X-Git-Url: https://git.saurik.com/apple/xnu.git/blobdiff_plain/316670eb35587141e969394ae8537d66b9211e80..HEAD:/bsd/netinet6/ip6_input.c?ds=sidebyside diff --git a/bsd/netinet6/ip6_input.c b/bsd/netinet6/ip6_input.c index 9fcc992cd..e1b9091a5 100644 --- a/bsd/netinet6/ip6_input.c +++ b/bsd/netinet6/ip6_input.c @@ -1,8 +1,8 @@ /* - * Copyright (c) 2003-2012 Apple Inc. All rights reserved. + * Copyright (c) 2003-2020 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 @@ -11,10 +11,10 @@ * 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, @@ -22,11 +22,9 @@ * 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@ */ -/* $FreeBSD: src/sys/netinet6/ip6_input.c,v 1.11.2.10 2001/07/24 19:10:18 brooks Exp $ */ -/* $KAME: ip6_input.c,v 1.194 2001/05/27 13:28:35 itojun Exp $ */ /* * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. @@ -92,7 +90,6 @@ * @(#)ip_input.c 8.2 (Berkeley) 1/4/94 */ - #include #include #include @@ -109,9 +106,11 @@ #include #include #include -#include +#include +#include #include +#include #include #include @@ -121,14 +120,18 @@ #include #include #include +#include +#include #include #include #if INET #include #include -#endif /*INET*/ +#endif /* INET */ +#include #include +#include #include #include #include @@ -136,121 +139,165 @@ #include #include #include -#include +#include #if IPSEC #include -#if INET6 #include -#endif extern int ipsec_bypass; -#endif - -#include +#endif /* IPSEC */ #if DUMMYNET -#include #include #endif /* DUMMYNET */ -#include - -#include - /* we need it for NLOOP. */ #include "loop.h" -#include - #if PF #include #endif /* PF */ -extern struct domain inet6domain; -extern struct ip6protosw inet6sw[]; +#include -struct ip6protosw * ip6_protox[IPPROTO_MAX]; -static int ip6qmaxlen = IFQ_MAXLEN; +struct ip6protosw *ip6_protox[IPPROTO_MAX]; -static lck_grp_attr_t *in6_ifaddr_rwlock_grp_attr; -static lck_grp_t *in6_ifaddr_rwlock_grp; -static lck_attr_t *in6_ifaddr_rwlock_attr; +static lck_grp_attr_t *in6_ifaddr_rwlock_grp_attr; +static lck_grp_t *in6_ifaddr_rwlock_grp; +static lck_attr_t *in6_ifaddr_rwlock_attr; decl_lck_rw_data(, in6_ifaddr_rwlock); /* Protected by in6_ifaddr_rwlock */ -struct in6_ifaddr *in6_ifaddrs = NULL; +struct in6_ifaddrhead in6_ifaddrhead; +struct in6_ifaddrhashhead * in6_ifaddrhashtbl; +uint32_t in6_ifaddrhmask; -int ip6_forward_srcrt; /* XXX */ -int ip6_sourcecheck; /* XXX */ -int ip6_sourcecheck_interval; /* XXX */ -const int int6intrq_present = 1; +#define IN6ADDR_NHASH 61 +u_int32_t in6addr_nhash = 0; /* hash table size */ +u_int32_t in6addr_hashp = 0; /* next largest prime */ -int ip6_ours_check_algorithm; -#define IN6_IFSTAT_REQUIRE_ALIGNED_64(f) \ +#define IN6_IFSTAT_REQUIRE_ALIGNED_64(f) \ _CASSERT(!(offsetof(struct in6_ifstat, f) % sizeof (uint64_t))) -#define ICMP6_IFSTAT_REQUIRE_ALIGNED_64(f) \ +#define ICMP6_IFSTAT_REQUIRE_ALIGNED_64(f) \ _CASSERT(!(offsetof(struct icmp6_ifstat, f) % sizeof (uint64_t))) -#if IPFW2 -/* firewall hooks */ -ip6_fw_chk_t *ip6_fw_chk_ptr; -ip6_fw_ctl_t *ip6_fw_ctl_ptr; -int ip6_fw_enable = 1; -#endif - struct ip6stat ip6stat; -#ifdef __APPLE__ -struct ifqueue ip6intrq; -decl_lck_mtx_data(, ip6_init_mutex); decl_lck_mtx_data(, proxy6_lock); -decl_lck_mtx_data(, dad6_mutex_data); -decl_lck_mtx_data(, nd6_mutex_data); -decl_lck_mtx_data(, prefix6_mutex_data); -decl_lck_mtx_data(, scope6_mutex_data); -lck_mtx_t *dad6_mutex = &dad6_mutex_data; -lck_mtx_t *nd6_mutex = &nd6_mutex_data; -lck_mtx_t *prefix6_mutex = &prefix6_mutex_data; -lck_mtx_t *scope6_mutex = &scope6_mutex_data; +decl_lck_mtx_data(static, dad6_mutex_data); +decl_lck_mtx_data(static, nd6_mutex_data); +decl_lck_mtx_data(static, prefix6_mutex_data); +lck_mtx_t *dad6_mutex = &dad6_mutex_data; +lck_mtx_t *nd6_mutex = &nd6_mutex_data; +lck_mtx_t *prefix6_mutex = &prefix6_mutex_data; #ifdef ENABLE_ADDRSEL -decl_lck_mtx_data(, addrsel_mutex_data); -lck_mtx_t *addrsel_mutex = &addrsel_mutex_data; -#endif -decl_lck_rw_data(, in6_ifs_rwlock); -decl_lck_rw_data(, icmp6_ifs_rwlock); -lck_attr_t *ip6_mutex_attr; -lck_grp_t *ip6_mutex_grp; -lck_grp_attr_t *ip6_mutex_grp_attr; -extern lck_mtx_t *inet6_domain_mutex; +decl_lck_mtx_data(static, addrsel_mutex_data); +lck_mtx_t *addrsel_mutex = &addrsel_mutex_data; #endif +static lck_attr_t *ip6_mutex_attr; +static lck_grp_t *ip6_mutex_grp; +static lck_grp_attr_t *ip6_mutex_grp_attr; + extern int loopattach_done; extern void addrsel_policy_init(void); +static int sysctl_reset_ip6_input_stats SYSCTL_HANDLER_ARGS; +static int sysctl_ip6_input_measure_bins SYSCTL_HANDLER_ARGS; +static int sysctl_ip6_input_getperf SYSCTL_HANDLER_ARGS; static void ip6_init_delayed(void); -static struct ip6aux *ip6_setdstifaddr(struct mbuf *, struct in6_ifaddr *); - static int ip6_hopopts_input(u_int32_t *, u_int32_t *, struct mbuf **, int *); -#if PULLDOWN_TEST -static struct mbuf *ip6_pullexthdr(struct mbuf *, size_t, int); -#endif -#ifdef __APPLE__ -void gifattach(void); -void stfattach(void); -#endif +static void in6_ifaddrhashtbl_init(void); + +#if NSTF +extern void stfattach(void); +#endif /* NSTF */ SYSCTL_DECL(_net_inet6_ip6); -int ip6_doscopedroute = 1; -SYSCTL_INT(_net_inet6_ip6, OID_AUTO, scopedroute, CTLFLAG_RD | CTLFLAG_LOCKED, - &ip6_doscopedroute, 0, "Enable IPv6 scoped routing"); +static uint32_t ip6_adj_clear_hwcksum = 0; +SYSCTL_UINT(_net_inet6_ip6, OID_AUTO, adj_clear_hwcksum, + CTLFLAG_RW | CTLFLAG_LOCKED, &ip6_adj_clear_hwcksum, 0, + "Invalidate hwcksum info when adjusting length"); + +static uint32_t ip6_adj_partial_sum = 1; +SYSCTL_UINT(_net_inet6_ip6, OID_AUTO, adj_partial_sum, + CTLFLAG_RW | CTLFLAG_LOCKED, &ip6_adj_partial_sum, 0, + "Perform partial sum adjustment of trailing bytes at IP layer"); + +static int ip6_input_measure = 0; +SYSCTL_PROC(_net_inet6_ip6, OID_AUTO, input_perf, + CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_LOCKED, + &ip6_input_measure, 0, sysctl_reset_ip6_input_stats, "I", "Do time measurement"); + +static uint64_t ip6_input_measure_bins = 0; +SYSCTL_PROC(_net_inet6_ip6, OID_AUTO, input_perf_bins, + CTLTYPE_QUAD | CTLFLAG_RW | CTLFLAG_LOCKED, &ip6_input_measure_bins, 0, + sysctl_ip6_input_measure_bins, "I", + "bins for chaining performance data histogram"); + +static net_perf_t net_perf; +SYSCTL_PROC(_net_inet6_ip6, OID_AUTO, input_perf_data, + CTLTYPE_STRUCT | CTLFLAG_RD | CTLFLAG_LOCKED, + 0, 0, sysctl_ip6_input_getperf, "S,net_perf", + "IP6 input performance data (struct net_perf, net/net_perf.h)"); -int ip6_restrictrecvif = 1; -SYSCTL_INT(_net_inet6_ip6, OID_AUTO, restrictrecvif, - CTLFLAG_RW | CTLFLAG_LOCKED, &ip6_restrictrecvif, 0, - "Enable inbound interface restrictions"); +/* + * ip6_checkinterface controls the receive side of the models for multihoming + * that are discussed in RFC 1122. + * + * sysctl_ip6_checkinterface values are: + * IP6_CHECKINTERFACE_WEAK_ES: + * This corresponds to the Weak End-System model where incoming packets from + * any interface are accepted provided the destination address of the incoming packet + * is assigned to some interface. + * + * IP6_CHECKINTERFACE_HYBRID_ES: + * The Hybrid End-System model use the Strong End-System for tunnel interfaces + * (ipsec and utun) and the weak End-System model for other interfaces families. + * This prevents a rogue middle box to probe for signs of TCP connections + * that use the tunnel interface. + * + * IP6_CHECKINTERFACE_STRONG_ES: + * The Strong model model requires the packet arrived on an interface that + * is assigned the destination address of the packet. + * + * Since the routing table and transmit implementation do not implement the Strong ES model, + * setting this to a value different from IP6_CHECKINTERFACE_WEAK_ES may lead to unexpected results. + * + * When forwarding is enabled, the system reverts to the Weak ES model as a router + * is expected by design to receive packets from several interfaces to the same address. + */ +#define IP6_CHECKINTERFACE_WEAK_ES 0 +#define IP6_CHECKINTERFACE_HYBRID_ES 1 +#define IP6_CHECKINTERFACE_STRONG_ES 2 + +static int ip6_checkinterface = IP6_CHECKINTERFACE_HYBRID_ES; + +static int sysctl_ip6_checkinterface SYSCTL_HANDLER_ARGS; +SYSCTL_PROC(_net_inet6_ip6, OID_AUTO, check_interface, + CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_LOCKED, + 0, 0, sysctl_ip6_checkinterface, "I", "Verify packet arrives on correct interface"); + +#if (DEBUG || DEVELOPMENT) +#define IP6_CHECK_IFDEBUG 1 +#else +#define IP6_CHECK_IFDEBUG 0 +#endif /* (DEBUG || DEVELOPMENT) */ +static int ip6_checkinterface_debug = IP6_CHECK_IFDEBUG; +SYSCTL_INT(_net_inet6_ip6, OID_AUTO, checkinterface_debug, CTLFLAG_RW | CTLFLAG_LOCKED, + &ip6_checkinterface_debug, IP6_CHECK_IFDEBUG, ""); + +typedef enum ip6_check_if_result { + IP6_CHECK_IF_NONE = 0, + IP6_CHECK_IF_OURS = 1, + IP6_CHECK_IF_DROP = 2, + IP6_CHECK_IF_FORWARD = 3 +} ip6_check_if_result_t; + +static ip6_check_if_result_t ip6_input_check_interface(struct mbuf *, struct ip6_hdr *, struct ifnet *, struct route_in6 *rin6, struct ifnet **); /* * On platforms which require strict alignment (currently for anything but @@ -264,36 +311,47 @@ SYSCTL_INT(_net_inet6_ip6, OID_AUTO, restrictrecvif, * load/store operations on the fields in IPv6 headers. */ #if defined(__i386__) || defined(__x86_64__) -#define IP6_HDR_ALIGNMENT_FIXUP(_m, _ifp, _action) do { } while (0) +#define IP6_HDR_ALIGNMENT_FIXUP(_m, _ifp, _action) do { } while (0) #else /* !