/*
- * Copyright (c) 2000-2009 Apple Inc. All rights reserved.
+ * Copyright (c) 2000-2018 Apple Inc. All rights reserved.
*
* @APPLE_OSREFERENCE_LICENSE_HEADER_START@
- *
+ *
* This file contains Original Code and/or Modifications of Original Code
* as defined in and that are subject to the Apple Public Source License
* Version 2.0 (the 'License'). You may not use this file except in
* unlawful or unlicensed copies of an Apple operating system, or to
* circumvent, violate, or enable the circumvention or violation of, any
* terms of an Apple operating system software license agreement.
- *
+ *
* Please obtain a copy of the License at
* http://www.opensource.apple.com/apsl/ and read it before using this file.
- *
+ *
* The Original Code and all software distributed under the License are
* distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
* EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
* FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
* Please see the License for the specific language governing rights and
* limitations under the License.
- *
+ *
* @APPLE_OSREFERENCE_LICENSE_HEADER_END@
*/
/*
#include <sys/kernel.h>
#include <sys/malloc.h>
#include <sys/mbuf.h>
+#include <sys/mcache.h>
#include <sys/proc.h>
#include <sys/domain.h>
#include <sys/protosw.h>
#include <pexpert/pexpert.h>
#include <net/if.h>
+#include <net/net_api_stats.h>
#include <net/route.h>
#define _IP_VHL
#include <netinet/in.h>
#include <netinet/in_systm.h>
+#include <netinet/in_tclass.h>
#include <netinet/ip.h>
#include <netinet/in_pcb.h>
#include <netinet/in_var.h>
#include <netinet/ip_var.h>
-#include <netinet/ip_mroute.h>
#if INET6
#include <netinet6/in6_pcb.h>
int rip_bind(struct socket *, struct sockaddr *, struct proc *);
int rip_connect(struct socket *, struct sockaddr *, struct proc *);
int rip_shutdown(struct socket *);
-
-#if IPSEC
-extern int ipsec_bypass;
-#endif
struct inpcbhead ripcb;
struct inpcbinfo ripcbinfo;
/* control hooks for ipfw and dummynet */
#if IPFIREWALL
ip_fw_ctl_t *ip_fw_ctl_ptr;
+#endif /* IPFIREWALL */
#if DUMMYNET
ip_dn_ctl_t *ip_dn_ctl_ptr;
#endif /* DUMMYNET */
-#endif /* IPFIREWALL */
/*
* Nominal space allocated to a raw ip socket.
* Initialize raw connection block q.
*/
void
-rip_init()
+rip_init(struct protosw *pp, struct domain *dp)
{
- struct inpcbinfo *pcbinfo;
+#pragma unused(dp)
+ static int rip_initialized = 0;
+ struct inpcbinfo *pcbinfo;
+
+ VERIFY((pp->pr_flags & (PR_INITIALIZED|PR_ATTACHED)) == PR_ATTACHED);
+
+ if (rip_initialized)
+ return;
+ rip_initialized = 1;
LIST_INIT(&ripcb);
- ripcbinfo.listhead = &ripcb;
+ ripcbinfo.ipi_listhead = &ripcb;
/*
* XXX We don't use the hash list for raw IP, but it's easier
* to allocate a one entry hash list than it is to check all
- * over the place for hashbase == NULL.
+ * over the place for ipi_hashbase == NULL.
*/
- ripcbinfo.hashbase = hashinit(1, M_PCB, &ripcbinfo.hashmask);
- ripcbinfo.porthashbase = hashinit(1, M_PCB, &ripcbinfo.porthashmask);
+ ripcbinfo.ipi_hashbase = hashinit(1, M_PCB, &ripcbinfo.ipi_hashmask);
+ ripcbinfo.ipi_porthashbase = hashinit(1, M_PCB, &ripcbinfo.ipi_porthashmask);
- ripcbinfo.ipi_zone = (void *) zinit(sizeof(struct inpcb),
- (4096 * sizeof(struct inpcb)),
- 4096, "ripzone");
+ ripcbinfo.ipi_zone = zinit(sizeof(struct inpcb),
+ (4096 * sizeof(struct inpcb)), 4096, "ripzone");
pcbinfo = &ripcbinfo;
/*
* allocate lock group attribute and group for udp pcb mutexes
*/
- pcbinfo->mtx_grp_attr = lck_grp_attr_alloc_init();
+ pcbinfo->ipi_lock_grp_attr = lck_grp_attr_alloc_init();
+ pcbinfo->ipi_lock_grp = lck_grp_alloc_init("ripcb", pcbinfo->ipi_lock_grp_attr);
- pcbinfo->mtx_grp = lck_grp_alloc_init("ripcb", pcbinfo->mtx_grp_attr);
-
/*
* allocate the lock attribute for udp pcb mutexes
*/
- pcbinfo->mtx_attr = lck_attr_alloc_init();
-
- if ((pcbinfo->mtx = lck_rw_alloc_init(pcbinfo->mtx_grp, pcbinfo->mtx_attr)) == NULL)
- return; /* pretty much dead if this fails... */
+ pcbinfo->ipi_lock_attr = lck_attr_alloc_init();
+ if ((pcbinfo->ipi_lock = lck_rw_alloc_init(pcbinfo->ipi_lock_grp,
+ pcbinfo->ipi_lock_attr)) == NULL) {
+ panic("%s: unable to allocate PCB lock\n", __func__);
+ /* NOTREACHED */
+ }
+ in_pcbinfo_attach(&ripcbinfo);
}
static struct sockaddr_in ripsrc = { sizeof(ripsrc), AF_INET , 0, {0}, {0,0,0,0,0,0,0,0,} };
* mbuf chain.
