Git-Mirrors/monero
1
0
Fork 0
mirror of https://github.com/monero-project/monero.git synced 2024-12-30 15:06:14 -05:00
Code Issues Packages Projects Releases Wiki Activity
92f2f687b9
monero/external/unbound/daemon/worker.c
Erik de Castro Lopo a85b5759f3 Upgrade unbound library 
These files were pulled from the 1.6.3 release tarball.

This new version builds against OpenSSL version 1.1 which will be
the default in the new Debian Stable which is due to be released
RealSoonNow (tm).
2017-06-17 23:04:00 +10:00

1884 lines
60 KiB
C
Raw Blame History

/*
* daemon/worker.c - worker that handles a pending list of requests.
*
* Copyright (c) 2007, NLnet Labs. All rights reserved.
*
* This software is open source.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
*
* Redistributions of source code must retain the above copyright notice,
* this list of conditions and the following disclaimer.
*
* 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.
*
* Neither the name of the NLNET LABS 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 COPYRIGHT HOLDERS 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 COPYRIGHT
* HOLDER 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.
*/
/**
* \file
*
* This file implements the worker that handles callbacks on events, for
* pending requests.
*/
#include "config.h"
#include "util/log.h"
#include "util/net_help.h"
#include "util/random.h"
#include "daemon/worker.h"
#include "daemon/daemon.h"
#include "daemon/remote.h"
#include "daemon/acl_list.h"
#include "util/netevent.h"
#include "util/config_file.h"
#include "util/module.h"
#include "util/regional.h"
#include "util/storage/slabhash.h"
#include "services/listen_dnsport.h"
#include "services/outside_network.h"
#include "services/outbound_list.h"
#include "services/cache/rrset.h"
#include "services/cache/infra.h"
#include "services/cache/dns.h"
#include "services/mesh.h"
#include "services/localzone.h"
#include "util/data/msgparse.h"
#include "util/data/msgencode.h"
#include "util/data/dname.h"
#include "util/fptr_wlist.h"
#include "util/tube.h"
#include "iterator/iter_fwd.h"
#include "iterator/iter_hints.h"
#include "validator/autotrust.h"
#include "validator/val_anchor.h"
#include "respip/respip.h"
#include "libunbound/context.h"
#include "libunbound/libworker.h"
#include "sldns/sbuffer.h"
#include "sldns/wire2str.h"
#include "util/shm_side/shm_main.h"
#include "dnscrypt/dnscrypt.h"
#ifdef HAVE_SYS_TYPES_H
# include <sys/types.h>
#endif
#ifdef HAVE_NETDB_H
#include <netdb.h>
#endif
#include <signal.h>
#ifdef UB_ON_WINDOWS
#include "winrc/win_svc.h"
#endif
/** Size of an UDP datagram */
#define NORMAL_UDP_SIZE 512 /* bytes */
/** ratelimit for error responses */
#define ERROR_RATELIMIT 100 /* qps */
/**
* seconds to add to prefetch leeway. This is a TTL that expires old rrsets
* earlier than they should in order to put the new update into the cache.
* This additional value is to make sure that if not all TTLs are equal in
* the message to be updated(and replaced), that rrsets with up to this much
* extra TTL are also replaced. This means that the resulting new message
* will have (most likely) this TTL at least, avoiding very small 'split
* second' TTLs due to operators choosing relative primes for TTLs (or so).
* Also has to be at least one to break ties (and overwrite cached entry).
*/
#define PREFETCH_EXPIRY_ADD 60
/** Report on memory usage by this thread and global */
static void
worker_mem_report(struct worker* ATTR_UNUSED(worker),
struct serviced_query* ATTR_UNUSED(cur_serv))
{
#ifdef UNBOUND_ALLOC_STATS
/* measure memory leakage */
extern size_t unbound_mem_alloc, unbound_mem_freed;
/* debug func in validator module */
size_t total, front, back, mesh, msg, rrset, infra, ac, superac;
size_t me, iter, val, anch;
int i;
#ifdef CLIENT_SUBNET
size_t subnet = 0;
#endif /* CLIENT_SUBNET */
if(verbosity < VERB_ALGO)
return;
front = listen_get_mem(worker->front);
back = outnet_get_mem(worker->back);
msg = slabhash_get_mem(worker->env.msg_cache);
rrset = slabhash_get_mem(&worker->env.rrset_cache->table);
infra = infra_get_mem(worker->env.infra_cache);
mesh = mesh_get_mem(worker->env.mesh);
ac = alloc_get_mem(&worker->alloc);
superac = alloc_get_mem(&worker->daemon->superalloc);
anch = anchors_get_mem(worker->env.anchors);
iter = 0;
val = 0;
for(i=0; i<worker->env.mesh->mods.num; i++) {
fptr_ok(fptr_whitelist_mod_get_mem(worker->env.mesh->
mods.mod[i]->get_mem));
if(strcmp(worker->env.mesh->mods.mod[i]->name, "validator")==0)
val += (*worker->env.mesh->mods.mod[i]->get_mem)
(&worker->env, i);
#ifdef CLIENT_SUBNET
else if(strcmp(worker->env.mesh->mods.mod[i]->name,
"subnet")==0)
subnet += (*worker->env.mesh->mods.mod[i]->get_mem)
(&worker->env, i);
#endif /* CLIENT_SUBNET */
else iter += (*worker->env.mesh->mods.mod[i]->get_mem)
(&worker->env, i);
}
me = sizeof(*worker) + sizeof(*worker->base) + sizeof(*worker->comsig)
+ comm_point_get_mem(worker->cmd_com)
+ sizeof(worker->rndstate)
+ regional_get_mem(worker->scratchpad)
+ sizeof(*worker->env.scratch_buffer)
+ sldns_buffer_capacity(worker->env.scratch_buffer)
+ forwards_get_mem(worker->env.fwds)
+ hints_get_mem(worker->env.hints);
if(worker->thread_num == 0)
me += acl_list_get_mem(worker->daemon->acl);
if(cur_serv) {
me += serviced_get_mem(cur_serv);
}
total = front+back+mesh+msg+rrset+infra+iter+val+ac+superac+me;
#ifdef CLIENT_SUBNET
total += subnet;
log_info("Memory conditions: %u front=%u back=%u mesh=%u msg=%u "
"rrset=%u infra=%u iter=%u val=%u subnet=%u anchors=%u "
"alloccache=%u globalalloccache=%u me=%u",
(unsigned)total, (unsigned)front, (unsigned)back,
(unsigned)mesh, (unsigned)msg, (unsigned)rrset, (unsigned)infra,
(unsigned)iter, (unsigned)val,
(unsigned)subnet, (unsigned)anch, (unsigned)ac,
(unsigned)superac, (unsigned)me);
#else /* no CLIENT_SUBNET */
log_info("Memory conditions: %u front=%u back=%u mesh=%u msg=%u "
"rrset=%u infra=%u iter=%u val=%u anchors=%u "
"alloccache=%u globalalloccache=%u me=%u",
(unsigned)total, (unsigned)front, (unsigned)back,
(unsigned)mesh, (unsigned)msg, (unsigned)rrset,
(unsigned)infra, (unsigned)iter, (unsigned)val, (unsigned)anch,
(unsigned)ac, (unsigned)superac, (unsigned)me);
#endif /* CLIENT_SUBNET */
log_info("Total heap memory estimate: %u total-alloc: %u "
"total-free: %u", (unsigned)total,
(unsigned)unbound_mem_alloc, (unsigned)unbound_mem_freed);
#else /* no UNBOUND_ALLOC_STATS */
size_t val = 0;
#ifdef CLIENT_SUBNET
size_t subnet = 0;
#endif /* CLIENT_SUBNET */
int i;
if(verbosity < VERB_QUERY)
return;
for(i=0; i<worker->env.mesh->mods.num; i++) {
fptr_ok(fptr_whitelist_mod_get_mem(worker->env.mesh->
mods.mod[i]->get_mem));
if(strcmp(worker->env.mesh->mods.mod[i]->name, "validator")==0)
val += (*worker->env.mesh->mods.mod[i]->get_mem)
(&worker->env, i);
#ifdef CLIENT_SUBNET
else if(strcmp(worker->env.mesh->mods.mod[i]->name,
"subnet")==0)
subnet += (*worker->env.mesh->mods.mod[i]->get_mem)
(&worker->env, i);
#endif /* CLIENT_SUBNET */
}
#ifdef CLIENT_SUBNET
verbose(VERB_QUERY, "cache memory msg=%u rrset=%u infra=%u val=%u "
"subnet=%u",
(unsigned)slabhash_get_mem(worker->env.msg_cache),
(unsigned)slabhash_get_mem(&worker->env.rrset_cache->table),
(unsigned)infra_get_mem(worker->env.infra_cache),
(unsigned)val, (unsigned)subnet);
#else /* no CLIENT_SUBNET */
verbose(VERB_QUERY, "cache memory msg=%u rrset=%u infra=%u val=%u",
