/*
- * Copyright (c) 2000-2011 Apple Inc. All rights reserved.
+ * Copyright (c) 2000-2019 Apple Inc. All rights reserved.
*
* @APPLE_OSREFERENCE_LICENSE_HEADER_START@
- *
+ *
* This file contains Original Code and/or Modifications of Original Code
* as defined in and that are subject to the Apple Public Source License
* Version 2.0 (the 'License'). You may not use this file except in
* unlawful or unlicensed copies of an Apple operating system, or to
* circumvent, violate, or enable the circumvention or violation of, any
* terms of an Apple operating system software license agreement.
- *
+ *
* Please obtain a copy of the License at
* http://www.opensource.apple.com/apsl/ and read it before using this file.
- *
+ *
* The Original Code and all software distributed under the License are
* distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
* EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
* FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
* Please see the License for the specific language governing rights and
* limitations under the License.
- *
+ *
* @APPLE_OSREFERENCE_LICENSE_HEADER_END@
*/
/*
* SUCH DAMAGE.
*
* @(#)in.c 8.4 (Berkeley) 1/9/95
- * $FreeBSD: src/sys/netinet/in.c,v 1.44.2.5 2001/08/13 16:26:17 ume Exp $
*/
#include <sys/param.h>
#include <sys/kern_event.h>
#include <sys/syslog.h>
#include <sys/mcache.h>
-#include <kern/zalloc.h>
+#include <sys/protosw.h>
+#include <sys/file.h>
+#include <kern/zalloc.h>
#include <pexpert/pexpert.h>
+#include <os/log.h>
#include <net/if.h>
#include <net/if_types.h>
#include <net/route.h>
#include <net/kpi_protocol.h>
+#include <net/dlil.h>
+#include <net/if_llatbl.h>
+#include <net/if_arp.h>
+#if PF
+#include <net/pfvar.h>
+#endif /* PF */
#include <netinet/in.h>
#include <netinet/in_var.h>
#include <netinet/in_pcb.h>
-
#include <netinet/igmp_var.h>
-#include <net/dlil.h>
-
#include <netinet/ip_var.h>
-
#include <netinet/tcp.h>
#include <netinet/tcp_timer.h>
#include <netinet/tcp_var.h>
-
-#include <sys/file.h>
-
-#if PF
-#include <net/pfvar.h>
-#endif /* PF */
-
-static int in_mask2len(struct in_addr *);
-static void in_len2mask(struct in_addr *, int);
-static int in_lifaddr_ioctl(struct socket *, u_long, struct if_laddrreq *,
- struct ifnet *, struct proc *);
-static int in_setrouter(struct ifnet *, int);
-
-static void in_socktrim(struct sockaddr_in *);
-static int in_ifinit(struct ifnet *,
- struct in_ifaddr *, struct sockaddr_in *, int);
-
-#define IA_HASH_INIT(ia) { \
- (ia)->ia_hash.tqe_next = (void *)(uintptr_t)-1; \
- (ia)->ia_hash.tqe_prev = (void *)(uintptr_t)-1; \
+#include <netinet/if_ether.h>
+
+static int inctl_associd(struct socket *, u_long, caddr_t);
+static int inctl_connid(struct socket *, u_long, caddr_t);
+static int inctl_conninfo(struct socket *, u_long, caddr_t);
+static int inctl_autoaddr(struct ifnet *, struct ifreq *);
+static int inctl_arpipll(struct ifnet *, struct ifreq *);
+static int inctl_setrouter(struct ifnet *, struct ifreq *);
+static int inctl_ifaddr(struct ifnet *, struct in_ifaddr *, u_long,
+ struct ifreq *);
+static int inctl_ifdstaddr(struct ifnet *, struct in_ifaddr *, u_long,
+ struct ifreq *);
+static int inctl_ifbrdaddr(struct ifnet *, struct in_ifaddr *, u_long,
+ struct ifreq *);
+static int inctl_ifnetmask(struct ifnet *, struct in_ifaddr *, u_long,
+ struct ifreq *);
+
+static void in_socktrim(struct sockaddr_in *);
+static int in_ifinit(struct ifnet *, struct in_ifaddr *,
+ struct sockaddr_in *, int);
+
+#define IA_HASH_INIT(ia) { \
+ (ia)->ia_hash.tqe_next = (void *)(uintptr_t)-1; \
+ (ia)->ia_hash.tqe_prev = (void *)(uintptr_t)-1; \
}
-#define IA_IS_HASHED(ia) \
- (!((ia)->ia_hash.tqe_next == (void *)(uintptr_t)-1 || \
+#define IA_IS_HASHED(ia) \
+ (!((ia)->ia_hash.tqe_next == (void *)(uintptr_t)-1 || \
(ia)->ia_hash.tqe_prev == (void *)(uintptr_t)-1))
static void in_iahash_remove(struct in_ifaddr *);
static void in_ifaddr_free(struct ifaddr *);
static void in_ifaddr_trace(struct ifaddr *, int);
+static int in_getassocids(struct socket *, uint32_t *, user_addr_t);
+static int in_getconnids(struct socket *, sae_associd_t, uint32_t *, user_addr_t);
+
+/* IPv4 Layer 2 neighbor cache management routines */
+static void in_lltable_destroy_lle_unlocked(struct llentry *lle);
+static void in_lltable_destroy_lle(struct llentry *lle);
+static struct llentry *in_lltable_new(struct in_addr addr4, u_int flags);
+static int in_lltable_match_prefix(const struct sockaddr *saddr,
+ const struct sockaddr *smask, u_int flags, struct llentry *lle);
+static void in_lltable_free_entry(struct lltable *llt, struct llentry *lle);
+static int in_lltable_rtcheck(struct ifnet *ifp, u_int flags, const struct sockaddr *l3addr);
+static inline uint32_t in_lltable_hash_dst(const struct in_addr dst, uint32_t hsize);
+static uint32_t in_lltable_hash(const struct llentry *lle, uint32_t hsize);
+static void in_lltable_fill_sa_entry(const struct llentry *lle, struct sockaddr *sa);
+static inline struct llentry * in_lltable_find_dst(struct lltable *llt, struct in_addr dst);
+static void in_lltable_delete_entry(struct lltable *llt, struct llentry *lle);
+static struct llentry * in_lltable_alloc(struct lltable *llt, u_int flags, const struct sockaddr *l3addr);
+static struct llentry * in_lltable_lookup(struct lltable *llt, u_int flags, const struct sockaddr *l3addr);
+static int in_lltable_dump_entry(struct lltable *llt, struct llentry *lle, struct sysctl_req *wr);
+static struct lltable * in_lltattach(struct ifnet *ifp);
+
static int subnetsarelocal = 0;
-SYSCTL_INT(_net_inet_ip, OID_AUTO, subnets_are_local, CTLFLAG_RW | CTLFLAG_LOCKED,
- &subnetsarelocal, 0, "");
+SYSCTL_INT(_net_inet_ip, OID_AUTO, subnets_are_local,
+ CTLFLAG_RW | CTLFLAG_LOCKED, &subnetsarelocal, 0, "");
/* Track whether or not the SIOCARPIPLL ioctl has been called */
-__private_extern__ u_int32_t ipv4_ll_arp_aware = 0;
+u_int32_t ipv4_ll_arp_aware = 0;
-#define INIFA_TRACE_HIST_SIZE 32 /* size of trace history */
+#define INIFA_TRACE_HIST_SIZE 32 /* size of trace history */
/* For gdb */
__private_extern__ unsigned int inifa_trace_hist_size = INIFA_TRACE_HIST_SIZE;
struct in_ifaddr_dbg {
- struct in_ifaddr inifa; /* in_ifaddr */
- struct in_ifaddr inifa_old; /* saved in_ifaddr */
- u_int16_t inifa_refhold_cnt; /* # of IFA_ADDREF */
- u_int16_t inifa_refrele_cnt; /* # of IFA_REMREF */
+ struct in_ifaddr inifa; /* in_ifaddr */
+ struct in_ifaddr inifa_old; /* saved in_ifaddr */
+ u_int16_t inifa_refhold_cnt; /* # of IFA_ADDREF */
+ u_int16_t inifa_refrele_cnt; /* # of IFA_REMREF */
/*
* Alloc and free callers.
*/
- ctrace_t inifa_alloc;
- ctrace_t inifa_free;
+ ctrace_t inifa_alloc;
+ ctrace_t inifa_free;
/*
* Circular lists of IFA_ADDREF and IFA_REMREF callers.
*/
- ctrace_t inifa_refhold[INIFA_TRACE_HIST_SIZE];
- ctrace_t inifa_refrele[INIFA_TRACE_HIST_SIZE];
+ ctrace_t inifa_refhold[INIFA_TRACE_HIST_SIZE];
+ ctrace_t inifa_refrele[INIFA_TRACE_HIST_SIZE];
/*
* Trash list linkage
*/
static decl_lck_mtx_data(, inifa_trash_lock);
#if DEBUG
-static unsigned int inifa_debug = 1; /* debugging (enabled) */
+static unsigned int inifa_debug = 1; /* debugging (enabled) */
#else
-static unsigned int inifa_debug; /* debugging (disabled) */
+static unsigned int inifa_debug; /* debugging (disabled) */
#endif /* !DEBUG */
-static unsigned int inifa_size; /* size of zone element */
-static struct zone *inifa_zone; /* zone for in_ifaddr */
+static unsigned int inifa_size; /* size of zone element */
+static struct zone *inifa_zone; /* zone for in_ifaddr */
+
+#define INIFA_ZONE_MAX 64 /* maximum elements in zone */
+#define INIFA_ZONE_NAME "in_ifaddr" /* zone name */
-#define INIFA_ZONE_MAX 64 /* maximum elements in zone */
-#define INIFA_ZONE_NAME "in_ifaddr" /* zone name */
+static const unsigned int in_extra_size = sizeof(struct in_ifextra);
+static const unsigned int in_extra_bufsize = in_extra_size +
+ sizeof(void *) + sizeof(uint64_t);
/*
* Return 1 if the address is
struct sockaddr_in sin;
int local = 0;
- if (ntohl(in.s_addr) == INADDR_LOOPBACK || IN_LINKLOCAL(ntohl(in.s_addr))) {
+ if (ntohl(in.s_addr) == INADDR_LOOPBACK ||
+ IN_LINKLOCAL(ntohl(in.s_addr))) {
local = 1;
} else if (ntohl(in.s_addr) >= INADDR_UNSPEC_GROUP &&
- ntohl(in.s_addr) <= INADDR_MAX_LOCAL_GROUP) {
- local = 1;
+ ntohl(in.s_addr) <= INADDR_MAX_LOCAL_GROUP) {
+ local = 1;
} else {
sin.sin_family = AF_INET;
- sin.sin_len = sizeof (sin);
+ sin.sin_len = sizeof(sin);
sin.sin_addr = in;
rt = rtalloc1((struct sockaddr *)&sin, 0, 0);
if (rt != NULL) {
RT_LOCK_SPIN(rt);
if (rt->rt_gateway->sa_family == AF_LINK ||
- (rt->rt_ifp->if_flags & IFF_LOOPBACK))
+ (rt->rt_ifp->if_flags & IFF_LOOPBACK)) {
local = 1;
+ }
RT_UNLOCK(rt);
rtfree(rt);
} else {
local = in_localaddr(in);
}
}
- return (local);
+ return local;
}
/*
* Return 1 if an internet address is for a ``local'' host
* (one to which we have a connection). If subnetsarelocal
- * is true, this includes other subnets of the local net.
- * Otherwise, it includes only the directly-connected (sub)nets.
+ * is true, this includes other subnets of the local net,
+ * otherwise, it includes the directly-connected (sub)nets.
+ * The IPv4 link local prefix 169.254/16 is also included.
*/
int
in_localaddr(struct in_addr in)
u_int32_t i = ntohl(in.s_addr);
struct in_ifaddr *ia;
+ if (IN_LINKLOCAL(i)) {
+ return 1;
+ }
+
if (subnetsarelocal) {
lck_rw_lock_shared(in_ifaddr_rwlock);
- for (ia = in_ifaddrhead.tqh_first; ia;
- ia = ia->ia_link.tqe_next) {
+ for (ia = in_ifaddrhead.tqh_first; ia != NULL;
+ ia = ia->ia_link.tqe_next) {
IFA_LOCK(&ia->ia_ifa);
if ((i & ia->ia_netmask) == ia->ia_net) {
IFA_UNLOCK(&ia->ia_ifa);
lck_rw_done(in_ifaddr_rwlock);
- return (1);
+ return 1;
}
IFA_UNLOCK(&ia->ia_ifa);
}
lck_rw_done(in_ifaddr_rwlock);
} else {
lck_rw_lock_shared(in_ifaddr_rwlock);
- for (ia = in_ifaddrhead.tqh_first; ia;
- ia = ia->ia_link.tqe_next) {
+ for (ia = in_ifaddrhead.tqh_first; ia != NULL;
+ ia = ia->ia_link.tqe_next) {
IFA_LOCK(&ia->ia_ifa);
if ((i & ia->ia_subnetmask) == ia->ia_subnet) {
IFA_UNLOCK(&ia->ia_ifa);
lck_rw_done(in_ifaddr_rwlock);
- return (1);
+ return 1;
}
IFA_UNLOCK(&ia->ia_ifa);
}
lck_rw_done(in_ifaddr_rwlock);
}
- return (0);
+ return 0;
}
/*
* that may not be forwarded, or whether datagrams to that destination
* may be forwarded.
*/
-int
+boolean_t
in_canforward(struct in_addr in)
{
u_int32_t i = ntohl(in.s_addr);
u_int32_t net;
- if (IN_EXPERIMENTAL(i) || IN_MULTICAST(i))
- return (0);
+ if (IN_EXPERIMENTAL(i) || IN_MULTICAST(i)) {
+ return FALSE;
+ }
if (IN_CLASSA(i)) {
net = i & IN_CLASSA_NET;
- if (net == 0 || net == (IN_LOOPBACKNET << IN_CLASSA_NSHIFT))
- return (0);
+ if (net == 0 || net == (IN_LOOPBACKNET << IN_CLASSA_NSHIFT)) {
+ return FALSE;
+ }
}
- return (1);
+ return TRUE;
}
/*
static void
in_socktrim(struct sockaddr_in *ap)
{
- char *cplim = (char *) &ap->sin_addr;
- char *cp = (char *) (&ap->sin_addr + 1);
-
- ap->sin_len = 0;
- while (--cp >= cplim)
- if (*cp) {
- (ap)->sin_len = cp - (char *) (ap) + 1;
- break;
- }
-}
-
-static int
-in_mask2len(struct in_addr *mask)
-{
- size_t x, y;
- u_char *p;
+ char *cplim = (char *)&ap->sin_addr;
+ char *cp = (char *)(&ap->sin_addr + 1);
- p = (u_char *)mask;
- for (x = 0; x < sizeof(*mask); x++) {
- if (p[x] != 0xff)
+ ap->sin_len = 0;
+ while (--cp >= cplim) {
+ if (*cp) {
+ (ap)->sin_len = cp - (char *)(ap) + 1;
break;
- }
- y = 0;
- if (x < sizeof(*mask)) {
- for (y = 0; y < 8; y++) {
- if ((p[x] & (0x80 >> y)) == 0)
- break;
}
}
- return x * 8 + y;
-}
-
-static void
-in_len2mask(struct in_addr *mask, int len)
-{
- int i;
- u_char *p;
-
- p = (u_char *)mask;
- bzero(mask, sizeof(*mask));
- for (i = 0; i < len / 8; i++)
- p[i] = 0xff;
- if (len % 8)
- p[i] = (0xff00 >> (len % 8)) & 0xff;
}
-static int in_interfaces; /* number of external internet interfaces */
+static int in_interfaces; /* number of external internet interfaces */
static int
in_domifattach(struct ifnet *ifp)
{
int error;
- if ((error = proto_plumb(PF_INET, ifp)) && error != EEXIST)
- log(LOG_ERR, "%s: proto_plumb returned %d if=%s%d\n",
- __func__, error, ifp->if_name, ifp->if_unit);
+ VERIFY(ifp != NULL);
+
+ if ((error = proto_plumb(PF_INET, ifp)) && error != EEXIST) {
+ log(LOG_ERR, "%s: proto_plumb returned %d if=%s\n",
+ __func__, error, if_name(ifp));
+ } else if (error == 0 && ifp->if_inetdata == NULL) {
+ void **pbuf, *base;
+ struct in_ifextra *ext;
+ int errorx;
+
+ if ((ext = (struct in_ifextra *)_MALLOC(in_extra_bufsize,
+ M_IFADDR, M_WAITOK | M_ZERO)) == NULL) {
+ error = ENOMEM;
+ errorx = proto_unplumb(PF_INET, ifp);
+ if (errorx != 0) {
+ log(LOG_ERR,
+ "%s: proto_unplumb returned %d if=%s%d\n",
+ __func__, errorx, ifp->if_name,
+ ifp->if_unit);
+ }
+ goto done;
+ }
- return (error);
+ /* Align on 64-bit boundary */
+ base = (void *)P2ROUNDUP((intptr_t)ext + sizeof(uint64_t),
+ sizeof(uint64_t));
+ VERIFY(((intptr_t)base + in_extra_size) <=
+ ((intptr_t)ext + in_extra_bufsize));
+ pbuf = (void **)((intptr_t)base - sizeof(void *));
+ *pbuf = ext;
+ ifp->if_inetdata = base;
+ IN_IFEXTRA(ifp)->ii_llt = in_lltattach(ifp);
+ VERIFY(IS_P2ALIGNED(ifp->if_inetdata, sizeof(uint64_t)));
+ }
+done:
+ if (error == 0 && ifp->if_inetdata != NULL) {
+ /*
+ * Since the structure is never freed, we need to
+ * zero out its contents to avoid reusing stale data.
