tap_flags & TAP_VMNET) == 0) && - ((tp->tap_flags & TAP_READY) != TAP_READY)) { - struct mbuf *m; - - /* Unlocked read. */ - TAPDEBUG("%s not ready, tap_flags = 0x%x\n", ifp->if_xname, - tp->tap_flags); - - for (;;) { - IF_DEQUEUE(&ifp->if_snd, m); - if (m != NULL) { - m_freem(m); - if_inc_counter(ifp, IFCOUNTER_OERRORS, 1); - } else - break; - } - mtx_unlock(&tp->tap_mtx); - - return; - } - - ifp->if_drv_flags |= IFF_DRV_OACTIVE; - - if (!IFQ_IS_EMPTY(&ifp->if_snd)) { - if (tp->tap_flags & TAP_RWAIT) { - tp->tap_flags &= ~TAP_RWAIT; - wakeup(tp); - } - - if ((tp->tap_flags & TAP_ASYNC) && (tp->tap_sigio != NULL)) { - mtx_unlock(&tp->tap_mtx); - pgsigio(&tp->tap_sigio, SIGIO, 0); - mtx_lock(&tp->tap_mtx); - } - - selwakeuppri(&tp->tap_rsel, PZERO+1); - KNOTE_LOCKED(&tp->tap_rsel.si_note, 0); - if_inc_counter(ifp, IFCOUNTER_OPACKETS, 1); /* obytes are counted in ether_output */ - } - - ifp->if_drv_flags &= ~IFF_DRV_OACTIVE; - mtx_unlock(&tp->tap_mtx); -} /* tapifstart */ - - -/* - * tapioctl - * - * the cdevsw interface is now pretty minimal - */ -static int -tapioctl(struct cdev *dev, u_long cmd, caddr_t data, int flag, struct thread *td) -{ - struct ifreq ifr; - struct tap_softc *tp = dev->si_drv1; - struct ifnet *ifp = tp->tap_ifp; - struct tapinfo *tapp = NULL; - int f; - int error; -#if defined(COMPAT_FREEBSD6) || defined(COMPAT_FREEBSD5) || \ - defined(COMPAT_FREEBSD4) - int ival; -#endif - - switch (cmd) { - case TAPSIFINFO: - tapp = (struct tapinfo *)data; - if (ifp->if_type != tapp->type) - return (EPROTOTYPE); - mtx_lock(&tp->tap_mtx); - if (ifp->if_mtu != tapp->mtu) { - strlcpy(ifr.ifr_name, if_name(ifp), IFNAMSIZ); - ifr.ifr_mtu = tapp->mtu; - CURVNET_SET(ifp->if_vnet); - error = ifhwioctl(SIOCSIFMTU, ifp, - (caddr_t)&ifr, td); - CURVNET_RESTORE(); - if (error) { - mtx_unlock(&tp->tap_mtx); - return (error); - } - } - ifp->if_baudrate = tapp->baudrate; - mtx_unlock(&tp->tap_mtx); - break; - - case TAPGIFINFO: - tapp = (struct tapinfo *)data; - mtx_lock(&tp->tap_mtx); - tapp->mtu = ifp->if_mtu; - tapp->type = ifp->if_type; - tapp->baudrate = ifp->if_baudrate; - mtx_unlock(&tp->tap_mtx); - break; - - case TAPSDEBUG: - tapdebug = *(int *)data; - break; - - case TAPGDEBUG: - *(int *)data = tapdebug; - break; - - case TAPGIFNAME: { - struct ifreq *ifr = (struct ifreq *) data; - - strlcpy(ifr->ifr_name, ifp->if_xname, IFNAMSIZ); - } break; - - case FIONBIO: - break; - - case FIOASYNC: - mtx_lock(&tp->tap_mtx); - if (*(int *)data) - tp->tap_flags |= TAP_ASYNC; - else - tp->tap_flags &= ~TAP_ASYNC; - mtx_unlock(&tp->tap_mtx); - break; - - case FIONREAD: - if (!IFQ_IS_EMPTY(&ifp->if_snd)) { - struct mbuf *mb; - - IFQ_LOCK(&ifp->if_snd); - IFQ_POLL_NOLOCK(&ifp->if_snd, mb); - for (*(int *)data = 0; mb != NULL; - mb = mb->m_next) - *(int *)data += mb->m_len; - IFQ_UNLOCK(&ifp->if_snd); - } else - *(int *)data = 0; - 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 */ - -#if defined(COMPAT_FREEBSD6) || defined(COMPAT_FREEBSD5) || \ - defined(COMPAT_FREEBSD4) - case _IO('V', 0): - ival = IOCPARM_IVAL(data); - data = (caddr_t)&ival; - /* FALLTHROUGH */ -#endif - case VMIO_SIOCSIFFLAGS: /* VMware/VMnet SIOCSIFFLAGS */ - f = *(int *)data; - f &= 0x0fff; - f &= ~IFF_CANTCHANGE; - f |= IFF_UP; - - mtx_lock(&tp->tap_mtx); - ifp->if_flags = f | (ifp->if_flags & IFF_CANTCHANGE); - mtx_unlock(&tp->tap_mtx); - break; - - case SIOCGIFADDR: /* get MAC address of the remote side */ - mtx_lock(&tp->tap_mtx); - bcopy(tp->ether_addr, data, sizeof(tp->ether_addr)); - mtx_unlock(&tp->tap_mtx); - break; - - case SIOCSIFADDR: /* set MAC address of the remote side */ - mtx_lock(&tp->tap_mtx); - bcopy(data, tp->ether_addr, sizeof(tp->ether_addr)); - mtx_unlock(&tp->tap_mtx); - 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(struct cdev *dev, struct uio *uio, int flag) -{ - struct tap_softc *tp = dev->si_drv1; - struct ifnet *ifp = tp->tap_ifp; - struct mbuf *m = NULL; - int error = 0, len; - - TAPDEBUG("%s reading, minor = %#x\n", ifp->if_xname, dev2unit(dev)); - - mtx_lock(&tp->tap_mtx); - if ((tp->tap_flags & TAP_READY) != TAP_READY) { - mtx_unlock(&tp->tap_mtx); - - /* Unlocked read. */ - TAPDEBUG("%s not ready. minor = %#x, tap_flags = 0x%x\n", - ifp->if_xname, dev2unit(dev), tp->tap_flags); - - return (EHOSTDOWN); - } - - tp->tap_flags &= ~TAP_RWAIT; - - /* sleep until we get a packet */ - do { - IF_DEQUEUE(&ifp->if_snd, m); - - if (m == NULL) { - if (flag & O_NONBLOCK) { - mtx_unlock(&tp->tap_mtx); - return (EWOULDBLOCK); - } - - tp->tap_flags |= TAP_RWAIT; - error = mtx_sleep(tp, &tp->tap_mtx, PCATCH | (PZERO + 1), - "taprd", 0); - if (error) { - mtx_unlock(&tp->tap_mtx); - return (error); - } - } - } while (m == NULL); - mtx_unlock(&tp->tap_mtx); - - /* feed packet to bpf */ - BPF_MTAP(ifp, m); - - /* xfer packet to user space */ - while ((m != NULL) && (uio->uio_resid > 0) && (error == 0)) { - len = min(uio->uio_resid, m->m_len); - if (len == 0) - break; - - error = uiomove(mtod(m, void *), len, uio); - m = m_free(m); - } - - if (m != NULL) { - TAPDEBUG("%s dropping mbuf, minor = %#x\n", ifp->if_xname, - dev2unit(dev)); - m_freem(m); - } - - return (error); -} /* tapread */ - - -/* - * tapwrite - * - * the cdevsw write interface - an atomic write is a packet - or else! - */ -static int -tapwrite(struct cdev *dev, struct uio *uio, int flag) -{ - struct ether_header *eh; - struct tap_softc *tp = dev->si_drv1; - struct ifnet *ifp = tp->tap_ifp; - struct mbuf *m; - - TAPDEBUG("%s writing, minor = %#x\n", - ifp->if_xname, dev2unit(dev)); - - if (uio->uio_resid == 0) - return (0); - - if ((uio->uio_resid < 0) || (uio->uio_resid > TAPMRU)) { - TAPDEBUG("%s invalid packet len = %zd, minor = %#x\n", - ifp->if_xname, uio->uio_resid, dev2unit(dev)); - - return (EIO); - } - - if ((m = m_uiotombuf(uio, M_NOWAIT, 0, ETHER_ALIGN, - M_PKTHDR)) == NULL) { - if_inc_counter(ifp, IFCOUNTER_IERRORS, 1); - return (ENOBUFS); - } - - m->m_pkthdr.rcvif = ifp; - - /* - * Only pass a unicast frame to ether_input(), if it would actually - * have been received by non-virtual hardware. - */ - if (m->m_len < sizeof(struct ether_header)) { - m_freem(m); - return (0); - } - eh = mtod(m, struct ether_header *); - - if (eh && (ifp->if_flags & IFF_PROMISC) == 0 && - !ETHER_IS_MULTICAST(eh->ether_dhost) && - bcmp(eh->ether_dhost, IF_LLADDR(ifp), ETHER_ADDR_LEN) != 0) { - m_freem(m); - return (0); - } - - /* Pass packet up to parent. */ - CURVNET_SET(ifp->if_vnet); - (*ifp->if_input)(ifp, m); - CURVNET_RESTORE(); - if_inc_counter(ifp, IFCOUNTER_IPACKETS, 1); /* ibytes are counted in parent */ - - 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(struct cdev *dev, int events, struct thread *td) -{ - struct tap_softc *tp = dev->si_drv1; - struct ifnet *ifp = tp->tap_ifp; - int revents = 0; - - TAPDEBUG("%s polling, minor = %#x\n", - ifp->if_xname, dev2unit(dev)); - - if (events & (POLLIN | POLLRDNORM)) { - IFQ_LOCK(&ifp->if_snd); - if (!IFQ_IS_EMPTY(&ifp->if_snd)) { - TAPDEBUG("%s have data in queue. len = %d, " \ - "minor = %#x\n", ifp->if_xname, - ifp->if_snd.ifq_len, dev2unit(dev)); - - revents |= (events & (POLLIN | POLLRDNORM)); - } else { - TAPDEBUG("%s waiting for data, minor = %#x\n", - ifp->if_xname, dev2unit(dev)); - - selrecord(td, &tp->tap_rsel); - } - IFQ_UNLOCK(&ifp->if_snd); - } - - if (events & (POLLOUT | POLLWRNORM)) - revents |= (events & (POLLOUT | POLLWRNORM)); - - return (revents); -} /* tappoll */ - - -/* - * tap_kqfilter - * - * support for kevent() system call - */ -static int -tapkqfilter(struct cdev *dev, struct knote *kn) -{ - struct tap_softc *tp = dev->si_drv1; - struct ifnet *ifp = tp->tap_ifp; - - switch (kn->kn_filter) { - case EVFILT_READ: - TAPDEBUG("%s kqfilter: EVFILT_READ, minor = %#x\n", - ifp->if_xname, dev2unit(dev)); - kn->kn_fop = &tap_read_filterops; - break; - - case EVFILT_WRITE: - TAPDEBUG("%s kqfilter: EVFILT_WRITE, minor = %#x\n", - ifp->if_xname, dev2unit(dev)); - kn->kn_fop = &tap_write_filterops; - break; - - default: - TAPDEBUG("%s kqfilter: invalid filter, minor = %#x\n", - ifp->if_xname, dev2unit(dev)); - return (EINVAL); - /* NOT REACHED */ - } - - kn->kn_hook = tp; - knlist_add(&tp->tap_rsel.si_note, kn, 0); - - return (0); -} /* tapkqfilter */ - - -/* - * tap_kqread - * - * Return true if there is data in the interface queue - */ -static int -tapkqread(struct knote *kn, long hint) -{ - int ret; - struct tap_softc *tp = kn->kn_hook; - struct cdev *dev = tp->tap_dev; - struct ifnet *ifp = tp->tap_ifp; - - if ((kn->kn_data = ifp->if_snd.ifq_len) > 0) { - TAPDEBUG("%s have data in queue. len = %d, minor = %#x\n", - ifp->if_xname, ifp->if_snd.ifq_len, dev2unit(dev)); - ret = 1; - } else { - TAPDEBUG("%s waiting for data, minor = %#x\n", - ifp->if_xname, dev2unit(dev)); - ret = 0; - } - - return (ret); -} /* tapkqread */ - - -/* - * tap_kqwrite - * - * Always can write. Return the MTU in kn->data - */ -static int -tapkqwrite(struct knote *kn, long hint) -{ - struct tap_softc *tp = kn->kn_hook; - struct ifnet *ifp = tp->tap_ifp; - - kn->kn_data = ifp->if_mtu; - - return (1); -} /* tapkqwrite */ - - -static void -tapkqdetach(struct knote *kn) -{ - struct tap_softc *tp = kn->kn_hook; - - knlist_remove(&tp->tap_rsel.si_note, kn, 0); -} /* tapkqdetach */ - diff --git a/sys/net/if_tap.h b/sys/net/if_tap.h index 34f44b38d9e2..90772af71497 100644 --- a/sys/net/if_tap.h +++ b/sys/net/if_tap.h @@ -40,23 +40,21 @@ #ifndef _NET_IF_TAP_H_ #define _NET_IF_TAP_H_ -/* refer to if_tapvar.h for the softc stuff */ +#include /* maximum receive packet size (hard limit) */ #define TAPMRU 16384 -struct