NetBSD/sys/net/if_tun.c

1123 lines
22 KiB
C
Raw Normal View History

/* $NetBSD: if_tun.c,v 1.140 2017/10/25 08:12:40 maya Exp $ */
1993-03-21 12:45:37 +03:00
/*
* Copyright (c) 1988, Julian Onions <jpo@cs.nott.ac.uk>
* Nottingham University 1987.
1993-03-21 12:45:37 +03:00
*
* 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 its
1993-03-21 12:45:37 +03:00
* roots in a similar driver written by Phil Cockcroft (formerly) at
1996-09-07 16:40:22 +04:00
* UCL. This driver is based much more on read/write/poll mode of
1993-03-21 12:45:37 +03:00
* operation though.
*/
/*
* tun - tunnel software network interface.
*/
2001-11-13 02:49:33 +03:00
#include <sys/cdefs.h>
__KERNEL_RCSID(0, "$NetBSD: if_tun.c,v 1.140 2017/10/25 08:12:40 maya Exp $");
2015-08-25 01:21:26 +03:00
#ifdef _KERNEL_OPT
1998-07-05 04:51:04 +04:00
#include "opt_inet.h"
2015-08-25 01:21:26 +03:00
#endif
1998-07-05 04:51:04 +04:00
#include <sys/param.h>
#include <sys/buf.h>
#include <sys/conf.h>
#include <sys/cpu.h>
#include <sys/device.h>
2017-01-23 18:32:04 +03:00
#include <sys/file.h>
#include <sys/ioctl.h>
#include <sys/kauth.h>
#include <sys/kmem.h>
#include <sys/lwp.h>
2017-01-23 18:32:04 +03:00
#include <sys/mbuf.h>
#include <sys/module.h>
2017-01-23 18:32:04 +03:00
#include <sys/mutex.h>
#include <sys/poll.h>
#include <sys/select.h>
#include <sys/signalvar.h>
#include <sys/socket.h>
1993-12-24 06:20:59 +03:00
2017-01-23 18:32:04 +03:00
#include <net/bpf.h>
#include <net/if.h>
#include <net/if_types.h>
#include <net/route.h>
1993-03-21 12:45:37 +03:00
#ifdef INET
#include <netinet/in.h>
#include <netinet/in_systm.h>
#include <netinet/in_var.h>
#include <netinet/ip.h>
#include <netinet/if_inarp.h>
1993-03-21 12:45:37 +03:00
#endif
#include <net/if_tun.h>
#include "ioconf.h"
1996-10-13 06:10:01 +04:00
#define TUNDEBUG if (tundebug) printf
int tundebug = 0;
1993-03-21 12:45:37 +03:00
extern int ifqmaxlen;
static LIST_HEAD(, tun_softc) tun_softc_list;
static LIST_HEAD(, tun_softc) tunz_softc_list;
static kmutex_t tun_softc_lock;
1993-03-21 12:45:37 +03:00
static int tun_ioctl(struct ifnet *, u_long, void *);
static int tun_output(struct ifnet *, struct mbuf *,
const struct sockaddr *, const struct rtentry *rt);
static int tun_clone_create(struct if_clone *, int);
static int tun_clone_destroy(struct ifnet *);
1996-02-14 00:59:53 +03:00
static struct if_clone tun_cloner =
IF_CLONE_INITIALIZER("tun", tun_clone_create, tun_clone_destroy);
static void tunattach0(struct tun_softc *);
static void tun_enable(struct tun_softc *, const struct ifaddr *);
static void tun_i_softintr(void *);
static void tun_o_softintr(void *);
#ifdef ALTQ
static void tunstart(struct ifnet *);
#endif
static struct tun_softc *tun_find_unit(dev_t);
static struct tun_softc *tun_find_zunit(int);
1993-03-21 12:45:37 +03:00
static dev_type_open(tunopen);
static dev_type_close(tunclose);
static dev_type_read(tunread);
static dev_type_write(tunwrite);
static dev_type_ioctl(tunioctl);
static dev_type_poll(tunpoll);
static dev_type_kqfilter(tunkqfilter);
const struct cdevsw tun_cdevsw = {
.d_open = tunopen,
.d_close = tunclose,
.d_read = tunread,
.d_write = tunwrite,
.d_ioctl = tunioctl,
.d_stop = nostop,
.d_tty = notty,
.d_poll = tunpoll,
.d_mmap = nommap,
.d_kqfilter = tunkqfilter,
.d_discard = nodiscard,
.d_flag = D_OTHER | D_MPSAFE
};
2016-09-05 05:25:37 +03:00
#ifdef _MODULE
devmajor_t tun_bmajor = -1, tun_cmajor = -1;
#endif
void
tunattach(int unused)
{
/*
* Nothing to do here, initialization is handled by the
2016-09-05 04:57:54 +03:00
* module initialization code in tuninit() below).
*/
}
static void
tuninit(void)
{
mutex_init(&tun_softc_lock, MUTEX_DEFAULT, IPL_NET);
LIST_INIT(&tun_softc_list);
LIST_INIT(&tunz_softc_list);
if_clone_attach(&tun_cloner);
2016-09-05 05:25:37 +03:00
#ifdef _MODULE
devsw_attach("tun", NULL, &tun_bmajor, &tun_cdevsw, &tun_cmajor);
#endif
}
static int
tundetach(void)
{
int error = 0;
if (!LIST_EMPTY(&tun_softc_list) || !LIST_EMPTY(&tunz_softc_list))
error = EBUSY;
2016-09-05 05:25:37 +03:00
#ifdef _MODULE
if (error == 0)
error = devsw_detach(NULL, &tun_cdevsw);
#endif
if (error == 0) {
if_clone_detach(&tun_cloner);
mutex_destroy(&tun_softc_lock);
}
return error;
}
/*
* Find driver instance from dev_t.