+ * A little redundant with allocation above, but it
+ * keeps the code simpler for all cases.
+ */
+ bzero(ifp->if_inetdata, in_extra_size);
+ }
+ return error;
}
-/*
- * Generic internet control operations (ioctl's).
- * Ifp is 0 if not an interface-specific ioctl.
- *
- * Returns: 0 Success
- * EINVAL
- * EADDRNOTAVAIL
- * EDESTADDRREQ
- * EPERM
- * ENOBUFS
- * EBUSY
- * EOPNOTSUPP
- * proc_suser:EPERM
- * suser:EPERM
- * in_lifaddr_ioctl:???
- * dlil_ioctl:???
- * in_ifinit:???
- * dlil_plumb_protocol:???
- * dlil_unplumb_protocol:???
- */
-/* ARGSUSED */
-int
-in_control(struct socket *so, u_long cmd, caddr_t data, struct ifnet *ifp,
- struct proc *p)
+static __attribute__((noinline)) int
+inctl_associd(struct socket *so, u_long cmd, caddr_t data)
{
- struct in_ifaddr *ia = NULL;
- struct ifaddr *ifa;
- struct sockaddr_in oldaddr;
int error = 0;
- int hostIsNew, maskIsNew;
- struct kev_msg ev_msg;
- struct kev_in_data in_event_data;
+ union {
+ struct so_aidreq32 a32;
+ struct so_aidreq64 a64;
+ } u;
- bzero(&in_event_data, sizeof (struct kev_in_data));
- bzero(&ev_msg, sizeof (struct kev_msg));
+ VERIFY(so != NULL);
switch (cmd) {
- case SIOCALIFADDR: /* struct if_laddrreq */
- case SIOCDLIFADDR: /* struct if_laddrreq */
- if ((error = proc_suser(p)) != 0)
- return (error);
- /* FALLTHRU */
- case SIOCGLIFADDR: { /* struct if_laddrreq */
- struct if_laddrreq iflr;
+ case SIOCGASSOCIDS32: /* struct so_aidreq32 */
+ bcopy(data, &u.a32, sizeof(u.a32));
+ error = in_getassocids(so, &u.a32.sar_cnt, u.a32.sar_aidp);
+ if (error == 0) {
+ bcopy(&u.a32, data, sizeof(u.a32));
+ }
+ break;
- if (ifp == NULL)
- return (EINVAL);
+ case SIOCGASSOCIDS64: /* struct so_aidreq64 */
+ bcopy(data, &u.a64, sizeof(u.a64));
+ error = in_getassocids(so, &u.a64.sar_cnt, u.a64.sar_aidp);
+ if (error == 0) {
+ bcopy(&u.a64, data, sizeof(u.a64));
+ }
+ break;
- bcopy(data, &iflr, sizeof (iflr));
- error = in_lifaddr_ioctl(so, cmd, &iflr, ifp, p);
- bcopy(&iflr, data, sizeof (iflr));
- return (error);
- }
+ default:
+ VERIFY(0);
+ /* NOTREACHED */
}
- /*
- * Find address for this interface, if it exists.
- *
- * If an alias address was specified, find that one instead of
- * the first one on the interface.
- */
- if (ifp != NULL) {
- struct in_ifaddr *iap;
- struct sockaddr_in sin;
-
- bcopy(&((struct ifreq *)(void *)data)->ifr_addr,
- &sin, sizeof (sin));
+ return error;
+}
- lck_rw_lock_shared(in_ifaddr_rwlock);
- for (iap = in_ifaddrhead.tqh_first; iap != NULL;
- iap = iap->ia_link.tqe_next) {
- if (iap->ia_ifp != ifp)
- continue;
+static __attribute__((noinline)) int
+inctl_connid(struct socket *so, u_long cmd, caddr_t data)
+{
+ int error = 0;
+ union {
+ struct so_cidreq32 c32;
+ struct so_cidreq64 c64;
+ } u;
- IFA_LOCK(&iap->ia_ifa);
- if (sin.sin_addr.s_addr ==
- iap->ia_addr.sin_addr.s_addr) {
- ia = iap;
- IFA_UNLOCK(&iap->ia_ifa);
- break;
- } else if (ia == NULL) {
- ia = iap;
- if (sin.sin_family != AF_INET) {
- IFA_UNLOCK(&iap->ia_ifa);
- break;
- }
- }
- IFA_UNLOCK(&iap->ia_ifa);
- }
- /* take a reference on ia before releasing lock */
- if (ia != NULL)
- IFA_ADDREF(&ia->ia_ifa);
- lck_rw_done(in_ifaddr_rwlock);
- }
+ VERIFY(so != NULL);
switch (cmd) {
- case SIOCAUTOADDR: /* struct ifreq */
- case SIOCARPIPLL: /* struct ifreq */
- case SIOCSETROUTERMODE: /* struct ifreq */
- if ((error = proc_suser(p)) != 0) {
- goto done;
- }
- if (ifp == NULL) {
- error = EADDRNOTAVAIL;
- goto done;
+ case SIOCGCONNIDS32: /* struct so_cidreq32 */
+ bcopy(data, &u.c32, sizeof(u.c32));
+ error = in_getconnids(so, u.c32.scr_aid, &u.c32.scr_cnt,
+ u.c32.scr_cidp);
+ if (error == 0) {
+ bcopy(&u.c32, data, sizeof(u.c32));
}
break;
- case SIOCAIFADDR: /* struct ifaliasreq */
- case SIOCDIFADDR: { /* struct ifreq */
- struct sockaddr_in addr, dstaddr;
-
- if (ifp == NULL) {
- error = EADDRNOTAVAIL;
- goto done;
+ case SIOCGCONNIDS64: /* struct so_cidreq64 */
+ bcopy(data, &u.c64, sizeof(u.c64));
+ error = in_getconnids(so, u.c64.scr_aid, &u.c64.scr_cnt,
+ u.c64.scr_cidp);
+ if (error == 0) {
+ bcopy(&u.c64, data, sizeof(u.c64));
}
+ break;
- if (cmd == SIOCAIFADDR) {
- bcopy(&((struct in_aliasreq *)(void *)data)->
- ifra_addr, &addr, sizeof (addr));
- bcopy(&((struct in_aliasreq *)(void *)data)->
- ifra_dstaddr, &dstaddr, sizeof (dstaddr));
- } else {
- VERIFY(cmd == SIOCDIFADDR);
- bcopy(&((struct ifreq *)(void *)data)->ifr_addr,
- &addr, sizeof (addr));
- bzero(&dstaddr, sizeof (dstaddr));
- }
+ default:
+ VERIFY(0);
+ /* NOTREACHED */
+ }
- if (addr.sin_family == AF_INET) {
- struct in_ifaddr *oia;
+ return error;
+}
- lck_rw_lock_shared(in_ifaddr_rwlock);
- for (oia = ia; ia; ia = ia->ia_link.tqe_next) {
- IFA_LOCK(&ia->ia_ifa);
- if (ia->ia_ifp == ifp &&
- ia->ia_addr.sin_addr.s_addr ==
- addr.sin_addr.s_addr) {
- IFA_ADDREF_LOCKED(&ia->ia_ifa);
- IFA_UNLOCK(&ia->ia_ifa);
- break;
- }
- IFA_UNLOCK(&ia->ia_ifa);
- }
- lck_rw_done(in_ifaddr_rwlock);
- if (oia != NULL)
- IFA_REMREF(&oia->ia_ifa);
- if ((ifp->if_flags & IFF_POINTOPOINT) &&
- (cmd == SIOCAIFADDR) &&
- (dstaddr.sin_addr.s_addr == INADDR_ANY)) {
- error = EDESTADDRREQ;
- goto done;
- }
- } else if (cmd == SIOCAIFADDR) {
- error = EINVAL;
- goto done;
- }
- if (cmd == SIOCDIFADDR && ia == NULL) {
- error = EADDRNOTAVAIL;
- goto done;
- }
- /* FALLTHROUGH */
- }
- case SIOCSIFADDR: /* struct ifreq */
- case SIOCSIFNETMASK: /* struct ifreq */
- case SIOCSIFDSTADDR: { /* struct ifreq */
- struct sockaddr_in addr;
+static __attribute__((noinline)) int
+inctl_conninfo(struct socket *so, u_long cmd, caddr_t data)
+{
+ int error = 0;
+ union {
+ struct so_cinforeq32 ci32;
+ struct so_cinforeq64 ci64;
+ } u;
- if (cmd == SIOCAIFADDR) {
- /* fell thru from above; just repeat it */
- bcopy(&((struct in_aliasreq *)(void *)data)->
- ifra_addr, &addr, sizeof (addr));
- } else {
- VERIFY(cmd == SIOCDIFADDR || cmd == SIOCSIFADDR ||
- cmd == SIOCSIFNETMASK || cmd == SIOCSIFDSTADDR);
- bcopy(&((struct ifreq *)(void *)data)->ifr_addr,
- &addr, sizeof (addr));
- }
+ VERIFY(so != NULL);
- /* socket is NULL if called from in_purgeaddrs() */
- if (so != NULL && (so->so_state & SS_PRIV) == 0) {
- error = EPERM;
- goto done;
- }
- /* in case it's NULL, make sure it came from the kernel */
- if (so == NULL && p != kernproc) {
- error = EPERM;
- goto done;
- }
- if (ifp == NULL) {
- error = EADDRNOTAVAIL;
- goto done;
- }
- if (addr.sin_family != AF_INET && cmd == SIOCSIFADDR) {
- error = EINVAL;
- goto done;
- }
- if (ia == NULL) {
- ia = in_ifaddr_alloc(M_WAITOK);
- if (ia == NULL) {
- error = ENOBUFS;
- goto done;
- }
- ifnet_lock_exclusive(ifp);
- ifa = &ia->ia_ifa;
- IFA_LOCK(ifa);
- /* Hold a reference for this routine */
- IFA_ADDREF_LOCKED(ifa);
- IA_HASH_INIT(ia);
- ifa->ifa_addr = (struct sockaddr *)&ia->ia_addr;
- ifa->ifa_dstaddr = (struct sockaddr *)&ia->ia_dstaddr;
- ifa->ifa_netmask = (struct sockaddr *)&ia->ia_sockmask;
- ia->ia_sockmask.sin_len = 8;
- if (ifp->if_flags & IFF_BROADCAST) {
- ia->ia_broadaddr.sin_len = sizeof (ia->ia_addr);
- ia->ia_broadaddr.sin_family = AF_INET;
- }
- ia->ia_ifp = ifp;
- if (!(ifp->if_flags & IFF_LOOPBACK))
- in_interfaces++;
- /* if_attach_ifa() holds a reference for ifa_link */
- if_attach_ifa(ifp, ifa);
- /*
- * If we have to go through in_ifinit(), make sure
- * to avoid installing route(s) based on this address
- * via PFC_IFUP event, before the link resolver (ARP)
- * initializes it.
- */
- if (cmd == SIOCAIFADDR || cmd == SIOCSIFADDR)
- ifa->ifa_debug |= IFD_NOTREADY;
- IFA_UNLOCK(ifa);
- ifnet_lock_done(ifp);
- lck_rw_lock_exclusive(in_ifaddr_rwlock);
- /* Hold a reference for ia_link */
- IFA_ADDREF(ifa);
- TAILQ_INSERT_TAIL(&in_ifaddrhead, ia, ia_link);
- lck_rw_done(in_ifaddr_rwlock);
- /* discard error */
- (void) in_domifattach(ifp);
- error = 0;
+ switch (cmd) {
+ case SIOCGCONNINFO32: /* struct so_cinforeq32 */
+ bcopy(data, &u.ci32, sizeof(u.ci32));
+ error = in_getconninfo(so, u.ci32.scir_cid, &u.ci32.scir_flags,
+ &u.ci32.scir_ifindex, &u.ci32.scir_error, u.ci32.scir_src,
+ &u.ci32.scir_src_len, u.ci32.scir_dst, &u.ci32.scir_dst_len,
+ &u.ci32.scir_aux_type, u.ci32.scir_aux_data,
+ &u.ci32.scir_aux_len);
+ if (error == 0) {
+ bcopy(&u.ci32, data, sizeof(u.ci32));
}
break;
- }
- case SIOCPROTOATTACH: /* struct ifreq */
- case SIOCPROTODETACH: /* struct ifreq */
- if ((error = proc_suser(p)) != 0) {
- goto done;
- }
- if (ifp == NULL) {
- error = EADDRNOTAVAIL;
- goto done;
+ case SIOCGCONNINFO64: /* struct so_cinforeq64 */
+ bcopy(data, &u.ci64, sizeof(u.ci64));
+ error = in_getconninfo(so, u.ci64.scir_cid, &u.ci64.scir_flags,
+ &u.ci64.scir_ifindex, &u.ci64.scir_error, u.ci64.scir_src,
+ &u.ci64.scir_src_len, u.ci64.scir_dst, &u.ci64.scir_dst_len,
+ &u.ci64.scir_aux_type, u.ci64.scir_aux_data,
+ &u.ci64.scir_aux_len);
+ if (error == 0) {
+ bcopy(&u.ci64, data, sizeof(u.ci64));
}
break;
- case SIOCSIFBRDADDR: /* struct ifreq */
- if ((so->so_state & SS_PRIV) == 0) {
- error = EPERM;
- goto done;
- }
- /* FALLTHROUGH */
- case SIOCGIFADDR: /* struct ifreq */
- case SIOCGIFNETMASK: /* struct ifreq */
- case SIOCGIFDSTADDR: /* struct ifreq */
- case SIOCGIFBRDADDR: /* struct ifreq */
- if (ia == NULL) {
- error = EADDRNOTAVAIL;
- goto done;
- }
- break;
+ default:
+ VERIFY(0);
+ /* NOTREACHED */
}
- switch (cmd) {
- case SIOCAUTOADDR: { /* struct ifreq */
- int intval;
-
- VERIFY(ifp != NULL);
- bcopy(&((struct ifreq *)(void *)data)->ifr_intval,
- &intval, sizeof (intval));
+ return error;
+}
- ifnet_lock_exclusive(ifp);
- if (intval) {
- /*
- * An interface in IPv4 router mode implies that it
- * is configured with a static IP address and should
- * not act as a DHCP client; prevent SIOCAUTOADDR from
- * being set in that mode.
- */
- if (ifp->if_eflags & IFEF_IPV4_ROUTER) {
- intval = 0; /* be safe; clear flag if set */
- error = EBUSY;
- } else {
- ifp->if_eflags |= IFEF_AUTOCONFIGURING;
- }
- }
- if (!intval)
- ifp->if_eflags &= ~IFEF_AUTOCONFIGURING;
- ifnet_lock_done(ifp);
- break;
- }
+/*
+ * Caller passes in the ioctl data pointer directly via "ifr", with the
+ * expectation that this routine always uses bcopy() or other byte-aligned
+ * memory accesses.
+ */
+static __attribute__((noinline)) int
+inctl_autoaddr(struct ifnet *ifp, struct ifreq *ifr)
+{
+ int error = 0, intval;
- case SIOCARPIPLL: { /* struct ifreq */
- int intval;
+ VERIFY(ifp != NULL);
- VERIFY(ifp != NULL);
- bcopy(&((struct ifreq *)(void *)data)->ifr_intval,
- &intval, sizeof (intval));
- ipv4_ll_arp_aware = 1;
+ bcopy(&ifr->ifr_intval, &intval, sizeof(intval));
- ifnet_lock_exclusive(ifp);
- if (intval) {
- /*
- * An interface in IPv4 router mode implies that it
- * is configured with a static IP address and should
- * not have to deal with IPv4 Link-Local Address;
- * prevent SIOCARPIPLL from being set in that mode.
- */
- if (ifp->if_eflags & IFEF_IPV4_ROUTER) {
- intval = 0; /* be safe; clear flag if set */
- error = EBUSY;
- } else {
- ifp->if_eflags |= IFEF_ARPLL;
- }
+ ifnet_lock_exclusive(ifp);
+ if (intval) {
+ /*
+ * An interface in IPv4 router mode implies that it
+ * is configured with a static IP address and should
+ * not act as a DHCP client; prevent SIOCAUTOADDR from
+ * being set in that mode.
