freebsd-skq/usr.sbin/ifmcstat/ifmcstat.c

1249 lines
28 KiB
C
Raw Normal View History

/* $KAME: ifmcstat.c,v 1.48 2006/11/15 05:13:59 itojun Exp $ */
/*-
* SPDX-License-Identifier: BSD-3-Clause
*
* Copyright (c) 2007-2009 Bruce Simpson.
* Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. Neither the name of the project nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include <sys/types.h>
#include <sys/param.h>
#include <sys/sysctl.h>
#include <sys/socket.h>
#include <sys/queue.h>
#include <sys/tree.h>
#include <net/if.h>
#include <net/if_types.h>
#include <net/if_dl.h>
#include <net/route.h>
#include <netinet/in.h>
#include <netinet/in_var.h>
#include <netinet/in_systm.h>
#include <netinet/ip.h>
#include <netinet/igmp.h>
#include <netinet/if_ether.h>
#include <netinet/igmp_var.h>
#ifdef INET6
#include <netinet/icmp6.h>
#include <netinet6/mld6_var.h>
#endif /* INET6 */
#include <arpa/inet.h>
#include <netdb.h>
#include <stddef.h>
#include <stdarg.h>
#include <stdint.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <ctype.h>
#include <err.h>
#include <errno.h>
#include <fcntl.h>
#include <limits.h>
#include <ifaddrs.h>
#include <sysexits.h>
#include <unistd.h>
#ifdef KVM
/*
* Currently the KVM build is broken. To be fixed it requires uncovering
* large amount of _KERNEL code in include files, and it is also very
* tentative to internal kernel ABI changes. If anyone wishes to restore
* it, please move it out of src/usr.sbin to src/tools/tools.
*/
#include <kvm.h>
#include <nlist.h>
#endif
/* XXX: This file currently assumes INET support in the base system. */
#ifndef INET
#define INET
#endif
extern void printb(const char *, unsigned int, const char *);
union sockunion {
struct sockaddr_storage ss;
struct sockaddr sa;
struct sockaddr_dl sdl;
#ifdef INET
struct sockaddr_in sin;
#endif
#ifdef INET6
struct sockaddr_in6 sin6;
#endif
};
typedef union sockunion sockunion_t;
uint32_t ifindex = 0;
int af = AF_UNSPEC;
#ifdef WITH_KVM
int Kflag = 0;
#endif
int vflag = 0;
#define sa_dl_equal(a1, a2) \
((((struct sockaddr_dl *)(a1))->sdl_len == \
((struct sockaddr_dl *)(a2))->sdl_len) && \
(bcmp(LLADDR((struct sockaddr_dl *)(a1)), \
LLADDR((struct sockaddr_dl *)(a2)), \
((struct sockaddr_dl *)(a1))->sdl_alen) == 0))
/*
* Most of the code in this utility is to support the use of KVM for
* post-mortem debugging of the multicast code.
*/
#ifdef WITH_KVM
#ifdef INET
static void if_addrlist(struct ifaddr *);
static struct in_multi *
in_multientry(struct in_multi *);
#endif /* INET */
#ifdef INET6
static void if6_addrlist(struct ifaddr *);
static struct in6_multi *
in6_multientry(struct in6_multi *);
#endif /* INET6 */
static void kread(u_long, void *, int);
static void ll_addrlist(struct ifaddr *);
static int ifmcstat_kvm(const char *kernel, const char *core);
#define KREAD(addr, buf, type) \
kread((u_long)addr, (void *)buf, sizeof(type))
kvm_t *kvmd;
struct nlist nl[] = {
{ "_ifnet", 0, 0, 0, 0, },
{ "", 0, 0, 0, 0, },
};
#define N_IFNET 0
#endif /* WITH_KVM */
static int ifmcstat_getifmaddrs(void);
#ifdef INET
static void in_ifinfo(struct igmp_ifinfo *);
static const char * inm_mode(u_int mode);
#endif
#ifdef INET6
Bite the bullet, and make the IPv6 SSM and MLDv2 mega-commit: import from p4 bms_netdev. Summary of changes: * Connect netinet6/in6_mcast.c to build. The legacy KAME KPIs are mostly preserved. * Eliminate now dead code from ip6_output.c. Don't do mbuf bingo, we are not going to do RFC 2292 style CMSG tricks for multicast options as they are not required by any current IPv6 normative reference. * Refactor transports (UDP, raw_ip6) to do own mcast filtering. SCTP, TCP unaffected by this change. * Add ip6_msource, in6_msource structs to in6_var.h. * Hookup mld_ifinfo state to in6_ifextra, allocate from domifattach path. * Eliminate IN6_LOOKUP_MULTI(), it is no longer referenced. Kernel consumers which need this should use in6m_lookup(). * Refactor IPv6 socket group memberships to use a vector (like IPv4). * Update ifmcstat(8) for IPv6 SSM. * Add witness lock order for IN6_MULTI_LOCK. * Move IN6_MULTI_LOCK out of lower ip6_output()/ip6_input() paths. * Introduce IP6STAT_ADD/SUB/INC/DEC as per rwatson's IPv4 cleanup. * Update carp(4) for new IPv6 SSM KPIs. * Virtualize ip6_mrouter socket. Changes mostly localized to IPv6 MROUTING. * Don't do a local group lookup in MROUTING. * Kill unused KAME prototypes in6_purgemkludge(), in6_restoremkludge(). * Preserve KAME DAD timer jitter behaviour in MLDv1 compatibility mode. * Bump __FreeBSD_version to 800084. * Update UPDATING. NOTE WELL: * This code hasn't been tested against real MLDv2 queriers (yet), although the on-wire protocol has been verified in Wireshark. * There are a few unresolved issues in the socket layer APIs to do with scope ID propagation. * There is a LOR present in ip6_output()'s use of in6_setscope() which needs to be resolved. See comments in mld6.c. This is believed to be benign and can't be avoided for the moment without re-introducing an indirect netisr. This work was mostly derived from the IGMPv3 implementation, and has been sponsored by a third party.
