version 1.275.4.1.8.2, 2011/01/07 03:17:44 |
version 1.316, 2014/05/29 23:02:48 |
|
|
__KERNEL_RCSID(0, "$NetBSD$"); |
__KERNEL_RCSID(0, "$NetBSD$"); |
|
|
#include "opt_inet.h" |
#include "opt_inet.h" |
|
#include "opt_compat_netbsd.h" |
#include "opt_gateway.h" |
#include "opt_gateway.h" |
#include "opt_pfil_hooks.h" |
|
#include "opt_ipsec.h" |
#include "opt_ipsec.h" |
#include "opt_mrouting.h" |
#include "opt_mrouting.h" |
#include "opt_mbuftrace.h" |
#include "opt_mbuftrace.h" |
Line 103 __KERNEL_RCSID(0, "$NetBSD$"); |
|
Line 103 __KERNEL_RCSID(0, "$NetBSD$"); |
|
|
|
#include <sys/param.h> |
#include <sys/param.h> |
#include <sys/systm.h> |
#include <sys/systm.h> |
#include <sys/malloc.h> |
|
#include <sys/mbuf.h> |
#include <sys/mbuf.h> |
#include <sys/domain.h> |
#include <sys/domain.h> |
#include <sys/protosw.h> |
#include <sys/protosw.h> |
Line 139 __KERNEL_RCSID(0, "$NetBSD$"); |
|
Line 138 __KERNEL_RCSID(0, "$NetBSD$"); |
|
#ifdef MROUTING |
#ifdef MROUTING |
#include <netinet/ip_mroute.h> |
#include <netinet/ip_mroute.h> |
#endif |
#endif |
|
#include <netinet/portalgo.h> |
|
|
#ifdef IPSEC |
#ifdef IPSEC |
#include <netinet6/ipsec.h> |
|
#include <netinet6/ipsec_private.h> |
|
#include <netkey/key.h> |
|
#endif |
|
#ifdef FAST_IPSEC |
|
#include <netipsec/ipsec.h> |
#include <netipsec/ipsec.h> |
#include <netipsec/key.h> |
#endif |
#endif /* FAST_IPSEC*/ |
|
|
|
#ifndef IPFORWARDING |
#ifndef IPFORWARDING |
#ifdef GATEWAY |
#ifdef GATEWAY |
Line 173 __KERNEL_RCSID(0, "$NetBSD$"); |
|
Line 167 __KERNEL_RCSID(0, "$NetBSD$"); |
|
#define IPMTUDISCTIMEOUT (10 * 60) /* as per RFC 1191 */ |
#define IPMTUDISCTIMEOUT (10 * 60) /* as per RFC 1191 */ |
#endif |
#endif |
|
|
|
#ifdef COMPAT_50 |
|
#include <compat/sys/time.h> |
|
#include <compat/sys/socket.h> |
|
#endif |
|
|
/* |
/* |
* Note: DIRECTED_BROADCAST is handled this way so that previous |
* Note: DIRECTED_BROADCAST is handled this way so that previous |
* configuration using this option will Just Work. |
* configuration using this option will Just Work. |
Line 213 int ip_do_randomid = 0; |
|
Line 212 int ip_do_randomid = 0; |
|
*/ |
*/ |
int ip_checkinterface = 0; |
int ip_checkinterface = 0; |
|
|
|
|
struct rttimer_queue *ip_mtudisc_timeout_q = NULL; |
struct rttimer_queue *ip_mtudisc_timeout_q = NULL; |
|
|
int ipqmaxlen = IFQ_MAXLEN; |
|
u_long in_ifaddrhash; /* size of hash table - 1 */ |
|
int in_ifaddrentries; /* total number of addrs */ |
|
struct in_ifaddrhead in_ifaddrhead; |
|
struct in_ifaddrhashhead *in_ifaddrhashtbl; |
|
u_long in_multihash; /* size of hash table - 1 */ |
|
int in_multientries; /* total number of addrs */ |
|
struct in_multihashhead *in_multihashtbl; |
|
struct ifqueue ipintrq; |
struct ifqueue ipintrq; |
|
|
|
ipid_state_t * ip_ids; |
uint16_t ip_id; |
uint16_t ip_id; |
|
|
percpu_t *ipstat_percpu; |
percpu_t *ipstat_percpu; |
|
|
#ifdef PFIL_HOOKS |
pfil_head_t *inet_pfil_hook; |
struct pfil_head inet_pfil_hook; |
|
#endif |
|
|
|
/* |
|
* 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; |
|
static int ip_nfragpackets; /* packets in reass queue */ |
|
static int ip_nfrags; /* total fragments in reass queues */ |
|
|
|
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(void); |
|
static inline void ipq_unlock(void); |
|
|
|
static inline int |
|
ipq_lock_try(void) |
|
{ |
|
int s; |
|
|
|
/* |
|
* Use splvm() -- we're blocking things that would cause |
|
* mbuf allocation. |
|
*/ |
|
s = splvm(); |
|
if (ipq_locked) { |
|
splx(s); |
|
return (0); |
|
} |
|
ipq_locked = 1; |
|
splx(s); |
|
return (1); |
|
} |
|
|
|
static inline void |
|
ipq_unlock(void) |
|
{ |
|
int s; |
|
|
|
s = splvm(); |
|
ipq_locked = 0; |
|
splx(s); |
|
} |
|
|
|
#ifdef DIAGNOSTIC |
|
#define IPQ_LOCK() \ |
|
do { \ |
|
if (ipq_lock_try() == 0) { \ |
|
printf("%s:%d: ipq already locked\n", __FILE__, __LINE__); \ |
|
panic("ipq_lock"); \ |
|
} \ |
|
} while (/*CONSTCOND*/ 0) |
|
#define IPQ_LOCK_CHECK() \ |
|
do { \ |
|
if (ipq_locked == 0) { \ |
|
printf("%s:%d: ipq lock not held\n", __FILE__, __LINE__); \ |
|
panic("ipq lock check"); \ |
|
} \ |
|
} while (/*CONSTCOND*/ 0) |
|
#else |
|
#define IPQ_LOCK() (void) ipq_lock_try() |
|
#define IPQ_LOCK_CHECK() /* nothing */ |
|
#endif |
|
|
|
#define IPQ_UNLOCK() ipq_unlock() |
|
|
|
struct pool inmulti_pool; |
|
struct pool ipqent_pool; |
|
|
|
#ifdef INET_CSUM_COUNTERS |
#ifdef INET_CSUM_COUNTERS |
#include <sys/device.h> |
#include <sys/device.h> |
Line 362 EVCNT_ATTACH_STATIC(ip_swcsum); |
|
Line 252 EVCNT_ATTACH_STATIC(ip_swcsum); |
|
* maintenance when the remote end is on a network that is not known |
* maintenance when the remote end is on a network that is not known |
* to us. |
* to us. |
*/ |
*/ |
int ip_nhops = 0; |
|
|
static int ip_nhops = 0; |
|
|
static struct ip_srcrt { |
static struct ip_srcrt { |
struct in_addr dst; /* final destination */ |
struct in_addr dst; /* final destination */ |
char nop; /* one NOP to align */ |
char nop; /* one NOP to align */ |
Line 370 static struct ip_srcrt { |
|
Line 262 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 int ip_drainwanted; |
|
|
static void save_rte(u_char *, struct in_addr); |
static void save_rte(u_char *, struct in_addr); |
|
|
#ifdef MBUFTRACE |
#ifdef MBUFTRACE |
Line 377 struct mowner ip_rx_mowner = MOWNER_INIT |
|
Line 271 struct mowner ip_rx_mowner = MOWNER_INIT |
|
struct mowner ip_tx_mowner = MOWNER_INIT("internet", "tx"); |
struct mowner ip_tx_mowner = MOWNER_INIT("internet", "tx"); |
#endif |
#endif |
|
|
/* |
static void ip_input(struct mbuf *); |
* Compute IP limits derived from the value of nmbclusters. |
static void ip_forward(struct mbuf *, int); |
*/ |
static bool ip_dooptions(struct mbuf *); |
static void |
static struct in_ifaddr *ip_rtaddr(struct in_addr); |
ip_nmbclusters_changed(void) |
static void sysctl_net_inet_ip_setup(struct sysctllog **); |
{ |
|
ip_maxfrags = nmbclusters / 4; |
|
ip_nmbclusters = nmbclusters; |
|
} |
|
|
|
/* |
/* |
* IP initialization: fill in IP protocol switch table. |
* IP initialization: fill in IP protocol switch table. |
|
|
ip_init(void) |
ip_init(void) |
{ |
{ |
const struct protosw *pr; |