+ */
+ if (ifp->if_eflags & IFEF_IPV4_ROUTER) {
+ intval = 0; /* be safe; clear flag if set */
+ error = EBUSY;
+ } else {
+ ifp->if_eflags |= IFEF_AUTOCONFIGURING;
}
- if (!intval)
- ifp->if_eflags &= ~IFEF_ARPLL;
- ifnet_lock_done(ifp);
- break;
}
+ if (!intval) {
+ ifp->if_eflags &= ~IFEF_AUTOCONFIGURING;
+ }
+ ifnet_lock_done(ifp);
- case SIOCGIFADDR: /* struct ifreq */
- VERIFY(ia != NULL);
- IFA_LOCK(&ia->ia_ifa);
- bcopy(&ia->ia_addr, &((struct ifreq *)(void *)data)->ifr_addr,
- sizeof (struct sockaddr_in));
- IFA_UNLOCK(&ia->ia_ifa);
- break;
+ return error;
+}
- case SIOCGIFBRDADDR: /* struct ifreq */
- VERIFY(ia != NULL);
- if ((ifp->if_flags & IFF_BROADCAST) == 0) {
- error = EINVAL;
- break;
- }
- IFA_LOCK(&ia->ia_ifa);
- bcopy(&ia->ia_broadaddr,
- &((struct ifreq *)(void *)data)->ifr_broadaddr,
- sizeof (struct sockaddr_in));
- IFA_UNLOCK(&ia->ia_ifa);
- break;
+/*
+ * Caller passes in the ioctl data pointer directly via "ifr", with the
+ * expectation that this routine always uses bcopy() or other byte-aligned
+ * memory accesses.
+ */
+static __attribute__((noinline)) int
+inctl_arpipll(struct ifnet *ifp, struct ifreq *ifr)
+{
+ int error = 0, intval;
- case SIOCGIFDSTADDR: /* struct ifreq */
- VERIFY(ia != NULL);
- if ((ifp->if_flags & IFF_POINTOPOINT) == 0) {
- error = EINVAL;
- break;
- }
- IFA_LOCK(&ia->ia_ifa);
- bcopy(&ia->ia_dstaddr,
- &((struct ifreq *)(void *)data)->ifr_dstaddr,
- sizeof (struct sockaddr_in));
- IFA_UNLOCK(&ia->ia_ifa);
- break;
+ VERIFY(ifp != NULL);
- case SIOCGIFNETMASK: /* struct ifreq */
- VERIFY(ia != NULL);
- IFA_LOCK(&ia->ia_ifa);
- bcopy(&ia->ia_sockmask,
- &((struct ifreq *)(void *)data)->ifr_addr,
- sizeof (struct sockaddr_in));
- IFA_UNLOCK(&ia->ia_ifa);
- break;
+ bcopy(&ifr->ifr_intval, &intval, sizeof(intval));
+ ipv4_ll_arp_aware = 1;
- case SIOCSIFDSTADDR: /* struct ifreq */
- VERIFY(ifp != NULL && ia != NULL);
- if ((ifp->if_flags & IFF_POINTOPOINT) == 0) {
- error = EINVAL;
- break;
- }
- IFA_LOCK(&ia->ia_ifa);
- oldaddr = ia->ia_dstaddr;
- bcopy(&((struct ifreq *)(void *)data)->ifr_dstaddr,
- &ia->ia_dstaddr, sizeof (struct sockaddr_in));
- if (ia->ia_dstaddr.sin_family == AF_INET)
- ia->ia_dstaddr.sin_len = sizeof (struct sockaddr_in);
- IFA_UNLOCK(&ia->ia_ifa);
+ ifnet_lock_exclusive(ifp);
+ if (intval) {
/*
- * NOTE: SIOCSIFDSTADDR is defined with struct ifreq
- * as parameter, but here we are sending it down
- * to the interface with a pointer to struct ifaddr,
- * for legacy reasons.
+ * An interface in IPv4 router mode implies that it
+ * is configured with a static IP address and should
+ * not have to deal with IPv4 Link-Local Address;
+ * prevent SIOCARPIPLL from being set in that mode.
*/
- error = ifnet_ioctl(ifp, PF_INET, SIOCSIFDSTADDR, ia);
- IFA_LOCK(&ia->ia_ifa);
- if (error == EOPNOTSUPP) {
- error = 0;
- }
- if (error) {
- ia->ia_dstaddr = oldaddr;
- IFA_UNLOCK(&ia->ia_ifa);
- break;
+ if (ifp->if_eflags & IFEF_IPV4_ROUTER) {
+ intval = 0; /* be safe; clear flag if set */
+ error = EBUSY;
+ } else {
+ ifp->if_eflags |= IFEF_ARPLL;
}
- IFA_LOCK_ASSERT_HELD(&ia->ia_ifa);
+ }
+ if (!intval) {
+ ifp->if_eflags &= ~IFEF_ARPLL;
+ }
+ ifnet_lock_done(ifp);
- ev_msg.vendor_code = KEV_VENDOR_APPLE;
- ev_msg.kev_class = KEV_NETWORK_CLASS;
- ev_msg.kev_subclass = KEV_INET_SUBCLASS;
+ return error;
+}
+
+/*
+ * Handle SIOCSETROUTERMODE to set or clear the IPv4 router mode flag on
+ * the interface. When in this mode, IPv4 Link-Local Address support is
+ * disabled in ARP, and DHCP client support is disabled in IP input; turning
+ * any of them on would cause an error to be returned. Entering or exiting
+ * this mode will result in the removal of IPv4 addresses currently configured
+ * on the interface.
+ *
+ * Caller passes in the ioctl data pointer directly via "ifr", with the
+ * expectation that this routine always uses bcopy() or other byte-aligned
+ * memory accesses.
+ */
+static __attribute__((noinline)) int
+inctl_setrouter(struct ifnet *ifp, struct ifreq *ifr)
+{
+ int error = 0, intval;
- ev_msg.event_code = KEV_INET_SIFDSTADDR;
+ VERIFY(ifp != NULL);
- if (ia->ia_ifa.ifa_dstaddr) {
- in_event_data.ia_dstaddr = ((struct sockaddr_in *)
- (void *)ia->ia_ifa.ifa_dstaddr)->sin_addr;
- } else {
- in_event_data.ia_dstaddr.s_addr = INADDR_ANY;
- }
+ /* Router mode isn't valid for loopback */
+ if (ifp->if_flags & IFF_LOOPBACK) {
+ return ENODEV;
+ }
- in_event_data.ia_addr = ia->ia_addr.sin_addr;
- in_event_data.ia_net = ia->ia_net;
- in_event_data.ia_netmask = ia->ia_netmask;
- in_event_data.ia_subnet = ia->ia_subnet;
- in_event_data.ia_subnetmask = ia->ia_subnetmask;
- in_event_data.ia_netbroadcast = ia->ia_netbroadcast;
- IFA_UNLOCK(&ia->ia_ifa);
- (void) strncpy(&in_event_data.link_data.if_name[0],
- ifp->if_name, IFNAMSIZ);
- in_event_data.link_data.if_family = ifp->if_family;
- in_event_data.link_data.if_unit = (u_int32_t) ifp->if_unit;
+ bcopy(&ifr->ifr_intval, &intval, sizeof(intval));
- ev_msg.dv[0].data_ptr = &in_event_data;
- ev_msg.dv[0].data_length = sizeof (struct kev_in_data);
- ev_msg.dv[1].data_length = 0;
+ ifnet_lock_exclusive(ifp);
+ if (intval) {
+ ifp->if_eflags |= IFEF_IPV4_ROUTER;
+ ifp->if_eflags &= ~(IFEF_ARPLL | IFEF_AUTOCONFIGURING);
+ } else {
+ ifp->if_eflags &= ~IFEF_IPV4_ROUTER;
+ }
+ ifnet_lock_done(ifp);
- kev_post_msg(&ev_msg);
+ /* purge all IPv4 addresses configured on this interface */
+ in_purgeaddrs(ifp);
- lck_mtx_lock(rnh_lock);
- IFA_LOCK(&ia->ia_ifa);
- if (ia->ia_flags & IFA_ROUTE) {
- ia->ia_ifa.ifa_dstaddr = (struct sockaddr *)&oldaddr;
- IFA_UNLOCK(&ia->ia_ifa);
- rtinit_locked(&(ia->ia_ifa), (int)RTM_DELETE, RTF_HOST);
- IFA_LOCK(&ia->ia_ifa);
- ia->ia_ifa.ifa_dstaddr =
- (struct sockaddr *)&ia->ia_dstaddr;
- IFA_UNLOCK(&ia->ia_ifa);
- rtinit_locked(&(ia->ia_ifa), (int)RTM_ADD,
- RTF_HOST|RTF_UP);
- } else {
- IFA_UNLOCK(&ia->ia_ifa);
- }
- lck_mtx_unlock(rnh_lock);
- break;
+ return error;
+}
- case SIOCSIFBRDADDR: /* struct ifreq */
- VERIFY(ia != NULL);
- if ((ifp->if_flags & IFF_BROADCAST) == 0) {
- error = EINVAL;
- break;
- }
- IFA_LOCK(&ia->ia_ifa);
- bcopy(&((struct ifreq *)(void *)data)->ifr_broadaddr,
- &ia->ia_broadaddr, sizeof (struct sockaddr_in));
+/*
+ * Caller passes in the ioctl data pointer directly via "ifr", with the
+ * expectation that this routine always uses bcopy() or other byte-aligned
+ * memory accesses.
+ */
+static __attribute__((noinline)) int
+inctl_ifaddr(struct ifnet *ifp, struct in_ifaddr *ia, u_long cmd,
+ struct ifreq *ifr)
+{
+ struct kev_in_data in_event_data;
+ struct kev_msg ev_msg;
+ struct sockaddr_in addr;
+ struct ifaddr *ifa;
+ int error = 0;
- ev_msg.vendor_code = KEV_VENDOR_APPLE;
- ev_msg.kev_class = KEV_NETWORK_CLASS;
- ev_msg.kev_subclass = KEV_INET_SUBCLASS;
+ VERIFY(ifp != NULL);
- ev_msg.event_code = KEV_INET_SIFBRDADDR;
+ bzero(&in_event_data, sizeof(struct kev_in_data));
+ bzero(&ev_msg, sizeof(struct kev_msg));
- if (ia->ia_ifa.ifa_dstaddr) {
- in_event_data.ia_dstaddr = ((struct sockaddr_in *)
- (void *)ia->ia_ifa.ifa_dstaddr)->sin_addr;
- } else {
- in_event_data.ia_dstaddr.s_addr = INADDR_ANY;
+ switch (cmd) {
+ case SIOCGIFADDR: /* struct ifreq */
+ if (ia == NULL) {
+ error = EADDRNOTAVAIL;
+ break;
}
- in_event_data.ia_addr = ia->ia_addr.sin_addr;
- in_event_data.ia_net = ia->ia_net;
- in_event_data.ia_netmask = ia->ia_netmask;
- in_event_data.ia_subnet = ia->ia_subnet;
- in_event_data.ia_subnetmask = ia->ia_subnetmask;
- in_event_data.ia_netbroadcast = ia->ia_netbroadcast;
+ IFA_LOCK(&ia->ia_ifa);
+ bcopy(&ia->ia_addr, &ifr->ifr_addr, sizeof(addr));
IFA_UNLOCK(&ia->ia_ifa);
- (void) strncpy(&in_event_data.link_data.if_name[0],
- ifp->if_name, IFNAMSIZ);
- in_event_data.link_data.if_family = ifp->if_family;
- in_event_data.link_data.if_unit = (u_int32_t) ifp->if_unit;
-
- ev_msg.dv[0].data_ptr = &in_event_data;
- ev_msg.dv[0].data_length = sizeof (struct kev_in_data);
- ev_msg.dv[1].data_length = 0;
-
- kev_post_msg(&ev_msg);
break;
- case SIOCSIFADDR: { /* struct ifreq */
- struct sockaddr_in addr;
-
- VERIFY(ifp != NULL && ia != NULL);
- bcopy(&((struct ifreq *)(void *)data)->ifr_addr,
- &addr, sizeof (addr));
+ case SIOCSIFADDR: /* struct ifreq */
+ VERIFY(ia != NULL);
+ bcopy(&ifr->ifr_addr, &addr, sizeof(addr));
/*
* If this is a new address, the reference count for the
* hash table has been taken at creation time above.
*/
error = in_ifinit(ifp, ia, &addr, 1);
-#if PF
- if (!error)
- (void) pf_ifaddr_hook(ifp, cmd);
-#endif /* PF */
- break;
- }
-
- case SIOCPROTOATTACH: /* struct ifreq */
- VERIFY(ifp != NULL);
- error = in_domifattach(ifp);
- break;
-
- case SIOCPROTODETACH: /* struct ifreq */
- VERIFY(ifp != NULL);
- /*
- * If an IPv4 address is still present, refuse to detach.
- */
- ifnet_lock_shared(ifp);
- TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) {
- IFA_LOCK(ifa);
- if (ifa->ifa_addr->sa_family == AF_INET) {
- IFA_UNLOCK(ifa);
- break;
- }
- IFA_UNLOCK(ifa);
- }
- ifnet_lock_done(ifp);
- if (ifa != NULL) {
- error = EBUSY;
- break;
- }
-
- error = proto_unplumb(PF_INET, ifp);
- break;
-
- case SIOCSETROUTERMODE: { /* struct ifreq */
- int intval;
-
- VERIFY(ifp != NULL);
- bcopy(&((struct ifreq *)(void *)data)->ifr_intval,
- &intval, sizeof (intval));
-
- error = in_setrouter(ifp, intval);
- break;
- }
-
- case SIOCSIFNETMASK: { /* struct ifreq */
- struct sockaddr_in addr;
- in_addr_t i;
-
- VERIFY(ifp != NULL && ia != NULL);
- bcopy(&((struct ifreq *)(void *)data)->ifr_addr,
- &addr, sizeof (addr));
- i = addr.sin_addr.s_addr;
-
- IFA_LOCK(&ia->ia_ifa);
- ia->ia_subnetmask = ntohl(ia->ia_sockmask.sin_addr.s_addr = i);
- ev_msg.vendor_code = KEV_VENDOR_APPLE;
- ev_msg.kev_class = KEV_NETWORK_CLASS;
- ev_msg.kev_subclass = KEV_INET_SUBCLASS;
-
- ev_msg.event_code = KEV_INET_SIFNETMASK;
-
- if (ia->ia_ifa.ifa_dstaddr) {
- in_event_data.ia_dstaddr = ((struct sockaddr_in *)
- (void *)ia->ia_ifa.ifa_dstaddr)->sin_addr;
- } else {
- in_event_data.ia_dstaddr.s_addr = INADDR_ANY;
+ if (error == 0) {
+ (void) ifnet_notify_address(ifp, AF_INET);
}
- in_event_data.ia_addr = ia->ia_addr.sin_addr;
- in_event_data.ia_net = ia->ia_net;
- in_event_data.ia_netmask = ia->ia_netmask;
- in_event_data.ia_subnet = ia->ia_subnet;
- in_event_data.ia_subnetmask = ia->ia_subnetmask;
- in_event_data.ia_netbroadcast = ia->ia_netbroadcast;
- IFA_UNLOCK(&ia->ia_ifa);
- (void) strncpy(&in_event_data.link_data.if_name[0],
- ifp->if_name, IFNAMSIZ);
- in_event_data.link_data.if_family = ifp->if_family;
- in_event_data.link_data.if_unit = (u_int32_t) ifp->if_unit;
-
- ev_msg.dv[0].data_ptr = &in_event_data;
- ev_msg.dv[0].data_length = sizeof (struct kev_in_data);
- ev_msg.dv[1].data_length = 0;
-
- kev_post_msg(&ev_msg);
break;
- }
- case SIOCAIFADDR: { /* struct ifaliasreq */
- struct sockaddr_in addr, broadaddr, mask;
+ case SIOCAIFADDR: { /* struct {if,in_}aliasreq */
+ struct in_aliasreq *ifra = (struct in_aliasreq *)ifr;
+ struct sockaddr_in broadaddr, mask;
+ int hostIsNew, maskIsNew;
- VERIFY(ifp != NULL && ia != NULL);
- bcopy(&((struct ifaliasreq *)(void *)data)->ifra_addr,
- &addr, sizeof (addr));
- bcopy(&((struct ifaliasreq *)(void *)data)->ifra_broadaddr,
- &broadaddr, sizeof (broadaddr));
- bcopy(&((struct ifaliasreq *)(void *)data)->ifra_mask,
- &mask, sizeof (mask));
+ VERIFY(ia != NULL);
+ bcopy(&ifra->ifra_addr, &addr, sizeof(addr));
+ bcopy(&ifra->ifra_broadaddr, &broadaddr, sizeof(broadaddr));
+ bcopy(&ifra->ifra_mask, &mask, sizeof(mask));
maskIsNew = 0;
hostIsNew = 1;
IFA_LOCK(&ia->ia_ifa);
ia->ia_sockmask = mask;
ia->ia_subnetmask =
- ntohl(ia->ia_sockmask.sin_addr.s_addr);
+ ntohl(ia->ia_sockmask.sin_addr.s_addr);
maskIsNew = 1;
}
if ((ifp->if_flags & IFF_POINTOPOINT) &&
in_ifscrub(ifp, ia, 0);
IFA_LOCK(&ia->ia_ifa);
ia->ia_dstaddr = broadaddr;
- ia->ia_dstaddr.sin_len = sizeof (struct sockaddr_in);
+ ia->ia_dstaddr.sin_len = sizeof(struct sockaddr_in);
maskIsNew = 1; /* We lie; but the effect's the same */
}
if (addr.sin_family == AF_INET && (hostIsNew || maskIsNew)) {
} else {
IFA_UNLOCK(&ia->ia_ifa);
}
-#if PF
- if (!error)
- (void) pf_ifaddr_hook(ifp, cmd);
-#endif /* PF */
+ if (error == 0) {
+ (void) ifnet_notify_address(ifp, AF_INET);
+ }
IFA_LOCK(&ia->ia_ifa);
if ((ifp->if_flags & IFF_BROADCAST) &&
- (broadaddr.sin_family == AF_INET))
+ (broadaddr.sin_family == AF_INET)) {
ia->ia_broadaddr = broadaddr;
+ }
/*
* Report event.