*/
void
-rip_input(m, iphlen)
- struct mbuf *m;
- int iphlen;
+rip_input(struct mbuf *m, int iphlen)
{
- register struct ip *ip = mtod(m, struct ip *);
- register struct inpcb *inp;
+ struct ip *ip = mtod(m, struct ip *);
+ struct inpcb *inp;
struct inpcb *last = 0;
struct mbuf *opts = 0;
- int skipit;
+ int skipit = 0, ret = 0;
+ struct ifnet *ifp = m->m_pkthdr.rcvif;
+
+ /* Expect 32-bit aligned data pointer on strict-align platforms */
+ MBUF_STRICT_DATA_ALIGNMENT_CHECK_32(m);
ripsrc.sin_addr = ip->ip_src;
- lck_rw_lock_shared(ripcbinfo.mtx);
+ lck_rw_lock_shared(ripcbinfo.ipi_lock);
LIST_FOREACH(inp, &ripcb, inp_list) {
#if INET6
if ((inp->inp_vflag & INP_IPV4) == 0)
if (inp->inp_faddr.s_addr &&
inp->inp_faddr.s_addr != ip->ip_src.s_addr)
continue;
+ if (inp_restricted_recv(inp, ifp))
+ continue;
if (last) {
struct mbuf *n = m_copy(m, 0, (int)M_COPYALL);
-
-#if IPSEC
- /* check AH/ESP integrity. */
+
skipit = 0;
- if (ipsec_bypass == 0 && n) {
- if (ipsec4_in_reject_so(n, last->inp_socket)) {
- m_freem(n);
- IPSEC_STAT_INCREMENT(ipsecstat.in_polvio);
- /* do not inject data to pcb */
- skipit = 1;
- }
- }
-#endif /*IPSEC*/
+
+#if NECP
+ if (n && !necp_socket_is_allowed_to_send_recv_v4(last, 0, 0,
+ &ip->ip_dst, &ip->ip_src, ifp, NULL, NULL)) {
+ m_freem(n);
+ /* do not inject data to pcb */
+ skipit = 1;
+ }
+#endif /* NECP */
#if CONFIG_MACF_NET
if (n && skipit == 0) {
if (mac_inpcb_check_deliver(last, n, AF_INET,
- SOCK_RAW) != 0)
+ SOCK_RAW) != 0) {
+ m_freem(n);
skipit = 1;
+ }
}
#endif
if (n && skipit == 0) {
int error = 0;
- if (last->inp_flags & INP_CONTROLOPTS ||
- last->inp_socket->so_options & SO_TIMESTAMP)
- ip_savecontrol(last, &opts, ip, n);
+ if ((last->inp_flags & INP_CONTROLOPTS) != 0 ||
+ (last->inp_socket->so_options & SO_TIMESTAMP) != 0 ||
+ (last->inp_socket->so_options & SO_TIMESTAMP_MONOTONIC) != 0) {
+ ret = ip_savecontrol(last, &opts, ip, n);
+ if (ret != 0) {
+ m_freem(n);
+ m_freem(opts);
+ last = inp;
+ continue;
+ }
+ }
if (last->inp_flags & INP_STRIPHDR) {
n->m_len -= iphlen;
n->m_pkthdr.len -= iphlen;
n->m_data += iphlen;
}
-// ###LOCK need to lock that socket?
+ so_recv_data_stat(last->inp_socket, m, 0);
if (sbappendaddr(&last->inp_socket->so_rcv,
(struct sockaddr *)&ripsrc, n,
opts, &error) != 0) {
sorwakeup(last->inp_socket);
- }
- else {
+ } else {
if (error) {
/* should notify about lost packet */
- kprintf("rip_input can't append to socket\n");
+ ipstat.ips_raw_sappend_fail++;
}
}
opts = 0;
}
last = inp;
}
- lck_rw_done(ripcbinfo.mtx);
-#if IPSEC
- /* check AH/ESP integrity. */
+
skipit = 0;
- if (ipsec_bypass == 0 && last) {
- if (ipsec4_in_reject_so(m, last->inp_socket)) {
- m_freem(m);
- IPSEC_STAT_INCREMENT(ipsecstat.in_polvio);
- OSAddAtomic(1, &ipstat.ips_delivered);
- /* do not inject data to pcb */
- skipit = 1;
- }
- }
-#endif /*IPSEC*/
+#if NECP
+ if (last && !necp_socket_is_allowed_to_send_recv_v4(last, 0, 0,
+ &ip->ip_dst, &ip->ip_src, ifp, NULL, NULL)) {
+ m_freem(m);
+ OSAddAtomic(1, &ipstat.ips_delivered);
+ /* do not inject data to pcb */
+ skipit = 1;
+ }
+#endif /* NECP */
#if CONFIG_MACF_NET
if (last && skipit == 0) {
- if (mac_inpcb_check_deliver(last, m, AF_INET, SOCK_RAW) != 0)
+ if (mac_inpcb_check_deliver(last, m, AF_INET, SOCK_RAW) != 0) {
skipit = 1;
+ m_freem(m);
+ }
}
#endif
if (skipit == 0) {
if (last) {
- if (last->inp_flags & INP_CONTROLOPTS ||
- last->inp_socket->so_options & SO_TIMESTAMP)
- ip_savecontrol(last, &opts, ip, m);
+ if ((last->inp_flags & INP_CONTROLOPTS) != 0 ||
+ (last->inp_socket->so_options & SO_TIMESTAMP) != 0 ||
+ (last->inp_socket->so_options & SO_TIMESTAMP_MONOTONIC) != 0) {
+ ret = ip_savecontrol(last, &opts, ip, m);
+ if (ret != 0) {
+ m_freem(m);
+ m_freem(opts);
+ goto unlock;
+ }
+ }
if (last->inp_flags & INP_STRIPHDR) {
m->m_len -= iphlen;
m->m_pkthdr.len -= iphlen;
m->m_data += iphlen;
}
+ so_recv_data_stat(last->inp_socket, m, 0);
if (sbappendaddr(&last->inp_socket->so_rcv,
(struct sockaddr *)&ripsrc, m, opts, NULL) != 0) {
sorwakeup(last->inp_socket);
} else {
- kprintf("rip_input(2) can't append to socket\n");
+ ipstat.ips_raw_sappend_fail++;
}
} else {
m_freem(m);
OSAddAtomic(-1, &ipstat.ips_delivered);
}
}
+unlock:
+ /*
+ * Keep the list locked because socket filter may force the socket lock
+ * to be released when calling sbappendaddr() -- see rdar://7627704
+ */
+ lck_rw_done(ripcbinfo.ipi_lock);
}
/*
* Tack on options user may have setup with control call.
