-/* $KAME: frag6.c,v 1.23 2000/02/28 16:18:11 itojun Exp $ */
+/*
+ * Copyright (c) 2000-2010 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
+ * compliance with the License. The rights granted to you under the License
+ * may not be used to create, or enable the creation or redistribution of,
+ * 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,
+ * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
+ * 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/frag6.c,v 1.2.2.5 2001/07/03 11:01:50 ume Exp $ */
+/* $KAME: frag6.c,v 1.31 2001/05/17 13:45:34 jinmei Exp $ */
/*
* Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
#include <sys/kernel.h>
#include <sys/syslog.h>
#include <kern/queue.h>
+#include <kern/locks.h>
#include <net/if.h>
#include <net/route.h>
#include <netinet/in.h>
#include <netinet/in_var.h>
+#include <netinet/ip.h>
#include <netinet/ip6.h>
#include <netinet6/ip6_var.h>
-#if !(defined(__FreeBSD__) && __FreeBSD__ >= 3) && !defined(__OpenBSD__) && !(defined(__bsdi__) && _BSDI_VERSION >= 199802) && !defined(__APPLE__)
-#include <netinet6/in6_pcb.h>
-#endif
#include <netinet/icmp6.h>
#include <net/net_osdep.h>
*/
#define IN6_IFSTAT_STRICT
-static void frag6_enq __P((struct ip6asfrag *, struct ip6asfrag *));
-static void frag6_deq __P((struct ip6asfrag *));
-static void frag6_insque __P((struct ip6q *, struct ip6q *));
-static void frag6_remque __P((struct ip6q *));
-static void frag6_freef __P((struct ip6q *));
+static void frag6_enq(struct ip6asfrag *, struct ip6asfrag *);
+static void frag6_deq(struct ip6asfrag *);
+static void frag6_insque(struct ip6q *, struct ip6q *);
+static void frag6_remque(struct ip6q *);
+static void frag6_freef(struct ip6q *);
+/* XXX we eventually need splreass6, or some real semaphore */
int frag6_doing_reass;
u_int frag6_nfragpackets;
+static u_int frag6_nfrags;
struct ip6q ip6q; /* ip6 reassemble queue */
-/* FreeBSD tweak */
-#if !defined(M_FTABLE) && (defined(__FreeBSD__) && __FreeBSD__ >= 3)
-MALLOC_DEFINE(M_FTABLE, "fragment", "fragment reassembly header");
-#endif
-
-#ifndef offsetof /* XXX */
-#define offsetof(type, member) ((size_t)(&((type *)0)->member))
-#endif
+extern lck_mtx_t *inet6_domain_mutex;
/*
* Initialise reassembly queue and fragment identifier.
*/
{
struct timeval tv;
+ ip6_maxfragpackets = nmbclusters / 32;
+ ip6_maxfrags = nmbclusters / 4;
+
/*
* in many cases, random() here does NOT return random number
* as initialization during bootstrap time occur in fixed order.
*/
microtime(&tv);
- ip6q.ip6q_next = ip6q.ip6q_prev = &ip6q;
ip6_id = random() ^ tv.tv_usec;
+ ip6q.ip6q_next = ip6q.ip6q_prev = &ip6q;
}
/*
*/
/*
* Fragment input
+ * NOTE: this function is called with the inet6_domain_mutex held from ip6_input.
+ * inet6_domain_mutex is protecting he frag6 queue manipulation.