* Returns with tp locked (if found).
*/
static struct tun_softc *
tun_find_unit(dev_t dev)
{
struct tun_softc *tp;
int unit = minor(dev);
mutex_enter(&tun_softc_lock);
LIST_FOREACH(tp, &tun_softc_list, tun_list)
if (unit == tp->tun_unit)
break;
if (tp)
mutex_enter(&tp->tun_lock);
mutex_exit(&tun_softc_lock);
2017-01-23 18:32:04 +03:00
return tp;
}
/*
* Find zombie driver instance by unit number.
* Remove tp from list and return it unlocked (if found).
*/
static struct tun_softc *
tun_find_zunit(int unit)
{
struct tun_softc *tp;
mutex_enter(&tun_softc_lock);
LIST_FOREACH(tp, &tunz_softc_list, tun_list)
if (unit == tp->tun_unit)
break;
if (tp)
LIST_REMOVE(tp, tun_list);
mutex_exit(&tun_softc_lock);
2017-01-27 00:38:11 +03:00
KASSERTMSG(!tp || (tp->tun_flags & (TUN_INITED|TUN_OPEN)) == TUN_OPEN,
"tun%d: inconsistent flags: %x", unit, tp->tun_flags);
2017-01-23 18:32:04 +03:00
return tp;
}
static int
tun_clone_create(struct if_clone *ifc, int unit)
{
struct tun_softc *tp;
if ((tp = tun_find_zunit(unit)) == NULL) {
tp = kmem_zalloc(sizeof(*tp), KM_SLEEP);
tp->tun_unit = unit;
mutex_init(&tp->tun_lock, MUTEX_DEFAULT, IPL_NET);
cv_init(&tp->tun_cv, "tunread");
selinit(&tp->tun_rsel);
selinit(&tp->tun_wsel);
} else {
/* Revive tunnel instance; clear ifp part */
(void)memset(&tp->tun_if, 0, sizeof(struct ifnet));
}
if_initname(&tp->tun_if, ifc->ifc_name, unit);
tunattach0(tp);
tp->tun_flags |= TUN_INITED;
tp->tun_osih = softint_establish(SOFTINT_CLOCK, tun_o_softintr, tp);
tp->tun_isih = softint_establish(SOFTINT_CLOCK, tun_i_softintr, tp);
mutex_enter(&tun_softc_lock);
LIST_INSERT_HEAD(&tun_softc_list, tp, tun_list);
mutex_exit(&tun_softc_lock);
2017-01-23 18:32:04 +03:00
return 0;
}
static void
tunattach0(struct tun_softc *tp)
{
struct ifnet *ifp;
ifp = &tp->tun_if;
ifp->if_softc = tp;
ifp->if_mtu = TUNMTU;
ifp->if_ioctl = tun_ioctl;
ifp->if_output = tun_output;
#ifdef ALTQ
ifp->if_start = tunstart;
#endif
ifp->if_flags = IFF_POINTOPOINT;
ifp->if_type = IFT_TUNNEL;
ifp->if_snd.ifq_maxlen = ifqmaxlen;
ifp->if_collisions = 0;
ifp->if_ierrors = 0;
ifp->if_oerrors = 0;
ifp->if_ipackets = 0;
ifp->if_opackets = 0;
ifp->if_ibytes = 0;
ifp->if_obytes = 0;
ifp->if_dlt = DLT_NULL;
IFQ_SET_READY(&ifp->if_snd);
if_attach(ifp);
if_alloc_sadl(ifp);
bpf_attach(ifp, DLT_NULL, sizeof(uint32_t));
}
static int
tun_clone_destroy(struct ifnet *ifp)
{
struct tun_softc *tp = (void *)ifp;
bool zombie = false;
IF_PURGE(&ifp->if_snd);
ifp->if_flags &= ~IFF_RUNNING;
mutex_enter(&tun_softc_lock);
mutex_enter(&tp->tun_lock);
LIST_REMOVE(tp, tun_list);
if (tp->tun_flags & TUN_OPEN) {
/* Hang on to storage until last close. */
tp->tun_flags &= ~TUN_INITED;
LIST_INSERT_HEAD(&tunz_softc_list, tp, tun_list);
zombie = true;
}
mutex_exit(&tun_softc_lock);
if (tp->tun_flags & TUN_RWAIT) {
tp->tun_flags &= ~TUN_RWAIT;
cv_broadcast(&tp->tun_cv);
}
selnotify(&tp->tun_rsel, 0, 0);
mutex_exit(&tp->tun_lock);
if (tp->tun_flags & TUN_ASYNC && tp->tun_pgid)
fownsignal(tp->tun_pgid, SIGIO, POLL_HUP, 0, NULL);
bpf_detach(ifp);
if_detach(ifp);
if (!zombie) {
seldestroy(&tp->tun_rsel);
seldestroy(&tp->tun_wsel);
softint_disestablish(tp->tun_osih);
softint_disestablish(tp->tun_isih);
mutex_destroy(&tp->tun_lock);
cv_destroy(&tp->tun_cv);
kmem_free(tp, sizeof(*tp));