__i386__ && !__x86_64__ */ -#define IP6_HDR_ALIGNMENT_FIXUP(_m, _ifp, _action) do { \ - if (!IP6_HDR_ALIGNED_P(mtod(_m, caddr_t))) { \ - struct mbuf *_n; \ - struct ifnet *__ifp = (_ifp); \ - atomic_add_64(&(__ifp)->if_alignerrs, 1); \ - if (((_m)->m_flags & M_PKTHDR) && \ - (_m)->m_pkthdr.header != NULL) \ - (_m)->m_pkthdr.header = NULL; \ - _n = m_defrag_offset(_m, max_linkhdr, M_NOWAIT); \ - if (_n == NULL) { \ - ip6stat.ip6s_toosmall++; \ - m_freem(_m); \ - (_m) = NULL; \ - _action \ - } else { \ - VERIFY(_n != (_m)); \ - (_m) = _n; \ - } \ - } \ +#define IP6_HDR_ALIGNMENT_FIXUP(_m, _ifp, _action) do { \ + if (!IP6_HDR_ALIGNED_P(mtod(_m, caddr_t))) { \ + struct mbuf *_n; \ + struct ifnet *__ifp = (_ifp); \ + atomic_add_64(&(__ifp)->if_alignerrs, 1); \ + if (((_m)->m_flags & M_PKTHDR) && \ + (_m)->m_pkthdr.pkt_hdr != NULL) \ + (_m)->m_pkthdr.pkt_hdr = NULL; \ + _n = m_defrag_offset(_m, max_linkhdr, M_NOWAIT); \ + if (_n == NULL) { \ + ip6stat.ip6s_toosmall++; \ + m_freem(_m); \ + (_m) = NULL; \ + _action; \ + } else { \ + VERIFY(_n != (_m)); \ + (_m) = _n; \ + } \ + } \ } while (0) #endif /* !__i386__ && !__x86_64__ */ static void -ip6_proto_input( - __unused protocol_family_t protocol, - mbuf_t packet) +ip6_proto_input(protocol_family_t protocol, mbuf_t packet) { +#pragma unused(protocol) +#if INET + struct timeval start_tv; + if (ip6_input_measure) { + net_perf_start_time(&net_perf, &start_tv); + } +#endif /* INET */ ip6_input(packet); +#if INET + if (ip6_input_measure) { + net_perf_measure_time(&net_perf, &start_tv, 1); + net_perf_histogram(&net_perf, 1); + } +#endif /* INET */ } /* @@ -301,29 +359,63 @@ ip6_proto_input( * All protocols not implemented in kernel go to raw IP6 protocol handler. */ void -ip6_init() +ip6_init(struct ip6protosw *pp, struct domain *dp) { - struct ip6protosw *pr; - int i; + static int ip6_initialized = 0; + struct protosw *pr; struct timeval tv; + int i; + domain_unguard_t unguard; - PE_parse_boot_argn("net.inet6.ip6.scopedroute", &ip6_doscopedroute, - sizeof (ip6_doscopedroute)); + domain_proto_mtx_lock_assert_held(); + VERIFY((pp->pr_flags & (PR_INITIALIZED | PR_ATTACHED)) == PR_ATTACHED); -#if DIAGNOSTIC - if (sizeof(struct protosw) != sizeof(struct ip6protosw)) - panic("sizeof(protosw) != sizeof(ip6protosw)"); -#endif - pr = (struct ip6protosw *)pffindproto_locked(PF_INET6, IPPROTO_RAW, SOCK_RAW); - if (pr == 0) - panic("ip6_init"); - for (i = 0; i < IPPROTO_MAX; i++) - ip6_protox[i] = pr; - for (pr = (struct ip6protosw*)inet6domain.dom_protosw; pr; pr = pr->pr_next) { - if(!(pr->pr_domain)) continue; /* If uninitialized, skip */ - if (pr->pr_domain->dom_family == PF_INET6 && - pr->pr_protocol && pr->pr_protocol != IPPROTO_RAW) { - ip6_protox[pr->pr_protocol] = pr; + _CASSERT((sizeof(struct ip6_hdr) + + sizeof(struct icmp6_hdr)) <= _MHLEN); + + if (ip6_initialized) { + return; + } + ip6_initialized = 1; + + eventhandler_lists_ctxt_init(&in6_evhdlr_ctxt); + (void)EVENTHANDLER_REGISTER(&in6_evhdlr_ctxt, in6_event, + in6_eventhdlr_callback, eventhandler_entry_dummy_arg, + EVENTHANDLER_PRI_ANY); + + eventhandler_lists_ctxt_init(&in6_clat46_evhdlr_ctxt); + (void)EVENTHANDLER_REGISTER(&in6_clat46_evhdlr_ctxt, in6_clat46_event, + in6_clat46_eventhdlr_callback, eventhandler_entry_dummy_arg, + EVENTHANDLER_PRI_ANY); + + for (i = 0; i < IN6_EVENT_MAX; i++) { + VERIFY(in6_event2kev_array[i].in6_event_code == i); + } + + pr = pffindproto_locked(PF_INET6, IPPROTO_RAW, SOCK_RAW); + if (pr == NULL) { + panic("%s: Unable to find [PF_INET6,IPPROTO_RAW,SOCK_RAW]\n", + __func__); + /* NOTREACHED */ + } + + /* Initialize the entire ip6_protox[] array to IPPROTO_RAW. */ + for (i = 0; i < IPPROTO_MAX; i++) { + ip6_protox[i] = (struct ip6protosw *)pr; + } + /* + * Cycle through IP protocols and put them into the appropriate place + * in ip6_protox[], skipping protocols IPPROTO_{IP,RAW}. + */ + VERIFY(dp == inet6domain && dp->dom_family == PF_INET6); + TAILQ_FOREACH(pr, &dp->dom_protosw, pr_entry) { + VERIFY(pr->pr_domain == dp); + if (pr->pr_protocol != 0 && pr->pr_protocol != IPPROTO_RAW) { + /* Be careful to only index valid IP protocols. */ + if (pr->pr_protocol < IPPROTO_MAX) { + ip6_protox[pr->pr_protocol] = + (struct ip6protosw *)pr; + } } } @@ -335,21 +427,13 @@ ip6_init() lck_mtx_init(dad6_mutex, ip6_mutex_grp, ip6_mutex_attr); lck_mtx_init(nd6_mutex, ip6_mutex_grp, ip6_mutex_attr); lck_mtx_init(prefix6_mutex, ip6_mutex_grp, ip6_mutex_attr); - lck_mtx_init(scope6_mutex, ip6_mutex_grp, ip6_mutex_attr); + scope6_init(ip6_mutex_grp, ip6_mutex_attr); #ifdef ENABLE_ADDRSEL lck_mtx_init(addrsel_mutex, ip6_mutex_grp, ip6_mutex_attr); #endif lck_mtx_init(&proxy6_lock, ip6_mutex_grp, ip6_mutex_attr); - lck_mtx_init(&ip6_init_mutex, ip6_mutex_grp, ip6_mutex_attr); - - lck_rw_init(&in6_ifs_rwlock, ip6_mutex_grp, ip6_mutex_attr); - lck_rw_init(&icmp6_ifs_rwlock, ip6_mutex_grp, ip6_mutex_attr); - - inet6domain.dom_flags = DOM_REENTRANT; - - ip6intrq.ifq_maxlen = ip6qmaxlen; in6_ifaddr_rwlock_grp_attr = lck_grp_attr_alloc_init(); in6_ifaddr_rwlock_grp = lck_grp_alloc_init("in6_ifaddr_rwlock", @@ -358,6 +442,9 @@ ip6_init() lck_rw_init(&in6_ifaddr_rwlock, in6_ifaddr_rwlock_grp, in6_ifaddr_rwlock_attr); + TAILQ_INIT(&in6_ifaddrhead); + in6_ifaddrhashtbl_init(); + IN6_IFSTAT_REQUIRE_ALIGNED_64(ifs6_in_receive); IN6_IFSTAT_REQUIRE_ALIGNED_64(ifs6_in_hdrerr); IN6_IFSTAT_REQUIRE_ALIGNED_64(ifs6_in_toobig); @@ -377,7 +464,7 @@ ip6_init() IN6_IFSTAT_REQUIRE_ALIGNED_64(ifs6_reass_ok); IN6_IFSTAT_REQUIRE_ALIGNED_64(ifs6_reass_fail); IN6_IFSTAT_REQUIRE_ALIGNED_64(ifs6_in_mcast); - IN6_IFSTAT_REQUIRE_ALIGNED_64(ifs6_out_mcast); + IN6_IFSTAT_REQUIRE_ALIGNED_64(ifs6_out_mcast); ICMP6_IFSTAT_REQUIRE_ALIGNED_64(ifs6_in_msg); ICMP6_IFSTAT_REQUIRE_ALIGNED_64(ifs6_in_error); @@ -415,20 +502,27 @@ ip6_init() ICMP6_IFSTAT_REQUIRE_ALIGNED_64(ifs6_out_mldreport); ICMP6_IFSTAT_REQUIRE_ALIGNED_64(ifs6_out_mlddone); + getmicrotime(&tv); + ip6_desync_factor = + (RandomULong() ^ tv.tv_usec) % MAX_TEMP_DESYNC_FACTOR; + + PE_parse_boot_argn("ip6_checkinterface", &i, sizeof(i)); + switch (i) { + case IP6_CHECKINTERFACE_WEAK_ES: + case IP6_CHECKINTERFACE_HYBRID_ES: + case IP6_CHECKINTERFACE_STRONG_ES: + ip6_checkinterface = i; + break; + default: + break; + } + in6_ifaddr_init(); ip6_moptions_init(); nd6_init(); frag6_init(); - icmp6_init(); + icmp6_init(NULL, dp); addrsel_policy_init(); - /* - * in many cases, random() here does NOT return random number - * as initialization during bootstrap time occur in fixed order. - */ - microtime(&tv); - ip6_flow_seq = random() ^ tv.tv_usec; - microtime(&tv); - ip6_desync_factor = (random() ^ tv.tv_usec) % MAX_TEMP_DESYNC_FACTOR; /* * P2P interfaces often route the local address to the loopback @@ -437,32 +531,266 @@ ip6_init() */ net_init_add(ip6_init_delayed); - domain_proto_mtx_unlock(TRUE); - proto_register_input(PF_INET6, ip6_proto_input, NULL, 0); - domain_proto_mtx_lock(); + unguard = domain_unguard_deploy(); + i = proto_register_input(PF_INET6, ip6_proto_input, NULL, 0); + if (i != 0) { + panic("%s: failed to register PF_INET6 protocol: %d\n", + __func__, i); + /* NOTREACHED */ + } + domain_unguard_release(unguard); } static void ip6_init_delayed(void) { - (void) in6_ifattach(lo_ifp, NULL, NULL); - -#ifdef __APPLE__ - /* nd6_timer_init */ - timeout(nd6_timer, (caddr_t)0, hz); + (void) in6_ifattach_prelim(lo_ifp); /* timer for regeneranation of temporary addresses randomize ID */ - timeout(in6_tmpaddrtimer, (caddr_t)0, - (ip6_temp_preferred_lifetime - ip6_desync_factor - - ip6_temp_regen_advance) * hz); + timeout(in6_tmpaddrtimer, NULL, + (ip6_temp_preferred_lifetime - ip6_desync_factor - + ip6_temp_regen_advance) * hz); -#if NGIF - gifattach(); -#endif #if NSTF stfattach(); +#endif /* NSTF */ +} + +static void +ip6_input_adjust(struct mbuf *m, struct ip6_hdr *ip6, uint32_t plen, + struct ifnet *inifp) +{ + boolean_t adjust = TRUE; + uint32_t tot_len = sizeof(*ip6) + plen; + + ASSERT(m_pktlen(m) > tot_len); + + /* + * Invalidate hardware checksum info if ip6_adj_clear_hwcksum + * is set; useful to handle buggy drivers. Note that this + * should not be enabled by default, as we may get here due + * to link-layer padding. + */ + if (ip6_adj_clear_hwcksum && + (m->m_pkthdr.csum_flags & CSUM_DATA_VALID) && + !(inifp->if_flags & IFF_LOOPBACK) && + !