freebsd-dev/lib/libc/net/name6.c
2004-07-21 17:26:40 +00:00

2394 lines
56 KiB
C

/* $KAME: name6.c,v 1.25 2000/06/26 16:44:40 itojun Exp $ */
/*
* Copyright (C) 1995, 1996, 1997, 1998, and 1999 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.
*/
/*
* ++Copyright++ 1985, 1988, 1993
* -
* Copyright (c) 1985, 1988, 1993
* The Regents of the University of California. 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. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by the University of
* California, Berkeley and its contributors.
* 4. Neither the name of the University 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 REGENTS 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 REGENTS 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.
* -
* Portions Copyright (c) 1993 by Digital Equipment Corporation.
*
* Permission to use, copy, modify, and distribute this software for any
* purpose with or without fee is hereby granted, provided that the above
* copyright notice and this permission notice appear in all copies, and that
* the name of Digital Equipment Corporation not be used in advertising or
* publicity pertaining to distribution of the document or software without
* specific, written prior permission.
*
* THE SOFTWARE IS PROVIDED "AS IS" AND DIGITAL EQUIPMENT CORP. DISCLAIMS ALL
* WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES
* OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL DIGITAL EQUIPMENT
* CORPORATION BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL
* DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR
* PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS
* ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS
* SOFTWARE.
* -
* --Copyright--
*/
/*
* Atsushi Onoe <onoe@sm.sony.co.jp>
*/
/*
* TODO for thread safe
* use mutex for _hostconf, _hostconf_init.
* rewrite resolvers to be thread safe
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include "namespace.h"
#include <sys/param.h>
#include <sys/socket.h>
#include <sys/time.h>
#include <sys/queue.h>
#include <netinet/in.h>
#ifdef INET6
#include <net/if.h>
#include <net/if_var.h>
#include <sys/sysctl.h>
#include <sys/ioctl.h>
#include <netinet6/in6_var.h> /* XXX */
#endif
#include <arpa/inet.h>
#include <arpa/nameser.h>
#include <errno.h>
#include <netdb.h>
#include <resolv.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <stdarg.h>
#include <nsswitch.h>
#include <pthread.h>
#include <unistd.h>
#include "un-namespace.h"
#ifndef _PATH_HOSTS
#define _PATH_HOSTS "/etc/hosts"
#endif
#ifndef MAXALIASES
#define MAXALIASES 10
#endif
#ifndef MAXADDRS
#define MAXADDRS 20
#endif
#ifndef MAXDNAME
#define MAXDNAME 1025
#endif
#ifdef INET6
#define ADDRLEN(af) ((af) == AF_INET6 ? sizeof(struct in6_addr) : \
sizeof(struct in_addr))
#else
#define ADDRLEN(af) sizeof(struct in_addr)
#endif
#define MAPADDR(ab, ina) \
do { \
memcpy(&(ab)->map_inaddr, ina, sizeof(struct in_addr)); \
memset((ab)->map_zero, 0, sizeof((ab)->map_zero)); \
memset((ab)->map_one, 0xff, sizeof((ab)->map_one)); \
} while (0)
#define MAPADDRENABLED(flags) \
(((flags) & AI_V4MAPPED) || \
(((flags) & AI_V4MAPPED_CFG) && _mapped_addr_enabled()))
union inx_addr {
struct in_addr in_addr;
#ifdef INET6
struct in6_addr in6_addr;
#endif
struct {
u_char mau_zero[10];
u_char mau_one[2];
struct in_addr mau_inaddr;
} map_addr_un;
#define map_zero map_addr_un.mau_zero
#define map_one map_addr_un.mau_one
#define map_inaddr map_addr_un.mau_inaddr
};
struct policyqueue {
TAILQ_ENTRY(policyqueue) pc_entry;
#ifdef INET6
struct in6_addrpolicy pc_policy;
#endif
};
TAILQ_HEAD(policyhead, policyqueue);
#define AIO_SRCFLAG_DEPRECATED 0x1
struct hp_order {
union {
struct sockaddr_storage aiou_ss;
struct sockaddr aiou_sa;
} aio_src_un;
#define aio_srcsa aio_src_un.aiou_sa
u_int32_t aio_srcflag;
int aio_srcscope;
int aio_dstscope;
struct policyqueue *aio_srcpolicy;
struct policyqueue *aio_dstpolicy;
union {
struct sockaddr_storage aiou_ss;
struct sockaddr aiou_sa;
} aio_un;
#define aio_sa aio_un.aiou_sa
int aio_matchlen;
u_char *aio_h_addr;
};
static struct hostent *_hpcopy(struct hostent *hp, int *errp);
static struct hostent *_hpaddr(int af, const char *name, void *addr, int *errp);
static struct hostent *_hpmerge(struct hostent *hp1, struct hostent *hp2, int *errp);
#ifdef INET6
static struct hostent *_hpmapv6(struct hostent *hp, int *errp);
#endif
static struct hostent *_hpsort(struct hostent *hp);
static struct hostent *_ghbyname(const char *name, int af, int flags, int *errp);
static char *_hgetword(char **pp);
static int _mapped_addr_enabled(void);
static struct hostent *_hpreorder(struct hostent *hp);
static int get_addrselectpolicy(struct policyhead *);
static void free_addrselectpolicy(struct policyhead *);
static struct policyqueue *match_addrselectpolicy(struct sockaddr *,
struct policyhead *);
static void set_source(struct hp_order *, struct policyhead *);
static int matchlen(struct sockaddr *, struct sockaddr *);
static int comp_dst(const void *, const void *);
static int gai_addr2scopetype(struct sockaddr *);
static FILE *_files_open(int *errp);
static int _files_ghbyname(void *, void *, va_list);
static int _files_ghbyaddr(void *, void *, va_list);
#ifdef YP
static int _nis_ghbyname(void *, void *, va_list);
static int _nis_ghbyaddr(void *, void *, va_list);
#endif
static int _dns_ghbyname(void *, void *, va_list);
static int _dns_ghbyaddr(void *, void *, va_list);
static void _dns_shent(int stayopen) __unused;
static void _dns_ehent(void) __unused;
#ifdef ICMPNL
static int _icmp_ghbyaddr(void *, void *, va_list);
#endif /* ICMPNL */
/*
* XXX: Many dependencies are not thread-safe. So, we share lock between
* getaddrinfo() and getipnodeby*(). Still, we cannot use
* getaddrinfo() and getipnodeby*() in conjunction with other
* functions which call them.
*/
#include "libc_private.h"
extern pthread_mutex_t __getaddrinfo_thread_lock;
#define THREAD_LOCK() \
if (__isthreaded) _pthread_mutex_lock(&__getaddrinfo_thread_lock);
#define THREAD_UNLOCK() \
if (__isthreaded) _pthread_mutex_unlock(&__getaddrinfo_thread_lock);
/* Host lookup order if nsswitch.conf is broken or nonexistant */
static const ns_src default_src[] = {
{ NSSRC_FILES, NS_SUCCESS },
{ NSSRC_DNS, NS_SUCCESS },
#ifdef ICMPNL
#define NSSRC_ICMP "icmp"
{ NSSRC_ICMP, NS_SUCCESS },
#endif
{ 0 }
};
/*
* Check if kernel supports mapped address.
* implementation dependent
*/
#ifdef __KAME__
#include <sys/sysctl.h>
#endif /* __KAME__ */
static int
_mapped_addr_enabled(void)
{
/* implementation dependent check */
#if defined(__KAME__) && defined(IPV6CTL_MAPPED_ADDR)
int mib[4];
size_t len;
int val;
mib[0] = CTL_NET;
mib[1] = PF_INET6;
mib[2] = IPPROTO_IPV6;
mib[3] = IPV6CTL_MAPPED_ADDR;
len = sizeof(val);
if (sysctl(mib, 4, &val, &len, 0, 0) == 0 && val != 0)
return 1;
#endif /* __KAME__ && IPV6CTL_MAPPED_ADDR */
return 0;
}
/*
* Functions defined in RFC2553
* getipnodebyname, getipnodebyaddr, freehostent
*/
static struct hostent *
_ghbyname(const char *name, int af, int flags, int *errp)
{
struct hostent *hp;
int rval;
static const ns_dtab dtab[] = {
NS_FILES_CB(_files_ghbyname, NULL)
{ NSSRC_DNS, _dns_ghbyname, NULL },
NS_NIS_CB(_nis_ghbyname, NULL)
{ 0 }
};
if (flags & AI_ADDRCONFIG) {
int s;
/*
* TODO:
* Note that implementation dependent test for address
* configuration should be done everytime called
* (or apropriate interval),
* because addresses will be dynamically assigned or deleted.
