version 1.186, 2003/11/24 20:54:59 |
version 1.231.2.1, 2006/10/22 06:07:28 |
Line 121 __KERNEL_RCSID(0, "$NetBSD$"); |
|
Line 121 __KERNEL_RCSID(0, "$NetBSD$"); |
|
#include <sys/kernel.h> |
#include <sys/kernel.h> |
#include <sys/pool.h> |
#include <sys/pool.h> |
#include <sys/sysctl.h> |
#include <sys/sysctl.h> |
|
#include <sys/kauth.h> |
|
|
#include <net/if.h> |
#include <net/if.h> |
#include <net/if_dl.h> |
#include <net/if_dl.h> |
Line 131 __KERNEL_RCSID(0, "$NetBSD$"); |
|
Line 132 __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 199 int ip_mtudisc_timeout = IPMTUDISCTIMEOU |
|
Line 201 int ip_mtudisc_timeout = IPMTUDISCTIMEOU |
|
int ipprintfs = 0; |
int ipprintfs = 0; |
#endif |
#endif |
|
|
#ifdef RANDOM_IP_ID |
|
int ip_do_randomid = 0; |
int ip_do_randomid = 0; |
#endif |
|
|
|
/* |
/* |
* XXX - Setting ip_checkinterface mostly implements the receive side of |
* XXX - Setting ip_checkinterface mostly implements the receive side of |
Line 221 int ip_checkinterface = 0; |
|
Line 221 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_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 */ |
|
|
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; |
int ip_nfragpackets = 0; |
static int ip_nfragpackets; /* packets in reass queue */ |
int ip_maxfragpackets = 200; |
static int ip_nfrags; /* total fragments in reass queues */ |
|
|
static __inline int ipq_lock_try __P((void)); |
int ip_maxfragpackets = 200; /* limit on packets. XXX sysctl */ |
static __inline void ipq_unlock __P((void)); |
int ip_maxfrags; /* limit on fragments. XXX sysctl */ |
|
|
static __inline int |
|
ipq_lock_try() |
/* |
|
* 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(void); |
|
static inline void ipq_unlock(void); |
|
|
|
static inline int |
|
ipq_lock_try(void) |
{ |
{ |
int s; |
int s; |
|
|
|
|
return (1); |
return (1); |
} |
} |
|
|
static __inline void |
static inline void |
ipq_unlock() |
ipq_unlock(void) |
{ |
{ |
int s; |
int s; |
|
|
|
|
|
|
#define IPQ_UNLOCK() ipq_unlock() |
#define IPQ_UNLOCK() ipq_unlock() |
|
|
struct pool inmulti_pool; |
POOL_INIT(inmulti_pool, sizeof(struct in_multi), 0, 0, 0, "inmltpl", NULL); |
struct pool ipqent_pool; |
POOL_INIT(ipqent_pool, sizeof(struct ipqent), 0, 0, 0, "ipqepl", NULL); |
|
|
#ifdef INET_CSUM_COUNTERS |
#ifdef INET_CSUM_COUNTERS |
#include <sys/device.h> |
#include <sys/device.h> |
Line 312 struct evcnt ip_swcsum = EVCNT_INITIALIZ |
|
Line 349 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 333 static struct ip_srcrt { |
|
Line 374 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 __P((u_char *, struct in_addr)); |
static void save_rte(u_char *, struct in_addr); |
|
|
#ifdef MBUFTRACE |
#ifdef MBUFTRACE |
struct mowner ip_rx_mowner = { "internet", "rx" }; |
struct mowner ip_rx_mowner = MOWNER_INIT("internet", "rx"); |
struct mowner ip_tx_mowner = { "internet", "tx" }; |
struct mowner ip_tx_mowner = MOWNER_INIT("internet", "tx"); |
#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() |
ip_init(void) |
{ |
{ |
struct protosw *pr; |
const 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); |
|
ip_id = time.tv_sec & 0xfffff; |
for (i = 0; i < IPREASS_NHASH; i++) |
|