tapinfo { - int baudrate; /* linespeed */ - short mtu; /* maximum transmission unit */ - u_char type; /* ethernet, tokenring, etc. */ - u_char dummy; /* place holder */ -}; +#define tapinfo tuninfo -/* ioctl's for get/set debug */ -#define TAPSDEBUG _IOW('t', 90, int) -#define TAPGDEBUG _IOR('t', 89, int) -#define TAPSIFINFO _IOW('t', 91, struct tapinfo) -#define TAPGIFINFO _IOR('t', 92, struct tapinfo) +/* + * ioctl's for get/set debug; these are aliases of TUN* ioctls, see net/if_tun.h + * for details. + */ +#define TAPSDEBUG TUNSDEBUG +#define TAPGDEBUG TUNGDEBUG +#define TAPSIFINFO TUNSIFINFO +#define TAPGIFINFO TUNGIFINFO #define TAPGIFNAME _IOR('t', 93, struct ifreq) /* VMware ioctl's */ diff --git a/sys/net/if_tapvar.h b/sys/net/if_tapvar.h deleted file mode 100644 index f5cf9f3e5cec..000000000000 --- a/sys/net/if_tapvar.h +++ /dev/null @@ -1,71 +0,0 @@ -/*- - * SPDX-License-Identifier: BSD-2-Clause-FreeBSD - * - * 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) 1998 Brian Somers - * All rights reserved. - * - * Copyright (c) 1988, Julian Onions - * Nottingham University 1987. - */ - -/* - * $FreeBSD$ - * $Id: if_tapvar.h,v 0.6 2000/07/11 02:16:08 max Exp $ - */ - -#ifndef _NET_IF_TAPVAR_H_ -#define _NET_IF_TAPVAR_H_ - -/* - * tap_mtx locks tap_flags, tap_pid. tap_next locked with global tapmtx. - * Other fields locked by owning subsystems. - */ -struct tap_softc { - struct ifnet *tap_ifp; - u_short tap_flags; /* misc flags */ -#define TAP_OPEN (1 << 0) -#define TAP_INITED (1 << 1) -#define TAP_RWAIT (1 << 2) -#define TAP_ASYNC (1 << 3) -#define TAP_READY (TAP_OPEN|TAP_INITED) -#define TAP_VMNET (1 << 4) - - u_int8_t ether_addr[ETHER_ADDR_LEN]; /* ether addr of the remote side */ - - pid_t tap_pid; /* PID of process to open */ - struct sigio *tap_sigio; /* information for async I/O */ - struct selinfo tap_rsel; /* read select */ - - SLIST_ENTRY(tap_softc) tap_next; /* next device in chain */ - struct cdev *tap_dev; - struct mtx tap_mtx; /* per-softc mutex */ -}; - -#endif /* !_NET_IF_TAPVAR_H_ */ diff --git a/sys/net/if_tun.c b/sys/net/if_tun.c deleted file mode 100644 index ca7639351cd3..000000000000 --- a/sys/net/if_tun.c +++ /dev/null @@ -1,1112 +0,0 @@ -/* $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 "opt_inet6.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 - -#include -#include -#include -#include -#include -#include -#include -#ifdef INET -#include -#endif -#include -#include - -#include -#include - -#include - -/* - * tun_list is protected by global tunmtx. Other mutable fields are - * protected by tun->tun_mtx, or by their owning subsystem. tun_dev is - * static for the duration of a tunnel interface. - */ -struct tun_softc { - TAILQ_ENTRY(tun_softc) tun_list; - struct cdev *tun_dev; - u_short tun_flags; /* misc flags */ -#define TUN_OPEN 0x0001 -#define TUN_INITED 0x0002 -#define TUN_RCOLL 0x0004 -#define TUN_IASET 0x0008 -#define TUN_DSTADDR 0x0010 -#define TUN_LMODE 0x0020 -#define TUN_RWAIT 0x0040 -#define TUN_ASYNC 0x0080 -#define TUN_IFHEAD 0x0100 -#define TUN_DYING 0x0200 - -#define TUN_READY (TUN_OPEN | TUN_INITED) - - pid_t tun_pid; /* owning pid */ - struct ifnet *tun_ifp; /* the interface */ - struct sigio *tun_sigio; /* information for async I/O */ - struct selinfo tun_rsel; /* read select */ - struct mtx tun_mtx; /* protect mutable softc fields */ - struct cv tun_cv; /* protect against ref'd dev destroy */ -}; -#define TUN2IFP(sc) ((sc)->tun_ifp) - -#define TUNDEBUG if (tundebug) if_printf - -/* - * All mutable global variables in if_tun are locked using tunmtx, with - * the exception of tundebug, which is used unlocked, and tunclones, - * which is static after setup. - */ -static struct mtx tunmtx; -static eventhandler_tag tag; -static const char tunname[] = "tun"; -static MALLOC_DEFINE(M_TUN, tunname, "Tunnel Interface"); -static int tundebug = 0; -static int tundclone = 1; -static struct clonedevs *tunclones; -static TAILQ_HEAD(,tun_softc) tunhead = TAILQ_HEAD_INITIALIZER(tunhead); -SYSCTL_INT(_debug, OID_AUTO, if_tun_debug, CTLFLAG_RW, &tundebug, 0, ""); - -static struct sx tun_ioctl_sx; -SX_SYSINIT(tun_ioctl_sx, &tun_ioctl_sx, "tun_ioctl"); - -SYSCTL_DECL(_net_link); -static SYSCTL_NODE(_net_link, OID_AUTO, tun, CTLFLAG_RW, 0, - "IP tunnel software network interface."); -SYSCTL_INT(_net_link_tun, OID_AUTO, devfs_cloning, CTLFLAG_RWTUN, &tundclone, 0, - "Enable legacy devfs interface creation."); - -static void tunclone(void *arg, struct ucred *cred, char *name, - int namelen, struct cdev **dev); -static void tuncreate(const char *name, struct cdev *dev); -static int tunifioctl(struct ifnet *, u_long, caddr_t); -static void tuninit(struct ifnet *); -static int tunmodevent(module_t, int, void *); -static int tunoutput(struct ifnet *, struct mbuf *, - const struct sockaddr *, struct route *ro); -static void tunstart(struct ifnet *); - -static int tun_clone_match(struct if_clone *ifc, const char *name); -static int tun_clone_create(struct if_clone *, char *, size_t, caddr_t); -static int tun_clone_destroy(struct if_clone *, struct ifnet *); -static struct unrhdr *tun_unrhdr; -VNET_DEFINE_STATIC(struct if_clone *, tun_cloner); -#define V_tun_cloner VNET(tun_cloner) - -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; -static d_kqfilter_t tunkqfilter; - -static int tunkqread(struct knote *, long); -static int tunkqwrite(struct knote *, long); -static void tunkqdetach(struct knote *); - -static struct filterops tun_read_filterops = { - .f_isfd = 1, - .f_attach = NULL, - .f_detach = tunkqdetach, - .f_event = tunkqread, -}; - -static struct filterops tun_write_filterops = { - .f_isfd = 1, - .f_attach = NULL, - .f_detach = tunkqdetach, - .f_event = tunkqwrite, -}; - -static struct cdevsw tun_cdevsw = { - .d_version = D_VERSION, - .d_flags = D_NEEDMINOR, - .d_open = tunopen, - .d_close = tunclose, - .d_read = tunread, - .d_write = tunwrite, - .d_ioctl = tunioctl, - .d_poll = tunpoll, - .d_kqfilter = tunkqfilter, - .d_name = tunname, -}; - -static int -tun_clone_match(struct if_clone *ifc, const char *name) -{ - if (strncmp(tunname, name, 3) == 0 && - (name[3] == '\0' || isdigit(name[3]))) - return (1); - - return (0); -} - -static int -tun_clone_create(struct if_clone *ifc, char *name, size_t len, caddr_t params) -{ - struct cdev *dev; - int err, unit, i; - - err = ifc_name2unit(name, &unit); - if (err != 0) - return (err); - - if (unit != -1) { - /* If this unit number is still available that/s okay. */ - if (alloc_unr_specific(tun_unrhdr, unit) == -1) - return (EEXIST); - } else { - unit = alloc_unr(tun_unrhdr); - } - - snprintf(name, IFNAMSIZ, "%s%d", tunname, unit); - - /* find any existing device, or allocate new unit number */ - i = clone_create(&tunclones, &tun_cdevsw, &unit, &dev, 0); - if (i) { - /* No preexisting struct cdev *, create one */ - dev = make_dev(&tun_cdevsw, unit, - UID_UUCP, GID_DIALER, 0600, "%s%d", tunname, unit); - } - tuncreate(tunname, dev); - - return (0); -} - -static void -tunclone(void *arg, struct ucred *cred, char *name, int namelen, - struct cdev **dev) -{ - char devname[SPECNAMELEN + 1]; - int u, i, append_unit; - - if (*dev != NULL) - return; - - /* - * If tun cloning is enabled, only the superuser can create an - * interface. - */ - if (!tundclone || priv_check_cred(cred, PRIV_NET_IFCREATE) != 0) - return; - - if (strcmp(name, tunname) == 0) { - u = -1; - } else if (dev_stdclone(name, NULL, tunname, &u) != 1) - return; /* Don't recognise the name */ - if (u != -1 && u > IF_MAXUNIT) - return; /* Unit number too high */ - - if (u == -1) - append_unit = 1; - else - append_unit = 0; - - CURVNET_SET(CRED_TO_VNET(cred)); - /* find any existing device, or allocate new unit number */ - i = clone_create(&tunclones, &tun_cdevsw, &u, dev, 0); - if (i) { - if (append_unit) { - namelen = snprintf(devname, sizeof(devname), "%s%d", - name, u); - name = devname; - } - /* No preexisting struct cdev *, create one */ - *dev = make_dev_credf(MAKEDEV_REF, &tun_cdevsw, u, cred, - UID_UUCP, GID_DIALER, 0600, "%s", name); - } - - if_clone_create(name, namelen, NULL); - CURVNET_RESTORE(); -} - -static void -tun_destroy(struct tun_softc *tp) -{ - struct cdev *dev; - - mtx_lock(&tp->tun_mtx); - tp->tun_flags |= TUN_DYING; - if ((tp->tun_flags & TUN_OPEN) != 0) - cv_wait_unlock(&tp->tun_cv, &tp->tun_mtx); - else - mtx_unlock(&tp->tun_mtx); - - CURVNET_SET(TUN2IFP(tp)->if_vnet); - sx_xlock(&tun_ioctl_sx); - TUN2IFP(tp)->if_softc = NULL; - sx_xunlock(&tun_ioctl_sx); - - dev = tp->tun_dev; - bpfdetach(TUN2IFP(tp)); - if_detach(TUN2IFP(tp)); - free_unr(tun_unrhdr, TUN2IFP(tp)->if_dunit); - if_free(TUN2IFP(tp)); - destroy_dev(dev); - seldrain(&tp->tun_rsel); - knlist_clear(&tp->tun_rsel.si_note, 0); - knlist_destroy(&tp->tun_rsel.si_note); - mtx_destroy(&tp->tun_mtx); - cv_destroy(&tp->tun_cv); - free(tp, M_TUN); - CURVNET_RESTORE(); -} - -static int -tun_clone_destroy(struct if_clone *ifc, struct ifnet *ifp) -{ - struct tun_softc *tp = ifp->if_softc; - - mtx_lock(&tunmtx); - TAILQ_REMOVE(&tunhead, tp, tun_list); - mtx_unlock(&tunmtx); - tun_destroy(tp); - - return (0); -} - -static void -vnet_tun_init(const void *unused __unused) -{ - V_tun_cloner = if_clone_advanced(tunname, 0, tun_clone_match, - tun_clone_create, tun_clone_destroy); -} -VNET_SYSINIT(vnet_tun_init, SI_SUB_PROTO_IF, SI_ORDER_ANY, - vnet_tun_init, NULL); - -static void -vnet_tun_uninit(const void *unused __unused) -{ - if_clone_detach(V_tun_cloner); -} -VNET_SYSUNINIT(vnet_tun_uninit, SI_SUB_PROTO_IF, SI_ORDER_ANY, - vnet_tun_uninit, NULL); - -static void -tun_uninit(const