(unsigned)slabhash_get_mem(worker->env.msg_cache),
(unsigned)slabhash_get_mem(&worker->env.rrset_cache->table),
(unsigned)infra_get_mem(worker->env.infra_cache),
(unsigned)val);
#endif /* CLIENT_SUBNET */
#endif /* UNBOUND_ALLOC_STATS */
}
void
worker_send_cmd(struct worker* worker, enum worker_commands cmd)
{
uint32_t c = (uint32_t)htonl(cmd);
if(!tube_write_msg(worker->cmd, (uint8_t*)&c, sizeof(c), 0)) {
log_err("worker send cmd %d failed", (int)cmd);
}
}
int
worker_handle_reply(struct comm_point* c, void* arg, int error,
struct comm_reply* reply_info)
{
struct module_qstate* q = (struct module_qstate*)arg;
struct worker* worker = q->env->worker;
struct outbound_entry e;
e.qstate = q;
e.qsent = NULL;
if(error != 0) {
mesh_report_reply(worker->env.mesh, &e, reply_info, error);
worker_mem_report(worker, NULL);
return 0;
}
/* sanity check. */
if(!LDNS_QR_WIRE(sldns_buffer_begin(c->buffer))
|| LDNS_OPCODE_WIRE(sldns_buffer_begin(c->buffer)) !=
LDNS_PACKET_QUERY
|| LDNS_QDCOUNT(sldns_buffer_begin(c->buffer)) > 1) {
/* error becomes timeout for the module as if this reply
* never arrived. */
mesh_report_reply(worker->env.mesh, &e, reply_info,
NETEVENT_TIMEOUT);
worker_mem_report(worker, NULL);
return 0;
}
mesh_report_reply(worker->env.mesh, &e, reply_info, NETEVENT_NOERROR);
worker_mem_report(worker, NULL);
return 0;
}
int
worker_handle_service_reply(struct comm_point* c, void* arg, int error,
struct comm_reply* reply_info)
{
struct outbound_entry* e = (struct outbound_entry*)arg;
struct worker* worker = e->qstate->env->worker;
struct serviced_query *sq = e->qsent;
verbose(VERB_ALGO, "worker svcd callback for qstate %p", e->qstate);
if(error != 0) {
mesh_report_reply(worker->env.mesh, e, reply_info, error);
worker_mem_report(worker, sq);
return 0;
}
/* sanity check. */
if(!LDNS_QR_WIRE(sldns_buffer_begin(c->buffer))
|| LDNS_OPCODE_WIRE(sldns_buffer_begin(c->buffer)) !=
LDNS_PACKET_QUERY
|| LDNS_QDCOUNT(sldns_buffer_begin(c->buffer)) > 1) {
/* error becomes timeout for the module as if this reply
* never arrived. */
verbose(VERB_ALGO, "worker: bad reply handled as timeout");
mesh_report_reply(worker->env.mesh, e, reply_info,
NETEVENT_TIMEOUT);
worker_mem_report(worker, sq);
return 0;
}
mesh_report_reply(worker->env.mesh, e, reply_info, NETEVENT_NOERROR);
worker_mem_report(worker, sq);
return 0;
}
/** ratelimit error replies
* @param worker: the worker struct with ratelimit counter
* @param err: error code that would be wanted.
* @return value of err if okay, or -1 if it should be discarded instead.
*/
static int
worker_err_ratelimit(struct worker* worker, int err)
{
if(worker->err_limit_time == *worker->env.now) {
/* see if limit is exceeded for this second */
if(worker->err_limit_count++ > ERROR_RATELIMIT)
return -1;
} else {
/* new second, new limits */
worker->err_limit_time = *worker->env.now;
worker->err_limit_count = 1;
}
return err;
}
/** check request sanity.
* @param pkt: the wire packet to examine for sanity.
* @param worker: parameters for checking.
* @return error code, 0 OK, or -1 discard.
*/
static int
worker_check_request(sldns_buffer* pkt, struct worker* worker)
{
if(sldns_buffer_limit(pkt) < LDNS_HEADER_SIZE) {
verbose(VERB_QUERY, "request too short, discarded");
return -1;
}
if(sldns_buffer_limit(pkt) > NORMAL_UDP_SIZE &&
worker->daemon->cfg->harden_large_queries) {
verbose(VERB_QUERY, "request too large, discarded");
return -1;
}
if(LDNS_QR_WIRE(sldns_buffer_begin(pkt))) {
verbose(VERB_QUERY, "request has QR bit on, discarded");
return -1;
}
if(LDNS_TC_WIRE(sldns_buffer_begin(pkt))) {
LDNS_TC_CLR(sldns_buffer_begin(pkt));
verbose(VERB_QUERY, "request bad, has TC bit on");
return worker_err_ratelimit(worker, LDNS_RCODE_FORMERR);
}
if(LDNS_OPCODE_WIRE(sldns_buffer_begin(pkt)) != LDNS_PACKET_QUERY) {
verbose(VERB_QUERY, "request unknown opcode %d",
LDNS_OPCODE_WIRE(sldns_buffer_begin(pkt)));
return worker_err_ratelimit(worker, LDNS_RCODE_NOTIMPL);
}
if(LDNS_QDCOUNT(sldns_buffer_begin(pkt)) != 1) {
verbose(VERB_QUERY, "request wrong nr qd=%d",
LDNS_QDCOUNT(sldns_buffer_begin(pkt)));
return worker_err_ratelimit(worker, LDNS_RCODE_FORMERR);
}
if(LDNS_ANCOUNT(sldns_buffer_begin(pkt)) != 0) {
verbose(VERB_QUERY, "request wrong nr an=%d",
LDNS_ANCOUNT(sldns_buffer_begin(pkt)));
return worker_err_ratelimit(worker, LDNS_RCODE_FORMERR);
}
if(LDNS_NSCOUNT(sldns_buffer_begin(pkt)) != 0) {
verbose(VERB_QUERY, "request wrong nr ns=%d",
LDNS_NSCOUNT(sldns_buffer_begin(pkt)));
return worker_err_ratelimit(worker, LDNS_RCODE_FORMERR);
}
if(LDNS_ARCOUNT(sldns_buffer_begin(pkt)) > 1) {
verbose(VERB_QUERY, "request wrong nr ar=%d",
LDNS_ARCOUNT(sldns_buffer_begin(pkt)));
return worker_err_ratelimit(worker, LDNS_RCODE_FORMERR);
}
return 0;
}
void
worker_handle_control_cmd(struct tube* ATTR_UNUSED(tube), uint8_t* msg,
size_t len, int error, void* arg)
{
struct worker* worker = (struct worker*)arg;
enum worker_commands cmd;
if(error != NETEVENT_NOERROR) {
free(msg);
if(error == NETEVENT_CLOSED)
comm_base_exit(worker->base);
else log_info("control event: %d", error);
return;
}
if(len != sizeof(uint32_t)) {
fatal_exit("bad control msg length %d", (int)len);
}
cmd = sldns_read_uint32(msg);
free(msg);
switch(cmd) {
case worker_cmd_quit:
verbose(VERB_ALGO, "got control cmd quit");
comm_base_exit(worker->base);
break;
case worker_cmd_stats:
verbose(VERB_ALGO, "got control cmd stats");
server_stats_reply(worker, 1);
break;
case worker_cmd_stats_noreset:
verbose(VERB_ALGO, "got control cmd stats_noreset");
server_stats_reply(worker, 0);
break;
case worker_cmd_remote:
verbose(VERB_ALGO, "got control cmd remote");
daemon_remote_exec(worker);
break;
default:
log_err("bad command %d", (int)cmd);
break;
}
}
/** check if a delegation is secure */
static enum sec_status
check_delegation_secure(struct reply_info *rep)
{
/* return smallest security status */
size_t i;
enum sec_status sec = sec_status_secure;
enum sec_status s;
size_t num = rep->an_numrrsets + rep->ns_numrrsets;
/* check if answer and authority are OK */
for(i=0; i<num; i++) {
s = ((struct packed_rrset_data*)rep->rrsets[i]->entry.data)
->security;
if(s < sec)
sec = s;
}
/* in additional, only unchecked triggers revalidation */
for(i=num; i<rep->rrset_count; i++) {
s = ((struct packed_rrset_data*)rep->rrsets[i]->entry.data)
->security;
if(s == sec_status_unchecked)
return s;
}
return sec;
}
/** remove nonsecure from a delegation referral additional section */
static void
deleg_remove_nonsecure_additional(struct reply_info* rep)
{
/* we can simply edit it, since we are working in the scratch region */
size_t i;
enum sec_status s;
for(i = rep->an_numrrsets+rep->ns_numrrsets; i<rep->rrset_count; i++) {
s = ((struct packed_rrset_data*)rep->rrsets[i]->entry.data)
->security;
if(s != sec_status_secure) {
memmove(rep->rrsets+i, rep->rrsets+i+1,
sizeof(struct ub_packed_rrset_key*)*
(rep->rrset_count - i - 1));
rep->ar_numrrsets--;
rep->rrset_count--;
i--;
}
}
}
/** answer nonrecursive query from the cache */
static int
answer_norec_from_cache(struct worker* worker, struct query_info* qinfo,
uint16_t id, uint16_t flags, struct comm_reply* repinfo,
struct edns_data* edns)
{
/* for a nonrecursive query return either:
* o an error (servfail; we try to avoid this)
* o a delegation (closest we have; this routine tries that)
* o the answer (checked by answer_from_cache)
*
* So, grab a delegation from the rrset cache.