*/
int
-frag6_input(mp, offp, proto)
- struct mbuf **mp;
- int *offp, proto;
+frag6_input(struct mbuf **mp, int *offp, int proto)
{
+#pragma unused(proto)
struct mbuf *m = *mp, *t;
struct ip6_hdr *ip6;
struct ip6_frag *ip6f;
int first_frag = 0;
int fragoff, frgpartlen; /* must be larger than u_int16_t */
struct ifnet *dstifp;
+ struct ifaddr *ifa = NULL;
+ u_int8_t ecn, ecn0;
+
#ifdef IN6_IFSTAT_STRICT
- static struct route_in6 ro;
+ struct route_in6 ro;
struct sockaddr_in6 *dst;
#endif
ip6 = mtod(m, struct ip6_hdr *);
#ifndef PULLDOWN_TEST
- IP6_EXTHDR_CHECK(m, offset, sizeof(struct ip6_frag), IPPROTO_DONE);
+ IP6_EXTHDR_CHECK(m, offset, sizeof(struct ip6_frag), return IPPROTO_DONE);
ip6f = (struct ip6_frag *)((caddr_t)ip6 + offset);
#else
IP6_EXTHDR_GET(ip6f, struct ip6_frag *, m, offset, sizeof(*ip6f));
dstifp = NULL;
#ifdef IN6_IFSTAT_STRICT
/* find the destination interface of the packet. */
+ bzero(&ro, sizeof (ro));
dst = (struct sockaddr_in6 *)&ro.ro_dst;
- if (ro.ro_rt
- && ((ro.ro_rt->rt_flags & RTF_UP) == 0
- || !IN6_ARE_ADDR_EQUAL(&dst->sin6_addr, &ip6->ip6_dst))) {
- RTFREE(ro.ro_rt);
- ro.ro_rt = (struct rtentry *)0;
- }
- if (ro.ro_rt == NULL) {
- bzero(dst, sizeof(*dst));
- dst->sin6_family = AF_INET6;
- dst->sin6_len = sizeof(struct sockaddr_in6);
- dst->sin6_addr = ip6->ip6_dst;
- }
-#ifndef __bsdi__
+ dst->sin6_family = AF_INET6;
+ dst->sin6_len = sizeof (struct sockaddr_in6);
+ dst->sin6_addr = ip6->ip6_dst;
+
rtalloc((struct route *)&ro);
-#else
- rtcalloc((struct route *)&ro);
-#endif
- if (ro.ro_rt != NULL && ro.ro_rt->rt_ifa != NULL)
- dstifp = ((struct in6_ifaddr *)ro.ro_rt->rt_ifa)->ia_ifp;
+ if (ro.ro_rt != NULL) {
+ RT_LOCK(ro.ro_rt);
+ if ((ifa = ro.ro_rt->rt_ifa) != NULL) {
+ IFA_ADDREF(ifa);
+ dstifp = ((struct in6_ifaddr *)ro.ro_rt->rt_ifa)->ia_ifp;
+ }
+ RT_UNLOCK(ro.ro_rt);
+ rtfree(ro.ro_rt);
+ ro.ro_rt = NULL;
+ }
#else
/* we are violating the spec, this is not the destination interface */
if ((m->m_flags & M_PKTHDR) != 0)
if (ip6->ip6_plen == 0) {
icmp6_error(m, ICMP6_PARAM_PROB, ICMP6_PARAMPROB_HEADER, offset);
in6_ifstat_inc(dstifp, ifs6_reass_fail);
+ if (ifa != NULL)
+ IFA_REMREF(ifa);
return IPPROTO_DONE;
}
ICMP6_PARAMPROB_HEADER,
offsetof(struct ip6_hdr, ip6_plen));
in6_ifstat_inc(dstifp, ifs6_reass_fail);
+ if (ifa != NULL)
+ IFA_REMREF(ifa);
return IPPROTO_DONE;
}
/* offset now points to data portion */
offset += sizeof(struct ip6_frag);
+ frag6_doing_reass = 1;
+
+ /*
+ * Enforce upper bound on number of fragments.
+ * If maxfrag is 0, never accept fragments.
+ * If maxfrag is -1, accept all fragments without limitation.
+ */
+ if (ip6_maxfrags < 0)
+ ;
+ else if (frag6_nfrags >= (u_int)ip6_maxfrags)
+ goto dropfrag;
+
for (q6 = ip6q.ip6q_next; q6 != &ip6q; q6 = q6->ip6q_next)
if (ip6f->ip6f_ident == q6->ip6q_ident &&
IN6_ARE_ADDR_EQUAL(&ip6->ip6_src, &q6->ip6q_src) &&
* the first fragment to arrive, create a reassembly queue.