const struct protosw *pr; |
int i; |
|
|
|
pool_init(&inmulti_pool, sizeof(struct in_multi), 0, 0, 0, "inmltpl", |
in_init(); |
NULL, IPL_SOFTNET); |
sysctl_net_inet_ip_setup(NULL); |
pool_init(&ipqent_pool, sizeof(struct ipqent), 0, 0, 0, "ipqepl", |
|
NULL, IPL_VM); |
|
|
|
pr = pffindproto(PF_INET, IPPROTO_RAW, SOCK_RAW); |
pr = pffindproto(PF_INET, IPPROTO_RAW, SOCK_RAW); |
if (pr == 0) |
KASSERT(pr != NULL); |
panic("ip_init"); |
|
for (i = 0; i < IPPROTO_MAX; i++) |
for (u_int i = 0; i < IPPROTO_MAX; i++) { |
ip_protox[i] = pr - inetsw; |
ip_protox[i] = pr - inetsw; |
|
} |
for (pr = inetdomain.dom_protosw; |
for (pr = inetdomain.dom_protosw; |
pr < inetdomain.dom_protoswNPROTOSW; pr++) |
pr < inetdomain.dom_protoswNPROTOSW; pr++) |
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; |
|
|
for (i = 0; i < IPREASS_NHASH; i++) |
ip_reass_init(); |
LIST_INIT(&ipq[i]); |
|
|
|
ip_initid(); |
ip_ids = ip_id_init(); |
ip_id = time_second & 0xfffff; |
ip_id = time_second & 0xfffff; |
|
|
ipintrq.ifq_maxlen = ipqmaxlen; |
ipintrq.ifq_maxlen = IFQ_MAXLEN; |
ip_nmbclusters_changed(); |
|
|
|
TAILQ_INIT(&in_ifaddrhead); |
|
in_ifaddrhashtbl = hashinit(IN_IFADDR_HASH_SIZE, HASH_LIST, true, |
|
&in_ifaddrhash); |
|
in_multihashtbl = hashinit(IN_IFADDR_HASH_SIZE, HASH_LIST, true, |
|
&in_multihash); |
|
ip_mtudisc_timeout_q = rt_timer_queue_create(ip_mtudisc_timeout); |
ip_mtudisc_timeout_q = rt_timer_queue_create(ip_mtudisc_timeout); |
#ifdef GATEWAY |
#ifdef GATEWAY |
ipflow_init(ip_hashsize); |
ipflow_init(); |
#endif |
#endif |
|
|
#ifdef PFIL_HOOKS |
|
/* Register our Packet Filter hook. */ |
/* Register our Packet Filter hook. */ |
inet_pfil_hook.ph_type = PFIL_TYPE_AF; |
inet_pfil_hook = pfil_head_create(PFIL_TYPE_AF, (void *)AF_INET); |
inet_pfil_hook.ph_af = AF_INET; |
KASSERT(inet_pfil_hook != NULL); |
i = pfil_head_register(&inet_pfil_hook); |
|
if (i != 0) |
|
printf("ip_init: WARNING: unable to register pfil hook, " |
|
"error %d\n", i); |
|
#endif /* PFIL_HOOKS */ |
|
|
|
#ifdef MBUFTRACE |
#ifdef MBUFTRACE |
MOWNER_ATTACH(&ip_tx_mowner); |
MOWNER_ATTACH(&ip_tx_mowner); |
Line 454 struct sockaddr_in ipaddr = { |
|
Line 329 struct sockaddr_in ipaddr = { |
|
.sin_len = sizeof(ipaddr), |
.sin_len = sizeof(ipaddr), |
.sin_family = AF_INET, |
.sin_family = AF_INET, |
}; |
}; |
struct route ipforward_rt; |
|
|
static struct route ipforward_rt; |
|
|
/* |
/* |
* IP software interrupt routine |
* IP software interrupt routine |
|
|
{ |
{ |
int s; |
int s; |
struct mbuf *m; |
struct mbuf *m; |
|
struct ifqueue lcl_intrq; |
|
|
|
memset(&lcl_intrq, 0, sizeof(lcl_intrq)); |
|
|
mutex_enter(softnet_lock); |
mutex_enter(softnet_lock); |
KERNEL_LOCK(1, NULL); |
KERNEL_LOCK(1, NULL); |
while (!IF_IS_EMPTY(&ipintrq)) { |
if (!IF_IS_EMPTY(&ipintrq)) { |
s = splnet(); |
s = splnet(); |
IF_DEQUEUE(&ipintrq, m); |
|
|
/* Take existing queue onto stack */ |
|
lcl_intrq = ipintrq; |
|
|
|
/* Zero out global queue, preserving maxlen and drops */ |
|
ipintrq.ifq_head = NULL; |
|
ipintrq.ifq_tail = NULL; |
|
ipintrq.ifq_len = 0; |
|
ipintrq.ifq_maxlen = lcl_intrq.ifq_maxlen; |
|
ipintrq.ifq_drops = lcl_intrq.ifq_drops; |
|
|
splx(s); |
splx(s); |
|
} |
|
KERNEL_UNLOCK_ONE(NULL); |
|
while (!IF_IS_EMPTY(&lcl_intrq)) { |
|
IF_DEQUEUE(&lcl_intrq, m); |
if (m == NULL) |
if (m == NULL) |
break; |
break; |
ip_input(m); |
ip_input(m); |
} |
} |
KERNEL_UNLOCK_ONE(NULL); |
|
mutex_exit(softnet_lock); |
mutex_exit(softnet_lock); |
} |
} |
|
|
|
|
* Ip input routine. Checksum and byte swap header. If fragmented |
* Ip input routine. Checksum and byte swap header. If fragmented |
* try to reassemble. Process options. Pass to next level. |
* try to reassemble. Process options. Pass to next level. |
*/ |
*/ |
void |
static void |
ip_input(struct mbuf *m) |
ip_input(struct mbuf *m) |
{ |
{ |
struct ip *ip = NULL; |
struct ip *ip = NULL; |
struct ipq *fp; |
|
struct in_ifaddr *ia; |
struct in_ifaddr *ia; |
struct ifaddr *ifa; |
struct ifaddr *ifa; |
struct ipqent *ipqe; |
int hlen = 0, len; |
int hlen = 0, mff, len; |
|
int downmatch; |
int downmatch; |
int checkif; |
int checkif; |
int srcrt = 0; |
int srcrt = 0; |
int s; |
ifnet_t *ifp; |
u_int hash; |
|
#ifdef FAST_IPSEC |
|
struct m_tag *mtag; |
|
struct tdb_ident *tdbi; |
|
struct secpolicy *sp; |
|
int error; |
|
#endif /* FAST_IPSEC */ |
|
|
|
MCLAIM(m, &ip_rx_mowner); |
MCLAIM(m, &ip_rx_mowner); |
#ifdef DIAGNOSTIC |
KASSERT((m->m_flags & M_PKTHDR) != 0); |
if ((m->m_flags & M_PKTHDR) == 0) |
ifp = m->m_pkthdr.rcvif; |
panic("ipintr no HDR"); |
|
#endif |
|
|
|
/* |
/* |
* If no IP addresses have been set yet but the interfaces |
* If no IP addresses have been set yet but the interfaces |
Line 547 ip_input(struct mbuf *m) |
|
Line 428 ip_input(struct mbuf *m) |
|
goto bad; |
goto bad; |
} |
} |
if (hlen > m->m_len) { |
if (hlen > m->m_len) { |
if ((m = m_pullup(m, hlen)) == 0) { |
if ((m = m_pullup(m, hlen)) == NULL) { |
IP_STATINC(IP_STAT_BADHLEN); |
IP_STATINC(IP_STAT_BADHLEN); |
return; |
return; |
} |
} |
Line 566 ip_input(struct mbuf *m) |
|
Line 447 ip_input(struct mbuf *m) |
|
/* 127/8 must not appear on wire - RFC1122 */ |
/* 127/8 must not appear on wire - RFC1122 */ |
if ((ntohl(ip->ip_dst.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET || |
if ((ntohl(ip->ip_dst.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET || |
(ntohl(ip->ip_src.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET) { |
(ntohl(ip->ip_src.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET) { |
if ((m->m_pkthdr.rcvif->if_flags & IFF_LOOPBACK) == 0) { |
if ((ifp->if_flags & IFF_LOOPBACK) == 0) { |
IP_STATINC(IP_STAT_BADADDR); |
IP_STATINC(IP_STAT_BADADDR); |
goto bad; |
goto bad; |
} |
} |
} |
} |
|
|
switch (m->m_pkthdr.csum_flags & |
switch (m->m_pkthdr.csum_flags & |
((m->m_pkthdr.rcvif->if_csum_flags_rx & M_CSUM_IPv4) | |
((ifp->if_csum_flags_rx & M_CSUM_IPv4) | |
M_CSUM_IPv4_BAD)) { |
M_CSUM_IPv4_BAD)) { |
case M_CSUM_IPv4|M_CSUM_IPv4_BAD: |
case M_CSUM_IPv4|M_CSUM_IPv4_BAD: |
INET_CSUM_COUNTER_INCR(&ip_hwcsum_bad); |
INET_CSUM_COUNTER_INCR(&ip_hwcsum_bad); |
Line 589 ip_input(struct mbuf *m) |