*/
if ((error == 0) || (error == EEXIST)) {
- ev_msg.vendor_code = KEV_VENDOR_APPLE;
- ev_msg.kev_class = KEV_NETWORK_CLASS;
- ev_msg.kev_subclass = KEV_INET_SUBCLASS;
+ ev_msg.vendor_code = KEV_VENDOR_APPLE;
+ ev_msg.kev_class = KEV_NETWORK_CLASS;
+ ev_msg.kev_subclass = KEV_INET_SUBCLASS;
- if (hostIsNew)
+ if (hostIsNew) {
ev_msg.event_code = KEV_INET_NEW_ADDR;
- else
+ } else {
ev_msg.event_code = KEV_INET_CHANGED_ADDR;
+ }
if (ia->ia_ifa.ifa_dstaddr) {
in_event_data.ia_dstaddr =
} else {
in_event_data.ia_dstaddr.s_addr = INADDR_ANY;
}
- in_event_data.ia_addr = ia->ia_addr.sin_addr;
- in_event_data.ia_net = ia->ia_net;
- in_event_data.ia_netmask = ia->ia_netmask;
- in_event_data.ia_subnet = ia->ia_subnet;
- in_event_data.ia_subnetmask = ia->ia_subnetmask;
- in_event_data.ia_netbroadcast = ia->ia_netbroadcast;
+ in_event_data.ia_addr = ia->ia_addr.sin_addr;
+ in_event_data.ia_net = ia->ia_net;
+ in_event_data.ia_netmask = ia->ia_netmask;
+ in_event_data.ia_subnet = ia->ia_subnet;
+ in_event_data.ia_subnetmask = ia->ia_subnetmask;
+ in_event_data.ia_netbroadcast = ia->ia_netbroadcast;
IFA_UNLOCK(&ia->ia_ifa);
- (void) strncpy(&in_event_data.link_data.if_name[0],
+ (void) strlcpy(&in_event_data.link_data.if_name[0],
ifp->if_name, IFNAMSIZ);
in_event_data.link_data.if_family = ifp->if_family;
in_event_data.link_data.if_unit = ifp->if_unit;
- ev_msg.dv[0].data_ptr = &in_event_data;
- ev_msg.dv[0].data_length = sizeof (struct kev_in_data);
+ ev_msg.dv[0].data_ptr = &in_event_data;
+ ev_msg.dv[0].data_length = sizeof(struct kev_in_data);
ev_msg.dv[1].data_length = 0;
- kev_post_msg(&ev_msg);
+ dlil_post_complete_msg(ifp, &ev_msg);
} else {
IFA_UNLOCK(&ia->ia_ifa);
}
break;
}
- case SIOCDIFADDR: /* struct ifreq */
- VERIFY(ifp != NULL && ia != NULL);
+ case SIOCDIFADDR: /* struct ifreq */
+ VERIFY(ia != NULL);
error = ifnet_ioctl(ifp, PF_INET, SIOCDIFADDR, ia);
- if (error == EOPNOTSUPP)
+ if (error == EOPNOTSUPP) {
error = 0;
+ }
if (error != 0) {
break;
}
/* Fill out the kernel event information */
- ev_msg.vendor_code = KEV_VENDOR_APPLE;
- ev_msg.kev_class = KEV_NETWORK_CLASS;
- ev_msg.kev_subclass = KEV_INET_SUBCLASS;
+ ev_msg.vendor_code = KEV_VENDOR_APPLE;
+ ev_msg.kev_class = KEV_NETWORK_CLASS;
+ ev_msg.kev_subclass = KEV_INET_SUBCLASS;
- ev_msg.event_code = KEV_INET_ADDR_DELETED;
+ ev_msg.event_code = KEV_INET_ADDR_DELETED;
IFA_LOCK(&ia->ia_ifa);
if (ia->ia_ifa.ifa_dstaddr) {
- in_event_data.ia_dstaddr = ((struct sockaddr_in *)
- (void *)ia->ia_ifa.ifa_dstaddr)->sin_addr;
+ in_event_data.ia_dstaddr = ((struct sockaddr_in *)
+ (void *)ia->ia_ifa.ifa_dstaddr)->sin_addr;
} else {
in_event_data.ia_dstaddr.s_addr = INADDR_ANY;
}
- in_event_data.ia_addr = ia->ia_addr.sin_addr;
- in_event_data.ia_net = ia->ia_net;
- in_event_data.ia_netmask = ia->ia_netmask;
- in_event_data.ia_subnet = ia->ia_subnet;
- in_event_data.ia_subnetmask = ia->ia_subnetmask;
- in_event_data.ia_netbroadcast = ia->ia_netbroadcast;
+ in_event_data.ia_addr = ia->ia_addr.sin_addr;
+ in_event_data.ia_net = ia->ia_net;
+ in_event_data.ia_netmask = ia->ia_netmask;
+ in_event_data.ia_subnet = ia->ia_subnet;
+ in_event_data.ia_subnetmask = ia->ia_subnetmask;
+ in_event_data.ia_netbroadcast = ia->ia_netbroadcast;
IFA_UNLOCK(&ia->ia_ifa);
- (void) strncpy(&in_event_data.link_data.if_name[0],
+ (void) strlcpy(&in_event_data.link_data.if_name[0],
ifp->if_name, IFNAMSIZ);
in_event_data.link_data.if_family = ifp->if_family;
- in_event_data.link_data.if_unit = (u_int32_t) ifp->if_unit;
+ in_event_data.link_data.if_unit = (u_int32_t)ifp->if_unit;
ev_msg.dv[0].data_ptr = &in_event_data;
ev_msg.dv[0].data_length = sizeof(struct kev_in_data);
IFA_REMREF(ifa);
TAILQ_REMOVE(&in_ifaddrhead, ia, ia_link);
IFA_LOCK(ifa);
- if (IA_IS_HASHED(ia))
+ if (IA_IS_HASHED(ia)) {
in_iahash_remove(ia);
+ }
IFA_UNLOCK(ifa);
lck_rw_done(in_ifaddr_rwlock);
/* Our reference to this address is dropped at the bottom */
IFA_UNLOCK(ifa);
- /*
- * If the interface supports multicast, and no address is left,
- * remove the "all hosts" multicast group from that interface.
- */
- if ((ifp->if_flags & IFF_MULTICAST) != 0 ||
- ifp->if_allhostsinm != NULL ) {
+ /* invalidate route caches */
+ routegenid_inet_update();
+ /*
+ * If the interface supports multicast, and no address is left,
+ * remove the "all hosts" multicast group from that interface.
+ */
+ if ((ifp->if_flags & IFF_MULTICAST) ||
+ ifp->if_allhostsinm != NULL) {
TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) {
IFA_LOCK(ifa);
if (ifa->ifa_addr->sa_family == AF_INET) {
}
/* Post the kernel event */
- kev_post_msg(&ev_msg);
+ dlil_post_complete_msg(ifp, &ev_msg);
/*
* See if there is any IPV4 address left and if so,
* for legacy reasons.
*/
error = ifnet_ioctl(ifp, PF_INET, SIOCSIFADDR, ifa);
- if (error == EOPNOTSUPP)
+ if (error == EOPNOTSUPP) {
error = 0;
+ }
+
+ /* Release reference from ifa_ifpgetprimary() */
+ IFA_REMREF(ifa);
+ }
+ (void) ifnet_notify_address(ifp, AF_INET);
+ break;
+
+ default:
+ VERIFY(0);
+ /* NOTREACHED */
+ }
+
+ return error;
+}
+
+/*
+ * Caller passes in the ioctl data pointer directly via "ifr", with the
+ * expectation that this routine always uses bcopy() or other byte-aligned
+ * memory accesses.
+ */
+static __attribute__((noinline)) int
+inctl_ifdstaddr(struct ifnet *ifp, struct in_ifaddr *ia, u_long cmd,
+ struct ifreq *ifr)
+{
+ struct kev_in_data in_event_data;
+ struct kev_msg ev_msg;
+ struct sockaddr_in dstaddr;
+ int error = 0;
+
+ VERIFY(ifp != NULL);
+
+ if (!(ifp->if_flags & IFF_POINTOPOINT)) {
+ return EINVAL;
+ }
+
+ bzero(&in_event_data, sizeof(struct kev_in_data));
+ bzero(&ev_msg, sizeof(struct kev_msg));
+
+ switch (cmd) {
+ case SIOCGIFDSTADDR: /* struct ifreq */
+ if (ia == NULL) {
+ error = EADDRNOTAVAIL;
+ break;
+ }
+ IFA_LOCK(&ia->ia_ifa);
+ bcopy(&ia->ia_dstaddr, &ifr->ifr_dstaddr, sizeof(dstaddr));
+ IFA_UNLOCK(&ia->ia_ifa);
+ break;
+
+ case SIOCSIFDSTADDR: /* struct ifreq */
+ VERIFY(ia != NULL);
+ IFA_LOCK(&ia->ia_ifa);
+ dstaddr = ia->ia_dstaddr;
+ bcopy(&ifr->ifr_dstaddr, &ia->ia_dstaddr, sizeof(dstaddr));
+ if (ia->ia_dstaddr.sin_family == AF_INET) {
+ ia->ia_dstaddr.sin_len = sizeof(struct sockaddr_in);
+ }
+ IFA_UNLOCK(&ia->ia_ifa);
+ /*
+ * NOTE: SIOCSIFDSTADDR is defined with struct ifreq
+ * as parameter, but here we are sending it down
+ * to the interface with a pointer to struct ifaddr,
+ * for legacy reasons.
+ */
+ error = ifnet_ioctl(ifp, PF_INET, SIOCSIFDSTADDR, ia);
+ IFA_LOCK(&ia->ia_ifa);
+ if (error == EOPNOTSUPP) {
+ error = 0;
+ }
+ if (error != 0) {
+ ia->ia_dstaddr = dstaddr;
+ IFA_UNLOCK(&ia->ia_ifa);
+ break;
+ }
+ IFA_LOCK_ASSERT_HELD(&ia->ia_ifa);
+
+ ev_msg.vendor_code = KEV_VENDOR_APPLE;
+ ev_msg.kev_class = KEV_NETWORK_CLASS;
+ ev_msg.kev_subclass = KEV_INET_SUBCLASS;
+
+ ev_msg.event_code = KEV_INET_SIFDSTADDR;
+
+ if (ia->ia_ifa.ifa_dstaddr) {
+ in_event_data.ia_dstaddr = ((struct sockaddr_in *)
+ (void *)ia->ia_ifa.ifa_dstaddr)->sin_addr;
+ } else {
+ in_event_data.ia_dstaddr.s_addr = INADDR_ANY;
+ }
+
+ in_event_data.ia_addr = ia->ia_addr.sin_addr;
+ in_event_data.ia_net = ia->ia_net;
+ in_event_data.ia_netmask = ia->ia_netmask;
+ in_event_data.ia_subnet = ia->ia_subnet;
+ in_event_data.ia_subnetmask = ia->ia_subnetmask;
+ in_event_data.ia_netbroadcast = ia->ia_netbroadcast;
+ IFA_UNLOCK(&ia->ia_ifa);
+ (void) strlcpy(&in_event_data.link_data.if_name[0],
+ ifp->if_name, IFNAMSIZ);
+ in_event_data.link_data.if_family = ifp->if_family;
+ in_event_data.link_data.if_unit = (u_int32_t)ifp->if_unit;
+
+ ev_msg.dv[0].data_ptr = &in_event_data;
+ ev_msg.dv[0].data_length = sizeof(struct kev_in_data);
+ ev_msg.dv[1].data_length = 0;
+
+ dlil_post_complete_msg(ifp, &ev_msg);
+
+ lck_mtx_lock(rnh_lock);
+ IFA_LOCK(&ia->ia_ifa);
+ if (ia->ia_flags & IFA_ROUTE) {
+ ia->ia_ifa.ifa_dstaddr = (struct sockaddr *)&dstaddr;
+ IFA_UNLOCK(&ia->ia_ifa);
+ rtinit_locked(&(ia->ia_ifa), (int)RTM_DELETE, RTF_HOST);
+ IFA_LOCK(&ia->ia_ifa);
+ ia->ia_ifa.ifa_dstaddr =
+ (struct sockaddr *)&ia->ia_dstaddr;
+ IFA_UNLOCK(&ia->ia_ifa);
+ rtinit_locked(&(ia->ia_ifa), (int)RTM_ADD,
+ RTF_HOST | RTF_UP);
+ } else {
+ IFA_UNLOCK(&ia->ia_ifa);
+ }
+ lck_mtx_unlock(rnh_lock);
+ break;
+
+
+
+ default:
+ VERIFY(0);
+ /* NOTREACHED */
+ }
+
+ return error;
+}
+
+/*
+ * Caller passes in the ioctl data pointer directly via "ifr", with the
+ * expectation that this routine always uses bcopy() or other byte-aligned
+ * memory accesses.
+ */
+static __attribute__((noinline)) int
+inctl_ifbrdaddr(struct ifnet *ifp, struct in_ifaddr *ia, u_long cmd,
+ struct ifreq *ifr)
+{
+ struct kev_in_data in_event_data;
+ struct kev_msg ev_msg;
+ int error = 0;
+
+ VERIFY(ifp != NULL);
+
+ if (ia == NULL) {
+ return EADDRNOTAVAIL;
+ }
+
+ if (!(ifp->if_flags & IFF_BROADCAST)) {
+ return EINVAL;
+ }
+
+ bzero(&in_event_data, sizeof(struct kev_in_data));
+ bzero(&ev_msg, sizeof(struct kev_msg));
+
+ switch (cmd) {
+ case SIOCGIFBRDADDR: /* struct ifreq */
+ IFA_LOCK(&ia->ia_ifa);
+ bcopy(&ia->ia_broadaddr, &ifr->ifr_broadaddr,
+ sizeof(struct sockaddr_in));
+ IFA_UNLOCK(&ia->ia_ifa);
+ break;
+
+ case SIOCSIFBRDADDR: /* struct ifreq */
+ IFA_LOCK(&ia->ia_ifa);
+ bcopy(&ifr->ifr_broadaddr, &ia->ia_broadaddr,
+ sizeof(struct sockaddr_in));
+
+ ev_msg.vendor_code = KEV_VENDOR_APPLE;
+ ev_msg.kev_class = KEV_NETWORK_CLASS;
+ ev_msg.kev_subclass = KEV_INET_SUBCLASS;
+
+ ev_msg.event_code = KEV_INET_SIFBRDADDR;
+
+ if (ia->ia_ifa.ifa_dstaddr) {
+ in_event_data.ia_dstaddr = ((struct sockaddr_in *)
+ (void *)ia->ia_ifa.ifa_dstaddr)->sin_addr;
+ } else {
+ in_event_data.ia_dstaddr.s_addr = INADDR_ANY;
+ }
+ in_event_data.ia_addr = ia->ia_addr.sin_addr;
+ in_event_data.ia_net = ia->ia_net;
+ in_event_data.ia_netmask = ia->ia_netmask;
+ in_event_data.ia_subnet = ia->ia_subnet;
+ in_event_data.ia_subnetmask = ia->ia_subnetmask;
+ in_event_data.ia_netbroadcast = ia->ia_netbroadcast;
+ IFA_UNLOCK(&ia->ia_ifa);
+ (void) strlcpy(&in_event_data.link_data.if_name[0],
+ ifp->if_name, IFNAMSIZ);
+ in_event_data.link_data.if_family = ifp->if_family;
+ in_event_data.link_data.if_unit = (u_int32_t)ifp->if_unit;
+
+ ev_msg.dv[0].data_ptr = &in_event_data;
+ ev_msg.dv[0].data_length = sizeof(struct kev_in_data);
+ ev_msg.dv[1].data_length = 0;
+
+ dlil_post_complete_msg(ifp, &ev_msg);
+ break;
+
+ default:
+ VERIFY(0);
+ /* NOTREACHED */
+ }
+
+ return error;
+}
+
+/*
+ * Caller passes in the ioctl data pointer directly via "ifr", with the
+ * expectation that this routine always uses bcopy() or other byte-aligned
+ * memory accesses.