void *unused __unused) -{ - struct tun_softc *tp; - - EVENTHANDLER_DEREGISTER(dev_clone, tag); - drain_dev_clone_events(); - - mtx_lock(&tunmtx); - while ((tp = TAILQ_FIRST(&tunhead)) != NULL) { - TAILQ_REMOVE(&tunhead, tp, tun_list); - mtx_unlock(&tunmtx); - tun_destroy(tp); - mtx_lock(&tunmtx); - } - mtx_unlock(&tunmtx); - delete_unrhdr(tun_unrhdr); - clone_cleanup(&tunclones); - mtx_destroy(&tunmtx); -} -SYSUNINIT(tun_uninit, SI_SUB_PROTO_IF, SI_ORDER_ANY, tun_uninit, NULL); - -static int -tunmodevent(module_t mod, int type, void *data) -{ - - switch (type) { - case MOD_LOAD: - mtx_init(&tunmtx, "tunmtx", NULL, MTX_DEF); - clone_setup(&tunclones); - tun_unrhdr = new_unrhdr(0, IF_MAXUNIT, &tunmtx); - tag = EVENTHANDLER_REGISTER(dev_clone, tunclone, 0, 1000); - if (tag == NULL) - return (ENOMEM); - break; - case MOD_UNLOAD: - /* See tun_uninit, so it's done after the vnet_sysuninit() */ - break; - default: - return EOPNOTSUPP; - } - return 0; -} - -static moduledata_t tun_mod = { - "if_tun", - tunmodevent, - 0 -}; - -DECLARE_MODULE(if_tun, tun_mod, SI_SUB_PSEUDO, SI_ORDER_ANY); -MODULE_VERSION(if_tun, 1); - -static void -tunstart(struct ifnet *ifp) -{ - struct tun_softc *tp = ifp->if_softc; - struct mbuf *m; - - TUNDEBUG(ifp,"%s starting\n", ifp->if_xname); - if (ALTQ_IS_ENABLED(&ifp->if_snd)) { - IFQ_LOCK(&ifp->if_snd); - IFQ_POLL_NOLOCK(&ifp->if_snd, m); - if (m == NULL) { - IFQ_UNLOCK(&ifp->if_snd); - return; - } - IFQ_UNLOCK(&ifp->if_snd); - } - - mtx_lock(&tp->tun_mtx); - if (tp->tun_flags & TUN_RWAIT) { - tp->tun_flags &= ~TUN_RWAIT; - wakeup(tp); - } - selwakeuppri(&tp->tun_rsel, PZERO + 1); - KNOTE_LOCKED(&tp->tun_rsel.si_note, 0); - if (tp->tun_flags & TUN_ASYNC && tp->tun_sigio) { - mtx_unlock(&tp->tun_mtx); - pgsigio(&tp->tun_sigio, SIGIO, 0); - } else - mtx_unlock(&tp->tun_mtx); -} - -/* XXX: should return an error code so it can fail. */ -static void -tuncreate(const char *name, struct cdev *dev) -{ - struct tun_softc *sc; - struct ifnet *ifp; - - sc = malloc(sizeof(*sc), M_TUN, M_WAITOK | M_ZERO); - mtx_init(&sc->tun_mtx, "tun_mtx", NULL, MTX_DEF); - cv_init(&sc->tun_cv, "tun_condvar"); - sc->tun_flags = TUN_INITED; - sc->tun_dev = dev; - mtx_lock(&tunmtx); - TAILQ_INSERT_TAIL(&tunhead, sc, tun_list); - mtx_unlock(&tunmtx); - - ifp = sc->tun_ifp = if_alloc(IFT_PPP); - if (ifp == NULL) - panic("%s%d: failed to if_alloc() interface.\n", - name, dev2unit(dev)); - if_initname(ifp, name, dev2unit(dev)); - 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_softc = sc; - IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen); - ifp->if_snd.ifq_drv_maxlen = 0; - IFQ_SET_READY(&ifp->if_snd); - knlist_init_mtx(&sc->tun_rsel.si_note, &sc->tun_mtx); - ifp->if_capabilities |= IFCAP_LINKSTATE; - ifp->if_capenable |= IFCAP_LINKSTATE; - - if_attach(ifp); - bpfattach(ifp, DLT_NULL, sizeof(u_int32_t)); - dev->si_drv1 = sc; - TUNDEBUG(ifp, "interface %s is created, minor = %#x\n", - ifp->if_xname, dev2unit(dev)); -} - -static int -tunopen(struct cdev *dev, int flag, int mode, struct thread *td) -{ - struct ifnet *ifp; - struct tun_softc *tp; - - /* - * XXXRW: Non-atomic test and set of dev->si_drv1 requires - * synchronization. - */ - tp = dev->si_drv1; - if (!tp) { - tuncreate(tunname, dev); - tp = dev->si_drv1; - } - - mtx_lock(&tp->tun_mtx); - if ((tp->tun_flags & (TUN_OPEN | TUN_DYING)) != 0) { - mtx_unlock(&tp->tun_mtx); - return (EBUSY); - } - - tp->tun_pid = td->td_proc->p_pid; - tp->tun_flags |= TUN_OPEN; - ifp = TUN2IFP(tp); - if_link_state_change(ifp, LINK_STATE_UP); - TUNDEBUG(ifp, "open\n"); - mtx_unlock(&tp->tun_mtx); - - return (0); -} - -/* - * tunclose - close the device - mark i/f down & delete - * routing info - */ -static int -tunclose(struct cdev *dev, int foo, int bar, struct thread *td) -{ - struct tun_softc *tp; - struct ifnet *ifp; - - tp = dev->si_drv1; - ifp = TUN2IFP(tp); - - mtx_lock(&tp->tun_mtx); - /* - * Simply close the device if this isn't the controlling process. This - * may happen if, for instance, the tunnel has been handed off to - * another process. The original controller should be able to close it - * without putting us into an inconsistent state. - */ - if (td->td_proc->p_pid != tp->tun_pid) { - mtx_unlock(&tp->tun_mtx); - return (0); - } - - /* - * junk all pending output - */ - CURVNET_SET(ifp->if_vnet); - IFQ_PURGE(&ifp->if_snd); - - if (ifp->if_flags & IFF_UP) { - mtx_unlock(&tp->tun_mtx); - if_down(ifp); - mtx_lock(&tp->tun_mtx); - } - - /* Delete all addresses and routes which reference this interface. */ - if (ifp->if_drv_flags & IFF_DRV_RUNNING) { - struct ifaddr *ifa; - - ifp->if_drv_flags &= ~IFF_DRV_RUNNING; - mtx_unlock(&tp->tun_mtx); - CK_STAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { - /* deal w/IPv4 PtP destination; unlocked read */ - if (ifa->ifa_addr->sa_family == AF_INET) { - rtinit(ifa, (int)RTM_DELETE, - tp->tun_flags & TUN_DSTADDR ? RTF_HOST : 0); - } else { - rtinit(ifa, (int)RTM_DELETE, 0); - } - } - if_purgeaddrs(ifp); - mtx_lock(&tp->tun_mtx); - } - if_link_state_change(ifp, LINK_STATE_DOWN); - CURVNET_RESTORE(); - - funsetown(&tp->tun_sigio); - selwakeuppri(&tp->tun_rsel, PZERO + 1); - KNOTE_LOCKED(&tp->tun_rsel.si_note, 0); - TUNDEBUG (ifp, "closed\n"); - tp->tun_flags &= ~TUN_OPEN; - tp->tun_pid = 0; - - cv_broadcast(&tp->tun_cv); - mtx_unlock(&tp->tun_mtx); - return (0); -} - -static void -tuninit(struct ifnet *ifp) -{ - struct tun_softc *tp = ifp->if_softc; -#ifdef INET - struct ifaddr *ifa; -#endif - - TUNDEBUG(ifp, "tuninit\n"); - - mtx_lock(&tp->tun_mtx); - ifp->if_flags |= IFF_UP; - ifp->if_drv_flags |= IFF_DRV_RUNNING; - getmicrotime(&ifp->if_lastchange); - -#ifdef INET - if_addr_rlock(ifp); - CK_STAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { - 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; - } - } - if_addr_runlock(ifp); -#endif - mtx_unlock(&tp->tun_mtx); -} - -/* - * Process an ioctl request. - */ -static int -tunifioctl(struct ifnet *ifp, u_long cmd, caddr_t data) -{ - struct ifreq *ifr = (struct ifreq *)data; - struct tun_softc *tp; - struct ifstat *ifs; - int error = 0; - - sx_xlock(&tun_ioctl_sx); - tp = ifp->if_softc; - if (tp == NULL) { - error = ENXIO; - goto bad; - } - switch(cmd) { - case SIOCGIFSTATUS: - ifs = (struct ifstat *)data; - mtx_lock(&tp->tun_mtx); - if (tp->tun_pid) - snprintf(ifs->ascii, sizeof(ifs->ascii), - "\tOpened by PID %d\n", tp->tun_pid); - else - ifs->ascii[0] = '\0'; - mtx_unlock(&tp->tun_mtx); - break; - case SIOCSIFADDR: - tuninit(ifp); - TUNDEBUG(ifp, "address set\n"); - break; - case SIOCSIFMTU: - ifp->if_mtu = ifr->ifr_mtu; - TUNDEBUG(ifp, "mtu set\n"); - break; - case SIOCSIFFLAGS: - case SIOCADDMULTI: - case SIOCDELMULTI: - break; - default: - error = EINVAL; - } -bad: - sx_xunlock(&tun_ioctl_sx); - return (error); -} - -/* - * tunoutput - queue packets from higher level ready to put out. - */ -static int -tunoutput(struct ifnet *ifp, struct mbuf *m0, const struct sockaddr *dst, - struct route *ro) -{ - struct tun_softc *tp = ifp->if_softc; - u_short cached_tun_flags; - int error; - u_int32_t af; - - TUNDEBUG (ifp, "tunoutput\n"); - -#ifdef MAC - error = mac_ifnet_check_transmit(ifp, m0); - if (error) { - m_freem(m0); - return (error); - } -#endif - - /* Could be unlocked read? */ - mtx_lock(&tp->tun_mtx); - cached_tun_flags = tp->tun_flags; - mtx_unlock(&tp->tun_mtx); - if ((cached_tun_flags & TUN_READY) != TUN_READY) { - TUNDEBUG (ifp, "not ready 0%o\n", tp->tun_flags); - m_freem (m0); - return (EHOSTDOWN); - } - - if ((ifp->if_flags & IFF_UP) != IFF_UP) { - m_freem (m0); - return (EHOSTDOWN); - } - - /* BPF writes need to be handled specially. */ - if (dst->sa_family == AF_UNSPEC) - bcopy(dst->sa_data, &af, sizeof(af)); - else - af = dst->sa_family; - - if (bpf_peers_present(ifp->if_bpf)) - bpf_mtap2(ifp->if_bpf, &af, sizeof(af), m0); - - /* prepend sockaddr? this may abort if the mbuf allocation fails */ - if (cached_tun_flags & TUN_LMODE) { - /* allocate space for sockaddr */ - M_PREPEND(m0, dst->sa_len, M_NOWAIT); - - /* if allocation failed drop packet */ - if (m0 == NULL) { - if_inc_counter(ifp, IFCOUNTER_IQDROPS, 1); - if_inc_counter(ifp, IFCOUNTER_OERRORS, 1); - return (ENOBUFS); - } else { - bcopy(dst, m0->m_data, dst->sa_len); - } - } - - if (cached_tun_flags & TUN_IFHEAD) { - /* Prepend the address family */ - M_PREPEND(m0, 4, M_NOWAIT); - - /* if allocation failed drop packet */ - if (m0 == NULL) { - if_inc_counter(ifp, IFCOUNTER_IQDROPS, 1); - if_inc_counter(ifp, IFCOUNTER_OERRORS, 1); - return (ENOBUFS); - } else - *(u_int32_t *)m0->m_data = htonl(af); - } else { -#ifdef INET - if (af != AF_INET) -#endif - { - m_freem(m0); - return (EAFNOSUPPORT); - } - } - - error = (ifp->if_transmit)(ifp, m0); - if (error) - return (ENOBUFS); - if_inc_counter(ifp, IFCOUNTER_OPACKETS, 1); - return (0); -} - -/* - * the cdevsw interface is now pretty minimal. - */ -static int -tunioctl(struct cdev *dev, u_long cmd, caddr_t data, int flag, - struct thread *td) -{ - struct ifreq ifr; - struct tun_softc *tp = dev->si_drv1; - struct tuninfo *tunp; - int error; - - switch (cmd) { - case TUNSIFINFO: - tunp = (struct tuninfo *)data; - if (TUN2IFP(tp)->if_type != tunp->type) - return (EPROTOTYPE); - mtx_lock(&tp->tun_mtx); - if (TUN2IFP(tp)->if_mtu != tunp->mtu) { - strlcpy(ifr.ifr_name, if_name(TUN2IFP(tp)), IFNAMSIZ); - ifr.ifr_mtu = tunp->mtu; - CURVNET_SET(TUN2IFP(tp)->if_vnet); - error = ifhwioctl(SIOCSIFMTU, TUN2IFP(tp), - (caddr_t)&ifr, td); - CURVNET_RESTORE(); - if (error) { - mtx_unlock(&tp->tun_mtx); - return (error); - } - } - TUN2IFP(tp)->if_baudrate = tunp->baudrate; - mtx_unlock(&tp->tun_mtx); - break; - case TUNGIFINFO: - tunp = (struct tuninfo *)data; - mtx_lock(&tp->tun_mtx); - tunp->mtu = TUN2IFP(tp)->if_mtu; - tunp->type = TUN2IFP(tp)->if_type; - tunp->baudrate = TUN2IFP(tp)->if_baudrate; - mtx_unlock(&tp->tun_mtx); - break; - case TUNSDEBUG: - tundebug = *(int *)data; - break; - case TUNGDEBUG: - *(int *)data = tundebug; - break; - case TUNSLMODE: - mtx_lock(&tp->tun_mtx); - if (*(int *)data) { - tp->tun_flags |= TUN_LMODE; - tp->tun_flags &= ~TUN_IFHEAD; - } else - tp->tun_flags &= ~TUN_LMODE; - mtx_unlock(&tp->tun_mtx); - break; - case TUNSIFHEAD: - mtx_lock(&tp->tun_mtx); - if (*(int *)data) { - tp->tun_flags |= TUN_IFHEAD; - tp->tun_flags &= ~TUN_LMODE; - } else - tp->tun_flags &= ~TUN_IFHEAD; - mtx_unlock(&tp->tun_mtx); - break; - case TUNGIFHEAD: - mtx_lock(&tp->tun_mtx); - *(int *)data = (tp->tun_flags & TUN_IFHEAD) ? 