nfsmapid_resolv.c revision 7c478bd95313f5f23a4c958a745db2134aa03244
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License, Version 1.0 only
* (the "License"). You may not use this file except in compliance
* with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright 2005 Sun Microsystems, Inc. All rights reserved.
* Use is subject to license terms.
*/
#pragma ident "%Z%%M% %I% %E% SMI"
/*
* PSARC/2004/154 nfsmapid DNS enhancements implementation.
*
* As per RFC 3050, file owner and group attributes in version 4 of the
* NFS protocol are no longer exchanged between client and server as 32
* bit integral values. Instead, owner and group file attributes are
* exchanged between client and server as UTF8 strings of form
*
* 'user@domain' (ie. "joeblow@central.sun.com")
* 'group@domain' (ie. "staff@central.sun.com")
*
* being able to describe a user with a unique string identifier provides
* a much more powerful and administrative friendly way of dealing with
* with issues of correctly mapping user and group ownership in a cross-
* domain environment has proven a difficult problem to solve, since
* dealing with different permutations of client naming configurations
* (ie. NIS only, LDAP only, etc.) have bloated the problem. Thus, users
* utilizing clients and servers that have the 'domain' portion of the
* UTF8 attribute string configured differently than its peer server and
* client accordingly, will experience watching their files owned by the
* 'nobody' user and group. This is due to the fact that the 'domain's
* don't match and the nfsmapid daemon treats the attribute strings as
* in the executing daemon's system). Please refer to PSARC/2004/154 for
* further background and motivation for these enhancements.
*
* The latest implementation of the nfsmapid daemon relies on a DNS TXT
* record. The behavior of nfsmapid is to first use the NFSMAPID_DOMAIN
* set, then the nfsmapid daemon queries the configured DNS domain server
* for the _nfsv4idmapdomain TXT record. If the record exists, then the
* record's value is used as the 'domain' portion of the UTF8 attribute
* strings. If the TXT record has not been configured in the DNS server,
* then the daemon falls back to using the DNS domain name itself as the
* 'domain' portion of the attribute strings. Lastly, if the configured
* DNS server is unresponsive, the nfsmapid daemon falls back to using
* the DNS domain name as the 'domain' portion of the attribute strings,
* and fires up a query thread to keep contacting the DNS server until
* it responds with either a TXT record, or a lack thereof, in which
* case, nfsmapid just continues to utilize the DNS domain name.
*/
#define __NFSMAPID_RES_IMPL
#include "nfsmapid_resolv.h"
/*
* DEBUG Only
* Decode any resolver errors and print out message to log
*/
static int
resolv_error(void)
{
switch (h_errno) {
case NETDB_INTERNAL:
break;
case HOST_NOT_FOUND:
(void) rw_rdlock(&s_dns_impl_lock);
#ifdef DEBUG
#endif
(void) rw_unlock(&s_dns_impl_lock);
break;
case TRY_AGAIN:
/*
* Nameserver is not responding.
* Try again after a given timeout.
*/
(void) rw_rdlock(&s_dns_impl_lock);
#ifdef DEBUG
#endif
(void) rw_unlock(&s_dns_impl_lock);
break;
case NO_RECOVERY:
/*
* This msg only really happens once, due
* to s_dns_disabled flag (see below)
*/
break;
case NO_DATA:
/*
* No entries in the nameserver for
* the specific record or record type.
*/
(void) rw_rdlock(&s_dns_impl_lock);
#ifdef DEBUG
s_dname);
#endif
(void) rw_unlock(&s_dns_impl_lock);
break;
case NETDB_SUCCESS:
default:
break;
}
return (h_errno);
}
/*
* Reset the global state variables used for the TXT record.
* Having these values reset to zero helps nfsmapid confirm
* that a valid DNS TXT record was not found; in which case,
* it would fall back to using the configured DNS domain name.
*
* If a valid DNS TXT record _was_ found, but subsequent contact
* to the DNS server is somehow hindered, the previous DNS TXT
* RR value continues to be used. Thus, in such instances, we
* forego clearing the global config variables so nfsmapid can
* continue to use a valid DNS TXT RR while contact to the DNS
* server is reestablished.