LIST_INIT(&ipq[i]); |
|
|
|
ip_id = time_second & 0xfffff; |
|
|
ipintrq.ifq_maxlen = ipqmaxlen; |
ipintrq.ifq_maxlen = ipqmaxlen; |
|
ip_nmbclusters_changed(); |
|
|
TAILQ_INIT(&in_ifaddrhead); |
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); |
|
|
"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); |
#endif /* MBUFTRACE */ |
#endif /* MBUFTRACE */ |
} |
} |
|
|
struct sockaddr_in ipaddr = { sizeof(ipaddr), AF_INET }; |
struct sockaddr_in ipaddr = { |
|
.sin_len = sizeof(ipaddr), |
|
.sin_family = AF_INET, |
|
}; |
struct route ipforward_rt; |
struct route ipforward_rt; |
|
|
/* |
/* |
* IP software interrupt routine |
* IP software interrupt routine |
*/ |
*/ |
void |
void |
ipintr() |
ipintr(void) |
{ |
{ |
int s; |
int s; |
struct mbuf *m; |
struct mbuf *m; |
Line 439 ip_input(struct mbuf *m) |
|
Line 488 ip_input(struct mbuf *m) |
|
int downmatch; |
int downmatch; |
int checkif; |
int checkif; |
int srcrt = 0; |
int srcrt = 0; |
|
int s; |
|
u_int hash; |
#ifdef FAST_IPSEC |
#ifdef FAST_IPSEC |
struct m_tag *mtag; |
struct m_tag *mtag; |
struct tdb_ident *tdbi; |
struct tdb_ident *tdbi; |
struct secpolicy *sp; |
struct secpolicy *sp; |
int s, error; |
int error; |
#endif /* FAST_IPSEC */ |
#endif /* FAST_IPSEC */ |
|
|
MCLAIM(m, &ip_rx_mowner); |
MCLAIM(m, &ip_rx_mowner); |
Line 527 ip_input(struct mbuf *m) |
|
Line 578 ip_input(struct mbuf *m) |
|
break; |
break; |
|
|
default: |
default: |
/* Must compute it ourselves. */ |
/* |
INET_CSUM_COUNTER_INCR(&ip_swcsum); |
* Must compute it ourselves. Maybe skip checksum on |
if (in_cksum(m, hlen) != 0) |
* loopback interfaces. |
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 563 ip_input(struct mbuf *m) |
|
Line 620 ip_input(struct mbuf *m) |
|
m_adj(m, len - m->m_pkthdr.len); |
m_adj(m, len - m->m_pkthdr.len); |
} |
} |
|
|
#ifdef IPSEC |
#if defined(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 604 ip_input(struct mbuf *m) |
|
Line 661 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 666 ip_input(struct mbuf *m) |
|
Line 737 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) { |
TAILQ_FOREACH(ifa, &m->m_pkthdr.rcvif->if_addrlist, ifa_list) { |
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; |
ia = ifatoia(ifa); |
ia = ifatoia(ifa); |
Line 693 ip_input(struct mbuf *m) |
|
Line 764 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 778 ip_input(struct mbuf *m) |
|
Line 841 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); |
Line 797 ip_input(struct mbuf *m) |
|
Line 860 ip_input(struct mbuf *m) |
|
ipstat.ips_cantforward++; |
ipstat.ips_cantforward++; |
goto bad; |
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 */ |
#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(); |
LIST_FOREACH(fp, &ipq, ipq_q) |
hash = IPREASS_HASH(ip->ip_src.s_addr, ip->ip_id); |
|
/* 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: |
|
|
|
|
*/ |
*/ |
if (mff || ip->ip_off != htons(0)) { |
if (mff || ip->ip_off != htons(0)) { |
ipstat.ips_fragments++; |
ipstat.ips_fragments++; |
|
s = splvm(); |
ipqe = pool_get(&ipqent_pool, PR_NOWAIT); |
ipqe = pool_get(&ipqent_pool, PR_NOWAIT); |
|
splx(s); |
if (ipqe == NULL) { |
if (ipqe == NULL) { |