1 : 0; - mtx_unlock(&tp->tun_mtx); - break; - case TUNSIFMODE: - /* deny this if UP */ - if (TUN2IFP(tp)->if_flags & IFF_UP) - return(EBUSY); - - switch (*(int *)data & ~IFF_MULTICAST) { - case IFF_POINTOPOINT: - case IFF_BROADCAST: - mtx_lock(&tp->tun_mtx); - TUN2IFP(tp)->if_flags &= - ~(IFF_BROADCAST|IFF_POINTOPOINT|IFF_MULTICAST); - TUN2IFP(tp)->if_flags |= *(int *)data; - mtx_unlock(&tp->tun_mtx); - break; - default: - return(EINVAL); - } - break; - case TUNSIFPID: - mtx_lock(&tp->tun_mtx); - tp->tun_pid = curthread->td_proc->p_pid; - mtx_unlock(&tp->tun_mtx); - break; - case FIONBIO: - break; - case FIOASYNC: - mtx_lock(&tp->tun_mtx); - if (*(int *)data) - tp->tun_flags |= TUN_ASYNC; - else - tp->tun_flags &= ~TUN_ASYNC; - mtx_unlock(&tp->tun_mtx); - break; - case FIONREAD: - if (!IFQ_IS_EMPTY(&TUN2IFP(tp)->if_snd)) { - struct mbuf *mb; - IFQ_LOCK(&TUN2IFP(tp)->if_snd); - IFQ_POLL_NOLOCK(&TUN2IFP(tp)->if_snd, mb); - for (*(int *)data = 0; mb != NULL; mb = mb->m_next) - *(int *)data += mb->m_len; - IFQ_UNLOCK(&TUN2IFP(tp)->if_snd); - } else - *(int *)data = 0; - 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(struct cdev *dev, struct uio *uio, int flag) -{ - struct tun_softc *tp = dev->si_drv1; - struct ifnet *ifp = TUN2IFP(tp); - struct mbuf *m; - int error=0, len; - - TUNDEBUG (ifp, "read\n"); - mtx_lock(&tp->tun_mtx); - if ((tp->tun_flags & TUN_READY) != TUN_READY) { - mtx_unlock(&tp->tun_mtx); - TUNDEBUG (ifp, "not ready 0%o\n", tp->tun_flags); - return (EHOSTDOWN); - } - - tp->tun_flags &= ~TUN_RWAIT; - - do { - IFQ_DEQUEUE(&ifp->if_snd, m); - if (m == NULL) { - if (flag & O_NONBLOCK) { - mtx_unlock(&tp->tun_mtx); - return (EWOULDBLOCK); - } - tp->tun_flags |= TUN_RWAIT; - error = mtx_sleep(tp, &tp->tun_mtx, PCATCH | (PZERO + 1), - "tunread", 0); - if (error != 0) { - mtx_unlock(&tp->tun_mtx); - return (error); - } - } - } while (m == NULL); - mtx_unlock(&tp->tun_mtx); - - while (m && uio->uio_resid > 0 && error == 0) { - len = min(uio->uio_resid, m->m_len); - if (len != 0) - error = uiomove(mtod(m, void *), len, uio); - m = m_free(m); - } - - if (m) { - TUNDEBUG(ifp, "Dropping mbuf\n"); - m_freem(m); - } - return (error); -} - -/* - * the cdevsw write interface - an atomic write is a packet - or else! - */ -static int -tunwrite(struct cdev *dev, struct uio *uio, int flag) -{ - struct tun_softc *tp = dev->si_drv1; - struct ifnet *ifp = TUN2IFP(tp); - struct mbuf *m; - uint32_t family, mru; - int isr; - - TUNDEBUG(ifp, "tunwrite\n"); - - if ((ifp->if_flags & IFF_UP) != IFF_UP) - /* ignore silently */ - return (0); - - if (uio->uio_resid == 0) - return (0); - - mru = TUNMRU; - if (tp->tun_flags & TUN_IFHEAD) - mru += sizeof(family); - if (uio->uio_resid < 0 || uio->uio_resid > mru) { - TUNDEBUG(ifp, "len=%zd!\n", uio->uio_resid); - return (EIO); - } - - if ((m = m_uiotombuf(uio, M_NOWAIT, 0, 0, M_PKTHDR)) == NULL) { - if_inc_counter(ifp, IFCOUNTER_IERRORS, 1); - return (ENOBUFS); - } - - m->m_pkthdr.rcvif = ifp; -#ifdef MAC - mac_ifnet_create_mbuf(ifp, m); -#endif - - /* Could be unlocked read? */ - mtx_lock(&tp->tun_mtx); - if (tp->tun_flags & TUN_IFHEAD) { - mtx_unlock(&tp->tun_mtx); - if (m->m_len < sizeof(family) && - (m = m_pullup(m, sizeof(family))) == NULL) - return (ENOBUFS); - family = ntohl(*mtod(m, u_int32_t *)); - m_adj(m, sizeof(family)); - } else { - mtx_unlock(&tp->tun_mtx); - family = AF_INET; - } - - BPF_MTAP2(ifp, &family, sizeof(family), m); - - switch (family) { -#ifdef INET - case AF_INET: - isr = NETISR_IP; - break; -#endif -#ifdef INET6 - case AF_INET6: - isr = NETISR_IPV6; - break; -#endif - default: - m_freem(m); - return (EAFNOSUPPORT); - } - random_harvest_queue(m, sizeof(*m), RANDOM_NET_TUN); - if_inc_counter(ifp, IFCOUNTER_IBYTES, m->m_pkthdr.len); - if_inc_counter(ifp, IFCOUNTER_IPACKETS, 1); - CURVNET_SET(ifp->if_vnet); - M_SETFIB(m, ifp->if_fib); - netisr_dispatch(isr, m); - CURVNET_RESTORE(); - return (0); -} - -/* - * 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(struct cdev *dev, int events, struct thread *td) -{ - struct tun_softc *tp = dev->si_drv1; - struct ifnet *ifp = TUN2IFP(tp); - int revents = 0; - struct mbuf *m; - - TUNDEBUG(ifp, "tunpoll\n"); - - if (events & (POLLIN | POLLRDNORM)) { - IFQ_LOCK(&ifp->if_snd); - IFQ_POLL_NOLOCK(&ifp->if_snd, m); - if (m != NULL) { - TUNDEBUG(ifp, "tunpoll q=%d\n", ifp->if_snd.ifq_len); - revents |= events & (POLLIN | POLLRDNORM); - } else { - TUNDEBUG(ifp, "tunpoll waiting\n"); - selrecord(td, &tp->tun_rsel); - } - IFQ_UNLOCK(&ifp->if_snd); - } - if (events & (POLLOUT | POLLWRNORM)) - revents |= events & (POLLOUT | POLLWRNORM); - - return (revents); -} - -/* - * tunkqfilter - support for the kevent() system call. - */ -static int -tunkqfilter(struct cdev *dev, struct knote *kn) -{ - struct tun_softc *tp = dev->si_drv1; - struct ifnet *ifp = TUN2IFP(tp); - - switch(kn->kn_filter) { - case EVFILT_READ: - TUNDEBUG(ifp, "%s kqfilter: EVFILT_READ, minor = %#x\n", - ifp->if_xname, dev2unit(dev)); - kn->kn_fop = &tun_read_filterops; - break; - - case EVFILT_WRITE: - TUNDEBUG(ifp, "%s kqfilter: EVFILT_WRITE, minor = %#x\n", - ifp->if_xname, dev2unit(dev)); - kn->kn_fop = &tun_write_filterops; - break; - - default: - TUNDEBUG(ifp, "%s kqfilter: invalid filter, minor = %#x\n", - ifp->if_xname, dev2unit(dev)); - return(EINVAL); - } - - kn->kn_hook = tp; - knlist_add(&tp->tun_rsel.si_note, kn, 0); - - return (0); -} - -/* - * Return true of there is data in the interface queue. - */ -static int -tunkqread(struct knote *kn, long hint) -{ - int ret; - struct tun_softc *tp = kn->kn_hook; - struct cdev *dev = tp->tun_dev; - struct ifnet *ifp = TUN2IFP(tp); - - if ((kn->kn_data = ifp->if_snd.ifq_len) > 0) { - TUNDEBUG(ifp, - "%s have data in the queue. Len = %d, minor = %#x\n", - ifp->if_xname, ifp->if_snd.ifq_len, dev2unit(dev)); - ret = 1; - } else { - TUNDEBUG(ifp, - "%s waiting for data, minor = %#x\n", ifp->if_xname, - dev2unit(dev)); - ret = 0; - } - - return (ret); -} - -/* - * Always can write, always return MTU in kn->data. - */ -static int -tunkqwrite(struct knote *kn, long hint) -{ - struct tun_softc *tp = kn->kn_hook; - struct ifnet *ifp = TUN2IFP(tp); - - kn->kn_data = ifp->if_mtu; - - return (1); -} - -static void -tunkqdetach(struct knote *kn) -{ - struct tun_softc *tp = kn->kn_hook; - - knlist_remove(&tp->tun_rsel.si_note, kn, 0); -} diff --git a/sys/net/if_tuntap.c b/sys/net/if_tuntap.c new file mode 100644 index 000000000000..41fc48cdcad7 --- /dev/null +++ b/sys/net/if_tuntap.c @@ -0,0 +1,1716 @@ +/* $NetBSD: if_tun.c,v 1.14 1994/06/29 06:36:25 cgd Exp $ */ +/*- + * SPDX-License-Identifier: BSD-2-Clause-FreeBSD + * + * Copyright (C) 1999-2000 by Maksim Yevmenkin + * All rights reserved. + * Copyright (c) 2019 Kyle Evans + * + * 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. + * + * 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 "opt_inet6.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 + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#ifdef INET +#include +#endif +#include +#include +#include + +#include +#include +#include + +struct tuntap_driver; + +/* + * tun_list is protected by global tunmtx. Other mutable fields are + * protected by tun->tun_mtx, or by their owning subsystem. tun_dev is + * static for the duration of a tunnel interface. + */ +struct tuntap_softc { + TAILQ_ENTRY(tuntap_softc) tun_list; + struct cdev *tun_dev; + u_short tun_flags; /* misc flags */ +#define TUN_OPEN 0x0001 +#define TUN_INITED 0x0002 +#define TUN_RCOLL 0x0004 +#define TUN_IASET 0x0008 +#define TUN_DSTADDR 0x0010 +#define TUN_LMODE 0x0020 +#define TUN_RWAIT 0x0040 +#define TUN_ASYNC 0x0080 +#define TUN_IFHEAD 0x0100 +#define TUN_DYING 0x0200 +#define TUN_L2 0x0400 +#define TUN_VMNET 0x0800 + +#define TUN_READY (TUN_OPEN | TUN_INITED) + + pid_t tun_pid; /* owning pid */ + struct ifnet *tun_ifp; /* the interface */ + struct sigio *tun_sigio; /* information for async I/O */ + struct tuntap_driver *tun_drv; /* appropriate driver */ + struct selinfo tun_rsel; /* read select */ + struct mtx tun_mtx; /* protect mutable softc fields */ + struct cv tun_cv; /* protect against ref'd dev destroy */ + struct ether_addr tun_ether; /* remote address */ +}; +#define TUN2IFP(sc) ((sc)->tun_ifp) + +#define TUNDEBUG if (tundebug) if_printf + +#define TUN_LOCK(tp) mtx_lock(&(tp)->tun_mtx) +#define TUN_UNLOCK(tp) mtx_unlock(&(tp)->tun_mtx) + +#define TUN_VMIO_FLAG_MASK 0x0fff + +/* + * All mutable global variables in if_tun are locked using tunmtx, with + * the exception of tundebug, which is used unlocked, and the drivers' *clones, + * which are static after setup. + */ +static struct mtx tunmtx; +static