*/
if (af == AF_UNSPEC) {
if ((s = _socket(AF_INET6, SOCK_DGRAM, 0)) < 0)
af = AF_INET;
else {
_close(s);
if ((s = _socket(AF_INET, SOCK_DGRAM, 0)) < 0)
af = AF_INET6;
else
_close(s);
}
}
if (af != AF_UNSPEC) {
if ((s = _socket(af, SOCK_DGRAM, 0)) < 0)
return NULL;
_close(s);
}
}
rval = _nsdispatch(&hp, dtab, NSDB_HOSTS, "ghbyname", default_src,
name, af, errp);
return (rval == NS_SUCCESS) ? hp : NULL;
}
/* getipnodebyname() internal routine for multiple query(PF_UNSPEC) support. */
struct hostent *
_getipnodebyname_multi(const char *name, int af, int flags, int *errp)
{
struct hostent *hp;
union inx_addr addrbuf;
/* XXX: PF_UNSPEC is only supposed to be passed from getaddrinfo() */
if (af != AF_INET
#ifdef INET6
&& af != AF_INET6
#endif
&& af != PF_UNSPEC
)
{
*errp = NO_RECOVERY;
return NULL;
}
#ifdef INET6
/* special case for literal address */
if (inet_pton(AF_INET6, name, &addrbuf) == 1) {
if (af != AF_INET6) {
*errp = HOST_NOT_FOUND;
return NULL;
}
return _hpaddr(af, name, &addrbuf, errp);
}
#endif
if (inet_aton(name, (struct in_addr *)&addrbuf) == 1) {
if (af != AF_INET) {
if (MAPADDRENABLED(flags)) {
MAPADDR(&addrbuf, &addrbuf.in_addr);
} else {
*errp = HOST_NOT_FOUND;
return NULL;
}
}
return _hpaddr(af, name, &addrbuf, errp);
}
*errp = HOST_NOT_FOUND;
hp = _ghbyname(name, af, flags, errp);
#ifdef INET6
if (af == AF_INET6
&& ((flags & AI_ALL) || hp == NULL)
&& (MAPADDRENABLED(flags))) {
struct hostent *hp2 = _ghbyname(name, AF_INET, flags, errp);
if (hp == NULL)
hp = _hpmapv6(hp2, errp);
else {
if (hp2 && strcmp(hp->h_name, hp2->h_name) != 0) {
freehostent(hp2);
hp2 = NULL;
}
hp = _hpmerge(hp, hp2, errp);
}
}
#endif
return _hpreorder(_hpsort(hp));
}
struct hostent *
getipnodebyname(const char *name, int af, int flags, int *errp)
{
if (af != AF_INET
#ifdef INET6
&& af != AF_INET6
#endif
)
{
*errp = NO_RECOVERY;
return NULL;
}
return(_getipnodebyname_multi(name, af ,flags, errp));
}
struct hostent *
getipnodebyaddr(const void *src, size_t len, int af, int *errp)
{
struct hostent *hp;
int rval;
#ifdef INET6
struct in6_addr addrbuf;
#else
struct in_addr addrbuf;
#endif
static const ns_dtab dtab[] = {
NS_FILES_CB(_files_ghbyaddr, NULL)
{ NSSRC_DNS, _dns_ghbyaddr, NULL },
NS_NIS_CB(_nis_ghbyaddr, NULL)
#ifdef ICMPNL
{ NSSRC_ICMP, _icmp_ghbyaddr, NULL },
#endif
{ 0 }
};
*errp = HOST_NOT_FOUND;
switch (af) {
case AF_INET:
if (len != sizeof(struct in_addr)) {
*errp = NO_RECOVERY;
return NULL;
}
if ((long)src & ~(sizeof(struct in_addr) - 1)) {
memcpy(&addrbuf, src, len);
src = &addrbuf;
}
if (((struct in_addr *)src)->s_addr == 0)
return NULL;
break;
#ifdef INET6
case AF_INET6:
if (len != sizeof(struct in6_addr)) {
*errp = NO_RECOVERY;
return NULL;
}
if ((long)src & ~(sizeof(struct in6_addr) / 2 - 1)) { /*XXX*/
memcpy(&addrbuf, src, len);
src = &addrbuf;
}
if (IN6_IS_ADDR_UNSPECIFIED((struct in6_addr *)src))
return NULL;
if (IN6_IS_ADDR_V4MAPPED((struct in6_addr *)src)
|| IN6_IS_ADDR_V4COMPAT((struct in6_addr *)src)) {
src = (char *)src +
(sizeof(struct in6_addr) - sizeof(struct in_addr));
af = AF_INET;
len = sizeof(struct in_addr);
}
break;
#endif
default:
*errp = NO_RECOVERY;
return NULL;
}
rval = _nsdispatch(&hp, dtab, NSDB_HOSTS, "ghbyaddr", default_src,
src, len, af, errp);
return (rval == NS_SUCCESS) ? hp : NULL;
}
void
freehostent(struct hostent *ptr)
{
free(ptr);
}
#if 0
/* XXX: should be deprecated */
struct hostent *
getnodebyname(const char *name, int af, int flags)
{
return getipnodebyname(name, af, flags, &h_errno);
}
#ifdef __warn_references
__warn_references(getnodebyname,
"warning: getnodebyname() deprecated, "
"should use getaddrinfo() or getipnodebyname()");
#endif
struct hostent *
getnodebyaddr(const void *src, size_t len, int af)
{
return getipnodebyaddr(src, len, af, &h_errno);
}
#ifdef __warn_references
__warn_references(getnodebyaddr,
"warning: getnodebyaddr() deprecated, "
"should use getnameinfo() or getipnodebyaddr()");
#endif
#endif
/*
* Private utility functions
*/
/*
* _hpcopy: allocate and copy hostent structure
*/
static struct hostent *
_hpcopy(struct hostent *hp, int *errp)
{
struct hostent *nhp;
char *cp, **pp;
int size, addrsize;
int nalias = 0, naddr = 0;
int al_off;
int i;
if (hp == NULL)
return hp;
/* count size to be allocated */
size = sizeof(struct hostent);
if (hp->h_name != NULL)
size += strlen(hp->h_name) + 1;
if ((pp = hp->h_aliases) != NULL) {
for (i = 0; *pp != NULL; i++, pp++) {
if (**pp != '\0') {
size += strlen(*pp) + 1;
nalias++;
}
}
}
/* adjust alignment */
size = ALIGN(size);
al_off = size;
size += sizeof(char *) * (nalias + 1);
addrsize = ALIGN(hp->h_length);
if ((pp = hp->h_addr_list) != NULL) {
while (*pp++ != NULL)
naddr++;
}
size += addrsize * naddr;
size += sizeof(char *) * (naddr + 1);
/* copy */
if ((nhp = (struct hostent *)malloc(size)) == NULL) {
*errp = TRY_AGAIN;
return NULL;
}
cp = (char *)&nhp[1];
if (hp->h_name != NULL) {
nhp->h_name = cp;
strcpy(cp, hp->h_name);
cp += strlen(cp) + 1;
} else
nhp->h_name = NULL;
nhp->h_aliases = (char **)((char *)nhp + al_off);
if ((pp = hp->h_aliases) != NULL) {
for (i = 0; *pp != NULL; pp++) {
if (**pp != '\0') {
nhp->h_aliases[i++] = cp;
strcpy(cp, *pp);
cp += strlen(cp) + 1;
}
}
}
nhp->h_aliases[nalias] = NULL;
cp = (char *)&nhp->h_aliases[nalias + 1];
nhp->h_addrtype = hp->h_addrtype;
nhp->h_length = hp->h_length;
nhp->h_addr_list = (char **)cp;
if ((pp = hp->h_addr_list) != NULL) {
cp = (char *)&nhp->h_addr_list[naddr + 1];
for (i = 0; *pp != NULL; pp++) {
nhp->h_addr_list[i++] = cp;
memcpy(cp, *pp, hp->h_length);
cp += addrsize;
}
}
nhp->h_addr_list[naddr] = NULL;
return nhp;
}
/*
* _hpaddr: construct hostent structure with one address
*/
static struct hostent *
_hpaddr(int af, const char *name, void *addr, int *errp)
{
struct hostent *hp, hpbuf;
char *addrs[2];
hp = &hpbuf;
hp->h_name = (char *)name;
hp->h_aliases = NULL;
hp->h_addrtype = af;
hp->h_length = ADDRLEN(af);
hp->h_addr_list = addrs;
addrs[0] = (char *)addr;
addrs[1] = NULL;
return _hpcopy(hp, errp);
}
/*
* _hpmerge: merge 2 hostent structure, arguments will be freed
*/
static struct hostent *
_hpmerge(struct hostent *hp1, struct hostent *hp2, int *errp)
{
int i, j;
int naddr, nalias;
char **pp;