*/
int
-rip_output(m, so, dst)
- register struct mbuf *m;
- struct socket *so;
- u_int32_t dst;
+rip_output(
+ struct mbuf *m,
+ struct socket *so,
+ u_int32_t dst,
+ struct mbuf *control)
{
- register struct ip *ip;
- register struct inpcb *inp = sotoinpcb(so);
+ struct ip *ip;
+ struct inpcb *inp = sotoinpcb(so);
int flags = (so->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST;
struct ip_out_args ipoa;
+ struct ip_moptions *imo;
int error = 0;
- /* If socket was bound to an ifindex, tell ip_output about it */
- ipoa.ipoa_ifscope = (inp->inp_flags & INP_BOUND_IF) ?
- inp->inp_boundif : IFSCOPE_NONE;
+ bzero(&ipoa, sizeof(ipoa));
+ ipoa.ipoa_boundif = IFSCOPE_NONE;
+ ipoa.ipoa_flags = IPOAF_SELECT_SRCIF;
+
+ int sotc = SO_TC_UNSPEC;
+ int netsvctype = _NET_SERVICE_TYPE_UNSPEC;
+
+
+ if (control != NULL) {
+ sotc = so_tc_from_control(control, &netsvctype);
+
+ m_freem(control);
+ control = NULL;
+ }
+ if (sotc == SO_TC_UNSPEC) {
+ sotc = so->so_traffic_class;
+ netsvctype = so->so_netsvctype;
+ }
+
+ if (inp == NULL
+#if NECP
+ || (necp_socket_should_use_flow_divert(inp))
+#endif /* NECP */
+ ) {
+ if (m != NULL)
+ m_freem(m);
+ VERIFY(control == NULL);
+ return (inp == NULL ? EINVAL : EPROTOTYPE);
+ }
+
flags |= IP_OUTARGS;
+ /* If socket was bound to an ifindex, tell ip_output about it */
+ if (inp->inp_flags & INP_BOUND_IF) {
+ ipoa.ipoa_boundif = inp->inp_boundifp->if_index;
+ ipoa.ipoa_flags |= IPOAF_BOUND_IF;
+ }
+ if (INP_NO_CELLULAR(inp))
+ ipoa.ipoa_flags |= IPOAF_NO_CELLULAR;
+ if (INP_NO_EXPENSIVE(inp))
+ ipoa.ipoa_flags |= IPOAF_NO_EXPENSIVE;
+ if (INP_AWDL_UNRESTRICTED(inp))
+ ipoa.ipoa_flags |= IPOAF_AWDL_UNRESTRICTED;
+ ipoa.ipoa_sotc = sotc;
+ ipoa.ipoa_netsvctype = netsvctype;
+
+ if (inp->inp_flowhash == 0)
+ inp->inp_flowhash = inp_calc_flowhash(inp);
/*
* If the user handed us a complete IP packet, use it.
m_freem(m);
return(EMSGSIZE);
}
- M_PREPEND(m, sizeof(struct ip), M_WAIT);
+ M_PREPEND(m, sizeof(struct ip), M_WAIT, 1);
if (m == NULL)
return ENOBUFS;
ip = mtod(m, struct ip *);
m_freem(m);
return EINVAL;
}
- if (ip->ip_id == 0)
-#if RANDOM_IP_ID
+ if (ip->ip_id == 0 && !(rfc6864 && IP_OFF_IS_ATOMIC(ntohs(ip->ip_off))))
ip->ip_id = ip_randomid();
-#else
- ip->ip_id = htons(ip_id++);
-#endif
/* XXX prevent ip_output from overwriting header fields */
flags |= IP_RAWOUTPUT;
OSAddAtomic(1, &ipstat.ips_rawout);
}
+ if (inp->inp_laddr.s_addr != INADDR_ANY)
+ ipoa.ipoa_flags |= IPOAF_BOUND_SRCADDR;
+
+#if NECP
+ {
+ necp_kernel_policy_id policy_id;
+ u_int32_t route_rule_id;
+
+ /*
+ * We need a route to perform NECP route rule checks
+ */
+ if (net_qos_policy_restricted != 0 &&
+ ROUTE_UNUSABLE(&inp->inp_route)) {
+ struct sockaddr_in to;
+ struct sockaddr_in from;
+ struct in_addr laddr = ip->ip_src;
+
+ ROUTE_RELEASE(&inp->inp_route);
+
+ bzero(&from, sizeof(struct sockaddr_in));
+ from.sin_family = AF_INET;
+ from.sin_len = sizeof(struct sockaddr_in);
+ from.sin_addr = laddr;
+
+ bzero(&to, sizeof(struct sockaddr_in));
+ to.sin_family = AF_INET;
+ to.sin_len = sizeof(struct sockaddr_in);
+ to.sin_addr.s_addr = ip->ip_dst.s_addr;
+
+ if ((error = in_pcbladdr(inp, (struct sockaddr *)&to,
+ &laddr, ipoa.ipoa_boundif, NULL, 1)) != 0) {
+ printf("%s in_pcbladdr(%p) error %d\n",
+ __func__, inp, error);
+ m_freem(m);
+ return (error);
+ }
+
+ inp_update_necp_policy(inp, (struct sockaddr *)&from,
+ (struct sockaddr *)&to, ipoa.ipoa_boundif);
+ inp->inp_policyresult.results.qos_marking_gencount = 0;
+ }
+