ipstat.ips_rcvmemdrop++; |
ipstat.ips_rcvmemdrop++; |
IPQ_UNLOCK(); |
IPQ_UNLOCK(); |
|
|
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); |
m = ip_reass(ipqe, fp, &ipq[hash]); |
if (m == 0) { |
if (m == 0) { |
IPQ_UNLOCK(); |
IPQ_UNLOCK(); |
return; |
return; |
|
|
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) { |
/* |
/* |
|
|
* 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(ipqe, fp) |
ip_reass(struct ipqent *ipqe, struct ipq *fp, struct ipqhead *ipqhead) |
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; |
struct ip *ip; |
struct ip *ip; |
struct mbuf *t; |
struct mbuf *t; |
int hlen = ipqe->ipqe_ip->ip_hl << 2; |
int hlen = ipqe->ipqe_ip->ip_hl << 2; |
int i, next; |
int i, next, s; |
|
|
IPQ_LOCK_CHECK(); |
IPQ_LOCK_CHECK(); |
|
|
Line 988 ip_reass(ipqe, fp) |
|
Line 1069 ip_reass(ipqe, fp) |
|
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 1007 ip_reass(ipqe, fp) |
|
Line 1102 ip_reass(ipqe, fp) |
|
M_FTABLE, M_NOWAIT); |
M_FTABLE, M_NOWAIT); |
if (fp == NULL) |
if (fp == NULL) |
goto dropfrag; |
goto dropfrag; |
LIST_INSERT_HEAD(&ipq, fp, ipq_q); |
LIST_INSERT_HEAD(ipqhead, 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 1016 ip_reass(ipqe, fp) |
|
Line 1112 ip_reass(ipqe, fp) |
|
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 1065 ip_reass(ipqe, fp) |
|
Line 1163 ip_reass(ipqe, fp) |
|
nq = TAILQ_NEXT(q, ipqe_q); |
nq = TAILQ_NEXT(q, ipqe_q); |
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); |
|
s = splvm(); |
pool_put(&ipqent_pool, q); |
pool_put(&ipqent_pool, q); |
|
splx(s); |
|
fp->ipq_nfrags--; |
|
ip_nfrags--; |
} |
} |
|
|
insert: |
insert: |
|
|
m->m_next = 0; |
m->m_next = 0; |
m_cat(m, t); |
m_cat(m, t); |
nq = TAILQ_NEXT(q, ipqe_q); |
nq = TAILQ_NEXT(q, ipqe_q); |
|
s = splvm(); |
pool_put(&ipqent_pool, q); |
pool_put(&ipqent_pool, q); |
|
splx(s); |
for (q = nq; q != NULL; q = nq) { |
for (q = nq; q != NULL; q = nq) { |
t = q->ipqe_m; |
t = q->ipqe_m; |
nq = TAILQ_NEXT(q, ipqe_q); |
nq = TAILQ_NEXT(q, ipqe_q); |
|
s = splvm(); |
pool_put(&ipqent_pool, q); |
pool_put(&ipqent_pool, q); |
|
splx(s); |
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); |
|
s = splvm(); |
pool_put(&ipqent_pool, ipqe); |
pool_put(&ipqent_pool, ipqe); |
|
splx(s); |
return (0); |
return (0); |
} |
} |
|
|
|
|
* associated datagrams. |
* associated datagrams. |
*/ |
*/ |
void |
void |
ip_freef(fp) |
ip_freef(struct ipq *fp) |
struct ipq *fp; |
|
{ |
{ |
struct ipqent *q, *p; |
struct ipqent *q, *p; |
|
u_int nfrags = 0; |
|
int s; |
|
|
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); |
|
s = splvm(); |
pool_put(&ipqent_pool, q); |
pool_put(&ipqent_pool, q); |
|
splx(s); |
} |
} |
|
|
|
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() |
ip_slowtimo(void) |
{ |
{ |
struct ipq *fp, *nfp; |
static u_int dropscanidx = 0; |
|
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) { |
|
nfp = LIST_NEXT(fp, ipq_q); |
/* Age TTL of all fragments by 1 tick .*/ |
if (--fp->ipq_ttl == 0) { |
median_ttl = ip_reass_ttl_decr(1); |
ipstat.ips_fragtimeout++; |
|
ip_freef(fp); |
/* 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_ttl_decr(median_ttl); |
|
|
/* |
/* |