+ */
+static __attribute__((noinline)) int
+inctl_ifnetmask(struct ifnet *ifp, struct in_ifaddr *ia, u_long cmd,
+ struct ifreq *ifr)
+{
+ struct kev_in_data in_event_data;
+ struct kev_msg ev_msg;
+ struct sockaddr_in mask;
+ int error = 0;
+
+ VERIFY(ifp != NULL);
+
+ bzero(&in_event_data, sizeof(struct kev_in_data));
+ bzero(&ev_msg, sizeof(struct kev_msg));
+
+ switch (cmd) {
+ case SIOCGIFNETMASK: /* struct ifreq */
+ if (ia == NULL) {
+ error = EADDRNOTAVAIL;
+ break;
+ }
+ IFA_LOCK(&ia->ia_ifa);
+ bcopy(&ia->ia_sockmask, &ifr->ifr_addr, sizeof(mask));
+ IFA_UNLOCK(&ia->ia_ifa);
+ break;
+
+ case SIOCSIFNETMASK: { /* struct ifreq */
+ in_addr_t i;
+
+ bcopy(&ifr->ifr_addr, &mask, sizeof(mask));
+ i = mask.sin_addr.s_addr;
+
+ VERIFY(ia != NULL);
+ IFA_LOCK(&ia->ia_ifa);
+ ia->ia_subnetmask = ntohl(ia->ia_sockmask.sin_addr.s_addr = i);
+ ev_msg.vendor_code = KEV_VENDOR_APPLE;
+ ev_msg.kev_class = KEV_NETWORK_CLASS;
+ ev_msg.kev_subclass = KEV_INET_SUBCLASS;
+
+ ev_msg.event_code = KEV_INET_SIFNETMASK;
+
+ if (ia->ia_ifa.ifa_dstaddr) {
+ in_event_data.ia_dstaddr = ((struct sockaddr_in *)
+ (void *)ia->ia_ifa.ifa_dstaddr)->sin_addr;
+ } else {
+ in_event_data.ia_dstaddr.s_addr = INADDR_ANY;
+ }
+ in_event_data.ia_addr = ia->ia_addr.sin_addr;
+ in_event_data.ia_net = ia->ia_net;
+ in_event_data.ia_netmask = ia->ia_netmask;
+ in_event_data.ia_subnet = ia->ia_subnet;
+ in_event_data.ia_subnetmask = ia->ia_subnetmask;
+ in_event_data.ia_netbroadcast = ia->ia_netbroadcast;
+ IFA_UNLOCK(&ia->ia_ifa);
+ (void) strlcpy(&in_event_data.link_data.if_name[0],
+ ifp->if_name, IFNAMSIZ);
+ in_event_data.link_data.if_family = ifp->if_family;
+ in_event_data.link_data.if_unit = (u_int32_t)ifp->if_unit;
+
+ ev_msg.dv[0].data_ptr = &in_event_data;
+ ev_msg.dv[0].data_length = sizeof(struct kev_in_data);
+ ev_msg.dv[1].data_length = 0;
+
+ dlil_post_complete_msg(ifp, &ev_msg);
+ break;
+ }
+
+ default:
+ VERIFY(0);
+ /* NOTREACHED */
+ }
+
+ return error;
+}
+
+/*
+ * Generic INET control operations (ioctl's).
+ *
+ * ifp is NULL if not an interface-specific ioctl.
+ *
+ * Most of the routines called to handle the ioctls would end up being
+ * tail-call optimized, which unfortunately causes this routine to
+ * consume too much stack space; this is the reason for the "noinline"
+ * attribute used on those routines.
+ *
+ * If called directly from within the networking stack (as opposed to via
+ * pru_control), the socket parameter may be NULL.
+ */
+int
+in_control(struct socket *so, u_long cmd, caddr_t data, struct ifnet *ifp,
+ struct proc *p)
+{
+ struct ifreq *ifr = (struct ifreq *)(void *)data;
+ struct sockaddr_in addr, dstaddr;
+ struct sockaddr_in sin, *sa = NULL;
+ boolean_t privileged = (proc_suser(p) == 0);
+ boolean_t so_unlocked = FALSE;
+ struct in_ifaddr *ia = NULL;
+ struct ifaddr *ifa;
+ int error = 0;
+
+ /* In case it's NULL, make sure it came from the kernel */
+ VERIFY(so != NULL || p == kernproc);
+
+ /*
+ * ioctls which don't require ifp, but require socket.
+ */
+ switch (cmd) {
+ case SIOCGASSOCIDS32: /* struct so_aidreq32 */
+ case SIOCGASSOCIDS64: /* struct so_aidreq64 */
+ return inctl_associd(so, cmd, data);
+ /* NOTREACHED */
+
+ case SIOCGCONNIDS32: /* struct so_cidreq32 */
+ case SIOCGCONNIDS64: /* struct so_cidreq64 */
+ return inctl_connid(so, cmd, data);
+ /* NOTREACHED */
+
+ case SIOCGCONNINFO32: /* struct so_cinforeq32 */
+ case SIOCGCONNINFO64: /* struct so_cinforeq64 */
+ return inctl_conninfo(so, cmd, data);
+ /* NOTREACHED */
+ }
+
+ /*
+ * The rest of ioctls require ifp; reject if we don't have one;
+ * return ENXIO to be consistent with ifioctl().
+ */
+ if (ifp == NULL) {
+ return ENXIO;
+ }
- /* Release reference from ifa_ifpgetprimary() */
- IFA_REMREF(ifa);
+ /*
+ * ioctls which require ifp but not interface address.
+ */
+ switch (cmd) {
+ case SIOCAUTOADDR: /* struct ifreq */
+ if (!privileged) {
+ return EPERM;
}
-#if PF
- (void) pf_ifaddr_hook(ifp, cmd);
-#endif /* PF */
- break;
+ return inctl_autoaddr(ifp, ifr);
+ /* NOTREACHED */
-#ifdef __APPLE__
- case SIOCSETOT: { /* int */
- /*
- * Inspiration from tcp_ctloutput() and ip_ctloutput()
- * Special ioctl for OpenTransport sockets
- */
- struct inpcb *inp, *cloned_inp;
- int error2 = 0;
- int cloned_fd;
+ case SIOCARPIPLL: /* struct ifreq */
+ if (!privileged) {
+ return EPERM;
+ }
+ return inctl_arpipll(ifp, ifr);
+ /* NOTREACHED */
- bcopy(data, &cloned_fd, sizeof (cloned_fd));
+ case SIOCSETROUTERMODE: /* struct ifreq */
+ if (!privileged) {
+ return EPERM;
+ }
+ return inctl_setrouter(ifp, ifr);
+ /* NOTREACHED */
- inp = sotoinpcb(so);
- if (inp == NULL) {
- break;
+ case SIOCPROTOATTACH: /* struct ifreq */
+ if (!privileged) {
+ return EPERM;
+ }
+ return in_domifattach(ifp);
+ /* NOTREACHED */
+
+ case SIOCPROTODETACH: /* struct ifreq */
+ if (!privileged) {
+ return EPERM;
}
- /* let's make sure it's either -1 or a valid file descriptor */
- if (cloned_fd != -1) {
- struct socket *cloned_so;
- error2 = file_socket(cloned_fd, &cloned_so);
- if (error2) {
+ /*
+ * If an IPv4 address is still present, refuse to detach.
+ */
+ ifnet_lock_shared(ifp);
+ TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) {
+ IFA_LOCK(ifa);
+ if (ifa->ifa_addr->sa_family == AF_INET) {
+ IFA_UNLOCK(ifa);
break;
}
- cloned_inp = sotoinpcb(cloned_so);
- file_drop(cloned_fd);
- } else {
- cloned_inp = NULL;
+ IFA_UNLOCK(ifa);
}
+ ifnet_lock_done(ifp);
+ return (ifa == NULL) ? proto_unplumb(PF_INET, ifp) : EBUSY;
+ /* NOTREACHED */
+ }
- if (cloned_inp == NULL) {
- /* OT always uses IP_PORTRANGE_HIGH */
- inp->inp_flags &= ~(INP_LOWPORT);
- inp->inp_flags |= INP_HIGHPORT;
- /*
- * For UDP, OT allows broadcast by default;
- * for TCP we want to see MSG_OOB when we
- * receive urgent data.
- */
- if (so->so_type == SOCK_DGRAM)
- so->so_options |= SO_BROADCAST;
- else if (so->so_type == SOCK_STREAM)
- so->so_options |= SO_WANTOOBFLAG;
- } else {
- inp->inp_ip_tos = cloned_inp->inp_ip_tos;
- inp->inp_ip_ttl = cloned_inp->inp_ip_ttl;
- inp->inp_flags = cloned_inp->inp_flags;
+ /*
+ * ioctls which require interface address; obtain sockaddr_in.
+ */
+ switch (cmd) {
+ case SIOCAIFADDR: /* struct {if,in_}aliasreq */
+ if (!privileged) {
+ return EPERM;
+ }
+ bcopy(&((struct in_aliasreq *)(void *)data)->ifra_addr,
+ &sin, sizeof(sin));
+ sa = &sin;
+ break;
- /* Multicast options */
- if (cloned_inp->inp_moptions != NULL)
- error2 = imo_clone(cloned_inp, inp);
+ case SIOCDIFADDR: /* struct ifreq */
+ case SIOCSIFADDR: /* struct ifreq */
+ case SIOCSIFDSTADDR: /* struct ifreq */
+ case SIOCSIFNETMASK: /* struct ifreq */
+ case SIOCSIFBRDADDR: /* struct ifreq */
+ if (!privileged) {
+ return EPERM;
}
+ /* FALLTHRU */
+ case SIOCGIFADDR: /* struct ifreq */
+ case SIOCGIFDSTADDR: /* struct ifreq */
+ case SIOCGIFNETMASK: /* struct ifreq */
+ case SIOCGIFBRDADDR: /* struct ifreq */
+ bcopy(&ifr->ifr_addr, &sin, sizeof(sin));
+ sa = &sin;
break;
}
-#endif /* __APPLE__ */
- default:
- error = EOPNOTSUPP;
- }
- done:
- if (ia != NULL) {
- IFA_REMREF(&ia->ia_ifa);
- }
- return (error);
-}
+ /*
+ * Find address for this interface, if it exists.
+ *
+ * If an alias address was specified, find that one instead of
+ * the first one on the interface, if possible.
+ */
+ VERIFY(ia == NULL);
+ if (sa != NULL) {
+ struct in_ifaddr *iap;
-/*
- * SIOC[GAD]LIFADDR.
- * SIOCGLIFADDR: get first address. (?!?)
- * SIOCGLIFADDR with IFLR_PREFIX:
- * get first address that matches the specified prefix.
- * SIOCALIFADDR: add the specified address.
- * SIOCALIFADDR with IFLR_PREFIX:
- * EINVAL since we can't deduce hostid part of the address.
- * SIOCDLIFADDR: delete the specified address.
- * SIOCDLIFADDR with IFLR_PREFIX:
- * delete the first address that matches the specified prefix.
- * return values:
- * EINVAL on invalid parameters
- * EADDRNOTAVAIL on prefix match failed/specified address not found
- * other values may be returned from in_ioctl()
- */
-static int
-in_lifaddr_ioctl(struct socket *so, u_long cmd, struct if_laddrreq *iflr,
- struct ifnet *ifp, struct proc *p)
-{
- struct ifaddr *ifa;
+ /*
+ * Any failures from this point on must take into account
+ * a non-NULL "ia" with an outstanding reference count, and
+ * therefore requires IFA_REMREF. Jump to "done" label
+ * instead of calling return if "ia" is valid.
+ */
+ lck_rw_lock_shared(in_ifaddr_rwlock);
+ TAILQ_FOREACH(iap, INADDR_HASH(sa->sin_addr.s_addr), ia_hash) {
+ IFA_LOCK(&iap->ia_ifa);
+ if (iap->ia_ifp == ifp &&
+ iap->ia_addr.sin_addr.s_addr ==
+ sa->sin_addr.s_addr) {
+ ia = iap;
+ IFA_ADDREF_LOCKED(&iap->ia_ifa);
+ IFA_UNLOCK(&iap->ia_ifa);
+ break;
+ }
+ IFA_UNLOCK(&iap->ia_ifa);
+ }
+ lck_rw_done(in_ifaddr_rwlock);
- VERIFY(ifp != NULL);
+ if (ia == NULL) {
+ ifnet_lock_shared(ifp);
+ TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) {
+ iap = ifatoia(ifa);
+ IFA_LOCK(&iap->ia_ifa);
+ if (iap->ia_addr.sin_family == AF_INET) {
+ ia = iap;
+ IFA_ADDREF_LOCKED(&iap->ia_ifa);
+ IFA_UNLOCK(&iap->ia_ifa);
+ break;
+ }
+ IFA_UNLOCK(&iap->ia_ifa);
+ }
+ ifnet_lock_done(ifp);
+ }
+ }
- switch (cmd) {
- case SIOCGLIFADDR:
- /* address must be specified on GET with IFLR_PREFIX */
- if ((iflr->flags & IFLR_PREFIX) == 0)
- break;
- /*FALLTHROUGH*/
- case SIOCALIFADDR:
- case SIOCDLIFADDR:
- /* address must be specified on ADD and DELETE */
- if (iflr->addr.ss_family != AF_INET)
- return EINVAL;
- if (iflr->addr.ss_len != sizeof(struct sockaddr_in))
- return EINVAL;
- /* XXX need improvement */
- if (iflr->dstaddr.ss_family
- && iflr->dstaddr.ss_family != AF_INET)
- return EINVAL;
- if (iflr->dstaddr.ss_family
- && iflr->dstaddr.ss_len != sizeof(struct sockaddr_in))
- return EINVAL;
- break;
- default: /*shouldn't happen*/
- return EOPNOTSUPP;
+ /*
+ * Unlock the socket since ifnet_ioctl() may be invoked by
+ * one of the ioctl handlers below. Socket will be re-locked
+ * prior to returning.