2009-04-29 19:19:13 +00:00
static void in6_ifinfo(struct mld_ifinfo *);
static const char * inet6_n2a(struct in6_addr *, uint32_t);
#endif
int main(int, char **);
static void
usage()
{
fprintf(stderr,
"usage: ifmcstat [-i interface] [-f address family]"
" [-v]"
#ifdef WITH_KVM
" [-K] [-M core] [-N system]"
#endif
"\n");
exit(EX_USAGE);
}
static const char *options = "i:f:vM:N:"
#ifdef WITH_KVM
"K"
#endif
;
int
main(int argc, char **argv)
{
int c, error;
#ifdef WITH_KVM
const char *kernel = NULL;
const char *core = NULL;
#endif
while ((c = getopt(argc, argv, options)) != -1) {
switch (c) {
case 'i':
if ((ifindex = if_nametoindex(optarg)) == 0) {
fprintf(stderr, "%s: unknown interface\n",
optarg);
2009-02-15 15:26:42 +00:00
exit(EX_NOHOST);
}
break;
case 'f':
#ifdef INET
if (strcmp(optarg, "inet") == 0) {
af = AF_INET;
break;
}
#endif
#ifdef INET6
if (strcmp(optarg, "inet6") == 0) {
af = AF_INET6;
break;
}
#endif
if (strcmp(optarg, "link") == 0) {
af = AF_LINK;
break;
}
fprintf(stderr, "%s: unknown address family\n", optarg);
2009-02-15 15:26:42 +00:00
exit(EX_USAGE);
/*NOTREACHED*/
break;
#ifdef WITH_KVM
case 'K':
++Kflag;
break;
#endif
case 'v':
++vflag;
break;
#ifdef WITH_KVM
case 'M':
core = strdup(optarg);
break;
case 'N':
kernel = strdup(optarg);
break;
#endif
default:
usage();
break;
/*NOTREACHED*/
}
}
if (af == AF_LINK && vflag)
usage();
#ifdef WITH_KVM
Merge final round of MLD changes from p4: ip6_input.c, in6.h: * Add netinet6-specific mbuf flag M_RTALERT_MLD, shadowing M_PROTO6. * Always set this flag if HBH Router Alert option is present for MLD, even when not forwarding. icmp6.c: * In icmp6_input(), spell m->m_pkthdr.rcvif as ifp to be consistent. * Use scope ID for verifying input. Do not apply SSM filters here, no inpcb. * Check for M_RTALERT_MLD when validating MLD traffic, as we can't see IPv6 hop options outside of ip6_input(). in6_mcast.c: * Use KAME scope/zone ID in in6_multi. * Update net.inet6.ip6.mcast.filters implementation to use scope IDs for comparisons. * Fix scope ID treatment in multicast socket option processing. Scope IDs passed in from userland will be ignored as other less ambiguous APIs exist for specifying the link. * Tighten userland input checks in IPv6 SSM delta and full-state ops. * Source filter embedded scope IDs need to be revisited, for now just clear them and ignore them on input. * Adapt KAME behaviour of looking up the scope ID in the default zone for multicast leaves, when the interface is ambiguous. mld6.c: * Tighten origin checks on MLD traffic as per RFC3810 Section 6.2: * ip6_src MAY be the unspecified address for MLDv1 reports. * ip6_src MAY have link-local address scope for MLDv1 reports, MLDv1 queries, and MLDv2 queries. * Perform address field validation *before* accepting queries. * Use KAME scope/zone ID in query/report processing. * Break const correctness for mld_v1_input_report(), mld_v1_input_query() as we temporarily modify the input mbuf chain. * Clear the scope ID before handoff to userland MLD daemon. * Fix MLDv1 old querier present timer processing. With the protocol defaults, hosts should revert to MLDv2 after 260s. * Add net.inet6.mld.v1enable sysctl, default to on. ifmcstat.c: * Use sysctl by default; -K requests kvm(3) if so compiled. mld.4: * Connect man page to build. Tested using PCS.
2009-05-27 18:57:13 +00:00
if (Kflag)
error = ifmcstat_kvm(kernel, core);
/*
* If KVM failed, and user did not explicitly specify a core file,
* or force KVM backend to be disabled, try the sysctl backend.
*/
Merge final round of MLD changes from p4: ip6_input.c, in6.h: * Add netinet6-specific mbuf flag M_RTALERT_MLD, shadowing M_PROTO6. * Always set this flag if HBH Router Alert option is present for MLD, even when not forwarding. icmp6.c: * In icmp6_input(), spell m->m_pkthdr.rcvif as ifp to be consistent. * Use scope ID for verifying input. Do not apply SSM filters here, no inpcb. * Check for M_RTALERT_MLD when validating MLD traffic, as we can't see IPv6 hop options outside of ip6_input(). in6_mcast.c: * Use KAME scope/zone ID in in6_multi. * Update net.inet6.ip6.mcast.filters implementation to use scope IDs for comparisons. * Fix scope ID treatment in multicast socket option processing. Scope IDs passed in from userland will be ignored as other less ambiguous APIs exist for specifying the link. * Tighten userland input checks in IPv6 SSM delta and full-state ops. * Source filter embedded scope IDs need to be revisited, for now just clear them and ignore them on input. * Adapt KAME behaviour of looking up the scope ID in the default zone for multicast leaves, when the interface is ambiguous. mld6.c: * Tighten origin checks on MLD traffic as per RFC3810 Section 6.2: * ip6_src MAY be the unspecified address for MLDv1 reports. * ip6_src MAY have link-local address scope for MLDv1 reports, MLDv1 queries, and MLDv2 queries. * Perform address field validation *before* accepting queries. * Use KAME scope/zone ID in query/report processing. * Break const correctness for mld_v1_input_report(), mld_v1_input_query() as we temporarily modify the input mbuf chain. * Clear the scope ID before handoff to userland MLD daemon. * Fix MLDv1 old querier present timer processing. With the protocol defaults, hosts should revert to MLDv2 after 260s. * Add net.inet6.mld.v1enable sysctl, default to on. ifmcstat.c: * Use sysctl by default; -K requests kvm(3) if so compiled. mld.4: * Connect man page to build. Tested using PCS.
2009-05-27 18:57:13 +00:00
if (!Kflag || (error != 0 && (core == NULL && kernel == NULL)))
#endif
error = ifmcstat_getifmaddrs();
if (error != 0)
2009-02-15 15:26:42 +00:00
exit(EX_OSERR);
2009-02-15 15:26:42 +00:00
exit(EX_OK);
/*NOTREACHED*/
}
#ifdef INET
static void
in_ifinfo(struct igmp_ifinfo *igi)
{
printf("\t");
switch (igi->igi_version) {
case IGMP_VERSION_1:
case IGMP_VERSION_2:
case IGMP_VERSION_3:
printf("igmpv%d", igi->igi_version);
break;
default:
printf("igmpv?(%d)", igi->igi_version);
break;
}
if (igi->igi_flags)
printb(" flags", igi->igi_flags, "\020\1SILENT\2LOOPBACK");
if (igi->igi_version == IGMP_VERSION_3) {
printf(" rv %u qi %u qri %u uri %u",
igi->igi_rv, igi->igi_qi, igi->igi_qri, igi->igi_uri);
}
if (vflag >= 2) {
printf(" v1timer %u v2timer %u v3timer %u",
igi->igi_v1_timer, igi->igi_v2_timer, igi->igi_v3_timer);
}
printf("\n");
}
static const char *inm_modes[] = {
"undefined",
"include",
"exclude",
};
static const char *
inm_mode(u_int mode)
{
if (mode >= MCAST_UNDEFINED && mode <= MCAST_EXCLUDE)
return (inm_modes[mode]);
return (NULL);
}
#endif /* INET */
#ifdef WITH_KVM
static int
ifmcstat_kvm(const char *kernel, const char *core)
{
char buf[_POSIX2_LINE_MAX], ifname[IFNAMSIZ];
struct ifnet *ifp, *nifp, ifnet;
if ((kvmd = kvm_openfiles(kernel, core, NULL, O_RDONLY, buf)) ==