*/
first_frag = 1;
- frag6_nfragpackets++;
/*
* Enforce upper bound on number of fragmented packets
* If maxfrag is 0, never accept fragments.
* If maxfrag is -1, accept all fragments without limitation.
*/
- if (frag6_nfragpackets >= (u_int)ip6_maxfragpackets) {
- ip6stat.ip6s_fragoverflow++;
- in6_ifstat_inc(dstifp, ifs6_reass_fail);
- frag6_freef(ip6q.ip6q_prev);
- }
+ if (ip6_maxfragpackets < 0)
+ ;
+ else if (frag6_nfragpackets >= (u_int)ip6_maxfragpackets)
+ goto dropfrag;
+ frag6_nfragpackets++;
q6 = (struct ip6q *)_MALLOC(sizeof(struct ip6q), M_FTABLE,
M_DONTWAIT);
if (q6 == NULL)
/* ip6q_nxt will be filled afterwards, from 1st fragment */
q6->ip6q_down = q6->ip6q_up = (struct ip6asfrag *)q6;
-#if notyet
+#ifdef notyet
q6->ip6q_nxtp = (u_char *)nxtp;
#endif
q6->ip6q_ident = ip6f->ip6f_ident;
- q6->ip6q_arrive = 0; /* Is it used anywhere? */
q6->ip6q_ttl = IPV6_FRAGTTL;
q6->ip6q_src = ip6->ip6_src;
q6->ip6q_dst = ip6->ip6_dst;
+ q6->ip6q_ecn =
+ (ntohl(ip6->ip6_flow) >> 20) & IPTOS_ECN_MASK;
q6->ip6q_unfrglen = -1; /* The 1st fragment has not arrived. */
+
+ q6->ip6q_nfrag = 0;
}
/*
icmp6_error(m, ICMP6_PARAM_PROB, ICMP6_PARAMPROB_HEADER,
offset - sizeof(struct ip6_frag) +
offsetof(struct ip6_frag, ip6f_offlg));
+ frag6_doing_reass = 0;
+ if (ifa != NULL)
+ IFA_REMREF(ifa);
return(IPPROTO_DONE);
}
}
icmp6_error(m, ICMP6_PARAM_PROB, ICMP6_PARAMPROB_HEADER,
offset - sizeof(struct ip6_frag) +
offsetof(struct ip6_frag, ip6f_offlg));
+ frag6_doing_reass = 0;
+ if (ifa != NULL)
+ IFA_REMREF(ifa);
return(IPPROTO_DONE);
}
/*
/* dequeue the fragment. */
frag6_deq(af6);
- _FREE(af6, M_FTABLE);
+ FREE(af6, M_FTABLE);
/* adjust pointer. */
ip6err = mtod(merr, struct ip6_hdr *);
if (ip6af == NULL)
goto dropfrag;
bzero(ip6af, sizeof(*ip6af));
- ip6af->ip6af_head = ip6->ip6_flow;
- ip6af->ip6af_len = ip6->ip6_plen;
- ip6af->ip6af_nxt = ip6->ip6_nxt;
- ip6af->ip6af_hlim = ip6->ip6_hlim;
ip6af->ip6af_mff = ip6f->ip6f_offlg & IP6F_MORE_FRAG;
ip6af->ip6af_off = fragoff;
ip6af->ip6af_frglen = frgpartlen;
goto insert;
}
+ /*
+ * Handle ECN by comparing this segment with the first one;
+ * if CE is set, do not lose CE.
+ * drop if CE and not-ECT are mixed for the same packet.
+ */
+ ecn = (ntohl(ip6->ip6_flow) >> 20) & IPTOS_ECN_MASK;
+ ecn0 = q6->ip6q_ecn;
+ if (ecn == IPTOS_ECN_CE) {
+ if (ecn0 == IPTOS_ECN_NOTECT) {
+ FREE(ip6af, M_FTABLE);
+ goto dropfrag;
+ }
+ if (ecn0 != IPTOS_ECN_CE)
+ q6->ip6q_ecn = IPTOS_ECN_CE;
+ }
+ if (ecn == IPTOS_ECN_NOTECT && ecn0 != IPTOS_ECN_NOTECT) {
+ FREE(ip6af, M_FTABLE);
+ goto dropfrag;
+ }
+
/*
* Find a segment which begins after this one does.