|
Line 470 ip_input(struct mbuf *m) |
|
* Must compute it ourselves. Maybe skip checksum on |
* Must compute it ourselves. Maybe skip checksum on |
* loopback interfaces. |
* loopback interfaces. |
*/ |
*/ |
if (__predict_true(!(m->m_pkthdr.rcvif->if_flags & |
if (__predict_true(!(ifp->if_flags & IFF_LOOPBACK) || |
IFF_LOOPBACK) || ip_do_loopback_cksum)) { |
ip_do_loopback_cksum)) { |
INET_CSUM_COUNTER_INCR(&ip_swcsum); |
INET_CSUM_COUNTER_INCR(&ip_swcsum); |
if (in_cksum(m, hlen) != 0) |
if (in_cksum(m, hlen) != 0) |
goto badcsum; |
goto badcsum; |
Line 627 ip_input(struct mbuf *m) |
|
Line 508 ip_input(struct mbuf *m) |
|
m_adj(m, len - m->m_pkthdr.len); |
m_adj(m, len - m->m_pkthdr.len); |
} |
} |
|
|
#if defined(IPSEC) |
|
/* ipflow (IP fast forwarding) is not compatible with IPsec. */ |
|
m->m_flags &= ~M_CANFASTFWD; |
|
#else |
|
/* |
/* |
* Assume that we can create a fast-forward IP flow entry |
* Assume that we can create a fast-forward IP flow entry |
* based on this packet. |
* based on this packet. |
*/ |
*/ |
m->m_flags |= M_CANFASTFWD; |
m->m_flags |= M_CANFASTFWD; |
#endif |
|
|
|
#ifdef PFIL_HOOKS |
|
/* |
/* |
* Run through list of hooks for input packets. If there are any |
* Run through list of hooks for input packets. If there are any |
* filters which require that additional packets in the flow are |
* filters which require that additional packets in the flow are |
Line 646 ip_input(struct mbuf *m) |
|
Line 521 ip_input(struct mbuf *m) |
|
* Note that filters must _never_ set this flag, as another filter |
* Note that filters must _never_ set this flag, as another filter |
* in the list may have previously cleared it. |
* in the list may have previously cleared it. |
*/ |
*/ |
/* |
#if defined(IPSEC) |
* let ipfilter look at packet on the wire, |
|
* not the decapsulated packet. |
|
*/ |
|
#ifdef IPSEC |
|
if (!ipsec_getnhist(m)) |
|
#elif defined(FAST_IPSEC) |
|
if (!ipsec_indone(m)) |
if (!ipsec_indone(m)) |
#else |
#else |
if (1) |
if (1) |
#endif |
#endif |
{ |
{ |
struct in_addr odst; |
struct in_addr odst = ip->ip_dst; |
|
|
odst = ip->ip_dst; |
if (pfil_run_hooks(inet_pfil_hook, &m, ifp, PFIL_IN) != 0) |
if (pfil_run_hooks(&inet_pfil_hook, &m, m->m_pkthdr.rcvif, |
|
PFIL_IN) != 0) |
|
return; |
return; |
if (m == NULL) |
if (m == NULL) |
return; |
return; |
Line 684 ip_input(struct mbuf *m) |
|
Line 551 ip_input(struct mbuf *m) |
|
*/ |
*/ |
srcrt = (odst.s_addr != ip->ip_dst.s_addr); |
srcrt = (odst.s_addr != ip->ip_dst.s_addr); |
} |
} |
#endif /* PFIL_HOOKS */ |
|
|
|
#ifdef ALTQ |
#ifdef ALTQ |
/* XXX Temporary until ALTQ is changed to use a pfil hook */ |
/* XXX Temporary until ALTQ is changed to use a pfil hook */ |
Line 721 ip_input(struct mbuf *m) |
|
Line 587 ip_input(struct mbuf *m) |
|
* we get finer grain control. |
* we get finer grain control. |
*/ |
*/ |
checkif = ip_checkinterface && (ipforwarding == 0) && |
checkif = ip_checkinterface && (ipforwarding == 0) && |
(m->m_pkthdr.rcvif != NULL) && |
(ifp->if_flags & IFF_LOOPBACK) == 0; |
((m->m_pkthdr.rcvif->if_flags & IFF_LOOPBACK) == 0); |
|
|
|
/* |
/* |
* Check our list of addresses, to see if the packet is for us. |
* Check our list of addresses, to see if the packet is for us. |
Line 734 ip_input(struct mbuf *m) |
|
Line 599 ip_input(struct mbuf *m) |
|
downmatch = 0; |
downmatch = 0; |
LIST_FOREACH(ia, &IN_IFADDR_HASH(ip->ip_dst.s_addr), ia_hash) { |
LIST_FOREACH(ia, &IN_IFADDR_HASH(ip->ip_dst.s_addr), ia_hash) { |
if (in_hosteq(ia->ia_addr.sin_addr, ip->ip_dst)) { |
if (in_hosteq(ia->ia_addr.sin_addr, ip->ip_dst)) { |
if (checkif && ia->ia_ifp != m->m_pkthdr.rcvif) |
if (checkif && ia->ia_ifp != ifp) |
continue; |
continue; |
if ((ia->ia_ifp->if_flags & IFF_UP) != 0) |
if ((ia->ia_ifp->if_flags & IFF_UP) != 0) |
break; |
break; |
Line 744 ip_input(struct mbuf *m) |
|
Line 609 ip_input(struct mbuf *m) |
|
} |
} |
if (ia != NULL) |
if (ia != NULL) |
goto ours; |
goto ours; |
if (m->m_pkthdr.rcvif && m->m_pkthdr.rcvif->if_flags & IFF_BROADCAST) { |
if (ifp->if_flags & IFF_BROADCAST) { |
IFADDR_FOREACH(ifa, m->m_pkthdr.rcvif) { |
IFADDR_FOREACH(ifa, ifp) { |
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 767 ip_input(struct mbuf *m) |
|
Line 632 ip_input(struct mbuf *m) |
|
} |
} |
} |
} |
if (IN_MULTICAST(ip->ip_dst.s_addr)) { |
if (IN_MULTICAST(ip->ip_dst.s_addr)) { |
struct in_multi *inm; |
|
#ifdef MROUTING |
#ifdef MROUTING |
extern struct socket *ip_mrouter; |
extern struct socket *ip_mrouter; |
|
|
Line 784 ip_input(struct mbuf *m) |
|
Line 648 ip_input(struct mbuf *m) |
|
* as expected when ip_mforward() is called from |
* as expected when ip_mforward() is called from |
* ip_output().) |
* ip_output().) |
*/ |
*/ |
if (ip_mforward(m, m->m_pkthdr.rcvif) != 0) { |
if (ip_mforward(m, ifp) != 0) { |
IP_STATINC(IP_STAT_CANTFORWARD); |
IP_STATINC(IP_STAT_CANTFORWARD); |
m_freem(m); |
m_freem(m); |
return; |
return; |
Line 804 ip_input(struct mbuf *m) |
|
Line 668 ip_input(struct mbuf *m) |
|
* See if we belong to the destination multicast group on the |
* See if we belong to the destination multicast group on the |
* arrival interface. |
* arrival interface. |
*/ |
*/ |
IN_LOOKUP_MULTI(ip->ip_dst, m->m_pkthdr.rcvif, inm); |
if (!in_multi_group(ip->ip_dst, ifp, 0)) { |
if (inm == NULL) { |
|
IP_STATINC(IP_STAT_CANTFORWARD); |
IP_STATINC(IP_STAT_CANTFORWARD); |
m_freem(m); |
m_freem(m); |
return; |
return; |
Line 835 ip_input(struct mbuf *m) |
|
Line 698 ip_input(struct mbuf *m) |
|
return; |
return; |
} |
} |
#ifdef IPSEC |
#ifdef IPSEC |
if (ipsec4_in_reject(m, NULL)) { |
/* Perform IPsec, if any. */ |
IPSEC_STATINC(IPSEC_STAT_IN_POLVIO); |
if (ipsec4_input(m, IP_FORWARDING | (ip_directedbcast ? |
|
IP_ALLOWBROADCAST : 0)) != 0) { |
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) { |
|
IP_STATINC(IP_STAT_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); |
} |
} |
return; |
return; |
Line 896 ip_input(struct mbuf *m) |
|
Line 711 ip_input(struct mbuf *m) |
|
ours: |
ours: |
/* |
/* |
* If offset or IP_MF are set, must reassemble. |
* If offset or IP_MF are set, must reassemble. |
* Otherwise, nothing need be done. |