+ */
+ if (so != NULL) {
+ socket_unlock(so, 0);
+ so_unlocked = TRUE;
}
- if (sizeof(struct in_addr) * 8 < iflr->prefixlen)
- return EINVAL;
switch (cmd) {
- case SIOCALIFADDR:
- {
- struct in_aliasreq ifra;
-
- if (iflr->flags & IFLR_PREFIX)
- return EINVAL;
-
- /* copy args to in_aliasreq, perform ioctl(SIOCAIFADDR_IN6). */
- bzero(&ifra, sizeof(ifra));
- bcopy(iflr->iflr_name, ifra.ifra_name,
- sizeof(ifra.ifra_name));
-
- bcopy(&iflr->addr, &ifra.ifra_addr, iflr->addr.ss_len);
-
- if (iflr->dstaddr.ss_family) { /*XXX*/
- bcopy(&iflr->dstaddr, &ifra.ifra_dstaddr,
- iflr->dstaddr.ss_len);
- }
-
- ifra.ifra_mask.sin_family = AF_INET;
- ifra.ifra_mask.sin_len = sizeof(struct sockaddr_in);
- in_len2mask(&ifra.ifra_mask.sin_addr, iflr->prefixlen);
-
- return in_control(so, SIOCAIFADDR, (caddr_t)&ifra, ifp, p);
- }
- case SIOCGLIFADDR:
- case SIOCDLIFADDR:
- {
- struct in_ifaddr *ia;
- struct in_addr mask, candidate;
- struct in_addr match = { 0 };
- struct sockaddr_in *sin;
- int cmp;
-
- bzero(&mask, sizeof(mask));
- if (iflr->flags & IFLR_PREFIX) {
- /* lookup a prefix rather than address. */
- in_len2mask(&mask, iflr->prefixlen);
-
- sin = (struct sockaddr_in *)&iflr->addr;
- match.s_addr = sin->sin_addr.s_addr;
- match.s_addr &= mask.s_addr;
-
- /* if you set extra bits, that's wrong */
- if (match.s_addr != sin->sin_addr.s_addr)
- return EINVAL;
-
- cmp = 1;
+ case SIOCAIFADDR: /* struct {if,in_}aliasreq */
+ case SIOCDIFADDR: /* struct ifreq */
+ if (cmd == SIOCAIFADDR) {
+ bcopy(&((struct in_aliasreq *)(void *)data)->
+ ifra_addr, &addr, sizeof(addr));
+ bcopy(&((struct in_aliasreq *)(void *)data)->
+ ifra_dstaddr, &dstaddr, sizeof(dstaddr));
} else {
- if (cmd == SIOCGLIFADDR) {
- /* on getting an address, take the 1st match */
- cmp = 0; /*XXX*/
- } else {
- /* on deleting an address, do exact match */
- in_len2mask(&mask, 32);
- sin = (struct sockaddr_in *)&iflr->addr;
- match.s_addr = sin->sin_addr.s_addr;
-
- cmp = 1;
- }
+ VERIFY(cmd == SIOCDIFADDR);
+ bcopy(&((struct ifreq *)(void *)data)->ifr_addr,
+ &addr, sizeof(addr));
+ bzero(&dstaddr, sizeof(dstaddr));
}
- ifnet_lock_shared(ifp);
- TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) {
- IFA_LOCK(ifa);
- if (ifa->ifa_addr->sa_family != AF_INET6) {
- IFA_UNLOCK(ifa);
- continue;
+ if (addr.sin_family == AF_INET) {
+ struct in_ifaddr *oia;
+
+ lck_rw_lock_shared(in_ifaddr_rwlock);
+ for (oia = ia; ia; ia = ia->ia_link.tqe_next) {
+ IFA_LOCK(&ia->ia_ifa);
+ if (ia->ia_ifp == ifp &&
+ ia->ia_addr.sin_addr.s_addr ==
+ addr.sin_addr.s_addr) {
+ IFA_ADDREF_LOCKED(&ia->ia_ifa);
+ IFA_UNLOCK(&ia->ia_ifa);
+ break;
+ }
+ IFA_UNLOCK(&ia->ia_ifa);
}
- if (!cmp) {
- IFA_UNLOCK(ifa);
- break;
+ lck_rw_done(in_ifaddr_rwlock);
+ if (oia != NULL) {
+ IFA_REMREF(&oia->ia_ifa);
}
- candidate.s_addr = ((struct sockaddr_in *)&ifa->ifa_addr)->sin_addr.s_addr;
- candidate.s_addr &= mask.s_addr;
- IFA_UNLOCK(ifa);
- if (candidate.s_addr == match.s_addr)
- break;
+ if ((ifp->if_flags & IFF_POINTOPOINT) &&
+ (cmd == SIOCAIFADDR) &&
+ (dstaddr.sin_addr.s_addr == INADDR_ANY)) {
+ error = EDESTADDRREQ;
+ goto done;
+ }
+ } else if (cmd == SIOCAIFADDR) {
+ error = EINVAL;
+ goto done;
}
- if (ifa != NULL)
- IFA_ADDREF(ifa);
- ifnet_lock_done(ifp);
- if (!ifa)
- return EADDRNOTAVAIL;
- ia = (struct in_ifaddr *)ifa;
-
- if (cmd == SIOCGLIFADDR) {
- IFA_LOCK(ifa);
- /* fill in the if_laddrreq structure */
- bcopy(&ia->ia_addr, &iflr->addr, ia->ia_addr.sin_len);
+ if (cmd == SIOCDIFADDR) {
+ if (ia == NULL) {
+ error = EADDRNOTAVAIL;
+ goto done;
+ }
- if ((ifp->if_flags & IFF_POINTOPOINT) != 0) {
- bcopy(&ia->ia_dstaddr, &iflr->dstaddr,
- ia->ia_dstaddr.sin_len);
- } else
- bzero(&iflr->dstaddr, sizeof(iflr->dstaddr));
+ IFA_LOCK(&ia->ia_ifa);
+ /*
+ * Avoid the race condition seen when two
+ * threads process SIOCDIFADDR command
+ * at the same time.
+ */
+ while (ia->ia_ifa.ifa_debug & IFD_DETACHING) {
+ os_log(OS_LOG_DEFAULT,
+ "Another thread is already attempting to "
+ "delete IPv4 address: %s on interface %s. "
+ "Go to sleep and check again after the operation is done",
+ inet_ntoa(sa->sin_addr), ia->ia_ifp->if_xname);
+ ia->ia_ifa.ifa_del_waiters++;
+ (void) msleep(ia->ia_ifa.ifa_del_wc, &ia->ia_ifa.ifa_lock, (PZERO - 1),
+ __func__, NULL);
+ IFA_LOCK_ASSERT_HELD(&ia->ia_ifa);
+ }
- iflr->prefixlen =
- in_mask2len(&ia->ia_sockmask.sin_addr);
+ if ((ia->ia_ifa.ifa_debug & IFD_ATTACHED) == 0) {
+ error = EADDRNOTAVAIL;
+ IFA_UNLOCK(&ia->ia_ifa);
+ goto done;
+ }
- iflr->flags = 0; /*XXX*/
+ ia->ia_ifa.ifa_debug |= IFD_DETACHING;
+ IFA_UNLOCK(&ia->ia_ifa);
+ }
- IFA_UNLOCK(ifa);
- IFA_REMREF(ifa);
- return 0;
+ /* FALLTHROUGH */
+ case SIOCSIFADDR: /* struct ifreq */
+ case SIOCSIFDSTADDR: /* struct ifreq */
+ case SIOCSIFNETMASK: /* struct ifreq */
+ if (cmd == SIOCAIFADDR) {
+ /* fell thru from above; just repeat it */
+ bcopy(&((struct in_aliasreq *)(void *)data)->
+ ifra_addr, &addr, sizeof(addr));
} else {
- struct in_aliasreq ifra;
-
- /* fill in_aliasreq and do ioctl(SIOCDIFADDR_IN6) */
- bzero(&ifra, sizeof(ifra));
- bcopy(iflr->iflr_name, ifra.ifra_name,
- sizeof(ifra.ifra_name));
+ VERIFY(cmd == SIOCDIFADDR || cmd == SIOCSIFADDR ||
+ cmd == SIOCSIFNETMASK || cmd == SIOCSIFDSTADDR);
+ bcopy(&((struct ifreq *)(void *)data)->ifr_addr,
+ &addr, sizeof(addr));
+ }
+ if (addr.sin_family != AF_INET && cmd == SIOCSIFADDR) {
+ error = EINVAL;
+ goto done;
+ }
+ if (ia == NULL) {
+ ia = in_ifaddr_alloc(M_WAITOK);
+ if (ia == NULL) {
+ error = ENOBUFS;
+ goto done;
+ }
+ ifnet_lock_exclusive(ifp);
+ ifa = &ia->ia_ifa;
IFA_LOCK(ifa);
- bcopy(&ia->ia_addr, &ifra.ifra_addr,
- ia->ia_addr.sin_len);
- if ((ifp->if_flags & IFF_POINTOPOINT) != 0) {
- bcopy(&ia->ia_dstaddr, &ifra.ifra_dstaddr,
- ia->ia_dstaddr.sin_len);
+ /* Hold a reference for this routine */
+ IFA_ADDREF_LOCKED(ifa);
+ IA_HASH_INIT(ia);
+ ifa->ifa_addr = (struct sockaddr *)&ia->ia_addr;
+ ifa->ifa_dstaddr = (struct sockaddr *)&ia->ia_dstaddr;
+ ifa->ifa_netmask = (struct sockaddr *)&ia->ia_sockmask;
+ ia->ia_sockmask.sin_len = 8;
+ if (ifp->if_flags & IFF_BROADCAST) {
+ ia->ia_broadaddr.sin_len = sizeof(ia->ia_addr);
+ ia->ia_broadaddr.sin_family = AF_INET;
+ }
+ ia->ia_ifp = ifp;
+ if (!(ifp->if_flags & IFF_LOOPBACK)) {
+ in_interfaces++;
+ }
+ /* if_attach_ifa() holds a reference for ifa_link */
+ if_attach_ifa(ifp, ifa);
+ /*
+ * If we have to go through in_ifinit(), make sure
+ * to avoid installing route(s) based on this address
+ * via PFC_IFUP event, before the link resolver (ARP)
+ * initializes it.
+ */
+ if (cmd == SIOCAIFADDR || cmd == SIOCSIFADDR) {
+ ifa->ifa_debug |= IFD_NOTREADY;
}
- bcopy(&ia->ia_sockmask, &ifra.ifra_dstaddr,
- ia->ia_sockmask.sin_len);
IFA_UNLOCK(ifa);
- IFA_REMREF(ifa);
- return in_control(so, SIOCDIFADDR, (caddr_t)&ifra,
- ifp, p);
+ ifnet_lock_done(ifp);
+ lck_rw_lock_exclusive(in_ifaddr_rwlock);
+ /* Hold a reference for ia_link */
+ IFA_ADDREF(ifa);
+ TAILQ_INSERT_TAIL(&in_ifaddrhead, ia, ia_link);
+ lck_rw_done(in_ifaddr_rwlock);
+ /* discard error */
+ (void) in_domifattach(ifp);
+ error = 0;
}
- }
+ break;
}
- return EOPNOTSUPP; /*just for safety*/
-}
+ switch (cmd) {
+ case SIOCGIFDSTADDR: /* struct ifreq */
+ case SIOCSIFDSTADDR: /* struct ifreq */
+ error = inctl_ifdstaddr(ifp, ia, cmd, ifr);
+ break;
-/*
- * Handle SIOCSETROUTERMODE to set or clear the IPv4 router mode flag on
- * the interface. When in this mode, IPv4 Link-Local Address support is
- * disabled in ARP, and DHCP client support is disabled in IP input; turning
- * any of them on would cause an error to be returned. Entering or exiting
- * this mode will result in the removal of IPv4 addresses currently configured
- * on the interface.
- */
-static int
-in_setrouter(struct ifnet *ifp, int enable)
-{
- if (ifp->if_flags & IFF_LOOPBACK)
- return (ENODEV);
+ case SIOCGIFBRDADDR: /* struct ifreq */
+ case SIOCSIFBRDADDR: /* struct ifreq */
+ error = inctl_ifbrdaddr(ifp, ia, cmd, ifr);
+ break;
- ifnet_lock_exclusive(ifp);
- if (enable) {
- ifp->if_eflags |= IFEF_IPV4_ROUTER;
- ifp->if_eflags &= ~(IFEF_ARPLL | IFEF_AUTOCONFIGURING);
- } else {
- ifp->if_eflags &= ~IFEF_IPV4_ROUTER;
+ case SIOCGIFNETMASK: /* struct ifreq */
+ case SIOCSIFNETMASK: /* struct ifreq */
+ error = inctl_ifnetmask(ifp, ia, cmd, ifr);
+ break;
+
+ case SIOCGIFADDR: /* struct ifreq */
+ case SIOCSIFADDR: /* struct ifreq */
+ case SIOCAIFADDR: /* struct {if,in_}aliasreq */
+ case SIOCDIFADDR: /* struct ifreq */
+ error = inctl_ifaddr(ifp, ia, cmd, ifr);
+ break;
+
+ default:
+ error = EOPNOTSUPP;
+ break;
}
- ifnet_lock_done(ifp);
- /* purge all IPv4 addresses configured on this interface */
- in_purgeaddrs(ifp);
+done:
+ if (ia != NULL) {
+ if (cmd == SIOCDIFADDR) {
+ IFA_LOCK(&ia->ia_ifa);
+ ia->ia_ifa.ifa_debug &= ~IFD_DETACHING;
+ if (ia->ia_ifa.ifa_del_waiters > 0) {
+ ia->ia_ifa.ifa_del_waiters = 0;
+ wakeup(ia->ia_ifa.ifa_del_wc);
+ }
+ IFA_UNLOCK(&ia->ia_ifa);
+ }
+ IFA_REMREF(&ia->ia_ifa);
+ }
+ if (so_unlocked) {
+ socket_lock(so, 0);
+ }
- return (0);
+ return error;
}
/*
return;
}
IFA_UNLOCK(&ia->ia_ifa);
- if (!locked)
+ if (!locked) {
lck_mtx_lock(rnh_lock);
- if (ifp->if_flags & (IFF_LOOPBACK|IFF_POINTOPOINT))
+ }
+ if (ifp->if_flags & (IFF_LOOPBACK | IFF_POINTOPOINT)) {
rtinit_locked(&(ia->ia_ifa), (int)RTM_DELETE, RTF_HOST);
- else
+ } else {
rtinit_locked(&(ia->ia_ifa), (int)RTM_DELETE, 0);
+ }
IFA_LOCK(&ia->ia_ifa);
ia->ia_flags &= ~IFA_ROUTE;
IFA_UNLOCK(&ia->ia_ifa);
- if (!locked)
+ if (!locked) {
lck_mtx_unlock(rnh_lock);
+ }
}
/*
static void
in_iahash_remove(struct in_ifaddr *ia)
{
- lck_rw_assert(in_ifaddr_rwlock, LCK_RW_ASSERT_EXCLUSIVE);
+ LCK_RW_ASSERT(in_ifaddr_rwlock, LCK_RW_ASSERT_EXCLUSIVE);
IFA_LOCK_ASSERT_HELD(&ia->ia_ifa);
if (!IA_IS_HASHED(ia)) {
static void
in_iahash_insert(struct in_ifaddr *ia)
{
- lck_rw_assert(in_ifaddr_rwlock, LCK_RW_ASSERT_EXCLUSIVE);
+ LCK_RW_ASSERT(in_ifaddr_rwlock, LCK_RW_ASSERT_EXCLUSIVE);
IFA_LOCK_ASSERT_HELD(&ia->ia_ifa);
if (ia->ia_addr.sin_family != AF_INET) {
panic("attempt to double-insert ia %p into hash table\n", ia);
/* NOTREACHED */
}
- TAILQ_INSERT_HEAD(INADDR_HASH(ia->ia_addr.sin_addr.s_addr), ia, ia_hash);
+ TAILQ_INSERT_HEAD(INADDR_HASH(ia->ia_addr.sin_addr.s_addr),
+ ia, ia_hash);
IFA_ADDREF_LOCKED(&ia->ia_ifa);
}
/*
- * Some point to point interfaces that are tunnels
- * borrow the address from an underlying interface (e.g.
- * VPN server). In order for source address selection logic to
- * find the underlying interface first, we add the address
+ * Some point to point interfaces that are tunnels borrow the address from
+ * an underlying interface (e.g. VPN server). In order for source address
+ * selection logic to find the underlying interface first, we add the address
* of borrowing point to point interfaces at the end of the list.
* (see rdar://6733789)
*
struct in_ifaddr *tmp_ifa;
struct ifnet *tmp_ifp;
- lck_rw_assert(in_ifaddr_rwlock, LCK_RW_ASSERT_EXCLUSIVE);
+ LCK_RW_ASSERT(in_ifaddr_rwlock, LCK_RW_ASSERT_EXCLUSIVE);
IFA_LOCK_ASSERT_HELD(&ia->ia_ifa);
if (ia->ia_addr.sin_family != AF_INET) {
* and routing table entry.
*/
static int
-in_ifinit(
- struct ifnet *ifp,
- struct in_ifaddr *ia,
- struct sockaddr_in *sin,
- int scrub)
+in_ifinit(struct ifnet *ifp, struct in_ifaddr *ia, struct sockaddr_in *sin,
+ int scrub)
{
u_int32_t i = ntohl(sin->sin_addr.s_addr);
struct sockaddr_in oldaddr;
in_iahash_remove(ia);
}
ia->ia_addr = *sin;
- ia->ia_addr.sin_len = sizeof (*sin);
- if ((ifp->if_flags & IFF_POINTOPOINT))
+ /*
+ * Interface addresses should not contain port or sin_zero information.
+ */
+ SIN(&ia->ia_addr)->sin_family = AF_INET;
+ SIN(&ia->ia_addr)->sin_len = sizeof(struct sockaddr_in);
+ SIN(&ia->ia_addr)->sin_port = 0;
+ bzero(&SIN(&ia->ia_addr)->sin_zero, sizeof(sin->sin_zero));
+ if ((ifp->if_flags & IFF_POINTOPOINT)) {
in_iahash_insert_ptp(ia);
- else
+ } else {
in_iahash_insert(ia);
+ }
IFA_UNLOCK(&ia->ia_ifa);
lck_rw_done(in_ifaddr_rwlock);
ifa0 = ifa_ifpgetprimary(ifp, AF_INET);
cmd = (&ia->ia_ifa == ifa0) ? SIOCSIFADDR : SIOCAIFADDR;
error = ifnet_ioctl(ifp, PF_INET, cmd, ia);
- if (error == EOPNOTSUPP)
+ if (error == EOPNOTSUPP) {
error = 0;
+ }
/*
* If we've just sent down SIOCAIFADDR, send another ioctl down
* for SIOCSIFADDR for the first IPV4 address of the interface,
* for legacy reasons.
