| version 1.169.2.8, 2005/11/10 14:11:07 |
version 1.170, 2003/07/03 05:03:53 |
|
|
| * 2. Redistributions in binary form must reproduce the above copyright |
* 2. Redistributions in binary form must reproduce the above copyright |
| * notice, this list of conditions and the following disclaimer in the |
* notice, this list of conditions and the following disclaimer in the |
| * documentation and/or other materials provided with the distribution. |
* documentation and/or other materials provided with the distribution. |
| * 3. Neither the name of the University nor the names of its contributors |
* 3. All advertising materials mentioning features or use of this software |
| |
* must display the following acknowledgement: |
| |
* This product includes software developed by the University of |
| |
* California, Berkeley and its contributors. |
| |
* 4. Neither the name of the University nor the names of its contributors |
| * may be used to endorse or promote products derived from this software |
* may be used to endorse or promote products derived from this software |
| * without specific prior written permission. |
* without specific prior written permission. |
| * |
* |
|
|
| #include <sys/cdefs.h> |
#include <sys/cdefs.h> |
| __KERNEL_RCSID(0, "$NetBSD$"); |
__KERNEL_RCSID(0, "$NetBSD$"); |
| |
|
| #include "opt_inet.h" |
|
| #include "opt_gateway.h" |
#include "opt_gateway.h" |
| #include "opt_pfil_hooks.h" |
#include "opt_pfil_hooks.h" |
| #include "opt_ipsec.h" |
#include "opt_ipsec.h" |
| Line 131 __KERNEL_RCSID(0, "$NetBSD$"); |
|
| Line 134 __KERNEL_RCSID(0, "$NetBSD$"); |
|
| #include <netinet/in_systm.h> |
#include <netinet/in_systm.h> |
| #include <netinet/ip.h> |
#include <netinet/ip.h> |
| #include <netinet/in_pcb.h> |
#include <netinet/in_pcb.h> |
| #include <netinet/in_proto.h> |
|
| #include <netinet/in_var.h> |
#include <netinet/in_var.h> |
| #include <netinet/ip_var.h> |
#include <netinet/ip_var.h> |
| #include <netinet/ip_icmp.h> |
#include <netinet/ip_icmp.h> |
| Line 149 __KERNEL_RCSID(0, "$NetBSD$"); |
|
| Line 151 __KERNEL_RCSID(0, "$NetBSD$"); |
|
| #include <netinet6/ipsec.h> |
#include <netinet6/ipsec.h> |
| #include <netkey/key.h> |
#include <netkey/key.h> |
| #endif |
#endif |
| #ifdef FAST_IPSEC |
|
| #include <netipsec/ipsec.h> |
|
| #include <netipsec/key.h> |
|
| #endif /* FAST_IPSEC*/ |
|
| |
|
| #ifndef IPFORWARDING |
#ifndef IPFORWARDING |
| #ifdef GATEWAY |
#ifdef GATEWAY |
| Line 199 int ip_mtudisc_timeout = IPMTUDISCTIMEOU |
|
| Line 197 int ip_mtudisc_timeout = IPMTUDISCTIMEOU |
|
| #ifdef DIAGNOSTIC |
#ifdef DIAGNOSTIC |
| int ipprintfs = 0; |
int ipprintfs = 0; |
| #endif |
#endif |
| |
|
| int ip_do_randomid = 0; |
|
| |
|
| /* |
/* |
| * XXX - Setting ip_checkinterface mostly implements the receive side of |
* XXX - Setting ip_checkinterface mostly implements the receive side of |
| * the Strong ES model described in RFC 1122, but since the routing table |
* the Strong ES model described in RFC 1122, but since the routing table |
| Line 220 int ip_checkinterface = 0; |
|
| Line 215 int ip_checkinterface = 0; |
|
| |
|
| struct rttimer_queue *ip_mtudisc_timeout_q = NULL; |
struct rttimer_queue *ip_mtudisc_timeout_q = NULL; |
| |
|
| |
extern struct domain inetdomain; |
| int ipqmaxlen = IFQ_MAXLEN; |
int ipqmaxlen = IFQ_MAXLEN; |
| u_long in_ifaddrhash; /* size of hash table - 1 */ |
u_long in_ifaddrhash; /* size of hash table - 1 */ |
| int in_ifaddrentries; /* total number of addrs */ |
int in_ifaddrentries; /* total number of addrs */ |
| struct in_ifaddrhead in_ifaddrhead; |
struct in_ifaddrhead in_ifaddr; |
| struct in_ifaddrhashhead *in_ifaddrhashtbl; |
struct in_ifaddrhashhead *in_ifaddrhashtbl; |
| u_long in_multihash; /* size of hash table - 1 */ |
u_long in_multihash; /* size of hash table - 1 */ |
| int in_multientries; /* total number of addrs */ |
int in_multientries; /* total number of addrs */ |
| |
struct in_multihead in_multi; |
| struct in_multihashhead *in_multihashtbl; |
struct in_multihashhead *in_multihashtbl; |
| struct ifqueue ipintrq; |
struct ifqueue ipintrq; |
| struct ipstat ipstat; |
struct ipstat ipstat; |
| uint16_t ip_id; |
u_int16_t ip_id; |
| |
|
| #ifdef PFIL_HOOKS |
#ifdef PFIL_HOOKS |
| struct pfil_head inet_pfil_hook; |
struct pfil_head inet_pfil_hook; |
| #endif |
#endif |
| |
|
| /* |
struct ipqhead ipq; |
| * Cached copy of nmbclusters. If nbclusters is different, |
|
| * recalculate IP parameters derived from nmbclusters. |
|
| */ |
|
| static int ip_nmbclusters; /* copy of nmbclusters */ |
|
| static void ip_nmbclusters_changed(void); /* recalc limits */ |
|
| |
|
| #define CHECK_NMBCLUSTER_PARAMS() \ |
|
| do { \ |
|
| if (__predict_false(ip_nmbclusters != nmbclusters)) \ |
|
| ip_nmbclusters_changed(); \ |
|
| } while (/*CONSTCOND*/0) |
|
| |
|
| /* IP datagram reassembly queues (hashed) */ |
|
| #define IPREASS_NHASH_LOG2 6 |
|
| #define IPREASS_NHASH (1 << IPREASS_NHASH_LOG2) |
|
| #define IPREASS_HMASK (IPREASS_NHASH - 1) |
|
| #define IPREASS_HASH(x,y) \ |
|
| (((((x) & 0xF) | ((((x) >> 8) & 0xF) << 4)) ^ (y)) & IPREASS_HMASK) |
|
| struct ipqhead ipq[IPREASS_NHASH]; |
|
| int ipq_locked; |
int ipq_locked; |
| static int ip_nfragpackets; /* packets in reass queue */ |
int ip_nfragpackets = 0; |
| static int ip_nfrags; /* total fragments in reass queues */ |
int ip_maxfragpackets = 200; |
| |
|
| int ip_maxfragpackets = 200; /* limit on packets. XXX sysctl */ |
|
| int ip_maxfrags; /* limit on fragments. XXX sysctl */ |
|
| |
|
| |
|
| /* |
|
| * Additive-Increase/Multiplicative-Decrease (AIMD) strategy for |
|
| * IP reassembly queue buffer managment. |
|
| * |
|
| * We keep a count of total IP fragments (NB: not fragmented packets!) |
|
| * awaiting reassembly (ip_nfrags) and a limit (ip_maxfrags) on fragments. |
|
| * If ip_nfrags exceeds ip_maxfrags the limit, we drop half the |
|
| * total fragments in reassembly queues.This AIMD policy avoids |
|
| * repeatedly deleting single packets under heavy fragmentation load |
|
| * (e.g., from lossy NFS peers). |
|
| */ |
|
| static u_int ip_reass_ttl_decr(u_int ticks); |
|
| static void ip_reass_drophalf(void); |
|
| |
|
| |
static __inline int ipq_lock_try __P((void)); |
| static __inline int ipq_lock_try(void); |
static __inline void ipq_unlock __P((void)); |
| static __inline void ipq_unlock(void); |
|
| |
|
| static __inline int |