*/
* If the incoming framgent overlaps some existing fragments in
* the reassembly queue, drop it, since it is dangerous to override
* existing fragments from a security point of view.
+ * We don't know which fragment is the bad guy - here we trust
+ * fragment that came in earlier, with no real reason.
+ *
+ * Note: due to changes after disabling this part, mbuf passed to
+ * m_adj() below now does not meet the requirement.
*/
if (af6->ip6af_up != (struct ip6asfrag *)q6) {
i = af6->ip6af_up->ip6af_off + af6->ip6af_up->ip6af_frglen
- ip6af->ip6af_off;
if (i > 0) {
+#if 0 /* suppress the noisy log */
log(LOG_ERR, "%d bytes of a fragment from %s "
"overlaps the previous fragment\n",
i, ip6_sprintf(&q6->ip6q_src));
+#endif
+ FREE(ip6af, M_FTABLE);
goto dropfrag;
}
}
if (af6 != (struct ip6asfrag *)q6) {
i = (ip6af->ip6af_off + ip6af->ip6af_frglen) - af6->ip6af_off;
if (i > 0) {
+#if 0 /* suppress the noisy log */
log(LOG_ERR, "%d bytes of a fragment from %s "
"overlaps the succeeding fragment",
i, ip6_sprintf(&q6->ip6q_src));
+#endif
+ FREE(ip6af, M_FTABLE);
goto dropfrag;
}
}
* the most recently active fragmented packet.
*/
frag6_enq(ip6af, af6->ip6af_up);
+ frag6_nfrags++;
+ q6->ip6q_nfrag++;
#if 0 /* xxx */
if (q6 != ip6q.ip6q_next) {
frag6_remque(q6);
af6 = af6->ip6af_down) {
if (af6->ip6af_off != next) {
frag6_doing_reass = 0;
+ if (ifa != NULL)
+ IFA_REMREF(ifa);
return IPPROTO_DONE;
}
next += af6->ip6af_frglen;
}
if (af6->ip6af_up->ip6af_mff) {
frag6_doing_reass = 0;
+ if (ifa != NULL)
+ IFA_REMREF(ifa);
return IPPROTO_DONE;
}
t = t->m_next;
t->m_next = IP6_REASS_MBUF(af6);
m_adj(t->m_next, af6->ip6af_offset);
- _FREE(af6, M_FTABLE);
+ FREE(af6, M_FTABLE);
af6 = af6dwn;
}
/* adjust offset to point where the original next header starts */
offset = ip6af->ip6af_offset - sizeof(struct ip6_frag);
- _FREE(ip6af, M_FTABLE);
+ FREE(ip6af, M_FTABLE);
ip6 = mtod(m, struct ip6_hdr *);
ip6->ip6_plen = htons((u_short)next + offset - sizeof(struct ip6_hdr));
ip6->ip6_src = q6->ip6q_src;
ip6->ip6_dst = q6->ip6q_dst;
+ if (q6->ip6q_ecn == IPTOS_ECN_CE)
+ ip6->ip6_flow |= htonl(IPTOS_ECN_CE << 20);
+
nxt = q6->ip6q_nxt;
#if notyet
*q6->ip6q_nxtp = (u_char)(nxt & 0xff);
#endif
- /*
- * Delete frag6 header with as a few cost as possible.