NULL) {
perror("kvm_openfiles");
return (-1);
}
if (kvm_nlist(kvmd, nl) < 0) {
perror("kvm_nlist");
return (-1);
}
if (nl[N_IFNET].n_value == 0) {
printf("symbol %s not found\n", nl[N_IFNET].n_name);
return (-1);
}
KREAD(nl[N_IFNET].n_value, &ifp, struct ifnet *);
while (ifp) {
KREAD(ifp, &ifnet, struct ifnet);
nifp = ifnet.if_link.tqe_next;
if (ifindex && ifindex != ifnet.if_index)
goto next;
printf("%s:\n", if_indextoname(ifnet.if_index, ifname));
#ifdef INET
if_addrlist(TAILQ_FIRST(&ifnet.if_addrhead));
#endif
#ifdef INET6
if6_addrlist(TAILQ_FIRST(&ifnet.if_addrhead));
#endif
if (vflag)
ll_addrlist(TAILQ_FIRST(&ifnet.if_addrhead));
2009-02-15 15:26:42 +00:00
next:
ifp = nifp;
}
return (0);
}
static void
kread(u_long addr, void *buf, int len)
{
if (kvm_read(kvmd, addr, buf, len) != len) {
perror("kvm_read");
2009-02-15 15:26:42 +00:00
exit(EX_OSERR);
}
}
static void
ll_addrlist(struct ifaddr *ifap)
{
char addrbuf[NI_MAXHOST];
struct ifaddr ifa;
struct sockaddr sa;
struct sockaddr_dl sdl;
struct ifaddr *ifap0;
if (af && af != AF_LINK)
return;
ifap0 = ifap;
while (ifap) {
KREAD(ifap, &ifa, struct ifaddr);
if (ifa.ifa_addr == NULL)
goto nextifap;
KREAD(ifa.ifa_addr, &sa, struct sockaddr);
if (sa.sa_family != PF_LINK)
goto nextifap;
KREAD(ifa.ifa_addr, &sdl, struct sockaddr_dl);
if (sdl.sdl_alen == 0)
goto nextifap;
addrbuf[0] = '\0';
getnameinfo((struct sockaddr *)&sdl, sdl.sdl_len,
addrbuf, sizeof(addrbuf), NULL, 0, NI_NUMERICHOST);
printf("\tlink %s\n", addrbuf);
nextifap:
ifap = ifa.ifa_link.tqe_next;
}
if (ifap0) {
struct ifnet ifnet;
struct ifmultiaddr ifm, *ifmp = 0;
KREAD(ifap0, &ifa, struct ifaddr);
KREAD(ifa.ifa_ifp, &ifnet, struct ifnet);
if (TAILQ_FIRST(&ifnet.if_multiaddrs))
ifmp = TAILQ_FIRST(&ifnet.if_multiaddrs);
while (ifmp) {
KREAD(ifmp, &ifm, struct ifmultiaddr);
if (ifm.ifma_addr == NULL)
goto nextmulti;
KREAD(ifm.ifma_addr, &sa, struct sockaddr);
if (sa.sa_family != AF_LINK)
goto nextmulti;
KREAD(ifm.ifma_addr, &sdl, struct sockaddr_dl);
addrbuf[0] = '\0';
getnameinfo((struct sockaddr *)&sdl,
sdl.sdl_len, addrbuf, sizeof(addrbuf),
NULL, 0, NI_NUMERICHOST);
printf("\t\tgroup %s refcnt %d\n",
addrbuf, ifm.ifma_refcount);
nextmulti:
ifmp = TAILQ_NEXT(&ifm, ifma_link);
}
}
}
#ifdef INET6
static void
if6_addrlist(struct ifaddr *ifap)
{
Bite the bullet, and make the IPv6 SSM and MLDv2 mega-commit: import from p4 bms_netdev. Summary of changes: * Connect netinet6/in6_mcast.c to build. The legacy KAME KPIs are mostly preserved. * Eliminate now dead code from ip6_output.c. Don't do mbuf bingo, we are not going to do RFC 2292 style CMSG tricks for multicast options as they are not required by any current IPv6 normative reference. * Refactor transports (UDP, raw_ip6) to do own mcast filtering. SCTP, TCP unaffected by this change. * Add ip6_msource, in6_msource structs to in6_var.h. * Hookup mld_ifinfo state to in6_ifextra, allocate from domifattach path. * Eliminate IN6_LOOKUP_MULTI(), it is no longer referenced. Kernel consumers which need this should use in6m_lookup(). * Refactor IPv6 socket group memberships to use a vector (like IPv4). * Update ifmcstat(8) for IPv6 SSM. * Add witness lock order for IN6_MULTI_LOCK. * Move IN6_MULTI_LOCK out of lower ip6_output()/ip6_input() paths. * Introduce IP6STAT_ADD/SUB/INC/DEC as per rwatson's IPv4 cleanup. * Update carp(4) for new IPv6 SSM KPIs. * Virtualize ip6_mrouter socket. Changes mostly localized to IPv6 MROUTING. * Don't do a local group lookup in MROUTING. * Kill unused KAME prototypes in6_purgemkludge(), in6_restoremkludge(). * Preserve KAME DAD timer jitter behaviour in MLDv1 compatibility mode. * Bump __FreeBSD_version to 800084. * Update UPDATING. NOTE WELL: * This code hasn't been tested against real MLDv2 queriers (yet), although the on-wire protocol has been verified in Wireshark. * There are a few unresolved issues in the socket layer APIs to do with scope ID propagation. * There is a LOR present in ip6_output()'s use of in6_setscope() which needs to be resolved. See comments in mld6.c. This is believed to be benign and can't be avoided for the moment without re-introducing an indirect netisr. This work was mostly derived from the IGMPv3 implementation, and has been sponsored by a third party.
2009-04-29 19:19:13 +00:00
struct ifnet ifnet;
struct ifaddr ifa;
struct sockaddr sa;
struct in6_ifaddr if6a;
struct ifaddr *ifap0;
if (af && af != AF_INET6)
return;
ifap0 = ifap;
while (ifap) {
KREAD(ifap, &ifa, struct ifaddr);
if (ifa.ifa_addr == NULL)
goto nextifap;
KREAD(ifa.ifa_addr, &sa, struct sockaddr);
if (sa.sa_family != PF_INET6)
goto nextifap;
KREAD(ifap, &if6a, struct in6_ifaddr);
printf("\tinet6 %s\n", inet6_n2a(&if6a.ia_addr.sin6_addr,
if6a.ia_addr.sin6_scope_id));
Bite the bullet, and make the IPv6 SSM and MLDv2 mega-commit: import from p4 bms_netdev. Summary of changes: * Connect netinet6/in6_mcast.c to build. The legacy KAME KPIs are mostly preserved. * Eliminate now dead code from ip6_output.c. Don't do mbuf bingo, we are not going to do RFC 2292 style CMSG tricks for multicast options as they are not required by any current IPv6 normative reference. * Refactor transports (UDP, raw_ip6) to do own mcast filtering. SCTP, TCP unaffected by this change. * Add ip6_msource, in6_msource structs to in6_var.h. * Hookup mld_ifinfo state to in6_ifextra, allocate from domifattach path. * Eliminate IN6_LOOKUP_MULTI(), it is no longer referenced. Kernel consumers which need this should use in6m_lookup(). * Refactor IPv6 socket group memberships to use a vector (like IPv4). * Update ifmcstat(8) for IPv6 SSM. * Add witness lock order for IN6_MULTI_LOCK. * Move IN6_MULTI_LOCK out of lower ip6_output()/ip6_input() paths. * Introduce IP6STAT_ADD/SUB/INC/DEC as per rwatson's IPv4 cleanup. * Update carp(4) for new IPv6 SSM KPIs. * Virtualize ip6_mrouter socket. Changes mostly localized to IPv6 MROUTING. * Don't do a local group lookup in MROUTING. * Kill unused KAME prototypes in6_purgemkludge(), in6_restoremkludge(). * Preserve KAME DAD timer jitter behaviour in MLDv1 compatibility mode. * Bump __FreeBSD_version to 800084. * Update UPDATING. NOTE WELL: * This code hasn't been tested against real MLDv2 queriers (yet), although the on-wire protocol has been verified in Wireshark. * There are a few unresolved issues in the socket layer APIs to do with scope ID propagation. * There is a LOR present in ip6_output()'s use of in6_setscope() which needs to be resolved. See comments in mld6.c. This is believed to be benign and can't be avoided for the moment without re-introducing an indirect netisr. This work was mostly derived from the IGMPv3 implementation, and has been sponsored by a third party.
2009-04-29 19:19:13 +00:00
/*
* Print per-link MLD information, if available.