static __inline int |
| ipq_lock_try(void) |
ipq_lock_try() |
| { |
{ |
| int s; |
int s; |
| |
|
| Line 302 ipq_lock_try(void) |
|
| Line 261 ipq_lock_try(void) |
|
| } |
} |
| |
|
| static __inline void |
static __inline void |
| ipq_unlock(void) |
ipq_unlock() |
| { |
{ |
| int s; |
int s; |
| |
|
|
|
| |
|
| #define IPQ_UNLOCK() ipq_unlock() |
#define IPQ_UNLOCK() ipq_unlock() |
| |
|
| POOL_INIT(inmulti_pool, sizeof(struct in_multi), 0, 0, 0, "inmltpl", NULL); |
struct pool inmulti_pool; |
| POOL_INIT(ipqent_pool, sizeof(struct ipqent), 0, 0, 0, "ipqepl", NULL); |
struct pool ipqent_pool; |
| |
|
| #ifdef INET_CSUM_COUNTERS |
#ifdef INET_CSUM_COUNTERS |
| #include <sys/device.h> |
#include <sys/device.h> |
| Line 348 struct evcnt ip_swcsum = EVCNT_INITIALIZ |
|
| Line 307 struct evcnt ip_swcsum = EVCNT_INITIALIZ |
|
| |
|
| #define INET_CSUM_COUNTER_INCR(ev) (ev)->ev_count++ |
#define INET_CSUM_COUNTER_INCR(ev) (ev)->ev_count++ |
| |
|
| EVCNT_ATTACH_STATIC(ip_hwcsum_bad); |
|
| EVCNT_ATTACH_STATIC(ip_hwcsum_ok); |
|
| EVCNT_ATTACH_STATIC(ip_swcsum); |
|
| |
|
| #else |
#else |
| |
|
| #define INET_CSUM_COUNTER_INCR(ev) /* nothing */ |
#define INET_CSUM_COUNTER_INCR(ev) /* nothing */ |
| Line 373 static struct ip_srcrt { |
|
| Line 328 static struct ip_srcrt { |
|
| struct in_addr route[MAX_IPOPTLEN/sizeof(struct in_addr)]; |
struct in_addr route[MAX_IPOPTLEN/sizeof(struct in_addr)]; |
| } ip_srcrt; |
} ip_srcrt; |
| |
|
| static void save_rte(u_char *, struct in_addr); |
static void save_rte __P((u_char *, struct in_addr)); |
| |
|
| #ifdef MBUFTRACE |
#ifdef MBUFTRACE |
| struct mowner ip_rx_mowner = { "internet", "rx" }; |
struct mowner ip_rx_mowner = { "internet", "rx" }; |
| Line 381 struct mowner ip_tx_mowner = { "internet |
|
| Line 336 struct mowner ip_tx_mowner = { "internet |
|
| #endif |
#endif |
| |
|
| /* |
/* |
| * Compute IP limits derived from the value of nmbclusters. |
|
| */ |
|
| static void |
|
| ip_nmbclusters_changed(void) |
|
| { |
|
| ip_maxfrags = nmbclusters / 4; |
|
| ip_nmbclusters = nmbclusters; |
|
| } |
|
| |
|
| /* |
|
| * IP initialization: fill in IP protocol switch table. |
* IP initialization: fill in IP protocol switch table. |
| * All protocols not implemented in kernel go to raw IP protocol handler. |
* All protocols not implemented in kernel go to raw IP protocol handler. |
| */ |
*/ |
| void |
void |
| ip_init(void) |
ip_init() |
| { |
{ |
| const struct protosw *pr; |
struct protosw *pr; |
| int i; |
int i; |
| |
|
| |
pool_init(&inmulti_pool, sizeof(struct in_multi), 0, 0, 0, "inmltpl", |
| |
NULL); |
| |
pool_init(&ipqent_pool, sizeof(struct ipqent), 0, 0, 0, "ipqepl", |
| |
NULL); |
| |
|
| pr = pffindproto(PF_INET, IPPROTO_RAW, SOCK_RAW); |
pr = pffindproto(PF_INET, IPPROTO_RAW, SOCK_RAW); |
| if (pr == 0) |
if (pr == 0) |
| panic("ip_init"); |
panic("ip_init"); |
|
|
| if (pr->pr_domain->dom_family == PF_INET && |
if (pr->pr_domain->dom_family == PF_INET && |
| pr->pr_protocol && pr->pr_protocol != IPPROTO_RAW) |
pr->pr_protocol && pr->pr_protocol != IPPROTO_RAW) |
| ip_protox[pr->pr_protocol] = pr - inetsw; |
ip_protox[pr->pr_protocol] = pr - inetsw; |
| |
LIST_INIT(&ipq); |
| for (i = 0; i < IPREASS_NHASH; i++) |
ip_id = time.tv_sec & 0xffff; |
| LIST_INIT(&ipq[i]); |
|
| |
|
| ip_id = time.tv_sec & 0xfffff; |
|
| |
|
| ipintrq.ifq_maxlen = ipqmaxlen; |
ipintrq.ifq_maxlen = ipqmaxlen; |
| ip_nmbclusters_changed(); |
TAILQ_INIT(&in_ifaddr); |
| |
|
| TAILQ_INIT(&in_ifaddrhead); |
|
| in_ifaddrhashtbl = hashinit(IN_IFADDR_HASH_SIZE, HASH_LIST, M_IFADDR, |
in_ifaddrhashtbl = hashinit(IN_IFADDR_HASH_SIZE, HASH_LIST, M_IFADDR, |
| M_WAITOK, &in_ifaddrhash); |
M_WAITOK, &in_ifaddrhash); |
| in_multihashtbl = hashinit(IN_IFADDR_HASH_SIZE, HASH_LIST, M_IPMADDR, |
in_multihashtbl = hashinit(IN_IFADDR_HASH_SIZE, HASH_LIST, M_IPMADDR, |
|
|
| "error %d\n", i); |
"error %d\n", i); |
| #endif /* PFIL_HOOKS */ |
#endif /* PFIL_HOOKS */ |
| |
|
| |
#ifdef INET_CSUM_COUNTERS |
| |
evcnt_attach_static(&ip_hwcsum_bad); |
| |
evcnt_attach_static(&ip_hwcsum_ok); |
| |
evcnt_attach_static(&ip_swcsum); |
| |
#endif /* INET_CSUM_COUNTERS */ |
| |
|
| #ifdef MBUFTRACE |
#ifdef MBUFTRACE |
| MOWNER_ATTACH(&ip_tx_mowner); |
MOWNER_ATTACH(&ip_tx_mowner); |
| MOWNER_ATTACH(&ip_rx_mowner); |
MOWNER_ATTACH(&ip_rx_mowner); |
| Line 452 struct route ipforward_rt; |
|
| Line 402 struct route ipforward_rt; |
|
| * IP software interrupt routine |
* IP software interrupt routine |
| */ |
*/ |
| void |
void |
| ipintr(void) |
ipintr() |
| { |
{ |
| int s; |
int s; |
| struct mbuf *m; |
struct mbuf *m; |
| Line 484 ip_input(struct mbuf *m) |
|
| Line 434 ip_input(struct mbuf *m) |
|
| int downmatch; |
int downmatch; |
| int checkif; |
int checkif; |
| int srcrt = 0; |
int srcrt = 0; |
| u_int hash; |
|
| #ifdef FAST_IPSEC |
|
| struct m_tag *mtag; |
|
| struct tdb_ident *tdbi; |
|
| struct secpolicy *sp; |
|
| int s, error; |
|
| #endif /* FAST_IPSEC */ |
|
| |
|
| MCLAIM(m, &ip_rx_mowner); |
MCLAIM(m, &ip_rx_mowner); |
| #ifdef DIAGNOSTIC |
#ifdef DIAGNOSTIC |
| if ((m->m_flags & M_PKTHDR) == 0) |
if ((m->m_flags & M_PKTHDR) == 0) |
| panic("ipintr no HDR"); |
panic("ipintr no HDR"); |
| #endif |
#endif |
| |
#ifdef 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 |
| |
|
| /* |
/* |
| * If no IP addresses have been set yet but the interfaces |
* If no IP addresses have been set yet but the interfaces |
| * are receiving, can't do anything with incoming packets yet. |
* are receiving, can't do anything with incoming packets yet. |
| */ |
*/ |
| if (TAILQ_FIRST(&in_ifaddrhead) == 0) |
if (TAILQ_FIRST(&in_ifaddr) == 0) |
| goto bad; |
goto bad; |
| ipstat.ips_total++; |
ipstat.ips_total++; |
| /* |
/* |
| Line 573 ip_input(struct mbuf *m) |
|
| Line 526 ip_input(struct mbuf *m) |
|
| break; |
break; |
| |
|
| default: |
default: |
| /* |
/* Must compute it ourselves. */ |
| * Must compute it ourselves. Maybe skip checksum on |
INET_CSUM_COUNTER_INCR(&ip_swcsum); |
| * loopback interfaces. |
if (in_cksum(m, hlen) != 0) |
| */ |
goto bad; |
| if (__predict_true(!(m->m_pkthdr.rcvif->if_flags & |
|
| IFF_LOOPBACK) || ip_do_loopback_cksum)) { |
|
| INET_CSUM_COUNTER_INCR(&ip_swcsum); |
|
| if (in_cksum(m, hlen) != 0) |
|
| goto badcsum; |
|
| } |
|
| break; |
break; |
| } |
} |
| |
|
| Line 615 ip_input(struct mbuf *m) |
|
| Line 562 ip_input(struct mbuf *m) |
|