(m->m_pkthdr.pkt_flags & PKTF_LOOP)) { + m->m_pkthdr.csum_flags &= ~CSUM_DATA_VALID; + m->m_pkthdr.csum_data = 0; + ip6stat.ip6s_adj_hwcsum_clr++; + } + + /* + * If partial checksum information is available, subtract + * out the partial sum of postpended extraneous bytes, and + * update the checksum metadata accordingly. By doing it + * here, the upper layer transport only needs to adjust any + * prepended extraneous bytes (else it will do both.) + */ + if (ip6_adj_partial_sum && + (m->m_pkthdr.csum_flags & (CSUM_DATA_VALID | CSUM_PARTIAL)) == + (CSUM_DATA_VALID | CSUM_PARTIAL)) { + m->m_pkthdr.csum_rx_val = m_adj_sum16(m, + m->m_pkthdr.csum_rx_start, m->m_pkthdr.csum_rx_start, + (tot_len - m->m_pkthdr.csum_rx_start), + m->m_pkthdr.csum_rx_val); + } else if ((m->m_pkthdr.csum_flags & + (CSUM_DATA_VALID | CSUM_PARTIAL)) == + (CSUM_DATA_VALID | CSUM_PARTIAL)) { + /* + * If packet has partial checksum info and we decided not + * to subtract the partial sum of postpended extraneous + * bytes here (not the default case), leave that work to + * be handled by the other layers. For now, only TCP, UDP + * layers are capable of dealing with this. For all other + * protocols (including fragments), trim and ditch the + * partial sum as those layers might not implement partial + * checksumming (or adjustment) at all. + */ + if (ip6->ip6_nxt == IPPROTO_TCP || + ip6->ip6_nxt == IPPROTO_UDP) { + adjust = FALSE; + } else { + m->m_pkthdr.csum_flags &= ~CSUM_DATA_VALID; + m->m_pkthdr.csum_data = 0; + ip6stat.ip6s_adj_hwcsum_clr++; + } + } + + if (adjust) { + ip6stat.ip6s_adj++; + if (m->m_len == m->m_pkthdr.len) { + m->m_len = tot_len; + m->m_pkthdr.len = tot_len; + } else { + m_adj(m, tot_len - m->m_pkthdr.len); + } + } +} +static ip6_check_if_result_t +ip6_input_check_interface(struct mbuf *m, struct ip6_hdr *ip6, struct ifnet *inifp, struct route_in6 *rin6, struct ifnet **deliverifp) +{ + struct in6_ifaddr *ia6 = NULL; + struct in6_addr tmp_dst = ip6->ip6_dst; /* copy to avoid unaligned access */ + struct in6_ifaddr *best_ia6 = NULL; + ip6_check_if_result_t result = IP6_CHECK_IF_NONE; + + *deliverifp = NULL; + + /* + * Check for exact addresses in the hash bucket. + */ + lck_rw_lock_shared(&in6_ifaddr_rwlock); + TAILQ_FOREACH(ia6, IN6ADDR_HASH(&tmp_dst), ia6_hash) { + /* + * TODO: should we accept loopbacl + */ + if (IN6_ARE_ADDR_EQUAL(&ia6->ia_addr.sin6_addr, &tmp_dst)) { + if ((ia6->ia6_flags & (IN6_IFF_NOTREADY | IN6_IFF_CLAT46))) { + continue; + } + best_ia6 = ia6; + if (ia6->ia_ifp == inifp) { + /* + * TODO: should we also accept locally originated packets + * or from loopback ??? + */ + break; + } + /* + * Continue the loop in case there's a exact match with another + * interface + */ + } + } + if (best_ia6 != NULL) { + if (best_ia6->ia_ifp != inifp && ip6_forwarding == 0 && + ((ip6_checkinterface == IP6_CHECKINTERFACE_HYBRID_ES && + (best_ia6->ia_ifp->if_family == IFNET_FAMILY_IPSEC || + best_ia6->ia_ifp->if_family == IFNET_FAMILY_UTUN)) || + ip6_checkinterface == IP6_CHECKINTERFACE_STRONG_ES)) { + /* + * Drop when interface address check is strict and forwarding + * is disabled + */ + result = IP6_CHECK_IF_DROP; + } else { + result = IP6_CHECK_IF_OURS; + *deliverifp = best_ia6->ia_ifp; + ip6_setdstifaddr_info(m, 0, best_ia6); + } + } + lck_rw_done(&in6_ifaddr_rwlock); + + if (result == IP6_CHECK_IF_NONE) { + /* + * Slow path: route lookup. + */ + struct sockaddr_in6 *dst6; + + dst6 = SIN6(&rin6->ro_dst); + dst6->sin6_len = sizeof(struct sockaddr_in6); + dst6->sin6_family = AF_INET6; + dst6->sin6_addr = ip6->ip6_dst; + + rtalloc_scoped_ign((struct route *)rin6, + RTF_PRCLONING, IFSCOPE_NONE); + if (rin6->ro_rt != NULL) { + RT_LOCK_SPIN(rin6->ro_rt); + } + +#define rt6_key(r) (SIN6((r)->rt_nodes->rn_key)) + + /* + * Accept the packet if the forwarding interface to the destination + * according to the routing table is the loopback interface, + * unless the associated route has a gateway. + * Note that this approach causes to accept a packet if there is a + * route to the loopback interface for the destination of the packet. + * But we think it's even useful in some situations, e.g. when using + * a special daemon which wants to intercept the packet. + * + * XXX: some OSes automatically make a cloned route for the destination + * of an outgoing packet. If the outgoing interface of the packet + * is a loopback one, the kernel would consider the packet to be + * accepted, even if we have no such address assinged on the interface. + * We check the cloned flag of the route entry to reject such cases, + * assuming that route entries for our own addresses are not made by + * cloning (it should be true because in6_addloop explicitly installs + * the host route). However, we might have to do an explicit check + * while it would be less efficient. Or, should we rather install a + * reject route for such a case? + */ + if (rin6->ro_rt != NULL && + (rin6->ro_rt->rt_flags & (RTF_HOST | RTF_GATEWAY)) == RTF_HOST && +#if RTF_WASCLONED + !(rin6->ro_rt->rt_flags & RTF_WASCLONED) && #endif -#endif /* __APPLE__ */ + rin6->ro_rt->rt_ifp->if_type == IFT_LOOP) { + ia6 = (struct in6_ifaddr *)rin6->ro_rt->rt_ifa; + /* + * Packets to a tentative, duplicated, or somehow invalid + * address must not be accepted. + * + * For performance, test without acquiring the address lock; + * a lot of things in the address are set once and never + * changed (e.g. ia_ifp.) + */ + if (!(ia6->ia6_flags & IN6_IFF_NOTREADY)) { + /* this address is ready */ + result = IP6_CHECK_IF_OURS; + *deliverifp = ia6->ia_ifp; /* correct? */ + /* + * record dst address information into mbuf. + */ + (void) ip6_setdstifaddr_info(m, 0, ia6); + } + } + + if (rin6->ro_rt != NULL) { + RT_UNLOCK(rin6->ro_rt); + } + } + + if (result == IP6_CHECK_IF_NONE) { + if (ip6_forwarding == 0) { + result = IP6_CHECK_IF_DROP; + } else { + result = IP6_CHECK_IF_FORWARD; + ip6_setdstifaddr_info(m, inifp->if_index, NULL); + } + } + + if (result == IP6_CHECK_IF_OURS && *deliverifp != inifp) { + ASSERT(*deliverifp != NULL); + ip6stat.ip6s_rcv_if_weak_match++; + + /* Logging is too noisy when forwarding is enabled */ + if (ip6_checkinterface_debug != IP6_CHECKINTERFACE_WEAK_ES && ip6_forwarding != 0) { + char src_str[MAX_IPv6_STR_LEN]; + char dst_str[MAX_IPv6_STR_LEN]; + + inet_ntop(AF_INET6, &ip6->ip6_src, src_str, sizeof(src_str)); + inet_ntop(AF_INET6, &ip6->ip6_dst, dst_str, sizeof(dst_str)); + os_log_info(OS_LOG_DEFAULT, + "%s: weak ES interface match to %s for packet from %s to %s proto %u received via %s", + __func__, (*deliverifp)->if_xname, src_str, dst_str, ip6->ip6_nxt, inifp->if_xname); + } + } else if (result == IP6_CHECK_IF_DROP) { + ip6stat.ip6s_rcv_if_no_match++; + if (ip6_checkinterface_debug > 0) { + char src_str[MAX_IPv6_STR_LEN]; + char dst_str[MAX_IPv6_STR_LEN]; + + inet_ntop(AF_INET6, &ip6->ip6_src, src_str, sizeof(src_str)); + inet_ntop(AF_INET6, &ip6->ip6_dst, dst_str, sizeof(dst_str)); + os_log(OS_LOG_DEFAULT, + "%s: no interface match for packet from %s to %s proto %u received via %s", + __func__, src_str, dst_str, ip6->ip6_nxt, inifp->if_xname); + } + } + + return result; } void @@ -473,41 +801,53 @@ ip6_input(struct mbuf *m) u_int32_t plen; u_int32_t rtalert = ~0; int nxt = 0, ours = 0; - struct ifnet *deliverifp = NULL; - ipfilter_t inject_ipfref = 0; - int seen; - struct in6_ifaddr *ia6 = NULL; - struct route_in6 ip6_forward_rt; - struct sockaddr_in6 *dst6; + struct ifnet *inifp, *deliverifp = NULL; + ipfilter_t inject_ipfref = NULL; + int seen = 1; #if DUMMYNET - struct m_tag *tag; - struct ip_fw_args args; - - bzero(&args, sizeof(struct ip_fw_args)); + struct m_tag *tag; + struct ip_fw_args args = {}; #endif /* DUMMYNET */ + struct route_in6 rin6 = {}; - bzero(&ip6_forward_rt, sizeof(ip6_forward_rt)); - - /* Check if the packet we received is valid after interface filter + /* + * Check if the packet we received is valid after interface filter * processing */ MBUF_INPUT_CHECK(m, m->m_pkthdr.rcvif); + inifp = m->m_pkthdr.rcvif; + VERIFY(inifp != NULL); /* Perform IP header alignment fixup, if needed */ - IP6_HDR_ALIGNMENT_FIXUP(m, m->m_pkthdr.rcvif, return;); + IP6_HDR_ALIGNMENT_FIXUP(m, inifp, return ); + + m->m_pkthdr.pkt_flags &= ~PKTF_FORWARDED; +#if IPSEC + /* + * should the inner packet be considered authentic? + * see comment in ah4_input(). + */ + m->m_flags &= ~M_AUTHIPHDR; + m->m_flags &= ~M_AUTHIPDGM; +#endif /* IPSEC */ + + /* + * make sure we don't have onion peering information into m_aux. + */ + ip6_delaux(m); #if DUMMYNET if ((tag = m_tag_locate(m, KERNEL_MODULE_TAG_ID, KERNEL_TAG_TYPE_DUMMYNET, NULL)) != NULL) { - struct dn_pkt_tag *dn_tag; + struct dn_pkt_tag *dn_tag; + + dn_tag = (struct dn_pkt_tag *)(tag + 1); - dn_tag = (struct dn_pkt_tag *)(tag+1); - args.fwa_pf_rule = dn_tag->dn_pf_rule; - + m_tag_delete(m, tag); } - + if (args.fwa_pf_rule) { ip6 = mtod(m, struct ip6_hdr *); /* In case PF got disabled */ @@ -516,84 +856,66 @@ ip6_input(struct mbuf *m) #endif /* DUMMYNET */ /* - * No need to proccess packet twice if we've - * already seen it + * No need to proccess packet twice if we've already seen it. */ inject_ipfref = ipf_get_inject_filter(m); - if (inject_ipfref != 0) { + if (inject_ipfref != NULL) { ip6 = mtod(m, struct ip6_hdr *); nxt = ip6->ip6_nxt; seen = 0; goto injectit; - } else + } else { seen = 1; - -#if IPSEC - /* - * should the inner packet be considered authentic? - * see comment in ah4_input(). - */ - if (m) { - m->m_flags &= ~M_AUTHIPHDR; - m->m_flags &= ~M_AUTHIPDGM; } -#endif - - /* - * make sure we don't have onion peering information into m_aux. - */ - ip6_delaux(m); /* * mbuf statistics */ if (m->m_flags & M_EXT) { - if (m->m_next) + if (m->m_next != NULL) { ip6stat.ip6s_mext2m++; - else + } else { ip6stat.ip6s_mext1++; + } } else { -#define M2MMAX (sizeof(ip6stat.ip6s_m2m)/sizeof(ip6stat.ip6s_m2m[0])) - if (m->m_next) { - if (m->m_flags & M_LOOP) { - ip6stat.ip6s_m2m[ifnet_index(lo_ifp)]++; /* XXX */ - } else if (m->m_pkthdr.rcvif->if_index < M2MMAX) - ip6stat.ip6s_m2m[m->m_pkthdr.rcvif->if_index]++; - else +#define M2MMAX (sizeof (ip6stat.ip6s_m2m) / sizeof (ip6stat.ip6s_m2m[0])) + if (m->m_next != NULL) { + if (m->m_pkthdr.pkt_flags & PKTF_LOOP) { + /* XXX */ + ip6stat.ip6s_m2m[ifnet_index(lo_ifp)]++; + } else if (inifp->if_index < M2MMAX) { + ip6stat.ip6s_m2m[inifp->if_index]++; + } else { ip6stat.ip6s_m2m[0]++; - } else + } + } else { ip6stat.ip6s_m1++; + } #undef M2MMAX } /* - * Drop the packet if IPv6 operation is disabled on the IF; - * accessing the flag is done without acquiring nd_ifinfo lock - * for performance reasons. + * Drop the packet if IPv6 operation is disabled on the interface. */ - lck_rw_lock_shared(nd_if_rwlock); - if (m->m_pkthdr.rcvif->if_index < nd_ifinfo_indexlim && - (nd_ifinfo[m->m_pkthdr.rcvif->if_index].flags & ND6_IFF_IFDISABLED)) { - lck_rw_done(nd_if_rwlock); + if (inifp->if_eflags & IFEF_IPV6_DISABLED) { goto bad; } - lck_rw_done(nd_if_rwlock); - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_receive); + in6_ifstat_inc_na(inifp, ifs6_in_receive); ip6stat.ip6s_total++; -#ifndef PULLDOWN_TEST /* * L2 bridge code and some other code can return mbuf chain * that does not conform to KAME requirement. too bad. * XXX: fails to join if interface MTU > MCLBYTES. jumbogram? */ - if (m && m->m_next != NULL && m->m_pkthdr.len < MCLBYTES) { + if (m->m_next != NULL && m->m_pkthdr.len < MCLBYTES) { struct mbuf *n; - MGETHDR(n, M_DONTWAIT, MT_HEADER); /* MAC-OK */ - if (n) + MGETHDR(n, M_DONTWAIT, MT_HEADER); /* MAC-OK */ + if (n) { M_COPY_PKTHDR(n, m); + } if (n && m->m_pkthdr.len > MHLEN) { MCLGET(n, M_DONTWAIT); if ((n->m_flags & M_EXT) == 0) { @@ -601,21 +923,18 @@ ip6_input(struct mbuf *m) n = NULL; } } - if (n == NULL) + if (n == NULL) { goto bad; + } m_copydata(m, 0, m->m_pkthdr.len, mtod(n, caddr_t)); n->m_len = m->m_pkthdr.len; m_freem(m); m = n; } - IP6_EXTHDR_CHECK(m, 0, sizeof(struct ip6_hdr), - {goto done;}); -#endif + IP6_EXTHDR_CHECK(m, 0, sizeof(struct ip6_hdr), { goto done; }); if (m->m_len < sizeof(struct ip6_hdr)) { - struct ifnet *inifp; - inifp = m->m_pkthdr.rcvif; if ((m = m_pullup(m, sizeof(struct ip6_hdr))) == 0) { ip6stat.ip6s_toosmall++; in6_ifstat_inc(inifp, ifs6_in_hdrerr); @@ -627,43 +946,32 @@ ip6_input(struct mbuf *m) if ((ip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION) { ip6stat.ip6s_badvers++; - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_hdrerr); + in6_ifstat_inc(inifp, ifs6_in_hdrerr); goto bad; } ip6stat.ip6s_nxthist[ip6->ip6_nxt]++; -#if IPFW2 - /* - * Check with the firewall... - */ - if (ip6_fw_enable && ip6_fw_chk_ptr) { - u_short port = 0; - /* If ipfw says divert, we have to just drop packet */ - /* use port as a dummy argument */ - if ((*ip6_fw_chk_ptr)(&ip6, NULL, &port, &m)) { - m_freem(m); - m = NULL; - } - if (!m) - goto done; - } -#endif - /* * Check against address spoofing/corruption. */ + if (!(m->m_pkthdr.pkt_flags & PKTF_LOOP) && + IN6_IS_ADDR_LOOPBACK(&ip6->ip6_src)) { + ip6stat.ip6s_badscope++; + in6_ifstat_inc(inifp, ifs6_in_addrerr); + goto bad; + } if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_src) || IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_dst)) { /* * XXX: "badscope" is not very suitable for a multicast source. */ ip6stat.ip6s_badscope++; - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr); + in6_ifstat_inc(inifp, ifs6_in_addrerr); goto bad; } if (IN6_IS_ADDR_MC_INTFACELOCAL(&ip6->ip6_dst) && - !(m->m_flags & M_LOOP)) { + !(m->m_pkthdr.pkt_flags & PKTF_LOOP)) { /* * In this case, the packet should come from the loopback * interface. However, we cannot just check the if_flags, @@ -671,7 +979,7 @@ ip6_input(struct mbuf *m) * as the outgoing/incoming interface. */ ip6stat.ip6s_badscope++; - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr); + in6_ifstat_inc(inifp, ifs6_in_addrerr); goto bad; } @@ -690,7 +998,7 @@ ip6_input(struct mbuf *m) if (IN6_IS_ADDR_V4MAPPED(&ip6->ip6_src) || IN6_IS_ADDR_V4MAPPED(&ip6->ip6_dst)) { ip6stat.ip6s_badscope++; - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr); + in6_ifstat_inc(inifp, ifs6_in_addrerr); goto bad; } #if 0 @@ -704,7 +1012,7 @@ ip6_input(struct mbuf *m) if (IN6_IS_ADDR_V4COMPAT(&ip6->ip6_src) || IN6_IS_ADDR_V4COMPAT(&ip6->ip6_dst)) { ip6stat.ip6s_badscope++; - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr); + in6_ifstat_inc(inifp, ifs6_in_addrerr); goto bad; } #endif @@ -722,8 +1030,7 @@ ip6_input(struct mbuf *m) */ if (nstat_collect) { struct rtentry *rte = - ifnet_cached_rtlookup_inet6(m->m_pkthdr.rcvif, - &ip6->ip6_src); + ifnet_cached_rtlookup_inet6(inifp, &ip6->ip6_src); if (rte != NULL) { nstat_route_rx(rte, 1, m->m_pkthdr.len, 0); rtfree(rte); @@ -732,19 +1039,20 @@ ip6_input(struct mbuf *m) #if DUMMYNET check_with_pf: -#endif +#endif /* DUMMYNET */ #if PF /* Invoke inbound packet filter */ if (PF_IS_ENABLED) { int error; #if DUMMYNET - error = pf_af_hook(m->m_pkthdr.rcvif, NULL, &m, AF_INET6, TRUE, &args); -#else - error = pf_af_hook(m->m_pkthdr.rcvif, NULL, &m, AF_INET6, TRUE, NULL); -#endif + error = pf_af_hook(inifp, NULL, &m, AF_INET6, TRUE, &args); +#else /* !DUMMYNET */ + error = pf_af_hook(inifp, NULL, &m, AF_INET6, TRUE, NULL); +#endif /* !DUMMYNET */ if (error != 0 || m == NULL) { if (m != NULL) { - panic("%s: unexpected packet %p\n", __func__, m); + panic("%s: unexpected packet %p\n", + __func__, m); /* NOTREACHED */ } /* Already freed by callee */ @@ -755,183 +1063,114 @@ check_with_pf: #endif /* PF */ /* drop packets if interface ID portion is already filled */ - if ((m->m_pkthdr.rcvif->if_flags & IFF_LOOPBACK) == 0) { + if (!(inifp->if_flags & IFF_LOOPBACK) && + !(m->m_pkthdr.pkt_flags & PKTF_LOOP)) { if (IN6_IS_SCOPE_LINKLOCAL(&ip6->ip6_src) && ip6->ip6_src.s6_addr16[1]) { ip6stat.ip6s_badscope++; goto bad; } - if (IN6_IS_SCOPE_LINKLOCAL(&ip6->ip6_dst) && + if (IN6_IS_SCOPE_EMBED(&ip6->ip6_dst) && ip6->ip6_dst.s6_addr16[1]) { ip6stat.ip6s_badscope++; goto bad; } } - if (IN6_IS_SCOPE_LINKLOCAL(&ip6->ip6_src)) - ip6->ip6_src.s6_addr16[1] - = htons(m->m_pkthdr.rcvif->if_index); - if (IN6_IS_SCOPE_LINKLOCAL(&ip6->ip6_dst)) - ip6->ip6_dst.s6_addr16[1] - = htons(m->m_pkthdr.rcvif->if_index); + if (m->m_pkthdr.pkt_flags & PKTF_IFAINFO) { + if (IN6_IS_SCOPE_LINKLOCAL(&ip6->ip6_src)) { + ip6->ip6_src.s6_addr16[1] = + htons(m->m_pkthdr.src_ifindex); + } + if (IN6_IS_SCOPE_EMBED(&ip6->ip6_dst)) { + ip6->ip6_dst.s6_addr16[1] = + htons(m->m_pkthdr.dst_ifindex); + } + } else { + if (IN6_IS_SCOPE_LINKLOCAL(&ip6->ip6_src)) { + ip6->ip6_src.s6_addr16[1] = htons(inifp->if_index); + } + if (IN6_IS_SCOPE_EMBED(&ip6->ip6_dst)) { + ip6->ip6_dst.s6_addr16[1] = htons(inifp->if_index); + } + } /* * Multicast check */ if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) { - struct in6_multi *in6m = NULL; - struct ifnet *ifp = m->m_pkthdr.rcvif; + struct in6_multi *in6m = NULL; - in6_ifstat_inc(ifp, ifs6_in_mcast); + in6_ifstat_inc_na(inifp, ifs6_in_mcast); /* * See if we belong to the destination multicast group on the * arrival interface. */ in6_multihead_lock_shared(); - IN6_LOOKUP_MULTI(&ip6->ip6_dst, ifp, in6m); + IN6_LOOKUP_MULTI(&ip6->ip6_dst, inifp, in6m); in6_multihead_lock_done(); if (in6m != NULL) { IN6M_REMREF(in6m); ours = 1; - } else if (!nd6_prproxy -#if MROUTING - && !ip6_mrouter -#endif - ) { + } else if (!nd6_prproxy) { ip6stat.ip6s_notmember++; ip6stat.ip6s_cantforward++; - in6_ifstat_inc(ifp, ifs6_in_discard); + in6_ifstat_inc(inifp, ifs6_in_discard); goto bad; } - deliverifp = ifp; - goto hbhcheck; - } - - /* - * Unicast check - */ - dst6 = (struct sockaddr_in6 *)&ip6_forward_rt.ro_dst; - dst6->sin6_len = sizeof(struct sockaddr_in6); - dst6->sin6_family = AF_INET6; - dst6->sin6_addr = ip6->ip6_dst; - - rtalloc_scoped_ign((struct route *)&ip6_forward_rt, - RTF_PRCLONING, IFSCOPE_NONE); - if (ip6_forward_rt.ro_rt != NULL) - RT_LOCK(ip6_forward_rt.ro_rt); - -#define rt6_key(r) ((struct sockaddr_in6 *)((r)->rt_nodes->rn_key)) - - /* - * Accept the packet if the forwarding interface to the destination - * according to the routing table is the loopback interface, - * unless the associated route has a gateway. - * Note that this approach causes to accept a packet if there is a - * route to the loopback interface for the destination of the packet. - * But we think it's even useful in some situations, e.g. when using - * a special daemon which wants to intercept the packet. - * - * XXX: some OSes automatically make a cloned route for the destination - * of an outgoing packet. If the outgoing interface of the packet - * is a loopback one, the kernel would consider the packet to be - * accepted, even if we have no such address assinged on the interface. - * We check the cloned flag of the route entry to reject such cases, - * assuming that route entries for our own addresses are not made by - * cloning (it should be true because in6_addloop explicitly installs - * the host route). However, we might have to do an explicit check - * while it would be less efficient. Or, should we rather install a - * reject route for such a case? - */ - if (ip6_forward_rt.ro_rt != NULL && - (ip6_forward_rt.ro_rt->rt_flags & - (RTF_HOST|RTF_GATEWAY)) == RTF_HOST && -#if RTF_WASCLONED - !