struct hostent *hp, hpbuf;
char *aliases[MAXALIASES + 1], *addrs[MAXADDRS + 1];
union inx_addr addrbuf[MAXADDRS];
if (hp1 == NULL)
return hp2;
if (hp2 == NULL)
return hp1;
#define HP(i) (i == 1 ? hp1 : hp2)
hp = &hpbuf;
hp->h_name = (hp1->h_name != NULL ? hp1->h_name : hp2->h_name);
hp->h_aliases = aliases;
nalias = 0;
for (i = 1; i <= 2; i++) {
if ((pp = HP(i)->h_aliases) == NULL)
continue;
for (; nalias < MAXALIASES && *pp != NULL; pp++) {
/* check duplicates */
for (j = 0; j < nalias; j++)
if (strcasecmp(*pp, aliases[j]) == 0)
break;
if (j == nalias)
aliases[nalias++] = *pp;
}
}
aliases[nalias] = NULL;
#ifdef INET6
if (hp1->h_length != hp2->h_length) {
hp->h_addrtype = AF_INET6;
hp->h_length = sizeof(struct in6_addr);
} else {
#endif
hp->h_addrtype = hp1->h_addrtype;
hp->h_length = hp1->h_length;
#ifdef INET6
}
#endif
hp->h_addr_list = addrs;
naddr = 0;
for (i = 1; i <= 2; i++) {
if ((pp = HP(i)->h_addr_list) == NULL)
continue;
if (HP(i)->h_length == hp->h_length) {
while (naddr < MAXADDRS && *pp != NULL)
addrs[naddr++] = *pp++;
} else {
/* copy IPv4 addr as mapped IPv6 addr */
while (naddr < MAXADDRS && *pp != NULL) {
MAPADDR(&addrbuf[naddr], *pp++);
addrs[naddr] = (char *)&addrbuf[naddr];
naddr++;
}
}
}
addrs[naddr] = NULL;
hp = _hpcopy(hp, errp);
freehostent(hp1);
freehostent(hp2);
return hp;
}
/*
* _hpmapv6: convert IPv4 hostent into IPv4-mapped IPv6 addresses
*/
#ifdef INET6
static struct hostent *
_hpmapv6(struct hostent *hp, int *errp)
{
struct hostent *hp6;
if (hp == NULL)
return NULL;
if (hp->h_addrtype == AF_INET6)
return hp;
/* make dummy hostent to convert IPv6 address */
if ((hp6 = (struct hostent *)malloc(sizeof(struct hostent))) == NULL) {
*errp = TRY_AGAIN;
return NULL;
}
hp6->h_name = NULL;
hp6->h_aliases = NULL;
hp6->h_addrtype = AF_INET6;
hp6->h_length = sizeof(struct in6_addr);
hp6->h_addr_list = NULL;
return _hpmerge(hp6, hp, errp);
}
#endif
/*
* _hpsort: sort address by sortlist
*/
static struct hostent *
_hpsort(struct hostent *hp)
{
int i, j, n;
u_char *ap, *sp, *mp, **pp;
char t;
char order[MAXADDRS];
int nsort = _res.nsort;
if (hp == NULL || hp->h_addr_list[1] == NULL || nsort == 0)
return hp;
for (i = 0; (ap = (u_char *)hp->h_addr_list[i]); i++) {
for (j = 0; j < nsort; j++) {
#ifdef INET6
if (_res_ext.sort_list[j].af != hp->h_addrtype)
continue;
sp = (u_char *)&_res_ext.sort_list[j].addr;
mp = (u_char *)&_res_ext.sort_list[j].mask;
#else
sp = (u_char *)&_res.sort_list[j].addr;
mp = (u_char *)&_res.sort_list[j].mask;
#endif
for (n = 0; n < hp->h_length; n++) {
if ((ap[n] & mp[n]) != sp[n])
break;
}
if (n == hp->h_length)
break;
}
order[i] = j;
}
n = i;
pp = (u_char **)hp->h_addr_list;
for (i = 0; i < n - 1; i++) {
for (j = i + 1; j < n; j++) {
if (order[i] > order[j]) {
ap = pp[i];
pp[i] = pp[j];
pp[j] = ap;
t = order[i];
order[i] = order[j];
order[j] = t;
}
}
}
return hp;
}
static char *
_hgetword(char **pp)
{
char c, *p, *ret;
const char *sp;
static const char sep[] = "# \t\n";
ret = NULL;
for (p = *pp; (c = *p) != '\0'; p++) {
for (sp = sep; *sp != '\0'; sp++) {
if (c == *sp)
break;
}
if (c == '#')
p[1] = '\0'; /* ignore rest of line */
if (ret == NULL) {
if (*sp == '\0')
ret = p;
} else {
if (*sp != '\0') {
*p++ = '\0';
break;
}
}
}
*pp = p;
if (ret == NULL || *ret == '\0')
return NULL;
return ret;
}
/*
* _hpreorder: sort address by default address selection
*/
static struct hostent *
_hpreorder(struct hostent *hp)
{
struct hp_order *aio;
int i, n;
u_char *ap;
struct sockaddr *sa;
struct policyhead policyhead;
if (hp == NULL)
return hp;
switch (hp->h_addrtype) {
case AF_INET:
#ifdef INET6
case AF_INET6:
#endif
break;
default:
free_addrselectpolicy(&policyhead);
return hp;
}
/* count the number of addrinfo elements for sorting. */
for (n = 0; hp->h_addr_list[n] != NULL; n++)
;
/*
* If the number is small enough, we can skip the reordering process.
*/
if (n <= 1)
return hp;
/* allocate a temporary array for sort and initialization of it. */
if ((aio = malloc(sizeof(*aio) * n)) == NULL)
return hp; /* give up reordering */
memset(aio, 0, sizeof(*aio) * n);
/* retrieve address selection policy from the kernel */
TAILQ_INIT(&policyhead);
if (!get_addrselectpolicy(&policyhead)) {
/* no policy is installed into kernel, we don't sort. */
free(aio);
return hp;
}
for (i = 0; i < n; i++) {
ap = (u_char *)hp->h_addr_list[i];
aio[i].aio_h_addr = ap;
sa = &aio[i].aio_sa;
switch (hp->h_addrtype) {
case AF_INET:
sa->sa_family = AF_INET;
sa->sa_len = sizeof(struct sockaddr_in);
memcpy(&((struct sockaddr_in *)sa)->sin_addr, ap,
sizeof(struct in_addr));
break;
#ifdef INET6
case AF_INET6:
if (IN6_IS_ADDR_V4MAPPED((struct in6_addr *)ap)) {
sa->sa_family = AF_INET;
sa->sa_len = sizeof(struct sockaddr_in);
memcpy(&((struct sockaddr_in *)sa)->sin_addr,
&ap[12], sizeof(struct in_addr));
} else {
sa->sa_family = AF_INET6;
sa->sa_len = sizeof(struct sockaddr_in6);
memcpy(&((struct sockaddr_in6 *)sa)->sin6_addr,
ap, sizeof(struct in6_addr));
}
break;
#endif
}
aio[i].aio_dstscope = gai_addr2scopetype(sa);
aio[i].aio_dstpolicy = match_addrselectpolicy(sa, &policyhead);
set_source(&aio[i], &policyhead);
}
/* perform sorting. */
qsort(aio, n, sizeof(*aio), comp_dst);
/* reorder the h_addr_list. */
for (i = 0; i < n; i++)
hp->h_addr_list[i] = aio[i].aio_h_addr;
/* cleanup and return */
free(aio);
free_addrselectpolicy(&policyhead);
return hp;
}
static int
get_addrselectpolicy(head)
struct policyhead *head;
{
#ifdef INET6
int mib[] = { CTL_NET, PF_INET6, IPPROTO_IPV6, IPV6CTL_ADDRCTLPOLICY };
size_t l;
char *buf;
struct in6_addrpolicy *pol, *ep;
if (sysctl(mib, sizeof(mib) / sizeof(mib[0]), NULL, &l, NULL, 0) < 0)
return (0);
if ((buf = malloc(l)) == NULL)
return (0);
if (sysctl(mib, sizeof(mib) / sizeof(mib[0]), buf, &l, NULL, 0) < 0) {
free(buf);
return (0);
}
ep = (struct in6_addrpolicy *)(buf + l);
for (pol = (struct in6_addrpolicy *)buf; pol + 1 <= ep; pol++) {
struct policyqueue *new;
if ((new = malloc(sizeof(*new))) == NULL) {
free_addrselectpolicy(head); /* make the list empty */
break;
}
new->pc_policy = *pol;
TAILQ_INSERT_TAIL(head, new, pc_entry);
}
free(buf);
return (1);
#else
return (0);
#endif
}
static void
free_addrselectpolicy(head)
struct policyhead *head;