| m_adj(m, len - m->m_pkthdr.len); |
m_adj(m, len - m->m_pkthdr.len); |
| } |
} |
| |
|
| #if defined(IPSEC) |
#ifdef IPSEC |
| /* ipflow (IP fast forwarding) is not compatible with IPsec. */ |
/* ipflow (IP fast forwarding) is not compatible with IPsec. */ |
| m->m_flags &= ~M_CANFASTFWD; |
m->m_flags &= ~M_CANFASTFWD; |
| #else |
#else |
| Line 640 ip_input(struct mbuf *m) |
|
| Line 587 ip_input(struct mbuf *m) |
|
| */ |
*/ |
| #ifdef IPSEC |
#ifdef IPSEC |
| if (!ipsec_getnhist(m)) |
if (!ipsec_getnhist(m)) |
| #elif defined(FAST_IPSEC) |
|
| if (!ipsec_indone(m)) |
|
| #else |
#else |
| if (1) |
if (1) |
| #endif |
#endif |
| Line 656 ip_input(struct mbuf *m) |
|
| Line 601 ip_input(struct mbuf *m) |
|
| return; |
return; |
| ip = mtod(m, struct ip *); |
ip = mtod(m, struct ip *); |
| hlen = ip->ip_hl << 2; |
hlen = ip->ip_hl << 2; |
| /* |
|
| * XXX The setting of "srcrt" here is to prevent ip_forward() |
|
| * from generating ICMP redirects for packets that have |
|
| * been redirected by a hook back out on to the same LAN that |
|
| * they came from and is not an indication that the packet |
|
| * is being inffluenced by source routing options. This |
|
| * allows things like |
|
| * "rdr tlp0 0/0 port 80 -> 1.1.1.200 3128 tcp" |
|
| * where tlp0 is both on the 1.1.1.0/24 network and is the |
|
| * default route for hosts on 1.1.1.0/24. Of course this |
|
| * also requires a "map tlp0 ..." to complete the story. |
|
| * One might argue whether or not this kind of network config. |
|
| * should be supported in this manner... |
|
| */ |
|
| srcrt = (odst.s_addr != ip->ip_dst.s_addr); |
srcrt = (odst.s_addr != ip->ip_dst.s_addr); |
| } |
} |
| #endif /* PFIL_HOOKS */ |
#endif /* PFIL_HOOKS */ |
| Line 733 ip_input(struct mbuf *m) |
|
| Line 664 ip_input(struct mbuf *m) |
|
| if (ia != NULL) |
if (ia != NULL) |
| goto ours; |
goto ours; |
| if (m->m_pkthdr.rcvif->if_flags & IFF_BROADCAST) { |
if (m->m_pkthdr.rcvif->if_flags & IFF_BROADCAST) { |
| IFADDR_FOREACH(ifa, m->m_pkthdr.rcvif) { |
TAILQ_FOREACH(ifa, &m->m_pkthdr.rcvif->if_addrlist, ifa_list) { |
| if (ifa->ifa_addr->sa_family != AF_INET) |
if (ifa->ifa_addr->sa_family != AF_INET) |
| continue; |
continue; |
| ia = ifatoia(ifa); |
ia = ifatoia(ifa); |
| Line 759 ip_input(struct mbuf *m) |
|
| Line 690 ip_input(struct mbuf *m) |
|
| #ifdef MROUTING |
#ifdef MROUTING |
| extern struct socket *ip_mrouter; |
extern struct socket *ip_mrouter; |
| |
|
| |
if (M_READONLY(m)) { |
| |
if ((m = m_pullup(m, hlen)) == 0) { |
| |
ipstat.ips_toosmall++; |
| |
return; |
| |
} |
| |
ip = mtod(m, struct ip *); |
| |
} |
| |
|
| if (ip_mrouter) { |
if (ip_mrouter) { |
| /* |
/* |
| * If we are acting as a multicast router, all |
* If we are acting as a multicast router, all |
| Line 828 ip_input(struct mbuf *m) |
|
| Line 767 ip_input(struct mbuf *m) |
|
| goto bad; |
goto bad; |
| } |
} |
| #endif |
#endif |
| #ifdef FAST_IPSEC |
|
| mtag = m_tag_find(m, PACKET_TAG_IPSEC_IN_DONE, NULL); |
|
| s = splsoftnet(); |
|
| if (mtag != NULL) { |
|
| tdbi = (struct tdb_ident *)(mtag + 1); |
|
| sp = ipsec_getpolicy(tdbi, IPSEC_DIR_INBOUND); |
|
| } else { |
|
| sp = ipsec_getpolicybyaddr(m, IPSEC_DIR_INBOUND, |
|
| IP_FORWARDING, &error); |
|
| } |
|
| if (sp == NULL) { /* NB: can happen if error */ |
|
| splx(s); |
|
| /*XXX error stat???*/ |
|
| DPRINTF(("ip_input: no SP for forwarding\n")); /*XXX*/ |
|
| goto bad; |
|
| } |
|
| |
|
| /* |
|
| * Check security policy against packet attributes. |
|
| */ |
|
| error = ipsec_in_reject(sp, m); |
|
| KEY_FREESP(&sp); |
|
| splx(s); |
|
| if (error) { |
|
| ipstat.ips_cantforward++; |
|
| goto bad; |
|
| } |
|
| |
|
| /* |
|
| * Peek at the outbound SP for this packet to determine if |
|
| * it's a Fast Forward candidate. |
|
| */ |
|
| mtag = m_tag_find(m, PACKET_TAG_IPSEC_PENDING_TDB, NULL); |
|
| if (mtag != NULL) |
|
| m->m_flags &= ~M_CANFASTFWD; |
|
| else { |
|
| s = splsoftnet(); |
|
| sp = ipsec4_checkpolicy(m, IPSEC_DIR_OUTBOUND, |
|
| (IP_FORWARDING | |
|
| (ip_directedbcast ? IP_ALLOWBROADCAST : 0)), |
|
| &error, NULL); |
|
| if (sp != NULL) { |
|
| m->m_flags &= ~M_CANFASTFWD; |
|
| KEY_FREESP(&sp); |
|
| } |
|
| splx(s); |
|
| } |
|
| #endif /* FAST_IPSEC */ |
|
| |
|
| ip_forward(m, srcrt); |
ip_forward(m, srcrt); |
| } |
} |
|
|
| * but it's not worth the time; just let them time out.) |
* but it's not worth the time; just let them time out.) |
| */ |
*/ |
| if (ip->ip_off & ~htons(IP_DF|IP_RF)) { |
if (ip->ip_off & ~htons(IP_DF|IP_RF)) { |
| |
if (M_READONLY(m)) { |
| |
if ((m = m_pullup(m, hlen)) == NULL) { |
| |
ipstat.ips_toosmall++; |
| |
goto bad; |
| |
} |
| |
ip = mtod(m, struct ip *); |
| |
} |
| |
|
| /* |
/* |
| * Look for queue of fragments |
* Look for queue of fragments |
| * of this datagram. |
* of this datagram. |
| */ |
*/ |
| IPQ_LOCK(); |
IPQ_LOCK(); |
| hash = IPREASS_HASH(ip->ip_src.s_addr, ip->ip_id); |
LIST_FOREACH(fp, &ipq, ipq_q) |
| /* XXX LIST_FOREACH(fp, &ipq[hash], ipq_q) */ |
|
| for (fp = LIST_FIRST(&ipq[hash]); fp != NULL; |
|
| fp = LIST_NEXT(fp, ipq_q)) { |
|
| if (ip->ip_id == fp->ipq_id && |
if (ip->ip_id == fp->ipq_id && |
| in_hosteq(ip->ip_src, fp->ipq_src) && |
in_hosteq(ip->ip_src, fp->ipq_src) && |
| in_hosteq(ip->ip_dst, fp->ipq_dst) && |
in_hosteq(ip->ip_dst, fp->ipq_dst) && |
| ip->ip_p == fp->ipq_p) |
ip->ip_p == fp->ipq_p) |
| goto found; |
goto found; |
| |
|
| } |
|
| fp = 0; |
fp = 0; |
| found: |
found: |
| |
|
|
|
| ipqe->ipqe_mff = mff; |
ipqe->ipqe_mff = mff; |
| ipqe->ipqe_m = m; |
ipqe->ipqe_m = m; |
| ipqe->ipqe_ip = ip; |
ipqe->ipqe_ip = ip; |
| m = ip_reass(ipqe, fp, &ipq[hash]); |
m = ip_reass(ipqe, fp); |
| if (m == 0) { |
if (m == 0) { |
| IPQ_UNLOCK(); |
IPQ_UNLOCK(); |
| return; |
return; |
|
|
| IPQ_UNLOCK(); |
IPQ_UNLOCK(); |
| } |
} |
| |
|
| #if defined(IPSEC) |
#ifdef IPSEC |
| /* |
/* |
| * enforce IPsec policy checking if we are seeing last header. |
* enforce IPsec policy checking if we are seeing last header. |
| * note that we do not visit this with protocols with pcb layer |
* note that we do not visit this with protocols with pcb layer |
|
|
| goto bad; |
goto bad; |
| } |
} |
| #endif |
#endif |
| #if FAST_IPSEC |
|
| /* |
|
| * enforce IPsec policy checking if we are seeing last header. |
|
| * note that we do not visit this with protocols with pcb layer |
|
| * code - like udp/tcp/raw ip. |
|
| */ |
|