(ip6_forward_rt.ro_rt->rt_flags & RTF_WASCLONED) && -#endif -#if 0 - /* - * The check below is redundant since the comparison of - * the destination and the key of the rtentry has - * already done through looking up the routing table. - */ - IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst, - &rt6_key(ip6_forward_rt.ro_rt)->sin6_addr) -#endif - ip6_forward_rt.ro_rt->rt_ifp->if_type == IFT_LOOP) { - ia6 = (struct in6_ifaddr *)ip6_forward_rt.ro_rt->rt_ifa; - + deliverifp = inifp; /* - * record address information into m_aux. + * record dst address information into mbuf, if we don't have one yet. + * note that we are unable to record it, if the address is not listed + * as our interface address (e.g. multicast addresses, etc.) */ - (void)ip6_setdstifaddr(m, ia6); - + if (deliverifp != NULL) { + struct in6_ifaddr *ia6 = NULL; + + ia6 = in6_ifawithifp(deliverifp, &ip6->ip6_dst); + if (ia6 != NULL) { + (void) ip6_setdstifaddr_info(m, 0, ia6); + IFA_REMREF(&ia6->ia_ifa); + } else { + (void) ip6_setdstifaddr_info(m, inifp->if_index, NULL); + } + } + goto hbhcheck; + } else { /* - * packets to a tentative, duplicated, or somehow invalid - * address must not be accepted. + * Unicast check */ - RT_CONVERT_LOCK(ip6_forward_rt.ro_rt); /* just in case */ - IFA_LOCK_SPIN(&ia6->ia_ifa); - if (!(ia6->ia6_flags & IN6_IFF_NOTREADY)) { - IFA_UNLOCK(&ia6->ia_ifa); - /* this address is ready */ + ip6_check_if_result_t check_if_result = IP6_CHECK_IF_NONE; + check_if_result = ip6_input_check_interface(m, ip6, inifp, &rin6, &deliverifp); + ASSERT(check_if_result != IP6_CHECK_IF_NONE); + if (check_if_result == IP6_CHECK_IF_OURS) { ours = 1; - deliverifp = ia6->ia_ifp; /* correct? */ - /* Count the packet in the ip address stats */ - - RT_UNLOCK(ip6_forward_rt.ro_rt); - ia6 = NULL; goto hbhcheck; + } else if (check_if_result == IP6_CHECK_IF_DROP) { + goto bad; } - IFA_UNLOCK(&ia6->ia_ifa); - RT_UNLOCK(ip6_forward_rt.ro_rt); - /* address is not ready, so discard the packet. */ - nd6log((LOG_INFO, - "ip6_input: packet to an unready address %s->%s\n", - ip6_sprintf(&ip6->ip6_src), - ip6_sprintf(&ip6->ip6_dst))); - ia6 = NULL; - goto bad; } - if (ip6_forward_rt.ro_rt != NULL) - RT_UNLOCK(ip6_forward_rt.ro_rt); - /* * Now there is no reason to process the packet if it's not our own * and we're not a router. */ if (!ip6_forwarding) { ip6stat.ip6s_cantforward++; - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_discard); - goto bad; - } - - hbhcheck: - /* - * record address information into m_aux, if we don't have one yet. - * note that we are unable to record it, if the address is not listed - * as our interface address (e.g. multicast addresses, etc.) - */ - if (deliverifp && (ia6 = ip6_getdstifaddr(m)) == NULL) { - ia6 = in6_ifawithifp(deliverifp, &ip6->ip6_dst); - if (ia6) { - if (!ip6_setdstifaddr(m, ia6)) { - /* - * XXX maybe we should drop the packet here, - * as we could not provide enough information - * to the upper layers. - */ - } - IFA_REMREF(&ia6->ia_ifa); - ia6 = NULL; + in6_ifstat_inc(inifp, ifs6_in_discard); + /* + * Raise a kernel event if the packet received on cellular + * interface is not intended for local host. + * For now limit it to ICMPv6 packets. + */ + if (inifp->if_type == IFT_CELLULAR && + ip6->ip6_nxt == IPPROTO_ICMPV6) { + in6_ifstat_inc(inifp, ifs6_cantfoward_icmp6); } + goto bad; } - if (ia6 != NULL) { - IFA_REMREF(&ia6->ia_ifa); - ia6 = NULL; - } - +hbhcheck: /* * Process Hop-by-Hop options header if it's contained. * m may be modified in ip6_hopopts_input(). @@ -941,11 +1180,17 @@ check_with_pf: if (ip6->ip6_nxt == IPPROTO_HOPOPTS) { struct ip6_hbh *hbh; + /* + * Mark the packet to imply that HBH option has been checked. + * This can only be true is the packet came in unfragmented + * or if the option is in the first fragment + */ + m->m_pkthdr.pkt_flags |= PKTF_HBH_CHKED; if (ip6_hopopts_input(&plen, &rtalert, &m, &off)) { -#if 0 /*touches NULL pointer*/ - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_discard); +#if 0 /* touches NULL pointer */ + in6_ifstat_inc(inifp, ifs6_in_discard); #endif - goto done; /* m have already been freed */ + goto done; /* m have already been freed */ } /* adjust pointer */ @@ -963,24 +1208,14 @@ check_with_pf: * (non-zero) payload length to the variable plen. */ ip6stat.ip6s_badoptions++; - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_discard); - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_hdrerr); - icmp6_error(m, ICMP6_PARAM_PROB, - ICMP6_PARAMPROB_HEADER, - (caddr_t)&ip6->ip6_plen - (caddr_t)ip6); + in6_ifstat_inc(inifp, ifs6_in_discard); + in6_ifstat_inc(inifp, ifs6_in_hdrerr); + icmp6_error(m, ICMP6_PARAM_PROB, ICMP6_PARAMPROB_HEADER, + (int)((caddr_t)&ip6->ip6_plen - (caddr_t)ip6)); goto done; } -#ifndef PULLDOWN_TEST /* ip6_hopopts_input() ensures that mbuf is contiguous */ hbh = (struct ip6_hbh *)(ip6 + 1); -#else - IP6_EXTHDR_GET(hbh, struct ip6_hbh *, m, sizeof(struct ip6_hdr), - sizeof(struct ip6_hbh)); - if (hbh == NULL) { - ip6stat.ip6s_tooshort++; - goto done; - } -#endif nxt = hbh->ip6h_nxt; /* @@ -1003,8 +1238,9 @@ check_with_pf: break; } } - } else + } else { nxt = ip6->ip6_nxt; + } /* * Check that the amount of data in the buffers @@ -1014,35 +1250,17 @@ check_with_pf: */ if (m->m_pkthdr.len - sizeof(struct ip6_hdr) < plen) { ip6stat.ip6s_tooshort++; - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_truncated); + in6_ifstat_inc(inifp, ifs6_in_truncated); goto bad; } if (m->m_pkthdr.len > sizeof(struct ip6_hdr) + plen) { - if (m->m_len == m->m_pkthdr.len) { - m->m_len = sizeof(struct ip6_hdr) + plen; - m->m_pkthdr.len = sizeof(struct ip6_hdr) + plen; - } else - m_adj(m, sizeof(struct ip6_hdr) + plen - m->m_pkthdr.len); + ip6_input_adjust(m, ip6, plen, inifp); } /* * Forward if desirable. */ if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) { - /* - * If we are acting as a multicast router, all - * incoming multicast packets are passed to the - * kernel-level multicast forwarding function. - * The packet is returned (relatively) intact; if - * ip6_mforward() returns a non-zero value, the packet - * must be discarded, else it may be accepted below. - */ -#if MROUTING - if (ip6_mrouter && ip6_mforward(ip6, m->m_pkthdr.rcvif, m)) { - ip6stat.ip6s_cantforward++; - goto bad; - } -#endif if (!ours && nd6_prproxy) { /* * If this isn't for us, this might be a Neighbor @@ -1053,10 +1271,11 @@ check_with_pf: */ ours = nd6_prproxy_isours(m, ip6, NULL, IFSCOPE_NONE); VERIFY(!ours || - (m->m_pkthdr.aux_flags & MAUXF_PROXY_DST)); + (m->m_pkthdr.pkt_flags & PKTF_PROXY_DST)); } - if (!ours) + if (!ours) { goto bad; + } } else if (!ours) { /* * The unicast forwarding function might return the packet @@ -1070,11 +1289,12 @@ check_with_pf: * proxied nodes on different links (src is link-local, dst * is target address.) */ - if ((m = ip6_forward(m, &ip6_forward_rt, 0)) == NULL) + if ((m = ip6_forward(m, &rin6, 0)) == NULL) { goto done; - VERIFY(ip6_forward_rt.ro_rt != NULL); - VERIFY(m->m_pkthdr.aux_flags & MAUXF_PROXY_DST); - deliverifp = ip6_forward_rt.ro_rt->rt_ifp; + } + VERIFY(rin6.ro_rt != NULL); + VERIFY(m->m_pkthdr.pkt_flags & PKTF_PROXY_DST); + deliverifp = rin6.ro_rt->rt_ifp; ours = 1; } @@ -1092,7 +1312,7 @@ check_with_pf: if (IN6_IS_ADDR_V4MAPPED(&ip6->ip6_src) || IN6_IS_ADDR_V4MAPPED(&ip6->ip6_dst)) { ip6stat.ip6s_badscope++; - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr); + in6_ifstat_inc(inifp, ifs6_in_addrerr); goto bad; } @@ -1100,7 +1320,7 @@ check_with_pf: * Tell launch routine the next header */ ip6stat.ip6s_delivered++; - in6_ifstat_inc(deliverifp, ifs6_in_deliver); + in6_ifstat_inc_na(deliverifp, ifs6_in_deliver); injectit: nest = 0; @@ -1110,12 +1330,26 @@ injectit: * we do it once for the outermost protocol, and we assume each * protocol handler wouldn't mess with the alignment afterwards. */ - IP6_HDR_ALIGNMENT_FIXUP(m, m->m_pkthdr.rcvif, return;); + IP6_HDR_ALIGNMENT_FIXUP(m, inifp, return ); while (nxt != IPPROTO_DONE) { struct ipfilter *filter; int (*pr_input)(struct mbuf **, int *, int); + /* + * This would imply either IPPROTO_HOPOPTS was not the first + * option or it did not come in the first fragment. + */ + if (nxt == IPPROTO_HOPOPTS && + (m->m_pkthdr.pkt_flags & PKTF_HBH_CHKED) == 0) { + /* + * This implies that HBH option was not contained + * in the first fragment + */ + ip6stat.ip6s_badoptions++; + goto bad; + } + if (ip6_hdrnestlimit && (++nest > ip6_hdrnestlimit)) { ip6stat.ip6s_toomanyhdr++; goto bad; @@ -1127,11 +1361,10 @@ injectit: */ if (m->m_pkthdr.len < off) { ip6stat.ip6s_tooshort++; - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_truncated); + in6_ifstat_inc(inifp, ifs6_in_truncated); goto bad; } - #if IPSEC /* * enforce IPsec policy checking if we are seeing last header. @@ -1143,26 +1376,27 @@ injectit: if (ipsec6_in_reject(m, NULL)) { IPSEC_STAT_INCREMENT(ipsec6stat.in_polvio); goto bad; - } + } } -#endif +#endif /* IPSEC */ /* * Call IP filter */ - if (!TAILQ_EMPTY(&ipv6_filters)) { + if (!TAILQ_EMPTY(&ipv6_filters) && !IFNET_IS_INTCOPROC(inifp)) { ipf_ref(); TAILQ_FOREACH(filter, &ipv6_filters, ipf_link) { if (seen == 0) { if ((struct ipfilter *)inject_ipfref == - filter) + filter) { seen = 1; + } } else if (filter->ipf_filter.ipf_input) { errno_t result; result = filter->ipf_filter.ipf_input( filter->ipf_filter.cookie, - (mbuf_t *)&m, off, nxt); + (mbuf_t *)&m, off, (uint8_t)nxt); if (result == EJUSTRETURN) { ipf_unref(); goto done; @@ -1177,8 +1411,8 @@ injectit: } DTRACE_IP6(receive, struct mbuf *, m, struct inpcb *, NULL, - struct ip6_hdr *, ip6, struct ifnet *, m->m_pkthdr.rcvif, - struct