* If we are over the maximum number of fragments |
* If we are over the maximum number of fragmented packets |
* (due to the limit being lowered), drain off |
* (due to the limit being lowered), drain off |
* enough to get down to the new limit. |
* enough to get down to the new limit. Start draining |
|
* from the reassembly hashqueue most recently drained. |
*/ |
*/ |
if (ip_maxfragpackets < 0) |
if (ip_maxfragpackets < 0) |
; |
; |
else { |
else { |
while (ip_nfragpackets > ip_maxfragpackets && LIST_FIRST(&ipq)) |
int wrapped = 0; |
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 |
|
|
* Drain off all datagram fragments. |
* Drain off all datagram fragments. |
*/ |
*/ |
void |
void |
ip_drain() |
ip_drain(void) |
{ |
{ |
|
|
/* |
/* |
|
|
if (ipq_lock_try() == 0) |
if (ipq_lock_try() == 0) |
return; |
return; |
|
|
while (LIST_FIRST(&ipq) != NULL) { |
/* |
ipstat.ips_fragdropped++; |
* Drop half the total fragments now. If more mbufs are needed, |
ip_freef(LIST_FIRST(&ipq)); |
* we will be called again soon. |
} |
*/ |
|
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(m) |
ip_dooptions(struct mbuf *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 1320 ip_dooptions(m) |
|
Line 1531 ip_dooptions(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_ifwithaddr(sintosa(&ipaddr))); |
ia = ifatoia(ifa_ifwithladdr(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(dst) |
ip_rtaddr(struct in_addr dst) |
struct in_addr dst; |
|
{ |
{ |
struct sockaddr_in *sin; |
struct sockaddr_in *sin; |
|
|
|
|
* 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(option, dst) |
save_rte(u_char *option, struct in_addr dst) |
u_char *option; |
|
struct in_addr dst; |
|
{ |
{ |
unsigned olen; |
unsigned olen; |
|
|
Line 1514 save_rte(option, dst) |
|
Line 1722 save_rte(option, dst) |
|
* 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() |
ip_srcroute(void) |
{ |
{ |
struct in_addr *p, *q; |
struct in_addr *p, *q; |
struct mbuf *m; |
struct mbuf *m; |
|
|
* XXX should be deleted; last arg currently ignored. |
* XXX should be deleted; last arg currently ignored. |
*/ |
*/ |
void |
void |
ip_stripoptions(m, mopt) |
ip_stripoptions(struct mbuf *m, struct mbuf *mopt __unused) |
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 1630 const int inetctlerrmap[PRC_NCMDS] = { |
|
Line 1836 const int inetctlerrmap[PRC_NCMDS] = { |
|
* via a source route. |
* via a source route. |
*/ |
*/ |
void |
void |
ip_forward(m, srcrt) |
ip_forward(struct mbuf *m, int 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; |
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 1657 ip_forward(m, srcrt) |
|
Line 1857 ip_forward(m, 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 1671 ip_forward(m, srcrt) |
|
Line 1871 ip_forward(m, 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 1686 ip_forward(m, srcrt) |
|
Line 1885 ip_forward(m, 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 1701 ip_forward(m, srcrt) |
|
Line 1900 ip_forward(m, 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 1757 ip_forward(m, srcrt) |
|
Line 1958 ip_forward(m, srcrt) |
|
} |
} |
if (mcopy == NULL) |
if (mcopy == NULL) |
return; |
return; |
destifp = NULL; |
|
|
|
switch (error) { |