*/
static void
resolv_txt_reset(void)
{
(void) rw_wrlock(&s_dns_impl_lock);
(void) rw_unlock(&s_dns_impl_lock);
(void) rw_wrlock(&dns_data_lock);
if (!dns_txt_cached) {
dns_txt_domain_len = 0;
}
(void) rw_unlock(&dns_data_lock);
}
/*
* Initialize resolver and populate &s_res struct
*
* DNS Domain is saved off sysdns_domain in case we
* need to fall back to using the DNS domain name as
* the v4 attribute string domain.
*/
int
resolv_init(void)
{
int n;
struct __res_state res;
(void) mutex_lock(&s_res_lock);
(void) mutex_unlock(&s_res_lock);
(void) resolv_error();
return (n);
}
(void) mutex_unlock(&s_res_lock);
(void) rw_wrlock(&s_dns_impl_lock);
(void) rw_unlock(&s_dns_impl_lock);
(void) rw_wrlock(&dns_data_lock);
(void) rw_unlock(&dns_data_lock);
return (0);
}
/*
* Search criteria assumptions:
*
* The onus will fall on the sysadmins to correctly configure the TXT
* record in the DNS domain where the box currently resides in order
* for the record to be found. However, if they sysadmin chooses to
* add the 'search' key to /etc/resolv.conf, then resolv_search()
* _will_ traverse up the DNS tree as specified in the 'search' key.
* Otherwise, we'll default the domain to the DNS domain itself.
*/
static int
resolv_search(void)
{
int len;
struct __res_state res;
(void) mutex_lock(&s_res_lock);
(void) mutex_unlock(&s_res_lock);
/*
* Avoid holding locks across the res_nsearch() call to
* prevent stalling threads during network partitions.
*/
return (resolv_error());
(void) rw_wrlock(&s_dns_impl_lock);
(void) rw_unlock(&s_dns_impl_lock);
return (NETDB_SUCCESS);
}
/*
* Skip one DNS record
*/
static uchar_t *
{
int t;
int dlen;
/*
* Skip compressed name
*/
errno = 0;
if ((t = dn_skipname(p, eom)) < 0) {
return (NULL);
}
/*
* Advance pointer and make sure
* we're still within the message
*/
p += t;
return (NULL);
/*
* Now, just skip over the rr fields
*/
p += INT16SZ; /* type */
p += INT16SZ; /* class */
p += INT32SZ; /* ttl */
p += dlen; /* dlen */
if (p > eom)
return (NULL);
return (p);
}
/*
* Process one TXT record.
*
* nfsmapid queries the DNS server for the specific _nfsv4idmapdomain
* TXT record. Thus, if the TXT record exists, the answer section of
* the DNS response carries the TXT record's value. Thus, we check that
* the value is indeed a valid domain and set the modular s_txt_rr
* global to the domain value.
*/
static void
{
char *rr_base = (char *)(p + 1);
char tmp_txt_rr[DNAMEMAX];
return; /* process next TXT RR */
/*
* make sure we have a clean buf since
* we may've processed several TXT rr's
*/
(void) rw_wrlock(&s_dns_impl_lock);
(void) rw_unlock(&s_dns_impl_lock);
/*
* If there is a record and it's a valid domain, we're done.
*/
(void) rw_wrlock(&s_dns_impl_lock);
(void) rw_unlock(&s_dns_impl_lock);
} else if (!(msg_done++ % NFSMAPID_SLOG_RATE)) {
/*
* Otherwise, log the error
*/
(void) rw_rdlock(&s_dns_impl_lock);
(void) rw_unlock(&s_dns_impl_lock);
}
}
/*
* Decode any answer received from the DNS server. This interface is
* capable of much more than just decoding TXT records. We maintain
* focus on TXT rr's for now, but this will probably change once we
* get the IETF approved application specific DNS RR.
*
* Here's an example of the TXT record we're decoding (as would appear
* in the DNS zone file):
*
* _nfsv4idmapdomain IN TXT "sun.com"
*
* Once the IETF application specific DNS RR is granted, we should only
* be changing the record flavor, but all should pretty much stay the
* same.
*/
static void
resolv_decode(void)
{
uchar_t *p;
int n;
int dlen;
int answer_len = 0;
/*
* Check the HEADER for any signs of errors
* and extract the answer counts for later.