- */
- if (offset < m->m_len) {
+ /* Delete frag6 header */
+ if (m->m_len >= offset + sizeof(struct ip6_frag)) {
+ /* This is the only possible case with !PULLDOWN_TEST */
ovbcopy((caddr_t)ip6, (caddr_t)ip6 + sizeof(struct ip6_frag),
offset);
m->m_data += sizeof(struct ip6_frag);
/* this comes with no copy if the boundary is on cluster */
if ((t = m_split(m, offset, M_DONTWAIT)) == NULL) {
frag6_remque(q6);
- _FREE(q6, M_FTABLE);
+ frag6_nfrags -= q6->ip6q_nfrag;
+ FREE(q6, M_FTABLE);
frag6_nfragpackets--;
goto dropfrag;
}
}
frag6_remque(q6);
- _FREE(q6, M_FTABLE);
+ frag6_nfrags -= q6->ip6q_nfrag;
+ FREE(q6, M_FTABLE);
frag6_nfragpackets--;
if (m->m_flags & M_PKTHDR) { /* Isn't it always true? */
*offp = offset;
frag6_doing_reass = 0;
+ if (ifa != NULL)
+ IFA_REMREF(ifa);
return nxt;
dropfrag:
in6_ifstat_inc(dstifp, ifs6_reass_fail);
ip6stat.ip6s_fragdropped++;
m_freem(m);
+ frag6_doing_reass = 0;
+ if (ifa != NULL)
+ IFA_REMREF(ifa);
return IPPROTO_DONE;
}
/* adjust pointer */
ip6 = mtod(m, struct ip6_hdr *);
- /* restoure source and destination addresses */
+ /* restore source and destination addresses */
ip6->ip6_src = q6->ip6q_src;
ip6->ip6_dst = q6->ip6q_dst;
-
icmp6_error(m, ICMP6_TIME_EXCEEDED,
ICMP6_TIME_EXCEED_REASSEMBLY, 0);
} else
m_freem(m);
- _FREE(af6, M_FTABLE);
+ FREE(af6, M_FTABLE);
}
frag6_remque(q6);
- _FREE(q6, M_FTABLE);
+ frag6_nfrags -= q6->ip6q_nfrag;
+ FREE(q6, M_FTABLE);
frag6_nfragpackets--;
}
}
/*
- * IP timer processing;
+ * IPv6 reassembling timer processing;
* if a timer expires on a reassembly
* queue, discard it.
*/
frag6_slowtimo()
{
struct ip6q *q6;
- int s;
-#ifdef __APPLE__
- boolean_t funnel_state;
- funnel_state = thread_set_funneled(TRUE);
-#endif
-#ifdef __NetBSD__
- s = splsoftnet();
-#else
- s = splnet();
-#endif
-#if 0
- extern struct route_in6 ip6_forward_rt;
-#endif
+ lck_mtx_lock(inet6_domain_mutex);
frag6_doing_reass = 1;
q6 = ip6q.ip6q_next;
* (due to the limit being lowered), drain off
* enough to get down to the new limit.
*/
- while (frag6_nfragpackets > (u_int)ip6_maxfragpackets) {
+ while (frag6_nfragpackets > (u_int)ip6_maxfragpackets &&
+ ip6q.ip6q_prev) {
ip6stat.ip6s_fragoverflow++;
/* XXX in6_ifstat_inc(ifp, ifs6_reass_fail) */
frag6_freef(ip6q.ip6q_prev);
}
frag6_doing_reass = 0;
-
-#if 0
- /*
- * Routing changes might produce a better route than we last used;
- * make sure we notice eventually, even if forwarding only for one
- * destination and the cache is never replaced.
- */
- if (ip6_forward_rt.ro_rt) {
- RTFREE(ip6_forward_rt.ro_rt);
- ip6_forward_rt.ro_rt = 0;
- }
- if (ipsrcchk_rt.ro_rt) {
- RTFREE(ipsrcchk_rt.ro_rt);
- ipsrcchk_rt.ro_rt = 0;
- }
-#endif
-
- splx(s);
-#ifdef __APPLE__
- (void) thread_set_funneled(funnel_state);
-#endif
+ lck_mtx_unlock(inet6_domain_mutex);
}
/*
{
if (frag6_doing_reass)
return;
+ lck_mtx_lock(inet6_domain_mutex);
while (ip6q.ip6q_next != &ip6q) {
ip6stat.ip6s_fragdropped++;
/* XXX in6_ifstat_inc(ifp, ifs6_reass_fail) */
frag6_freef(ip6q.ip6q_next);
}
+ lck_mtx_unlock(inet6_domain_mutex);
}