}
2017-01-23 18:32:04 +03:00
return 0;
}
/*
* tunnel open - must be superuser & the device must be
* configured in
*/
static int
tunopen(dev_t dev, int flag, int mode, struct lwp *l)
1993-03-21 12:45:37 +03:00
{
struct ifnet *ifp;
struct tun_softc *tp;
int error;
error = kauth_authorize_network(l->l_cred, KAUTH_NETWORK_INTERFACE_TUN,
KAUTH_REQ_NETWORK_INTERFACE_TUN_ADD, NULL, NULL, NULL);
if (error)
2017-01-23 18:32:04 +03:00
return error;
1993-03-21 12:45:37 +03:00
tp = tun_find_unit(dev);
if (tp == NULL) {
(void)tun_clone_create(&tun_cloner, minor(dev));
tp = tun_find_unit(dev);
if (tp == NULL) {
return ENXIO;
}
}
if (tp->tun_flags & TUN_OPEN) {
mutex_exit(&tp->tun_lock);
return EBUSY;
}
ifp = &tp->tun_if;
tp->tun_flags |= TUN_OPEN;
TUNDEBUG("%s: open\n", ifp->if_xname);
mutex_exit(&tp->tun_lock);
2017-01-23 18:32:04 +03:00
return error;
1993-03-21 12:45:37 +03:00
}
/*
* tunclose - close the device - mark i/f down & delete
* routing info
*/
int
tunclose(dev_t dev, int flag, int mode,
struct lwp *l)
1993-03-21 12:45:37 +03:00
{
struct tun_softc *tp;
struct ifnet *ifp;
if ((tp = tun_find_zunit(minor(dev))) != NULL) {
/* interface was "destroyed" before the close */
seldestroy(&tp->tun_rsel);
seldestroy(&tp->tun_wsel);
softint_disestablish(tp->tun_osih);
softint_disestablish(tp->tun_isih);
mutex_destroy(&tp->tun_lock);
kmem_free(tp, sizeof(*tp));
return 0;
}
if ((tp = tun_find_unit(dev)) == NULL)
goto out_nolock;
ifp = &tp->tun_if;
tp->tun_flags &= ~TUN_OPEN;
tp->tun_pgid = 0;
selnotify(&tp->tun_rsel, 0, 0);
TUNDEBUG ("%s: closed\n", ifp->if_xname);
mutex_exit(&tp->tun_lock);
/*
* junk all pending output
*/
IFQ_PURGE(&ifp->if_snd);
if (ifp->if_flags & IFF_UP) {
if_down(ifp);
if (ifp->if_flags & IFF_RUNNING) {
/* find internet addresses and delete routes */
2000-03-30 13:45:33 +04:00
struct ifaddr *ifa;
IFADDR_READER_FOREACH(ifa, ifp) {
#if defined(INET) || defined(INET6)
if (ifa->ifa_addr->sa_family == AF_INET ||
ifa->ifa_addr->sa_family == AF_INET6) {
rtinit(ifa, (int)RTM_DELETE,
tp->tun_flags & TUN_DSTADDR
? RTF_HOST
: 0);
}
#endif
}
}
}
out_nolock:
2017-01-23 18:32:04 +03:00
return 0;
1993-03-21 12:45:37 +03:00
}
static void
tun_enable(struct tun_softc *tp, const struct ifaddr *ifa)
1993-03-21 12:45:37 +03:00
{
struct ifnet *ifp = &tp->tun_if;
TUNDEBUG("%s: %s\n", __func__, ifp->if_xname);
mutex_enter(&tp->tun_lock);
ifp->if_flags |= IFF_UP | IFF_RUNNING;
tp->tun_flags &= ~(TUN_IASET|TUN_DSTADDR);
switch (ifa->ifa_addr->sa_family) {
#ifdef INET
case AF_INET: {
struct sockaddr_in *sin;
sin = satosin(ifa->ifa_addr);
if (sin && sin->sin_addr.s_addr)
tp->tun_flags |= TUN_IASET;
if (ifp->if_flags & IFF_POINTOPOINT) {
sin = satosin(ifa->ifa_dstaddr);
if (sin && sin->sin_addr.s_addr)
tp->tun_flags |= TUN_DSTADDR;
}
break;
}
#endif
#ifdef INET6
case AF_INET6: {
struct sockaddr_in6 *sin;
sin = satosin6(ifa->ifa_addr);
if (!IN6_IS_ADDR_UNSPECIFIED(&sin->sin6_addr))
tp->tun_flags |= TUN_IASET;
if (ifp->if_flags & IFF_POINTOPOINT) {
sin = satosin6(ifa->ifa_dstaddr);
if (sin && !IN6_IS_ADDR_UNSPECIFIED(&sin->sin6_addr))
tp->tun_flags |= TUN_DSTADDR;
} else
tp->tun_flags &= ~TUN_DSTADDR;
break;
}
#endif /* INET6 */
default:
break;
}
mutex_exit(&tp->tun_lock);
1993-03-21 12:45:37 +03:00
}
/*
* Process an ioctl request.