| if ((inetsw[ip_protox[ip->ip_p]].pr_flags & PR_LASTHDR) != 0) { |
|
| /* |
|
| * Check if the packet has already had IPsec processing |
|
| * done. If so, then just pass it along. This tag gets |
|
| * set during AH, ESP, etc. input handling, before the |
|
| * packet is returned to the ip input queue for delivery. |
|
| */ |
|
| mtag = m_tag_find(m, PACKET_TAG_IPSEC_IN_DONE, NULL); |
|
| s = splsoftnet(); |
|
| if (mtag != NULL) { |
|
| tdbi = (struct tdb_ident *)(mtag + 1); |
|
| sp = ipsec_getpolicy(tdbi, IPSEC_DIR_INBOUND); |
|
| } else { |
|
| sp = ipsec_getpolicybyaddr(m, IPSEC_DIR_INBOUND, |
|
| IP_FORWARDING, &error); |
|
| } |
|
| if (sp != NULL) { |
|
| /* |
|
| * Check security policy against packet attributes. |
|
| */ |
|
| error = ipsec_in_reject(sp, m); |
|
| KEY_FREESP(&sp); |
|
| } else { |
|
| /* XXX error stat??? */ |
|
| error = EINVAL; |
|
| DPRINTF(("ip_input: no SP, packet discarded\n"));/*XXX*/ |
|
| goto bad; |
|
| } |
|
| splx(s); |
|
| if (error) |
|
| goto bad; |
|
| } |
|
| #endif /* FAST_IPSEC */ |
|
| |
|
| /* |
/* |
| * Switch out to protocol's input routine. |
* Switch out to protocol's input routine. |
|
|
| * is given as fp; otherwise have to make a chain. |
* is given as fp; otherwise have to make a chain. |
| */ |
*/ |
| struct mbuf * |
struct mbuf * |
| ip_reass(struct ipqent *ipqe, struct ipq *fp, struct ipqhead *ipqhead) |
ip_reass(ipqe, fp) |
| |
struct ipqent *ipqe; |
| |
struct ipq *fp; |
| { |
{ |
| struct mbuf *m = ipqe->ipqe_m; |
struct mbuf *m = ipqe->ipqe_m; |
| struct ipqent *nq, *p, *q; |
struct ipqent *nq, *p, *q; |
| Line 1062 ip_reass(struct ipqent *ipqe, struct ipq |
|
| Line 918 ip_reass(struct ipqent *ipqe, struct ipq |
|
| m->m_data += hlen; |
m->m_data += hlen; |
| m->m_len -= hlen; |
m->m_len -= hlen; |
| |
|
| #ifdef notyet |
|
| /* make sure fragment limit is up-to-date */ |
|
| CHECK_NMBCLUSTER_PARAMS(); |
|
| |
|
| /* If we have too many fragments, drop the older half. */ |
|
| if (ip_nfrags >= ip_maxfrags) |
|
| ip_reass_drophalf(void); |
|
| #endif |
|
| |
|
| /* |
|
| * We are about to add a fragment; increment frag count. |
|
| */ |
|
| ip_nfrags++; |
|
| |
|
| /* |
/* |
| * If first fragment to arrive, create a reassembly queue. |
* If first fragment to arrive, create a reassembly queue. |
| */ |
*/ |
| Line 1095 ip_reass(struct ipqent *ipqe, struct ipq |
|
| Line 937 ip_reass(struct ipqent *ipqe, struct ipq |
|
| M_FTABLE, M_NOWAIT); |
M_FTABLE, M_NOWAIT); |
| if (fp == NULL) |
if (fp == NULL) |
| goto dropfrag; |
goto dropfrag; |
| LIST_INSERT_HEAD(ipqhead, fp, ipq_q); |
LIST_INSERT_HEAD(&ipq, fp, ipq_q); |
| fp->ipq_nfrags = 1; |
|
| fp->ipq_ttl = IPFRAGTTL; |
fp->ipq_ttl = IPFRAGTTL; |
| fp->ipq_p = ipqe->ipqe_ip->ip_p; |
fp->ipq_p = ipqe->ipqe_ip->ip_p; |
| fp->ipq_id = ipqe->ipqe_ip->ip_id; |
fp->ipq_id = ipqe->ipqe_ip->ip_id; |
| Line 1105 ip_reass(struct ipqent *ipqe, struct ipq |
|
| Line 946 ip_reass(struct ipqent *ipqe, struct ipq |
|
| fp->ipq_dst = ipqe->ipqe_ip->ip_dst; |
fp->ipq_dst = ipqe->ipqe_ip->ip_dst; |
| p = NULL; |
p = NULL; |
| goto insert; |
goto insert; |
| } else { |
|
| fp->ipq_nfrags++; |
|
| } |
} |
| |
|
| /* |
/* |
| Line 1157 ip_reass(struct ipqent *ipqe, struct ipq |
|
| Line 996 ip_reass(struct ipqent *ipqe, struct ipq |
|
| m_freem(q->ipqe_m); |
m_freem(q->ipqe_m); |
| TAILQ_REMOVE(&fp->ipq_fragq, q, ipqe_q); |
TAILQ_REMOVE(&fp->ipq_fragq, q, ipqe_q); |
| pool_put(&ipqent_pool, q); |
pool_put(&ipqent_pool, q); |
| fp->ipq_nfrags--; |
|
| ip_nfrags--; |
|
| } |
} |
| |
|
| insert: |
insert: |
|
|
| pool_put(&ipqent_pool, q); |
pool_put(&ipqent_pool, q); |
| m_cat(m, t); |
m_cat(m, t); |
| } |
} |
| ip_nfrags -= fp->ipq_nfrags; |
|
| |
|
| /* |
/* |
| * Create header for new ip packet by |
* Create header for new ip packet by |
|
|
| for (t = m; t; t = t->m_next) |
for (t = m; t; t = t->m_next) |
| plen += t->m_len; |
plen += t->m_len; |
| m->m_pkthdr.len = plen; |
m->m_pkthdr.len = plen; |
| m->m_pkthdr.csum_flags = 0; |
|
| } |
} |
| return (m); |
return (m); |
| |
|
| dropfrag: |
dropfrag: |
| if (fp != 0) |
|
| fp->ipq_nfrags--; |
|
| ip_nfrags--; |
|
| ipstat.ips_fragdropped++; |
ipstat.ips_fragdropped++; |
| m_freem(m); |
m_freem(m); |
| pool_put(&ipqent_pool, ipqe); |
pool_put(&ipqent_pool, ipqe); |
|
|
| * associated datagrams. |
* associated datagrams. |
| */ |
*/ |
| void |
void |
| ip_freef(struct ipq *fp) |
ip_freef(fp) |
| |
struct ipq *fp; |
| { |
{ |
| struct ipqent *q, *p; |
struct ipqent *q, *p; |
| u_int nfrags = 0; |
|
| |
|
| IPQ_LOCK_CHECK(); |
IPQ_LOCK_CHECK(); |
| |
|
| for (q = TAILQ_FIRST(&fp->ipq_fragq); q != NULL; q = p) { |
for (q = TAILQ_FIRST(&fp->ipq_fragq); q != NULL; q = p) { |
| p = TAILQ_NEXT(q, ipqe_q); |
p = TAILQ_NEXT(q, ipqe_q); |
| m_freem(q->ipqe_m); |
m_freem(q->ipqe_m); |
| nfrags++; |
|
| TAILQ_REMOVE(&fp->ipq_fragq, q, ipqe_q); |
TAILQ_REMOVE(&fp->ipq_fragq, q, ipqe_q); |
| pool_put(&ipqent_pool, q); |
pool_put(&ipqent_pool, q); |
| } |
} |
| |
|
| if (nfrags != fp->ipq_nfrags) |
|
| printf("ip_freef: nfrags %d != %d\n", fp->ipq_nfrags, nfrags); |
|
| ip_nfrags -= nfrags; |
|
| LIST_REMOVE(fp, ipq_q); |
LIST_REMOVE(fp, ipq_q); |
| FREE(fp, M_FTABLE); |
FREE(fp, M_FTABLE); |
| ip_nfragpackets--; |
ip_nfragpackets--; |
| } |
} |
| |
|
| /* |
/* |
| * IP reassembly TTL machinery for multiplicative drop. |
|
| */ |
|
| static u_int fragttl_histo[(IPFRAGTTL+1)]; |
|
| |
|
| |
|
| /* |
|
| * Decrement TTL of all reasembly queue entries by `ticks'. |
|
| * Count number of distinct fragments (as opposed to partial, fragmented |
|
| * datagrams) in the reassembly queue. While we traverse the entire |
|
| * reassembly queue, compute and return the median TTL over all fragments. |
|
| */ |
|
| static u_int |
|
| ip_reass_ttl_decr(u_int ticks) |
|
| { |
|
| u_int nfrags, median, dropfraction, keepfraction; |
|
| struct ipq *fp, *nfp; |
|
| int i; |
|
| |
|
| nfrags = 0; |
|
| memset(fragttl_histo, 0, sizeof fragttl_histo); |
|
| |
|
| for (i = 0; i < IPREASS_NHASH; i++) { |
|
| for (fp = LIST_FIRST(&ipq[i]); fp != NULL; fp = nfp) { |
|
| fp->ipq_ttl = ((fp->ipq_ttl <= ticks) ? |
|
| 0 : fp->ipq_ttl - ticks); |
|
| nfp = LIST_NEXT(fp, ipq_q); |
|
| if (fp->ipq_ttl == 0) { |
|
| ipstat.ips_fragtimeout++; |
|
| ip_freef(fp); |
|
| } else { |
|
| nfrags += fp->ipq_nfrags; |
|
| fragttl_histo[fp->ipq_ttl] += fp->ipq_nfrags; |
|
| } |
|
| } |
|
| } |
|
| |
|