* Then check if it needs validation, if so, this routine fails,
* so that iterator can prime and validator can verify rrsets.
*/
uint16_t udpsize = edns->udp_size;
int secure = 0;
time_t timenow = *worker->env.now;
int must_validate = (!(flags&BIT_CD) || worker->env.cfg->ignore_cd)
&& worker->env.need_to_validate;
struct dns_msg *msg = NULL;
struct delegpt *dp;
dp = dns_cache_find_delegation(&worker->env, qinfo->qname,
qinfo->qname_len, qinfo->qtype, qinfo->qclass,
worker->scratchpad, &msg, timenow);
if(!dp) { /* no delegation, need to reprime */
return 0;
}
/* In case we have a local alias, copy it into the delegation message.
* Shallow copy should be fine, as we'll be done with msg in this
* function. */
msg->qinfo.local_alias = qinfo->local_alias;
if(must_validate) {
switch(check_delegation_secure(msg->rep)) {
case sec_status_unchecked:
/* some rrsets have not been verified yet, go and
* let validator do that */
return 0;
case sec_status_bogus:
/* some rrsets are bogus, reply servfail */
edns->edns_version = EDNS_ADVERTISED_VERSION;
edns->udp_size = EDNS_ADVERTISED_SIZE;
edns->ext_rcode = 0;
edns->bits &= EDNS_DO;
if(!inplace_cb_reply_servfail_call(&worker->env, qinfo, NULL,
msg->rep, LDNS_RCODE_SERVFAIL, edns, worker->scratchpad))
return 0;
error_encode(repinfo->c->buffer, LDNS_RCODE_SERVFAIL,
&msg->qinfo, id, flags, edns);
if(worker->stats.extended) {
worker->stats.ans_bogus++;
worker->stats.ans_rcode[LDNS_RCODE_SERVFAIL]++;
}
return 1;
case sec_status_secure:
/* all rrsets are secure */
/* remove non-secure rrsets from the add. section*/
if(worker->env.cfg->val_clean_additional)
deleg_remove_nonsecure_additional(msg->rep);
secure = 1;
break;
case sec_status_indeterminate:
case sec_status_insecure:
default:
/* not secure */
secure = 0;
break;
}
}
/* return this delegation from the cache */
edns->edns_version = EDNS_ADVERTISED_VERSION;
edns->udp_size = EDNS_ADVERTISED_SIZE;
edns->ext_rcode = 0;
edns->bits &= EDNS_DO;
if(!inplace_cb_reply_cache_call(&worker->env, qinfo, NULL, msg->rep,
(int)(flags&LDNS_RCODE_MASK), edns, worker->scratchpad))
return 0;
msg->rep->flags |= BIT_QR|BIT_RA;
if(!reply_info_answer_encode(&msg->qinfo, msg->rep, id, flags,
repinfo->c->buffer, 0, 1, worker->scratchpad,
udpsize, edns, (int)(edns->bits & EDNS_DO), secure)) {
if(!inplace_cb_reply_servfail_call(&worker->env, qinfo, NULL, NULL,
LDNS_RCODE_SERVFAIL, edns, worker->scratchpad))
edns->opt_list = NULL;
error_encode(repinfo->c->buffer, LDNS_RCODE_SERVFAIL,
&msg->qinfo, id, flags, edns);
}
if(worker->stats.extended) {
if(secure) worker->stats.ans_secure++;
server_stats_insrcode(&worker->stats, repinfo->c->buffer);
}
return 1;
}
/** Apply, if applicable, a response IP action to a cached answer.
* If the answer is rewritten as a result of an action, '*encode_repp' will
* point to the reply info containing the modified answer. '*encode_repp' will
* be intact otherwise.
* It returns 1 on success, 0 otherwise. */
static int
apply_respip_action(struct worker* worker, const struct query_info* qinfo,
struct respip_client_info* cinfo, struct reply_info* rep,
struct comm_reply* repinfo, struct ub_packed_rrset_key** alias_rrset,
struct reply_info** encode_repp)
{
struct respip_action_info actinfo = {respip_none, NULL};
if(qinfo->qtype != LDNS_RR_TYPE_A &&
qinfo->qtype != LDNS_RR_TYPE_AAAA &&
qinfo->qtype != LDNS_RR_TYPE_ANY)
return 1;
if(!respip_rewrite_reply(qinfo, cinfo, rep, encode_repp, &actinfo,
alias_rrset, 0, worker->scratchpad))
return 0;
/* xxx_deny actions mean dropping the reply, unless the original reply
* was redirected to response-ip data. */
if((actinfo.action == respip_deny ||
actinfo.action == respip_inform_deny) &&
*encode_repp == rep)
*encode_repp = NULL;
/* If address info is returned, it means the action should be an
* 'inform' variant and the information should be logged. */
if(actinfo.addrinfo) {
respip_inform_print(actinfo.addrinfo, qinfo->qname,
qinfo->qtype, qinfo->qclass, qinfo->local_alias,
repinfo);
}
return 1;
}
/** answer query from the cache.
* Normally, the answer message will be built in repinfo->c->buffer; if the
* answer is supposed to be suppressed or the answer is supposed to be an
* incomplete CNAME chain, the buffer is explicitly cleared to signal the
* caller as such. In the latter case *partial_rep will point to the incomplete
* reply, and this function is (possibly) supposed to be called again with that
* *partial_rep value to complete the chain. In addition, if the query should
* be completely dropped, '*need_drop' will be set to 1. */
static int
answer_from_cache(struct worker* worker, struct query_info* qinfo,
struct respip_client_info* cinfo, int* need_drop,
struct ub_packed_rrset_key** alias_rrset,
struct reply_info** partial_repp,
struct reply_info* rep, uint16_t id, uint16_t flags,
struct comm_reply* repinfo, struct edns_data* edns)
{
time_t timenow = *worker->env.now;
uint16_t udpsize = edns->udp_size;
struct reply_info* encode_rep = rep;
struct reply_info* partial_rep = *partial_repp;
int secure;
int must_validate = (!(flags&BIT_CD) || worker->env.cfg->ignore_cd)
&& worker->env.need_to_validate;
*partial_repp = NULL; /* avoid accidental further pass */
if(worker->env.cfg->serve_expired) {
/* always lock rrsets, rep->ttl is ignored */
if(!rrset_array_lock(rep->ref, rep->rrset_count, 0))
return 0;
/* below, rrsets with ttl before timenow become TTL 0 in
* the response */
/* This response was served with zero TTL */
if (timenow >= rep->ttl) {
worker->stats.zero_ttl_responses++;
}
} else {
/* see if it is possible */
if(rep->ttl < timenow) {
/* the rrsets may have been updated in the meantime.