*/
static int
tun_ioctl(struct ifnet *ifp, u_long cmd, void *data)
1993-03-21 12:45:37 +03:00
{
struct tun_softc *tp = (struct tun_softc *)(ifp->if_softc);
struct ifreq *ifr = (struct ifreq *)data;
struct ifaddr *ifa = (struct ifaddr *)data;
int error = 0;
2003-05-02 07:15:23 +04:00
switch (cmd) {
*** Summary *** When a link-layer address changes (e.g., ifconfig ex0 link 02:de:ad:be:ef:02 active), send a gratuitous ARP and/or a Neighbor Advertisement to update the network-/link-layer address bindings on our LAN peers. Refuse a change of ethernet address to the address 00:00:00:00:00:00 or to any multicast/broadcast address. (Thanks matt@.) Reorder ifnet ioctl operations so that driver ioctls may inherit the functions of their "class"---ether_ioctl(), fddi_ioctl(), et cetera---and the class ioctls may inherit from the generic ioctl, ifioctl_common(), but both driver- and class-ioctls may override the generic behavior. Make network drivers share more code. Distinguish a "factory" link-layer address from others for the purposes of both protecting that address from deletion and computing EUI64. Return consistent, appropriate error codes from network drivers. Improve readability. KNF. *** Details *** In if_attach(), always initialize the interface ioctl routine, ifnet->if_ioctl, if the driver has not already initialized it. Delete if_ioctl == NULL tests everywhere else, because it cannot happen. In the ioctl routines of network interfaces, inherit common ioctl behaviors by calling either ifioctl_common() or whichever ioctl routine is appropriate for the class of interface---e.g., ether_ioctl() for ethernets. Stop (ab)using SIOCSIFADDR and start to use SIOCINITIFADDR. In the user->kernel interface, SIOCSIFADDR's argument was an ifreq, but on the protocol->ifnet interface, SIOCSIFADDR's argument was an ifaddr. That was confusing, and it would work against me as I make it possible for a network interface to overload most ioctls. On the protocol->ifnet interface, replace SIOCSIFADDR with SIOCINITIFADDR. In ifioctl(), return EPERM if userland tries to invoke SIOCINITIFADDR. In ifioctl(), give the interface the first shot at handling most interface ioctls, and give the protocol the second shot, instead of the other way around. Finally, let compatibility code (COMPAT_OSOCK) take a shot. Pull device initialization out of switch statements under SIOCINITIFADDR. For example, pull ..._init() out of any switch statement that looks like this: switch (...->sa_family) { case ...: ..._init(); ... break; ... default: ..._init(); ... break; } Rewrite many if-else clauses that handle all permutations of IFF_UP and IFF_RUNNING to use a switch statement, switch (x & (IFF_UP|IFF_RUNNING)) { case 0: ... break; case IFF_RUNNING: ... break; case IFF_UP: ... break; case IFF_UP|IFF_RUNNING: ... break; } unifdef lots of code containing #ifdef FreeBSD, #ifdef NetBSD, and #ifdef SIOCSIFMTU, especially in fwip(4) and in ndis(4). In ipw(4), remove an if_set_sadl() call that is out of place. In nfe(4), reuse the jumbo MTU logic in ether_ioctl(). Let ethernets register a callback for setting h/w state such as promiscuous mode and the multicast filter in accord with a change in the if_flags: ether_set_ifflags_cb() registers a callback that returns ENETRESET if the caller should reset the ethernet by calling if_init(), 0 on success, != 0 on failure. Pull common code from ex(4), gem(4), nfe(4), sip(4), tlp(4), vge(4) into ether_ioctl(), and register if_flags callbacks for those drivers. Return ENOTTY instead of EINVAL for inappropriate ioctls. In zyd(4), use ENXIO instead of ENOTTY to indicate that the device is not any longer attached. Add to if_set_sadl() a boolean 'factory' argument that indicates whether a link-layer address was assigned by the factory or some other source. In a comment, recommend using the factory address for generating an EUI64, and update in6_get_hw_ifid() to prefer a factory address to any other link-layer address. Add a routing message, RTM_LLINFO_UPD, that tells protocols to update the binding of network-layer addresses to link-layer addresses. Implement this message in IPv4 and IPv6 by sending a gratuitous ARP or a neighbor advertisement, respectively. Generate RTM_LLINFO_UPD messages on a change of an interface's link-layer address. In ether_ioctl(), do not let SIOCALIFADDR set a link-layer address that is broadcast/multicast or equal to 00:00:00:00:00:00. Make ether_ioctl() call ifioctl_common() to handle ioctls that it does not understand. In gif(4), initialize if_softc and use it, instead of assuming that the gif_softc and ifp overlap. Let ifioctl_common() handle SIOCGIFADDR. Sprinkle rtcache_invariants(), which checks on DIAGNOSTIC kernels that certain invariants on a struct route are satisfied. In agr(4), rewrite agr_ioctl_filter() to be a bit more explicit about the ioctls that we do not allow on an agr(4) member interface. bzero -> memset. Delete unnecessary casts to void *. Use sockaddr_in_init() and sockaddr_in6_init(). Compare pointers with NULL instead of "testing truth". Replace some instances of (type *)0 with NULL. Change some K&R prototypes to ANSI C, and join lines.