*/
if (ifa.ifa_ifp != NULL) {
struct in6_ifextra ie;
struct mld_ifinfo mli;
KREAD(ifa.ifa_ifp, &ifnet, struct ifnet);
KREAD(ifnet.if_afdata[AF_INET6], &ie,
struct in6_ifextra);
if (ie.mld_ifinfo != NULL) {
KREAD(ie.mld_ifinfo, &mli, struct mld_ifinfo);
in6_ifinfo(&mli);
}
}
nextifap:
ifap = ifa.ifa_link.tqe_next;
}
if (ifap0) {
struct ifnet ifnet;
struct ifmultiaddr ifm, *ifmp = 0;
struct sockaddr_dl sdl;
KREAD(ifap0, &ifa, struct ifaddr);
KREAD(ifa.ifa_ifp, &ifnet, struct ifnet);
if (TAILQ_FIRST(&ifnet.if_multiaddrs))
ifmp = TAILQ_FIRST(&ifnet.if_multiaddrs);
while (ifmp) {
KREAD(ifmp, &ifm, struct ifmultiaddr);
if (ifm.ifma_addr == NULL)
goto nextmulti;
KREAD(ifm.ifma_addr, &sa, struct sockaddr);
if (sa.sa_family != AF_INET6)
goto nextmulti;
(void)in6_multientry((struct in6_multi *)
ifm.ifma_protospec);
if (ifm.ifma_lladdr == 0)
goto nextmulti;
KREAD(ifm.ifma_lladdr, &sdl, struct sockaddr_dl);
printf("\t\t\tmcast-macaddr %s refcnt %d\n",
ether_ntoa((struct ether_addr *)LLADDR(&sdl)),
ifm.ifma_refcount);
nextmulti:
ifmp = TAILQ_NEXT(&ifm, ifma_link);
}
}
}
static struct in6_multi *
in6_multientry(struct in6_multi *mc)
{
struct in6_multi multi;
KREAD(mc, &multi, struct in6_multi);
printf("\t\tgroup %s", inet6_n2a(&multi.in6m_addr, 0));
printf(" refcnt %u\n", multi.in6m_refcount);
2009-02-15 15:26:42 +00:00
return (multi.in6m_entry.le_next);
}
#endif /* INET6 */
#ifdef INET
static void
if_addrlist(struct ifaddr *ifap)
{
struct ifaddr ifa;
struct ifnet ifnet;
struct sockaddr sa;
struct in_ifaddr ia;
struct ifaddr *ifap0;
if (af && af != AF_INET)
return;
ifap0 = ifap;
while (ifap) {
KREAD(ifap, &ifa, struct ifaddr);
if (ifa.ifa_addr == NULL)
goto nextifap;
KREAD(ifa.ifa_addr, &sa, struct sockaddr);
if (sa.sa_family != PF_INET)
goto nextifap;
KREAD(ifap, &ia, struct in_ifaddr);
printf("\tinet %s\n", inet_ntoa(ia.ia_addr.sin_addr));
/*
* Print per-link IGMP information, if available.
*/
if (ifa.ifa_ifp != NULL) {
struct in_ifinfo ii;
struct igmp_ifinfo igi;
KREAD(ifa.ifa_ifp, &ifnet, struct ifnet);
KREAD(ifnet.if_afdata[AF_INET], &ii, struct in_ifinfo);
if (ii.ii_igmp != NULL) {
KREAD(ii.ii_igmp, &igi, struct igmp_ifinfo);
in_ifinfo(&igi);
}
}
nextifap:
ifap = ifa.ifa_link.tqe_next;
}
if (ifap0) {
struct ifmultiaddr ifm, *ifmp = 0;
struct sockaddr_dl sdl;
KREAD(ifap0, &ifa, struct ifaddr);
KREAD(ifa.ifa_ifp, &ifnet, struct ifnet);
if (TAILQ_FIRST(&ifnet.if_multiaddrs))
ifmp = TAILQ_FIRST(&ifnet.if_multiaddrs);
while (ifmp) {
KREAD(ifmp, &ifm, struct ifmultiaddr);
if (ifm.ifma_addr == NULL)
goto nextmulti;
KREAD(ifm.ifma_addr, &sa, struct sockaddr);
if (sa.sa_family != AF_INET)
goto nextmulti;
(void)in_multientry((struct in_multi *)
ifm.ifma_protospec);
if (ifm.ifma_lladdr == 0)
goto nextmulti;
KREAD(ifm.ifma_lladdr, &sdl, struct sockaddr_dl);
printf("\t\t\tmcast-macaddr %s refcnt %d\n",
ether_ntoa((struct ether_addr *)LLADDR(&sdl)),
ifm.ifma_refcount);
nextmulti:
ifmp = TAILQ_NEXT(&ifm, ifma_link);
}
}
}
static const char *inm_states[] = {
"not-member",
"silent",
"idle",
"lazy",
"sleeping",
"awakening",
"query-pending",
"sg-query-pending",
"leaving"
};
static const char *
inm_state(u_int state)
{
if (state >= IGMP_NOT_MEMBER && state <= IGMP_LEAVING_MEMBER)
return (inm_states[state]);
return (NULL);
}
#if 0
static struct ip_msource *
ims_min_kvm(struct in_multi *pinm)
{
struct ip_msource ims0;
struct ip_msource *tmp, *parent;
parent = NULL;
tmp = RB_ROOT(&pinm->inm_srcs);
while (tmp) {
parent = tmp;
KREAD(tmp, &ims0, struct ip_msource);
tmp = RB_LEFT(&ims0, ims_link);
}
return (parent); /* kva */
}
/* XXX This routine is buggy. See RB_NEXT in sys/tree.h. */
static struct ip_msource *
ims_next_kvm(struct ip_msource *ims)
{
struct ip_msource ims0, ims1;
struct ip_msource *tmp;
KREAD(ims, &ims0, struct ip_msource);
if (RB_RIGHT(&ims0, ims_link)) {
ims = RB_RIGHT(&ims0, ims_link);
KREAD(ims, &ims1, struct ip_msource);
while ((tmp = RB_LEFT(&ims1, ims_link))) {
KREAD(tmp, &ims0, struct ip_msource);
ims = RB_LEFT(&ims0, ims_link);
}
} else {
tmp = RB_PARENT(&ims0, ims_link);
if (tmp) {
KREAD(tmp, &ims1, struct ip_msource);
if (ims == RB_LEFT(&ims1, ims_link))
ims = tmp;
} else {
while ((tmp = RB_PARENT(&ims0, ims_link))) {
KREAD(tmp, &ims1, struct ip_msource);
if (ims == RB_RIGHT(&ims1, ims_link)) {
ims = tmp;
KREAD(ims, &ims0, struct ip_msource);
} else
break;
}
ims = RB_PARENT(&ims0, ims_link);
}
}
return (ims); /* kva */
}
static void
inm_print_sources_kvm(struct in_multi *pinm)
{
struct ip_msource ims0;
struct ip_msource *ims;
struct in_addr src;
int cnt;
uint8_t fmode;
cnt = 0;
fmode = pinm->inm_st[1].iss_fmode;
if (fmode == MCAST_UNDEFINED)
return;
for (ims = ims_min_kvm(pinm); ims != NULL; ims = ims_next_kvm(ims)) {
if (cnt == 0)
printf(" srcs ");
KREAD(ims, &ims0, struct ip_msource);
/* Only print sources in-mode at t1. */
if (fmode != ims_get_mode(pinm, ims, 1))
continue;
src.s_addr = htonl(ims0.ims_haddr);
printf("%s%s", (cnt++ == 0 ? "" : ","), inet_ntoa(src));
}
}
#endif
static struct in_multi *
in_multientry(struct in_multi *pinm)
{
struct in_multi inm;
const char *state, *mode;
KREAD(pinm, &inm, struct in_multi);
printf("\t\tgroup %s", inet_ntoa(inm.inm_addr));
printf(" refcnt %u", inm.inm_refcount);
state = inm_state(inm.inm_state);
if (state)
printf(" state %s", state);