| KASSERT(ip_nfrags == nfrags); |
|
| |
|
| /* Find median (or other drop fraction) in histogram. */ |
|
| dropfraction = (ip_nfrags / 2); |
|
| keepfraction = ip_nfrags - dropfraction; |
|
| for (i = IPFRAGTTL, median = 0; i >= 0; i--) { |
|
| median += fragttl_histo[i]; |
|
| if (median >= keepfraction) |
|
| break; |
|
| } |
|
| |
|
| /* Return TTL of median (or other fraction). */ |
|
| return (u_int)i; |
|
| } |
|
| |
|
| void |
|
| ip_reass_drophalf(void) |
|
| { |
|
| |
|
| u_int median_ticks; |
|
| /* |
|
| * Compute median TTL of all fragments, and count frags |
|
| * with that TTL or lower (roughly half of all fragments). |
|
| */ |
|
| median_ticks = ip_reass_ttl_decr(0); |
|
| |
|
| /* Drop half. */ |
|
| median_ticks = ip_reass_ttl_decr(median_ticks); |
|
| |
|
| } |
|
| |
|
| /* |
|
| * IP timer processing; |
* IP timer processing; |
| * if a timer expires on a reassembly |
* if a timer expires on a reassembly |
| * queue, discard it. |
* queue, discard it. |
| */ |
*/ |
| void |
void |
| ip_slowtimo(void) |
ip_slowtimo() |
| { |
{ |
| static u_int dropscanidx = 0; |
struct ipq *fp, *nfp; |
| u_int i; |
|
| u_int median_ttl; |
|
| int s = splsoftnet(); |
int s = splsoftnet(); |
| |
|
| IPQ_LOCK(); |
IPQ_LOCK(); |
| |
for (fp = LIST_FIRST(&ipq); fp != NULL; fp = nfp) { |
| /* Age TTL of all fragments by 1 tick .*/ |
nfp = LIST_NEXT(fp, ipq_q); |
| median_ttl = ip_reass_ttl_decr(1); |
if (--fp->ipq_ttl == 0) { |
| |
ipstat.ips_fragtimeout++; |
| /* make sure fragment limit is up-to-date */ |
ip_freef(fp); |
| CHECK_NMBCLUSTER_PARAMS(); |
} |
| |
} |
| /* If we have too many fragments, drop the older half. */ |
|
| if (ip_nfrags > ip_maxfrags) |
|
| ip_reass_ttl_decr(median_ttl); |
|
| |
|
| /* |
/* |
| * If we are over the maximum number of fragmented packets |
* If we are over the maximum number of fragments |
| * (due to the limit being lowered), drain off |
* (due to the limit being lowered), drain off |
| * enough to get down to the new limit. Start draining |
* enough to get down to the new limit. |
| * from the reassembly hashqueue most recently drained. |
|
| */ |
*/ |
| if (ip_maxfragpackets < 0) |
if (ip_maxfragpackets < 0) |
| ; |
; |
| else { |
else { |
| int wrapped = 0; |
while (ip_nfragpackets > ip_maxfragpackets && LIST_FIRST(&ipq)) |
| |
ip_freef(LIST_FIRST(&ipq)); |
| i = dropscanidx; |
|
| while (ip_nfragpackets > ip_maxfragpackets && wrapped == 0) { |
|
| while (LIST_FIRST(&ipq[i]) != NULL) |
|
| ip_freef(LIST_FIRST(&ipq[i])); |
|
| if (++i >= IPREASS_NHASH) { |
|
| i = 0; |
|
| } |
|
| /* |
|
| * Dont scan forever even if fragment counters are |
|
| * wrong: stop after scanning entire reassembly queue. |
|
| */ |
|
| if (i == dropscanidx) |
|
| wrapped = 1; |
|
| } |
|
| dropscanidx = i; |
|
| } |
} |
| IPQ_UNLOCK(); |
IPQ_UNLOCK(); |
| #ifdef GATEWAY |
#ifdef GATEWAY |
| Line 1399 ip_slowtimo(void) |
|
| Line 1136 ip_slowtimo(void) |
|
| * Drain off all datagram fragments. |
* Drain off all datagram fragments. |
| */ |
*/ |
| void |
void |
| ip_drain(void) |
ip_drain() |
| { |
{ |
| |
|
| /* |
/* |
|
|
| if (ipq_lock_try() == 0) |
if (ipq_lock_try() == 0) |
| return; |
return; |
| |
|
| /* |
while (LIST_FIRST(&ipq) != NULL) { |
| * Drop half the total fragments now. If more mbufs are needed, |
ipstat.ips_fragdropped++; |
| * we will be called again soon. |
ip_freef(LIST_FIRST(&ipq)); |
| */ |
} |
| ip_reass_drophalf(); |
|
| |
|
| IPQ_UNLOCK(); |
IPQ_UNLOCK(); |
| } |
} |
|
|
| * 0 if the packet should be processed further. |
* 0 if the packet should be processed further. |
| */ |
*/ |
| int |
int |
| ip_dooptions(struct mbuf *m) |
ip_dooptions(m) |
| |
struct mbuf *m; |
| { |
{ |
| struct ip *ip = mtod(m, struct ip *); |
struct ip *ip = mtod(m, struct ip *); |
| u_char *cp, *cp0; |
u_char *cp, *cp0; |
| Line 1513 ip_dooptions(struct mbuf *m) |
|
| Line 1250 ip_dooptions(struct mbuf *m) |
|
| bcopy((caddr_t)(cp + off), (caddr_t)&ipaddr.sin_addr, |
bcopy((caddr_t)(cp + off), (caddr_t)&ipaddr.sin_addr, |
| sizeof(ipaddr.sin_addr)); |
sizeof(ipaddr.sin_addr)); |
| if (opt == IPOPT_SSRR) |
if (opt == IPOPT_SSRR) |
| ia = ifatoia(ifa_ifwithladdr(sintosa(&ipaddr))); |
ia = ifatoia(ifa_ifwithaddr(sintosa(&ipaddr))); |
| else |
else |
| ia = ip_rtaddr(ipaddr.sin_addr); |
ia = ip_rtaddr(ipaddr.sin_addr); |
| if (ia == 0) { |
if (ia == 0) { |
|
|
| * return internet address info of interface to be used to get there. |
* return internet address info of interface to be used to get there. |
| */ |
*/ |
| struct in_ifaddr * |
struct in_ifaddr * |
| ip_rtaddr(struct in_addr dst) |
ip_rtaddr(dst) |
| |
struct in_addr dst; |
| { |
{ |
| struct sockaddr_in *sin; |
struct sockaddr_in *sin; |
| |
|
| Line 1682 ip_rtaddr(struct in_addr dst) |
|
| Line 1420 ip_rtaddr(struct in_addr dst) |
|
| * to be picked up later by ip_srcroute if the receiver is interested. |
* to be picked up later by ip_srcroute if the receiver is interested. |
| */ |
*/ |
| void |
void |
| save_rte(u_char *option, struct in_addr dst) |
save_rte(option, dst) |
| |
u_char *option; |
| |
struct in_addr dst; |
| { |
{ |
| unsigned olen; |
unsigned olen; |
| |
|
| Line 1704 save_rte(u_char *option, struct in_addr |
|
| Line 1444 save_rte(u_char *option, struct in_addr |
|
| * The first hop is placed before the options, will be removed later. |
* The first hop is placed before the options, will be removed later. |
| */ |
*/ |
| struct mbuf * |
struct mbuf * |
| ip_srcroute(void) |
ip_srcroute() |
| { |
{ |
| struct in_addr *p, *q; |
struct in_addr *p, *q; |
| struct mbuf *m; |
struct mbuf *m; |
| Line 1776 ip_srcroute(void) |
|
| Line 1516 ip_srcroute(void) |
|
| * XXX should be deleted; last arg currently ignored. |
* XXX should be deleted; last arg currently ignored. |
| */ |
*/ |
| void |
void |
| ip_stripoptions(struct mbuf *m, struct mbuf *mopt) |
ip_stripoptions(m, mopt) |
| |
struct mbuf *m; |
| |
struct mbuf *mopt; |
| { |
{ |
| int i; |
int i; |
| struct ip *ip = mtod(m, struct ip *); |
struct ip *ip = mtod(m, struct ip *); |
| Line 1818 const int inetctlerrmap[PRC_NCMDS] = { |
|
| Line 1560 const int inetctlerrmap[PRC_NCMDS] = { |
|
| * via a source route. |
* via a source route. |
| */ |
*/ |
| void |
void |
| ip_forward(struct mbuf *m, int srcrt) |
ip_forward(m, srcrt) |
| |
struct mbuf *m; |
| |
int srcrt; |
| { |
{ |
| struct ip *ip = mtod(m, struct ip *); |
struct ip *ip = mtod(m, struct ip *); |