2008-11-07 03:20:01 +03:00
case SIOCINITIFADDR:
tun_enable(tp, ifa);
ifa->ifa_rtrequest = p2p_rtrequest;
TUNDEBUG("%s: address set\n", ifp->if_xname);
break;
case SIOCSIFBRDADDR:
TUNDEBUG("%s: broadcast address set\n", ifp->if_xname);
break;
case SIOCSIFMTU:
if (ifr->ifr_mtu > TUNMTU || ifr->ifr_mtu < 576) {
error = EINVAL;
break;
}
TUNDEBUG("%s: interface mtu set\n", ifp->if_xname);
if ((error = ifioctl_common(ifp, cmd, data)) == ENETRESET)
error = 0;
break;
case SIOCADDMULTI:
case SIOCDELMULTI:
if (ifr == NULL) {
error = EAFNOSUPPORT; /* XXX */
break;
}
switch (ifreq_getaddr(cmd, ifr)->sa_family) {
#ifdef INET
case AF_INET:
break;
#endif
#ifdef INET6
case AF_INET6:
break;
#endif
default:
error = EAFNOSUPPORT;
break;
}
break;
default:
*** Summary *** When a link-layer address changes (e.g., ifconfig ex0 link 02:de:ad:be:ef:02 active), send a gratuitous ARP and/or a Neighbor Advertisement to update the network-/link-layer address bindings on our LAN peers. Refuse a change of ethernet address to the address 00:00:00:00:00:00 or to any multicast/broadcast address. (Thanks matt@.) Reorder ifnet ioctl operations so that driver ioctls may inherit the functions of their "class"---ether_ioctl(), fddi_ioctl(), et cetera---and the class ioctls may inherit from the generic ioctl, ifioctl_common(), but both driver- and class-ioctls may override the generic behavior. Make network drivers share more code. Distinguish a "factory" link-layer address from others for the purposes of both protecting that address from deletion and computing EUI64. Return consistent, appropriate error codes from network drivers. Improve readability. KNF. *** Details *** In if_attach(), always initialize the interface ioctl routine, ifnet->if_ioctl, if the driver has not already initialized it. Delete if_ioctl == NULL tests everywhere else, because it cannot happen. In the ioctl routines of network interfaces, inherit common ioctl behaviors by calling either ifioctl_common() or whichever ioctl routine is appropriate for the class of interface---e.g., ether_ioctl() for ethernets. Stop (ab)using SIOCSIFADDR and start to use SIOCINITIFADDR. In the user->kernel interface, SIOCSIFADDR's argument was an ifreq, but on the protocol->ifnet interface, SIOCSIFADDR's argument was an ifaddr. That was confusing, and it would work against me as I make it possible for a network interface to overload most ioctls. On the protocol->ifnet interface, replace SIOCSIFADDR with SIOCINITIFADDR. In ifioctl(), return EPERM if userland tries to invoke SIOCINITIFADDR. In ifioctl(), give the interface the first shot at handling most interface ioctls, and give the protocol the second shot, instead of the other way around. Finally, let compatibility code (COMPAT_OSOCK) take a shot. Pull device initialization out of switch statements under SIOCINITIFADDR. For example, pull ..._init() out of any switch statement that looks like this: switch (...->sa_family) { case ...: ..._init(); ... break; ... default: ..._init(); ... break; } Rewrite many if-else clauses that handle all permutations of IFF_UP and IFF_RUNNING to use a switch statement, switch (x & (IFF_UP|IFF_RUNNING)) { case 0: ... break; case IFF_RUNNING: ... break; case IFF_UP: ... break; case IFF_UP|IFF_RUNNING: ... break; } unifdef lots of code containing #ifdef FreeBSD, #ifdef NetBSD, and #ifdef SIOCSIFMTU, especially in fwip(4) and in ndis(4). In ipw(4), remove an if_set_sadl() call that is out of place. In nfe(4), reuse the jumbo MTU logic in ether_ioctl(). Let ethernets register a callback for setting h/w state such as promiscuous mode and the multicast filter in accord with a change in the if_flags: ether_set_ifflags_cb() registers a callback that returns ENETRESET if the caller should reset the ethernet by calling if_init(), 0 on success, != 0 on failure. Pull common code from ex(4), gem(4), nfe(4), sip(4), tlp(4), vge(4) into ether_ioctl(), and register if_flags callbacks for those drivers. Return ENOTTY instead of EINVAL for inappropriate ioctls. In zyd(4), use ENXIO instead of ENOTTY to indicate that the device is not any longer attached. Add to if_set_sadl() a boolean 'factory' argument that indicates whether a link-layer address was assigned by the factory or some other source. In a comment, recommend using the factory address for generating an EUI64, and update in6_get_hw_ifid() to prefer a factory address to any other link-layer address. Add a routing message, RTM_LLINFO_UPD, that tells protocols to update the binding of network-layer addresses to link-layer addresses. Implement this message in IPv4 and IPv6 by sending a gratuitous ARP or a neighbor advertisement, respectively. Generate RTM_LLINFO_UPD messages on a change of an interface's link-layer address. In ether_ioctl(), do not let SIOCALIFADDR set a link-layer address that is broadcast/multicast or equal to 00:00:00:00:00:00. Make ether_ioctl() call ifioctl_common() to handle ioctls that it does not understand. In gif(4), initialize if_softc and use it, instead of assuming that the gif_softc and ifp overlap. Let ifioctl_common() handle SIOCGIFADDR. Sprinkle rtcache_invariants(), which checks on DIAGNOSTIC kernels that certain invariants on a struct route are satisfied. In agr(4), rewrite agr_ioctl_filter() to be a bit more explicit about the ioctls that we do not allow on an agr(4) member interface. bzero -> memset. Delete unnecessary casts to void *. Use sockaddr_in_init() and sockaddr_in6_init(). Compare pointers with NULL instead of "testing truth". Replace some instances of (type *)0 with NULL. Change some K&R prototypes to ANSI C, and join lines.
2008-11-07 03:20:01 +03:00
error = ifioctl_common(ifp, cmd, data);
}
2017-01-23 18:32:04 +03:00
return error;
1993-03-21 12:45:37 +03:00
}
/*
1996-02-14 00:59:53 +03:00
* tun_output - queue packets from higher level ready to put out.
1993-03-21 12:45:37 +03:00
*/
static int
tun_output(struct ifnet *ifp, struct mbuf *m0, const struct sockaddr *dst,
const struct rtentry *rt)
1993-03-21 12:45:37 +03:00
{
struct tun_softc *tp = ifp->if_softc;
int error;
#if defined(INET) || defined(INET6)
int mlen;
uint32_t *af;
#endif
mutex_enter(&tp->tun_lock);
TUNDEBUG ("%s: tun_output\n", ifp->if_xname);
if ((tp->tun_flags & TUN_READY) != TUN_READY) {
TUNDEBUG ("%s: not ready 0%o\n", ifp->if_xname,
tp->tun_flags);
error = EHOSTDOWN;
mutex_exit(&tp->tun_lock);
goto out;
}
// XXXrmind
mutex_exit(&tp->tun_lock);
/*
* if the queueing discipline needs packet classification,
* do it before prepending link headers.