*/
(void) rw_rdlock(&s_dns_impl_lock);
(void) rw_unlock(&s_dns_impl_lock);
return;
}
/*
* skip query entries
*/
errno = 0;
while (qd_cnt-- > 0) {
n = dn_skipname(p, eom);
if (n < 0) {
return;
}
p += n;
p += INT16SZ; /* type */
p += INT16SZ; /* class */
}
#ifdef DEBUG
/*
* If debugging... print query only once.
* NOTE: Don't advance pointer... this is done
* in while() loop on a per record basis !
*/
if (n < 0) {
(void) resolv_error();
return;
}
#endif
/*
* Process actual answer(s).
*/
/* skip the name field */
if (n < 0) {
(void) resolv_error();
return;
}
p += n;
return;
return;
switch (type) {
case T_TXT:
resolve_process_txt(p, dlen);
break;
default:
/*
* Advance to next answer record for any
* other record types. Again, this will
* probably change (see block comment).
*/
p += dlen;
break;
}
}
/*
* Skip name server and additional records for now.
*/
if (cnt > 0) {
p = resolv_skip_rr(p, eom);
if (p == NULL)
return;
}
}
}
/*
* If a valid TXT record entry exists, s_txt_rr contains the domain
* value (as set in resolv_process_txt) and we extract the value into
* dns_txt_domain (the exported global). If there was _no_ valid TXT
* entry, we simply return and check_domain() will default to the
* DNS domain since we did resolv_txt_reset() first.
*/
static void
{
(void) rw_rdlock(&s_dns_impl_lock);
if (s_txt_rr[0] != '\0') {
(void) rw_wrlock(&dns_data_lock);
s_txt_rr);
dns_txt_cached = 1;
(void) rw_unlock(&dns_data_lock);
}
(void) rw_unlock(&s_dns_impl_lock);
}
/*
* Thread to keep pinging DNS server for TXT record if nfsmapid's
* initial attempt at contact with server failed. We could potentially
* have a substantial number of NFSv4 clients and having all of them
* hammering on an already unresponsive DNS server would not help
* things. So, we limit the number of live query threads to at most
* 1 at any one time to keep things from getting out of hand.
*/
/* ARGSUSED */
void *
resolv_query_thread(void *arg)
{
#ifdef DEBUG
char *whoami = "query_thread";
#endif
(void) rw_rdlock(&s_dns_impl_lock);
(void) rw_unlock(&s_dns_impl_lock);
for (;;) {
(void) resolv_init();
switch (resolv_search()) {
case NETDB_SUCCESS:
/*
* This is a bit different than what we
* do in get_dns_txt_domain(). Here, the
* thread _must_ update the global state
* if a new TXT record was found.
*/
(void) rw_rdlock(&dns_data_lock);
if (dns_txt_domain_len != 0) {
/*
* Update global state and only
* flush the cache if there were
* any updates to cur_domain
*/
(void) rw_wrlock(&domain_cfg_lock);
(void) strncpy(cur_domain,
DNAMEMAX-1);
(void) rw_unlock(&domain_cfg_lock);
idmap_kcall(-1);
}
(void) rw_unlock(&dns_data_lock);
goto thr_okay;
case NO_DATA:
/*
* DNS is up now, but does not have
* the NFSV4IDMAPDOMAIN TXT record.
*/
NULL);
goto thr_reset;
case NO_RECOVERY:
/*
* Non-Recoverable error occurred. No sense
* in keep pinging the DNS server at this
* point, so we disable any further contact.
*/
(void) rw_wrlock(&s_dns_impl_lock);
(void) rw_unlock(&s_dns_impl_lock);
goto thr_reset;
case HOST_NOT_FOUND:
/*
* Authoritative NS not responding...
* keep trying for non-authoritative reply
*/
/*FALLTHROUGH*/
case TRY_AGAIN:
/* keep trying */
break;
case NETDB_INTERNAL:
default:
IDMAP_DBG("%s: Internal resolver error: %s",
goto thr_reset;
}
}
(void) rw_wrlock(&dns_data_lock);
dns_txt_cached = 0;
(void) rw_unlock(&dns_data_lock);
/* mark thread as done */
(void) rw_wrlock(&s_dns_impl_lock);
(void) rw_unlock(&s_dns_impl_lock);
/*NOTREACHED*/
return (NULL);
}
/*
* nfsmapid's interface into the resolver for getting the TXT record.