| struct sockaddr_in *sin; |
struct sockaddr_in *sin; |
| struct rtentry *rt; |
struct rtentry *rt; |
| int error, type = 0, code = 0, destmtu = 0; |
int error, type = 0, code = 0; |
| struct mbuf *mcopy; |
struct mbuf *mcopy; |
| n_long dest; |
n_long dest; |
| |
struct ifnet *destifp; |
| |
#ifdef IPSEC |
| |
struct ifnet dummyifp; |
| |
#endif |
| |
|
| /* |
/* |
| * We are now in the output path. |
* We are now in the output path. |
| Line 1853 ip_forward(struct mbuf *m, int srcrt) |
|
| Line 1601 ip_forward(struct mbuf *m, int srcrt) |
|
| icmp_error(m, ICMP_TIMXCEED, ICMP_TIMXCEED_INTRANS, dest, 0); |
icmp_error(m, ICMP_TIMXCEED, ICMP_TIMXCEED_INTRANS, dest, 0); |
| return; |
return; |
| } |
} |
| |
ip->ip_ttl -= IPTTLDEC; |
| |
|
| sin = satosin(&ipforward_rt.ro_dst); |
sin = satosin(&ipforward_rt.ro_dst); |
| if ((rt = ipforward_rt.ro_rt) == 0 || |
if ((rt = ipforward_rt.ro_rt) == 0 || |
| Line 1867 ip_forward(struct mbuf *m, int srcrt) |
|
| Line 1616 ip_forward(struct mbuf *m, int srcrt) |
|
| |
|
| rtalloc(&ipforward_rt); |
rtalloc(&ipforward_rt); |
| if (ipforward_rt.ro_rt == 0) { |
if (ipforward_rt.ro_rt == 0) { |
| icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_NET, dest, 0); |
icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_HOST, dest, 0); |
| return; |
return; |
| } |
} |
| rt = ipforward_rt.ro_rt; |
rt = ipforward_rt.ro_rt; |
| Line 1882 ip_forward(struct mbuf *m, int srcrt) |
|
| Line 1631 ip_forward(struct mbuf *m, int srcrt) |
|
| if (mcopy) |
if (mcopy) |
| mcopy = m_pullup(mcopy, ip->ip_hl << 2); |
mcopy = m_pullup(mcopy, ip->ip_hl << 2); |
| |
|
| ip->ip_ttl -= IPTTLDEC; |
|
| |
|
| /* |
/* |
| * If forwarding packet using same interface that it came in on, |
* If forwarding packet using same interface that it came in on, |
| * perhaps should send a redirect to sender to shortcut a hop. |
* perhaps should send a redirect to sender to shortcut a hop. |
| Line 1917 ip_forward(struct mbuf *m, int srcrt) |
|
| Line 1664 ip_forward(struct mbuf *m, int srcrt) |
|
| } |
} |
| } |
} |
| |
|
| |
#ifdef IPSEC |
| |
/* Don't lookup socket in forwarding case */ |
| |
(void)ipsec_setsocket(m, NULL); |
| |
#endif |
| error = ip_output(m, (struct mbuf *)0, &ipforward_rt, |
error = ip_output(m, (struct mbuf *)0, &ipforward_rt, |
| (IP_FORWARDING | (ip_directedbcast ? IP_ALLOWBROADCAST : 0)), |
(IP_FORWARDING | (ip_directedbcast ? IP_ALLOWBROADCAST : 0)), 0); |
| (struct ip_moptions *)NULL, (struct socket *)NULL); |
|
| |
|
| if (error) |
if (error) |
| ipstat.ips_cantforward++; |
ipstat.ips_cantforward++; |
| else { |
else { |
| Line 1940 ip_forward(struct mbuf *m, int srcrt) |
|
| Line 1689 ip_forward(struct mbuf *m, int srcrt) |
|
| } |
} |
| if (mcopy == NULL) |
if (mcopy == NULL) |
| return; |
return; |
| |
destifp = NULL; |
| |
|
| switch (error) { |
switch (error) { |
| |
|
| Line 1959 ip_forward(struct mbuf *m, int srcrt) |
|
| Line 1709 ip_forward(struct mbuf *m, int srcrt) |
|
| case EMSGSIZE: |
case EMSGSIZE: |
| type = ICMP_UNREACH; |
type = ICMP_UNREACH; |
| code = ICMP_UNREACH_NEEDFRAG; |
code = ICMP_UNREACH_NEEDFRAG; |
| #if !defined(IPSEC) && !defined(FAST_IPSEC) |
#ifndef IPSEC |
| if (ipforward_rt.ro_rt) |
if (ipforward_rt.ro_rt) |
| destmtu = ipforward_rt.ro_rt->rt_ifp->if_mtu; |
destifp = ipforward_rt.ro_rt->rt_ifp; |
| #else |
#else |
| /* |
/* |
| * If the packet is routed over IPsec tunnel, tell the |
* If the packet is routed over IPsec tunnel, tell the |
| Line 1980 ip_forward(struct mbuf *m, int srcrt) |
|
| Line 1730 ip_forward(struct mbuf *m, int srcrt) |
|
| &ipsecerror); |
&ipsecerror); |
| |
|
| if (sp == NULL) |
if (sp == NULL) |
| destmtu = ipforward_rt.ro_rt->rt_ifp->if_mtu; |
destifp = ipforward_rt.ro_rt->rt_ifp; |
| else { |
else { |
| /* count IPsec header size */ |
/* count IPsec header size */ |
| ipsechdr = ipsec4_hdrsiz(mcopy, |
ipsechdr = ipsec4_hdrsiz(mcopy, |
| Line 1989 ip_forward(struct mbuf *m, int srcrt) |
|
| Line 1739 ip_forward(struct mbuf *m, int srcrt) |
|
| /* |
/* |
| * find the correct route for outer IPv4 |
* find the correct route for outer IPv4 |
| * header, compute tunnel MTU. |
* header, compute tunnel MTU. |
| |
* |
| |
* XXX BUG ALERT |
| |
* The "dummyifp" code relies upon the fact |
| |
* that icmp_error() touches only ifp->if_mtu. |
| */ |
*/ |
| |
/*XXX*/ |
| |
destifp = NULL; |
| if (sp->req != NULL |
if (sp->req != NULL |
| && sp->req->sav != NULL |
&& sp->req->sav != NULL |
| && sp->req->sav->sah != NULL) { |
&& sp->req->sav->sah != NULL) { |
| ro = &sp->req->sav->sah->sa_route; |
ro = &sp->req->sav->sah->sa_route; |
| if (ro->ro_rt && ro->ro_rt->rt_ifp) { |
if (ro->ro_rt && ro->ro_rt->rt_ifp) { |
| destmtu = |
dummyifp.if_mtu = |
| ro->ro_rt->rt_rmx.rmx_mtu ? |
ro->ro_rt->rt_rmx.rmx_mtu ? |
| ro->ro_rt->rt_rmx.rmx_mtu : |
ro->ro_rt->rt_rmx.rmx_mtu : |
| ro->ro_rt->rt_ifp->if_mtu; |
ro->ro_rt->rt_ifp->if_mtu; |
| destmtu -= ipsechdr; |
dummyifp.if_mtu -= ipsechdr; |
| |
destifp = &dummyifp; |
| } |
} |
| } |
} |
| |
|
| #ifdef IPSEC |
|
| key_freesp(sp); |
key_freesp(sp); |
| #else |
|
| KEY_FREESP(&sp); |
|
| #endif |
|
| } |
} |
| } |
} |
| #endif /*IPSEC*/ |
#endif /*IPSEC*/ |
| Line 2032 ip_forward(struct mbuf *m, int srcrt) |
|
| Line 1784 ip_forward(struct mbuf *m, int srcrt) |
|
| break; |
break; |
| #endif |
#endif |
| } |
} |
| icmp_error(mcopy, type, code, dest, destmtu); |
icmp_error(mcopy, type, code, dest, destifp); |
| } |
} |
| |
|
| void |
void |
| ip_savecontrol(struct inpcb *inp, struct mbuf **mp, struct ip *ip, |
ip_savecontrol(inp, mp, ip, m) |
| struct mbuf *m) |
struct inpcb *inp; |
| |
struct mbuf **mp; |
| |
struct ip *ip; |
| |
struct mbuf *m; |
| { |
{ |
| |
|
| if (inp->inp_socket->so_options & SO_TIMESTAMP) { |
if (inp->inp_socket->so_options & SO_TIMESTAMP) { |
| Line 2092 ip_savecontrol(struct inpcb *inp, struct |
|
| Line 1847 ip_savecontrol(struct inpcb *inp, struct |
|
| } |
} |
| } |
} |
| |
|
| /* |
int |
| * sysctl helper routine for net.inet.ip.mtudisctimeout. checks the |
ip_sysctl(name, namelen, oldp, oldlenp, newp, newlen) |
| * range of the new value and tweaks timers if it changes. |
int *name; |
| */ |
u_int namelen; |
| static int |
void *oldp; |
| sysctl_net_inet_ip_pmtudto(SYSCTLFN_ARGS) |
size_t *oldlenp; |
| |
void *newp; |
| |
size_t newlen; |
| { |
{ |
| int error, tmp; |
extern int subnetsarelocal, hostzeroisbroadcast; |