* we will refetch the message format from the
* authoritative server
*/
return 0;
}
if(!rrset_array_lock(rep->ref, rep->rrset_count, timenow))
return 0;
/* locked and ids and ttls are OK. */
}
/* check CNAME chain (if any) */
if(rep->an_numrrsets > 0 && (rep->rrsets[0]->rk.type ==
htons(LDNS_RR_TYPE_CNAME) || rep->rrsets[0]->rk.type ==
htons(LDNS_RR_TYPE_DNAME))) {
if(!reply_check_cname_chain(qinfo, rep)) {
/* cname chain invalid, redo iterator steps */
verbose(VERB_ALGO, "Cache reply: cname chain broken");
bail_out:
rrset_array_unlock_touch(worker->env.rrset_cache,
worker->scratchpad, rep->ref, rep->rrset_count);
return 0;
}
}
/* check security status of the cached answer */
if( rep->security == sec_status_bogus && must_validate) {
/* BAD cached */
edns->edns_version = EDNS_ADVERTISED_VERSION;
edns->udp_size = EDNS_ADVERTISED_SIZE;
edns->ext_rcode = 0;
edns->bits &= EDNS_DO;
if(!inplace_cb_reply_servfail_call(&worker->env, qinfo, NULL, rep,
LDNS_RCODE_SERVFAIL, edns, worker->scratchpad))
goto bail_out;
error_encode(repinfo->c->buffer, LDNS_RCODE_SERVFAIL,
qinfo, id, flags, edns);
rrset_array_unlock_touch(worker->env.rrset_cache,
worker->scratchpad, rep->ref, rep->rrset_count);
if(worker->stats.extended) {
worker->stats.ans_bogus ++;
worker->stats.ans_rcode[LDNS_RCODE_SERVFAIL] ++;
}
return 1;
} else if( rep->security == sec_status_unchecked && must_validate) {
verbose(VERB_ALGO, "Cache reply: unchecked entry needs "
"validation");
goto bail_out; /* need to validate cache entry first */
} else if(rep->security == sec_status_secure) {
if(reply_all_rrsets_secure(rep))
secure = 1;
else {
if(must_validate) {
verbose(VERB_ALGO, "Cache reply: secure entry"
" changed status");
goto bail_out; /* rrset changed, re-verify */
}
secure = 0;
}
} else secure = 0;
edns->edns_version = EDNS_ADVERTISED_VERSION;
edns->udp_size = EDNS_ADVERTISED_SIZE;
edns->ext_rcode = 0;
edns->bits &= EDNS_DO;
if(!inplace_cb_reply_cache_call(&worker->env, qinfo, NULL, rep,
(int)(flags&LDNS_RCODE_MASK), edns, worker->scratchpad))
goto bail_out;
*alias_rrset = NULL; /* avoid confusion if caller set it to non-NULL */
if(worker->daemon->use_response_ip && !partial_rep &&
!apply_respip_action(worker, qinfo, cinfo, rep, repinfo, alias_rrset,
&encode_rep)) {
goto bail_out;
} else if(partial_rep &&
!respip_merge_cname(partial_rep, qinfo, rep, cinfo,
must_validate, &encode_rep, worker->scratchpad)) {
goto bail_out;
}
if(encode_rep != rep)
secure = 0; /* if rewritten, it can't be considered "secure" */
if(!encode_rep || *alias_rrset) {
sldns_buffer_clear(repinfo->c->buffer);
sldns_buffer_flip(repinfo->c->buffer);
if(!encode_rep)
*need_drop = 1;
else {
/* If a partial CNAME chain is found, we first need to
* make a copy of the reply in the scratchpad so we
* can release the locks and lookup the cache again. */
*partial_repp = reply_info_copy(encode_rep, NULL,
worker->scratchpad);
if(!*partial_repp)
goto bail_out;
}
} else if(!reply_info_answer_encode(qinfo, encode_rep, id, flags,
repinfo->c->buffer, timenow, 1, worker->scratchpad,
udpsize, edns, (int)(edns->bits & EDNS_DO), secure)) {
if(!inplace_cb_reply_servfail_call(&worker->env, qinfo, NULL, NULL,
LDNS_RCODE_SERVFAIL, edns, worker->scratchpad))
edns->opt_list = NULL;
error_encode(repinfo->c->buffer, LDNS_RCODE_SERVFAIL,
qinfo, id, flags, edns);
}
/* cannot send the reply right now, because blocking network syscall
* is bad while holding locks. */
rrset_array_unlock_touch(worker->env.rrset_cache, worker->scratchpad,
rep->ref, rep->rrset_count);
if(worker->stats.extended) {
if(secure) worker->stats.ans_secure++;
server_stats_insrcode(&worker->stats, repinfo->c->buffer);
}
/* go and return this buffer to the client */
return 1;
}
/** Reply to client and perform prefetch to keep cache up to date.
* If the buffer for the reply is empty, it indicates that only prefetch is
* necessary and the reply should be suppressed (because it's dropped or
* being deferred). */
static void
reply_and_prefetch(struct worker* worker, struct query_info* qinfo,
uint16_t flags, struct comm_reply* repinfo, time_t leeway)
{
/* first send answer to client to keep its latency
* as small as a cachereply */
if(sldns_buffer_limit(repinfo->c->buffer) != 0)
comm_point_send_reply(repinfo);
server_stats_prefetch(&worker->stats, worker);
/* create the prefetch in the mesh as a normal lookup without
* client addrs waiting, which has the cache blacklisted (to bypass
* the cache and go to the network for the data). */
/* this (potentially) runs the mesh for the new query */
mesh_new_prefetch(worker->env.mesh, qinfo, flags, leeway +
PREFETCH_EXPIRY_ADD);
}
/**
* Fill CH class answer into buffer. Keeps query.
* @param pkt: buffer
* @param str: string to put into text record (<255).
* array of strings, every string becomes a text record.
* @param num: number of strings in array.
* @param edns: edns reply information.
* @param worker: worker with scratch region.
*/
static void
chaos_replystr(sldns_buffer* pkt, char** str, int num, struct edns_data* edns,
struct worker* worker)
{
int i;
unsigned int rd = LDNS_RD_WIRE(sldns_buffer_begin(pkt));
unsigned int cd = LDNS_CD_WIRE(sldns_buffer_begin(pkt));
sldns_buffer_clear(pkt);
sldns_buffer_skip(pkt, (ssize_t)sizeof(uint16_t)); /* skip id */
sldns_buffer_write_u16(pkt, (uint16_t)(BIT_QR|BIT_RA));
if(rd) LDNS_RD_SET(sldns_buffer_begin(pkt));
if(cd) LDNS_CD_SET(sldns_buffer_begin(pkt));
sldns_buffer_write_u16(pkt, 1); /* qdcount */
sldns_buffer_write_u16(pkt, (uint16_t)num); /* ancount */
sldns_buffer_write_u16(pkt, 0); /* nscount */
sldns_buffer_write_u16(pkt, 0); /* arcount */
(void)query_dname_len(pkt); /* skip qname */
sldns_buffer_skip(pkt, (ssize_t)sizeof(uint16_t)); /* skip qtype */
sldns_buffer_skip(pkt, (ssize_t)sizeof(uint16_t)); /* skip qclass */
for(i=0; i<num; i++) {
size_t len = strlen(str[i]);
if(len>255) len=255; /* cap size of TXT record */
sldns_buffer_write_u16(pkt, 0xc00c); /* compr ptr to query */
sldns_buffer_write_u16(pkt, LDNS_RR_TYPE_TXT);
sldns_buffer_write_u16(pkt, LDNS_RR_CLASS_CH);
sldns_buffer_write_u32(pkt, 0); /* TTL */
sldns_buffer_write_u16(pkt, sizeof(uint8_t) + len);
sldns_buffer_write_u8(pkt, len);
sldns_buffer_write(pkt, str[i], len);
}
sldns_buffer_flip(pkt);
edns->edns_version = EDNS_ADVERTISED_VERSION;
edns->udp_size = EDNS_ADVERTISED_SIZE;
edns->bits &= EDNS_DO;
if(!inplace_cb_reply_local_call(&worker->env, NULL, NULL, NULL,
LDNS_RCODE_NOERROR, edns, worker->scratchpad))
edns->opt_list = NULL;
attach_edns_record(pkt, edns);
}
/** Reply with one string */
static void
chaos_replyonestr(sldns_buffer* pkt, const char* str, struct edns_data* edns,
struct worker* worker)
{
chaos_replystr(pkt, (char**)&str, 1, edns, worker);
}
/**
* Create CH class trustanchor answer.
