version 1.211, 2005/02/03 22:56:42 |
version 1.226, 2006/05/08 18:50:12 |
Line 131 __KERNEL_RCSID(0, "$NetBSD$"); |
|
Line 131 __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 200 int ipprintfs = 0; |
|
Line 201 int ipprintfs = 0; |
|
#endif |
#endif |
|
|
int ip_do_randomid = 0; |
int ip_do_randomid = 0; |
int ip_do_loopback_cksum = 0; |
|
|
|
/* |
/* |
* XXX - Setting ip_checkinterface mostly implements the receive side of |
* XXX - Setting ip_checkinterface mostly implements the receive side of |
Line 223 struct rttimer_queue *ip_mtudisc_timeout |
|
Line 223 struct rttimer_queue *ip_mtudisc_timeout |
|
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_ifaddrhead; |
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 */ |
|
|
(((((x) & 0xF) | ((((x) >> 8) & 0xF) << 4)) ^ (y)) & IPREASS_HMASK) |
(((((x) & 0xF) | ((((x) >> 8) & 0xF) << 4)) ^ (y)) & IPREASS_HMASK) |
struct ipqhead ipq[IPREASS_NHASH]; |
struct ipqhead ipq[IPREASS_NHASH]; |
int ipq_locked; |
int ipq_locked; |
static int ip_nfragpackets; /* packets in reass queue */ |
static int ip_nfragpackets; /* packets in reass queue */ |
static int ip_nfrags; /* total fragments in reass queues */ |
static int ip_nfrags; /* total fragments in reass queues */ |
|
|
int ip_maxfragpackets = 200; /* limit on packets. XXX sysctl */ |
int ip_maxfragpackets = 200; /* limit on packets. XXX sysctl */ |
Line 267 int ip_maxfrags; /* limit on fr |
|
Line 267 int ip_maxfrags; /* limit on fr |
|
/* |
/* |
* Additive-Increase/Multiplicative-Decrease (AIMD) strategy for |
* Additive-Increase/Multiplicative-Decrease (AIMD) strategy for |
* IP reassembly queue buffer managment. |
* IP reassembly queue buffer managment. |
* |
* |
* We keep a count of total IP fragments (NB: not fragmented packets!) |
* We keep a count of total IP fragments (NB: not fragmented packets!) |
* awaiting reassembly (ip_nfrags) and a limit (ip_maxfrags) on fragments. |
* awaiting reassembly (ip_nfrags) and a limit (ip_maxfrags) on fragments. |
* If ip_nfrags exceeds ip_maxfrags the limit, we drop half the |
* If ip_nfrags exceeds ip_maxfrags the limit, we drop half the |
Line 275 int ip_maxfrags; /* limit on fr |
|
Line 275 int ip_maxfrags; /* limit on fr |
|
* repeatedly deleting single packets under heavy fragmentation load |
* repeatedly deleting single packets under heavy fragmentation load |
* (e.g., from lossy NFS peers). |
* (e.g., from lossy NFS peers). |
*/ |
*/ |
static u_int ip_reass_ttl_decr(u_int ticks); |
static u_int ip_reass_ttl_decr(u_int ticks); |
static void ip_reass_drophalf(void); |
static void ip_reass_drophalf(void); |
|
|
|
|
static __inline int ipq_lock_try(void); |
static inline int ipq_lock_try(void); |
static __inline void ipq_unlock(void); |
static inline void ipq_unlock(void); |
|
|
static __inline int |
static inline int |
ipq_lock_try(void) |
ipq_lock_try(void) |
{ |
{ |
int s; |
int s; |
Line 301 ipq_lock_try(void) |
|
Line 301 ipq_lock_try(void) |
|
return (1); |
return (1); |
} |
} |
|
|
static __inline void |
static inline void |
ipq_unlock(void) |
ipq_unlock(void) |
{ |
{ |
int s; |
int s; |
Line 668 ip_input(struct mbuf *m) |
|
Line 668 ip_input(struct mbuf *m) |
|
* default route for hosts on 1.1.1.0/24. Of course this |
* default route for hosts on 1.1.1.0/24. Of course this |
* also requires a "map tlp0 ..." to complete the story. |
* also requires a "map tlp0 ..." to complete the story. |
* One might argue whether or not this kind of network config. |
* One might argue whether or not this kind of network config. |
* should be supported in this manner... |
* should be supported in this manner... |
*/ |
*/ |
srcrt = (odst.s_addr != ip->ip_dst.s_addr); |
srcrt = (odst.s_addr != ip->ip_dst.s_addr); |
} |
} |
Line 732 ip_input(struct mbuf *m) |
|