*/
IFQ_CLASSIFY(&ifp->if_snd, m0, dst->sa_family);
bpf_mtap_af(ifp, dst->sa_family, m0);
switch(dst->sa_family) {
#ifdef INET6
case AF_INET6:
#endif
1993-03-21 12:45:37 +03:00
#ifdef INET
case AF_INET:
#endif
#if defined(INET) || defined(INET6)
if (tp->tun_flags & TUN_PREPADDR) {
/* Simple link-layer header */
M_PREPEND(m0, dst->sa_len, M_DONTWAIT);
if (m0 == NULL) {
IF_DROP(&ifp->if_snd);
error = ENOBUFS;
goto out;
}
memcpy(mtod(m0, char *), dst, dst->sa_len);
}
if (tp->tun_flags & TUN_IFHEAD) {
2006-03-03 22:46:35 +03:00
/* Prepend the address family */
2006-03-03 22:57:37 +03:00
M_PREPEND(m0, sizeof(*af), M_DONTWAIT);
if (m0 == NULL) {
IF_DROP(&ifp->if_snd);
error = ENOBUFS;
goto out;
}
af = mtod(m0,uint32_t *);
*af = htonl(dst->sa_family);
} else {
#ifdef INET
if (dst->sa_family != AF_INET)
#endif
{
error = EAFNOSUPPORT;
goto out;
}
}
/* FALLTHROUGH */
case AF_UNSPEC:
IFQ_ENQUEUE(&ifp->if_snd, m0, error);
if (error) {
ifp->if_collisions++;
error = EAFNOSUPPORT;
m0 = NULL;
goto out;
}
mlen = m0->m_pkthdr.len;
ifp->if_opackets++;
ifp->if_obytes += mlen;
break;
1993-03-21 12:45:37 +03:00
#endif
default:
error = EAFNOSUPPORT;
goto out;
}
mutex_enter(&tp->tun_lock);
if (tp->tun_flags & TUN_RWAIT) {
tp->tun_flags &= ~TUN_RWAIT;
cv_broadcast(&tp->tun_cv);
}
if (tp->tun_flags & TUN_ASYNC && tp->tun_pgid)
softint_schedule(tp->tun_isih);
selnotify(&tp->tun_rsel, 0, 0);
mutex_exit(&tp->tun_lock);
out:
if (error && m0) {
m_freem(m0);
}
return 0;
1993-03-21 12:45:37 +03:00
}
static void
tun_i_softintr(void *cookie)
{
struct tun_softc *tp = cookie;
if (tp->tun_flags & TUN_ASYNC && tp->tun_pgid)
fownsignal(tp->tun_pgid, SIGIO, POLL_IN, POLLIN|POLLRDNORM,
NULL);
}
static void
tun_o_softintr(void *cookie)
{
struct tun_softc *tp = cookie;
if (tp->tun_flags & TUN_ASYNC && tp->tun_pgid)
fownsignal(tp->tun_pgid, SIGIO, POLL_OUT, POLLOUT|POLLWRNORM,
NULL);
}
1993-03-21 12:45:37 +03:00
/*
* the cdevsw interface is now pretty minimal.
*/
int
tunioctl(dev_t dev, u_long cmd, void *data, int flag, struct lwp *l)
1993-03-21 12:45:37 +03:00
{
struct tun_softc *tp;
int error = 0;
tp = tun_find_unit(dev);
/* interface was "destroyed" already */
if (tp == NULL) {
return ENXIO;
}
switch (cmd) {
case TUNSDEBUG:
tundebug = *(int *)data;
break;
case TUNGDEBUG:
*(int *)data = tundebug;
break;
case TUNSIFMODE:
switch (*(int *)data & (IFF_POINTOPOINT|IFF_BROADCAST)) {
case IFF_POINTOPOINT:
case IFF_BROADCAST:
if (tp->tun_if.if_flags & IFF_UP) {
error = EBUSY;
goto out;
}
tp->tun_if.if_flags &=
~(IFF_BROADCAST|IFF_POINTOPOINT|IFF_MULTICAST);
tp->tun_if.if_flags |= *(int *)data;
break;
default:
error = EINVAL;
goto out;
}
break;
case TUNSLMODE:
if (*(int *)data) {
tp->tun_flags |= TUN_PREPADDR;
tp->tun_flags &= ~TUN_IFHEAD;
} else
tp->tun_flags &= ~TUN_PREPADDR;
break;
case TUNSIFHEAD:
if (*(int *)data) {
tp->tun_flags |= TUN_IFHEAD;
tp->tun_flags &= ~TUN_PREPADDR;
} else
tp->tun_flags &= ~TUN_IFHEAD;
break;
case TUNGIFHEAD:
*(int *)data = (tp->tun_flags & TUN_IFHEAD);
break;
case FIONBIO:
if (*(int *)data)
tp->tun_flags |= TUN_NBIO;
else
tp->tun_flags &= ~TUN_NBIO;
break;
case FIOASYNC:
if (*(int *)data)
tp->tun_flags |= TUN_ASYNC;
else
tp->tun_flags &= ~TUN_ASYNC;
break;
case FIONREAD:
if (tp->tun_if.if_snd.ifq_head)
*(int *)data = tp->tun_if.if_snd.ifq_head->m_pkthdr.len;
else
*(int *)data = 0;
break;
case TIOCSPGRP:
case FIOSETOWN:
error = fsetown(&tp->tun_pgid, cmd, data);
break;
case TIOCGPGRP:
case FIOGETOWN:
error = fgetown(tp->tun_pgid, cmd, data);
break;
default:
error = ENOTTY;
}
out:
mutex_exit(&tp->tun_lock);
2017-01-23 18:32:04 +03:00
return error;
1993-03-21 12:45:37 +03:00
}
/*
* The cdevsw read interface - reads a packet at a time, or at
* least as much of a packet as can be read.