* @param pkt: buffer
* @param edns: edns reply information.
* @param w: worker with scratch region.
*/
static void
chaos_trustanchor(sldns_buffer* pkt, struct edns_data* edns, struct worker* w)
{
#define TA_RESPONSE_MAX_TXT 16 /* max number of TXT records */
#define TA_RESPONSE_MAX_TAGS 32 /* max number of tags printed per zone */
char* str_array[TA_RESPONSE_MAX_TXT];
uint16_t tags[TA_RESPONSE_MAX_TAGS];
int num = 0;
struct trust_anchor* ta;
if(!w->env.need_to_validate) {
/* no validator module, reply no trustanchors */
chaos_replystr(pkt, NULL, 0, edns, w);
return;
}
/* fill the string with contents */
lock_basic_lock(&w->env.anchors->lock);
RBTREE_FOR(ta, struct trust_anchor*, w->env.anchors->tree) {
char* str;
size_t i, numtag, str_len = 255;
if(num == TA_RESPONSE_MAX_TXT) continue;
str = (char*)regional_alloc(w->scratchpad, str_len);
if(!str) continue;
lock_basic_lock(&ta->lock);
numtag = anchor_list_keytags(ta, tags, TA_RESPONSE_MAX_TAGS);
if(numtag == 0) {
/* empty, insecure point */
lock_basic_unlock(&ta->lock);
continue;
}
str_array[num] = str;
num++;
/* spool name of anchor */
(void)sldns_wire2str_dname_buf(ta->name, ta->namelen, str, str_len);
str_len -= strlen(str); str += strlen(str);
/* spool tags */
for(i=0; i<numtag; i++) {
snprintf(str, str_len, " %u", (unsigned)tags[i]);
str_len -= strlen(str); str += strlen(str);
}
lock_basic_unlock(&ta->lock);
}
lock_basic_unlock(&w->env.anchors->lock);
chaos_replystr(pkt, str_array, num, edns, w);
regional_free_all(w->scratchpad);
}
/**
* Answer CH class queries.
* @param w: worker
* @param qinfo: query info. Pointer into packet buffer.
* @param edns: edns info from query.
* @param pkt: packet buffer.
* @return: true if a reply is to be sent.
*/
static int
answer_chaos(struct worker* w, struct query_info* qinfo,
struct edns_data* edns, sldns_buffer* pkt)
{
struct config_file* cfg = w->env.cfg;
if(qinfo->qtype != LDNS_RR_TYPE_ANY && qinfo->qtype != LDNS_RR_TYPE_TXT)
return 0;
if(query_dname_compare(qinfo->qname,
(uint8_t*)"\002id\006server") == 0 ||
query_dname_compare(qinfo->qname,
(uint8_t*)"\010hostname\004bind") == 0)
{
if(cfg->hide_identity)
return 0;
if(cfg->identity==NULL || cfg->identity[0]==0) {
char buf[MAXHOSTNAMELEN+1];
if (gethostname(buf, MAXHOSTNAMELEN) == 0) {
buf[MAXHOSTNAMELEN] = 0;
chaos_replyonestr(pkt, buf, edns, w);
} else {
log_err("gethostname: %s", strerror(errno));
chaos_replyonestr(pkt, "no hostname", edns, w);
}
}
else chaos_replyonestr(pkt, cfg->identity, edns, w);
return 1;
}
if(query_dname_compare(qinfo->qname,
(uint8_t*)"\007version\006server") == 0 ||
query_dname_compare(qinfo->qname,
(uint8_t*)"\007version\004bind") == 0)
{
if(cfg->hide_version)
return 0;
if(cfg->version==NULL || cfg->version[0]==0)
chaos_replyonestr(pkt, PACKAGE_STRING, edns, w);
else chaos_replyonestr(pkt, cfg->version, edns, w);
return 1;
}
if(query_dname_compare(qinfo->qname,
(uint8_t*)"\013trustanchor\007unbound") == 0)
{
if(cfg->hide_trustanchor)
return 0;
chaos_trustanchor(pkt, edns, w);
return 1;
}
return 0;
}
static int
deny_refuse(struct comm_point* c, enum acl_access acl,
enum acl_access deny, enum acl_access refuse,
struct worker* worker, struct comm_reply* repinfo)
{
if(acl == deny) {
comm_point_drop_reply(repinfo);
if(worker->stats.extended)
worker->stats.unwanted_queries++;
return 0;
} else if(acl == refuse) {
log_addr(VERB_ALGO, "refused query from",
&repinfo->addr, repinfo->addrlen);
log_buf(VERB_ALGO, "refuse", c->buffer);
if(worker->stats.extended)
worker->stats.unwanted_queries++;
if(worker_check_request(c->buffer, worker) == -1) {
comm_point_drop_reply(repinfo);
return 0; /* discard this */
}
sldns_buffer_set_limit(c->buffer, LDNS_HEADER_SIZE);
sldns_buffer_write_at(c->buffer, 4,
(uint8_t*)"\0\0\0\0\0\0\0\0", 8);
LDNS_QR_SET(sldns_buffer_begin(c->buffer));
LDNS_RCODE_SET(sldns_buffer_begin(c->buffer),
LDNS_RCODE_REFUSED);
sldns_buffer_set_position(c->buffer, LDNS_HEADER_SIZE);
sldns_buffer_flip(c->buffer);
return 1;
}
return -1;
}
static int
deny_refuse_all(struct comm_point* c, enum acl_access acl,
struct worker* worker, struct comm_reply* repinfo)
{
return deny_refuse(c, acl, acl_deny, acl_refuse, worker, repinfo);
}
static int
deny_refuse_non_local(struct comm_point* c, enum acl_access acl,
struct worker* worker, struct comm_reply* repinfo)
{
return deny_refuse(c, acl, acl_deny_non_local, acl_refuse_non_local, worker, repinfo);
}
int
worker_handle_request(struct comm_point* c, void* arg, int error,
struct comm_reply* repinfo)
{
struct worker* worker = (struct worker*)arg;
int ret;
hashvalue_type h;
struct lruhash_entry* e;
struct query_info qinfo;
struct edns_data edns;
enum acl_access acl;
struct acl_addr* acladdr;
int rc = 0;
int need_drop = 0;
/* We might have to chase a CNAME chain internally, in which case
* we'll have up to two replies and combine them to build a complete
* answer. These variables control this case. */
struct ub_packed_rrset_key* alias_rrset = NULL;
struct reply_info* partial_rep = NULL;
struct query_info* lookup_qinfo = &qinfo;
struct query_info qinfo_tmp; /* placeholdoer for lookup_qinfo */
struct respip_client_info* cinfo = NULL, cinfo_tmp;
if(error != NETEVENT_NOERROR) {
/* some bad tcp query DNS formats give these error calls */
verbose(VERB_ALGO, "handle request called with err=%d", error);
return 0;
}
#ifdef USE_DNSCRYPT
repinfo->max_udp_size = worker->daemon->cfg->max_udp_size;
if(!dnsc_handle_curved_request(worker->daemon->dnscenv, repinfo)) {
worker->stats.num_query_dnscrypt_crypted_malformed++;
return 0;
}
if(c->dnscrypt && !repinfo->is_dnscrypted) {
char buf[LDNS_MAX_DOMAINLEN+1];
// Check if this is unencrypted and asking for certs
if(worker_check_request(c->buffer, worker) != 0) {
verbose(VERB_ALGO, "dnscrypt: worker check request: bad query.");
log_addr(VERB_CLIENT,"from",&repinfo->addr, repinfo->addrlen);
comm_point_drop_reply(repinfo);
return 0;
}
if(!query_info_parse(&qinfo, c->buffer)) {
verbose(VERB_ALGO, "dnscrypt: worker parse request: formerror.");
log_addr(VERB_CLIENT,"from",&repinfo->addr, repinfo->addrlen);
comm_point_drop_reply(repinfo);
return 0;
}
dname_str(qinfo.qname, buf);
if(!(qinfo.qtype == LDNS_RR_TYPE_TXT &&
strcasecmp(buf, worker->daemon->dnscenv->provider_name) == 0)) {
verbose(VERB_ALGO,
"dnscrypt: not TXT %s. Receive: %s %s",
worker->daemon->dnscenv->provider_name,
sldns_rr_descript(qinfo.qtype)->_name,
buf);