| struct sysctlnode node; |
|
| |
|
| node = *rnode; |
|
| tmp = ip_mtudisc_timeout; |
|
| node.sysctl_data = &tmp; |
|
| error = sysctl_lookup(SYSCTLFN_CALL(&node)); |
|
| if (error || newp == NULL) |
|
| return (error); |
|
| if (tmp < 0) |
|
| return (EINVAL); |
|
| |
|
| ip_mtudisc_timeout = tmp; |
int error, old; |
| rt_timer_queue_change(ip_mtudisc_timeout_q, ip_mtudisc_timeout); |
|
| |
|
| return (0); |
|
| } |
|
| |
|
| |
/* All sysctl names at this level are terminal. */ |
| |
if (namelen != 1) |
| |
return (ENOTDIR); |
| |
|
| |
switch (name[0]) { |
| |
case IPCTL_FORWARDING: |
| |
return (sysctl_int(oldp, oldlenp, newp, newlen, &ipforwarding)); |
| |
case IPCTL_SENDREDIRECTS: |
| |
return (sysctl_int(oldp, oldlenp, newp, newlen, |
| |
&ipsendredirects)); |
| |
case IPCTL_DEFTTL: |
| |
return (sysctl_int(oldp, oldlenp, newp, newlen, &ip_defttl)); |
| |
#ifdef notyet |
| |
case IPCTL_DEFMTU: |
| |
return (sysctl_int(oldp, oldlenp, newp, newlen, &ip_mtu)); |
| |
#endif |
| |
case IPCTL_FORWSRCRT: |
| |
/* Don't allow this to change in a secure environment. */ |
| |
if (securelevel > 0) |
| |
return (sysctl_rdint(oldp, oldlenp, newp, |
| |
ip_forwsrcrt)); |
| |
else |
| |
return (sysctl_int(oldp, oldlenp, newp, newlen, |
| |
&ip_forwsrcrt)); |
| |
case IPCTL_DIRECTEDBCAST: |
| |
return (sysctl_int(oldp, oldlenp, newp, newlen, |
| |
&ip_directedbcast)); |
| |
case IPCTL_ALLOWSRCRT: |
| |
return (sysctl_int(oldp, oldlenp, newp, newlen, |
| |
&ip_allowsrcrt)); |
| |
case IPCTL_SUBNETSARELOCAL: |
| |
return (sysctl_int(oldp, oldlenp, newp, newlen, |
| |
&subnetsarelocal)); |
| |
case IPCTL_MTUDISC: |
| |
error = sysctl_int(oldp, oldlenp, newp, newlen, |
| |
&ip_mtudisc); |
| |
if (error == 0 && ip_mtudisc == 0) |
| |
rt_timer_queue_remove_all(ip_mtudisc_timeout_q, TRUE); |
| |
return error; |
| |
case IPCTL_ANONPORTMIN: |
| |
old = anonportmin; |
| |
error = sysctl_int(oldp, oldlenp, newp, newlen, &anonportmin); |
| |
if (anonportmin >= anonportmax || anonportmin < 0 |
| |
|| anonportmin > 65535 |
| |
#ifndef IPNOPRIVPORTS |
| |
|| anonportmin < IPPORT_RESERVED |
| |
#endif |
| |
) { |
| |
anonportmin = old; |
| |
return (EINVAL); |
| |
} |
| |
return (error); |
| |
case IPCTL_ANONPORTMAX: |
| |
old = anonportmax; |
| |
error = sysctl_int(oldp, oldlenp, newp, newlen, &anonportmax); |
| |
if (anonportmin >= anonportmax || anonportmax < 0 |
| |
|| anonportmax > 65535 |
| |
#ifndef IPNOPRIVPORTS |
| |
|| anonportmax < IPPORT_RESERVED |
| |
#endif |
| |
) { |
| |
anonportmax = old; |
| |
return (EINVAL); |
| |
} |
| |
return (error); |
| |
case IPCTL_MTUDISCTIMEOUT: |
| |
old = ip_mtudisc_timeout; |
| |
error = sysctl_int(oldp, oldlenp, newp, newlen, |
| |
&ip_mtudisc_timeout); |
| |
if (ip_mtudisc_timeout < 0) { |
| |
ip_mtudisc_timeout = old; |
| |
return (EINVAL); |
| |
} |
| |
if (error == 0) |
| |
rt_timer_queue_change(ip_mtudisc_timeout_q, |
| |
ip_mtudisc_timeout); |
| |
return (error); |
| #ifdef GATEWAY |
#ifdef GATEWAY |
| /* |
case IPCTL_MAXFLOWS: |
| * sysctl helper routine for net.inet.ip.maxflows. apparently if |
{ |
| * maxflows is even looked up, we "reap flows". |
int s; |
| */ |
|
| static int |
|
| sysctl_net_inet_ip_maxflows(SYSCTLFN_ARGS) |
|
| { |
|
| int s; |
|
| |
|
| s = sysctl_lookup(SYSCTLFN_CALL(rnode)); |
|
| if (s) |
|
| return (s); |
|
| |
|
| s = splsoftnet(); |
|
| ipflow_reap(0); |
|
| splx(s); |
|
| |
|
| return (0); |
|
| } |
|
| #endif /* GATEWAY */ |
|
| |
|
| |
error = sysctl_int(oldp, oldlenp, newp, newlen, |
| |
&ip_maxflows); |
| |
s = splsoftnet(); |
| |
ipflow_reap(0); |
| |
splx(s); |
| |
return (error); |
| |
} |
| |
#endif |
| |
case IPCTL_HOSTZEROBROADCAST: |
| |
return (sysctl_int(oldp, oldlenp, newp, newlen, |
| |
&hostzeroisbroadcast)); |
| |
#if NGIF > 0 |
| |
case IPCTL_GIF_TTL: |
| |
return (sysctl_int(oldp, oldlenp, newp, newlen, |
| |
&ip_gif_ttl)); |
| |
#endif |
| |
|
| SYSCTL_SETUP(sysctl_net_inet_ip_setup, "sysctl net.inet.ip subtree setup") |
#if NGRE > 0 |
| { |
case IPCTL_GRE_TTL: |
| extern int subnetsarelocal, hostzeroisbroadcast; |
return (sysctl_int(oldp, oldlenp, newp, newlen, |
| |
&ip_gre_ttl)); |
| |
#endif |
| |
|
| sysctl_createv(clog, 0, NULL, NULL, |
|
| CTLFLAG_PERMANENT, |
|
| CTLTYPE_NODE, "net", NULL, |
|
| NULL, 0, NULL, 0, |
|
| CTL_NET, CTL_EOL); |
|
| sysctl_createv(clog, 0, NULL, NULL, |
|
| CTLFLAG_PERMANENT, |
|
| CTLTYPE_NODE, "inet", |
|
| SYSCTL_DESCR("PF_INET related settings"), |
|
| NULL, 0, NULL, 0, |
|
| CTL_NET, PF_INET, CTL_EOL); |
|
| sysctl_createv(clog, 0, NULL, NULL, |
|
| CTLFLAG_PERMANENT, |
|
| CTLTYPE_NODE, "ip", |
|
| SYSCTL_DESCR("IPv4 related settings"), |
|
| NULL, 0, NULL, 0, |
|
| CTL_NET, PF_INET, IPPROTO_IP, CTL_EOL); |
|
| |
|
| sysctl_createv(clog, 0, NULL, NULL, |
|
| CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
|
| CTLTYPE_INT, "forwarding", |
|
| SYSCTL_DESCR("Enable forwarding of INET datagrams"), |
|
| NULL, 0, &ipforwarding, 0, |
|
| CTL_NET, PF_INET, IPPROTO_IP, |
|
| IPCTL_FORWARDING, CTL_EOL); |
|
| sysctl_createv(clog, 0, NULL, NULL, |
|
| CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
|
| CTLTYPE_INT, "redirect", |
|
| SYSCTL_DESCR("Enable sending of ICMP redirect messages"), |
|
| NULL, 0, &ipsendredirects, 0, |
|
| CTL_NET, PF_INET, IPPROTO_IP, |
|
| IPCTL_SENDREDIRECTS, CTL_EOL); |
|
| sysctl_createv(clog, 0, NULL, NULL, |
|
| CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
|
| CTLTYPE_INT, "ttl", |
|
| SYSCTL_DESCR("Default TTL for an INET datagram"), |
|
| NULL, 0, &ip_defttl, 0, |
|
| CTL_NET, PF_INET, IPPROTO_IP, |
|
| IPCTL_DEFTTL, CTL_EOL); |
|
| #ifdef IPCTL_DEFMTU |
|
| sysctl_createv(clog, 0, NULL, NULL, |
|
| CTLFLAG_PERMANENT /* |CTLFLAG_READWRITE? */, |
|
| CTLTYPE_INT, "mtu", |
|
| SYSCTL_DESCR("Default MTA for an INET route"), |
|
| NULL, 0, &ip_mtu, 0, |
|
| CTL_NET, PF_INET, IPPROTO_IP, |
|
| IPCTL_DEFMTU, CTL_EOL); |
|
| #endif /* IPCTL_DEFMTU */ |
|
| sysctl_createv(clog, 0, NULL, NULL, |
|
| CTLFLAG_PERMANENT|CTLFLAG_READONLY1, |
|
| CTLTYPE_INT, "forwsrcrt", |
|
| SYSCTL_DESCR("Enable forwarding of source-routed " |
|
| "datagrams"), |
|
| NULL, 0, &ip_forwsrcrt, 0, |
|
| CTL_NET, PF_INET, IPPROTO_IP, |
|
| IPCTL_FORWSRCRT, CTL_EOL); |
|
| sysctl_createv(clog, 0, NULL, NULL, |
|
| CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
|
| CTLTYPE_INT, "directed-broadcast", |
|
| SYSCTL_DESCR("Enable forwarding of broadcast datagrams"), |