|
* (We could look in the reassembly queue to see |
|
* if the packet was previously fragmented, |
|
* 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)) { |
uint16_t off; |
|
/* |
/* |
* Prevent TCP blind data attacks by not allowing non-initial |
* Pass to IP reassembly mechanism. |
* fragments to start at less than 68 bytes (minimal fragment |
|
* size) and making sure the first fragment is at least 68 |
|
* bytes. |
|
*/ |
*/ |
off = (ntohs(ip->ip_off) & IP_OFFMASK) << 3; |
if (ip_reass_packet(&m, ip) != 0) { |
if ((off > 0 ? off + hlen : len) < IP_MINFRAGSIZE - 1) { |
/* Failed; invalid fragment(s) or packet. */ |
IP_STATINC(IP_STAT_BADFRAGS); |
|
goto bad; |
goto bad; |
} |
} |
/* |
if (m == NULL) { |
* Look for queue of fragments |
/* More fragments should come; silently return. */ |
* of this datagram. |
return; |
*/ |
|
IPQ_LOCK(); |
|
hash = IPREASS_HASH(ip->ip_src.s_addr, ip->ip_id); |
|
LIST_FOREACH(fp, &ipq[hash], ipq_q) { |
|
if (ip->ip_id == fp->ipq_id && |
|
in_hosteq(ip->ip_src, fp->ipq_src) && |
|
in_hosteq(ip->ip_dst, fp->ipq_dst) && |
|
ip->ip_p == fp->ipq_p) { |
|
/* |
|
* Make sure the TOS is matches previous |
|
* fragments. |
|
*/ |
|
if (ip->ip_tos != fp->ipq_tos) { |
|
IP_STATINC(IP_STAT_BADFRAGS); |
|
IPQ_UNLOCK(); |
|
goto bad; |
|
} |
|
goto found; |
|
} |
|
} |
|
fp = 0; |
|
found: |
|
|
|
/* |
|
* Adjust ip_len to not reflect header, |
|
* set ipqe_mff if more fragments are expected, |
|
* convert offset of this to bytes. |
|
*/ |
|
ip->ip_len = htons(ntohs(ip->ip_len) - hlen); |
|
mff = (ip->ip_off & htons(IP_MF)) != 0; |
|
if (mff) { |
|
/* |
|
* Make sure that fragments have a data length |
|
* that's a non-zero multiple of 8 bytes. |
|
*/ |
|
if (ntohs(ip->ip_len) == 0 || |
|
(ntohs(ip->ip_len) & 0x7) != 0) { |
|
IP_STATINC(IP_STAT_BADFRAGS); |
|
IPQ_UNLOCK(); |
|
goto bad; |
|
} |
|
} |
} |
ip->ip_off = htons((ntohs(ip->ip_off) & IP_OFFMASK) << 3); |
|
|
|
/* |
/* |
* If datagram marked as having more fragments |
* Reassembly is done, we have the final packet. |
* or if this is not the first fragment, |
* Updated cached data in local variable(s). |
* attempt reassembly; if it succeeds, proceed. |
|
*/ |
*/ |
if (mff || ip->ip_off != htons(0)) { |
ip = mtod(m, struct ip *); |
IP_STATINC(IP_STAT_FRAGMENTS); |
hlen = ip->ip_hl << 2; |
s = splvm(); |
|
ipqe = pool_get(&ipqent_pool, PR_NOWAIT); |
|
splx(s); |
|
if (ipqe == NULL) { |
|
IP_STATINC(IP_STAT_RCVMEMDROP); |
|
IPQ_UNLOCK(); |
|
goto bad; |
|
} |
|
ipqe->ipqe_mff = mff; |
|
ipqe->ipqe_m = m; |
|
ipqe->ipqe_ip = ip; |
|
m = ip_reass(ipqe, fp, &ipq[hash]); |
|
if (m == 0) { |
|
IPQ_UNLOCK(); |
|
return; |
|
} |
|
IP_STATINC(IP_STAT_REASSEMBLED); |
|
ip = mtod(m, struct ip *); |
|
hlen = ip->ip_hl << 2; |
|
ip->ip_len = htons(ntohs(ip->ip_len) + hlen); |
|
} else |
|
if (fp) |
|
ip_freef(fp); |
|
IPQ_UNLOCK(); |
|
} |
} |
|
|
#if defined(IPSEC) |
#ifdef 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 && |
|
ipsec4_in_reject(m, NULL)) { |
|
IPSEC_STATINC(IPSEC_STAT_IN_POLVIO); |
|
goto bad; |
|
} |
|
#endif |
|
#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 |
* code - like udp/tcp/raw ip. |
* code - like UDP/TCP/raw IP. |
*/ |
*/ |
if ((inetsw[ip_protox[ip->ip_p]].pr_flags & PR_LASTHDR) != 0) { |
if ((inetsw[ip_protox[ip->ip_p]].pr_flags & PR_LASTHDR) != 0) { |
/* |
if (ipsec4_input(m, 0) != 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*/ |
|
} |
|
splx(s); |
|
if (error) |
|
goto bad; |
goto bad; |
|
} |
} |
} |
#endif /* FAST_IPSEC */ |
#endif |
|
|
/* |
/* |
* Switch out to protocol's input routine. |
* Switch out to protocol's input routine. |
|
|
} |
} |
|
|
/* |
/* |
* Take incoming datagram fragment and try to |
* IP timer processing. |
* reassemble it into whole datagram. If a chain for |
|
* reassembly of this datagram already exists, then it |
|
* is given as fp; otherwise have to make a chain. |
|
*/ |
|
struct mbuf * |
|
ip_reass(struct ipqent *ipqe, struct ipq *fp, struct ipqhead *ipqhead) |
|
{ |
|
struct mbuf *m = ipqe->ipqe_m; |
|
struct ipqent *nq, *p, *q; |
|
struct ip *ip; |
|
struct mbuf *t; |
|
int hlen = ipqe->ipqe_ip->ip_hl << 2; |
|
int i, next, s; |
|
|
|
IPQ_LOCK_CHECK(); |
|
|
|
/* |
|
* Presence of header sizes in mbufs |
|
* would confuse code below. |
|
*/ |
|
m->m_data += 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 (fp == 0) { |
|
/* |
|
* Enforce upper bound on number of fragmented packets |
|
* for which we attempt reassembly; |
|
* If maxfrag is 0, never accept fragments. |
|
* If maxfrag is -1, accept all fragments without limitation. |
|
*/ |
|
if (ip_maxfragpackets < 0) |
|
; |
|
else if (ip_nfragpackets >= ip_maxfragpackets) |
|
goto dropfrag; |
|
ip_nfragpackets++; |
|
MALLOC(fp, struct ipq *, sizeof (struct ipq), |
|
M_FTABLE, M_NOWAIT); |
|
if (fp == NULL) |
|
goto dropfrag; |
|
LIST_INSERT_HEAD(ipqhead, fp, ipq_q); |
|
fp->ipq_nfrags = 1; |
|
fp->ipq_ttl = IPFRAGTTL; |
|
fp->ipq_p = ipqe->ipqe_ip->ip_p; |
|
fp->ipq_id = ipqe->ipqe_ip->ip_id; |
|
fp->ipq_tos = ipqe->ipqe_ip->ip_tos; |
|
TAILQ_INIT(&fp->ipq_fragq); |
|
fp->ipq_src = ipqe->ipqe_ip->ip_src; |
|
fp->ipq_dst = ipqe->ipqe_ip->ip_dst; |
|
p = NULL; |
|
goto insert; |
|
} else { |
|
fp->ipq_nfrags++; |
|
} |
|
|
|
/* |
|
* Find a segment which begins after this one does. |
|
*/ |
|
for (p = NULL, q = TAILQ_FIRST(&fp->ipq_fragq); q != NULL; |
|
p = q, q = TAILQ_NEXT(q, ipqe_q)) |
|
if (ntohs(q->ipqe_ip->ip_off) > ntohs(ipqe->ipqe_ip->ip_off)) |
|
break; |
|
|
|
/* |
|
* If there is a preceding segment, it may provide some of |
|
* our data already. If so, drop the data from the incoming |
|
* segment. If it provides all of our data, drop us. |
|
*/ |
|
if (p != NULL) { |
|
i = ntohs(p->ipqe_ip->ip_off) + ntohs(p->ipqe_ip->ip_len) - |
|
ntohs(ipqe->ipqe_ip->ip_off); |
|
if (i > 0) { |
|
if (i >= ntohs(ipqe->ipqe_ip->ip_len)) |
|
goto dropfrag; |
|
m_adj(ipqe->ipqe_m, i); |
|
ipqe->ipqe_ip->ip_off = |
|
htons(ntohs(ipqe->ipqe_ip->ip_off) + i); |
|
ipqe->ipqe_ip->ip_len = |
|
htons(ntohs(ipqe->ipqe_ip->ip_len) - i); |
|
} |
|
} |
|
|
|
/* |
|
* While we overlap succeeding segments trim them or, |
|