{
struct policyqueue *ent, *nent;
for (ent = TAILQ_FIRST(head); ent; ent = nent) {
nent = TAILQ_NEXT(ent, pc_entry);
TAILQ_REMOVE(head, ent, pc_entry);
free(ent);
}
}
static struct policyqueue *
match_addrselectpolicy(addr, head)
struct sockaddr *addr;
struct policyhead *head;
{
#ifdef INET6
struct policyqueue *ent, *bestent = NULL;
struct in6_addrpolicy *pol;
int matchlen, bestmatchlen = -1;
u_char *mp, *ep, *k, *p, m;
struct sockaddr_in6 key;
switch(addr->sa_family) {
case AF_INET6:
key = *(struct sockaddr_in6 *)addr;
break;
case AF_INET:
/* convert the address into IPv4-mapped IPv6 address. */
memset(&key, 0, sizeof(key));
key.sin6_family = AF_INET6;
key.sin6_len = sizeof(key);
key.sin6_addr.s6_addr[10] = 0xff;
key.sin6_addr.s6_addr[11] = 0xff;
memcpy(&key.sin6_addr.s6_addr[12],
&((struct sockaddr_in *)addr)->sin_addr, 4);
break;
default:
return(NULL);
}
for (ent = TAILQ_FIRST(head); ent; ent = TAILQ_NEXT(ent, pc_entry)) {
pol = &ent->pc_policy;
matchlen = 0;
mp = (u_char *)&pol->addrmask.sin6_addr;
ep = mp + 16; /* XXX: scope field? */
k = (u_char *)&key.sin6_addr;
p = (u_char *)&pol->addr.sin6_addr;
for (; mp < ep && *mp; mp++, k++, p++) {
m = *mp;
if ((*k & m) != *p)
goto next; /* not match */
if (m == 0xff) /* short cut for a typical case */
matchlen += 8;
else {
while (m >= 0x80) {
matchlen++;
m <<= 1;
}
}
}
/* matched. check if this is better than the current best. */
if (matchlen > bestmatchlen) {
bestent = ent;
bestmatchlen = matchlen;
}
next:
continue;
}
return(bestent);
#else
return(NULL);
#endif
}
static void
set_source(aio, ph)
struct hp_order *aio;
struct policyhead *ph;
{
struct sockaddr_storage ss = aio->aio_un.aiou_ss;
int s, srclen;
/* set unspec ("no source is available"), just in case */
aio->aio_srcsa.sa_family = AF_UNSPEC;
aio->aio_srcscope = -1;
switch(ss.ss_family) {
case AF_INET:
((struct sockaddr_in *)&ss)->sin_port = htons(1);
break;
#ifdef INET6
case AF_INET6:
((struct sockaddr_in6 *)&ss)->sin6_port = htons(1);
break;
#endif
default: /* ignore unsupported AFs explicitly */
return;
}
/* open a socket to get the source address for the given dst */
if ((s = _socket(ss.ss_family, SOCK_DGRAM, IPPROTO_UDP)) < 0)
return; /* give up */
if (_connect(s, (struct sockaddr *)&ss, ss.ss_len) < 0)
goto cleanup;
srclen = ss.ss_len;
if (_getsockname(s, &aio->aio_srcsa, &srclen) < 0) {
aio->aio_srcsa.sa_family = AF_UNSPEC;
goto cleanup;
}
aio->aio_srcscope = gai_addr2scopetype(&aio->aio_srcsa);
aio->aio_srcpolicy = match_addrselectpolicy(&aio->aio_srcsa, ph);
aio->aio_matchlen = matchlen(&aio->aio_srcsa, (struct sockaddr *)&ss);
#ifdef INET6
if (ss.ss_family == AF_INET6) {
struct in6_ifreq ifr6;
u_int32_t flags6;
/* XXX: interface name should not be hardcoded */
strncpy(ifr6.ifr_name, "lo0", sizeof(ifr6.ifr_name));
memset(&ifr6, 0, sizeof(ifr6));
memcpy(&ifr6.ifr_addr, &ss, ss.ss_len);
if (_ioctl(s, SIOCGIFAFLAG_IN6, &ifr6) == 0) {
flags6 = ifr6.ifr_ifru.ifru_flags6;
if ((flags6 & IN6_IFF_DEPRECATED))
aio->aio_srcflag |= AIO_SRCFLAG_DEPRECATED;
}
}
#endif
cleanup:
_close(s);
return;
}
static int
matchlen(src, dst)
struct sockaddr *src, *dst;
{
int match = 0;
u_char *s, *d;
u_char *lim, r;
int addrlen;
switch (src->sa_family) {
#ifdef INET6
case AF_INET6:
s = (u_char *)&((struct sockaddr_in6 *)src)->sin6_addr;
d = (u_char *)&((struct sockaddr_in6 *)dst)->sin6_addr;
addrlen = sizeof(struct in6_addr);
lim = s + addrlen;
break;
#endif
case AF_INET:
s = (u_char *)&((struct sockaddr_in6 *)src)->sin6_addr;
d = (u_char *)&((struct sockaddr_in6 *)dst)->sin6_addr;
addrlen = sizeof(struct in_addr);
lim = s + addrlen;
break;
default:
return(0);
}
while (s < lim)
if ((r = (*d++ ^ *s++)) != 0) {
while (r < addrlen * 8) {
match++;
r <<= 1;
}
break;
} else
match += 8;
return(match);
}
static int
comp_dst(arg1, arg2)
const void *arg1, *arg2;
{
const struct hp_order *dst1 = arg1, *dst2 = arg2;
/*
* Rule 1: Avoid unusable destinations.
* XXX: we currently do not consider if an appropriate route exists.
*/
if (dst1->aio_srcsa.sa_family != AF_UNSPEC &&
dst2->aio_srcsa.sa_family == AF_UNSPEC) {
return(-1);
}
if (dst1->aio_srcsa.sa_family == AF_UNSPEC &&
dst2->aio_srcsa.sa_family != AF_UNSPEC) {
return(1);
}
/* Rule 2: Prefer matching scope. */
if (dst1->aio_dstscope == dst1->aio_srcscope &&
dst2->aio_dstscope != dst2->aio_srcscope) {
return(-1);
}
if (dst1->aio_dstscope != dst1->aio_srcscope &&
dst2->aio_dstscope == dst2->aio_srcscope) {
return(1);
}
/* Rule 3: Avoid deprecated addresses. */
if (dst1->aio_srcsa.sa_family != AF_UNSPEC &&
dst2->aio_srcsa.sa_family != AF_UNSPEC) {
if (!(dst1->aio_srcflag & AIO_SRCFLAG_DEPRECATED) &&
(dst2->aio_srcflag & AIO_SRCFLAG_DEPRECATED)) {
return(-1);
}
if ((dst1->aio_srcflag & AIO_SRCFLAG_DEPRECATED) &&
!(dst2->aio_srcflag & AIO_SRCFLAG_DEPRECATED)) {
return(1);
}
}
/* Rule 4: Prefer home addresses. */
/* XXX: not implemented yet */
/* Rule 5: Prefer matching label. */
#ifdef INET6
if (dst1->aio_srcpolicy && dst1->aio_dstpolicy &&
dst1->aio_srcpolicy->pc_policy.label ==
dst1->aio_dstpolicy->pc_policy.label &&
(dst2->aio_srcpolicy == NULL || dst2->aio_dstpolicy == NULL ||
dst2->aio_srcpolicy->pc_policy.label !=
dst2->aio_dstpolicy->pc_policy.label)) {
return(-1);
}
if (dst2->aio_srcpolicy && dst2->aio_dstpolicy &&
dst2->aio_srcpolicy->pc_policy.label ==
dst2->aio_dstpolicy->pc_policy.label &&
(dst1->aio_srcpolicy == NULL || dst1->aio_dstpolicy == NULL ||
dst1->aio_srcpolicy->pc_policy.label !=
dst1->aio_dstpolicy->pc_policy.label)) {
return(1);
}
#endif
/* Rule 6: Prefer higher precedence. */
#ifdef INET6
if (dst1->aio_dstpolicy &&
(dst2->aio_dstpolicy == NULL ||
dst1->aio_dstpolicy->pc_policy.preced >
dst2->aio_dstpolicy->pc_policy.preced)) {
return(-1);
}
if (dst2->aio_dstpolicy &&
(dst1->aio_dstpolicy == NULL ||
dst2->aio_dstpolicy->pc_policy.preced >
dst1->aio_dstpolicy->pc_policy.preced)) {
return(1);
}
#endif
/* Rule 7: Prefer native transport. */
/* XXX: not implemented yet */
/* Rule 8: Prefer smaller scope. */
if (dst1->aio_dstscope >= 0 &&
dst1->aio_dstscope < dst2->aio_dstscope) {
return(-1);
}
if (dst2->aio_dstscope >= 0 &&
dst2->aio_dstscope < dst1->aio_dstscope) {
return(1);
}
/*
* Rule 9: Use longest matching prefix.