comm_point_drop_reply(repinfo);
worker->stats.num_query_dnscrypt_cleartext++;
return 0;
}
worker->stats.num_query_dnscrypt_cert++;
sldns_buffer_rewind(c->buffer);
} else if(c->dnscrypt && repinfo->is_dnscrypted) {
worker->stats.num_query_dnscrypt_crypted++;
}
#endif
#ifdef USE_DNSTAP
if(worker->dtenv.log_client_query_messages)
dt_msg_send_client_query(&worker->dtenv, &repinfo->addr, c->type,
c->buffer);
#endif
acladdr = acl_addr_lookup(worker->daemon->acl, &repinfo->addr,
repinfo->addrlen);
acl = acl_get_control(acladdr);
if((ret=deny_refuse_all(c, acl, worker, repinfo)) != -1)
{
if(ret == 1)
goto send_reply;
return ret;
}
if((ret=worker_check_request(c->buffer, worker)) != 0) {
verbose(VERB_ALGO, "worker check request: bad query.");
log_addr(VERB_CLIENT,"from",&repinfo->addr, repinfo->addrlen);
if(ret != -1) {
LDNS_QR_SET(sldns_buffer_begin(c->buffer));
LDNS_RCODE_SET(sldns_buffer_begin(c->buffer), ret);
return 1;
}
comm_point_drop_reply(repinfo);
return 0;
}
worker->stats.num_queries++;
/* check if this query should be dropped based on source ip rate limiting */
if(!infra_ip_ratelimit_inc(worker->env.infra_cache, repinfo,
*worker->env.now)) {
/* See if we are passed through with slip factor */
if(worker->env.cfg->ip_ratelimit_factor != 0 &&
ub_random_max(worker->env.rnd,
worker->env.cfg->ip_ratelimit_factor) == 1) {
char addrbuf[128];
addr_to_str(&repinfo->addr, repinfo->addrlen,
addrbuf, sizeof(addrbuf));
verbose(VERB_OPS, "ip_ratelimit allowed through for ip address %s ",
addrbuf);
} else {
worker->stats.num_queries_ip_ratelimited++;
comm_point_drop_reply(repinfo);
return 0;
}
}
/* see if query is in the cache */
if(!query_info_parse(&qinfo, c->buffer)) {
verbose(VERB_ALGO, "worker parse request: formerror.");
log_addr(VERB_CLIENT,"from",&repinfo->addr, repinfo->addrlen);
if(worker_err_ratelimit(worker, LDNS_RCODE_FORMERR) == -1) {
comm_point_drop_reply(repinfo);
return 0;
}
sldns_buffer_rewind(c->buffer);
LDNS_QR_SET(sldns_buffer_begin(c->buffer));
LDNS_RCODE_SET(sldns_buffer_begin(c->buffer),
LDNS_RCODE_FORMERR);
server_stats_insrcode(&worker->stats, c->buffer);
goto send_reply;
}
if(worker->env.cfg->log_queries) {
char ip[128];
addr_to_str(&repinfo->addr, repinfo->addrlen, ip, sizeof(ip));
log_nametypeclass(0, ip, qinfo.qname, qinfo.qtype, qinfo.qclass);
}
if(qinfo.qtype == LDNS_RR_TYPE_AXFR ||
qinfo.qtype == LDNS_RR_TYPE_IXFR) {
verbose(VERB_ALGO, "worker request: refused zone transfer.");
log_addr(VERB_CLIENT,"from",&repinfo->addr, repinfo->addrlen);
sldns_buffer_rewind(c->buffer);
LDNS_QR_SET(sldns_buffer_begin(c->buffer));
LDNS_RCODE_SET(sldns_buffer_begin(c->buffer),
LDNS_RCODE_REFUSED);
if(worker->stats.extended) {
worker->stats.qtype[qinfo.qtype]++;
server_stats_insrcode(&worker->stats, c->buffer);
}
goto send_reply;
}
if(qinfo.qtype == LDNS_RR_TYPE_OPT ||
qinfo.qtype == LDNS_RR_TYPE_TSIG ||
qinfo.qtype == LDNS_RR_TYPE_TKEY ||
qinfo.qtype == LDNS_RR_TYPE_MAILA ||
qinfo.qtype == LDNS_RR_TYPE_MAILB ||
(qinfo.qtype >= 128 && qinfo.qtype <= 248)) {
verbose(VERB_ALGO, "worker request: formerror for meta-type.");
log_addr(VERB_CLIENT,"from",&repinfo->addr, repinfo->addrlen);
if(worker_err_ratelimit(worker, LDNS_RCODE_FORMERR) == -1) {
comm_point_drop_reply(repinfo);
return 0;
}
sldns_buffer_rewind(c->buffer);
LDNS_QR_SET(sldns_buffer_begin(c->buffer));
LDNS_RCODE_SET(sldns_buffer_begin(c->buffer),
LDNS_RCODE_FORMERR);
if(worker->stats.extended) {
worker->stats.qtype[qinfo.qtype]++;
server_stats_insrcode(&worker->stats, c->buffer);
}
goto send_reply;
}
if((ret=parse_edns_from_pkt(c->buffer, &edns, worker->scratchpad)) != 0) {
struct edns_data reply_edns;
verbose(VERB_ALGO, "worker parse edns: formerror.");
log_addr(VERB_CLIENT,"from",&repinfo->addr, repinfo->addrlen);
memset(&reply_edns, 0, sizeof(reply_edns));
reply_edns.edns_present = 1;
reply_edns.udp_size = EDNS_ADVERTISED_SIZE;
LDNS_RCODE_SET(sldns_buffer_begin(c->buffer), ret);
error_encode(c->buffer, ret, &qinfo,
*(uint16_t*)(void *)sldns_buffer_begin(c->buffer),
sldns_buffer_read_u16_at(c->buffer, 2), &reply_edns);
regional_free_all(worker->scratchpad);
server_stats_insrcode(&worker->stats, c->buffer);
goto send_reply;
}
if(edns.edns_present && edns.edns_version != 0) {
edns.ext_rcode = (uint8_t)(EDNS_RCODE_BADVERS>>4);
edns.edns_version = EDNS_ADVERTISED_VERSION;
edns.udp_size = EDNS_ADVERTISED_SIZE;
edns.bits &= EDNS_DO;
edns.opt_list = NULL;
verbose(VERB_ALGO, "query with bad edns version.");
log_addr(VERB_CLIENT,"from",&repinfo->addr, repinfo->addrlen);
error_encode(c->buffer, EDNS_RCODE_BADVERS&0xf, &qinfo,
*(uint16_t*)(void *)sldns_buffer_begin(c->buffer),
sldns_buffer_read_u16_at(c->buffer, 2), NULL);
attach_edns_record(c->buffer, &edns);
regional_free_all(worker->scratchpad);
goto send_reply;
}
if(edns.edns_present && edns.udp_size < NORMAL_UDP_SIZE &&
worker->daemon->cfg->harden_short_bufsize) {
verbose(VERB_QUERY, "worker request: EDNS bufsize %d ignored",
(int)edns.udp_size);
log_addr(VERB_CLIENT,"from",&repinfo->addr, repinfo->addrlen);
edns.udp_size = NORMAL_UDP_SIZE;
}
if(edns.udp_size > worker->daemon->cfg->max_udp_size &&
c->type == comm_udp) {
verbose(VERB_QUERY,
"worker request: max UDP reply size modified"
" (%d to max-udp-size)", (int)edns.udp_size);
log_addr(VERB_CLIENT,"from",&repinfo->addr, repinfo->addrlen);
edns.udp_size = worker->daemon->cfg->max_udp_size;
}
if(edns.udp_size < LDNS_HEADER_SIZE) {
verbose(VERB_ALGO, "worker request: edns is too small.");
log_addr(VERB_CLIENT, "from", &repinfo->addr, repinfo->addrlen);
LDNS_QR_SET(sldns_buffer_begin(c->buffer));
LDNS_TC_SET(sldns_buffer_begin(c->buffer));
LDNS_RCODE_SET(sldns_buffer_begin(c->buffer),
LDNS_RCODE_SERVFAIL);
sldns_buffer_set_position(c->buffer, LDNS_HEADER_SIZE);
sldns_buffer_write_at(c->buffer, 4,
(uint8_t*)"\0\0\0\0\0\0\0\0", 8);
sldns_buffer_flip(c->buffer);
regional_free_all(worker->scratchpad);
goto send_reply;
}
if(worker->stats.extended)
server_stats_insquery(&worker->stats, c, qinfo.qtype,
qinfo.qclass, &edns, repinfo);
if(c->type != comm_udp)
edns.udp_size = 65535; /* max size for TCP replies */
if(qinfo.qclass == LDNS_RR_CLASS_CH && answer_chaos(worker, &qinfo,
&edns, c->buffer)) {
server_stats_insrcode(&worker->stats, c->buffer);
regional_free_all(worker->scratchpad);
goto send_reply;
}
if(local_zones_answer(worker->daemon->local_zones, &worker->env, &qinfo,
&edns, c->buffer, worker->scratchpad, repinfo, acladdr->taglist,