switch (error) { |
|
|
Line 1779 ip_forward(m, srcrt) |
|
Line 1979 ip_forward(m, 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 1798 ip_forward(m, srcrt) |
|
Line 1998 ip_forward(m, 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 1807 ip_forward(m, srcrt) |
|
Line 2007 ip_forward(m, 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 1856 ip_forward(m, srcrt) |
|
Line 2050 ip_forward(m, srcrt) |
|
break; |
break; |
#endif |
#endif |
} |
} |
icmp_error(mcopy, type, code, dest, destifp); |
icmp_error(mcopy, type, code, dest, destmtu); |
} |
} |
|
|
void |
void |
ip_savecontrol(inp, mp, ip, m) |
ip_savecontrol(struct inpcb *inp, struct mbuf **mp, struct ip *ip, |
struct inpcb *inp; |
struct mbuf *m) |
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 1919 ip_savecontrol(inp, mp, ip, m) |
|
Line 2110 ip_savecontrol(inp, mp, ip, m) |
|
} |
} |
} |
} |
|
|
int |
/* |
ip_sysctl(name, namelen, oldp, oldlenp, newp, newlen) |
* sysctl helper routine for net.inet.ip.forwsrcrt. |
int *name; |
*/ |
u_int namelen; |
static int |
void *oldp; |
sysctl_net_inet_ip_forwsrcrt(SYSCTLFN_ARGS) |
size_t *oldlenp; |
|
void *newp; |
|
size_t newlen; |
|
{ |
{ |
extern int subnetsarelocal, hostzeroisbroadcast; |
int error, tmp; |
|
struct sysctlnode node; |
int error, old; |
|
|
|
/* All sysctl names (except ifq.*) at this level are terminal. */ |
node = *rnode; |
if ((namelen != 1) && !(namelen == 2 && name[0] == IPCTL_IFQ)) |
tmp = ip_forwsrcrt; |
return (ENOTDIR); |
node.sysctl_data = &tmp; |
|
error = sysctl_lookup(SYSCTLFN_CALL(&node)); |
switch (name[0]) { |
if (error || newp == NULL) |
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); |
return (error); |
case IPCTL_MTUDISCTIMEOUT: |
|
old = ip_mtudisc_timeout; |
if (kauth_authorize_network(l->l_cred, KAUTH_NETWORK_FORWSRCRT, |
error = sysctl_int(oldp, oldlenp, newp, newlen, |
0, NULL, NULL, NULL)) |
&ip_mtudisc_timeout); |
return (EPERM); |
if (ip_mtudisc_timeout < 0) { |
|
ip_mtudisc_timeout = old; |
ip_forwsrcrt = tmp; |
return (EINVAL); |
|
} |
return (0); |
if (error == 0) |
} |
rt_timer_queue_change(ip_mtudisc_timeout_q, |
|
ip_mtudisc_timeout); |
/* |
|
* sysctl helper routine for net.inet.ip.mtudisctimeout. checks the |
|
* range of the new value and tweaks timers if it changes. |
|
*/ |
|
static int |
|
sysctl_net_inet_ip_pmtudto(SYSCTLFN_ARGS) |
|
{ |
|
int error, tmp; |
|
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); |
return (error); |
|
if (tmp < 0) |
|
return (EINVAL); |
|
|
|
ip_mtudisc_timeout = tmp; |
|
rt_timer_queue_change(ip_mtudisc_timeout_q, ip_mtudisc_timeout); |
|
|
|
return (0); |
|
} |
|
|
#ifdef GATEWAY |
#ifdef GATEWAY |
case IPCTL_MAXFLOWS: |
/* |
{ |
* sysctl helper routine for net.inet.ip.maxflows. apparently if |
int s; |
* maxflows is even looked up, we "reap flows". |
|
*/ |
|
static int |
|
sysctl_net_inet_ip_maxflows(SYSCTLFN_ARGS) |
|
{ |
|
int s; |
|
|
error = sysctl_int(oldp, oldlenp, newp, newlen, |
s = sysctl_lookup(SYSCTLFN_CALL(rnode)); |
&ip_maxflows); |
if (s) |
s = splsoftnet(); |
return (s); |
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 |
|
|
|
#if NGRE > 0 |
s = splsoftnet(); |
case IPCTL_GRE_TTL: |
ipflow_reap(0); |