* if they are completely covered, dequeue them. |
|
*/ |
|
for (; q != NULL && |
|
ntohs(ipqe->ipqe_ip->ip_off) + ntohs(ipqe->ipqe_ip->ip_len) > |
|
ntohs(q->ipqe_ip->ip_off); q = nq) { |
|
i = (ntohs(ipqe->ipqe_ip->ip_off) + |
|
ntohs(ipqe->ipqe_ip->ip_len)) - ntohs(q->ipqe_ip->ip_off); |
|
if (i < ntohs(q->ipqe_ip->ip_len)) { |
|
q->ipqe_ip->ip_len = |
|
htons(ntohs(q->ipqe_ip->ip_len) - i); |
|
q->ipqe_ip->ip_off = |
|
htons(ntohs(q->ipqe_ip->ip_off) + i); |
|
m_adj(q->ipqe_m, i); |
|
break; |
|
} |
|
nq = TAILQ_NEXT(q, ipqe_q); |
|
m_freem(q->ipqe_m); |
|
TAILQ_REMOVE(&fp->ipq_fragq, q, ipqe_q); |
|
s = splvm(); |
|
pool_put(&ipqent_pool, q); |
|
splx(s); |
|
fp->ipq_nfrags--; |
|
ip_nfrags--; |
|
} |
|
|
|
insert: |
|
/* |
|
* Stick new segment in its place; |
|
* check for complete reassembly. |
|
*/ |
|
if (p == NULL) { |
|
TAILQ_INSERT_HEAD(&fp->ipq_fragq, ipqe, ipqe_q); |
|
} else { |
|
TAILQ_INSERT_AFTER(&fp->ipq_fragq, p, ipqe, ipqe_q); |
|
} |
|
next = 0; |
|
for (p = NULL, q = TAILQ_FIRST(&fp->ipq_fragq); q != NULL; |
|
p = q, q = TAILQ_NEXT(q, ipqe_q)) { |
|
if (ntohs(q->ipqe_ip->ip_off) != next) |
|
return (0); |
|
next += ntohs(q->ipqe_ip->ip_len); |
|
} |
|
if (p->ipqe_mff) |
|
return (0); |
|
|
|
/* |
|
* Reassembly is complete. Check for a bogus message size and |
|
* concatenate fragments. |
|
*/ |
|
q = TAILQ_FIRST(&fp->ipq_fragq); |
|
ip = q->ipqe_ip; |
|
if ((next + (ip->ip_hl << 2)) > IP_MAXPACKET) { |
|
IP_STATINC(IP_STAT_TOOLONG); |
|
ip_freef(fp); |
|
return (0); |
|
} |
|
m = q->ipqe_m; |
|
t = m->m_next; |
|
m->m_next = 0; |
|
m_cat(m, t); |
|
nq = TAILQ_NEXT(q, ipqe_q); |
|
s = splvm(); |
|
pool_put(&ipqent_pool, q); |
|
splx(s); |
|
for (q = nq; q != NULL; q = nq) { |
|
t = q->ipqe_m; |
|
nq = TAILQ_NEXT(q, ipqe_q); |
|
s = splvm(); |
|
pool_put(&ipqent_pool, q); |
|
splx(s); |
|
m_cat(m, t); |
|
} |
|
ip_nfrags -= fp->ipq_nfrags; |
|
|
|
/* |
|
* Create header for new ip packet by |
|
* modifying header of first packet; |
|
* dequeue and discard fragment reassembly header. |
|
* Make header visible. |
|
*/ |
|
ip->ip_len = htons(next); |
|
ip->ip_src = fp->ipq_src; |
|
ip->ip_dst = fp->ipq_dst; |
|
LIST_REMOVE(fp, ipq_q); |
|
FREE(fp, M_FTABLE); |
|
ip_nfragpackets--; |
|
m->m_len += (ip->ip_hl << 2); |
|
m->m_data -= (ip->ip_hl << 2); |
|
/* some debugging cruft by sklower, below, will go away soon */ |
|
if (m->m_flags & M_PKTHDR) { /* XXX this should be done elsewhere */ |
|
int plen = 0; |
|
for (t = m; t; t = t->m_next) |
|
plen += t->m_len; |
|
m->m_pkthdr.len = plen; |
|
m->m_pkthdr.csum_flags = 0; |
|
} |
|
return (m); |
|
|
|
dropfrag: |
|
if (fp != 0) |
|
fp->ipq_nfrags--; |
|
ip_nfrags--; |
|
IP_STATINC(IP_STAT_FRAGDROPPED); |
|
m_freem(m); |
|
s = splvm(); |
|
pool_put(&ipqent_pool, ipqe); |
|
splx(s); |
|
return (0); |
|
} |
|
|
|
/* |
|
* Free a fragment reassembly header and all |
|
* associated datagrams. |
|
*/ |
|
void |
|
ip_freef(struct ipq *fp) |
|
{ |
|
struct ipqent *q, *p; |
|
u_int nfrags = 0; |
|
int s; |
|
|
|
IPQ_LOCK_CHECK(); |
|
|
|
for (q = TAILQ_FIRST(&fp->ipq_fragq); q != NULL; q = p) { |
|
p = TAILQ_NEXT(q, ipqe_q); |
|
m_freem(q->ipqe_m); |
|
nfrags++; |
|
TAILQ_REMOVE(&fp->ipq_fragq, q, ipqe_q); |
|
s = splvm(); |
|
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); |
|
FREE(fp, M_FTABLE); |
|
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) { |
|
IP_STATINC(IP_STAT_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; |
|
* if a timer expires on a reassembly |
|
* queue, discard it. |
|
*/ |
*/ |
void |
void |
ip_slowtimo(void) |
ip_slowtimo(void) |
{ |
{ |
static u_int dropscanidx = 0; |
|
u_int i; |
|
u_int median_ttl; |
|
|
|
mutex_enter(softnet_lock); |
mutex_enter(softnet_lock); |
KERNEL_LOCK(1, NULL); |
KERNEL_LOCK(1, NULL); |
|
|
IPQ_LOCK(); |
ip_reass_slowtimo(); |
|
|
/* Age TTL of all fragments by 1 tick .*/ |
|
median_ttl = ip_reass_ttl_decr(1); |
|
|
|
/* 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 fragmented packets |
|
* (due to the limit being lowered), drain off |
|
* enough to get down to the new limit. Start draining |
|
* from the reassembly hashqueue most recently drained. |
|
*/ |
|
if (ip_maxfragpackets < 0) |
|
; |
|
else { |
|
int wrapped = 0; |
|
|
|
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(); |
|
|
|
KERNEL_UNLOCK_ONE(NULL); |
KERNEL_UNLOCK_ONE(NULL); |
mutex_exit(softnet_lock); |
mutex_exit(softnet_lock); |
} |
} |
|
|
/* |
/* |
* Drain off all datagram fragments. Don't acquire softnet_lock as |
* IP drain processing. |
* can be called from hardware interrupt context. |
|
*/ |
*/ |
void |
void |
ip_drain(void) |
ip_drain(void) |
{ |
{ |
|
|
KERNEL_LOCK(1, NULL); |
KERNEL_LOCK(1, NULL); |
|
ip_reass_drain(); |
/* |
|
* We may be called from a device's interrupt context. If |
|
* the ipq is already busy, just bail out now. |
|
*/ |
|
if (ipq_lock_try() != 0) { |
|
/* |
|
* Drop half the total fragments now. If more mbufs are |
|
* needed, we will be called again soon. |
|
*/ |
|
ip_reass_drophalf(); |
|
IPQ_UNLOCK(); |
|
} |
|
|
|
KERNEL_UNLOCK_ONE(NULL); |
KERNEL_UNLOCK_ONE(NULL); |
} |
} |
|
|
/* |
/* |
* Do option processing on a datagram, |
* ip_dooptions: perform option processing on a datagram, possibly discarding |
* possibly discarding it if bad options are encountered, |
* it if bad options are encountered, or forwarding it if source-routed. |
* or forwarding it if source-routed. |
* |
* Returns 1 if packet has been forwarded/freed, |
* => Returns true if packet has been forwarded/freed. |
* 0 if the packet should be processed further. |
* => Returns false if the packet should be processed further. |
*/ |
*/ |
int |
static bool |
ip_dooptions(struct mbuf *m) |
ip_dooptions(struct mbuf *m) |
{ |
{ |
struct ip *ip = mtod(m, struct ip *); |
struct ip *ip = mtod(m, struct ip *); |
Line 1557 ip_dooptions(struct mbuf *m) |
|
Line 888 ip_dooptions(struct mbuf *m) |
|
/* |
/* |
* locate outgoing interface |
* locate outgoing interface |
*/ |
*/ |
bcopy((void *)(cp + off), (void *)&ipaddr.sin_addr, |
memcpy((void *)&ipaddr.sin_addr, (void *)(cp + off), |
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_ifwithladdr(sintosa(&ipaddr))); |