acladdr->taglen, acladdr->tag_actions,
acladdr->tag_actions_size, acladdr->tag_datas,
acladdr->tag_datas_size, worker->daemon->cfg->tagname,
worker->daemon->cfg->num_tags, acladdr->view)) {
regional_free_all(worker->scratchpad);
if(sldns_buffer_limit(c->buffer) == 0) {
comm_point_drop_reply(repinfo);
return 0;
}
server_stats_insrcode(&worker->stats, c->buffer);
goto send_reply;
}
/* We've looked in our local zones. If the answer isn't there, we
* might need to bail out based on ACLs now. */
if((ret=deny_refuse_non_local(c, acl, worker, repinfo)) != -1)
{
regional_free_all(worker->scratchpad);
if(ret == 1)
goto send_reply;
return ret;
}
/* If this request does not have the recursion bit set, verify
* ACLs allow the snooping. */
if(!(LDNS_RD_WIRE(sldns_buffer_begin(c->buffer))) &&
acl != acl_allow_snoop ) {
sldns_buffer_set_limit(c->buffer, LDNS_HEADER_SIZE);
sldns_buffer_write_at(c->buffer, 4,
(uint8_t*)"\0\0\0\0\0\0\0\0", 8);
LDNS_QR_SET(sldns_buffer_begin(c->buffer));
LDNS_RCODE_SET(sldns_buffer_begin(c->buffer),
LDNS_RCODE_REFUSED);
sldns_buffer_flip(c->buffer);
regional_free_all(worker->scratchpad);
server_stats_insrcode(&worker->stats, c->buffer);
log_addr(VERB_ALGO, "refused nonrec (cache snoop) query from",
&repinfo->addr, repinfo->addrlen);
goto send_reply;
}
/* If we've found a local alias, replace the qname with the alias
* target before resolving it. */
if(qinfo.local_alias) {
struct ub_packed_rrset_key* rrset = qinfo.local_alias->rrset;
struct packed_rrset_data* d = rrset->entry.data;
/* Sanity check: our current implementation only supports
* a single CNAME RRset as a local alias. */
if(qinfo.local_alias->next ||
rrset->rk.type != htons(LDNS_RR_TYPE_CNAME) ||
d->count != 1) {
log_err("assumption failure: unexpected local alias");
regional_free_all(worker->scratchpad);
return 0; /* drop it */
}
qinfo.qname = d->rr_data[0] + 2;
qinfo.qname_len = d->rr_len[0] - 2;
}
/* If we may apply IP-based actions to the answer, build the client
* information. As this can be expensive, skip it if there is
* absolutely no possibility of it. */
if(worker->daemon->use_response_ip &&
(qinfo.qtype == LDNS_RR_TYPE_A ||
qinfo.qtype == LDNS_RR_TYPE_AAAA ||
qinfo.qtype == LDNS_RR_TYPE_ANY)) {
cinfo_tmp.taglist = acladdr->taglist;
cinfo_tmp.taglen = acladdr->taglen;
cinfo_tmp.tag_actions = acladdr->tag_actions;
cinfo_tmp.tag_actions_size = acladdr->tag_actions_size;
cinfo_tmp.tag_datas = acladdr->tag_datas;
cinfo_tmp.tag_datas_size = acladdr->tag_datas_size;
cinfo_tmp.view = acladdr->view;
cinfo_tmp.respip_set = worker->daemon->respip_set;
cinfo = &cinfo_tmp;
}
lookup_cache:
/* Lookup the cache. In case we chase an intermediate CNAME chain
* this is a two-pass operation, and lookup_qinfo is different for
* each pass. We should still pass the original qinfo to
* answer_from_cache(), however, since it's used to build the reply. */
if(!edns_bypass_cache_stage(edns.opt_list, &worker->env)) {
h = query_info_hash(lookup_qinfo, sldns_buffer_read_u16_at(c->buffer, 2));
if((e=slabhash_lookup(worker->env.msg_cache, h, lookup_qinfo, 0))) {
/* answer from cache - we have acquired a readlock on it */
if(answer_from_cache(worker, &qinfo,
cinfo, &need_drop, &alias_rrset, &partial_rep,
(struct reply_info*)e->data,
*(uint16_t*)(void *)sldns_buffer_begin(c->buffer),
sldns_buffer_read_u16_at(c->buffer, 2), repinfo,
&edns)) {
/* prefetch it if the prefetch TTL expired.
* Note that if there is more than one pass
* its qname must be that used for cache
* lookup. */
if((worker->env.cfg->prefetch || worker->env.cfg->serve_expired)
&& *worker->env.now >=
((struct reply_info*)e->data)->prefetch_ttl) {
time_t leeway = ((struct reply_info*)e->
data)->ttl - *worker->env.now;
if(((struct reply_info*)e->data)->ttl
< *worker->env.now)
leeway = 0;
lock_rw_unlock(&e->lock);
reply_and_prefetch(worker, lookup_qinfo,
sldns_buffer_read_u16_at(c->buffer, 2),
repinfo, leeway);
if(!partial_rep) {
rc = 0;
regional_free_all(worker->scratchpad);
goto send_reply_rc;
}
} else if(!partial_rep) {
lock_rw_unlock(&e->lock);
regional_free_all(worker->scratchpad);
goto send_reply;
}
/* We've found a partial reply ending with an
* alias. Replace the lookup qinfo for the
* alias target and lookup the cache again to
* (possibly) complete the reply. As we're
* passing the "base" reply, there will be no
* more alias chasing. */
lock_rw_unlock(&e->lock);
memset(&qinfo_tmp, 0, sizeof(qinfo_tmp));
get_cname_target(alias_rrset, &qinfo_tmp.qname,
&qinfo_tmp.qname_len);
if(!qinfo_tmp.qname) {
log_err("unexpected: invalid answer alias");
regional_free_all(worker->scratchpad);
return 0; /* drop query */
}
qinfo_tmp.qtype = qinfo.qtype;
qinfo_tmp.qclass = qinfo.qclass;
lookup_qinfo = &qinfo_tmp;
goto lookup_cache;
}
verbose(VERB_ALGO, "answer from the cache failed");
lock_rw_unlock(&e->lock);
}
if(!LDNS_RD_WIRE(sldns_buffer_begin(c->buffer))) {
if(answer_norec_from_cache(worker, &qinfo,
*(uint16_t*)(void *)sldns_buffer_begin(c->buffer),
sldns_buffer_read_u16_at(c->buffer, 2), repinfo,
&edns)) {
regional_free_all(worker->scratchpad);
goto send_reply;
}
verbose(VERB_ALGO, "answer norec from cache -- "
"need to validate or not primed");
}
}
sldns_buffer_rewind(c->buffer);
server_stats_querymiss(&worker->stats, worker);
if(verbosity >= VERB_CLIENT) {
if(c->type == comm_udp)
log_addr(VERB_CLIENT, "udp request from",
&repinfo->addr, repinfo->addrlen);
else log_addr(VERB_CLIENT, "tcp request from",
&repinfo->addr, repinfo->addrlen);
}
/* grab a work request structure for this new request */
mesh_new_client(worker->env.mesh, &qinfo, cinfo,
sldns_buffer_read_u16_at(c->buffer, 2),
&edns, repinfo, *(uint16_t*)(void *)sldns_buffer_begin(c->buffer));
regional_free_all(worker->scratchpad);
worker_mem_report(worker, NULL);
return 0;
send_reply:
rc = 1;
send_reply_rc:
if(need_drop) {
comm_point_drop_reply(repinfo);
return 0;
}
#ifdef USE_DNSTAP
if(worker->dtenv.log_client_response_messages)
dt_msg_send_client_response(&worker->dtenv, &repinfo->addr,
c->type, c->buffer);
#endif
if(worker->env.cfg->log_replies)
{
struct timeval tv = {0, 0};
log_reply_info(0, &qinfo, &repinfo->addr, repinfo->addrlen,
tv, 1, c->buffer);
}
#ifdef USE_DNSCRYPT
if(!dnsc_handle_uncurved_request(repinfo)) {
return 0;
}
#endif
return rc;
}
void
worker_sighandler(int sig, void* arg)
{
/* note that log, print, syscalls here give race conditions.