eventhandler_tag tag; +static const char tunname[] = "tun"; +static const char tapname[] = "tap"; +static const char vmnetname[] = "vmnet"; +static MALLOC_DEFINE(M_TUN, tunname, "Tunnel Interface"); +static int tundebug = 0; +static int tundclone = 1; +static int tap_allow_uopen = 0; /* allow user open() */ +static int tapuponopen = 0; /* IFF_UP on open() */ +static int tapdclone = 1; /* enable devfs cloning */ + +static TAILQ_HEAD(,tuntap_softc) tunhead = TAILQ_HEAD_INITIALIZER(tunhead); +SYSCTL_INT(_debug, OID_AUTO, if_tun_debug, CTLFLAG_RW, &tundebug, 0, ""); + +static struct sx tun_ioctl_sx; +SX_SYSINIT(tun_ioctl_sx, &tun_ioctl_sx, "tun_ioctl"); + +SYSCTL_DECL(_net_link); +/* tun */ +static SYSCTL_NODE(_net_link, OID_AUTO, tun, CTLFLAG_RW, 0, + "IP tunnel software network interface."); +SYSCTL_INT(_net_link_tun, OID_AUTO, devfs_cloning, CTLFLAG_RWTUN, &tundclone, 0, + "Enable legacy devfs interface creation."); + +/* tap */ +static SYSCTL_NODE(_net_link, OID_AUTO, tap, CTLFLAG_RW, 0, + "Ethernet tunnel software network interface"); +SYSCTL_INT(_net_link_tap, OID_AUTO, user_open, CTLFLAG_RW, &tap_allow_uopen, 0, + "Allow user to open /dev/tap (based on node permissions)"); +SYSCTL_INT(_net_link_tap, OID_AUTO, up_on_open, CTLFLAG_RW, &tapuponopen, 0, + "Bring interface up when /dev/tap is opened"); +SYSCTL_INT(_net_link_tap, OID_AUTO, devfs_cloning, CTLFLAG_RWTUN, &tapdclone, 0, + "Enable legacy devfs interface creation"); +SYSCTL_INT(_net_link_tap, OID_AUTO, debug, CTLFLAG_RW, &tundebug, 0, ""); + +static int tuntap_name2info(const char *name, int *unit, int *flags); +static void tunclone(void *arg, struct ucred *cred, char *name, + int namelen, struct cdev **dev); +static void tuncreate(struct cdev *dev, struct tuntap_driver *); +static int tunifioctl(struct ifnet *, u_long, caddr_t); +static void tuninit(struct ifnet *); +static void tunifinit(void *xtp); +static int tuntapmodevent(module_t, int, void *); +static int tunoutput(struct ifnet *, struct mbuf *, + const struct sockaddr *, struct route *ro); +static void tunstart(struct ifnet *); +static void tunstart_l2(struct ifnet *); + +static int tun_clone_match(struct if_clone *ifc, const char *name); +static int tap_clone_match(struct if_clone *ifc, const char *name); +static int vmnet_clone_match(struct if_clone *ifc, const char *name); +static int tun_clone_create(struct if_clone *, char *, size_t, caddr_t); +static int tun_clone_destroy(struct if_clone *, 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; +static d_kqfilter_t tunkqfilter; + +static int tunkqread(struct knote *, long); +static int tunkqwrite(struct knote *, long); +static void tunkqdetach(struct knote *); + +static struct filterops tun_read_filterops = { + .f_isfd = 1, + .f_attach = NULL, + .f_detach = tunkqdetach, + .f_event = tunkqread, +}; + +static struct filterops tun_write_filterops = { + .f_isfd = 1, + .f_attach = NULL, + .f_detach = tunkqdetach, + .f_event = tunkqwrite, +}; + +#define TUN_DRIVER_IDENT_MASK (TUN_L2 | TUN_VMNET) + +static struct tuntap_driver { + int tun_flags; + struct unrhdr *unrhdr; + struct cdevsw cdevsw; + struct clonedevs *clones; + ifc_match_t *clone_match_fn; + ifc_create_t *clone_create_fn; + ifc_destroy_t *clone_destroy_fn; +} tuntap_drivers[] = { + { + .tun_flags = 0, + .cdevsw = { + .d_version = D_VERSION, + .d_flags = D_NEEDMINOR, + .d_open = tunopen, + .d_close = tunclose, + .d_read = tunread, + .d_write = tunwrite, + .d_ioctl = tunioctl, + .d_poll = tunpoll, + .d_kqfilter = tunkqfilter, + .d_name = tunname, + }, + .clone_match_fn = tun_clone_match, + .clone_create_fn = tun_clone_create, + .clone_destroy_fn = tun_clone_destroy, + }, + { + .tun_flags = TUN_L2, + .cdevsw = { + .d_version = D_VERSION, + .d_flags = D_NEEDMINOR, + .d_open = tunopen, + .d_close = tunclose, + .d_read = tunread, + .d_write = tunwrite, + .d_ioctl = tunioctl, + .d_poll = tunpoll, + .d_kqfilter = tunkqfilter, + .d_name = tapname, + }, + .clone_match_fn = tap_clone_match, + .clone_create_fn = tun_clone_create, + .clone_destroy_fn = tun_clone_destroy, + }, + { + .tun_flags = TUN_L2 | TUN_VMNET, + .cdevsw = { + .d_version = D_VERSION, + .d_flags = D_NEEDMINOR, + .d_open = tunopen, + .d_close = tunclose, + .d_read = tunread, + .d_write = tunwrite, + .d_ioctl = tunioctl, + .d_poll = tunpoll, + .d_kqfilter = tunkqfilter, + .d_name = vmnetname, + }, + .clone_match_fn = vmnet_clone_match, + .clone_create_fn = tun_clone_create, + .clone_destroy_fn = tun_clone_destroy, + }, +}; + +struct tuntap_driver_cloner { + SLIST_ENTRY(tuntap_driver_cloner) link; + struct tuntap_driver *drv; + struct if_clone *cloner; +}; + +VNET_DEFINE_STATIC(SLIST_HEAD(, tuntap_driver_cloner), tuntap_driver_cloners) = + SLIST_HEAD_INITIALIZER(tuntap_driver_cloners); + +#define V_tuntap_driver_cloners VNET(tuntap_driver_cloners) + +/* + * Sets unit and/or flags given the device name. Must be called with correct + * vnet context. + */ +static int +tuntap_name2info(const char *name, int *outunit, int *outflags) +{ + struct tuntap_driver *drv; + struct tuntap_driver_cloner *drvc; + char *dname; + int flags, unit; + bool found; + + if (name == NULL) + return (EINVAL); + + /* + * Needed for dev_stdclone, but dev_stdclone will not modify, it just + * wants to be able to pass back a char * through the second param. We + * will always set that as NULL here, so we'll fake it. + */ + dname = __DECONST(char *, name); + found = false; + + KASSERT(!SLIST_EMPTY(&V_tuntap_driver_cloners), + ("tuntap_driver_cloners failed to initialize")); + SLIST_FOREACH(drvc, &V_tuntap_driver_cloners, link) { + KASSERT(drvc->drv != NULL, + ("tuntap_driver_cloners entry not properly initialized")); + drv = drvc->drv; + + if (strcmp(name, drv->cdevsw.d_name) == 0) { + found = true; + unit = -1; + flags = drv->tun_flags; + break; + } + + if (dev_stdclone(dname, NULL, drv->cdevsw.d_name, &unit) == 1) { + found = true; + flags = drv->tun_flags; + break; + } + } + + if (!found) + return (ENXIO); + + if (outunit != NULL) + *outunit = unit; + if (outflags != NULL) + *outflags = flags; + return (0); +} + +/* + * Get driver information from a set of flags specified. Masks the identifying + * part of the flags and compares it against all of the available + * tuntap_drivers. Must be called with correct vnet context. + */ +static struct tuntap_driver * +tuntap_driver_from_flags(int tun_flags) +{ + struct tuntap_driver *drv; + struct tuntap_driver_cloner *drvc; + + KASSERT(!SLIST_EMPTY(&V_tuntap_driver_cloners), + ("tuntap_driver_cloners failed to initialize")); + SLIST_FOREACH(drvc, &V_tuntap_driver_cloners, link) { + KASSERT(drvc->drv != NULL, + ("tuntap_driver_cloners entry not properly initialized")); + drv = drvc->drv; + if ((tun_flags & TUN_DRIVER_IDENT_MASK) == drv->tun_flags) + return (drv); + } + + return (NULL); +} + + + +static int +tun_clone_match(struct if_clone *ifc, const char *name) +{ + int tunflags; + + if (tuntap_name2info(name, NULL, &tunflags) == 0) { + if ((tunflags & TUN_L2) == 0) + return (1); + } + + return (0); +} + +static int +tap_clone_match(struct if_clone *ifc, const char *name) +{ + int tunflags; + + if (tuntap_name2info(name, NULL, &tunflags) == 0) { + if ((tunflags & (TUN_L2 | TUN_VMNET)) == TUN_L2) + return (1); + } + + return (0); +} + +static int +vmnet_clone_match(struct if_clone *ifc, const char *name) +{ + int tunflags; + + if (tuntap_name2info(name, NULL, &tunflags) == 0) { + if ((tunflags & TUN_VMNET) != 0) + return (1); + } + + return (0); +} + +static int +tun_clone_create(struct if_clone *ifc, char *name, size_t len, caddr_t params) +{ + struct tuntap_driver *drv; + struct cdev *dev; + int err, i, tunflags, unit; + + tunflags = 0; + /* The name here tells us exactly what we're creating */ + err = tuntap_name2info(name, &unit, &tunflags); + if (err != 0) + return (err); + + drv = tuntap_driver_from_flags(tunflags); + if (drv == NULL) + return (ENXIO); + + if (unit != -1) { + /* If this unit number is still available that/s okay. */ + if (alloc_unr_specific(drv->unrhdr, unit) == -1) + return (EEXIST); + } else { + unit = alloc_unr(drv->unrhdr); + } + + snprintf(name, IFNAMSIZ, "%s%d", drv->cdevsw.d_name, unit); + + /* find any existing device, or allocate new unit number */ + i = clone_create(&drv->clones, &drv->cdevsw, &unit, &dev, 0); + if (i) { + /* No preexisting struct cdev *, create one */ + dev = make_dev(&drv->cdevsw, unit, UID_UUCP, GID_DIALER, 0600, + "%s%d", drv->cdevsw.d_name, unit); + } + + tuncreate(dev, drv); + + return (0); +} + +static void +tunclone(void *arg, struct ucred *cred, char *name, int namelen, + struct cdev **dev) +{ + char devname[SPECNAMELEN + 1]; + struct tuntap_driver *drv; + int append_unit, i, u, tunflags; + bool mayclone; + + if (*dev != NULL) + return; + + tunflags = 0; + CURVNET_SET(CRED_TO_VNET(cred)); + if (tuntap_name2info(name, &u, &tunflags) != 0) + goto out; /* Not recognized */ + + if (u != -1 && u > IF_MAXUNIT) + goto out; /* Unit number too high */ + + mayclone = priv_check_cred(cred, PRIV_NET_IFCREATE) == 0; + if ((tunflags & TUN_L2) != 0) { + /* tap/vmnet allow user open with a sysctl */ + mayclone = (mayclone || tap_allow_uopen) && tapdclone; + } else { + mayclone = mayclone && tundclone; + } + + /* + * If tun cloning is enabled, only the superuser can create an + * interface. + */ + if (!mayclone) + goto out; + + if (u == -1) + append_unit = 1; + else + append_unit = 0; + + drv = tuntap_driver_from_flags(tunflags); + if (drv == NULL) + goto out; + + /* find any existing device, or allocate new unit number */ + i = clone_create(&drv->clones, &drv->cdevsw, &u, dev, 