+ if (!necp_socket_is_allowed_to_send_recv_v4(inp, 0, 0,
+ &ip->ip_src, &ip->ip_dst, NULL, &policy_id, &route_rule_id)) {
+ m_freem(m);
+ return(EHOSTUNREACH);
+ }
+
+ necp_mark_packet_from_socket(m, inp, policy_id, route_rule_id);
+
+ if (net_qos_policy_restricted != 0) {
+ struct ifnet *rt_ifp = NULL;
+
+ if (inp->inp_route.ro_rt != NULL)
+ rt_ifp = inp->inp_route.ro_rt->rt_ifp;
+
+ necp_socket_update_qos_marking(inp, inp->inp_route.ro_rt,
+ NULL, route_rule_id);
+ }
+ }
+#endif /* NECP */
+ if ((so->so_flags1 & SOF1_QOSMARKING_ALLOWED))
+ ipoa.ipoa_flags |= IPOAF_QOSMARKING_ALLOWED;
+
#if IPSEC
- if (ipsec_bypass == 0 && ipsec_setsocket(m, so) != 0) {
+ if (inp->inp_sp != NULL && ipsec_setsocket(m, so) != 0) {
m_freem(m);
return ENOBUFS;
}
#endif /*IPSEC*/
- if (inp->inp_route.ro_rt != NULL &&
- inp->inp_route.ro_rt->generation_id != route_generation) {
- rtfree(inp->inp_route.ro_rt);
- inp->inp_route.ro_rt = NULL;
- }
+ if (ROUTE_UNUSABLE(&inp->inp_route))
+ ROUTE_RELEASE(&inp->inp_route);
-#if PKT_PRIORITY
- if (soisbackground(so))
- m_prio_background(m);
-#endif /* PKT_PRIORITY */
+ set_packet_service_class(m, so, sotc, 0);
+ m->m_pkthdr.pkt_flowsrc = FLOWSRC_INPCB;
+ m->m_pkthdr.pkt_flowid = inp->inp_flowhash;
+ m->m_pkthdr.pkt_flags |= (PKTF_FLOW_ID | PKTF_FLOW_LOCALSRC |
+ PKTF_FLOW_RAWSOCK);
+ m->m_pkthdr.pkt_proto = inp->inp_ip_p;
#if CONFIG_MACF_NET
mac_mbuf_label_associate_inpcb(inp, m);
#endif
+ imo = inp->inp_moptions;
+ if (imo != NULL)
+ IMO_ADDREF(imo);
/*
* The domain lock is held across ip_output, so it is okay
* to pass the PCB cached route pointer directly to IP and
* the modules beneath it.
*/
+ // TODO: PASS DOWN ROUTE RULE ID
error = ip_output(m, inp->inp_options, &inp->inp_route, flags,
- inp->inp_moptions, &ipoa);
+ imo, &ipoa);
+
+ if (imo != NULL)
+ IMO_REMREF(imo);
+
+ if (inp->inp_route.ro_rt != NULL) {
+ struct rtentry *rt = inp->inp_route.ro_rt;
+ struct ifnet *outif;
+
+ if ((rt->rt_flags & (RTF_MULTICAST|RTF_BROADCAST)) ||
+ inp->inp_socket == NULL ||
+ !(inp->inp_socket->so_state & SS_ISCONNECTED)) {
+ rt = NULL; /* unusable */
+ }
+ /*
+ * Always discard the cached route for unconnected
+ * socket or if it is a multicast route.
+ */
+ if (rt == NULL)
+ ROUTE_RELEASE(&inp->inp_route);
+
+ /*
+ * If this is a connected socket and the destination
+ * route is unicast, update outif with that of the
+ * route interface used by IP.
+ */
+ if (rt != NULL &&
+ (outif = rt->rt_ifp) != inp->inp_last_outifp) {
+ inp->inp_last_outifp = outif;
+ }
+ } else {
+ ROUTE_RELEASE(&inp->inp_route);
+ }
-#if IFNET_ROUTE_REFCNT
/*
- * Always discard the cached route for unconnected socket
- * or if it is a non-unicast route.
+ * If output interface was cellular/expensive, and this socket is
+ * denied access to it, generate an event.
*/
- if (inp->inp_route.ro_rt != NULL &&
- ((inp->inp_route.ro_rt->rt_flags & (RTF_MULTICAST|RTF_BROADCAST)) ||
- inp->inp_socket == NULL ||
- inp->inp_socket->so_state != SS_ISCONNECTED)) {
- rtfree(inp->inp_route.ro_rt);
- inp->inp_route.ro_rt = NULL;
- }
-#endif /* IFNET_ROUTE_REFCNT */
+ if (error != 0 && (ipoa.ipoa_retflags & IPOARF_IFDENIED) &&
+ (INP_NO_CELLULAR(inp) || INP_NO_EXPENSIVE(inp)))
+ soevent(so, (SO_FILT_HINT_LOCKED|SO_FILT_HINT_IFDENIED));
return (error);
}
load_ipfw(void)
{
kern_return_t err;
-
+
ipfw_init();
-
+
#if DUMMYNET
if (!DUMMYNET_LOADED)
ip_dn_init();
#endif /* DUMMYNET */
err = 0;
-
+
return err == 0 && ip_fw_ctl_ptr == NULL ? -1 : err;
}
#endif /* IPFIREWALL */
* Raw IP socket option processing.