else
printf(" state (%d)", inm.inm_state);
mode = inm_mode(inm.inm_st[1].iss_fmode);
if (mode)
printf(" mode %s", mode);
else
printf(" mode (%d)", inm.inm_st[1].iss_fmode);
if (vflag >= 2) {
printf(" asm %u ex %u in %u rec %u",
(u_int)inm.inm_st[1].iss_asm,
(u_int)inm.inm_st[1].iss_ex,
(u_int)inm.inm_st[1].iss_in,
(u_int)inm.inm_st[1].iss_rec);
}
#if 0
/* Buggy. */
if (vflag)
inm_print_sources_kvm(&inm);
#endif
printf("\n");
return (NULL);
}
#endif /* INET */
#endif /* WITH_KVM */
#ifdef INET6
static void
in6_ifinfo(struct mld_ifinfo *mli)
{
printf("\t");
switch (mli->mli_version) {
case MLD_VERSION_1:
case MLD_VERSION_2:
printf("mldv%d", mli->mli_version);
break;
default:
printf("mldv?(%d)", mli->mli_version);
break;
}
if (mli->mli_flags)
printb(" flags", mli->mli_flags, "\020\1SILENT\2USEALLOW");
if (mli->mli_version == MLD_VERSION_2) {
printf(" rv %u qi %u qri %u uri %u",
mli->mli_rv, mli->mli_qi, mli->mli_qri, mli->mli_uri);
}
if (vflag >= 2) {
printf(" v1timer %u v2timer %u", mli->mli_v1_timer,
mli->mli_v2_timer);
}
printf("\n");
}
static const char *
inet6_n2a(struct in6_addr *p, uint32_t scope_id)
{
static char buf[NI_MAXHOST];
struct sockaddr_in6 sin6;
const int niflags = NI_NUMERICHOST;
memset(&sin6, 0, sizeof(sin6));
sin6.sin6_family = AF_INET6;
sin6.sin6_len = sizeof(struct sockaddr_in6);
sin6.sin6_addr = *p;
sin6.sin6_scope_id = scope_id;
if (getnameinfo((struct sockaddr *)&sin6, sin6.sin6_len,
2009-02-15 15:26:42 +00:00
buf, sizeof(buf), NULL, 0, niflags) == 0) {
return (buf);
} else {
return ("(invalid)");
}
}
#endif /* INET6 */
#ifdef INET
/*
* Retrieve per-group source filter mode and lists via sysctl.
*/
static void
inm_print_sources_sysctl(uint32_t ifindex, struct in_addr gina)
{
#define MAX_SYSCTL_TRY 5
int mib[7];
int ntry = 0;
size_t mibsize;
size_t len;
size_t needed;
size_t cnt;
int i;
char *buf;
struct in_addr *pina;
uint32_t *p;
uint32_t fmode;
const char *modestr;
mibsize = nitems(mib);
if (sysctlnametomib("net.inet.ip.mcast.filters", mib, &mibsize) == -1) {
perror("sysctlnametomib");
return;
}
needed = 0;
mib[5] = ifindex;
mib[6] = gina.s_addr; /* 32 bits wide */
mibsize = nitems(mib);
do {
if (sysctl(mib, mibsize, NULL, &needed, NULL, 0) == -1) {
perror("sysctl net.inet.ip.mcast.filters");
return;
}
if ((buf = malloc(needed)) == NULL) {
perror("malloc");
return;
}
if (sysctl(mib, mibsize, buf, &needed, NULL, 0) == -1) {
if (errno != ENOMEM || ++ntry >= MAX_SYSCTL_TRY) {
perror("sysctl");
goto out_free;
}
free(buf);
buf = NULL;
}
} while (buf == NULL);
len = needed;
if (len < sizeof(uint32_t)) {
perror("sysctl");
goto out_free;
}
p = (uint32_t *)buf;
fmode = *p++;
len -= sizeof(uint32_t);
modestr = inm_mode(fmode);
if (modestr)
printf(" mode %s", modestr);
else
printf(" mode (%u)", fmode);
if (vflag == 0)
goto out_free;
cnt = len / sizeof(struct in_addr);
pina = (struct in_addr *)p;
for (i = 0; i < cnt; i++) {
if (i == 0)
printf(" srcs ");
fprintf(stdout, "%s%s", (i == 0 ? "" : ","),
inet_ntoa(*pina++));
len -= sizeof(struct in_addr);
}
if (len > 0) {
fprintf(stderr, "warning: %u trailing bytes from %s\n",
(unsigned int)len, "net.inet.ip.mcast.filters");
}
out_free:
free(buf);
#undef MAX_SYSCTL_TRY
}
#endif /* INET */
Bite the bullet, and make the IPv6 SSM and MLDv2 mega-commit: import from p4 bms_netdev. Summary of changes: * Connect netinet6/in6_mcast.c to build. The legacy KAME KPIs are mostly preserved. * Eliminate now dead code from ip6_output.c. Don't do mbuf bingo, we are not going to do RFC 2292 style CMSG tricks for multicast options as they are not required by any current IPv6 normative reference. * Refactor transports (UDP, raw_ip6) to do own mcast filtering. SCTP, TCP unaffected by this change. * Add ip6_msource, in6_msource structs to in6_var.h. * Hookup mld_ifinfo state to in6_ifextra, allocate from domifattach path. * Eliminate IN6_LOOKUP_MULTI(), it is no longer referenced. Kernel consumers which need this should use in6m_lookup(). * Refactor IPv6 socket group memberships to use a vector (like IPv4). * Update ifmcstat(8) for IPv6 SSM. * Add witness lock order for IN6_MULTI_LOCK. * Move IN6_MULTI_LOCK out of lower ip6_output()/ip6_input() paths. * Introduce IP6STAT_ADD/SUB/INC/DEC as per rwatson's IPv4 cleanup. * Update carp(4) for new IPv6 SSM KPIs. * Virtualize ip6_mrouter socket. Changes mostly localized to IPv6 MROUTING. * Don't do a local group lookup in MROUTING. * Kill unused KAME prototypes in6_purgemkludge(), in6_restoremkludge(). * Preserve KAME DAD timer jitter behaviour in MLDv1 compatibility mode. * Bump __FreeBSD_version to 800084. * Update UPDATING. NOTE WELL: * This code hasn't been tested against real MLDv2 queriers (yet), although the on-wire protocol has been verified in Wireshark. * There are a few unresolved issues in the socket layer APIs to do with scope ID propagation. * There is a LOR present in ip6_output()'s use of in6_setscope() which needs to be resolved. See comments in mld6.c. This is believed to be benign and can't be avoided for the moment without re-introducing an indirect netisr. This work was mostly derived from the IGMPv3 implementation, and has been sponsored by a third party.
2009-04-29 19:19:13 +00:00
#ifdef INET6
/*
* Retrieve MLD per-group source filter mode and lists via sysctl.