* We compare the match length in a same AF only.
*/
if (dst1->aio_sa.sa_family == dst2->aio_sa.sa_family) {
if (dst1->aio_matchlen > dst2->aio_matchlen) {
return(-1);
}
if (dst1->aio_matchlen < dst2->aio_matchlen) {
return(1);
}
}
/* Rule 10: Otherwise, leave the order unchanged. */
return(-1);
}
/*
* Copy from scope.c.
* XXX: we should standardize the functions and link them as standard
* library.
*/
static int
gai_addr2scopetype(sa)
struct sockaddr *sa;
{
#ifdef INET6
struct sockaddr_in6 *sa6;
#endif
struct sockaddr_in *sa4;
switch(sa->sa_family) {
#ifdef INET6
case AF_INET6:
sa6 = (struct sockaddr_in6 *)sa;
if (IN6_IS_ADDR_MULTICAST(&sa6->sin6_addr)) {
/* just use the scope field of the multicast address */
return(sa6->sin6_addr.s6_addr[2] & 0x0f);
}
/*
* Unicast addresses: map scope type to corresponding scope
* value defined for multcast addresses.
* XXX: hardcoded scope type values are bad...
*/
if (IN6_IS_ADDR_LOOPBACK(&sa6->sin6_addr))
return(1); /* node local scope */
if (IN6_IS_ADDR_LINKLOCAL(&sa6->sin6_addr))
return(2); /* link-local scope */
if (IN6_IS_ADDR_SITELOCAL(&sa6->sin6_addr))
return(5); /* site-local scope */
return(14); /* global scope */
break;
#endif
case AF_INET:
/*
* IPv4 pseudo scoping according to RFC 3484.
*/
sa4 = (struct sockaddr_in *)sa;
/* IPv4 autoconfiguration addresses have link-local scope. */
if (((u_char *)&sa4->sin_addr)[0] == 169 &&
((u_char *)&sa4->sin_addr)[1] == 254)
return(2);
/* Private addresses have site-local scope. */
if (((u_char *)&sa4->sin_addr)[0] == 10 ||
(((u_char *)&sa4->sin_addr)[0] == 172 &&
(((u_char *)&sa4->sin_addr)[1] & 0xf0) == 16) ||
(((u_char *)&sa4->sin_addr)[0] == 192 &&
((u_char *)&sa4->sin_addr)[1] == 168))
return(14); /* XXX: It should be 5 unless NAT */
/* Loopback addresses have link-local scope. */
if (((u_char *)&sa4->sin_addr)[0] == 127)
return(2);
return(14);
break;
default:
errno = EAFNOSUPPORT; /* is this a good error? */
return(-1);
}
}
/*
* FILES (/etc/hosts)
*/
static FILE *
_files_open(int *errp)
{
FILE *fp;
fp = fopen(_PATH_HOSTS, "r");
if (fp == NULL)
*errp = NO_RECOVERY;
return fp;
}
static int
_files_ghbyname(void *rval, void *cb_data, va_list ap)
{
const char *name;
int af;
int *errp;
int match, nalias;
char *p, *line, *addrstr, *cname;
FILE *fp;
struct hostent *rethp, *hp, hpbuf;
char *aliases[MAXALIASES + 1], *addrs[2];
union inx_addr addrbuf;
char buf[BUFSIZ];
int af0;
name = va_arg(ap, const char *);
af = va_arg(ap, int);
errp = va_arg(ap, int *);
*(struct hostent **)rval = NULL;
if ((fp = _files_open(errp)) == NULL)
return NS_UNAVAIL;
rethp = hp = NULL;
af0 = af;
while (fgets(buf, sizeof(buf), fp)) {
line = buf;
if ((addrstr = _hgetword(&line)) == NULL
|| (cname = _hgetword(&line)) == NULL)
continue;
match = (strcasecmp(cname, name) == 0);
nalias = 0;
while ((p = _hgetword(&line)) != NULL) {
if (!match)
match = (strcasecmp(p, name) == 0);
if (nalias < MAXALIASES)
aliases[nalias++] = p;
}
if (!match)
continue;
switch (af0) {
case AF_INET:
if (inet_aton(addrstr, (struct in_addr *)&addrbuf)
!= 1) {
*errp = NO_DATA; /* name found */
continue;
}
af = af0;
break;
#ifdef INET6
case AF_INET6:
if (inet_pton(af, addrstr, &addrbuf) != 1) {
*errp = NO_DATA; /* name found */
continue;
}
af = af0;
break;
#endif
case AF_UNSPEC:
if (inet_aton(addrstr, (struct in_addr *)&addrbuf)
== 1) {
af = AF_INET;
break;
}
#ifdef INET6
if (inet_pton(AF_INET6, addrstr, &addrbuf) == 1) {
af = AF_INET6;
break;
}
#endif
*errp = NO_DATA; /* name found */
continue;
/* NOTREACHED */
}
hp = &hpbuf;
hp->h_name = cname;
hp->h_aliases = aliases;
aliases[nalias] = NULL;
hp->h_addrtype = af;
hp->h_length = ADDRLEN(af);
hp->h_addr_list = addrs;
addrs[0] = (char *)&addrbuf;
addrs[1] = NULL;
hp = _hpcopy(hp, errp);
rethp = _hpmerge(rethp, hp, errp);
}
fclose(fp);
*(struct hostent **)rval = rethp;
return (rethp != NULL) ? NS_SUCCESS : NS_NOTFOUND;
}
static int
_files_ghbyaddr(void *rval, void *cb_data, va_list ap)
{
const void *addr;
int addrlen;
int af;
int *errp;
int nalias;
char *p, *line;
FILE *fp;
struct hostent *hp, hpbuf;
char *aliases[MAXALIASES + 1], *addrs[2];
union inx_addr addrbuf;
char buf[BUFSIZ];
addr = va_arg(ap, const void *);
addrlen = va_arg(ap, int);
af = va_arg(ap, int);
errp = va_arg(ap, int *);
*(struct hostent**)rval = NULL;
if ((fp = _files_open(errp)) == NULL)
return NS_UNAVAIL;
hp = NULL;
while (fgets(buf, sizeof(buf), fp)) {
line = buf;
if ((p = _hgetword(&line)) == NULL
|| (af == AF_INET
? inet_aton(p, (struct in_addr *)&addrbuf)
: inet_pton(af, p, &addrbuf)) != 1
|| memcmp(addr, &addrbuf, addrlen) != 0
|| (p = _hgetword(&line)) == NULL)
continue;
hp = &hpbuf;
hp->h_name = p;
hp->h_aliases = aliases;
nalias = 0;
while ((p = _hgetword(&line)) != NULL) {
if (nalias < MAXALIASES)
aliases[nalias++] = p;
}
aliases[nalias] = NULL;
hp->h_addrtype = af;
hp->h_length = addrlen;
hp->h_addr_list = addrs;
addrs[0] = (char *)&addrbuf;
addrs[1] = NULL;
hp = _hpcopy(hp, errp);
break;
}
fclose(fp);
*(struct hostent **)rval = hp;
return (hp != NULL) ? NS_SUCCESS : NS_NOTFOUND;
}
#ifdef YP
/*
* NIS
*
* XXX actually a hack, these are INET4 specific.
*/
static int
_nis_ghbyname(void *rval, void *cb_data, va_list ap)
{
const char *name;
int af;
int *errp;
struct hostent *hp = NULL;
name = va_arg(ap, const char *);
af = va_arg(ap, int);
errp = va_arg(ap, int *);
if (af == AF_UNSPEC)
af = AF_INET;
if (af == AF_INET) {
THREAD_LOCK();
hp = _gethostbynisname(name, af);
if (hp != NULL)
hp = _hpcopy(hp, errp);
THREAD_UNLOCK();
}
*(struct hostent **)rval = hp;
return (hp != NULL) ? NS_SUCCESS : NS_NOTFOUND;
}
static int
_nis_ghbyaddr(void *rval, void *cb_data, va_list ap)
{
const void *addr;
int addrlen;
int af;
int *errp;
struct hostent *hp = NULL;
addr = va_arg(ap, const void *);
addrlen = va_arg(ap, int);
af = va_arg(ap, int);
if (af == AF_INET) {
THREAD_LOCK();
hp = _gethostbynisaddr(addr, addrlen, af);
if (hp != NULL)
hp = _hpcopy(hp, errp);
THREAD_UNLOCK();
}
*(struct hostent **)rval = hp;
return (hp != NULL) ? NS_SUCCESS : NS_NOTFOUND;
}
#endif
struct __res_type_list {
SLIST_ENTRY(__res_type_list) rtl_entry;
int rtl_type;
};
#define MAXPACKET (64*1024)
typedef union {
HEADER hdr;
u_char buf[MAXPACKET];
} querybuf;
static struct hostent *getanswer(const querybuf *, int, const char *, int,
struct hostent *, int *);
/*
* we don't need to take care about sorting, nor IPv4 mapped address here.