0); + if (i) { + if (append_unit) { + namelen = snprintf(devname, sizeof(devname), "%s%d", + name, u); + name = devname; + } + /* No preexisting struct cdev *, create one */ + *dev = make_dev_credf(MAKEDEV_REF, &drv->cdevsw, u, cred, + UID_UUCP, GID_DIALER, 0600, "%s", name); + } + + if_clone_create(name, namelen, NULL); +out: + CURVNET_RESTORE(); +} + +static void +tun_destroy(struct tuntap_softc *tp) +{ + struct cdev *dev; + + TUN_LOCK(tp); + tp->tun_flags |= TUN_DYING; + if ((tp->tun_flags & TUN_OPEN) != 0) + cv_wait_unlock(&tp->tun_cv, &tp->tun_mtx); + else + TUN_UNLOCK(tp); + + CURVNET_SET(TUN2IFP(tp)->if_vnet); + sx_xlock(&tun_ioctl_sx); + TUN2IFP(tp)->if_softc = NULL; + sx_xunlock(&tun_ioctl_sx); + + dev = tp->tun_dev; + bpfdetach(TUN2IFP(tp)); + if_detach(TUN2IFP(tp)); + free_unr(tp->tun_drv->unrhdr, TUN2IFP(tp)->if_dunit); + if_free(TUN2IFP(tp)); + destroy_dev(dev); + seldrain(&tp->tun_rsel); + knlist_clear(&tp->tun_rsel.si_note, 0); + knlist_destroy(&tp->tun_rsel.si_note); + mtx_destroy(&tp->tun_mtx); + cv_destroy(&tp->tun_cv); + free(tp, M_TUN); + CURVNET_RESTORE(); +} + +static int +tun_clone_destroy(struct if_clone *ifc __unused, struct ifnet *ifp) +{ + struct tuntap_softc *tp = ifp->if_softc; + + mtx_lock(&tunmtx); + TAILQ_REMOVE(&tunhead, tp, tun_list); + mtx_unlock(&tunmtx); + tun_destroy(tp); + + return (0); +} + +static void +vnet_tun_init(const void *unused __unused) +{ + struct tuntap_driver *drv; + struct tuntap_driver_cloner *drvc; + int i; + + for (i = 0; i < nitems(tuntap_drivers); ++i) { + drv = &tuntap_drivers[i]; + drvc = malloc(sizeof(*drvc), M_TUN, M_WAITOK | M_ZERO); + + drvc->drv = drv; + drvc->cloner = if_clone_advanced(drv->cdevsw.d_name, 0, + drv->clone_match_fn, drv->clone_create_fn, + drv->clone_destroy_fn); + SLIST_INSERT_HEAD(&V_tuntap_driver_cloners, drvc, link); + }; +} +VNET_SYSINIT(vnet_tun_init, SI_SUB_PROTO_IF, SI_ORDER_ANY, + vnet_tun_init, NULL); + +static void +vnet_tun_uninit(const void *unused __unused) +{ + struct tuntap_driver_cloner *drvc; + + while (!SLIST_EMPTY(&V_tuntap_driver_cloners)) { + drvc = SLIST_FIRST(&V_tuntap_driver_cloners); + SLIST_REMOVE_HEAD(&V_tuntap_driver_cloners, link); + + if_clone_detach(drvc->cloner); + free(drvc, M_TUN); + } +} +VNET_SYSUNINIT(vnet_tun_uninit, SI_SUB_PROTO_IF, SI_ORDER_ANY, + vnet_tun_uninit, NULL); + +static void +tun_uninit(const void *unused __unused) +{ + struct tuntap_driver *drv; + struct tuntap_softc *tp; + int i; + + EVENTHANDLER_DEREGISTER(dev_clone, tag); + drain_dev_clone_events(); + + mtx_lock(&tunmtx); + while ((tp = TAILQ_FIRST(&tunhead)) != NULL) { + TAILQ_REMOVE(&tunhead, tp, tun_list); + mtx_unlock(&tunmtx); + tun_destroy(tp); + mtx_lock(&tunmtx); + } + mtx_unlock(&tunmtx); + for (i = 0; i < nitems(tuntap_drivers); ++i) { + drv = &tuntap_drivers[i]; + delete_unrhdr(drv->unrhdr); + clone_cleanup(&drv->clones); + } + mtx_destroy(&tunmtx); +} +SYSUNINIT(tun_uninit, SI_SUB_PROTO_IF, SI_ORDER_ANY, tun_uninit, NULL); + +static int +tuntapmodevent(module_t mod, int type, void *data) +{ + struct tuntap_driver *drv; + int i; + + switch (type) { + case MOD_LOAD: + mtx_init(&tunmtx, "tunmtx", NULL, MTX_DEF); + for (i = 0; i < nitems(tuntap_drivers); ++i) { + drv = &tuntap_drivers[i]; + clone_setup(&drv->clones); + drv->unrhdr = new_unrhdr(0, IF_MAXUNIT, &tunmtx); + } + tag = EVENTHANDLER_REGISTER(dev_clone, tunclone, 0, 1000); + if (tag == NULL) + return (ENOMEM); + break; + case MOD_UNLOAD: + /* See tun_uninit, so it's done after the vnet_sysuninit() */ + break; + default: + return EOPNOTSUPP; + } + return 0; +} + +static moduledata_t tuntap_mod = { + "if_tuntap", + tuntapmodevent, + 0 +}; + +DECLARE_MODULE(if_tuntap, tuntap_mod, SI_SUB_PSEUDO, SI_ORDER_ANY); +MODULE_VERSION(if_tuntap, 1); + +static void +tunstart(struct ifnet *ifp) +{ + struct tuntap_softc *tp = ifp->if_softc; + struct mbuf *m; + + TUNDEBUG(ifp, "starting\n"); + if (ALTQ_IS_ENABLED(&ifp->if_snd)) { + IFQ_LOCK(&ifp->if_snd); + IFQ_POLL_NOLOCK(&ifp->if_snd, m); + if (m == NULL) { + IFQ_UNLOCK(&ifp->if_snd); + return; + } + IFQ_UNLOCK(&ifp->if_snd); + } + + TUN_LOCK(tp); + if (tp->tun_flags & TUN_RWAIT) { + tp->tun_flags &= ~TUN_RWAIT; + wakeup(tp); + } + selwakeuppri(&tp->tun_rsel, PZERO + 1); + KNOTE_LOCKED(&tp->tun_rsel.si_note, 0); + if (tp->tun_flags & TUN_ASYNC && tp->tun_sigio) { + TUN_UNLOCK(tp); + pgsigio(&tp->tun_sigio, SIGIO, 0); + } else + TUN_UNLOCK(tp); +} + +/* + * tunstart_l2 + * + * queue packets from higher level ready to put out + */ +static void +tunstart_l2(struct ifnet *ifp) +{ + struct tuntap_softc *tp = ifp->if_softc; + + TUNDEBUG(ifp, "starting\n"); + + /* + * do not junk pending output if we are in VMnet mode. + * XXX: can this do any harm because of queue overflow? + */ + + TUN_LOCK(tp); + if (((tp->tun_flags & TUN_VMNET) == 0) && + ((tp->tun_flags & TUN_READY) != TUN_READY)) { + struct mbuf *m; + + /* Unlocked read. */ + TUNDEBUG(ifp, "not ready, tun_flags = 0x%x\n", tp->tun_flags); + + for (;;) { + IF_DEQUEUE(&ifp->if_snd, m); + if (m != NULL) { + m_freem(m); + if_inc_counter(ifp, IFCOUNTER_OERRORS, 1); + } else + break; + } + TUN_UNLOCK(tp); + + return; + } + + ifp->if_drv_flags |= IFF_DRV_OACTIVE; + + if (!IFQ_IS_EMPTY(&ifp->if_snd)) { + if (tp->tun_flags & TUN_RWAIT) { + tp->tun_flags &= ~TUN_RWAIT; + wakeup(tp); + } + + if ((tp->tun_flags & TUN_ASYNC) && (tp->tun_sigio != NULL)) { + TUN_UNLOCK(tp); + pgsigio(&tp->tun_sigio, SIGIO, 0); + TUN_LOCK(tp); + } + + selwakeuppri(&tp->tun_rsel, PZERO+1); + KNOTE_LOCKED(&tp->tun_rsel.si_note, 0); + if_inc_counter(ifp, IFCOUNTER_OPACKETS, 1); /* obytes are counted in ether_output */ + } + + ifp->if_drv_flags &= ~IFF_DRV_OACTIVE; + TUN_UNLOCK(tp); +} /* tunstart_l2 */ + + +/* XXX: should return an error code so it can fail. */ +static void +tuncreate(struct cdev *dev, struct tuntap_driver *drv) +{ + struct tuntap_softc *sc; + struct ifnet *ifp; + struct ether_addr eaddr; + int iflags; + u_char type; + + sc = malloc(sizeof(*sc), M_TUN, M_WAITOK | M_ZERO); + mtx_init(&sc->tun_mtx, "tun_mtx", NULL, MTX_DEF); + cv_init(&sc->tun_cv, "tun_condvar"); + sc->tun_flags = drv->tun_flags; + sc->tun_dev = dev; + sc->tun_drv = drv; + mtx_lock(&tunmtx); + TAILQ_INSERT_TAIL(&tunhead, sc, tun_list); + mtx_unlock(&tunmtx); + + iflags = IFF_MULTICAST; + if ((sc->tun_flags & TUN_L2) != 0) { + type = IFT_ETHER; + iflags |= IFF_BROADCAST | IFF_SIMPLEX; + } else { + type = IFT_PPP; + iflags |= IFF_POINTOPOINT; + } + ifp = sc->tun_ifp = if_alloc(type); + if (ifp == NULL) + panic("%s%d: failed to if_alloc() interface.\n", + drv->cdevsw.d_name, dev2unit(dev)); + ifp->if_softc = sc; + if_initname(ifp, drv->cdevsw.d_name, dev2unit(dev)); + ifp->if_ioctl = tunifioctl; + ifp->if_flags = iflags; + IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen); + knlist_init_mtx(&sc->tun_rsel.si_note, &sc->tun_mtx); + ifp->if_capabilities |= IFCAP_LINKSTATE; + ifp->if_capenable |= IFCAP_LINKSTATE; + + if ((sc->tun_flags & TUN_L2) != 0) { + ifp->if_mtu = ETHERMTU; + ifp->if_init = tunifinit; + ifp->if_start = tunstart_l2; + + ether_gen_addr(ifp, &eaddr); + ether_ifattach(ifp, eaddr.octet); + } else { + ifp->if_mtu = TUNMTU; + ifp->if_start = tunstart; + ifp->if_output = tunoutput; + + ifp->if_snd.ifq_drv_maxlen = 0; + IFQ_SET_READY(&ifp->if_snd); + + if_attach(ifp); + bpfattach(ifp, DLT_NULL, sizeof(u_int32_t)); + } + dev->si_drv1 = sc; + + TUN_LOCK(sc); + sc->tun_flags |= TUN_INITED; + TUN_UNLOCK(sc); + + TUNDEBUG(ifp, "interface %s is created, minor = %#x\n", + ifp->if_xname, dev2unit(dev)); +} + +static int +tunopen(struct cdev *dev, int flag, int mode, struct thread *td) +{ + struct ifnet *ifp; + struct tuntap_driver *drv; + struct tuntap_softc *tp; + int error, tunflags; + + tunflags = 0; + CURVNET_SET(TD_TO_VNET(td)); + error = tuntap_name2info(dev->si_name, NULL, &tunflags); + if (error != 0) { + CURVNET_RESTORE(); + return (error); /* Shouldn't happen */ + } + + if ((tunflags & TUN_L2) != 0) { + /* Restrict? */ + if (tap_allow_uopen == 0) { + error = priv_check(td, PRIV_NET_TAP); + if (error != 0) { + CURVNET_RESTORE(); + return (error); + } + } + } + + /* + * XXXRW: Non-atomic test and set of dev->si_drv1 requires + * synchronization. + */ + tp = dev->si_drv1; + if (!tp) { + drv = tuntap_driver_from_flags(tunflags); + if (drv == NULL) { + CURVNET_RESTORE(); + return (ENXIO); + } + tuncreate(dev, drv); + tp = dev->si_drv1; + } + + TUN_LOCK(tp); + if ((tp->tun_flags & (TUN_OPEN | TUN_DYING)) != 0) { + TUN_UNLOCK(tp); + CURVNET_RESTORE(); + return (EBUSY); + } + + ifp = TUN2IFP(tp); + + if ((tp->tun_flags & TUN_L2) != 0) { + bcopy(IF_LLADDR(ifp), tp->tun_ether.octet, + sizeof(tp->tun_ether.octet)); + + ifp->if_drv_flags |= IFF_DRV_RUNNING; + ifp->if_drv_flags &= ~IFF_DRV_OACTIVE; + + if (tapuponopen) + ifp->if_flags |= IFF_UP; + } + + tp->tun_pid = td->td_proc->p_pid; + tp->tun_flags |= TUN_OPEN; + + if_link_state_change(ifp, LINK_STATE_UP); + TUNDEBUG(ifp, "open\n"); + TUN_UNLOCK(tp); + CURVNET_RESTORE(); + return (0); +} + +/* + * tunclose - close the device - mark i/f down & delete + * routing info + */ +static int +tunclose(struct cdev *dev, int foo, int bar, struct thread *td) +{ + struct tuntap_softc *tp; + struct ifnet *ifp; + bool l2tun; + + tp = dev->si_drv1; + ifp = TUN2IFP(tp); + + TUN_LOCK(tp); + /* + * Simply close the device if this isn't the controlling process. This + * may happen if, for instance, the tunnel has been handed off to + * another process. The original controller should be able to close it + * without putting us into an inconsistent