*/
error = ifnet_ioctl(ifp, PF_INET, SIOCSIFADDR, ifa0);
- if (error == EOPNOTSUPP)
+ if (error == EOPNOTSUPP) {
error = 0;
+ }
}
/* Release reference from ifa_ifpgetprimary() */
if (error) {
lck_rw_lock_exclusive(in_ifaddr_rwlock);
IFA_LOCK(&ia->ia_ifa);
- if (IA_IS_HASHED(ia))
+ if (IA_IS_HASHED(ia)) {
in_iahash_remove(ia);
+ }
ia->ia_addr = oldaddr;
if (oldremoved) {
- if ((ifp->if_flags & IFF_POINTOPOINT))
+ if ((ifp->if_flags & IFF_POINTOPOINT)) {
in_iahash_insert_ptp(ia);
- else
+ } else {
in_iahash_insert(ia);
+ }
}
IFA_UNLOCK(&ia->ia_ifa);
lck_rw_done(in_ifaddr_rwlock);
/* Release extra reference taken above */
IFA_REMREF(&ia->ia_ifa);
- return (error);
+ return error;
}
lck_mtx_lock(rnh_lock);
IFA_LOCK(&ia->ia_ifa);
ia->ia_ifa.ifa_addr = (struct sockaddr *)&ia->ia_addr;
}
IFA_LOCK_ASSERT_HELD(&ia->ia_ifa);
- if (IN_CLASSA(i))
+ if (IN_CLASSA(i)) {
ia->ia_netmask = IN_CLASSA_NET;
- else if (IN_CLASSB(i))
+ } else if (IN_CLASSB(i)) {
ia->ia_netmask = IN_CLASSB_NET;
- else
+ } else {
ia->ia_netmask = IN_CLASSC_NET;
+ }
/*
* The subnet mask usually includes at least the standard network part,
* but may may be smaller in the case of supernetting.
if (ia->ia_subnetmask == 0) {
ia->ia_subnetmask = ia->ia_netmask;
ia->ia_sockmask.sin_addr.s_addr = htonl(ia->ia_subnetmask);
- } else
+ } else {
ia->ia_netmask &= ia->ia_subnetmask;
+ }
ia->ia_net = i & ia->ia_netmask;
ia->ia_subnet = i & ia->ia_subnetmask;
in_socktrim(&ia->ia_sockmask);
ia->ia_ifa.ifa_metric = ifp->if_metric;
if (ifp->if_flags & IFF_BROADCAST) {
ia->ia_broadaddr.sin_addr.s_addr =
- htonl(ia->ia_subnet | ~ia->ia_subnetmask);
+ htonl(ia->ia_subnet | ~ia->ia_subnetmask);
ia->ia_netbroadcast.s_addr =
- htonl(ia->ia_net | ~ ia->ia_netmask);
+ htonl(ia->ia_net | ~ia->ia_netmask);
} else if (ifp->if_flags & IFF_LOOPBACK) {
ia->ia_ifa.ifa_dstaddr = ia->ia_ifa.ifa_addr;
flags |= RTF_HOST;
lck_mtx_unlock(rnh_lock);
/* Release extra reference taken above */
IFA_REMREF(&ia->ia_ifa);
- return (0);
+ return 0;
}
- ia->ia_dstaddr.sin_len = sizeof (*sin);
+ ia->ia_dstaddr.sin_len = sizeof(struct sockaddr_in);
flags |= RTF_HOST;
}
IFA_UNLOCK(&ia->ia_ifa);
+
if ((error = rtinit_locked(&(ia->ia_ifa), (int)RTM_ADD, flags)) == 0) {
IFA_LOCK(&ia->ia_ifa);
ia->ia_flags |= IFA_ROUTE;
lck_mtx_unlock(rnh_lock);
/* XXX check if the subnet route points to the same interface */
- if (error == EEXIST)
+ if (error == EEXIST) {
error = 0;
+ }
/*
* If the interface supports multicast, join the "all hosts"
inm = in_addmulti(&addr, ifp);
if (inm != NULL) {
- /* keep the reference on inm added by
- * in_addmulti above for storing the
- * pointer in allhostsinm
+ /*
+ * Keep the reference on inm added by
+ * in_addmulti above for storing the
+ * pointer in allhostsinm.
*/
ifp->if_allhostsinm = inm;
} else {
- printf("Failed to add membership to all-hosts multicast address on interface %s%d\n", ifp->if_name, ifp->if_unit);
+ printf("%s: failed to add membership to "
+ "all-hosts multicast address on %s\n",
+ __func__, if_name(ifp));
}
}
lck_mtx_unlock(&ifp->if_addrconfig_lock);
/* Release extra reference taken above */
IFA_REMREF(&ia->ia_ifa);
- return (error);
-}
+ if (error == 0) {
+ /* invalidate route caches */
+ routegenid_inet_update();
+ }
+
+ return error;
+}
/*
- * Return 1 if the address might be a local broadcast address.
+ * Return TRUE if the address might be a local broadcast address.
*/
-int
+boolean_t
in_broadcast(struct in_addr in, struct ifnet *ifp)
{
struct ifaddr *ifa;
u_int32_t t;
- if (in.s_addr == INADDR_BROADCAST || in.s_addr == INADDR_ANY)
- return (1);
- if ((ifp->if_flags & IFF_BROADCAST) == 0)
- return (0);
+ if (in.s_addr == INADDR_BROADCAST || in.s_addr == INADDR_ANY) {
+ return TRUE;
+ }
+ if (!(ifp->if_flags & IFF_BROADCAST)) {
+ return FALSE;
+ }
t = ntohl(in.s_addr);
+
/*
* Look through the list of addresses for a match
* with a broadcast address.
IFA_LOCK(ifa);
if (ifa->ifa_addr->sa_family == AF_INET &&
(in.s_addr == ia->ia_broadaddr.sin_addr.s_addr ||
- in.s_addr == ia->ia_netbroadcast.s_addr ||
- /*
- * Check for old-style (host 0) broadcast.
- */
- t == ia->ia_subnet || t == ia->ia_net) &&
- /*
- * Check for an all one subnetmask. These
- * only exist when an interface gets a secondary
- * address.
- */
- ia->ia_subnetmask != (u_int32_t)0xffffffff) {
+ in.s_addr == ia->ia_netbroadcast.s_addr ||
+ /*
+ * Check for old-style (host 0) broadcast.
+ */
+ t == ia->ia_subnet || t == ia->ia_net) &&
+ /*
+ * Check for an all one subnetmask. These
+ * only exist when an interface gets a secondary
+ * address.
+ */
+ ia->ia_subnetmask != (u_int32_t)0xffffffff) {
IFA_UNLOCK(ifa);
ifnet_lock_done(ifp);
- return (1);
+ return TRUE;
}
IFA_UNLOCK(ifa);
}
ifnet_lock_done(ifp);
- return (0);
+ return FALSE;
#undef ia
}
struct ifaddr **ifap;
int err, i;
+ VERIFY(ifp != NULL);
+
/*
* Be nice, and try the civilized way first. If we can't get
* rid of them this way, then do it the rough way. We must
* removed from the global list and arrays.
*/
err = ifnet_get_address_list_family_internal(ifp, &ifap, AF_INET, 1,
- M_WAITOK);
+ M_WAITOK, 0);
if (err == 0 && ifap != NULL) {
+ struct ifreq ifr;
+
+ bzero(&ifr, sizeof(ifr));
+ (void) snprintf(ifr.ifr_name, sizeof(ifr.ifr_name),
+ "%s", if_name(ifp));
+
for (i = 0; ifap[i] != NULL; i++) {
- struct ifaliasreq ifr;
struct ifaddr *ifa;
ifa = ifap[i];
- bzero(&ifr, sizeof (ifr));
IFA_LOCK(ifa);
- ifr.ifra_addr = *ifa->ifa_addr;
- if (ifa->ifa_dstaddr != NULL)
- ifr.ifra_broadaddr = *ifa->ifa_dstaddr;
+ bcopy(ifa->ifa_addr, &ifr.ifr_addr,
+ sizeof(struct sockaddr_in));
IFA_UNLOCK(ifa);
err = in_control(NULL, SIOCDIFADDR, (caddr_t)&ifr, ifp,
kernproc);
/* if we lost the race, ignore it */
- if (err == EADDRNOTAVAIL)
+ if (err == EADDRNOTAVAIL) {
err = 0;
+ }
if (err != 0) {
char s_addr[MAX_IPv4_STR_LEN];
char s_dstaddr[MAX_IPv4_STR_LEN];
d = &((struct sockaddr_in *)
(void *)ifa->ifa_dstaddr)->sin_addr;
(void) inet_ntop(AF_INET, &s->s_addr, s_addr,
- sizeof (s_addr));
+ sizeof(s_addr));
(void) inet_ntop(AF_INET, &d->s_addr, s_dstaddr,
- sizeof (s_dstaddr));
+ sizeof(s_dstaddr));
IFA_UNLOCK(ifa);
- printf("%s: SIOCDIFADDR ifp=%p ifa_addr=%s "
- "ifa_dstaddr=%s (err=%d)\n", __func__, ifp,
- s_addr, s_dstaddr, err);
+ printf("%s: SIOCDIFADDR ifp=%s ifa_addr=%s "
+ "ifa_dstaddr=%s (err=%d)\n", __func__,
+ ifp->if_xname, s_addr, s_dstaddr, err);
}
}
ifnet_free_address_list(ifap);
} else if (err != 0 && err != ENXIO) {
printf("%s: error retrieving list of AF_INET addresses for "
- "ifp=%p (err=%d)\n", __func__, ifp, err);
+ "ifp=%s (err=%d)\n", __func__, ifp->if_xname, err);
}
}
-int inet_aton(char *cp, struct in_addr *pin);
-int
-inet_aton(char * cp, struct in_addr * pin)
-{
- u_char * b = (unsigned char *)pin;
- int i;
- char * p;
-
- for (p = cp, i = 0; i < 4; i++) {
- u_int32_t l = strtoul(p, 0, 0);
- if (l > 255)
- return (FALSE);
- b[i] = l;
- p = strchr(p, '.');
- if (i < 3 && p == NULL)
- return (FALSE);
- p++;
- }
- return (TRUE);
-}
-
-int inet_ntoa2(struct in_addr * pin, char * cp, const int len);
-int inet_ntoa2(struct in_addr * pin, char * cp, const int len)
-{
- int ret;
-
- /* address is in network byte order */
- ret = snprintf(cp, len, "%u.%u.%u.%u", pin->s_addr & 0xFF,
- (pin->s_addr >> 8) & 0xFF, (pin->s_addr >> 16) & 0xFF,
- (pin->s_addr >> 24) & 0xFF);
-
- return ret < len ? TRUE : FALSE;
-}
-
/*
* Called as part of ip_init
*/
{
in_multi_init();
- PE_parse_boot_argn("ifa_debug", &inifa_debug, sizeof (inifa_debug));
+ PE_parse_boot_argn("ifa_debug", &inifa_debug, sizeof(inifa_debug));
- inifa_size = (inifa_debug == 0) ? sizeof (struct in_ifaddr) :
- sizeof (struct in_ifaddr_dbg);
+ inifa_size = (inifa_debug == 0) ? sizeof(struct in_ifaddr) :
+ sizeof(struct in_ifaddr_dbg);
inifa_zone = zinit(inifa_size, INIFA_ZONE_MAX * inifa_size,
0, INIFA_ZONE_NAME);
bzero(inifa, inifa_size);
inifa->ia_ifa.ifa_free = in_ifaddr_free;
inifa->ia_ifa.ifa_debug |= IFD_ALLOC;
+ inifa->ia_ifa.ifa_del_wc = &inifa->ia_ifa.ifa_debug;
+ inifa->ia_ifa.ifa_del_waiters = 0;
ifa_lock_init(&inifa->ia_ifa);
if (inifa_debug != 0) {
struct in_ifaddr_dbg *inifa_dbg =
ctrace_record(&inifa_dbg->inifa_alloc);
}
}
- return (inifa);
+ return inifa;
}
static void
if (ifa->ifa_refcnt != 0) {
panic("%s: ifa %p bad ref cnt", __func__, ifa);
/* NOTREACHED */
- } if (!(ifa->ifa_debug & IFD_ALLOC)) {
+ }
+ if (!(ifa->ifa_debug & IFD_ALLOC)) {
panic("%s: ifa %p cannot be freed", __func__, ifa);
/* NOTREACHED */
}
struct in_ifaddr_dbg *inifa_dbg = (struct in_ifaddr_dbg *)ifa;
ctrace_record(&inifa_dbg->inifa_free);
bcopy(&inifa_dbg->inifa, &inifa_dbg->inifa_old,
- sizeof (struct in_ifaddr));
+ sizeof(struct in_ifaddr));
if (ifa->ifa_debug & IFD_TRASHED) {
/* Become a regular mutex, just in case */
IFA_CONVERT_LOCK(ifa);
}
IFA_UNLOCK(ifa);
ifa_lock_destroy(ifa);
- bzero(ifa, sizeof (struct in_ifaddr));
+ bzero(ifa, sizeof(struct in_ifaddr));
zfree(inifa_zone, ifa);
}
idx = atomic_add_16_ov(cnt, 1) % INIFA_TRACE_HIST_SIZE;
ctrace_record(&tr[idx]);
}
+
+/*
+ * Handle SIOCGASSOCIDS ioctl for PF_INET domain.
+ */
+static int
+in_getassocids(struct socket *so, uint32_t *cnt, user_addr_t aidp)
+{
+ struct inpcb *inp = sotoinpcb(so);
+ sae_associd_t aid;
+
+ if (inp == NULL || inp->inp_state == INPCB_STATE_DEAD) {
+ return EINVAL;
+ }
+
+ /* INPCB has no concept of association */
+ aid = SAE_ASSOCID_ANY;
+ *cnt = 0;
+
+ /* just asking how many there are? */
+ if (aidp == USER_ADDR_NULL) {
+ return 0;
+ }
+
+ return copyout(&aid, aidp, sizeof(aid));
+}
+
+/*
+ * Handle SIOCGCONNIDS ioctl for PF_INET domain.
+ */
+static int
+in_getconnids(struct socket *so, sae_associd_t aid, uint32_t *cnt,
+ user_addr_t cidp)
+{
+ struct inpcb *inp = sotoinpcb(so);
+ sae_connid_t cid;
+
+ if (inp == NULL || inp->inp_state == INPCB_STATE_DEAD) {
+ return EINVAL;
+ }
+
+ if (aid != SAE_ASSOCID_ANY && aid != SAE_ASSOCID_ALL) {
+ return EINVAL;
+ }
+
+ /* if connected, return 1 connection count */
+ *cnt = ((so->so_state & SS_ISCONNECTED) ? 1 : 0);
+
+ /* just asking how many there are? */
+ if (cidp == USER_ADDR_NULL) {
+ return 0;
+ }
+
+ /* if INPCB is connected, assign it connid 1 */
+ cid = ((*cnt != 0) ? 1 : SAE_CONNID_ANY);
+
+ return copyout(&cid, cidp, sizeof(cid));
+}
+
+/*
+ * Handle SIOCGCONNINFO ioctl for PF_INET domain.
+ */
+int
+in_getconninfo(struct socket *so, sae_connid_t cid, uint32_t *flags,
+ uint32_t *ifindex, int32_t *soerror, user_addr_t src, socklen_t *src_len,
+ user_addr_t dst, socklen_t *dst_len, uint32_t *aux_type,
+ user_addr_t aux_data, uint32_t *aux_len)
+{
+ struct inpcb *inp = sotoinpcb(so);
+ struct sockaddr_in sin;
+ struct ifnet *ifp = NULL;
+ int error = 0;
+ u_int32_t copy_len = 0;
+
+ /*
+ * Don't test for INPCB_STATE_DEAD since this may be called
+ * after SOF_PCBCLEARING is set, e.g. after tcp_close().