*
2011-12-30 10:58:14 +00:00
* Note: The 128-bit IPv6 group address needs to be segmented into
Bite the bullet, and make the IPv6 SSM and MLDv2 mega-commit: import from p4 bms_netdev. Summary of changes: * Connect netinet6/in6_mcast.c to build. The legacy KAME KPIs are mostly preserved. * Eliminate now dead code from ip6_output.c. Don't do mbuf bingo, we are not going to do RFC 2292 style CMSG tricks for multicast options as they are not required by any current IPv6 normative reference. * Refactor transports (UDP, raw_ip6) to do own mcast filtering. SCTP, TCP unaffected by this change. * Add ip6_msource, in6_msource structs to in6_var.h. * Hookup mld_ifinfo state to in6_ifextra, allocate from domifattach path. * Eliminate IN6_LOOKUP_MULTI(), it is no longer referenced. Kernel consumers which need this should use in6m_lookup(). * Refactor IPv6 socket group memberships to use a vector (like IPv4). * Update ifmcstat(8) for IPv6 SSM. * Add witness lock order for IN6_MULTI_LOCK. * Move IN6_MULTI_LOCK out of lower ip6_output()/ip6_input() paths. * Introduce IP6STAT_ADD/SUB/INC/DEC as per rwatson's IPv4 cleanup. * Update carp(4) for new IPv6 SSM KPIs. * Virtualize ip6_mrouter socket. Changes mostly localized to IPv6 MROUTING. * Don't do a local group lookup in MROUTING. * Kill unused KAME prototypes in6_purgemkludge(), in6_restoremkludge(). * Preserve KAME DAD timer jitter behaviour in MLDv1 compatibility mode. * Bump __FreeBSD_version to 800084. * Update UPDATING. NOTE WELL: * This code hasn't been tested against real MLDv2 queriers (yet), although the on-wire protocol has been verified in Wireshark. * There are a few unresolved issues in the socket layer APIs to do with scope ID propagation. * There is a LOR present in ip6_output()'s use of in6_setscope() which needs to be resolved. See comments in mld6.c. This is believed to be benign and can't be avoided for the moment without re-introducing an indirect netisr. This work was mostly derived from the IGMPv3 implementation, and has been sponsored by a third party.
2009-04-29 19:19:13 +00:00
* 32-bit pieces for marshaling to sysctl. So the MIB name ends
* up looking like this:
* a.b.c.d.e.ifindex.g[0].g[1].g[2].g[3]
* Assumes that pgroup originated from the kernel, so its components
* are already in network-byte order.
*/
static void
in6m_print_sources_sysctl(uint32_t ifindex, struct in6_addr *pgroup)
{
#define MAX_SYSCTL_TRY 5
char addrbuf[INET6_ADDRSTRLEN];
int mib[10];
int ntry = 0;
int *pi;
size_t mibsize;
size_t len;
size_t needed;
size_t cnt;
int i;
char *buf;
struct in6_addr *pina;
uint32_t *p;
uint32_t fmode;
const char *modestr;
mibsize = nitems(mib);
Bite the bullet, and make the IPv6 SSM and MLDv2 mega-commit: import from p4 bms_netdev. Summary of changes: * Connect netinet6/in6_mcast.c to build. The legacy KAME KPIs are mostly preserved. * Eliminate now dead code from ip6_output.c. Don't do mbuf bingo, we are not going to do RFC 2292 style CMSG tricks for multicast options as they are not required by any current IPv6 normative reference. * Refactor transports (UDP, raw_ip6) to do own mcast filtering. SCTP, TCP unaffected by this change. * Add ip6_msource, in6_msource structs to in6_var.h. * Hookup mld_ifinfo state to in6_ifextra, allocate from domifattach path. * Eliminate IN6_LOOKUP_MULTI(), it is no longer referenced. Kernel consumers which need this should use in6m_lookup(). * Refactor IPv6 socket group memberships to use a vector (like IPv4). * Update ifmcstat(8) for IPv6 SSM. * Add witness lock order for IN6_MULTI_LOCK. * Move IN6_MULTI_LOCK out of lower ip6_output()/ip6_input() paths. * Introduce IP6STAT_ADD/SUB/INC/DEC as per rwatson's IPv4 cleanup. * Update carp(4) for new IPv6 SSM KPIs. * Virtualize ip6_mrouter socket. Changes mostly localized to IPv6 MROUTING. * Don't do a local group lookup in MROUTING. * Kill unused KAME prototypes in6_purgemkludge(), in6_restoremkludge(). * Preserve KAME DAD timer jitter behaviour in MLDv1 compatibility mode. * Bump __FreeBSD_version to 800084. * Update UPDATING. NOTE WELL: * This code hasn't been tested against real MLDv2 queriers (yet), although the on-wire protocol has been verified in Wireshark. * There are a few unresolved issues in the socket layer APIs to do with scope ID propagation. * There is a LOR present in ip6_output()'s use of in6_setscope() which needs to be resolved. See comments in mld6.c. This is believed to be benign and can't be avoided for the moment without re-introducing an indirect netisr. This work was mostly derived from the IGMPv3 implementation, and has been sponsored by a third party.
2009-04-29 19:19:13 +00:00
if (sysctlnametomib("net.inet6.ip6.mcast.filters", mib,
&mibsize) == -1) {
perror("sysctlnametomib");
return;
}
needed = 0;
mib[5] = ifindex;
pi = (int *)pgroup;
for (i = 0; i < 4; i++)
mib[6 + i] = *pi++;
mibsize = nitems(mib);
Bite the bullet, and make the IPv6 SSM and MLDv2 mega-commit: import from p4 bms_netdev. Summary of changes: * Connect netinet6/in6_mcast.c to build. The legacy KAME KPIs are mostly preserved. * Eliminate now dead code from ip6_output.c. Don't do mbuf bingo, we are not going to do RFC 2292 style CMSG tricks for multicast options as they are not required by any current IPv6 normative reference. * Refactor transports (UDP, raw_ip6) to do own mcast filtering. SCTP, TCP unaffected by this change. * Add ip6_msource, in6_msource structs to in6_var.h. * Hookup mld_ifinfo state to in6_ifextra, allocate from domifattach path. * Eliminate IN6_LOOKUP_MULTI(), it is no longer referenced. Kernel consumers which need this should use in6m_lookup(). * Refactor IPv6 socket group memberships to use a vector (like IPv4). * Update ifmcstat(8) for IPv6 SSM. * Add witness lock order for IN6_MULTI_LOCK. * Move IN6_MULTI_LOCK out of lower ip6_output()/ip6_input() paths. * Introduce IP6STAT_ADD/SUB/INC/DEC as per rwatson's IPv4 cleanup. * Update carp(4) for new IPv6 SSM KPIs. * Virtualize ip6_mrouter socket. Changes mostly localized to IPv6 MROUTING. * Don't do a local group lookup in MROUTING. * Kill unused KAME prototypes in6_purgemkludge(), in6_restoremkludge(). * Preserve KAME DAD timer jitter behaviour in MLDv1 compatibility mode. * Bump __FreeBSD_version to 800084. * Update UPDATING. NOTE WELL: * This code hasn't been tested against real MLDv2 queriers (yet), although the on-wire protocol has been verified in Wireshark. * There are a few unresolved issues in the socket layer APIs to do with scope ID propagation. * There is a LOR present in ip6_output()'s use of in6_setscope() which needs to be resolved. See comments in mld6.c. This is believed to be benign and can't be avoided for the moment without re-introducing an indirect netisr. This work was mostly derived from the IGMPv3 implementation, and has been sponsored by a third party.