*/
static struct hostent *
getanswer(answer, anslen, qname, qtype, template, errp)
const querybuf *answer;
int anslen;
const char *qname;
int qtype;
struct hostent *template;
int *errp;
{
const HEADER *hp;
const u_char *cp;
int n;
const u_char *eom, *erdata;
char *bp, *ep, **ap, **hap;
int type, class, ancount, qdcount;
int haveanswer, had_error;
char tbuf[MAXDNAME];
const char *tname;
int (*name_ok)(const char *);
static char *h_addr_ptrs[MAXADDRS + 1];
static char *host_aliases[MAXALIASES];
static char hostbuf[8*1024];
#define BOUNDED_INCR(x) \
do { \
cp += x; \
if (cp > eom) { \
*errp = NO_RECOVERY; \
return (NULL); \
} \
} while (0)
#define BOUNDS_CHECK(ptr, count) \
do { \
if ((ptr) + (count) > eom) { \
*errp = NO_RECOVERY; \
return (NULL); \
} \
} while (0)
/* XXX do {} while (0) cannot be put here */
#define DNS_ASSERT(x) \
{ \
if (!(x)) { \
cp += n; \
continue; \
} \
}
/* XXX do {} while (0) cannot be put here */
#define DNS_FATAL(x) \
{ \
if (!(x)) { \
had_error++; \
continue; \
} \
}
tname = qname;
template->h_name = NULL;
eom = answer->buf + anslen;
switch (qtype) {
case T_A:
case T_AAAA:
name_ok = res_hnok;
break;
case T_PTR:
name_ok = res_dnok;
break;
default:
return (NULL); /* XXX should be abort(); */
}
/*
* find first satisfactory answer
*/
hp = &answer->hdr;
ancount = ntohs(hp->ancount);
qdcount = ntohs(hp->qdcount);
bp = hostbuf;
ep = hostbuf + sizeof hostbuf;
cp = answer->buf;
BOUNDED_INCR(HFIXEDSZ);
if (qdcount != 1) {
*errp = NO_RECOVERY;
return (NULL);
}
n = dn_expand(answer->buf, eom, cp, bp, ep - bp);
if ((n < 0) || !(*name_ok)(bp)) {
*errp = NO_RECOVERY;
return (NULL);
}
BOUNDED_INCR(n + QFIXEDSZ);
if (qtype == T_A || qtype == T_AAAA) {
/* res_send() has already verified that the query name is the
* same as the one we sent; this just gets the expanded name
* (i.e., with the succeeding search-domain tacked on).
*/
n = strlen(bp) + 1; /* for the \0 */
if (n >= MAXHOSTNAMELEN) {
*errp = NO_RECOVERY;
return (NULL);
}
template->h_name = bp;
bp += n;
/* The qname can be abbreviated, but h_name is now absolute. */
qname = template->h_name;
}
ap = host_aliases;
*ap = NULL;
template->h_aliases = host_aliases;
hap = h_addr_ptrs;
*hap = NULL;
template->h_addr_list = h_addr_ptrs;
haveanswer = 0;
had_error = 0;
while (ancount-- > 0 && cp < eom && !had_error) {
n = dn_expand(answer->buf, eom, cp, bp, ep - bp);
DNS_FATAL(n >= 0);
DNS_FATAL((*name_ok)(bp));
cp += n; /* name */
BOUNDS_CHECK(cp, 3 * INT16SZ + INT32SZ);
type = _getshort(cp);
cp += INT16SZ; /* type */
class = _getshort(cp);
cp += INT16SZ + INT32SZ; /* class, TTL */
n = _getshort(cp);
cp += INT16SZ; /* len */
BOUNDS_CHECK(cp, n);
erdata = cp + n;
DNS_ASSERT(class == C_IN);
if ((qtype == T_A || qtype == T_AAAA) && type == T_CNAME) {
if (ap >= &host_aliases[MAXALIASES-1])
continue;
n = dn_expand(answer->buf, eom, cp, tbuf, sizeof tbuf);
DNS_FATAL(n >= 0);
DNS_FATAL((*name_ok)(tbuf));
cp += n;
if (cp != erdata) {
*errp = NO_RECOVERY;
return (NULL);
}
/* Store alias. */
*ap++ = bp;
n = strlen(bp) + 1; /* for the \0 */
DNS_FATAL(n < MAXHOSTNAMELEN);
bp += n;
/* Get canonical name. */
n = strlen(tbuf) + 1; /* for the \0 */
DNS_FATAL(n <= ep - bp);
DNS_FATAL(n < MAXHOSTNAMELEN);
strcpy(bp, tbuf);
template->h_name = bp;
bp += n;
continue;
}
if (qtype == T_PTR && type == T_CNAME) {
n = dn_expand(answer->buf, eom, cp, tbuf, sizeof tbuf);
if (n < 0 || !res_dnok(tbuf)) {
had_error++;
continue;
}
cp += n;
if (cp != erdata) {
*errp = NO_RECOVERY;
return (NULL);
}
/* Get canonical name. */
n = strlen(tbuf) + 1; /* for the \0 */
if (n > ep - bp || n >= MAXHOSTNAMELEN) {
had_error++;
continue;
}
strcpy(bp, tbuf);
tname = bp;
bp += n;
continue;
}
DNS_ASSERT(type == qtype);
switch (type) {
case T_PTR:
DNS_ASSERT(strcasecmp(tname, bp) == 0);
n = dn_expand(answer->buf, eom, cp, bp, ep - bp);
DNS_FATAL(n >= 0);
DNS_FATAL(res_hnok(bp));
#if MULTI_PTRS_ARE_ALIASES
cp += n;
if (cp != erdata) {
*errp = NO_RECOVERY;
return (NULL);
}
if (!haveanswer)
template->h_name = bp;
else if (ap < &host_aliases[MAXALIASES-1])
*ap++ = bp;
else
n = -1;
if (n != -1) {
n = strlen(bp) + 1; /* for the \0 */
if (n >= MAXHOSTNAMELEN) {
had_error++;
break;
}
bp += n;
}
break;
#else
template->h_name = bp;
*errp = NETDB_SUCCESS;
return (template);
#endif
case T_A:
case T_AAAA:
DNS_ASSERT(strcasecmp(template->h_name, bp) == 0);
DNS_ASSERT(n == template->h_length);
if (!haveanswer) {
int nn;
template->h_name = bp;
nn = strlen(bp) + 1; /* for the \0 */
bp += nn;
}
bp = (char *)ALIGN(bp);
DNS_FATAL(bp + n < ep);
DNS_ASSERT(hap < &h_addr_ptrs[MAXADDRS-1]);
#ifdef FILTER_V4MAPPED
if (type == T_AAAA) {
struct in6_addr in6;
memcpy(&in6, cp, sizeof(in6));
DNS_ASSERT(IN6_IS_ADDR_V4MAPPED(&in6) == 0);
}
#endif
bcopy(cp, *hap++ = bp, n);
bp += n;
cp += n;
if (cp != erdata) {
*errp = NO_RECOVERY;
return (NULL);
}
break;
default:
abort();
}
if (!had_error)
haveanswer++;
}
if (haveanswer) {
*ap = NULL;
*hap = NULL;
if (!template->h_name) {
n = strlen(qname) + 1; /* for the \0 */
if (n > ep - bp || n >= MAXHOSTNAMELEN)
goto no_recovery;
strcpy(bp, qname);
template->h_name = bp;
bp += n;
}
*errp = NETDB_SUCCESS;
return (template);
}
no_recovery:
*errp = NO_RECOVERY;
return (NULL);
#undef BOUNDED_INCR
#undef BOUNDS_CHECK
#undef DNS_ASSERT
#undef DNS_FATAL
}
/* res_search() variant with multiple query support. */
static struct hostent *
_res_search_multi(name, rtl, errp)
const char *name; /* domain name */
struct __res_type_list *rtl; /* list of query types */
int *errp;
{
const char *cp, * const *domain;
struct hostent *hp0 = NULL, *hp;
struct hostent hpbuf;
u_int dots;
int trailing_dot, ret, saved_herrno;
int got_nodata = 0, got_servfail = 0, tried_as_is = 0;
struct __res_type_list *rtl0 = rtl;
querybuf *buf;
if ((_res.options & RES_INIT) == 0 && res_init() == -1) {
*errp = NETDB_INTERNAL;
return (NULL);
}
dots = 0;
for (cp = name; *cp; cp++)
dots += (*cp == '.');
trailing_dot = 0;
if (cp > name && *--cp == '.')
trailing_dot++;
buf = malloc(sizeof(*buf));
if (buf == NULL) {
*errp = NETDB_INTERNAL;
return NULL;
}
/* If there aren't any dots, it could be a user-level alias */
if (!dots && (cp = hostalias(name)) != NULL) {
for(rtl = rtl0; rtl != NULL;
rtl = SLIST_NEXT(rtl, rtl_entry)) {
ret = res_query(cp, C_IN, rtl->rtl_type, buf->buf,
sizeof(buf->buf));
if (ret > 0 && ret < sizeof(buf->buf)) {
hpbuf.h_addrtype = (rtl->rtl_type == T_AAAA)
? AF_INET6 : AF_INET;
hpbuf.h_length = ADDRLEN(hpbuf.h_addrtype);
hp = getanswer(buf, ret, name, rtl->rtl_type,
&hpbuf, errp);
if (!hp)
continue;
hp = _hpcopy(&hpbuf, errp);
hp0 = _hpmerge(hp0, hp, errp);
} else
*errp = h_errno;
}
free(buf);
return (hp0);
}
/*
* If there are dots in the name already, let's just give it a try
* 'as is'. The threshold can be set with the "ndots" option.