*/
int
-rip_ctloutput(so, sopt)
- struct socket *so;
- struct sockopt *sopt;
+rip_ctloutput(struct socket *so, struct sockopt *sopt)
{
struct inpcb *inp = sotoinpcb(so);
int error, optval;
- if (sopt->sopt_level != IPPROTO_IP)
+ /* Allow <SOL_SOCKET,SO_FLUSH> at this level */
+ if (sopt->sopt_level != IPPROTO_IP &&
+ !(sopt->sopt_level == SOL_SOCKET && sopt->sopt_name == SO_FLUSH))
return (EINVAL);
error = 0;
error = sooptcopyout(sopt, &optval, sizeof optval);
break;
- case IP_STRIPHDR:
- optval = inp->inp_flags & INP_STRIPHDR;
- error = sooptcopyout(sopt, &optval, sizeof optval);
- break;
+ case IP_STRIPHDR:
+ optval = inp->inp_flags & INP_STRIPHDR;
+ error = sooptcopyout(sopt, &optval, sizeof optval);
+ break;
#if IPFIREWALL
case IP_FW_ADD:
#if DUMMYNET
case IP_DUMMYNET_GET:
+ if (!DUMMYNET_LOADED)
+ ip_dn_init();
if (DUMMYNET_LOADED)
error = ip_dn_ctl_ptr(sopt);
else
break ;
#endif /* DUMMYNET */
-#if MROUTING
- case MRT_INIT:
- case MRT_DONE:
- case MRT_ADD_VIF:
- case MRT_DEL_VIF:
- case MRT_ADD_MFC:
- case MRT_DEL_MFC:
- case MRT_VERSION:
- case MRT_ASSERT:
- error = ip_mrouter_get(so, sopt);
- break;
-#endif /* MROUTING */
-
default:
error = ip_ctloutput(so, sopt);
break;
inp->inp_flags &= ~INP_HDRINCL;
break;
- case IP_STRIPHDR:
- error = sooptcopyin(sopt, &optval, sizeof optval,
- sizeof optval);
- if (error)
- break;
- if (optval)
- inp->inp_flags |= INP_STRIPHDR;
- else
- inp->inp_flags &= ~INP_STRIPHDR;
- break;
-
+ case IP_STRIPHDR:
+ error = sooptcopyin(sopt, &optval, sizeof optval,
+ sizeof optval);
+ if (error)
+ break;
+ if (optval)
+ inp->inp_flags |= INP_STRIPHDR;
+ else
+ inp->inp_flags &= ~INP_STRIPHDR;
+ break;
#if IPFIREWALL
case IP_FW_ADD:
case IP_DUMMYNET_CONFIGURE:
case IP_DUMMYNET_DEL:
case IP_DUMMYNET_FLUSH:
+ if (!DUMMYNET_LOADED)
+ ip_dn_init();
if (DUMMYNET_LOADED)
error = ip_dn_ctl_ptr(sopt);
else
break ;
#endif
-#if MROUTING
- case IP_RSVP_ON:
- error = ip_rsvp_init(so);
- break;
-
- case IP_RSVP_OFF:
- error = ip_rsvp_done();
- break;
+ case SO_FLUSH:
+ if ((error = sooptcopyin(sopt, &optval, sizeof (optval),
+ sizeof (optval))) != 0)
+ break;
- /* XXX - should be combined */
- case IP_RSVP_VIF_ON:
- error = ip_rsvp_vif_init(so, sopt);
- break;
-
- case IP_RSVP_VIF_OFF:
- error = ip_rsvp_vif_done(so, sopt);
+ error = inp_flush(inp, optval);
break;
-
- case MRT_INIT:
- case MRT_DONE:
- case MRT_ADD_VIF:
- case MRT_DEL_VIF:
- case MRT_ADD_MFC:
- case MRT_DEL_MFC:
- case MRT_VERSION:
- case MRT_ASSERT:
- error = ip_mrouter_set(so, sopt);
- break;
-#endif /* MROUTING */
default:
error = ip_ctloutput(so, sopt);
rip_ctlinput(
int cmd,
struct sockaddr *sa,
- __unused void *vip)
+ __unused void *vip,
+ __unused struct ifnet *ifp)
{
- struct in_ifaddr *ia;
- struct ifnet *ifp;
- int err;
+ struct in_ifaddr *ia = NULL;
+ struct ifnet *iaifp = NULL;
+ int err = 0;
int flags, done = 0;
switch (cmd) {
lck_rw_lock_shared(in_ifaddr_rwlock);
for (ia = in_ifaddrhead.tqh_first; ia;
ia = ia->ia_link.tqe_next) {
- if (ia->ia_ifa.ifa_addr == sa
- && (ia->ia_flags & IFA_ROUTE)) {
+ IFA_LOCK(&ia->ia_ifa);
+ if (ia->ia_ifa.ifa_addr == sa &&
+ (ia->ia_flags & IFA_ROUTE)) {
done = 1;
- ifaref(&ia->ia_ifa);
+ IFA_ADDREF_LOCKED(&ia->ia_ifa);
+ IFA_UNLOCK(&ia->ia_ifa);
lck_rw_done(in_ifaddr_rwlock);
lck_mtx_lock(rnh_lock);
/*
*/
in_ifadown(&ia->ia_ifa, 1);
lck_mtx_unlock(rnh_lock);
- ifafree(&ia->ia_ifa);
+ IFA_REMREF(&ia->ia_ifa);
break;
}
+ IFA_UNLOCK(&ia->ia_ifa);
}
if (!done)
lck_rw_done(in_ifaddr_rwlock);
lck_rw_lock_shared(in_ifaddr_rwlock);