2009-04-29 19:19:13 +00:00
do {
if (sysctl(mib, mibsize, NULL, &needed, NULL, 0) == -1) {
perror("sysctl net.inet6.ip6.mcast.filters");
return;
}
if ((buf = malloc(needed)) == NULL) {
perror("malloc");
return;
}
if (sysctl(mib, mibsize, buf, &needed, NULL, 0) == -1) {
if (errno != ENOMEM || ++ntry >= MAX_SYSCTL_TRY) {
perror("sysctl");
goto out_free;
}
free(buf);
buf = NULL;
}
} while (buf == NULL);
len = needed;
if (len < sizeof(uint32_t)) {
perror("sysctl");
goto out_free;
}
p = (uint32_t *)buf;
fmode = *p++;
len -= sizeof(uint32_t);
modestr = inm_mode(fmode);
if (modestr)
printf(" mode %s", modestr);
else
printf(" mode (%u)", fmode);
if (vflag == 0)
goto out_free;
cnt = len / sizeof(struct in6_addr);
pina = (struct in6_addr *)p;
for (i = 0; i < cnt; i++) {
if (i == 0)
printf(" srcs ");
inet_ntop(AF_INET6, (const char *)pina++, addrbuf,
INET6_ADDRSTRLEN);
fprintf(stdout, "%s%s", (i == 0 ? "" : ","), addrbuf);
len -= sizeof(struct in6_addr);
}
if (len > 0) {
fprintf(stderr, "warning: %u trailing bytes from %s\n",
(unsigned int)len, "net.inet6.ip6.mcast.filters");
}
out_free:
free(buf);
#undef MAX_SYSCTL_TRY
}
#endif /* INET6 */
static int
ifmcstat_getifmaddrs(void)
{
char thisifname[IFNAMSIZ];
char addrbuf[NI_MAXHOST];
struct ifaddrs *ifap, *ifa;
struct ifmaddrs *ifmap, *ifma;
sockunion_t lastifasa;
sockunion_t *psa, *pgsa, *pllsa, *pifasa;
char *pcolon;
char *pafname;
uint32_t lastifindex, thisifindex;
int error;
error = 0;
ifap = NULL;
ifmap = NULL;
lastifindex = 0;
thisifindex = 0;
lastifasa.ss.ss_family = AF_UNSPEC;
if (getifaddrs(&ifap) != 0) {
warn("getifmaddrs");
return (-1);
}
if (getifmaddrs(&ifmap) != 0) {
warn("getifmaddrs");
error = -1;
goto out;
}
for (ifma = ifmap; ifma; ifma = ifma->ifma_next) {
error = 0;
if (ifma->ifma_name == NULL || ifma->ifma_addr == NULL)
continue;
psa = (sockunion_t *)ifma->ifma_name;
if (psa->sa.sa_family != AF_LINK) {
fprintf(stderr,
"WARNING: Kernel returned invalid data.\n");
error = -1;
break;
}
/* Filter on interface name. */
thisifindex = psa->sdl.sdl_index;
if (ifindex != 0 && thisifindex != ifindex)
continue;
/* Filter on address family. */
pgsa = (sockunion_t *)ifma->ifma_addr;
if (af != 0 && pgsa->sa.sa_family != af)
continue;
strlcpy(thisifname, link_ntoa(&psa->sdl), IFNAMSIZ);
pcolon = strchr(thisifname, ':');
if (pcolon)
*pcolon = '\0';
/* Only print the banner for the first ifmaddrs entry. */
if (lastifindex == 0 || lastifindex != thisifindex) {
lastifindex = thisifindex;
fprintf(stdout, "%s:\n", thisifname);
}
/*
* Currently, multicast joins only take place on the
* primary IPv4 address, and only on the link-local IPv6
* address, as per IGMPv2/3 and MLDv1/2 semantics.
* Therefore, we only look up the primary address on
* the first pass.
*/
pifasa = NULL;
for (ifa = ifap; ifa; ifa = ifa->ifa_next) {
if ((strcmp(ifa->ifa_name, thisifname) != 0) ||
(ifa->ifa_addr == NULL) ||
(ifa->ifa_addr->sa_family != pgsa->sa.sa_family))
continue;
/*
* For AF_INET6 only the link-local address should
* be returned. If built without IPv6 support,
* skip this address entirely.
*/
pifasa = (sockunion_t *)ifa->ifa_addr;
if (pifasa->sa.sa_family == AF_INET6
#ifdef INET6
&& !IN6_IS_ADDR_LINKLOCAL(&pifasa->sin6.sin6_addr)
#endif
) {
pifasa = NULL;
continue;
}
break;
}
if (pifasa == NULL)
continue; /* primary address not found */
if (!vflag && pifasa->sa.sa_family == AF_LINK)
continue;
/* Parse and print primary address, if not already printed. */
if (lastifasa.ss.ss_family == AF_UNSPEC ||
((lastifasa.ss.ss_family == AF_LINK &&
!sa_dl_equal(&lastifasa.sa, &pifasa->sa)) ||
!sa_equal(&lastifasa.sa, &pifasa->sa))) {
switch (pifasa->sa.sa_family) {
case AF_INET:
pafname = "inet";
break;
case AF_INET6:
pafname = "inet6";
break;
case AF_LINK:
pafname = "link";
break;
default:
pafname = "unknown";
break;
}
switch (pifasa->sa.sa_family) {
case AF_INET6:
#ifdef INET6
{
const char *p =
inet6_n2a(&pifasa->sin6.sin6_addr,
pifasa->sin6.sin6_scope_id);
strlcpy(addrbuf, p, sizeof(addrbuf));
break;
}
#else
/* FALLTHROUGH */
#endif
case AF_INET:
case AF_LINK:
error = getnameinfo(&pifasa->sa,
pifasa->sa.sa_len,
addrbuf, sizeof(addrbuf), NULL, 0,
NI_NUMERICHOST);
if (error)
perror("getnameinfo");
break;
default:
addrbuf[0] = '\0';
break;
}
fprintf(stdout, "\t%s %s", pafname, addrbuf);
#ifdef INET6
if (pifasa->sa.sa_family == AF_INET6 &&
pifasa->sin6.sin6_scope_id)
fprintf(stdout, " scopeid 0x%x",
pifasa->sin6.sin6_scope_id);
#endif
fprintf(stdout, "\n");
#ifdef INET
/*
* Print per-link IGMP information, if available.
*/
if (pifasa->sa.sa_family == AF_INET) {
struct igmp_ifinfo igi;
size_t mibsize, len;
int mib[5];
mibsize = nitems(mib);
if (sysctlnametomib("net.inet.igmp.ifinfo",
mib, &mibsize) == -1) {
perror("sysctlnametomib");
goto next_ifnet;
}
mib[mibsize] = thisifindex;
len = sizeof(struct igmp_ifinfo);
if (sysctl(mib, mibsize + 1, &igi, &len, NULL,
0) == -1) {
perror("sysctl net.inet.igmp.ifinfo");
goto next_ifnet;
}
in_ifinfo(&igi);
}
Bite the bullet, and make the IPv6 SSM and MLDv2 mega-commit: import from p4 bms_netdev. Summary of changes: * Connect netinet6/in6_mcast.c to build. The legacy KAME KPIs are mostly preserved. * Eliminate now dead code from ip6_output.c. Don't do mbuf bingo, we are not going to do RFC 2292 style CMSG tricks for multicast options as they are not required by any current IPv6 normative reference. * Refactor transports (UDP, raw_ip6) to do own mcast filtering. SCTP, TCP unaffected by this change. * Add ip6_msource, in6_msource structs to in6_var.h. * Hookup mld_ifinfo state to in6_ifextra, allocate from domifattach path. * Eliminate IN6_LOOKUP_MULTI(), it is no longer referenced. Kernel consumers which need this should use in6m_lookup(). * Refactor IPv6 socket group memberships to use a vector (like IPv4). * Update ifmcstat(8) for IPv6 SSM. * Add witness lock order for IN6_MULTI_LOCK. * Move IN6_MULTI_LOCK out of lower ip6_output()/ip6_input() paths. * Introduce IP6STAT_ADD/SUB/INC/DEC as per rwatson's IPv4 cleanup. * Update carp(4) for new IPv6 SSM KPIs. * Virtualize ip6_mrouter socket. Changes mostly localized to IPv6 MROUTING. * Don't do a local group lookup in MROUTING. * Kill unused KAME prototypes in6_purgemkludge(), in6_restoremkludge(). * Preserve KAME DAD timer jitter behaviour in MLDv1 compatibility mode. * Bump __FreeBSD_version to 800084. * Update UPDATING. NOTE WELL: * This code hasn't been tested against real MLDv2 queriers (yet), although the on-wire protocol has been verified in Wireshark. * There are a few unresolved issues in the socket layer APIs to do with scope ID propagation. * There is a LOR present in ip6_output()'s use of in6_setscope() which needs to be resolved. See comments in mld6.c. This is believed to be benign and can't be avoided for the moment without re-introducing an indirect netisr. This work was mostly derived from the IGMPv3 implementation, and has been sponsored by a third party.