+ */
+ if (inp == NULL) {
+ error = EINVAL;
+ goto out;
+ }
+
+ if (cid != SAE_CONNID_ANY && cid != SAE_CONNID_ALL && cid != 1) {
+ error = EINVAL;
+ goto out;
+ }
+
+ ifp = inp->inp_last_outifp;
+ *ifindex = ((ifp != NULL) ? ifp->if_index : 0);
+ *soerror = so->so_error;
+ *flags = 0;
+ if (so->so_state & SS_ISCONNECTED) {
+ *flags |= (CIF_CONNECTED | CIF_PREFERRED);
+ }
+ if (inp->inp_flags & INP_BOUND_IF) {
+ *flags |= CIF_BOUND_IF;
+ }
+ if (!(inp->inp_flags & INP_INADDR_ANY)) {
+ *flags |= CIF_BOUND_IP;
+ }
+ if (!(inp->inp_flags & INP_ANONPORT)) {
+ *flags |= CIF_BOUND_PORT;
+ }
+
+ bzero(&sin, sizeof(sin));
+ sin.sin_len = sizeof(sin);
+ sin.sin_family = AF_INET;
+
+ /* source address and port */
+ sin.sin_port = inp->inp_lport;
+ sin.sin_addr.s_addr = inp->inp_laddr.s_addr;
+ if (*src_len == 0) {
+ *src_len = sin.sin_len;
+ } else {
+ if (src != USER_ADDR_NULL) {
+ copy_len = min(*src_len, sizeof(sin));
+ error = copyout(&sin, src, copy_len);
+ if (error != 0) {
+ goto out;
+ }
+ *src_len = copy_len;
+ }
+ }
+
+ /* destination address and port */
+ sin.sin_port = inp->inp_fport;
+ sin.sin_addr.s_addr = inp->inp_faddr.s_addr;
+ if (*dst_len == 0) {
+ *dst_len = sin.sin_len;
+ } else {
+ if (dst != USER_ADDR_NULL) {
+ copy_len = min(*dst_len, sizeof(sin));
+ error = copyout(&sin, dst, copy_len);
+ if (error != 0) {
+ goto out;
+ }
+ *dst_len = copy_len;
+ }
+ }
+
+ if (SOCK_PROTO(so) == IPPROTO_TCP) {
+ struct conninfo_tcp tcp_ci;
+
+ *aux_type = CIAUX_TCP;
+ if (*aux_len == 0) {
+ *aux_len = sizeof(tcp_ci);
+ } else {
+ if (aux_data != USER_ADDR_NULL) {
+ copy_len = min(*aux_len, sizeof(tcp_ci));
+ bzero(&tcp_ci, sizeof(tcp_ci));
+ tcp_getconninfo(so, &tcp_ci);
+ error = copyout(&tcp_ci, aux_data, copy_len);
+ if (error != 0) {
+ goto out;
+ }
+ *aux_len = copy_len;
+ }
+ }
+ } else {
+ *aux_type = 0;
+ *aux_len = 0;
+ }
+
+out:
+ return error;
+}
+
+struct in_llentry {
+ struct llentry base;
+};
+
+#define IN_LLTBL_DEFAULT_HSIZE 32
+#define IN_LLTBL_HASH(k, h) \
+ ((((((((k) >> 8) ^ (k)) >> 8) ^ (k)) >> 8) ^ (k)) & ((h) - 1))
+
+/*
+ * Do actual deallocation of @lle.
+ */
+static void
+in_lltable_destroy_lle_unlocked(struct llentry *lle)
+{
+ LLE_LOCK_DESTROY(lle);
+ LLE_REQ_DESTROY(lle);
+ FREE(lle, M_LLTABLE);
+}
+
+/*
+ * Called by LLE_FREE_LOCKED when number of references
+ * drops to zero.
+ */
+static void
+in_lltable_destroy_lle(struct llentry *lle)
+{
+ LLE_WUNLOCK(lle);
+ in_lltable_destroy_lle_unlocked(lle);
+}
+
+static struct llentry *
+in_lltable_new(struct in_addr addr4, u_int flags)
+{
+#pragma unused(flags)
+ struct in_llentry *lle;
+
+ MALLOC(lle, struct in_llentry *, sizeof(struct in_llentry), M_LLTABLE, M_NOWAIT | M_ZERO);
+ if (lle == NULL) { /* NB: caller generates msg */
+ return NULL;
+ }
+
+ /*
+ * For IPv4 this will trigger "arpresolve" to generate
+ * an ARP request.
+ */
+ lle->base.la_expire = net_uptime(); /* mark expired */
+ lle->base.r_l3addr.addr4 = addr4;
+ lle->base.lle_refcnt = 1;
+ lle->base.lle_free = in_lltable_destroy_lle;
+
+ LLE_LOCK_INIT(&lle->base);
+ LLE_REQ_INIT(&lle->base);
+ //callout_init(&lle->base.lle_timer, 1);
+
+ return &lle->base;
+}
+
+#define IN_ARE_MASKED_ADDR_EQUAL(d, a, m) ( \
+ ((((d).s_addr ^ (a).s_addr) & (m).s_addr)) == 0 )
+
+static int
+in_lltable_match_prefix(const struct sockaddr *saddr,
+ const struct sockaddr *smask, u_int flags, struct llentry *lle)
+{
+ struct in_addr addr, mask, lle_addr;
+
+ addr = ((const struct sockaddr_in *)(const void *)saddr)->sin_addr;
+ mask = ((const struct sockaddr_in *)(const void *)smask)->sin_addr;
+ lle_addr.s_addr = ntohl(lle->r_l3addr.addr4.s_addr);
+
+ if (IN_ARE_MASKED_ADDR_EQUAL(lle_addr, addr, mask) == 0) {
+ return 0;
+ }
+
+ if (lle->la_flags & LLE_IFADDR) {
+ /*
+ * Delete LLE_IFADDR records IFF address & flag matches.
+ * Note that addr is the interface address within prefix
+ * being matched.
+ * Note also we should handle 'ifdown' cases without removing
+ * ifaddr macs.
+ */
+ if (addr.s_addr == lle_addr.s_addr && (flags & LLE_STATIC) != 0) {
+ return 1;
+ }
+ return 0;
+ }
+
+ /* flags & LLE_STATIC means deleting both dynamic and static entries */
+ if ((flags & LLE_STATIC) || !(lle->la_flags & LLE_STATIC)) {
+ return 1;
+ }
+
+ return 0;
+}
+
+static void
+in_lltable_free_entry(struct lltable *llt, struct llentry *lle)
+{
+ struct ifnet *ifp;
+ size_t pkts_dropped;
+
+ LLE_WLOCK_ASSERT(lle);
+ KASSERT(llt != NULL, ("lltable is NULL"));
+
+ /* Unlink entry from table if not already */
+ if ((lle->la_flags & LLE_LINKED) != 0) {
+ ifp = llt->llt_ifp;
+ IF_AFDATA_WLOCK_ASSERT(ifp, llt->llt_af);
+ lltable_unlink_entry(llt, lle);
+ }
+
+#if 0
+ /* cancel timer */
+ if (callout_stop(&lle->lle_timer) > 0) {
+ LLE_REMREF(lle);
+ }
+#endif
+ /* Drop hold queue */
+ pkts_dropped = llentry_free(lle);
+ arpstat.dropped += pkts_dropped;
+}
+
+
+static int
+in_lltable_rtcheck(struct ifnet *ifp, u_int flags, const struct sockaddr *l3addr)
+{
+#pragma unused(flags)
+ struct rtentry *rt;
+
+ KASSERT(l3addr->sa_family == AF_INET,
+ ("sin_family %d", l3addr->sa_family));
+
+ /* XXX rtalloc1 should take a const param */
+ rt = rtalloc1(__DECONST(struct sockaddr *, l3addr), 0, 0);
+ if (rt == NULL || (rt->rt_flags & RTF_GATEWAY) || rt->rt_ifp != ifp) {
+ log(LOG_INFO, "IPv4 address: \"%s\" is not on the network\n",
+ inet_ntoa(((const struct sockaddr_in *)(const void *)l3addr)->sin_addr));
+ if (rt != NULL) {
+ rtfree_locked(rt);
+ }
+ return EINVAL;
+ }
+ rtfree_locked(rt);
+ return 0;
+}
+
+static inline uint32_t
+in_lltable_hash_dst(const struct in_addr dst, uint32_t hsize)
+{
+ return IN_LLTBL_HASH(dst.s_addr, hsize);
+}
+
+static uint32_t
+in_lltable_hash(const struct llentry *lle, uint32_t hsize)
+{
+ return in_lltable_hash_dst(lle->r_l3addr.addr4, hsize);
+}
+
+
+static void
+in_lltable_fill_sa_entry(const struct llentry *lle, struct sockaddr *sa)
+{
+ struct sockaddr_in *sin;
+
+ sin = (struct sockaddr_in *)(void *)sa;
+ bzero(sin, sizeof(*sin));
+ sin->sin_family = AF_INET;
+ sin->sin_len = sizeof(*sin);
+ sin->sin_addr = lle->r_l3addr.addr4;
+}
+
+static inline struct llentry *
+in_lltable_find_dst(struct lltable *llt, struct in_addr dst)
+{
+ struct llentry *lle;
+ struct llentries *lleh;
+ u_int hashidx;
+
+ hashidx = in_lltable_hash_dst(dst, llt->llt_hsize);
+ lleh = &llt->lle_head[hashidx];
+ LIST_FOREACH(lle, lleh, lle_next) {
+ if (lle->la_flags & LLE_DELETED) {
+ continue;
+ }
+ if (lle->r_l3addr.addr4.s_addr == dst.s_addr) {
+ break;
+ }
+ }
+
+ return lle;
+}
+
+static void
+in_lltable_delete_entry(struct lltable *llt, struct llentry *lle)
+{
+#pragma unused(llt)
+ lle->la_flags |= LLE_DELETED;
+ //EVENTHANDLER_INVOKE(lle_event, lle, LLENTRY_DELETED);
+#ifdef DIAGNOSTIC
+ log(LOG_INFO, "ifaddr cache = %p is deleted\n", lle);
+#endif
+ llentry_free(lle);
+}
+
+static struct llentry *
+in_lltable_alloc(struct lltable *llt, u_int flags, const struct sockaddr *l3addr)
+{
+ const struct sockaddr_in *sin = (const struct sockaddr_in *) (const void *)l3addr;
+ struct ifnet *ifp = llt->llt_ifp;
+ struct llentry *lle;
+
+ KASSERT(l3addr->sa_family == AF_INET,
+ ("sin_family %d", l3addr->sa_family));
+
+ /*
+ * A route that covers the given address must have
+ * been installed 1st because we are doing a resolution,
+ * verify this.
+ */
+ if (!(flags & LLE_IFADDR) &&
+ in_lltable_rtcheck(ifp, flags, l3addr) != 0) {
+ return NULL;
+ }
+
+ lle = in_lltable_new(sin->sin_addr, flags);
+ if (lle == NULL) {
+ log(LOG_INFO, "lla_lookup: new lle malloc failed\n");
+ return NULL;
+ }
+ lle->la_flags = flags & ~LLE_CREATE;
+ if (flags & LLE_STATIC) {
+ lle->r_flags |= RLLE_VALID;
+ }
+ if ((flags & LLE_IFADDR) == LLE_IFADDR) {
+ lltable_set_entry_addr(ifp, lle, LLADDR(SDL(ifp->if_lladdr->ifa_addr)));
+ lle->la_flags |= LLE_STATIC;
+ lle->r_flags |= (RLLE_VALID | RLLE_IFADDR);
+ }
+ return lle;
+}
+
+/*
+ * Return NULL if not found or marked for deletion.
+ * If found return lle read locked.
+ */
+static struct llentry *
+in_lltable_lookup(struct lltable *llt, u_int flags, const struct sockaddr *l3addr)
+{
+ const struct sockaddr_in *sin = (const struct sockaddr_in *)(const void *)l3addr;
+ struct llentry *lle;
+
+ IF_AFDATA_WLOCK_ASSERT(llt->llt_ifp, llt->llt_af);
+
+ KASSERT(l3addr->sa_family == AF_INET,
+ ("sin_family %d", l3addr->sa_family));
+ lle = in_lltable_find_dst(llt, sin->sin_addr);
+
+ if (lle == NULL) {
+ return NULL;
+ }
+
+ KASSERT((flags & (LLE_UNLOCKED | LLE_EXCLUSIVE)) !=
+ (LLE_UNLOCKED | LLE_EXCLUSIVE), ("wrong lle request flags: 0x%X",
+ flags));
+
+ if (flags & LLE_UNLOCKED) {
+ return lle;
+ }
+
+ if (flags & LLE_EXCLUSIVE) {
+ LLE_WLOCK(lle);
+ } else {
+ LLE_RLOCK(lle);
+ }
+
+ return lle;
+}
+
+static int
+in_lltable_dump_entry(struct lltable *llt, struct llentry *lle,
+ struct sysctl_req *wr)
+{
+ struct ifnet *ifp = llt->llt_ifp;
+ /* XXX stack use */
+ struct {
+ struct rt_msghdr rtm;
+ struct sockaddr_in sin;
+ struct sockaddr_dl sdl;
+ } arpc;
+ struct sockaddr_dl *sdl;
+ int error;
+
+ bzero(&arpc, sizeof(arpc));
+ /* skip deleted entries */
+ if ((lle->la_flags & LLE_DELETED) == LLE_DELETED) {
+ return 0;
+ }
+ /* Skip if jailed and not a valid IP of the prison. */
+ lltable_fill_sa_entry(lle, (struct sockaddr *)&arpc.sin);
+ /*
+ * produce a msg made of:
+ * struct rt_msghdr;
+ * struct sockaddr_in; (IPv4)
+ * struct sockaddr_dl;
+ */
+ arpc.rtm.rtm_msglen = sizeof(arpc);
+ arpc.rtm.rtm_version = RTM_VERSION;
+ arpc.rtm.rtm_type = RTM_GET;
+ arpc.rtm.rtm_flags = RTF_UP;
+ arpc.rtm.rtm_addrs = RTA_DST | RTA_GATEWAY;
+
+ /* publish */
+ if (lle->la_flags & LLE_PUB) {
+ arpc.rtm.rtm_flags |= RTF_ANNOUNCE;
+ }
+
+ sdl = &arpc.sdl;
+ sdl->sdl_family = AF_LINK;
+ sdl->sdl_len = sizeof(*sdl);
+ sdl->sdl_index = ifp->if_index;
+ sdl->sdl_type = ifp->if_type;
+ if ((lle->la_flags & LLE_VALID) == LLE_VALID) {
+ sdl->sdl_alen = ifp->if_addrlen;
+ bcopy(&lle->ll_addr, LLADDR(sdl), ifp->if_addrlen);
+ } else {
+ sdl->sdl_alen = 0;
+ bzero(LLADDR(sdl), ifp->if_addrlen);
+ }
+
+ arpc.rtm.rtm_rmx.rmx_expire =
+ lle->la_flags & LLE_STATIC ? 0 : lle->la_expire;
+ arpc.rtm.rtm_flags |= (RTF_HOST | RTF_LLDATA);
+ if (lle->la_flags & LLE_STATIC) {
+ arpc.rtm.rtm_flags |= RTF_STATIC;
+ }
+ if (lle->la_flags & LLE_IFADDR) {
+ arpc.rtm.rtm_flags |= RTF_PINNED;
+ }
+ arpc.rtm.rtm_flags |= RTF_PINNED;
+ arpc.rtm.rtm_index = ifp->if_index;
+ error = SYSCTL_OUT(wr, &arpc, sizeof(arpc));
+
+ return error;
+}
+
+static struct lltable *
+in_lltattach(struct ifnet *ifp)
+{
+ struct lltable *llt;
+
+ llt = lltable_allocate_htbl(IN_LLTBL_DEFAULT_HSIZE);
+ llt->llt_af = AF_INET;
+ llt->llt_ifp = ifp;
+
+ llt->llt_lookup = in_lltable_lookup;
+ llt->llt_alloc_entry = in_lltable_alloc;
+ llt->llt_delete_entry = in_lltable_delete_entry;
+ llt->llt_dump_entry = in_lltable_dump_entry;
+ llt->llt_hash = in_lltable_hash;
+ llt->llt_fill_sa_entry = in_lltable_fill_sa_entry;
+ llt->llt_free_entry = in_lltable_free_entry;
+ llt->llt_match_prefix = in_lltable_match_prefix;
+ lltable_link(llt);
+
+ return llt;
+}
+
+struct in_ifaddr*
+inifa_ifpwithflag(struct ifnet * ifp, uint32_t flag)
+{
+ struct ifaddr *ifa;
+
+ ifnet_lock_shared(ifp);
+ TAILQ_FOREACH(ifa, &ifp->if_addrlist, ifa_link)
+ {
+ IFA_LOCK_SPIN(ifa);
+ if (ifa->ifa_addr->sa_family != AF_INET) {
+ IFA_UNLOCK(ifa);
+ continue;
+ }
+ if ((((struct in_ifaddr *)ifa)->ia_flags & flag) == flag) {
+ IFA_ADDREF_LOCKED(ifa);
+ IFA_UNLOCK(ifa);
+ break;
+ }
+ IFA_UNLOCK(ifa);
+ }
+ ifnet_lock_done(ifp);
+
+ return (struct in_ifaddr *)ifa;
+}
+
+struct in_ifaddr *
+inifa_ifpclatv4(struct ifnet * ifp)
+{
+ struct ifaddr *ifa;
+
+ ifnet_lock_shared(ifp);
+ TAILQ_FOREACH(ifa, &ifp->if_addrlist, ifa_link)
+ {
+ uint32_t addr = 0;
+ IFA_LOCK_SPIN(ifa);
+ if (ifa->ifa_addr->sa_family != AF_INET) {
+ IFA_UNLOCK(ifa);
+ continue;
+ }
+
+ addr = ntohl(SIN(ifa->ifa_addr)->sin_addr.s_addr);
+ if (!IN_LINKLOCAL(addr) &&
+ !IN_LOOPBACK(addr)) {
+ IFA_ADDREF_LOCKED(ifa);
+ IFA_UNLOCK(ifa);
+ break;
+ }
+ IFA_UNLOCK(ifa);
+ }
+ ifnet_lock_done(ifp);
+
+ return (struct in_ifaddr *)ifa;
+}