*
* Key concepts:
*
* o If the DNS server is available and the TXT record is found, we
* simply decode the output and fill the exported dns_txt_domain
* global, so our caller can configure the daemon appropriately.
*
* o If the TXT record is not found, then having done resolv_txt_reset()
* first will allow our caller to recognize that the exported globals
* are empty and thus configure nfsmapid to use the default DNS domain.
*
* o Having no /etc/resolv.conf file is pretty much a show stopper, since
* there is no name server address information. We return since we've
* already have reset the TXT global state.
*
* o If a previous call to the DNS server resulted in an unrecoverable
* error, then we disable further contact to the DNS server and return.
* Having the TXT global state already reset guarantees that our caller
* will fall back to the right configuration.
*
* o Query thread creation is throttled by s_dns_qthr_created. We mitigate
* the problem of an already unresponsive DNS server by allowing at most
* 1 outstanding query thread since we could potentially have a substantial
* amount of clients hammering on the same DNS server attempting to get
* the TXT record.
*/
void
get_dns_txt_domain(int sighup)
{
int err;
#ifdef DEBUG
char *whoami = "get_dns_txt_domain";
#endif
long thr_flags = THR_DETACHED;
/*
* We reset TXT variables first in case /etc/resolv.conf
* is missing or we've had unrecoverable resolver errors,
* we'll default to get_dns_domain(). If a previous DNS
* TXT RR was found, don't clear it until we're certain
* that contact can be made to the DNS server (see block
* comment atop resolv_txt_reset). If we're responding to
* a SIGHUP signal, force a reset of the cached copy.
*/
if (sighup) {
(void) rw_wrlock(&dns_data_lock);
dns_txt_cached = 0;
(void) rw_unlock(&dns_data_lock);
}
errno = 0;
/*
* If /etc/resolv.conf is not there, then we'll
* get the domain from domainname(1M). No real
* reason to query DNS or fire a thread since we
* have no nameserver addresses.
*/
goto txtclear;
}
(void) rw_rdlock(&s_dns_impl_lock);
if (s_dns_disabled) {
/*
* If there were non-recoverable problems with DNS,
* we have stopped querying DNS entirely. See
* NO_RECOVERY clause below.
*/
(void) rw_unlock(&s_dns_impl_lock);
return;
}
(void) rw_unlock(&s_dns_impl_lock);
(void) resolv_init();
switch (resolv_search()) {
case NETDB_SUCCESS:
/*
* If there _is_ a TXT record, we let
* our caller set the global state.
*/
return;
case TRY_AGAIN:
(void) rw_wrlock(&s_dns_impl_lock);
if (s_dns_qthr_created) {
/*
* We may have lots of clients, so we don't
* want to bog down the DNS server with tons
* of requests... lest it becomes even more
* unresponsive, so limit 1 thread to query
* DNS at a time.
*/
IDMAP_DBG("%s: query thread already active",
(void) rw_unlock(&s_dns_impl_lock);
return;
}
/*
* DNS did not respond ! Set timeout and kick off
* thread to try op again after s_dns_tout seconds.
* We've made sure that we don't have an already
* running thread above.
*/
if (!err) {
}
#ifdef DEBUG
else {
msg_done++;
if (!(msg_done % NFSMAPID_SLOG_RATE))
NULL);
}
#endif
(void) rw_unlock(&s_dns_impl_lock);
return;
case NO_RECOVERY:
(void) rw_wrlock(&s_dns_impl_lock);
(void) rw_unlock(&s_dns_impl_lock);
/*FALLTHROUGH*/
default:
/*
* For any other errors... DNS is responding, but
* either it has no data, or some other problem is
* occuring. At any rate, the TXT domain should not
* be used, so we default to the DNS domain.
*/
break;
}
(void) rw_wrlock(&dns_data_lock);
dns_txt_cached = 0;
(void) rw_unlock(&dns_data_lock);
}