Line 732 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 && m->m_pkthdr.rcvif->if_flags & IFF_BROADCAST) { |
IFADDR_FOREACH(ifa, m->m_pkthdr.rcvif) { |
IFADDR_FOREACH(ifa, m->m_pkthdr.rcvif) { |
if (ifa->ifa_addr->sa_family != AF_INET) |
if (ifa->ifa_addr->sa_family != AF_INET) |
continue; |
continue; |
Line 836 ip_input(struct mbuf *m) |
|
Line 836 ip_input(struct mbuf *m) |
|
sp = ipsec_getpolicy(tdbi, IPSEC_DIR_INBOUND); |
sp = ipsec_getpolicy(tdbi, IPSEC_DIR_INBOUND); |
} else { |
} else { |
sp = ipsec_getpolicybyaddr(m, IPSEC_DIR_INBOUND, |
sp = ipsec_getpolicybyaddr(m, IPSEC_DIR_INBOUND, |
IP_FORWARDING, &error); |
IP_FORWARDING, &error); |
} |
} |
if (sp == NULL) { /* NB: can happen if error */ |
if (sp == NULL) { /* NB: can happen if error */ |
splx(s); |
splx(s); |
|
|
goto bad; |
goto bad; |
} |
} |
#endif |
#endif |
#if FAST_IPSEC |
#ifdef FAST_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 |
|
|
* done. If so, then just pass it along. This tag gets |
* done. If so, then just pass it along. This tag gets |
* set during AH, ESP, etc. input handling, before the |
* set during AH, ESP, etc. input handling, before the |
* packet is returned to the ip input queue for delivery. |
* packet is returned to the ip input queue for delivery. |
*/ |
*/ |
mtag = m_tag_find(m, PACKET_TAG_IPSEC_IN_DONE, NULL); |
mtag = m_tag_find(m, PACKET_TAG_IPSEC_IN_DONE, NULL); |
s = splsoftnet(); |
s = splsoftnet(); |
if (mtag != NULL) { |
if (mtag != NULL) { |
|
|
sp = ipsec_getpolicy(tdbi, IPSEC_DIR_INBOUND); |
sp = ipsec_getpolicy(tdbi, IPSEC_DIR_INBOUND); |
} else { |
} else { |
sp = ipsec_getpolicybyaddr(m, IPSEC_DIR_INBOUND, |
sp = ipsec_getpolicybyaddr(m, IPSEC_DIR_INBOUND, |
IP_FORWARDING, &error); |
IP_FORWARDING, &error); |
} |
} |
if (sp != NULL) { |
if (sp != NULL) { |
/* |
/* |
Line 1075 ip_reass(struct ipqent *ipqe, struct ipq |
|
Line 1075 ip_reass(struct ipqent *ipqe, struct ipq |
|
* We are about to add a fragment; increment frag count. |
* We are about to add a fragment; increment frag count. |
*/ |
*/ |
ip_nfrags++; |
ip_nfrags++; |
|
|
/* |
/* |
* If first fragment to arrive, create a reassembly queue. |
* If first fragment to arrive, create a reassembly queue. |
*/ |
*/ |
|
|
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); |
|
|
Line 1285 ip_reass_ttl_decr(u_int ticks) |
|
Line 1286 ip_reass_ttl_decr(u_int ticks) |
|
u_int nfrags, median, dropfraction, keepfraction; |
u_int nfrags, median, dropfraction, keepfraction; |
struct ipq *fp, *nfp; |
struct ipq *fp, *nfp; |
int i; |
int i; |
|
|
nfrags = 0; |
nfrags = 0; |
memset(fragttl_histo, 0, sizeof fragttl_histo); |
memset(fragttl_histo, 0, sizeof fragttl_histo); |
|
|
for (i = 0; i < IPREASS_NHASH; i++) { |
for (i = 0; i < IPREASS_NHASH; i++) { |
for (fp = LIST_FIRST(&ipq[i]); fp != NULL; fp = nfp) { |
for (fp = LIST_FIRST(&ipq[i]); fp != NULL; fp = nfp) { |
fp->ipq_ttl = ((fp->ipq_ttl <= ticks) ? |
fp->ipq_ttl = ((fp->ipq_ttl <= ticks) ? |
Line 1822 ip_forward(struct mbuf *m, int srcrt) |
|
Line 1823 ip_forward(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; |
int error, type = 0, code = 0, destmtu = 0; |
struct mbuf *mcopy; |
struct mbuf *mcopy; |
n_long dest; |
n_long dest; |
struct ifnet *destifp; |
|
#if defined(IPSEC) || defined(FAST_IPSEC) |
|
struct ifnet dummyifp; |
|
#endif |
|
|
|
/* |
/* |
* We are now in the output path. |
* We are now in the output path. |
Line 1842 ip_forward(struct mbuf *m, int srcrt) |
|
Line 1839 ip_forward(struct mbuf *m, int srcrt) |
|
|
|
dest = 0; |
dest = 0; |
#ifdef DIAGNOSTIC |
#ifdef DIAGNOSTIC |
if (ipprintfs) |
if (ipprintfs) { |
printf("forward: src %2.2x dst %2.2x ttl %x\n", |
printf("forward: src %s ", inet_ntoa(ip->ip_src)); |
ntohl(ip->ip_src.s_addr), |
printf("dst %s ttl %x\n", inet_ntoa(ip->ip_dst), ip->ip_ttl); |