for (ia = in_ifaddrhead.tqh_first; ia;
ia = ia->ia_link.tqe_next) {
- if (ia->ia_ifa.ifa_addr == sa)
+ IFA_LOCK(&ia->ia_ifa);
+ if (ia->ia_ifa.ifa_addr == sa) {
+ /* keep it locked */
break;
+ }
+ IFA_UNLOCK(&ia->ia_ifa);
}
- if (ia == 0 || (ia->ia_flags & IFA_ROUTE)) {
+ if (ia == NULL || (ia->ia_flags & IFA_ROUTE) ||
+ (ia->ia_ifa.ifa_debug & IFD_NOTREADY)) {
+ if (ia != NULL)
+ IFA_UNLOCK(&ia->ia_ifa);
lck_rw_done(in_ifaddr_rwlock);
return;
}
- ifaref(&ia->ia_ifa);
+ IFA_ADDREF_LOCKED(&ia->ia_ifa);
+ IFA_UNLOCK(&ia->ia_ifa);
lck_rw_done(in_ifaddr_rwlock);
flags = RTF_UP;
- ifp = ia->ia_ifa.ifa_ifp;
+ iaifp = ia->ia_ifa.ifa_ifp;
- if ((ifp->if_flags & IFF_LOOPBACK)
- || (ifp->if_flags & IFF_POINTOPOINT))
+ if ((iaifp->if_flags & IFF_LOOPBACK)
+ || (iaifp->if_flags & IFF_POINTOPOINT))
flags |= RTF_HOST;
err = rtinit(&ia->ia_ifa, RTM_ADD, flags);
- if (err == 0)
+ if (err == 0) {
+ IFA_LOCK_SPIN(&ia->ia_ifa);
ia->ia_flags |= IFA_ROUTE;
- ifafree(&ia->ia_ifa);
+ IFA_UNLOCK(&ia->ia_ifa);
+ }
+ IFA_REMREF(&ia->ia_ifa);
break;
}
}
u_int32_t rip_sendspace = RIPSNDQ;
u_int32_t rip_recvspace = RIPRCVQ;
-SYSCTL_INT(_net_inet_raw, OID_AUTO, maxdgram, CTLFLAG_RW,
+SYSCTL_INT(_net_inet_raw, OID_AUTO, maxdgram, CTLFLAG_RW | CTLFLAG_LOCKED,
&rip_sendspace, 0, "Maximum outgoing raw IP datagram size");
-SYSCTL_INT(_net_inet_raw, OID_AUTO, recvspace, CTLFLAG_RW,
+SYSCTL_INT(_net_inet_raw, OID_AUTO, recvspace, CTLFLAG_RW | CTLFLAG_LOCKED,
&rip_recvspace, 0, "Maximum incoming raw IP datagram size");
+SYSCTL_UINT(_net_inet_raw, OID_AUTO, pcbcount, CTLFLAG_RD | CTLFLAG_LOCKED,
+ &ripcbinfo.ipi_count, 0, "Number of active PCBs");
static int
rip_attach(struct socket *so, int proto, struct proc *p)
inp = sotoinpcb(so);
if (inp == 0)
panic("rip_detach");
-#if MROUTING
- if (so == ip_mrouter)
- ip_mrouter_done();
- ip_rsvp_force_done(so);
- if (so == ip_rsvpd)
- ip_rsvp_done();
-#endif /* MROUTING */
in_pcbdetach(inp);
return 0;
}
}
__private_extern__ int
-rip_bind(struct socket *so, struct sockaddr *nam, __unused struct proc *p)
+rip_bind(struct socket *so, struct sockaddr *nam, struct proc *p)
{
+#pragma unused(p)
struct inpcb *inp = sotoinpcb(so);
- struct sockaddr_in *addr = (struct sockaddr_in *)nam;
+ struct sockaddr_in sin;
struct ifaddr *ifa = NULL;
+ struct ifnet *outif = NULL;
- if (nam->sa_len != sizeof(*addr))
- return EINVAL;
+ if (inp == NULL
+#if NECP
+ || (necp_socket_should_use_flow_divert(inp))
+#endif /* NECP */
+ )
+ return (inp == NULL ? EINVAL : EPROTOTYPE);
- if (TAILQ_EMPTY(&ifnet_head) || ((addr->sin_family != AF_INET) &&
- (addr->sin_family != AF_IMPLINK)) ||
- (addr->sin_addr.s_addr &&
- (ifa = ifa_ifwithaddr((struct sockaddr *)addr)) == 0)) {
- return EADDRNOTAVAIL;
- }
- else if (ifa) {
- ifafree(ifa);
- ifa = NULL;
+ if (nam->sa_len != sizeof (struct sockaddr_in))
+ return (EINVAL);
+
+ /* Sanitized local copy for interface address searches */
+ bzero(&sin, sizeof (sin));
+ sin.sin_family = AF_INET;
+ sin.sin_len = sizeof (struct sockaddr_in);
+ sin.sin_addr.s_addr = SIN(nam)->sin_addr.s_addr;
+
+ if (TAILQ_EMPTY(&ifnet_head) ||
+ (sin.sin_family != AF_INET && sin.sin_family != AF_IMPLINK) ||
+ (sin.sin_addr.s_addr && (ifa = ifa_ifwithaddr(SA(&sin))) == 0)) {
+ return (EADDRNOTAVAIL);
+ } else if (ifa) {
+ /*
+ * Opportunistically determine the outbound
+ * interface that may be used; this may not
+ * hold true if we end up using a route
+ * going over a different interface, e.g.
+ * when sending to a local address. This
+ * will get updated again after sending.