Line 1593 ip_dooptions(struct mbuf *m) |
|
Line 924 ip_dooptions(struct mbuf *m) |
|
off--; /* 0 origin */ |
off--; /* 0 origin */ |
if ((off + sizeof(struct in_addr)) > optlen) |
if ((off + sizeof(struct in_addr)) > optlen) |
break; |
break; |
bcopy((void *)(&ip->ip_dst), (void *)&ipaddr.sin_addr, |
memcpy((void *)&ipaddr.sin_addr, (void *)(&ip->ip_dst), |
sizeof(ipaddr.sin_addr)); |
sizeof(ipaddr.sin_addr)); |
/* |
/* |
* locate outgoing interface; if we're the destination, |
* locate outgoing interface; if we're the destination, |
Line 1660 ip_dooptions(struct mbuf *m) |
|
Line 991 ip_dooptions(struct mbuf *m) |
|
(u_char *)ip; |
(u_char *)ip; |
goto bad; |
goto bad; |
} |
} |
bcopy(cp0, &ipaddr.sin_addr, |
memcpy(&ipaddr.sin_addr, cp0, |
sizeof(struct in_addr)); |
sizeof(struct in_addr)); |
if (ifatoia(ifa_ifwithaddr(sintosa(&ipaddr))) |
if (ifatoia(ifa_ifwithaddr(sintosa(&ipaddr))) |
== NULL) |
== NULL) |
Line 1688 ip_dooptions(struct mbuf *m) |
|
Line 1019 ip_dooptions(struct mbuf *m) |
|
goto bad; |
goto bad; |
} |
} |
ip_forward(m, 1); |
ip_forward(m, 1); |
return (1); |
return true; |
} |
} |
return (0); |
return false; |
bad: |
bad: |
icmp_error(m, type, code, 0, 0); |
icmp_error(m, type, code, 0, 0); |
IP_STATINC(IP_STAT_BADOPTIONS); |
IP_STATINC(IP_STAT_BADOPTIONS); |
return (1); |
return true; |
} |
} |
|
|
/* |
/* |
* Given address of next destination (final or next hop), |
* ip_rtaddr: given address of next destination (final or next hop), |
* 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 * |
static struct in_ifaddr * |
ip_rtaddr(struct in_addr dst) |
ip_rtaddr(struct in_addr dst) |
{ |
{ |
struct rtentry *rt; |
struct rtentry *rt; |
Line 1719 ip_rtaddr(struct in_addr dst) |
|
Line 1050 ip_rtaddr(struct in_addr dst) |
|
} |
} |
|
|
/* |
/* |
* Save incoming source route for use in replies, |
* save_rte: save incoming source route for use in replies, to be picked |
* to be picked up later by ip_srcroute if the receiver is interested. |
* up later by ip_srcroute if the receiver is interested. |
*/ |
*/ |
void |
static void |
save_rte(u_char *option, struct in_addr dst) |
save_rte(u_char *option, struct in_addr dst) |
{ |
{ |
unsigned olen; |
unsigned olen; |
|
|
olen = option[IPOPT_OLEN]; |
olen = option[IPOPT_OLEN]; |
#ifdef DIAGNOSTIC |
|
if (ipprintfs) |
|
printf("save_rte: olen %d\n", olen); |
|
#endif /* 0 */ |
|
if (olen > sizeof(ip_srcrt) - (1 + sizeof(dst))) |
if (olen > sizeof(ip_srcrt) - (1 + sizeof(dst))) |
return; |
return; |
bcopy((void *)option, (void *)ip_srcrt.srcopt, olen); |
memcpy((void *)ip_srcrt.srcopt, (void *)option, olen); |
ip_nhops = (olen - IPOPT_OFFSET - 1) / sizeof(struct in_addr); |
ip_nhops = (olen - IPOPT_OFFSET - 1) / sizeof(struct in_addr); |
ip_srcrt.dst = dst; |
ip_srcrt.dst = dst; |
} |
} |
Line 1762 ip_srcroute(void) |
|
Line 1089 ip_srcroute(void) |
|
/* length is (nhops+1)*sizeof(addr) + sizeof(nop + srcrt header) */ |
/* length is (nhops+1)*sizeof(addr) + sizeof(nop + srcrt header) */ |
m->m_len = ip_nhops * sizeof(struct in_addr) + sizeof(struct in_addr) + |
m->m_len = ip_nhops * sizeof(struct in_addr) + sizeof(struct in_addr) + |
OPTSIZ; |
OPTSIZ; |
#ifdef DIAGNOSTIC |
|
if (ipprintfs) |
|
printf("ip_srcroute: nhops %d mlen %d", ip_nhops, m->m_len); |
|
#endif |
|
|
|
/* |
/* |
* First save first hop for return route |
* First save first hop for return route |
*/ |
*/ |
p = &ip_srcrt.route[ip_nhops - 1]; |
p = &ip_srcrt.route[ip_nhops - 1]; |
*(mtod(m, struct in_addr *)) = *p--; |
*(mtod(m, struct in_addr *)) = *p--; |
#ifdef DIAGNOSTIC |
|
if (ipprintfs) |
|
printf(" hops %x", ntohl(mtod(m, struct in_addr *)->s_addr)); |
|
#endif |
|
|
|
/* |
/* |
* Copy option fields and padding (nop) to mbuf. |
* Copy option fields and padding (nop) to mbuf. |
Line 1792 ip_srcroute(void) |
|
Line 1111 ip_srcroute(void) |
|
* reversing the path (pointers are now aligned). |
* reversing the path (pointers are now aligned). |
*/ |
*/ |
while (p >= ip_srcrt.route) { |
while (p >= ip_srcrt.route) { |
#ifdef DIAGNOSTIC |
|
if (ipprintfs) |
|
printf(" %x", ntohl(q->s_addr)); |
|
#endif |
|
*q++ = *p--; |
*q++ = *p--; |
} |
} |
/* |
/* |
* Last hop goes to final destination. |
* Last hop goes to final destination. |
*/ |
*/ |
*q = ip_srcrt.dst; |
*q = ip_srcrt.dst; |
#ifdef DIAGNOSTIC |
|
if (ipprintfs) |
|
printf(" %x\n", ntohl(q->s_addr)); |
|
#endif |
|
return (m); |
return (m); |
} |
} |
|
|
Line 1821 const int inetctlerrmap[PRC_NCMDS] = { |
|
Line 1132 const int inetctlerrmap[PRC_NCMDS] = { |
|
[PRC_PARAMPROB] = ENOPROTOOPT, |
[PRC_PARAMPROB] = ENOPROTOOPT, |
}; |
}; |
|
|
|
void |
|
ip_fasttimo(void) |
|
{ |
|
if (ip_drainwanted) { |
|
ip_drain(); |
|
ip_drainwanted = 0; |
|
} |
|
} |
|
|
|
void |
|
ip_drainstub(void) |
|
{ |
|
ip_drainwanted = 1; |
|
} |
|
|
/* |
/* |
* Forward a packet. If some error occurs return the sender |
* Forward a packet. If some error occurs return the sender |
* an icmp packet. Note we can't always generate a meaningful |
* an icmp packet. Note we can't always generate a meaningful |
Line 1835 const int inetctlerrmap[PRC_NCMDS] = { |
|
Line 1161 const int inetctlerrmap[PRC_NCMDS] = { |
|
* The srcrt parameter indicates whether the packet is being forwarded |
* The srcrt parameter indicates whether the packet is being forwarded |
* via a source route. |
* via a source route. |
*/ |
*/ |
void |
static void |
ip_forward(struct mbuf *m, int srcrt) |
ip_forward(struct mbuf *m, int srcrt) |
{ |
{ |
struct ip *ip = mtod(m, struct ip *); |
struct ip *ip = mtod(m, struct ip *); |
Line 1859 ip_forward(struct mbuf *m, int srcrt) |
|
Line 1185 ip_forward(struct mbuf *m, int srcrt) |
|
m->m_pkthdr.csum_flags = 0; |
m->m_pkthdr.csum_flags = 0; |
|
|
dest = 0; |
dest = 0; |
#ifdef DIAGNOSTIC |
|
if (ipprintfs) { |
|
printf("forward: src %s ", inet_ntoa(ip->ip_src)); |
|
printf("dst %s ttl %x\n", inet_ntoa(ip->ip_dst), ip->ip_ttl); |
|
} |
|
#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) { |
IP_STATINC(IP_STAT_CANTFORWARD); |
IP_STATINC(IP_STAT_CANTFORWARD); |
m_freem(m); |
m_freem(m); |
Line 1917 ip_forward(struct mbuf *m, int srcrt) |