return (sysctl_int(oldp, oldlenp, newp, newlen, |
splx(s); |
&ip_gre_ttl)); |
|
#endif |
|
|
|
#ifndef IPNOPRIVPORTS |
return (0); |
case IPCTL_LOWPORTMIN: |
} |
old = lowportmin; |
#endif /* GATEWAY */ |
error = sysctl_int(oldp, oldlenp, newp, newlen, &lowportmin); |
|
if (lowportmin >= lowportmax |
|
|| lowportmin > IPPORT_RESERVEDMAX |
|
|| lowportmin < IPPORT_RESERVEDMIN |
|
) { |
|
lowportmin = old; |
|
return (EINVAL); |
|
} |
|
return (error); |
|
case IPCTL_LOWPORTMAX: |
|
old = lowportmax; |
|
error = sysctl_int(oldp, oldlenp, newp, newlen, &lowportmax); |
|
if (lowportmin >= lowportmax |
|
|| lowportmax > IPPORT_RESERVEDMAX |
|
|| lowportmax < IPPORT_RESERVEDMIN |
|
) { |
|
lowportmax = old; |
|
return (EINVAL); |
|
} |
|
return (error); |
|
#endif |
|
|
|
case IPCTL_MAXFRAGPACKETS: |
|
return (sysctl_int(oldp, oldlenp, newp, newlen, |
|
&ip_maxfragpackets)); |
|
|
|
case IPCTL_CHECKINTERFACE: |
|
return (sysctl_int(oldp, oldlenp, newp, newlen, |
|
&ip_checkinterface)); |
|
|
|
case IPCTL_IFQ: |
|
return (sysctl_ifq(name + 1, namelen - 1, oldp, oldlenp, |
|
newp, newlen, &ipintrq)); |
|
|
|
case IPCTL_RANDOMID: |
|
#ifdef RANDOM_IP_ID |
|
return (sysctl_int(oldp, oldlenp, newp, newlen, |
|
&ip_do_randomid)); |
|
#else |
|
return (sysctl_rdint(oldp, oldlenp, newp, 0)); |
|
#endif |
|
|
|
default: |
SYSCTL_SETUP(sysctl_net_inet_ip_setup, "sysctl net.inet.ip subtree setup") |
return (EOPNOTSUPP); |
{ |
} |
extern int subnetsarelocal, hostzeroisbroadcast; |
/* NOTREACHED */ |
|
|
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_READWRITE, |
|
CTLTYPE_INT, "forwsrcrt", |
|
SYSCTL_DESCR("Enable forwarding of source-routed " |
|
"datagrams"), |
|
sysctl_net_inet_ip_forwsrcrt, 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 |
|
sysctl_createv(clog, 0, NULL, NULL, |
|
CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
|
CTLTYPE_INT, "lowportmin", |
|
SYSCTL_DESCR("Lowest privileged ephemeral port number " |
|
"to assign"), |
|
sysctl_net_inet_ip_ports, 0, &lowportmin, 0, |
|
CTL_NET, PF_INET, IPPROTO_IP, |
|
IPCTL_LOWPORTMIN, CTL_EOL); |
|
sysctl_createv(clog, 0, NULL, NULL, |
|
CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
|
CTLTYPE_INT, "lowportmax", |
|
SYSCTL_DESCR("Highest privileged ephemeral port number " |
|
"to assign"), |
|
sysctl_net_inet_ip_ports, 0, &lowportmax, 0, |
|
CTL_NET, PF_INET, IPPROTO_IP, |
|
IPCTL_LOWPORTMAX, CTL_EOL); |
|
#endif /* IPNOPRIVPORTS */ |
|
sysctl_createv(clog, 0, NULL, NULL, |
|
CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
|
CTLTYPE_INT, "maxfragpackets", |
|
SYSCTL_DESCR("Maximum number of fragments to retain for " |
|
"possible reassembly"), |
|
NULL, 0, &ip_maxfragpackets, 0, |
|
CTL_NET, PF_INET, IPPROTO_IP, |
|
IPCTL_MAXFRAGPACKETS, CTL_EOL); |
|
#if NGRE > 0 |
|
sysctl_createv(clog, 0, NULL, NULL, |
|
CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
|
CTLTYPE_INT, "grettl", |
|
SYSCTL_DESCR("Default TTL for a gre tunnel datagram"), |
|
NULL, 0, &ip_gre_ttl, 0, |
|
CTL_NET, PF_INET, IPPROTO_IP, |
|
IPCTL_GRE_TTL, CTL_EOL); |
|
#endif /* NGRE */ |
|
sysctl_createv(clog, 0, NULL, NULL, |
|
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); |
} |
} |