|
| NULL, 0, &ip_directedbcast, 0, |
|
| CTL_NET, PF_INET, IPPROTO_IP, |
|
| IPCTL_DIRECTEDBCAST, CTL_EOL); |
|
| sysctl_createv(clog, 0, NULL, NULL, |
|
| CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
|
| CTLTYPE_INT, "allowsrcrt", |
|
| SYSCTL_DESCR("Accept source-routed datagrams"), |
|
| NULL, 0, &ip_allowsrcrt, 0, |
|
| CTL_NET, PF_INET, IPPROTO_IP, |
|
| IPCTL_ALLOWSRCRT, CTL_EOL); |
|
| sysctl_createv(clog, 0, NULL, NULL, |
|
| CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
|
| CTLTYPE_INT, "subnetsarelocal", |
|
| SYSCTL_DESCR("Whether logical subnets are considered " |
|
| "local"), |
|
| NULL, 0, &subnetsarelocal, 0, |
|
| CTL_NET, PF_INET, IPPROTO_IP, |
|
| IPCTL_SUBNETSARELOCAL, CTL_EOL); |
|
| sysctl_createv(clog, 0, NULL, NULL, |
|
| CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
|
| CTLTYPE_INT, "mtudisc", |
|
| SYSCTL_DESCR("Use RFC1191 Path MTU Discovery"), |
|
| NULL, 0, &ip_mtudisc, 0, |
|
| CTL_NET, PF_INET, IPPROTO_IP, |
|
| IPCTL_MTUDISC, CTL_EOL); |
|
| sysctl_createv(clog, 0, NULL, NULL, |
|
| CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
|
| CTLTYPE_INT, "anonportmin", |
|
| SYSCTL_DESCR("Lowest ephemeral port number to assign"), |
|
| sysctl_net_inet_ip_ports, 0, &anonportmin, 0, |
|
| CTL_NET, PF_INET, IPPROTO_IP, |
|
| IPCTL_ANONPORTMIN, CTL_EOL); |
|
| sysctl_createv(clog, 0, NULL, NULL, |
|
| CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
|
| CTLTYPE_INT, "anonportmax", |
|
| SYSCTL_DESCR("Highest ephemeral port number to assign"), |
|
| sysctl_net_inet_ip_ports, 0, &anonportmax, 0, |
|
| CTL_NET, PF_INET, IPPROTO_IP, |
|
| IPCTL_ANONPORTMAX, CTL_EOL); |
|
| sysctl_createv(clog, 0, NULL, NULL, |
|
| CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
|
| CTLTYPE_INT, "mtudisctimeout", |
|
| SYSCTL_DESCR("Lifetime of a Path MTU Discovered route"), |
|
| sysctl_net_inet_ip_pmtudto, 0, &ip_mtudisc_timeout, 0, |
|
| CTL_NET, PF_INET, IPPROTO_IP, |
|
| IPCTL_MTUDISCTIMEOUT, CTL_EOL); |
|
| #ifdef GATEWAY |
|
| sysctl_createv(clog, 0, NULL, NULL, |
|
| CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
|
| CTLTYPE_INT, "maxflows", |
|
| SYSCTL_DESCR("Number of flows for fast forwarding"), |
|
| sysctl_net_inet_ip_maxflows, 0, &ip_maxflows, 0, |
|
| CTL_NET, PF_INET, IPPROTO_IP, |
|
| IPCTL_MAXFLOWS, CTL_EOL); |
|
| #endif /* GATEWAY */ |
|
| sysctl_createv(clog, 0, NULL, NULL, |
|
| CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
|
| CTLTYPE_INT, "hostzerobroadcast", |
|
| SYSCTL_DESCR("All zeroes address is broadcast address"), |
|
| NULL, 0, &hostzeroisbroadcast, 0, |
|
| CTL_NET, PF_INET, IPPROTO_IP, |
|
| IPCTL_HOSTZEROBROADCAST, CTL_EOL); |
|
| #if NGIF > 0 |
|
| sysctl_createv(clog, 0, NULL, NULL, |
|
| CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
|
| CTLTYPE_INT, "gifttl", |
|
| SYSCTL_DESCR("Default TTL for a gif tunnel datagram"), |
|
| NULL, 0, &ip_gif_ttl, 0, |
|
| CTL_NET, PF_INET, IPPROTO_IP, |
|
| IPCTL_GIF_TTL, CTL_EOL); |
|
| #endif /* NGIF */ |
|
| #ifndef IPNOPRIVPORTS |
#ifndef IPNOPRIVPORTS |
| sysctl_createv(clog, 0, NULL, NULL, |
case IPCTL_LOWPORTMIN: |
| CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
old = lowportmin; |
| CTLTYPE_INT, "lowportmin", |
error = sysctl_int(oldp, oldlenp, newp, newlen, &lowportmin); |
| SYSCTL_DESCR("Lowest privileged ephemeral port number " |
if (lowportmin >= lowportmax |
| "to assign"), |
|| lowportmin > IPPORT_RESERVEDMAX |
| sysctl_net_inet_ip_ports, 0, &lowportmin, 0, |
|| lowportmin < IPPORT_RESERVEDMIN |
| CTL_NET, PF_INET, IPPROTO_IP, |
) { |
| IPCTL_LOWPORTMIN, CTL_EOL); |
lowportmin = old; |
| sysctl_createv(clog, 0, NULL, NULL, |
return (EINVAL); |
| CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
} |
| CTLTYPE_INT, "lowportmax", |
return (error); |
| SYSCTL_DESCR("Highest privileged ephemeral port number " |
case IPCTL_LOWPORTMAX: |
| "to assign"), |
old = lowportmax; |
| sysctl_net_inet_ip_ports, 0, &lowportmax, 0, |
error = sysctl_int(oldp, oldlenp, newp, newlen, &lowportmax); |
| CTL_NET, PF_INET, IPPROTO_IP, |
if (lowportmin >= lowportmax |
| IPCTL_LOWPORTMAX, CTL_EOL); |
|| lowportmax > IPPORT_RESERVEDMAX |
| #endif /* IPNOPRIVPORTS */ |
|| lowportmax < IPPORT_RESERVEDMIN |
| sysctl_createv(clog, 0, NULL, NULL, |
) { |
| CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
lowportmax = old; |
| CTLTYPE_INT, "maxfragpackets", |
return (EINVAL); |
| SYSCTL_DESCR("Maximum number of fragments to retain for " |
} |
| "possible reassembly"), |
return (error); |
| NULL, 0, &ip_maxfragpackets, 0, |
#endif |
| CTL_NET, PF_INET, IPPROTO_IP, |
|
| IPCTL_MAXFRAGPACKETS, CTL_EOL); |
case IPCTL_MAXFRAGPACKETS: |
| #if NGRE > 0 |
return (sysctl_int(oldp, oldlenp, newp, newlen, |
| sysctl_createv(clog, 0, NULL, NULL, |
&ip_maxfragpackets)); |
| CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
|
| CTLTYPE_INT, "grettl", |
case IPCTL_CHECKINTERFACE: |
| SYSCTL_DESCR("Default TTL for a gre tunnel datagram"), |
return (sysctl_int(oldp, oldlenp, newp, newlen, |
| NULL, 0, &ip_gre_ttl, 0, |
&ip_checkinterface)); |
| CTL_NET, PF_INET, IPPROTO_IP, |
default: |
| IPCTL_GRE_TTL, CTL_EOL); |
return (EOPNOTSUPP); |
| #endif /* NGRE */ |
} |
| sysctl_createv(clog, 0, NULL, NULL, |
/* NOTREACHED */ |
| CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
|
| CTLTYPE_INT, "checkinterface", |
|
| SYSCTL_DESCR("Enable receive side of Strong ES model " |
|
| "from RFC1122"), |
|
| NULL, 0, &ip_checkinterface, 0, |
|
| CTL_NET, PF_INET, IPPROTO_IP, |
|
| IPCTL_CHECKINTERFACE, CTL_EOL); |
|
| sysctl_createv(clog, 0, NULL, NULL, |
|
| CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
|
| CTLTYPE_INT, "random_id", |
|
| SYSCTL_DESCR("Assign random ip_id values"), |
|
| NULL, 0, &ip_do_randomid, 0, |
|
| CTL_NET, PF_INET, IPPROTO_IP, |
|
| IPCTL_RANDOMID, CTL_EOL); |
|
| sysctl_createv(clog, 0, NULL, NULL, |
|
| CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
|
| CTLTYPE_INT, "do_loopback_cksum", |
|
| SYSCTL_DESCR("Perform IP checksum on loopback"), |
|
| NULL, 0, &ip_do_loopback_cksum, 0, |
|
| CTL_NET, PF_INET, IPPROTO_IP, |
|
| IPCTL_LOOPBACKCKSUM, CTL_EOL); |
|
| sysctl_createv(clog, 0, NULL, NULL, |
|
| CTLFLAG_PERMANENT, |
|
| CTLTYPE_STRUCT, "stats", |
|
| SYSCTL_DESCR("IP statistics"), |
|
| NULL, 0, &ipstat, sizeof(ipstat), |
|
| CTL_NET, PF_INET, IPPROTO_IP, IPCTL_STATS, |
|
| CTL_EOL); |
|
| } |
} |