*/
saved_herrno = -1;
if (dots >= _res.ndots) {
for(rtl = rtl0; rtl != NULL;
rtl = SLIST_NEXT(rtl, rtl_entry)) {
ret = res_querydomain(name, NULL, C_IN, rtl->rtl_type,
buf->buf, sizeof(buf->buf));
if (ret > 0 && ret < sizeof(buf->buf)) {
hpbuf.h_addrtype = (rtl->rtl_type == T_AAAA)
? AF_INET6 : AF_INET;
hpbuf.h_length = ADDRLEN(hpbuf.h_addrtype);
hp = getanswer(buf, ret, name, rtl->rtl_type,
&hpbuf, errp);
if (!hp)
continue;
hp = _hpcopy(&hpbuf, errp);
hp0 = _hpmerge(hp0, hp, errp);
} else
*errp = h_errno;
}
if (hp0 != NULL) {
free(buf);
return (hp0);
}
saved_herrno = *errp;
tried_as_is++;
}
/*
* We do at least one level of search if
* - there is no dot and RES_DEFNAME is set, or
* - there is at least one dot, there is no trailing dot,
* and RES_DNSRCH is set.
*/
if ((!dots && (_res.options & RES_DEFNAMES)) ||
(dots && !trailing_dot && (_res.options & RES_DNSRCH))) {
int done = 0;
for (domain = (const char * const *)_res.dnsrch;
*domain && !done;
domain++) {
for(rtl = rtl0; rtl != NULL;
rtl = SLIST_NEXT(rtl, rtl_entry)) {
ret = res_querydomain(name, *domain, C_IN,
rtl->rtl_type,
buf->buf, sizeof(buf->buf));
if (ret > 0 && ret < sizeof(buf->buf)) {
hpbuf.h_addrtype = (rtl->rtl_type == T_AAAA)
? AF_INET6 : AF_INET;
hpbuf.h_length = ADDRLEN(hpbuf.h_addrtype);
hp = getanswer(buf, ret, name,
rtl->rtl_type, &hpbuf, errp);
if (!hp)
continue;
hp = _hpcopy(&hpbuf, errp);
hp0 = _hpmerge(hp0, hp, errp);
} else
*errp = h_errno;
}
if (hp0 != NULL) {
free(buf);
return (hp0);
}
/*
* If no server present, give up.
* If name isn't found in this domain,
* keep trying higher domains in the search list
* (if that's enabled).
* On a NO_DATA error, keep trying, otherwise
* a wildcard entry of another type could keep us
* from finding this entry higher in the domain.
* If we get some other error (negative answer or
* server failure), then stop searching up,
* but try the input name below in case it's
* fully-qualified.
*/
if (errno == ECONNREFUSED) {
free(buf);
*errp = TRY_AGAIN;
return (NULL);
}
switch (*errp) {
case NO_DATA:
got_nodata++;
/* FALLTHROUGH */
case HOST_NOT_FOUND:
/* keep trying */
break;
case TRY_AGAIN:
if (buf->hdr.rcode == SERVFAIL) {
/* try next search element, if any */
got_servfail++;
break;
}
/* FALLTHROUGH */
default:
/* anything else implies that we're done */
done++;
}
/* if we got here for some reason other than DNSRCH,
* we only wanted one iteration of the loop, so stop.
*/
if (!(_res.options & RES_DNSRCH))
done++;
}
}
/*
* If we have not already tried the name "as is", do that now.
* note that we do this regardless of how many dots were in the
* name or whether it ends with a dot unless NOTLDQUERY is set.
*/
if (!tried_as_is && (dots || !(_res.options & RES_NOTLDQUERY))) {
for(rtl = rtl0; rtl != NULL;
rtl = SLIST_NEXT(rtl, rtl_entry)) {
ret = res_querydomain(name, NULL, C_IN, rtl->rtl_type,
buf->buf, sizeof(buf->buf));
if (ret > 0 && ret < sizeof(buf->buf)) {
hpbuf.h_addrtype = (rtl->rtl_type == T_AAAA)
? AF_INET6 : AF_INET;
hpbuf.h_length = ADDRLEN(hpbuf.h_addrtype);
hp = getanswer(buf, ret, name, rtl->rtl_type,
&hpbuf, errp);
if (!hp)
continue;
hp = _hpcopy(&hpbuf, errp);
hp0 = _hpmerge(hp0, hp, errp);
} else
*errp = h_errno;
}
if (hp0 != NULL) {
free(buf);
return (hp0);
}
}
free(buf);
/* if we got here, we didn't satisfy the search.
* if we did an initial full query, return that query's h_errno
* (note that we wouldn't be here if that query had succeeded).
* else if we ever got a nodata, send that back as the reason.
* else send back meaningless h_errno, that being the one from
* the last DNSRCH we did.
*/
if (saved_herrno != -1)
*errp = saved_herrno;
else if (got_nodata)
*errp = NO_DATA;
else if (got_servfail)
*errp = TRY_AGAIN;
return (NULL);
}
static int
_dns_ghbyname(void *rval, void *cb_data, va_list ap)
{
const char *name;
int af;
int *errp;
struct __res_type_list *rtl, rtl4;
#ifdef INET6
struct __res_type_list rtl6;
#endif
name = va_arg(ap, const char *);
af = va_arg(ap, int);
errp = va_arg(ap, int *);
#ifdef INET6
switch (af) {
case AF_UNSPEC:
SLIST_NEXT(&rtl4, rtl_entry) = NULL; rtl4.rtl_type = T_A;
SLIST_NEXT(&rtl6, rtl_entry) = &rtl4; rtl6.rtl_type = T_AAAA;
rtl = &rtl6;
break;
case AF_INET6:
SLIST_NEXT(&rtl6, rtl_entry) = NULL; rtl6.rtl_type = T_AAAA;
rtl = &rtl6;
break;
case AF_INET:
SLIST_NEXT(&rtl4, rtl_entry) = NULL; rtl4.rtl_type = T_A;
rtl = &rtl4;
break;
}
#else
SLIST_NEXT(&rtl4, rtl_entry) = NULL; rtl4.rtl_type = T_A;
rtl = &rtl4;
#endif
*(struct hostent **)rval = _res_search_multi(name, rtl, errp);
if (*(struct hostent **)rval != NULL)
return NS_SUCCESS;
else if (*errp == TRY_AGAIN)
return NS_TRYAGAIN;
else
return NS_NOTFOUND;
}
static int
_dns_ghbyaddr(void *rval, void *cb_data, va_list ap)
{
const void *addr;
int addrlen;
int af;
int *errp;
int n;
int err;
struct hostent *hp;
u_char c, *cp;
char *bp;
struct hostent hbuf;
int na;
#ifdef INET6
static const char hex[] = "0123456789abcdef";
#endif
querybuf *buf;
char qbuf[MAXDNAME+1];
char *hlist[2];
char *tld6[] = { "ip6.arpa", NULL };
char *tld4[] = { "in-addr.arpa", NULL };
char **tld;
addr = va_arg(ap, const void *);
addrlen = va_arg(ap, int);
af = va_arg(ap, int);
errp = va_arg(ap, int *);
*(struct hostent **)rval = NULL;
#ifdef INET6
/* XXX */
if (af == AF_INET6 && IN6_IS_ADDR_LINKLOCAL((struct in6_addr *)addr))
return NS_NOTFOUND;
#endif
switch (af) {
#ifdef INET6
case AF_INET6:
tld = tld6;
break;
#endif
case AF_INET:
tld = tld4;
break;
default:
return NS_NOTFOUND;
}
if ((_res.options & RES_INIT) == 0) {
if (res_init() < 0) {
*errp = h_errno;
return NS_UNAVAIL;
}
}
memset(&hbuf, 0, sizeof(hbuf));