|
Line 1237 ip_forward(struct mbuf *m, int srcrt) |
|
*/ |
*/ |
type = ICMP_REDIRECT; |
type = ICMP_REDIRECT; |
code = ICMP_REDIRECT_HOST; |
code = ICMP_REDIRECT_HOST; |
#ifdef DIAGNOSTIC |
|
if (ipprintfs) |
|
printf("redirect (%d) to %x\n", code, |
|
(u_int32_t)dest); |
|
#endif |
|
} |
} |
} |
} |
|
|
error = ip_output(m, NULL, &ipforward_rt, |
error = ip_output(m, NULL, &ipforward_rt, |
(IP_FORWARDING | (ip_directedbcast ? IP_ALLOWBROADCAST : 0)), |
(IP_FORWARDING | (ip_directedbcast ? IP_ALLOWBROADCAST : 0)), |
(struct ip_moptions *)NULL, (struct socket *)NULL); |
NULL, NULL); |
|
|
if (error) |
if (error) |
IP_STATINC(IP_STAT_CANTFORWARD); |
IP_STATINC(IP_STAT_CANTFORWARD); |
Line 1973 ip_forward(struct mbuf *m, int srcrt) |
|
Line 1288 ip_forward(struct mbuf *m, int srcrt) |
|
|
|
if ((rt = rtcache_validate(&ipforward_rt)) != NULL) |
if ((rt = rtcache_validate(&ipforward_rt)) != NULL) |
destmtu = rt->rt_ifp->if_mtu; |
destmtu = rt->rt_ifp->if_mtu; |
|
#ifdef IPSEC |
#if defined(IPSEC) || defined(FAST_IPSEC) |
(void)ipsec4_forward(mcopy, &destmtu); |
{ |
|
/* |
|
* If the packet is routed over IPsec tunnel, tell the |
|
* originator the tunnel MTU. |
|
* tunnel MTU = if MTU - sizeof(IP) - ESP/AH hdrsiz |
|
* XXX quickhack!!! |
|
*/ |
|
|
|
struct secpolicy *sp; |
|
int ipsecerror; |
|
size_t ipsechdr; |
|
struct route *ro; |
|
|
|
sp = ipsec4_getpolicybyaddr(mcopy, |
|
IPSEC_DIR_OUTBOUND, IP_FORWARDING, |
|
&ipsecerror); |
|
|
|
if (sp != NULL) { |
|
/* count IPsec header size */ |
|
ipsechdr = ipsec4_hdrsiz(mcopy, |
|
IPSEC_DIR_OUTBOUND, NULL); |
|
|
|
/* |
|
* find the correct route for outer IPv4 |
|
* header, compute tunnel MTU. |
|
*/ |
|
|
|
if (sp->req != NULL |
|
&& sp->req->sav != NULL |
|
&& sp->req->sav->sah != NULL) { |
|
ro = &sp->req->sav->sah->sa_route; |
|
rt = rtcache_validate(ro); |
|
if (rt && rt->rt_ifp) { |
|
destmtu = |
|
rt->rt_rmx.rmx_mtu ? |
|
rt->rt_rmx.rmx_mtu : |
|
rt->rt_ifp->if_mtu; |
|
destmtu -= ipsechdr; |
|
} |
|
} |
|
|
|
#ifdef IPSEC |
|
key_freesp(sp); |
|
#else |
|
KEY_FREESP(&sp); |
|
#endif |
#endif |
} |
|
} |
|
#endif /*defined(IPSEC) || defined(FAST_IPSEC)*/ |
|
IP_STATINC(IP_STAT_CANTFRAG); |
IP_STATINC(IP_STAT_CANTFRAG); |
break; |
break; |
|
|
case ENOBUFS: |
case ENOBUFS: |
#if 1 |
|
/* |
/* |
* a router should not generate ICMP_SOURCEQUENCH as |
* Do not generate ICMP_SOURCEQUENCH as required in RFC 1812, |
* required in RFC1812 Requirements for IP Version 4 Routers. |
* Requirements for IP Version 4 Routers. Source quench can |
* source quench could be a big problem under DoS attacks, |
* big problem under DoS attacks or if the underlying |
* or if the underlying interface is rate-limited. |
* interface is rate-limited. |
*/ |
*/ |
if (mcopy) |
if (mcopy) |
m_freem(mcopy); |
m_freem(mcopy); |
return; |
return; |
#else |
|
type = ICMP_SOURCEQUENCH; |
|
code = 0; |
|
break; |
|
#endif |
|
} |
} |
icmp_error(mcopy, type, code, dest, destmtu); |
icmp_error(mcopy, type, code, dest, destmtu); |
} |
} |
|
|
ip_savecontrol(struct inpcb *inp, struct mbuf **mp, struct ip *ip, |
ip_savecontrol(struct inpcb *inp, struct mbuf **mp, struct ip *ip, |
struct mbuf *m) |
struct mbuf *m) |
{ |
{ |
|
struct socket *so = inp->inp_socket; |
if (inp->inp_socket->so_options & SO_TIMESTAMP) { |
ifnet_t *ifp = m->m_pkthdr.rcvif; |
|
int inpflags = inp->inp_flags; |
|
|
|
if (so->so_options & SO_TIMESTAMP |
|
#ifdef SO_OTIMESTAMP |
|
|| so->so_options & SO_OTIMESTAMP |
|
#endif |
|
) { |
struct timeval tv; |
struct timeval tv; |
|
|
microtime(&tv); |
microtime(&tv); |
|
#ifdef SO_OTIMESTAMP |
|
if (so->so_options & SO_OTIMESTAMP) { |
|
struct timeval50 tv50; |
|
timeval_to_timeval50(&tv, &tv50); |
|
*mp = sbcreatecontrol((void *) &tv50, sizeof(tv50), |
|
SCM_OTIMESTAMP, SOL_SOCKET); |
|
} else |
|
#endif |
*mp = sbcreatecontrol((void *) &tv, sizeof(tv), |
*mp = sbcreatecontrol((void *) &tv, sizeof(tv), |
SCM_TIMESTAMP, SOL_SOCKET); |
SCM_TIMESTAMP, SOL_SOCKET); |
if (*mp) |
if (*mp) |
mp = &(*mp)->m_next; |
mp = &(*mp)->m_next; |
} |
} |
if (inp->inp_flags & INP_RECVDSTADDR) { |
if (inpflags & INP_RECVDSTADDR) { |
*mp = sbcreatecontrol((void *) &ip->ip_dst, |
*mp = sbcreatecontrol((void *) &ip->ip_dst, |
sizeof(struct in_addr), IP_RECVDSTADDR, IPPROTO_IP); |
sizeof(struct in_addr), IP_RECVDSTADDR, IPPROTO_IP); |
if (*mp) |
if (*mp) |
mp = &(*mp)->m_next; |
mp = &(*mp)->m_next; |
} |
} |
#ifdef notyet |
if (inpflags & INP_RECVPKTINFO) { |
/* |
struct in_pktinfo ipi; |
* XXX |
ipi.ipi_addr = ip->ip_src; |
* Moving these out of udp_input() made them even more broken |
ipi.ipi_ifindex = ifp->if_index; |
* than they already were. |
*mp = sbcreatecontrol((void *) &ipi, |
* - fenner@parc.xerox.com |
sizeof(ipi), IP_RECVPKTINFO, IPPROTO_IP); |
*/ |
|
/* options were tossed already */ |
|
if (inp->inp_flags & INP_RECVOPTS) { |
|
*mp = sbcreatecontrol((void *) opts_deleted_above, |
|
sizeof(struct in_addr), IP_RECVOPTS, IPPROTO_IP); |
|
if (*mp) |
if (*mp) |
mp = &(*mp)->m_next; |
mp = &(*mp)->m_next; |
} |
} |
/* ip_srcroute doesn't do what we want here, need to fix */ |
if (inpflags & INP_PKTINFO) { |
if (inp->inp_flags & INP_RECVRETOPTS) { |
struct in_pktinfo ipi; |
*mp = sbcreatecontrol((void *) ip_srcroute(), |
ipi.ipi_addr = ip->ip_dst; |
sizeof(struct in_addr), IP_RECVRETOPTS, IPPROTO_IP); |
ipi.ipi_ifindex = ifp->if_index; |
|
*mp = sbcreatecontrol((void *) &ipi, |
|
sizeof(ipi), IP_PKTINFO, IPPROTO_IP); |
if (*mp) |
if (*mp) |
mp = &(*mp)->m_next; |
mp = &(*mp)->m_next; |
} |
} |
#endif |
if (inpflags & INP_RECVIF) { |
if (inp->inp_flags & INP_RECVIF) { |
|
struct sockaddr_dl sdl; |
struct sockaddr_dl sdl; |
|
|
sockaddr_dl_init(&sdl, sizeof(sdl), |
sockaddr_dl_init(&sdl, sizeof(sdl), ifp ? |
(m->m_pkthdr.rcvif != NULL) |
ifp->if_index : 0, 0, NULL, 0, NULL, 0); |
? m->m_pkthdr.rcvif->if_index |
|
: 0, |
|
0, NULL, 0, NULL, 0); |
|
*mp = sbcreatecontrol(&sdl, sdl.sdl_len, IP_RECVIF, IPPROTO_IP); |
*mp = sbcreatecontrol(&sdl, sdl.sdl_len, IP_RECVIF, IPPROTO_IP); |
if (*mp) |
if (*mp) |
mp = &(*mp)->m_next; |
mp = &(*mp)->m_next; |
} |
} |
|
if (inpflags & INP_RECVTTL) { |
|