ip *, NULL, struct ip6_hdr *, ip6); + struct ip6_hdr *, ip6, struct ifnet *, inifp, + struct ip *, NULL, struct ip6_hdr *, ip6); if ((pr_input = ip6_protox[nxt]->pr_input) == NULL) { m_freem(m); @@ -1193,47 +1427,101 @@ injectit: } } done: - if (ip6_forward_rt.ro_rt != NULL) - rtfree(ip6_forward_rt.ro_rt); + ROUTE_RELEASE(&rin6); return; - bad: +bad: m_freem(m); goto done; } -/* - * set/grab in6_ifaddr correspond to IPv6 destination address. - * XXX backward compatibility wrapper - */ -static struct ip6aux * -ip6_setdstifaddr(struct mbuf *m, struct in6_ifaddr *ia6) +void +ip6_setsrcifaddr_info(struct mbuf *m, uint32_t src_idx, struct in6_ifaddr *ia6) +{ + VERIFY(m->m_flags & M_PKTHDR); + + /* + * If the source ifaddr is specified, pick up the information + * from there; otherwise just grab the passed-in ifindex as the + * caller may not have the ifaddr available. + */ + if (ia6 != NULL) { + m->m_pkthdr.pkt_flags |= PKTF_IFAINFO; + m->m_pkthdr.src_ifindex = ia6->ia_ifp->if_index; + + /* See IN6_IFF comments in in6_var.h */ + m->m_pkthdr.src_iff = (ia6->ia6_flags & 0xffff); + } else { + m->m_pkthdr.src_iff = 0; + m->m_pkthdr.src_ifindex = (uint16_t)src_idx; + if (src_idx != 0) { + m->m_pkthdr.pkt_flags |= PKTF_IFAINFO; + } + } +} + +void +ip6_setdstifaddr_info(struct mbuf *m, uint32_t dst_idx, struct in6_ifaddr *ia6) +{ + VERIFY(m->m_flags & M_PKTHDR); + + /* + * If the destination ifaddr is specified, pick up the information + * from there; otherwise just grab the passed-in ifindex as the + * caller may not have the ifaddr available. + */ + if (ia6 != NULL) { + m->m_pkthdr.pkt_flags |= PKTF_IFAINFO; + m->m_pkthdr.dst_ifindex = ia6->ia_ifp->if_index; + + /* See IN6_IFF comments in in6_var.h */ + m->m_pkthdr.dst_iff = (ia6->ia6_flags & 0xffff); + } else { + m->m_pkthdr.dst_iff = 0; + m->m_pkthdr.dst_ifindex = (uint16_t)dst_idx; + if (dst_idx != 0) { + m->m_pkthdr.pkt_flags |= PKTF_IFAINFO; + } + } +} + +int +ip6_getsrcifaddr_info(struct mbuf *m, uint32_t *src_idx, uint32_t *ia6f) { - struct ip6aux *n; + VERIFY(m->m_flags & M_PKTHDR); + + if (!(m->m_pkthdr.pkt_flags & PKTF_IFAINFO)) { + return -1; + } + + if (src_idx != NULL) { + *src_idx = m->m_pkthdr.src_ifindex; + } - n = ip6_addaux(m); - if (n != NULL) { - if (ia6 != NULL) - IFA_ADDREF(&ia6->ia_ifa); - if (n->ip6a_dstia6 != NULL) - IFA_REMREF(&n->ip6a_dstia6->ia_ifa); - n->ip6a_dstia6 = ia6; + if (ia6f != NULL) { + *ia6f = m->m_pkthdr.src_iff; } - return (struct ip6aux *)n; /* NULL if failed to set */ + + return 0; } -struct in6_ifaddr * -ip6_getdstifaddr(m) - struct mbuf *m; +int +ip6_getdstifaddr_info(struct mbuf *m, uint32_t *dst_idx, uint32_t *ia6f) { - struct ip6aux *n; + VERIFY(m->m_flags & M_PKTHDR); + + if (!(m->m_pkthdr.pkt_flags & PKTF_IFAINFO)) { + return -1; + } + + if (dst_idx != NULL) { + *dst_idx = m->m_pkthdr.dst_ifindex; + } - n = ip6_findaux(m); - if (n != NULL) { - if (n->ip6a_dstia6 != NULL) - IFA_ADDREF(&n->ip6a_dstia6->ia_ifa); - return (n->ip6a_dstia6); + if (ia6f != NULL) { + *ia6f = m->m_pkthdr.dst_iff; } - return (NULL); + + return 0; } /* @@ -1250,39 +1538,24 @@ ip6_hopopts_input(uint32_t *plenp, uint32_t *rtalertp, struct mbuf **mp, u_int8_t *opt; /* validation of the length of the header */ -#ifndef PULLDOWN_TEST - IP6_EXTHDR_CHECK(m, off, sizeof(*hbh), return -1); + IP6_EXTHDR_CHECK(m, off, sizeof(*hbh), return (-1)); hbh = (struct ip6_hbh *)(mtod(m, caddr_t) + off); hbhlen = (hbh->ip6h_len + 1) << 3; - IP6_EXTHDR_CHECK(m, off, hbhlen, return -1); + IP6_EXTHDR_CHECK(m, off, hbhlen, return (-1)); hbh = (struct ip6_hbh *)(mtod(m, caddr_t) + off); -#else - IP6_EXTHDR_GET(hbh, struct ip6_hbh *, m, - sizeof(struct ip6_hdr), sizeof(struct ip6_hbh)); - if (hbh == NULL) { - ip6stat.ip6s_tooshort++; - return -1; - } - hbhlen = (hbh->ip6h_len + 1) << 3; - IP6_EXTHDR_GET(hbh, struct ip6_hbh *, m, sizeof(struct ip6_hdr), - hbhlen); - if (hbh == NULL) { - ip6stat.ip6s_tooshort++; - return -1; - } -#endif off += hbhlen; hbhlen -= sizeof(struct ip6_hbh); opt = (u_int8_t *)hbh + sizeof(struct ip6_hbh); if (ip6_process_hopopts(m, (u_int8_t *)hbh + sizeof(struct ip6_hbh), - hbhlen, rtalertp, plenp) < 0) - return (-1); + hbhlen, rtalertp, plenp) < 0) { + return -1; + } *offp = off; *mp = m; - return (0); + return 0; } /* @@ -1296,12 +1569,8 @@ ip6_hopopts_input(uint32_t *plenp, uint32_t *rtalertp, struct mbuf **mp, * opthead + hbhlen is located in continuous memory region. */ int -ip6_process_hopopts(m, opthead, hbhlen, rtalertp, plenp) - struct mbuf *m; - u_int8_t *opthead; - int hbhlen; - u_int32_t *rtalertp; - u_int32_t *plenp; +ip6_process_hopopts(struct mbuf *m, u_int8_t *opthead, int hbhlen, + u_int32_t *rtalertp, u_int32_t *plenp) { struct ip6_hdr *ip6; int optlen = 0; @@ -1331,9 +1600,9 @@ ip6_process_hopopts(m, opthead, hbhlen, rtalertp, plenp) if (*(opt + 1) != IP6OPT_RTALERT_LEN - 2) { /* XXX stat */ icmp6_error(m, ICMP6_PARAM_PROB, - ICMP6_PARAMPROB_HEADER, - erroff + opt + 1 - opthead); - return(-1); + ICMP6_PARAMPROB_HEADER, + (int)(erroff + opt + 1 - opthead)); + return -1; } optlen = IP6OPT_RTALERT_LEN; bcopy((caddr_t)(opt + 2), (caddr_t)&rtalert_val, 2); @@ -1348,9 +1617,9 @@ ip6_process_hopopts(m, opthead, hbhlen, rtalertp, plenp) if (*(opt + 1) != IP6OPT_JUMBO_LEN - 2) { /* XXX stat */ icmp6_error(m, ICMP6_PARAM_PROB, - ICMP6_PARAMPROB_HEADER, - erroff + opt + 1 - opthead); - return(-1); + ICMP6_PARAMPROB_HEADER, + (int)(erroff + opt + 1 - opthead)); + return -1; } optlen = IP6OPT_JUMBO_LEN; @@ -1362,9 +1631,9 @@ ip6_process_hopopts(m, opthead, hbhlen, rtalertp, plenp) if (ip6->ip6_plen) { ip6stat.ip6s_badoptions++; icmp6_error(m, ICMP6_PARAM_PROB, - ICMP6_PARAMPROB_HEADER, - erroff + opt - opthead); - return(-1); + ICMP6_PARAMPROB_HEADER, + (int)(erroff + opt - opthead)); + return -1; } /* @@ -1386,9 +1655,9 @@ ip6_process_hopopts(m, opthead, hbhlen, rtalertp, plenp) if (*plenp != 0) { ip6stat.ip6s_badoptions++; icmp6_error(m, ICMP6_PARAM_PROB, - ICMP6_PARAMPROB_HEADER, - erroff + opt + 2 - opthead); - return(-1); + ICMP6_PARAMPROB_HEADER, + (int)(erroff + opt + 2 - opthead)); + return -1; } #endif @@ -1398,14 +1667,14 @@ ip6_process_hopopts(m, opthead, hbhlen, rtalertp, plenp) if (jumboplen <= IPV6_MAXPACKET) { ip6stat.ip6s_badoptions++; icmp6_error(m, ICMP6_PARAM_PROB, - ICMP6_PARAMPROB_HEADER, - erroff + opt + 2 - opthead); - return(-1); + ICMP6_PARAMPROB_HEADER, + (int)(erroff + opt + 2 - opthead)); + return -1; } *plenp = jumboplen; break; - default: /* unknown option */ + default: /* unknown option */ if (hbhlen < IP6OPT_MINLEN) { ip6stat.ip6s_toosmall++; goto bad; @@ -1413,18 +1682,18 @@ ip6_process_hopopts(m, opthead, hbhlen, rtalertp, plenp) optlen = ip6_unknown_opt(opt, m, erroff + opt - opthead); if (optlen == -1) { - return(-1); + return -1; } optlen += 2; break; } } - return(0); + return 0; - bad: +bad: m_freem(m); - return(-1); + return -1; } /* @@ -1434,34 +1703,38 @@ ip6_process_hopopts(m, opthead, hbhlen, rtalertp, plenp) * is not continuous in order to return an ICMPv6 error. */ int -ip6_unknown_opt(uint8_t *optp, struct mbuf *m, int off) +ip6_unknown_opt(uint8_t *optp, struct mbuf *m, size_t off) { struct ip6_hdr *ip6; switch (IP6OPT_TYPE(*optp)) { case IP6OPT_TYPE_SKIP: /* ignore the option */ - return((int)*(optp + 1)); - case IP6OPT_TYPE_DISCARD: /* silently discard */ + return (int)*(optp + 1); + + case IP6OPT_TYPE_DISCARD: /* silently discard */ m_freem(m); - return(-1); + return -1; + case IP6OPT_TYPE_FORCEICMP: /* send ICMP even if multicasted */ ip6stat.ip6s_badoptions++; - icmp6_error(m, ICMP6_PARAM_PROB, ICMP6_PARAMPROB_OPTION, off); - return(-1); + icmp6_error(m, ICMP6_PARAM_PROB, ICMP6_PARAMPROB_OPTION, (int)off); + return -1; + case IP6OPT_TYPE_ICMP: /* send ICMP if not multicasted */ ip6stat.ip6s_badoptions++; ip6 = mtod(m, struct ip6_hdr *); if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) || - (m->m_flags & (M_BCAST|M_MCAST))) + (m->m_flags & (M_BCAST | M_MCAST))) { m_freem(m); - else + } else { icmp6_error(m, ICMP6_PARAM_PROB, - ICMP6_PARAMPROB_OPTION, off); - return(-1); + ICMP6_PARAMPROB_OPTION, (int)off); + } + return -1; } - m_freem(m); /* XXX: NOTREACHED */ - return (-1); + m_freem(m); /* XXX: NOTREACHED */ + return -1; } /* @@ -1488,67 +1761,125 @@ ip6_savecontrol_v4(struct inpcb *inp, struct mbuf *m, struct mbuf **mp, if ((inp->inp_socket->so_options & SO_TIMESTAMP) != 0) { struct timeval tv; - microtime(&tv); - mp = sbcreatecontrol_mbuf((caddr_t) &tv, sizeof(tv), + getmicrotime(&tv); + mp = sbcreatecontrol_mbuf((caddr_t)&tv, sizeof(tv), SCM_TIMESTAMP, SOL_SOCKET, mp); - if (*mp == NULL) + if (*mp == NULL) { return NULL; + } } - if ((inp->inp_socket->so_options & SO_TIMESTAMP_MONOTONIC) != 0) { - uint64_t time; + if ((inp->inp_socket->so_options & SO_TIMESTAMP_MONOTONIC) != 0) { + uint64_t time; - time = mach_absolute_time(); - mp = sbcreatecontrol_mbuf((caddr_t) &time, sizeof(time), - SCM_TIMESTAMP_MONOTONIC, SOL_SOCKET, mp); + time = mach_absolute_time(); + mp = sbcreatecontrol_mbuf((caddr_t)&time, sizeof(time), + SCM_TIMESTAMP_MONOTONIC, SOL_SOCKET, mp); + if (*mp == NULL) { + return NULL; + } + } + if ((inp->inp_socket->so_options & SO_TIMESTAMP_CONTINUOUS) != 0) { + uint64_t time; - if (*mp == NULL) - return NULL; - } + time = mach_continuous_time(); + mp = sbcreatecontrol_mbuf((caddr_t)&time, sizeof(time), + SCM_TIMESTAMP_CONTINUOUS, SOL_SOCKET, mp); + if (*mp == NULL) { + return NULL; + } + } if ((inp->inp_socket->so_flags & SOF_RECV_TRAFFIC_CLASS) != 0) { int tc = m_get_traffic_class(m); - mp = sbcreatecontrol_mbuf((caddr_t) &tc, sizeof(tc), - SO_TRAFFIC_CLASS, SOL_SOCKET, mp); - if (*mp == NULL) - return (NULL); + mp = sbcreatecontrol_mbuf((caddr_t)&tc, sizeof(tc), + SO_TRAFFIC_CLASS, SOL_SOCKET, mp); + if (*mp == NULL) { + return NULL; + } } +#define IS2292(inp, x, y) (((inp)->inp_flags & IN6P_RFC2292) ? (x) : (y)) if ((ip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION) { - if (v4only != NULL) + if (v4only != NULL) { *v4only = 1; - return (mp); + } + + // Send ECN flags for v4-mapped addresses + if ((inp->inp_flags & IN6P_TCLASS) != 0) { + struct ip *ip_header = mtod(m, struct ip *); + + int tclass = (int)(ip_header->ip_tos); + mp = sbcreatecontrol_mbuf((caddr_t)&tclass, sizeof(tclass), + IPV6_TCLASS, IPPROTO_IPV6, mp); + if (*mp == NULL) { + return NULL; + } + } + + // Send IN6P_PKTINFO for v4-mapped address + if ((inp->inp_flags & IN6P_PKTINFO) != 0 +#if CONTENT_FILTER + /* Content Filter needs to see local address */ + || (inp->inp_socket->so_cfil_db != NULL) +#endif + ) { + struct in6_pktinfo pi6 = { + .ipi6_addr = IN6ADDR_V4MAPPED_INIT, + .ipi6_ifindex = (m && m->m_pkthdr.rcvif) ? m->m_pkthdr.rcvif->if_index : 0, + }; + + struct ip *ip_header = mtod(m, struct ip *); + bcopy(&ip_header->ip_dst, &pi6.ipi6_addr.s6_addr32[3], sizeof(struct in_addr)); + + mp = sbcreatecontrol_mbuf((caddr_t)&pi6, + sizeof(struct in6_pktinfo), + IS2292(inp, IPV6_2292PKTINFO, IPV6_PKTINFO), + IPPROTO_IPV6, mp); + if (*mp == NULL) { + return NULL; + } + } + return mp; } -#define IS2292(inp, x, y) (((inp)->inp_flags & IN6P_RFC2292) ? (x) : (y)) /* RFC 2292 sec. 5 */ - if ((inp->inp_flags & IN6P_PKTINFO) != 0) { + if ((inp->inp_flags & IN6P_PKTINFO) != 0 +#if CONTENT_FILTER + /* Content Filter needs to see local address */ + || (inp->inp_socket->so_cfil_db != NULL) +#endif + ) { struct in6_pktinfo pi6; bcopy(&ip6->ip6_dst, &pi6.ipi6_addr, sizeof(struct in6_addr)); - in6_clearscope(&pi6.ipi6_addr); /* XXX */ + in6_clearscope(&pi6.ipi6_addr); /* XXX */ pi6.ipi6_ifindex = (m && m->m_pkthdr.rcvif) ? m->m_pkthdr.rcvif->if_index : 0; - mp = sbcreatecontrol_mbuf((caddr_t) &pi6, + mp = sbcreatecontrol_mbuf((caddr_t)&pi6, sizeof(struct in6_pktinfo), - IS2292(inp, IPV6_2292PKTINFO, IPV6_PKTINFO), IPPROTO_IPV6, mp); - if (*mp == NULL) + IS2292(inp, IPV6_2292PKTINFO, IPV6_PKTINFO), + IPPROTO_IPV6, mp); + if (*mp == NULL) { return NULL; + } } if ((inp->inp_flags & IN6P_HOPLIMIT) != 0) { int hlim = ip6->ip6_hlim & 0xff; - mp = sbcreatecontrol_mbuf((caddr_t) &hlim, sizeof(int), + mp = sbcreatecontrol_mbuf((caddr_t)&hlim, sizeof(int), IS2292(inp, IPV6_2292HOPLIMIT, IPV6_HOPLIMIT), IPPROTO_IPV6, mp); - if (*mp == NULL) + if (*mp == NULL) { return NULL; + } } - if (v4only != NULL) + if (v4only != NULL) { *v4only = 0; - return (mp); + } + return mp; } int @@ -1560,12 +1891,14 @@ ip6_savecontrol(struct inpcb *in6p, struct mbuf *m, struct mbuf **mp) *mp = NULL; np = ip6_savecontrol_v4(in6p, m, mp, &v4only); - if (np == NULL) + if (np == NULL) { goto no_mbufs; + } mp = np; - if (v4only) - return(0); + if (v4only) { + return 0; + } if ((in6p->inp_flags & IN6P_TCLASS) != 0) { u_int32_t flowinfo; @@ -1575,10 +1908,11 @@ ip6_savecontrol(struct inpcb *in6p, struct mbuf *m, struct mbuf **mp) flowinfo >>= 20; tclass = flowinfo & 0xff; - mp = sbcreatecontrol_mbuf((caddr_t) &tclass, sizeof(tclass), + mp = sbcreatecontrol_mbuf((caddr_t)&tclass, sizeof(tclass), IPV6_TCLASS, IPPROTO_IPV6, mp); - if (*mp == NULL) + if (*mp == NULL) { goto no_mbufs; + } } /* @@ -1600,28 +1934,8 @@ ip6_savecontrol(struct inpcb *in6p, struct mbuf *m, struct mbuf **mp) if (ip6->ip6_nxt == IPPROTO_HOPOPTS) { struct ip6_hbh *hbh; int hbhlen = 0; -#if PULLDOWN_TEST - struct mbuf *ext; -#endif - -#ifndef PULLDOWN_TEST hbh = (struct ip6_hbh *)(ip6 + 1); hbhlen = (hbh->ip6h_len + 1) << 3; -#else - ext = ip6_pullexthdr(m, sizeof(struct ip6_hdr), - ip6->ip6_nxt); - if (ext == NULL) { - ip6stat.ip6s_tooshort++; - return(0); - } - hbh = mtod(ext, struct ip6_hbh *); - hbhlen = (hbh->ip6h_len + 1) << 3; - if (hbhlen != ext->m_len) { - m_freem(ext); - ip6stat.ip6s_tooshort++; - return(0); - } -#endif /* * XXX: We copy the whole header even if a @@ -1634,9 +1948,6 @@ ip6_savecontrol(struct inpcb *in6p, struct mbuf *m, struct mbuf **mp) IS2292(in6p, IPV6_2292HOPOPTS, IPV6_HOPOPTS), IPPROTO_IPV6, mp); -#if PULLDOWN_TEST - m_freem(ext); -#endif if (*mp == NULL) { goto no_mbufs; } @@ -1653,12 +1964,9 @@ ip6_savecontrol(struct inpcb *in6p, struct mbuf *m, struct mbuf **mp) * Note that the order of the headers remains in * the chain of ancillary data. */ - while (1) { /* is explicit loop prevention necessary? */ + while (1) { /* is explicit loop prevention necessary? */ struct ip6_ext *ip6e = NULL; int elen; -#if PULLDOWN_TEST - struct mbuf *ext = NULL; -#endif /* * if it is not an extension header, don't try to @@ -1674,61 +1982,41 @@ ip6_savecontrol(struct inpcb *in6p, struct mbuf *m, struct mbuf **mp) goto loopend; } -#ifndef PULLDOWN_TEST - if (off + sizeof(*ip6e) > m->m_len) + if (off + sizeof(*ip6e) > m->m_len) { goto loopend; + } ip6e = (struct ip6_ext *)(mtod(m, caddr_t) + off); - if (nxt == IPPROTO_AH) + if (nxt == IPPROTO_AH) { elen = (ip6e->ip6e_len + 2) << 2; - else + } else { elen = (ip6e->ip6e_len + 1) << 3; - if (off + elen > m->m_len) - goto loopend; -#else - ext = ip6_pullexthdr(m, off, nxt); - if (ext == NULL) { - ip6stat.ip6s_tooshort++; - return(0); } - ip6e = mtod(ext, struct ip6_ext *); - if (nxt == IPPROTO_AH) - elen = (ip6e->ip6e_len + 2) << 2; - else - elen = (ip6e->ip6e_len + 1) << 3; - if (elen != ext->m_len) { - m_freem(ext); - ip6stat.ip6s_tooshort++; - return(0); + if (off + elen > m->m_len) { + goto loopend; } -#endif switch (nxt) { case IPPROTO_DSTOPTS: - if (!(in6p->inp_flags & IN6P_DSTOPTS)) + if (!(in6p->inp_flags & IN6P_DSTOPTS)) { break; + } mp = sbcreatecontrol_mbuf((caddr_t)ip6e, elen, - IS2292(in6p, - IPV6_2292DSTOPTS, IPV6_DSTOPTS), - IPPROTO_IPV6, mp); - if (*mp == NULL) { -#if PULLDOWN_TEST - m_freem(ext); -#endif + IS2292(in6p, IPV6_2292DSTOPTS, + IPV6_DSTOPTS), IPPROTO_IPV6, mp); + if (*mp == NULL) { goto no_mbufs; } break; case IPPROTO_ROUTING: - if (!in6p->inp_flags & IN6P_RTHDR) + if (!(in6p->inp_flags & IN6P_RTHDR)) { break; + } mp = sbcreatecontrol_mbuf((caddr_t)ip6e, elen, IS2292(in6p, IPV6_2292RTHDR, IPV6_RTHDR), IPPROTO_IPV6, mp); if (*mp == NULL) { -#if PULLDOWN_TEST - m_freem(ext); -#endif goto no_mbufs; } break; @@ -1743,30 +2031,22 @@ ip6_savecontrol(struct inpcb *in6p, struct mbuf *m, struct mbuf **mp) * the code just in case (nxt overwritten or * other cases). */ -#if PULLDOWN_TEST - m_freem(ext); -#endif goto loopend; - } /* proceed with the next header. */ off += elen; nxt = ip6e->ip6e_nxt; ip6e = NULL; -#if PULLDOWN_TEST - m_freem(ext); - ext = NULL; -#endif } - loopend: +loopend: ; } - return(0); + return 0; no_mbufs: ip6stat.ip6s_pktdropcntrl++; /* XXX increment a stat to show the failure */ - return(ENOBUFS); + return ENOBUFS; } #undef IS2292 @@ -1779,89 +2059,48 @@ ip6_notify_pmtu(struct inpcb *in6p, struct sockaddr_in6 *dst, u_int32_t *mtu) so = in6p->inp_socket; - if (mtu == NULL) + if ((in6p->inp_flags & IN6P_MTU) == 0) { + return; + } + + if (mtu == NULL) { return; + } #ifdef DIAGNOSTIC - if (so == NULL) /* I believe this is impossible */ + if (so == NULL) { /* I believe this is impossible */ panic("ip6_notify_pmtu: socket is NULL"); + /* NOTREACHED */ + } #endif - bzero(&mtuctl, sizeof(mtuctl)); /* zero-clear for safety */ - mtuctl.ip6m_mtu = *mtu; - mtuctl.ip6m_addr = *dst; - if (sa6_recoverscope(&mtuctl.ip6m_addr, TRUE)) + if (IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_faddr) && + (so->so_proto == NULL || so->so_proto->pr_protocol == IPPROTO_TCP)) { return; + } - if ((m_mtu = sbcreatecontrol((caddr_t)&mtuctl, sizeof(mtuctl), - IPV6_PATHMTU, IPPROTO_IPV6)) == NULL) + if (!IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_faddr) && + !IN6_ARE_ADDR_EQUAL(&in6p->in6p_faddr, &dst->sin6_addr)) { return; - - if (sbappendaddr(&so->so_rcv, (struct sockaddr *)dst, NULL, m_mtu, NULL) - == 0) { - m_freem(m_mtu); - /* XXX: should count statistics */ - } else - sorwakeup(so); - - return; -} - -#if PULLDOWN_TEST -/* - * pull single extension header from mbuf chain. returns single mbuf that - * contains the result, or NULL on error. - */ -static struct mbuf * -ip6_pullexthdr(m, off, nxt) - struct mbuf *m; - size_t off; - int nxt; -{ - struct ip6_ext ip6e; - size_t elen; - struct mbuf *n; - -#if DIAGNOSTIC - switch (nxt) { - case IPPROTO_DSTOPTS: - case IPPROTO_ROUTING: - case IPPROTO_HOPOPTS: - case IPPROTO_AH: /* is it possible? */ - break; - default: - printf("ip6_pullexthdr: invalid nxt=%d\n", nxt); } -#endif - - m_copydata(m, off, sizeof(ip6e), (caddr_t)&ip6e); - if (nxt == IPPROTO_AH) - elen = (ip6e.ip6e_len + 2) << 2; - else - elen = (ip6e.ip6e_len + 1) << 3; - MGET(n, M_DONTWAIT, MT_DATA); - if (n && elen >= MLEN) { - MCLGET(n, M_DONTWAIT); - if ((n->m_flags & M_EXT) == 0) { - m_free(n); - n = NULL; - } + bzero(&mtuctl, sizeof(mtuctl)); /* zero-clear for safety */ + mtuctl.ip6m_mtu = *mtu; + mtuctl.ip6m_addr = *dst; + if (sa6_recoverscope(&mtuctl.ip6m_addr, TRUE)) { + return; } - if (!n) - return NULL; - n->m_len = 0; - if (elen >= M_TRAILINGSPACE(n)) { - m_free(n); - return NULL; + if ((m_mtu = sbcreatecontrol((caddr_t)&mtuctl, sizeof(mtuctl), + IPV6_PATHMTU, IPPROTO_IPV6)) == NULL) { + return; } - m_copydata(m, off, elen, mtod(n, caddr_t)); - n->m_len = elen; - return n; + if (sbappendaddr(&so->so_rcv, SA(dst), NULL, m_mtu, NULL) == 0) { + return; + } + sorwakeup(so); } -#endif /* * Get pointer to the previous header followed by the header @@ -1876,15 +2115,13 @@ ip6_pullexthdr(m, off, nxt) * we develop `neater' mechanism to process extension headers. */ char * -ip6_get_prevhdr(m, off) - struct mbuf *m; - int off; +ip6_get_prevhdr(struct mbuf *m, int off) { struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *); - if (off == sizeof(struct ip6_hdr)) - return((char *) &ip6->ip6_nxt); - else { + if (off == sizeof(struct ip6_hdr)) { + return (char *)&ip6->ip6_nxt; + } else { int len, nxt; struct ip6_ext *ip6e = NULL; @@ -1906,10 +2143,11 @@ ip6_get_prevhdr(m, off) } nxt = ip6e->ip6e_nxt; } - if (ip6e) - return((char *) &ip6e->ip6e_nxt); - else + if (ip6e) { + return (char *)&ip6e->ip6e_nxt; + } else { return NULL; + } } } @@ -1917,29 +2155,27 @@ ip6_get_prevhdr(m, off) * get next header offset. m will be retained. */ int -ip6_nexthdr(m, off, proto, nxtp) - struct mbuf *m; - int off; - int proto; - int *nxtp; +ip6_nexthdr(struct mbuf *m, int off, int proto, int *nxtp) { struct ip6_hdr ip6; struct ip6_ext ip6e; struct ip6_frag fh; /* just in case */ - if (m == NULL) - panic("ip6_nexthdr: m == NULL"); - if ((m->m_flags & M_PKTHDR) == 0 || m->m_pkthdr.len < off) + VERIFY(m != NULL); + if ((m->m_flags & M_PKTHDR) == 0 || m->m_pkthdr.len < off) { return -1; + } switch (proto) { case IPPROTO_IPV6: - if (m->m_pkthdr.len < off + sizeof(ip6)) + if (m->m_pkthdr.len < off + sizeof(ip6)) { return -1; + } m_copydata(m, off, sizeof(ip6), (caddr_t)&ip6); - if (nxtp) + if (nxtp) { *nxtp = ip6.ip6_nxt; + } off += sizeof(ip6); return off; @@ -1948,34 +2184,41 @@ ip6_nexthdr(m, off, proto, nxtp) * terminate parsing if it is not the first fragment, * it does not make sense to parse through it. */ - if (m->m_pkthdr.len < off + sizeof(fh)) + if (m->m_pkthdr.len < off + sizeof(fh)) { return -1; + } m_copydata(m, off, sizeof(fh), (caddr_t)&fh); /* IP6F_OFF_MASK = 0xfff8(BigEndian), 0xf8ff(LittleEndian) */ - if (fh.ip6f_offlg & IP6F_OFF_MASK) + if (fh.ip6f_offlg & IP6F_OFF_MASK) { return -1; - if (nxtp) + } + if (nxtp) { *nxtp = fh.ip6f_nxt; + } off += sizeof(struct ip6_frag); return off; case IPPROTO_AH: - if (m->m_pkthdr.len < off + sizeof(ip6e)) + if (m->m_pkthdr.len < off + sizeof(ip6e)) { return -1; + } m_copydata(m, off, sizeof(ip6e), (caddr_t)&ip6e); - if (nxtp) + if (nxtp) { *nxtp = ip6e.ip6e_nxt; + } off += (ip6e.ip6e_len + 2) << 2; return off; case IPPROTO_HOPOPTS: case IPPROTO_ROUTING: case IPPROTO_DSTOPTS: - if (m->m_pkthdr.len < off + sizeof(ip6e)) + if (m->m_pkthdr.len < off + sizeof(ip6e)) { return -1; + } m_copydata(m, off, sizeof(ip6e), (caddr_t)&ip6e); - if (nxtp) + if (nxtp) { *nxtp = ip6e.ip6e_nxt; + } off += (ip6e.ip6e_len + 1) << 3; return off; @@ -1988,19 +2231,13 @@ ip6_nexthdr(m, off, proto, nxtp) default: return -1; } - - return -1; } /* * get offset for the last header in the chain. m will be kept untainted. */ int -ip6_lasthdr(m, off, proto, nxtp) - struct mbuf *m; - int off; - int proto; - int *nxtp; +ip6_lasthdr(struct mbuf *m, int off, int proto, int *nxtp) { int newoff; int nxt; @@ -2011,22 +2248,61 @@ ip6_lasthdr(m, off, proto, nxtp) } while (1) { newoff = ip6_nexthdr(m, off, proto, nxtp); - if (newoff < 0) + if (newoff < 0) { return off; - else if (newoff < off) - return -1; /* invalid */ - else if (newoff == off) + } else if (newoff < off) { + return -1; /* invalid */ + } else if (newoff == off) { return newoff; + } off = newoff; proto = *nxtp; } } +boolean_t +ip6_pkt_has_ulp(struct mbuf *m) +{ + int off = 0, nxt = IPPROTO_NONE; + + off = ip6_lasthdr(m, 0, IPPROTO_IPV6, &nxt); + if (off < 0 || m->m_pkthdr.len < off) { + return FALSE; + } + + switch (nxt) { + case IPPROTO_TCP: + if (off + sizeof(struct tcphdr) > m->m_pkthdr.len) { + return FALSE; + } + break; + case IPPROTO_UDP: + if (off + sizeof(struct udphdr) > m->m_pkthdr.len) { + return FALSE; + } + break; + case IPPROTO_ICMPV6: + if (off + sizeof(uint32_t) > m->m_pkthdr.len) { + return FALSE; + } + break; + case IPPROTO_NONE: + return TRUE; + case IPPROTO_ESP: + return TRUE; + case IPPROTO_IPCOMP: + return TRUE; + default: + return FALSE; + } + return TRUE; +} + struct ip6aux * ip6_addaux(struct mbuf *m) { - struct m_tag *tag; + struct m_tag *tag; /* Check if one is already allocated */ tag = m_tag_locate(m, KERNEL_MODULE_TAG_ID, @@ -2034,7 +2310,7 @@ ip6_addaux(struct mbuf *m) if (tag == NULL) { /* Allocate a tag */ tag = m_tag_create(KERNEL_MODULE_TAG_ID, KERNEL_TAG_TYPE_INET6, - sizeof (struct ip6aux), M_DONTWAIT, m); + sizeof(struct ip6aux), M_DONTWAIT, m); /* Attach it to the mbuf */ if (tag) { @@ -2042,24 +2318,24 @@ ip6_addaux(struct mbuf *m) } } - return (tag ? (struct ip6aux *)(tag + 1) : NULL); + return tag ? (struct ip6aux *)(tag + 1) : NULL; } struct ip6aux * ip6_findaux(struct mbuf *m) { - struct m_tag *tag; + struct m_tag *tag; tag = m_tag_locate(m, KERNEL_MODULE_TAG_ID, KERNEL_TAG_TYPE_INET6, NULL); - return (tag ? (struct ip6aux *)(tag + 1) : NULL); + return tag ? (struct ip6aux *)(tag + 1) : NULL; } void ip6_delaux(struct mbuf *m) { - struct m_tag *tag; + struct m_tag *tag; tag = m_tag_locate(m, KERNEL_MODULE_TAG_ID, KERNEL_TAG_TYPE_INET6, NULL); @@ -2069,37 +2345,156 @@ ip6_delaux(struct mbuf *m) } /* - * Called by m_tag_free(). + * Drain callback */ void -ip6_destroyaux(struct ip6aux *n) +ip6_drain(void) { - if (n->ip6a_dstia6 != NULL) { - IFA_REMREF(&n->ip6a_dstia6->ia_ifa); - n->ip6a_dstia6 = NULL; - } + frag6_drain(); /* fragments */ + in6_rtqdrain(); /* protocol cloned routes */ + nd6_drain(NULL); /* cloned routes: ND6 */ } /* - * Called by m_tag_copy() + * System control for IP6 */ -void -ip6_copyaux(struct ip6aux *src, struct ip6aux *dst) + +u_char inet6ctlerrmap[PRC_NCMDS] = { + 0, 0, 0, 0, + 0, EMSGSIZE, EHOSTDOWN, EHOSTUNREACH, + EHOSTUNREACH, EHOSTUNREACH, ECONNREFUSED, ECONNREFUSED, + EMSGSIZE, EHOSTUNREACH, 0, 0, + 0, 0, 0, 0, + ENOPROTOOPT +}; + +static int +sysctl_reset_ip6_input_stats SYSCTL_HANDLER_ARGS +{ +#pragma unused(arg1, arg2) + int error, i; + + i = ip6_input_measure; + error = sysctl_handle_int(oidp, &i, 0, req); + if (error || req->newptr == USER_ADDR_NULL) { + goto done; + } + /* impose bounds */ + if (i < 0 || i > 1) { + error = EINVAL; + goto done; + } + if (ip6_input_measure != i && i == 1) { + net_perf_initialize(&net_perf, ip6_input_measure_bins); + } + ip6_input_measure = i; +done: + return error; +} + +static int +sysctl_ip6_input_measure_bins SYSCTL_HANDLER_ARGS +{ +#pragma unused(arg1, arg2) + int error; + uint64_t i; + + i = ip6_input_measure_bins; + error = sysctl_handle_quad(oidp, &i, 0, req); + if (error || req->newptr == USER_ADDR_NULL) { + goto done; + } + /* validate data */ + if (!net_perf_validate_bins(i)) { + error = EINVAL; + goto done; + } + ip6_input_measure_bins = i; +done: + return error; +} + +static int +sysctl_ip6_input_getperf SYSCTL_HANDLER_ARGS { - bcopy(src, dst, sizeof (*dst)); - if (dst->ip6a_dstia6 != NULL) - IFA_ADDREF(&dst->ip6a_dstia6->ia_ifa); +#pragma unused(oidp, arg1, arg2) + if (req->oldptr == USER_ADDR_NULL) { + req->oldlen = (size_t)sizeof(struct net_perf); + } + + return SYSCTL_OUT(req, &net_perf, MIN(sizeof(net_perf), req->oldlen)); } + /* - * System control for IP6 + * Initialize IPv6 source address hash table. */ +static void +in6_ifaddrhashtbl_init(void) +{ + int i, k, p; -u_char inet6ctlerrmap[PRC_NCMDS] = { - 0, 0, 0, 0, - 0, EMSGSIZE, EHOSTDOWN, EHOSTUNREACH, - EHOSTUNREACH, EHOSTUNREACH, ECONNREFUSED, ECONNREFUSED, - EMSGSIZE, EHOSTUNREACH, 0, 0, - 0, 0, 0, 0, - ENOPROTOOPT -}; + if (in6_ifaddrhashtbl != NULL) { + return; + } + + PE_parse_boot_argn("ina6ddr_nhash", &in6addr_nhash, + sizeof(in6addr_nhash)); + if (in6addr_nhash == 0) { + in6addr_nhash = IN6ADDR_NHASH; + } + + MALLOC(in6_ifaddrhashtbl, struct in6_ifaddrhashhead *, + in6addr_nhash * sizeof(*in6_ifaddrhashtbl), + M_IFADDR, M_WAITOK | M_ZERO); + if (in6_ifaddrhashtbl == NULL) { + panic("in6_ifaddrhashtbl allocation failed"); + } + + /* + * Generate the next largest prime greater than in6addr_nhash. + */ + k = (in6addr_nhash % 2 == 0) ? in6addr_nhash + 1 : in6addr_nhash + 2; + for (;;) { + p = 1; + for (i = 3; i * i <= k; i += 2) { + if (k % i == 0) { + p = 0; + } + } + if (p == 1) { + break; + } + k += 2; + } + in6addr_hashp = k; +} + +static int +sysctl_ip6_checkinterface SYSCTL_HANDLER_ARGS +{ +#pragma unused(arg1, arg2) + int error, i; + + i = ip6_checkinterface; + error = sysctl_handle_int(oidp, &i, 0, req); + if (error || req->newptr == USER_ADDR_NULL) { + return error; + } + + switch (i) { + case IP6_CHECKINTERFACE_WEAK_ES: + case IP6_CHECKINTERFACE_HYBRID_ES: + case IP6_CHECKINTERFACE_STRONG_ES: + if (ip6_checkinterface != i) { + ip6_checkinterface = i; + os_log(OS_LOG_DEFAULT, "%s: ip6_checkinterface is now %d\n", + __func__, ip6_checkinterface); + } + break; + default: + error = EINVAL; + break; + } + return error; +}