hbuf.h_name = NULL;
hbuf.h_addrtype = af;
hbuf.h_length = addrlen;
na = 0;
buf = malloc(sizeof(*buf));
if (buf == NULL) {
*errp = NETDB_INTERNAL;
return NS_UNAVAIL;
}
err = NS_SUCCESS;
for (/* nothing */; *tld; tld++) {
/*
* XXX assumes that MAXDNAME is big enough - error checks
* has been made by callers
*/
n = 0;
bp = qbuf;
cp = (u_char *)addr+addrlen-1;
switch (af) {
#ifdef INET6
case AF_INET6:
for (; n < addrlen; n++, cp--) {
c = *cp;
*bp++ = hex[c & 0xf];
*bp++ = '.';
*bp++ = hex[c >> 4];
*bp++ = '.';
}
strcpy(bp, *tld);
break;
#endif
case AF_INET:
for (; n < addrlen; n++, cp--) {
c = *cp;
if (c >= 100)
*bp++ = '0' + c / 100;
if (c >= 10)
*bp++ = '0' + (c % 100) / 10;
*bp++ = '0' + c % 10;
*bp++ = '.';
}
strcpy(bp, *tld);
break;
}
n = res_query(qbuf, C_IN, T_PTR, buf->buf, sizeof buf->buf);
if (n < 0) {
*errp = h_errno;
err = NS_UNAVAIL;
continue;
} else if (n > sizeof(buf->buf)) {
#if 0
errno = ERANGE; /* XXX is it OK to set errno here? */
#endif
*errp = NETDB_INTERNAL;
err = NS_UNAVAIL;
continue;
}
hp = getanswer(buf, n, qbuf, T_PTR, &hbuf, errp);
if (!hp) {
err = NS_NOTFOUND;
continue;
}
free(buf);
hbuf.h_addrtype = af;
hbuf.h_length = addrlen;
hbuf.h_addr_list = hlist;
hlist[0] = (char *)addr;
hlist[1] = NULL;
*(struct hostent **)rval = _hpcopy(&hbuf, errp);
return NS_SUCCESS;
}
free(buf);
return err;
}
static void
_dns_shent(int stayopen)
{
if ((_res.options & RES_INIT) == 0) {
if (res_init() < 0)
return;
}
if (stayopen)
_res.options |= RES_STAYOPEN | RES_USEVC;
}
static void
_dns_ehent(void)
{
_res.options &= ~(RES_STAYOPEN | RES_USEVC);
res_close();
}
#ifdef ICMPNL
/*
* experimental:
* draft-ietf-ipngwg-icmp-namelookups-02.txt
* ifindex is assumed to be encoded in addr.
*/
#include <sys/uio.h>
#include <netinet/ip6.h>
#include <netinet/icmp6.h>
struct _icmp_host_cache {
struct _icmp_host_cache *hc_next;
int hc_ifindex;
struct in6_addr hc_addr;
char *hc_name;
};
static char *
_icmp_fqdn_query(const struct in6_addr *addr, int ifindex)
{
int s;
struct icmp6_filter filter;
struct msghdr msg;
struct cmsghdr *cmsg;
struct in6_pktinfo *pkt;
char cbuf[256];
char buf[1024];
int cc;
struct icmp6_fqdn_query *fq;
struct icmp6_fqdn_reply *fr;
struct _icmp_host_cache *hc;
struct sockaddr_in6 sin6;
struct iovec iov;
fd_set s_fds, fds;
struct timeval tout;
int len;
char *name;
static struct _icmp_host_cache *hc_head;
THREAD_LOCK();
for (hc = hc_head; hc; hc = hc->hc_next) {
if (hc->hc_ifindex == ifindex
&& IN6_ARE_ADDR_EQUAL(&hc->hc_addr, addr)) {
THREAD_UNLOCK();
return hc->hc_name; /* XXX: never freed */
}
}
ICMP6_FILTER_SETBLOCKALL(&filter);
ICMP6_FILTER_SETPASS(ICMP6_FQDN_REPLY, &filter);
FD_ZERO(&s_fds);
tout.tv_sec = 0;
tout.tv_usec = 200000; /*XXX: 200ms*/
fq = (struct icmp6_fqdn_query *)buf;
fq->icmp6_fqdn_type = ICMP6_FQDN_QUERY;
fq->icmp6_fqdn_code = 0;
fq->icmp6_fqdn_cksum = 0;
fq->icmp6_fqdn_id = (u_short)getpid();
fq->icmp6_fqdn_unused = 0;
fq->icmp6_fqdn_cookie[0] = 0;
fq->icmp6_fqdn_cookie[1] = 0;
memset(&sin6, 0, sizeof(sin6));
sin6.sin6_family = AF_INET6;
sin6.sin6_addr = *addr;
memset(&msg, 0, sizeof(msg));
msg.msg_name = (caddr_t)&sin6;
msg.msg_namelen = sizeof(sin6);
msg.msg_iov = &iov;
msg.msg_iovlen = 1;
msg.msg_control = NULL;
msg.msg_controllen = 0;
iov.iov_base = (caddr_t)buf;
iov.iov_len = sizeof(struct icmp6_fqdn_query);
if (ifindex) {
msg.msg_control = cbuf;
msg.msg_controllen = sizeof(cbuf);
cmsg = CMSG_FIRSTHDR(&msg);
cmsg->cmsg_len = CMSG_LEN(sizeof(struct in6_pktinfo));
cmsg->cmsg_level = IPPROTO_IPV6;
cmsg->cmsg_type = IPV6_PKTINFO;
pkt = (struct in6_pktinfo *)&cmsg[1];
memset(&pkt->ipi6_addr, 0, sizeof(struct in6_addr));
pkt->ipi6_ifindex = ifindex;
cmsg = CMSG_NXTHDR(&msg, cmsg);
msg.msg_controllen = (char *)cmsg - cbuf;
}
if ((s = _socket(PF_INET6, SOCK_RAW, IPPROTO_ICMPV6)) < 0)
return NULL;
(void)_setsockopt(s, IPPROTO_ICMPV6, ICMP6_FILTER,
(char *)&filter, sizeof(filter));
cc = _sendmsg(s, &msg, 0);
if (cc < 0) {
_close(s);
return NULL;
}
FD_SET(s, &s_fds);
for (;;) {
fds = s_fds;
if (_select(s + 1, &fds, NULL, NULL, &tout) <= 0) {
_close(s);
return NULL;
}
len = sizeof(sin6);
cc = _recvfrom(s, buf, sizeof(buf), 0,
(struct sockaddr *)&sin6, &len);
if (cc <= 0) {
_close(s);
return NULL;
}
if (cc < sizeof(struct ip6_hdr) + sizeof(struct icmp6_hdr))
continue;
if (!IN6_ARE_ADDR_EQUAL(addr, &sin6.sin6_addr))
continue;
fr = (struct icmp6_fqdn_reply *)(buf + sizeof(struct ip6_hdr));
if (fr->icmp6_fqdn_type == ICMP6_FQDN_REPLY)
break;
}
_close(s);
if (fr->icmp6_fqdn_cookie[1] != 0) {
/* rfc1788 type */
name = buf + sizeof(struct ip6_hdr) + sizeof(struct icmp6_hdr) + 4;
len = (buf + cc) - name;
} else {
len = fr->icmp6_fqdn_namelen;
name = fr->icmp6_fqdn_name;
}
if (len <= 0)
return NULL;
name[len] = 0;
if ((hc = (struct _icmp_host_cache *)malloc(sizeof(*hc))) == NULL)
return NULL;
/* XXX: limit number of cached entries */
hc->hc_ifindex = ifindex;
hc->hc_addr = *addr;
hc->hc_name = strdup(name);
THREAD_LOCK();
hc->hc_next = hc_head;
hc_head = hc;
THREAD_UNLOCK();
return hc->hc_name;
}
static struct hostent *
_icmp_ghbyaddr(const void *addr, int addrlen, int af, int *errp)
{
char *hname;
int ifindex;
struct in6_addr addr6;
if (af != AF_INET6) {
/*
* Note: rfc1788 defines Who Are You for IPv4,
* but no one implements it.
*/
return NULL;
}
memcpy(&addr6, addr, addrlen);
ifindex = (addr6.s6_addr[2] << 8) | addr6.s6_addr[3];
addr6.s6_addr[2] = addr6.s6_addr[3] = 0;
if (!IN6_IS_ADDR_LINKLOCAL(&addr6))
return NULL; /*XXX*/
if ((hname = _icmp_fqdn_query(&addr6, ifindex)) == NULL)
return NULL;
return _hpaddr(af, hname, &addr6, errp);
}
#endif /* ICMPNL */