*mp = sbcreatecontrol((void *) &ip->ip_ttl, |
|
sizeof(uint8_t), IP_RECVTTL, IPPROTO_IP); |
|
if (*mp) |
|
mp = &(*mp)->m_next; |
|
} |
} |
} |
|
|
/* |
/* |
Line 2119 sysctl_net_inet_ip_forwsrcrt(SYSCTLFN_AR |
|
Line 1394 sysctl_net_inet_ip_forwsrcrt(SYSCTLFN_AR |
|
if (error || newp == NULL) |
if (error || newp == NULL) |
return (error); |
return (error); |
|
|
if (kauth_authorize_network(l->l_cred, KAUTH_NETWORK_FORWSRCRT, |
error = kauth_authorize_network(l->l_cred, KAUTH_NETWORK_FORWSRCRT, |
0, NULL, NULL, NULL)) |
0, NULL, NULL, NULL); |
return (EPERM); |
if (error) |
|
return (error); |
|
|
ip_forwsrcrt = tmp; |
ip_forwsrcrt = tmp; |
|
|
Line 2157 sysctl_net_inet_ip_pmtudto(SYSCTLFN_ARGS |
|
Line 1433 sysctl_net_inet_ip_pmtudto(SYSCTLFN_ARGS |
|
return (0); |
return (0); |
} |
} |
|
|
#ifdef GATEWAY |
|
/* |
|
* sysctl helper routine for net.inet.ip.maxflows. |
|
*/ |
|
static int |
|
sysctl_net_inet_ip_maxflows(SYSCTLFN_ARGS) |
|
{ |
|
int error; |
|
|
|
error = sysctl_lookup(SYSCTLFN_CALL(rnode)); |
|
if (error || newp == NULL) |
|
return (error); |
|
|
|
mutex_enter(softnet_lock); |
|
KERNEL_LOCK(1, NULL); |
|
|
|
ipflow_prune(); |
|
|
|
KERNEL_UNLOCK_ONE(NULL); |
|
mutex_exit(softnet_lock); |
|
|
|
return (0); |
|
} |
|
|
|
static int |
|
sysctl_net_inet_ip_hashsize(SYSCTLFN_ARGS) |
|
{ |
|
int error, tmp; |
|
struct sysctlnode node; |
|
|
|
node = *rnode; |
|
tmp = ip_hashsize; |
|
node.sysctl_data = &tmp; |
|
error = sysctl_lookup(SYSCTLFN_CALL(&node)); |
|
if (error || newp == NULL) |
|
return (error); |
|
|
|
if ((tmp & (tmp - 1)) == 0 && tmp != 0) { |
|
/* |
|
* Can only fail due to malloc() |
|
*/ |
|
mutex_enter(softnet_lock); |
|
KERNEL_LOCK(1, NULL); |
|
|
|
error = ipflow_invalidate_all(tmp); |
|
|
|
KERNEL_UNLOCK_ONE(NULL); |
|
mutex_exit(softnet_lock); |
|
|
|
} else { |
|
/* |
|
* EINVAL if not a power of 2 |
|
*/ |
|
error = EINVAL; |
|
} |
|
|
|
return error; |
|
} |
|
#endif /* GATEWAY */ |
|
|
|
static int |
static int |
sysctl_net_inet_ip_stats(SYSCTLFN_ARGS) |
sysctl_net_inet_ip_stats(SYSCTLFN_ARGS) |
{ |
{ |
Line 2224 sysctl_net_inet_ip_stats(SYSCTLFN_ARGS) |
|
Line 1440 sysctl_net_inet_ip_stats(SYSCTLFN_ARGS) |
|
return (NETSTAT_SYSCTL(ipstat_percpu, IP_NSTATS)); |
return (NETSTAT_SYSCTL(ipstat_percpu, IP_NSTATS)); |
} |
} |
|
|
SYSCTL_SETUP(sysctl_net_inet_ip_setup, "sysctl net.inet.ip subtree setup") |
static void |
|
sysctl_net_inet_ip_setup(struct sysctllog **clog) |
{ |
{ |
extern int subnetsarelocal, hostzeroisbroadcast; |
|
|
|
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, |
sysctl_createv(clog, 0, NULL, NULL, |
CTLFLAG_PERMANENT, |
CTLFLAG_PERMANENT, |
CTLTYPE_NODE, "inet", |
CTLTYPE_NODE, "inet", |
Line 2298 SYSCTL_SETUP(sysctl_net_inet_ip_setup, " |
|
Line 1508 SYSCTL_SETUP(sysctl_net_inet_ip_setup, " |
|
NULL, 0, &ip_allowsrcrt, 0, |
NULL, 0, &ip_allowsrcrt, 0, |
CTL_NET, PF_INET, IPPROTO_IP, |
CTL_NET, PF_INET, IPPROTO_IP, |
IPCTL_ALLOWSRCRT, CTL_EOL); |
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, |
sysctl_createv(clog, 0, NULL, NULL, |
CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
CTLTYPE_INT, "mtudisc", |
CTLTYPE_INT, "mtudisc", |
Line 2331 SYSCTL_SETUP(sysctl_net_inet_ip_setup, " |
|
Line 1534 SYSCTL_SETUP(sysctl_net_inet_ip_setup, " |
|
CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
CTLTYPE_INT, "mtudisctimeout", |
CTLTYPE_INT, "mtudisctimeout", |
SYSCTL_DESCR("Lifetime of a Path MTU Discovered route"), |
SYSCTL_DESCR("Lifetime of a Path MTU Discovered route"), |
sysctl_net_inet_ip_pmtudto, 0, &ip_mtudisc_timeout, 0, |
sysctl_net_inet_ip_pmtudto, 0, (void *)&ip_mtudisc_timeout, 0, |
CTL_NET, PF_INET, IPPROTO_IP, |
CTL_NET, PF_INET, IPPROTO_IP, |
IPCTL_MTUDISCTIMEOUT, CTL_EOL); |
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); |
|
sysctl_createv(clog, 0, NULL, NULL, |
|
CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
|
CTLTYPE_INT, "hashsize", |
|
SYSCTL_DESCR("Size of hash table for fast forwarding (IPv4)"), |
|
sysctl_net_inet_ip_hashsize, 0, &ip_hashsize, 0, |
|
CTL_NET, PF_INET, IPPROTO_IP, |
|
CTL_CREATE, 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 |
#if NGIF > 0 |
sysctl_createv(clog, 0, NULL, NULL, |
sysctl_createv(clog, 0, NULL, NULL, |
CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
Line 2384 SYSCTL_SETUP(sysctl_net_inet_ip_setup, " |
|
Line 1564 SYSCTL_SETUP(sysctl_net_inet_ip_setup, " |
|
CTL_NET, PF_INET, IPPROTO_IP, |
CTL_NET, PF_INET, IPPROTO_IP, |
IPCTL_LOWPORTMAX, CTL_EOL); |
IPCTL_LOWPORTMAX, CTL_EOL); |
#endif /* IPNOPRIVPORTS */ |
#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 |
#if NGRE > 0 |
sysctl_createv(clog, 0, NULL, NULL, |
sysctl_createv(clog, 0, NULL, NULL, |
CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
Line 2430 SYSCTL_SETUP(sysctl_net_inet_ip_setup, " |
|
Line 1602 SYSCTL_SETUP(sysctl_net_inet_ip_setup, " |
|
sysctl_net_inet_ip_stats, 0, NULL, 0, |
sysctl_net_inet_ip_stats, 0, NULL, 0, |
CTL_NET, PF_INET, IPPROTO_IP, IPCTL_STATS, |
CTL_NET, PF_INET, IPPROTO_IP, IPCTL_STATS, |
CTL_EOL); |
CTL_EOL); |
|
|
|
/* anonportalgo RFC6056 subtree */ |
|
const struct sysctlnode *portalgo_node; |
|
sysctl_createv(clog, 0, NULL, &portalgo_node, |
|
CTLFLAG_PERMANENT, |
|
CTLTYPE_NODE, "anonportalgo", |
|
SYSCTL_DESCR("Anonymous Port Algorithm Selection (RFC 6056)"), |
|
NULL, 0, NULL, 0, |
|
CTL_NET, PF_INET, IPPROTO_IP, CTL_CREATE, CTL_EOL); |
|
sysctl_createv(clog, 0, &portalgo_node, NULL, |
|
CTLFLAG_PERMANENT, |
|
CTLTYPE_STRING, "available", |
|
SYSCTL_DESCR("available algorithms"), |
|
sysctl_portalgo_available, 0, NULL, PORTALGO_MAXLEN, |
|
CTL_CREATE, CTL_EOL); |
|
sysctl_createv(clog, 0, &portalgo_node, NULL, |
|
CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
|
CTLTYPE_STRING, "selected", |
|
SYSCTL_DESCR("selected algorithm"), |
|
sysctl_portalgo_selected4, 0, NULL, PORTALGO_MAXLEN, |
|
CTL_CREATE, CTL_EOL); |
|
sysctl_createv(clog, 0, &portalgo_node, NULL, |
|
CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
|
CTLTYPE_STRUCT, "reserve", |
|
SYSCTL_DESCR("bitmap of reserved ports"), |
|
sysctl_portalgo_reserve4, 0, NULL, 0, |
|
CTL_CREATE, CTL_EOL); |
} |
} |
|
|
void |
void |