1993-03-21 12:45:37 +03:00
*/
int
tunread(dev_t dev, struct uio *uio, int ioflag)
1993-03-21 12:45:37 +03:00
{
struct tun_softc *tp;
struct ifnet *ifp;
struct mbuf *m, *m0;
int error = 0, len;
tp = tun_find_unit(dev);
/* interface was "destroyed" already */
if (tp == NULL) {
return ENXIO;
}
ifp = &tp->tun_if;
TUNDEBUG ("%s: read\n", ifp->if_xname);
if ((tp->tun_flags & TUN_READY) != TUN_READY) {
TUNDEBUG ("%s: not ready 0%o\n", ifp->if_xname, tp->tun_flags);
error = EHOSTDOWN;
goto out;
}
tp->tun_flags &= ~TUN_RWAIT;
do {
IFQ_DEQUEUE(&ifp->if_snd, m0);
if (m0 == 0) {
if (tp->tun_flags & TUN_NBIO) {
error = EWOULDBLOCK;
goto out;
}
tp->tun_flags |= TUN_RWAIT;
if (cv_wait_sig(&tp->tun_cv, &tp->tun_lock)) {
error = EINTR;
goto out;
}
}
} while (m0 == 0);
mutex_exit(&tp->tun_lock);
/* Copy the mbuf chain */
while (m0 && uio->uio_resid > 0 && error == 0) {
1994-05-26 04:47:19 +04:00
len = min(uio->uio_resid, m0->m_len);
if (len != 0)
error = uiomove(mtod(m0, void *), len, uio);
2016-10-02 17:16:02 +03:00
m0 = m = m_free(m0);
}
if (m0) {
TUNDEBUG("Dropping mbuf\n");
m_freem(m0);
}
if (error)
ifp->if_ierrors++;
2017-01-23 18:32:04 +03:00
return error;
out:
mutex_exit(&tp->tun_lock);
2017-01-23 18:32:04 +03:00
return error;
1993-03-21 12:45:37 +03:00
}
/*
* the cdevsw write interface - an atomic write is a packet - or else!
*/
int
tunwrite(dev_t dev, struct uio *uio, int ioflag)
1993-03-21 12:45:37 +03:00
{
struct tun_softc *tp;
struct ifnet *ifp;
struct mbuf *top, **mp, *m;
pktqueue_t *pktq;
struct sockaddr dst;
int error = 0, tlen, mlen;
uint32_t family;
tp = tun_find_unit(dev);
if (tp == NULL) {
/* Interface was "destroyed" already. */
return ENXIO;
}
/* Unlock until we've got the data */
mutex_exit(&tp->tun_lock);
ifp = &tp->tun_if;
TUNDEBUG("%s: tunwrite\n", ifp->if_xname);
if (tp->tun_flags & TUN_PREPADDR) {
if (uio->uio_resid < sizeof(dst)) {
error = EIO;
goto out0;
}
error = uiomove((void *)&dst, sizeof(dst), uio);
if (dst.sa_len > sizeof(dst)) {
/* Duh.. */
int n = dst.sa_len - sizeof(dst);
while (n--) {
char discard;
error = uiomove(&discard, 1, uio);
if (error) {
goto out0;
}
}
}
} else if (tp->tun_flags & TUN_IFHEAD) {
if (uio->uio_resid < sizeof(family)){
error = EIO;
goto out0;
}
error = uiomove((void *)&family, sizeof(family), uio);
dst.sa_family = ntohl(family);
} else {
#ifdef INET
dst.sa_family = AF_INET;
#endif
}
if (uio->uio_resid > TUNMTU) {
TUNDEBUG("%s: len=%lu!\n", ifp->if_xname,
(unsigned long)uio->uio_resid);
error = EIO;
goto out0;
}
switch (dst.sa_family) {
#ifdef INET
case AF_INET:
pktq = ip_pktq;
break;
#endif
#ifdef INET6
case AF_INET6:
pktq = ip6_pktq;
break;
#endif
default:
error = EAFNOSUPPORT;
goto out0;
}
tlen = uio->uio_resid;
/* get a header mbuf */
MGETHDR(m, M_DONTWAIT, MT_DATA);
if (m == NULL) {
return ENOBUFS;
}
mlen = MHLEN;
2004-03-01 16:54:02 +03:00
top = NULL;
mp = &top;
while (error == 0 && uio->uio_resid > 0) {
1994-05-26 04:47:19 +04:00
m->m_len = min(mlen, uio->uio_resid);
error = uiomove(mtod(m, void *), m->m_len, uio);
*mp = m;
mp = &m->m_next;
2004-03-01 16:54:02 +03:00
if (error == 0 && uio->uio_resid > 0) {
MGET(m, M_DONTWAIT, MT_DATA);
if (m == NULL) {
error = ENOBUFS;
break;
}
mlen = MLEN;
}
}
if (error) {
2004-03-01 16:54:02 +03:00
if (top != NULL)
m_freem(top);
ifp->if_ierrors++;
goto out0;
}
top->m_pkthdr.len = tlen;
m_set_rcvif(top, ifp);
bpf_mtap_af(ifp, dst.sa_family, top);
mutex_enter(&tp->tun_lock);
if ((tp->tun_flags & TUN_INITED) == 0) {
/* Interface was destroyed */
error = ENXIO;
goto out;
}
if (__predict_false(!pktq_enqueue(pktq, top, 0))) {
ifp->if_collisions++;
mutex_exit(&tp->tun_lock);
error = ENOBUFS;
m_freem(top);
goto out0;
}
ifp->if_ipackets++;
ifp->if_ibytes += tlen;
out:
mutex_exit(&tp->tun_lock);
out0:
2017-01-23 18:32:04 +03:00
return error;
1993-03-21 12:45:37 +03:00
}
#ifdef ALTQ
/*
* Start packet transmission on the interface.