+ */
+ IFA_LOCK(ifa);
+ outif = ifa->ifa_ifp;
+ IFA_UNLOCK(ifa);
+ IFA_REMREF(ifa);
}
- inp->inp_laddr = addr->sin_addr;
- return 0;
+ inp->inp_laddr = sin.sin_addr;
+ inp->inp_last_outifp = outif;
+
+ return (0);
}
__private_extern__ int
rip_connect(struct socket *so, struct sockaddr *nam, __unused struct proc *p)
{
struct inpcb *inp = sotoinpcb(so);
- struct sockaddr_in *addr = (struct sockaddr_in *)nam;
-
+ struct sockaddr_in *addr = (struct sockaddr_in *)(void *)nam;
+
+ if (inp == NULL
+#if NECP
+ || (necp_socket_should_use_flow_divert(inp))
+#endif /* NECP */
+ )
+ return (inp == NULL ? EINVAL : EPROTOTYPE);
if (nam->sa_len != sizeof(*addr))
return EINVAL;
if (TAILQ_EMPTY(&ifnet_head))
if ((addr->sin_family != AF_INET) &&
(addr->sin_family != AF_IMPLINK))
return EAFNOSUPPORT;
+
+ if (!(so->so_flags1 & SOF1_CONNECT_COUNTED)) {
+ so->so_flags1 |= SOF1_CONNECT_COUNTED;
+ INC_ATOMIC_INT64_LIM(net_api_stats.nas_socket_inet_dgram_connected);
+ }
+
inp->inp_faddr = addr->sin_addr;
soisconnected(so);
+
return 0;
}
}
__private_extern__ int
-rip_send(struct socket *so, __unused int flags, struct mbuf *m, struct sockaddr *nam,
- __unused struct mbuf *control, __unused struct proc *p)
+rip_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam,
+ struct mbuf *control, struct proc *p)
{
+#pragma unused(flags, p)
struct inpcb *inp = sotoinpcb(so);
- register u_int32_t dst;
+ u_int32_t dst;
+ int error = 0;
+
+ if (inp == NULL
+#if NECP
+ || (necp_socket_should_use_flow_divert(inp) && (error = EPROTOTYPE))
+#endif /* NECP */
+ ) {
+ if (inp == NULL)
+ error = EINVAL;
+ else
+ error = EPROTOTYPE;
+ goto bad;
+ }
if (so->so_state & SS_ISCONNECTED) {
- if (nam) {
- m_freem(m);
- return EISCONN;
+ if (nam != NULL) {
+ error = EISCONN;
+ goto bad;
}
dst = inp->inp_faddr.s_addr;
} else {
if (nam == NULL) {
- m_freem(m);
- return ENOTCONN;
+ error = ENOTCONN;
+ goto bad;
}
- dst = ((struct sockaddr_in *)nam)->sin_addr.s_addr;
+ dst = ((struct sockaddr_in *)(void *)nam)->sin_addr.s_addr;
}
- return rip_output(m, so, dst);
+ return (rip_output(m, so, dst, control));
+
+bad:
+ VERIFY(error != 0);
+
+ if (m != NULL)
+ m_freem(m);
+ if (control != NULL)
+ m_freem(control);
+
+ return (error);
}
/* note: rip_unlock is called from different protos instead of the generic socket_unlock,
- * it will handle the socket dealloc on last reference
+ * it will handle the socket dealloc on last reference
* */
int
rip_unlock(struct socket *so, int refcount, void *debug)
if (so->so_usecount == 0 && (inp->inp_wantcnt == WNT_STOPUSING)) {
/* cleanup after last reference */
lck_mtx_unlock(so->so_proto->pr_domain->dom_mtx);
- lck_rw_lock_exclusive(ripcbinfo.mtx);
+ lck_rw_lock_exclusive(ripcbinfo.ipi_lock);
if (inp->inp_state != INPCB_STATE_DEAD) {
#if INET6
- if (INP_CHECK_SOCKAF(so, AF_INET6))
+ if (SOCK_CHECK_DOM(so, PF_INET6))
in6_pcbdetach(inp);
else
#endif /* INET6 */
in_pcbdetach(inp);
}
in_pcbdispose(inp);
- lck_rw_done(ripcbinfo.mtx);
+ lck_rw_done(ripcbinfo.ipi_lock);
return(0);
}
}
* The process of preparing the TCB list is too time-consuming and
* resource-intensive to repeat twice on every request.
*/
- lck_rw_lock_exclusive(ripcbinfo.mtx);
+ lck_rw_lock_exclusive(ripcbinfo.ipi_lock);
if (req->oldptr == USER_ADDR_NULL) {
n = ripcbinfo.ipi_count;
req->oldidx = 2 * (sizeof xig)
+ (n + n/8) * sizeof(struct xinpcb);
- lck_rw_done(ripcbinfo.mtx);
+ lck_rw_done(ripcbinfo.ipi_lock);
return 0;
}
if (req->newptr != USER_ADDR_NULL) {
- lck_rw_done(ripcbinfo.mtx);
+ lck_rw_done(ripcbinfo.ipi_lock);
return EPERM;
}
*/
gencnt = ripcbinfo.ipi_gencnt;
n = ripcbinfo.ipi_count;
-
+
bzero(&xig, sizeof(xig));
xig.xig_len = sizeof xig;
xig.xig_count = n;
xig.xig_sogen = so_gencnt;
error = SYSCTL_OUT(req, &xig, sizeof xig);
if (error) {
- lck_rw_done(ripcbinfo.mtx);
+ lck_rw_done(ripcbinfo.ipi_lock);
return error;
}
/*
* We are done if there is no pcb
*/
if (n == 0) {
- lck_rw_done(ripcbinfo.mtx);
- return 0;
+ lck_rw_done(ripcbinfo.ipi_lock);
+ return 0;
}
inp_list = _MALLOC(n * sizeof *inp_list, M_TEMP, M_WAITOK);
if (inp_list == 0) {
- lck_rw_done(ripcbinfo.mtx);
+ lck_rw_done(ripcbinfo.ipi_lock);
return ENOMEM;
}
-
- for (inp = ripcbinfo.listhead->lh_first, i = 0; inp && i < n;
+
+ for (inp = ripcbinfo.ipi_listhead->lh_first, i = 0; inp && i < n;
inp = inp->inp_list.le_next) {
if (inp->inp_gencnt <= gencnt && inp->inp_state != INPCB_STATE_DEAD)
inp_list[i++] = inp;
error = SYSCTL_OUT(req, &xig, sizeof xig);
}
FREE(inp_list, M_TEMP);
- lck_rw_done(ripcbinfo.mtx);
+ lck_rw_done(ripcbinfo.ipi_lock);
return error;
}
-SYSCTL_PROC(_net_inet_raw, OID_AUTO/*XXX*/, pcblist, CTLFLAG_RD, 0, 0,
+SYSCTL_PROC(_net_inet_raw, OID_AUTO/*XXX*/, pcblist,
+ CTLTYPE_STRUCT | CTLFLAG_RD | CTLFLAG_LOCKED, 0, 0,
rip_pcblist, "S,xinpcb", "List of active raw IP sockets");
#if !CONFIG_EMBEDDED
* The process of preparing the TCB list is too time-consuming and
* resource-intensive to repeat twice on every request.