2009-04-29 19:19:13 +00:00
#endif /* INET */
#ifdef INET6
/*
* Print per-link MLD information, if available.
*/
if (pifasa->sa.sa_family == AF_INET6) {
struct mld_ifinfo mli;
size_t mibsize, len;
int mib[5];
mibsize = nitems(mib);
Bite the bullet, and make the IPv6 SSM and MLDv2 mega-commit: import from p4 bms_netdev. Summary of changes: * Connect netinet6/in6_mcast.c to build. The legacy KAME KPIs are mostly preserved. * Eliminate now dead code from ip6_output.c. Don't do mbuf bingo, we are not going to do RFC 2292 style CMSG tricks for multicast options as they are not required by any current IPv6 normative reference. * Refactor transports (UDP, raw_ip6) to do own mcast filtering. SCTP, TCP unaffected by this change. * Add ip6_msource, in6_msource structs to in6_var.h. * Hookup mld_ifinfo state to in6_ifextra, allocate from domifattach path. * Eliminate IN6_LOOKUP_MULTI(), it is no longer referenced. Kernel consumers which need this should use in6m_lookup(). * Refactor IPv6 socket group memberships to use a vector (like IPv4). * Update ifmcstat(8) for IPv6 SSM. * Add witness lock order for IN6_MULTI_LOCK. * Move IN6_MULTI_LOCK out of lower ip6_output()/ip6_input() paths. * Introduce IP6STAT_ADD/SUB/INC/DEC as per rwatson's IPv4 cleanup. * Update carp(4) for new IPv6 SSM KPIs. * Virtualize ip6_mrouter socket. Changes mostly localized to IPv6 MROUTING. * Don't do a local group lookup in MROUTING. * Kill unused KAME prototypes in6_purgemkludge(), in6_restoremkludge(). * Preserve KAME DAD timer jitter behaviour in MLDv1 compatibility mode. * Bump __FreeBSD_version to 800084. * Update UPDATING. NOTE WELL: * This code hasn't been tested against real MLDv2 queriers (yet), although the on-wire protocol has been verified in Wireshark. * There are a few unresolved issues in the socket layer APIs to do with scope ID propagation. * There is a LOR present in ip6_output()'s use of in6_setscope() which needs to be resolved. See comments in mld6.c. This is believed to be benign and can't be avoided for the moment without re-introducing an indirect netisr. This work was mostly derived from the IGMPv3 implementation, and has been sponsored by a third party.
2009-04-29 19:19:13 +00:00
if (sysctlnametomib("net.inet6.mld.ifinfo",
mib, &mibsize) == -1) {
perror("sysctlnametomib");
goto next_ifnet;
}
mib[mibsize] = thisifindex;
len = sizeof(struct mld_ifinfo);
if (sysctl(mib, mibsize + 1, &mli, &len, NULL,
0) == -1) {
perror("sysctl net.inet6.mld.ifinfo");
goto next_ifnet;
}
in6_ifinfo(&mli);
}
#endif /* INET6 */
#if defined(INET) || defined(INET6)
next_ifnet:
#endif
lastifasa = *pifasa;
}
/* Print this group address. */
#ifdef INET6
if (pgsa->sa.sa_family == AF_INET6) {
const char *p = inet6_n2a(&pgsa->sin6.sin6_addr,
pgsa->sin6.sin6_scope_id);
strlcpy(addrbuf, p, sizeof(addrbuf));
} else
#endif
{
error = getnameinfo(&pgsa->sa, pgsa->sa.sa_len,
addrbuf, sizeof(addrbuf), NULL, 0, NI_NUMERICHOST);
if (error)
perror("getnameinfo");
}
fprintf(stdout, "\t\tgroup %s", addrbuf);
#ifdef INET6
if (pgsa->sa.sa_family == AF_INET6 &&
pgsa->sin6.sin6_scope_id)
fprintf(stdout, " scopeid 0x%x",
pgsa->sin6.sin6_scope_id);
#endif
#ifdef INET
if (pgsa->sa.sa_family == AF_INET) {
inm_print_sources_sysctl(thisifindex,
pgsa->sin.sin_addr);
}
Bite the bullet, and make the IPv6 SSM and MLDv2 mega-commit: import from p4 bms_netdev. Summary of changes: * Connect netinet6/in6_mcast.c to build. The legacy KAME KPIs are mostly preserved. * Eliminate now dead code from ip6_output.c. Don't do mbuf bingo, we are not going to do RFC 2292 style CMSG tricks for multicast options as they are not required by any current IPv6 normative reference. * Refactor transports (UDP, raw_ip6) to do own mcast filtering. SCTP, TCP unaffected by this change. * Add ip6_msource, in6_msource structs to in6_var.h. * Hookup mld_ifinfo state to in6_ifextra, allocate from domifattach path. * Eliminate IN6_LOOKUP_MULTI(), it is no longer referenced. Kernel consumers which need this should use in6m_lookup(). * Refactor IPv6 socket group memberships to use a vector (like IPv4). * Update ifmcstat(8) for IPv6 SSM. * Add witness lock order for IN6_MULTI_LOCK. * Move IN6_MULTI_LOCK out of lower ip6_output()/ip6_input() paths. * Introduce IP6STAT_ADD/SUB/INC/DEC as per rwatson's IPv4 cleanup. * Update carp(4) for new IPv6 SSM KPIs. * Virtualize ip6_mrouter socket. Changes mostly localized to IPv6 MROUTING. * Don't do a local group lookup in MROUTING. * Kill unused KAME prototypes in6_purgemkludge(), in6_restoremkludge(). * Preserve KAME DAD timer jitter behaviour in MLDv1 compatibility mode. * Bump __FreeBSD_version to 800084. * Update UPDATING. NOTE WELL: * This code hasn't been tested against real MLDv2 queriers (yet), although the on-wire protocol has been verified in Wireshark. * There are a few unresolved issues in the socket layer APIs to do with scope ID propagation. * There is a LOR present in ip6_output()'s use of in6_setscope() which needs to be resolved. See comments in mld6.c. This is believed to be benign and can't be avoided for the moment without re-introducing an indirect netisr. This work was mostly derived from the IGMPv3 implementation, and has been sponsored by a third party.
2009-04-29 19:19:13 +00:00
#endif
#ifdef INET6
if (pgsa->sa.sa_family == AF_INET6) {
in6m_print_sources_sysctl(thisifindex,
&pgsa->sin6.sin6_addr);
}
#endif
fprintf(stdout, "\n");
/* Link-layer mapping, if present. */
pllsa = (sockunion_t *)ifma->ifma_lladdr;
if (pllsa != NULL) {
error = getnameinfo(&pllsa->sa, pllsa->sa.sa_len,
addrbuf, sizeof(addrbuf), NULL, 0, NI_NUMERICHOST);
fprintf(stdout, "\t\t\tmcast-macaddr %s\n", addrbuf);
}
}
out:
if (ifmap != NULL)
freeifmaddrs(ifmap);
if (ifap != NULL)
freeifaddrs(ifap);
return (error);
}