ntohl(ip->ip_dst.s_addr), ip->ip_ttl); |
} |
#endif |
#endif |
if (m->m_flags & (M_BCAST|M_MCAST) || in_canforward(ip->ip_dst) == 0) { |
if (m->m_flags & (M_BCAST|M_MCAST) || in_canforward(ip->ip_dst) == 0) { |
ipstat.ips_cantforward++; |
ipstat.ips_cantforward++; |
Line 1856 ip_forward(struct mbuf *m, int srcrt) |
|
Line 1853 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 1871 ip_forward(struct mbuf *m, int srcrt) |
|
Line 1867 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_HOST, dest, 0); |
icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_NET, dest, 0); |
return; |
return; |
} |
} |
rt = ipforward_rt.ro_rt; |
rt = ipforward_rt.ro_rt; |
Line 1886 ip_forward(struct mbuf *m, int srcrt) |
|
Line 1882 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 1942 ip_forward(struct mbuf *m, int srcrt) |
|
Line 1940 ip_forward(struct mbuf *m, int srcrt) |
|
} |
} |
if (mcopy == NULL) |
if (mcopy == NULL) |
return; |
return; |
destifp = NULL; |
|
|
|
switch (error) { |
switch (error) { |
|
|
Line 1964 ip_forward(struct mbuf *m, int srcrt) |
|
Line 1961 ip_forward(struct mbuf *m, int srcrt) |
|
code = ICMP_UNREACH_NEEDFRAG; |
code = ICMP_UNREACH_NEEDFRAG; |
#if !defined(IPSEC) && !defined(FAST_IPSEC) |
#if !defined(IPSEC) && !defined(FAST_IPSEC) |
if (ipforward_rt.ro_rt) |
if (ipforward_rt.ro_rt) |
destifp = ipforward_rt.ro_rt->rt_ifp; |
destmtu = ipforward_rt.ro_rt->rt_ifp->if_mtu; |
#else |
#else |
/* |
/* |
* If the packet is routed over IPsec tunnel, tell the |
* If the packet is routed over IPsec tunnel, tell the |
Line 1983 ip_forward(struct mbuf *m, int srcrt) |
|
Line 1980 ip_forward(struct mbuf *m, int srcrt) |
|
&ipsecerror); |
&ipsecerror); |
|
|
if (sp == NULL) |
if (sp == NULL) |
destifp = ipforward_rt.ro_rt->rt_ifp; |
destmtu = ipforward_rt.ro_rt->rt_ifp->if_mtu; |
else { |
else { |
/* count IPsec header size */ |
/* count IPsec header size */ |
ipsechdr = ipsec4_hdrsiz(mcopy, |
ipsechdr = ipsec4_hdrsiz(mcopy, |
Line 1992 ip_forward(struct mbuf *m, int srcrt) |
|
Line 1989 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) { |
dummyifp.if_mtu = |
destmtu = |
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; |
dummyifp.if_mtu -= ipsechdr; |
destmtu -= ipsechdr; |
destifp = &dummyifp; |
|
} |
} |
} |
} |
|
|
Line 2041 ip_forward(struct mbuf *m, int srcrt) |
|
Line 2032 ip_forward(struct mbuf *m, int srcrt) |
|
break; |
break; |
#endif |
#endif |
} |
} |
icmp_error(mcopy, type, code, dest, destifp); |
icmp_error(mcopy, type, code, dest, destmtu); |
} |
} |
|
|
void |
void |
Line 2139 sysctl_net_inet_ip_maxflows(SYSCTLFN_ARG |
|
Line 2130 sysctl_net_inet_ip_maxflows(SYSCTLFN_ARG |
|
s = sysctl_lookup(SYSCTLFN_CALL(rnode)); |
s = sysctl_lookup(SYSCTLFN_CALL(rnode)); |
if (s) |
if (s) |
return (s); |
return (s); |
|
|
s = splsoftnet(); |
s = splsoftnet(); |
ipflow_reap(0); |
ipflow_reap(0); |
splx(s); |
splx(s); |
Line 2170 SYSCTL_SETUP(sysctl_net_inet_ip_setup, " |
|
Line 2161 SYSCTL_SETUP(sysctl_net_inet_ip_setup, " |
|
SYSCTL_DESCR("IPv4 related settings"), |
SYSCTL_DESCR("IPv4 related settings"), |
NULL, 0, NULL, 0, |
NULL, 0, NULL, 0, |
CTL_NET, PF_INET, IPPROTO_IP, CTL_EOL); |
CTL_NET, PF_INET, IPPROTO_IP, CTL_EOL); |
|
|
sysctl_createv(clog, 0, NULL, NULL, |
sysctl_createv(clog, 0, NULL, NULL, |
CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
CTLTYPE_INT, "forwarding", |
CTLTYPE_INT, "forwarding", |
Line 2341 SYSCTL_SETUP(sysctl_net_inet_ip_setup, " |
|
Line 2332 SYSCTL_SETUP(sysctl_net_inet_ip_setup, " |
|
NULL, 0, &ip_do_loopback_cksum, 0, |
NULL, 0, &ip_do_loopback_cksum, 0, |
CTL_NET, PF_INET, IPPROTO_IP, |
CTL_NET, PF_INET, IPPROTO_IP, |
IPCTL_LOOPBACKCKSUM, CTL_EOL); |
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); |
} |
} |