*/
- lck_rw_lock_exclusive(ripcbinfo.mtx);
+ lck_rw_lock_exclusive(ripcbinfo.ipi_lock);
if (req->oldptr == USER_ADDR_NULL) {
n = ripcbinfo.ipi_count;
req->oldidx = 2 * (sizeof xig)
+ (n + n/8) * sizeof(struct xinpcb64);
- lck_rw_done(ripcbinfo.mtx);
+ lck_rw_done(ripcbinfo.ipi_lock);
return 0;
}
if (req->newptr != USER_ADDR_NULL) {
- lck_rw_done(ripcbinfo.mtx);
+ lck_rw_done(ripcbinfo.ipi_lock);
return EPERM;
}
xig.xig_sogen = so_gencnt;
error = SYSCTL_OUT(req, &xig, sizeof xig);
if (error) {
- lck_rw_done(ripcbinfo.mtx);
+ lck_rw_done(ripcbinfo.ipi_lock);
return error;
}
/*
* We are done if there is no pcb
*/
if (n == 0) {
- lck_rw_done(ripcbinfo.mtx);
+ lck_rw_done(ripcbinfo.ipi_lock);
return 0;
}
inp_list = _MALLOC(n * sizeof *inp_list, M_TEMP, M_WAITOK);
if (inp_list == 0) {
- lck_rw_done(ripcbinfo.mtx);
+ lck_rw_done(ripcbinfo.ipi_lock);
return ENOMEM;
}
- for (inp = ripcbinfo.listhead->lh_first, i = 0; inp && i < n;
+ for (inp = ripcbinfo.ipi_listhead->lh_first, i = 0; inp && i < n;
inp = inp->inp_list.le_next) {
if (inp->inp_gencnt <= gencnt && inp->inp_state != INPCB_STATE_DEAD)
inp_list[i++] = inp;
error = SYSCTL_OUT(req, &xig, sizeof xig);
}
FREE(inp_list, M_TEMP);
- lck_rw_done(ripcbinfo.mtx);
+ lck_rw_done(ripcbinfo.ipi_lock);
return error;
}
-SYSCTL_PROC(_net_inet_raw, OID_AUTO, pcblist64, CTLFLAG_RD, 0, 0,
+SYSCTL_PROC(_net_inet_raw, OID_AUTO, pcblist64,
+ CTLTYPE_STRUCT | CTLFLAG_RD | CTLFLAG_LOCKED, 0, 0,
rip_pcblist64, "S,xinpcb64", "List of active raw IP sockets");
#endif /* !CONFIG_EMBEDDED */
+
+static int
+rip_pcblist_n SYSCTL_HANDLER_ARGS
+{
+#pragma unused(oidp, arg1, arg2)
+ int error = 0;
+
+ error = get_pcblist_n(IPPROTO_IP, req, &ripcbinfo);
+
+ return error;
+}
+
+SYSCTL_PROC(_net_inet_raw, OID_AUTO, pcblist_n,
+ CTLTYPE_STRUCT | CTLFLAG_RD | CTLFLAG_LOCKED, 0, 0,
+ rip_pcblist_n, "S,xinpcb_n", "List of active raw IP sockets");
+
struct pr_usrreqs rip_usrreqs = {
- rip_abort, pru_accept_notsupp, rip_attach, rip_bind, rip_connect,
- pru_connect2_notsupp, in_control, rip_detach, rip_disconnect,
- pru_listen_notsupp, in_setpeeraddr, pru_rcvd_notsupp,
- pru_rcvoob_notsupp, rip_send, pru_sense_null, rip_shutdown,
- in_setsockaddr, sosend, soreceive, pru_sopoll_notsupp
+ .pru_abort = rip_abort,
+ .pru_attach = rip_attach,
+ .pru_bind = rip_bind,
+ .pru_connect = rip_connect,
+ .pru_control = in_control,
+ .pru_detach = rip_detach,
+ .pru_disconnect = rip_disconnect,
+ .pru_peeraddr = in_getpeeraddr,
+ .pru_send = rip_send,
+ .pru_shutdown = rip_shutdown,
+ .pru_sockaddr = in_getsockaddr,
+ .pru_sosend = sosend,
+ .pru_soreceive = soreceive,
};
/* DSEP Review Done pl-20051213-v02 @3253 */
projects / apple / xnu.git / blobdiff