state. + */ + if (td->td_proc->p_pid != tp->tun_pid) { + TUN_UNLOCK(tp); + return (0); + } + + /* + * junk all pending output + */ + CURVNET_SET(ifp->if_vnet); + + l2tun = false; + if ((tp->tun_flags & TUN_L2) != 0) { + l2tun = true; + IF_DRAIN(&ifp->if_snd); + } else { + IFQ_PURGE(&ifp->if_snd); + } + + /* For vmnet, we won't do most of the address/route bits */ + if ((tp->tun_flags & TUN_VMNET) != 0) + goto out; + + if (ifp->if_flags & IFF_UP) { + TUN_UNLOCK(tp); + if_down(ifp); + TUN_LOCK(tp); + } + + /* Delete all addresses and routes which reference this interface. */ + if (ifp->if_drv_flags & IFF_DRV_RUNNING) { + struct ifaddr *ifa; + + ifp->if_drv_flags &= ~IFF_DRV_RUNNING; + TUN_UNLOCK(tp); + CK_STAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { + /* deal w/IPv4 PtP destination; unlocked read */ + if (!l2tun && ifa->ifa_addr->sa_family == AF_INET) { + rtinit(ifa, (int)RTM_DELETE, + tp->tun_flags & TUN_DSTADDR ? RTF_HOST : 0); + } else { + rtinit(ifa, (int)RTM_DELETE, 0); + } + } + if_purgeaddrs(ifp); + TUN_LOCK(tp); + } + +out: + if_link_state_change(ifp, LINK_STATE_DOWN); + CURVNET_RESTORE(); + + funsetown(&tp->tun_sigio); + selwakeuppri(&tp->tun_rsel, PZERO + 1); + KNOTE_LOCKED(&tp->tun_rsel.si_note, 0); + TUNDEBUG (ifp, "closed\n"); + tp->tun_flags &= ~TUN_OPEN; + tp->tun_pid = 0; + + cv_broadcast(&tp->tun_cv); + TUN_UNLOCK(tp); + return (0); +} + +static void +tuninit(struct ifnet *ifp) +{ + struct tuntap_softc *tp = ifp->if_softc; +#ifdef INET + struct ifaddr *ifa; +#endif + + TUNDEBUG(ifp, "tuninit\n"); + + TUN_LOCK(tp); + ifp->if_drv_flags |= IFF_DRV_RUNNING; + if ((tp->tun_flags & TUN_L2) == 0) { + ifp->if_flags |= IFF_UP; + getmicrotime(&ifp->if_lastchange); +#ifdef INET + if_addr_rlock(ifp); + CK_STAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { + 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; + } + } + if_addr_runlock(ifp); +#endif + TUN_UNLOCK(tp); + } else { + ifp->if_drv_flags &= ~IFF_DRV_OACTIVE; + TUN_UNLOCK(tp); + /* attempt to start output */ + tunstart_l2(ifp); + } + +} + +/* + * Used only for l2 tunnel. + */ +static void +tunifinit(void *xtp) +{ + struct tuntap_softc *tp; + + tp = (struct tuntap_softc *)xtp; + tuninit(tp->tun_ifp); +} + +/* + * Process an ioctl request. + */ +static int +tunifioctl(struct ifnet *ifp, u_long cmd, caddr_t data) +{ + struct ifreq *ifr = (struct ifreq *)data; + struct tuntap_softc *tp; + struct ifstat *ifs; + struct ifmediareq *ifmr; + int dummy, error = 0; + bool l2tun; + + ifmr = NULL; + sx_xlock(&tun_ioctl_sx); + tp = ifp->if_softc; + if (tp == NULL) { + error = ENXIO; + goto bad; + } + l2tun = (tp->tun_flags & TUN_L2) != 0; + switch(cmd) { + case SIOCGIFSTATUS: + ifs = (struct ifstat *)data; + TUN_LOCK(tp); + if (tp->tun_pid) + snprintf(ifs->ascii, sizeof(ifs->ascii), + "\tOpened by PID %d\n", tp->tun_pid); + else + ifs->ascii[0] = '\0'; + TUN_UNLOCK(tp); + break; + case SIOCSIFADDR: + if (l2tun) + error = ether_ioctl(ifp, cmd, data); + else + tuninit(ifp); + if (error == 0) + TUNDEBUG(ifp, "address set\n"); + break; + case SIOCSIFMTU: + ifp->if_mtu = ifr->ifr_mtu; + TUNDEBUG(ifp, "mtu set\n"); + break; + case SIOCSIFFLAGS: + case SIOCADDMULTI: + case SIOCDELMULTI: + break; + case SIOCGIFMEDIA: + if (!l2tun) { + error = EINVAL; + break; + } + + ifmr = (struct ifmediareq *)data; + dummy = ifmr->ifm_count; + ifmr->ifm_count = 1; + ifmr->ifm_status = IFM_AVALID; + ifmr->ifm_active = IFM_ETHER; + if (tp->tun_flags & TUN_OPEN) + ifmr->ifm_status |= IFM_ACTIVE; + ifmr->ifm_current = ifmr->ifm_active; + if (dummy >= 1) { + int media = IFM_ETHER; + error = copyout(&media, ifmr->ifm_ulist, sizeof(int)); + } + break; + default: + if (l2tun) { + error = ether_ioctl(ifp, cmd, data); + } else { + error = EINVAL; + } + } +bad: + sx_xunlock(&tun_ioctl_sx); + return (error); +} + +/* + * tunoutput - queue packets from higher level ready to put out. + */ +static int +tunoutput(struct ifnet *ifp, struct mbuf *m0, const struct sockaddr *dst, + struct route *ro) +{ + struct tuntap_softc *tp = ifp->if_softc; + u_short cached_tun_flags; + int error; + u_int32_t af; + + TUNDEBUG (ifp, "tunoutput\n"); + +#ifdef MAC + error = mac_ifnet_check_transmit(ifp, m0); + if (error) { + m_freem(m0); + return (error); + } +#endif + + /* Could be unlocked read? */ + TUN_LOCK(tp); + cached_tun_flags = tp->tun_flags; + TUN_UNLOCK(tp); + if ((cached_tun_flags & TUN_READY) != TUN_READY) { + TUNDEBUG (ifp, "not ready 0%o\n", tp->tun_flags); + m_freem (m0); + return (EHOSTDOWN); + } + + if ((ifp->if_flags & IFF_UP) != IFF_UP) { + m_freem (m0); + return (EHOSTDOWN); + } + + /* BPF writes need to be handled specially. */ + if (dst->sa_family == AF_UNSPEC) + bcopy(dst->sa_data, &af, sizeof(af)); + else + af = dst->sa_family; + + if (bpf_peers_present(ifp->if_bpf)) + bpf_mtap2(ifp->if_bpf, &af, sizeof(af), m0); + + /* prepend sockaddr? this may abort if the mbuf allocation fails */ + if (cached_tun_flags & TUN_LMODE) { + /* allocate space for sockaddr */ + M_PREPEND(m0, dst->sa_len, M_NOWAIT); + + /* if allocation failed drop packet */ + if (m0 == NULL) { + if_inc_counter(ifp, IFCOUNTER_IQDROPS, 1); + if_inc_counter(ifp, IFCOUNTER_OERRORS, 1); + return (ENOBUFS); + } else { + bcopy(dst, m0->m_data, dst->sa_len); + } + } + + if (cached_tun_flags & TUN_IFHEAD) { + /* Prepend the address family */ + M_PREPEND(m0, 4, M_NOWAIT); + + /* if allocation failed drop packet */ + if (m0 == NULL) { + if_inc_counter(ifp, IFCOUNTER_IQDROPS, 1); + if_inc_counter(ifp, IFCOUNTER_OERRORS, 1); + return (ENOBUFS); + } else + *(u_int32_t *)m0->m_data = htonl(af); + } else { +#ifdef INET + if (af != AF_INET) +#endif + { + m_freem(m0); + return (EAFNOSUPPORT); + } + } + + error = (ifp->if_transmit)(ifp, m0); + if (error) + return (ENOBUFS); + if_inc_counter(ifp, IFCOUNTER_OPACKETS, 1); + return (0); +} + +/* + * the cdevsw interface is now pretty minimal. + */ +static int +tunioctl(struct cdev *dev, u_long cmd, caddr_t data, int flag, + struct thread *td) +{ + struct ifreq ifr, *ifrp; + struct tuntap_softc *tp = dev->si_drv1; + struct tuninfo *tunp; + int error, iflags; +#if defined(COMPAT_FREEBSD6) || defined(COMPAT_FREEBSD5) || \ + defined(COMPAT_FREEBSD4) + int ival; +#endif + bool l2tun; + + l2tun = (tp->tun_flags & TUN_L2) != 0; + if (l2tun) { + /* tap specific ioctls */ + switch(cmd) { + case TAPGIFNAME: + ifrp = (struct ifreq *)data; + strlcpy(ifrp->ifr_name, TUN2IFP(tp)->if_xname, + IFNAMSIZ); + + return (0); + /* VMware/VMnet port ioctl's */ +#if defined(COMPAT_FREEBSD6) || defined(COMPAT_FREEBSD5) || \ + defined(COMPAT_FREEBSD4) + case _IO('V', 0): + ival = IOCPARM_IVAL(data); + data = (caddr_t)&ival; + /* FALLTHROUGH */ +#endif + case VMIO_SIOCSIFFLAGS: /* VMware/VMnet SIOCSIFFLAGS */ + iflags = *(int *)data; + iflags &= TUN_VMIO_FLAG_MASK; + iflags &= ~IFF_CANTCHANGE; + iflags |= IFF_UP; + + TUN_LOCK(tp); + TUN2IFP(tp)->if_flags = iflags | + (TUN2IFP(tp)->if_flags & IFF_CANTCHANGE); + TUN_UNLOCK(tp); + + return (0); + case SIOCGIFADDR: /* get MAC address of the remote side */ + TUN_LOCK(tp); + bcopy(&tp->tun_ether.octet, data, + sizeof(tp->tun_ether.octet)); + TUN_UNLOCK(tp); + + return (0); + case SIOCSIFADDR: /* set MAC address of the remote side */ + TUN_LOCK(tp); + bcopy(data, &tp->tun_ether.octet, + sizeof(tp->tun_ether.octet)); + TUN_UNLOCK(tp); + + return (0); + } + + /* Fall through to the common ioctls if unhandled */ + } else { + switch (cmd) { + case TUNSLMODE: + TUN_LOCK(tp); + if (*(int *)data) { + tp->tun_flags |= TUN_LMODE; + tp->tun_flags &= ~TUN_IFHEAD; + } else + tp->tun_flags &= ~TUN_LMODE; + TUN_UNLOCK(tp); + + return (0); + case TUNSIFHEAD: + TUN_LOCK(tp); + if (*(int *)data) { + tp->tun_flags |= TUN_IFHEAD; + tp->tun_flags &= ~TUN_LMODE; + } else + tp->tun_flags &= ~TUN_IFHEAD; + TUN_UNLOCK(tp); + + return (0); + case TUNGIFHEAD: + TUN_LOCK(tp); + *(int *)data = (tp->tun_flags & TUN_IFHEAD) ? 1 : 0; + TUN_UNLOCK(tp); + + return (0); + case TUNSIFMODE: + /* deny this if UP */ + if (TUN2IFP(tp)->if_flags & IFF_UP) + return (EBUSY); + + switch (*(int *)data & ~IFF_MULTICAST) { + case IFF_POINTOPOINT: + case IFF_BROADCAST: + TUN_LOCK(tp); + TUN2IFP(tp)->if_flags &= + ~(IFF_BROADCAST|IFF_POINTOPOINT|IFF_MULTICAST); + TUN2IFP(tp)->if_flags |= *(int *)data; + TUN_UNLOCK(tp); + + break; + default: + return (EINVAL); + } + + return (0); + case TUNSIFPID: + TUN_LOCK(tp); + tp->tun_pid = curthread->td_proc->p_pid; + TUN_UNLOCK(tp); + + return (0); + } + /* Fall through to the common ioctls if unhandled */ + } + + switch (cmd) { + case TUNSIFINFO: + tunp = (struct tuninfo *)data; + if (TUN2IFP(tp)->if_type != tunp->type) + return (EPROTOTYPE); + TUN_LOCK(tp); + if (TUN2IFP(tp)->if_mtu != tunp->mtu) { + strlcpy(ifr.ifr_name, if_name(TUN2IFP(tp)), IFNAMSIZ); + ifr.ifr_mtu = tunp->mtu; + CURVNET_SET(TUN2IFP(tp)->if_vnet); + error = ifhwioctl(SIOCSIFMTU, TUN2IFP(tp), + (caddr_t)&ifr, td); + CURVNET_RESTORE(); + if (error) { + TUN_UNLOCK(tp); + return (error); + } + } + TUN2IFP(tp)->if_baudrate = tunp->baudrate; + TUN_UNLOCK(tp); + break; + case TUNGIFINFO: + tunp = (struct tuninfo *)data; + TUN_LOCK(tp); + tunp->mtu = TUN2IFP(tp)->if_mtu; + tunp->type = TUN2IFP(tp)->if_type; + tunp->baudrate = TUN2IFP(tp)->if_baudrate; + TUN_UNLOCK(tp); + break; + case TUNSDEBUG: + tundebug = *(int *)data; + break; + case TUNGDEBUG: + *(int *)data = tundebug; + break; + case FIONBIO: + break; + case FIOASYNC: + TUN_LOCK(tp); + if (*(int *)data) + tp->tun_flags |= TUN_ASYNC; + else + tp->tun_flags &= ~TUN_ASYNC; + TUN_UNLOCK(tp); + break; + case FIONREAD: + if (!IFQ_IS_EMPTY(&TUN2IFP(tp)->if_snd)) { + struct mbuf *mb; + IFQ_LOCK(&TUN2IFP(tp)->if_snd); + IFQ_POLL_NOLOCK(&TUN2IFP(tp)->if_snd, mb); + for (*(int *)data = 0; mb != NULL; mb = mb->m_next) + *(int *)data += mb->m_len; + IFQ_UNLOCK(&TUN2IFP(tp)->if_snd); + } else + *(int *)data = 0; + 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(struct cdev *dev, struct uio *uio, int flag) +{ + struct tuntap_softc *tp = dev->si_drv1; + struct ifnet *ifp = TUN2IFP(tp); + struct mbuf *m; + int error=0, len; + + TUNDEBUG (ifp, "read\n"); + TUN_LOCK(tp); + if ((tp->tun_flags & TUN_READY) != TUN_READY) { + TUN_UNLOCK(tp); + TUNDEBUG (ifp, "not ready 0%o\n", tp->tun_flags); + return (EHOSTDOWN); + } + + tp->tun_flags &= ~TUN_RWAIT; + + do { + IFQ_DEQUEUE(&ifp->if_snd, m); + if (m == NULL) { + if (flag & O_NONBLOCK) { + TUN_UNLOCK(tp); + return (EWOULDBLOCK); + } + tp->tun_flags |= TUN_RWAIT; + error = mtx_sleep(tp, &tp->tun_mtx, PCATCH | (PZERO + 1), + "tunread", 0); + if (error != 0) { + TUN_UNLOCK(tp); + return (error); + } + } + } while (m == NULL); + TUN_UNLOCK(tp); + + if ((tp->tun_flags & TUN_L2) != 0) + BPF_MTAP(ifp, m); + + while (m && uio->uio_resid > 0 && error == 0) { + len = min(uio->uio_resid, m->m_len); + if (len != 0) + error = uiomove(mtod(m, void *), len, uio); + m = m_free(m); + } + + if (m) { + TUNDEBUG(ifp, "Dropping mbuf\n"); + m_freem(m); + } + return (error); +} + +static int +tunwrite_l2(struct tuntap_softc *tp, struct mbuf *m) +{ + struct ether_header *eh; + struct ifnet *ifp; + + ifp = TUN2IFP(tp); + + /* + * Only pass a unicast frame to ether_input(), if it would + * actually have been received by non-virtual hardware. + */ + if (m->m_len < sizeof(struct ether_header)) { + m_freem(m); + return (0); + } + + eh = mtod(m, struct ether_header *); + + if (eh && (ifp->if_flags & IFF_PROMISC) == 0 && + !ETHER_IS_MULTICAST(eh->ether_dhost) && + bcmp(eh->ether_dhost, IF_LLADDR(ifp), ETHER_ADDR_LEN) != 0) { + m_freem(m); + return (0); + } + + /* Pass packet up to parent. */ + CURVNET_SET(ifp->if_vnet); + (*ifp->if_input)(ifp, m); + CURVNET_RESTORE(); + /* ibytes are counted in parent */ + if_inc_counter(ifp, IFCOUNTER_IPACKETS, 1); + return (0); +} + +static int +tunwrite_l3(struct tuntap_softc *tp, struct mbuf *m) +{ + struct ifnet *ifp; + int family, isr; + + ifp = TUN2IFP(tp); + /* Could be unlocked read? */ + TUN_LOCK(tp); + if (tp->tun_flags & TUN_IFHEAD) { + TUN_UNLOCK(tp); + if (m->m_len < sizeof(family) && + (m = m_pullup(m, sizeof(family))) == NULL) + return (ENOBUFS); + family = ntohl(*mtod(m, u_int32_t *)); + m_adj(m, sizeof(family)); + } else { + TUN_UNLOCK(tp); + family = AF_INET; + } + + BPF_MTAP2(ifp, &family, sizeof(family), m); + + switch (family) { +#ifdef INET + case AF_INET: + isr = NETISR_IP; + break; +#endif +#ifdef INET6 + case AF_INET6: + isr = NETISR_IPV6; + break; +#endif + default: + m_freem(m); + return (EAFNOSUPPORT); + } + random_harvest_queue(m, sizeof(*m), RANDOM_NET_TUN); + if_inc_counter(ifp, IFCOUNTER_IBYTES, m->m_pkthdr.len); + if_inc_counter(ifp, IFCOUNTER_IPACKETS, 1); + CURVNET_SET(ifp->if_vnet); + M_SETFIB(m, ifp->if_fib); + netisr_dispatch(isr, m); + CURVNET_RESTORE(); + return (0); +} + +/* + * the cdevsw write interface - an atomic write is a packet - or else! + */ +static int +tunwrite(struct cdev *dev, struct uio *uio, int flag) +{ + struct tuntap_softc *tp; + struct ifnet *ifp; + struct mbuf *m; + uint32_t mru; + int align; + bool l2tun; + + tp = dev->si_drv1; + ifp = TUN2IFP(tp); + TUNDEBUG(ifp, "tunwrite\n"); + if ((ifp->if_flags & IFF_UP) != IFF_UP) + /* ignore silently */ + return (0); + + if (uio->uio_resid == 0) + return (0); + + l2tun = (tp->tun_flags & TUN_L2) != 0; + align = 0; + mru = l2tun ? TAPMRU : TUNMRU; + if (l2tun) + align = ETHER_ALIGN; + else if ((tp->tun_flags & TUN_IFHEAD) != 0) + mru += sizeof(uint32_t); /* family */ + if (uio->uio_resid < 0 || uio->uio_resid > mru) { + TUNDEBUG(ifp, "len=%zd!\n", uio->uio_resid); + return (EIO); + } + + if ((m = m_uiotombuf(uio, M_NOWAIT, 0, align, M_PKTHDR)) == NULL) { + if_inc_counter(ifp, IFCOUNTER_IERRORS, 1); + return (ENOBUFS); + } + + m->m_pkthdr.rcvif = ifp; +#ifdef MAC + mac_ifnet_create_mbuf(ifp, m); +#endif + + if (l2tun) + return (tunwrite_l2(tp, m)); + + return (tunwrite_l3(tp, m)); +} + +/* + * 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(struct cdev *dev, int events, struct thread *td) +{ + struct tuntap_softc *tp = dev->si_drv1; + struct ifnet *ifp = TUN2IFP(tp); + int revents = 0; + + TUNDEBUG(ifp, "tunpoll\n"); + + if (events & (POLLIN | POLLRDNORM)) { + IFQ_LOCK(&ifp->if_snd); + if (!IFQ_IS_EMPTY(&ifp->if_snd)) { + TUNDEBUG(ifp, "tunpoll q=%d\n", ifp->if_snd.ifq_len); + revents |= events & (POLLIN | POLLRDNORM); + } else { + TUNDEBUG(ifp, "tunpoll waiting\n"); + selrecord(td, &tp->tun_rsel); + } + IFQ_UNLOCK(&ifp->if_snd); + } + if (events & (POLLOUT | POLLWRNORM)) + revents |= events & (POLLOUT | POLLWRNORM); + + return (revents); +} + +/* + * tunkqfilter - support for the kevent() system call. + */ +static int +tunkqfilter(struct cdev *dev, struct knote *kn) +{ + struct tuntap_softc *tp = dev->si_drv1; + struct ifnet *ifp = TUN2IFP(tp); + + switch(kn->kn_filter) { + case EVFILT_READ: + TUNDEBUG(ifp, "%s kqfilter: EVFILT_READ, minor = %#x\n", + ifp->if_xname, dev2unit(dev)); + kn->kn_fop = &tun_read_filterops; + break; + + case EVFILT_WRITE: + TUNDEBUG(ifp, "%s kqfilter: EVFILT_WRITE, minor = %#x\n", + ifp->if_xname, dev2unit(dev)); + kn->kn_fop = &tun_write_filterops; + break; + + default: + TUNDEBUG(ifp, "%s kqfilter: invalid filter, minor = %#x\n", + ifp->if_xname, dev2unit(dev)); + return(EINVAL); + } + + kn->kn_hook = tp; + knlist_add(&tp->tun_rsel.si_note, kn, 0); + + return (0); +} + +/* + * Return true of there is data in the interface queue. + */ +static int +tunkqread(struct knote *kn, long hint) +{ + int ret; + struct tuntap_softc *tp = kn->kn_hook; + struct cdev *dev = tp->tun_dev; + struct ifnet *ifp = TUN2IFP(tp); + + if ((kn->kn_data = ifp->if_snd.ifq_len) > 0) { + TUNDEBUG(ifp, + "%s have data in the queue. Len = %d, minor = %#x\n", + ifp->if_xname, ifp->if_snd.ifq_len, dev2unit(dev)); + ret = 1; + } else { + TUNDEBUG(ifp, + "%s waiting for data, minor = %#x\n", ifp->if_xname, + dev2unit(dev)); + ret = 0; + } + + return (ret); +} + +/* + * Always can write, always return MTU in kn->data. + */ +static int +tunkqwrite(struct knote *kn, long hint) +{ + struct tuntap_softc *tp = kn->kn_hook; + struct ifnet *ifp = TUN2IFP(tp); + + kn->kn_data = ifp->if_mtu; + + return (1); +} + +static void +tunkqdetach(struct knote *kn) +{ + struct tuntap_softc *tp = kn->kn_hook; + + knlist_remove(&tp->tun_rsel.si_note, kn, 0); +} diff --git a/sys/powerpc/conf/GENERIC b/sys/powerpc/conf/GENERIC index d17ba4dd965e..35e5880a2605 100644 --- a/sys/powerpc/conf/GENERIC +++ b/sys/powerpc/conf/GENERIC @@ -167,7 +167,7 @@ device loop # Network loopback device random # Entropy device device ether # Ethernet support device vlan # 802.1Q VLAN support -device tun # Packet tunnel. +device tuntap # Packet tunnel. device md # Memory "disks" device ofwd # Open Firmware disks device gif # IPv6 and IPv4 tunneling diff --git a/sys/powerpc/conf/GENERIC64 b/sys/powerpc/conf/GENERIC64 index cbeacabdbea6..7bc22c8e5545 100644 --- a/sys/powerpc/conf/GENERIC64 +++ b/sys/powerpc/conf/GENERIC64 @@ -186,7 +186,7 @@ device loop # Network loopback device random # Entropy device device ether # Ethernet support device vlan # 802.1Q VLAN support -device tun # Packet tunnel. +device tuntap # Packet tunnel. device md # Memory "disks" device ofwd # Open Firmware disks device gif # IPv6 and IPv4 tunneling diff --git a/sys/powerpc/conf/MPC85XX b/sys/powerpc/conf/MPC85XX index 40ad47a5c90c..3eeba5f54ec4 100644 --- a/sys/powerpc/conf/MPC85XX +++ b/sys/powerpc/conf/MPC85XX @@ -103,7 +103,7 @@ device spibus device spigen device tsec device dpaa -device tun +device tuntap device uart options USB_DEBUG # enable debug msgs #device uhci diff --git a/sys/powerpc/conf/MPC85XXSPE b/sys/powerpc/conf/MPC85XXSPE index 1421db6742d0..1d649c1b36b9 100644 --- a/sys/powerpc/conf/MPC85XXSPE +++ b/sys/powerpc/conf/MPC85XXSPE @@ -104,7 +104,7 @@ device spibus device spigen device tsec device dpaa -device tun +device tuntap device uart options USB_DEBUG # enable debug msgs #device uhci diff --git a/sys/powerpc/conf/QORIQ64 b/sys/powerpc/conf/QORIQ64 index 4493c30d5afa..eecc0eb8e06b 100644 --- a/sys/powerpc/conf/QORIQ64 +++ b/sys/powerpc/conf/QORIQ64 @@ -107,7 +107,7 @@ device sdhci device sec device spibus device spigen -device tun +device tuntap device uart options USB_DEBUG # enable debug msgs #device uhci diff --git a/sys/riscv/conf/GENERIC b/sys/riscv/conf/GENERIC index e304ebab9a59..33c234b32074 100644 --- a/sys/riscv/conf/GENERIC +++ b/sys/riscv/conf/GENERIC @@ -135,7 +135,7 @@ device loop # Network loopback device random # Entropy device device ether # Ethernet support device vlan # 802.1Q VLAN support -device tun # Packet tunnel. +device tuntap # Packet tunnel. device md # Memory "disks" device gif # IPv6 and IPv4 tunneling device firmware # firmware assist module diff --git a/sys/sparc64/conf/GENERIC b/sys/sparc64/conf/GENERIC index cd4128c74cc0..4ccd401fb9ad 100644 --- a/sys/sparc64/conf/GENERIC +++ b/sys/sparc64/conf/GENERIC @@ -236,7 +236,7 @@ device loop # Network loopback device random # Entropy device device ether # Ethernet support device vlan # 802.1Q VLAN support -device tun # Packet tunnel. +device tuntap # Packet tunnel. device md # Memory "disks" device gif # IPv6 and IPv4 tunneling device firmware # firmware assist module