* when the interface queue is rate-limited by ALTQ or TBR,
* if_start is needed to drain packets from the queue in order
* to notify readers when outgoing packets become ready.
*/
static void
tunstart(struct ifnet *ifp)
{
struct tun_softc *tp = ifp->if_softc;
if (!ALTQ_IS_ENABLED(&ifp->if_snd) && !TBR_IS_ENABLED(&ifp->if_snd))
return;
mutex_enter(&tp->tun_lock);
if (!IF_IS_EMPTY(&ifp->if_snd)) {
if (tp->tun_flags & TUN_RWAIT) {
tp->tun_flags &= ~TUN_RWAIT;
cv_broadcast(&tp->tun_cv);
}
if (tp->tun_flags & TUN_ASYNC && tp->tun_pgid)
softint_schedule(tp->tun_osih);
selnotify(&tp->tun_rsel, 0, 0);
}
mutex_exit(&tp->tun_lock);
}
#endif /* ALTQ */
1993-03-21 12:45:37 +03:00
/*
1996-09-07 16:40:22 +04:00
* 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.
1993-03-21 12:45:37 +03:00
*/
int
tunpoll(dev_t dev, int events, struct lwp *l)
1993-03-21 12:45:37 +03:00
{
struct tun_softc *tp;
struct ifnet *ifp;
int revents = 0;
tp = tun_find_unit(dev);
if (tp == NULL) {
/* Interface was "destroyed" already. */
return 0;
}
ifp = &tp->tun_if;
1996-09-07 16:40:22 +04:00
TUNDEBUG("%s: tunpoll\n", ifp->if_xname);
if (events & (POLLIN | POLLRDNORM)) {
if (!IFQ_IS_EMPTY(&ifp->if_snd)) {
1996-09-07 16:40:22 +04:00
TUNDEBUG("%s: tunpoll q=%d\n", ifp->if_xname,
ifp->if_snd.ifq_len);
1996-09-07 16:40:22 +04:00
revents |= events & (POLLIN | POLLRDNORM);
} else {
TUNDEBUG("%s: tunpoll waiting\n", ifp->if_xname);
2005-12-11 15:16:03 +03:00
selrecord(l, &tp->tun_rsel);
}
}
1996-09-07 16:40:22 +04:00
if (events & (POLLOUT | POLLWRNORM))
revents |= events & (POLLOUT | POLLWRNORM);
mutex_exit(&tp->tun_lock);
2017-01-23 18:32:04 +03:00
return revents;
1993-03-21 12:45:37 +03:00
}
static void
filt_tunrdetach(struct knote *kn)
{
struct tun_softc *tp = kn->kn_hook;
mutex_enter(&tp->tun_lock);
2002-11-26 21:51:18 +03:00
SLIST_REMOVE(&tp->tun_rsel.sel_klist, kn, knote, kn_selnext);
mutex_exit(&tp->tun_lock);
}
static int
filt_tunread(struct knote *kn, long hint)
{
struct tun_softc *tp = kn->kn_hook;
struct ifnet *ifp = &tp->tun_if;
struct mbuf *m;
mutex_enter(&tp->tun_lock);
IF_POLL(&ifp->if_snd, m);
if (m == NULL) {
mutex_exit(&tp->tun_lock);
2017-01-23 18:32:04 +03:00
return 0;
}
for (kn->kn_data = 0; m != NULL; m = m->m_next)
kn->kn_data += m->m_len;
mutex_exit(&tp->tun_lock);
2017-01-23 18:32:04 +03:00
return 1;
}
static const struct filterops tunread_filtops = {
.f_isfd = 1,
.f_attach = NULL,
.f_detach = filt_tunrdetach,
.f_event = filt_tunread,
};
static const struct filterops tun_seltrue_filtops = {
.f_isfd = 1,
.f_attach = NULL,
.f_detach = filt_tunrdetach,
.f_event = filt_seltrue,
};
int
tunkqfilter(dev_t dev, struct knote *kn)
{
struct tun_softc *tp;
struct klist *klist;
int rv = 0;
tp = tun_find_unit(dev);
if (tp == NULL)
goto out_nolock;
switch (kn->kn_filter) {
case EVFILT_READ:
2002-11-26 21:51:18 +03:00
klist = &tp->tun_rsel.sel_klist;
kn->kn_fop = &tunread_filtops;
break;
case EVFILT_WRITE:
2002-11-26 21:51:18 +03:00
klist = &tp->tun_rsel.sel_klist;
kn->kn_fop = &tun_seltrue_filtops;
break;
default:
rv = EINVAL;
goto out;
}
kn->kn_hook = tp;
SLIST_INSERT_HEAD(klist, kn, kn_selnext);
out:
mutex_exit(&tp->tun_lock);
out_nolock:
2017-01-23 18:32:04 +03:00
return rv;
}
/*
* Module infrastructure
*/
#include "if_module.h"
IF_MODULE(MODULE_CLASS_DRIVER, tun, "")