* And cause hangups if the log-lock is held by the application. */
struct worker* worker = (struct worker*)arg;
switch(sig) {
#ifdef SIGHUP
case SIGHUP:
comm_base_exit(worker->base);
break;
#endif
case SIGINT:
worker->need_to_exit = 1;
comm_base_exit(worker->base);
break;
#ifdef SIGQUIT
case SIGQUIT:
worker->need_to_exit = 1;
comm_base_exit(worker->base);
break;
#endif
case SIGTERM:
worker->need_to_exit = 1;
comm_base_exit(worker->base);
break;
default:
/* unknown signal, ignored */
break;
}
}
/** restart statistics timer for worker, if enabled */
static void
worker_restart_timer(struct worker* worker)
{
if(worker->env.cfg->stat_interval > 0) {
struct timeval tv;
#ifndef S_SPLINT_S
tv.tv_sec = worker->env.cfg->stat_interval;
tv.tv_usec = 0;
#endif
comm_timer_set(worker->stat_timer, &tv);
}
}
void worker_stat_timer_cb(void* arg)
{
struct worker* worker = (struct worker*)arg;
server_stats_log(&worker->stats, worker, worker->thread_num);
mesh_stats(worker->env.mesh, "mesh has");
worker_mem_report(worker, NULL);
/* SHM is enabled, process data to SHM */
if (worker->daemon->cfg->shm_enable) {
shm_main_run(worker);
}
if(!worker->daemon->cfg->stat_cumulative) {
worker_stats_clear(worker);
}
/* start next timer */
worker_restart_timer(worker);
}
void worker_probe_timer_cb(void* arg)
{
struct worker* worker = (struct worker*)arg;
struct timeval tv;
#ifndef S_SPLINT_S
tv.tv_sec = (time_t)autr_probe_timer(&worker->env);
tv.tv_usec = 0;
#endif
if(tv.tv_sec != 0)
comm_timer_set(worker->env.probe_timer, &tv);
}
struct worker*
worker_create(struct daemon* daemon, int id, int* ports, int n)
{
unsigned int seed;
struct worker* worker = (struct worker*)calloc(1,
sizeof(struct worker));
if(!worker)
return NULL;
worker->numports = n;
worker->ports = (int*)memdup(ports, sizeof(int)*n);
if(!worker->ports) {
free(worker);
return NULL;
}
worker->daemon = daemon;
worker->thread_num = id;
if(!(worker->cmd = tube_create())) {
free(worker->ports);
free(worker);
return NULL;
}
/* create random state here to avoid locking trouble in RAND_bytes */
seed = (unsigned int)time(NULL) ^ (unsigned int)getpid() ^
(((unsigned int)worker->thread_num)<<17);
/* shift thread_num so it does not match out pid bits */
if(!(worker->rndstate = ub_initstate(seed, daemon->rand))) {
seed = 0;
log_err("could not init random numbers.");
tube_delete(worker->cmd);
free(worker->ports);
free(worker);
return NULL;
}
seed = 0;
#ifdef USE_DNSTAP
if(daemon->cfg->dnstap) {
log_assert(daemon->dtenv != NULL);
memcpy(&worker->dtenv, daemon->dtenv, sizeof(struct dt_env));
if(!dt_init(&worker->dtenv))
fatal_exit("dt_init failed");
}
#endif
return worker;
}
int
worker_init(struct worker* worker, struct config_file *cfg,
struct listen_port* ports, int do_sigs)
{
#ifdef USE_DNSTAP
struct dt_env* dtenv = &worker->dtenv;
#else
void* dtenv = NULL;
#endif
worker->need_to_exit = 0;
worker->base = comm_base_create(do_sigs);
if(!worker->base) {
log_err("could not create event handling base");
worker_delete(worker);
return 0;
}
comm_base_set_slow_accept_handlers(worker->base, &worker_stop_accept,
&worker_start_accept, worker);
if(do_sigs) {
#ifdef SIGHUP
ub_thread_sig_unblock(SIGHUP);
#endif
ub_thread_sig_unblock(SIGINT);
#ifdef SIGQUIT
ub_thread_sig_unblock(SIGQUIT);
#endif
ub_thread_sig_unblock(SIGTERM);
#ifndef LIBEVENT_SIGNAL_PROBLEM
worker->comsig = comm_signal_create(worker->base,
worker_sighandler, worker);
if(!worker->comsig
#ifdef SIGHUP
|| !comm_signal_bind(worker->comsig, SIGHUP)
#endif
#ifdef SIGQUIT
|| !comm_signal_bind(worker->comsig, SIGQUIT)
#endif
|| !comm_signal_bind(worker->comsig, SIGTERM)
|| !comm_signal_bind(worker->comsig, SIGINT)) {
log_err("could not create signal handlers");
worker_delete(worker);
return 0;
}
#endif /* LIBEVENT_SIGNAL_PROBLEM */
if(!daemon_remote_open_accept(worker->daemon->rc,
worker->daemon->rc_ports, worker)) {
worker_delete(worker);
return 0;
}
#ifdef UB_ON_WINDOWS
wsvc_setup_worker(worker);
#endif /* UB_ON_WINDOWS */
} else { /* !do_sigs */
worker->comsig = NULL;
}
worker->front = listen_create(worker->base, ports,
cfg->msg_buffer_size, (int)cfg->incoming_num_tcp,
worker->daemon->listen_sslctx, dtenv, worker_handle_request,
worker);
if(!worker->front) {
log_err("could not create listening sockets");
worker_delete(worker);
return 0;
}
worker->back = outside_network_create(worker->base,
cfg->msg_buffer_size, (size_t)cfg->outgoing_num_ports,
cfg->out_ifs, cfg->num_out_ifs, cfg->do_ip4, cfg->do_ip6,
cfg->do_tcp?cfg->outgoing_num_tcp:0,
worker->daemon->env->infra_cache, worker->rndstate,
cfg->use_caps_bits_for_id, worker->ports, worker->numports,
cfg->unwanted_threshold, cfg->outgoing_tcp_mss,
&worker_alloc_cleanup, worker,
cfg->do_udp, worker->daemon->connect_sslctx, cfg->delay_close,
dtenv);
if(!worker->back) {
log_err("could not create outgoing sockets");
worker_delete(worker);
return 0;
}
/* start listening to commands */
if(!tube_setup_bg_listen(worker->cmd, worker->base,
&worker_handle_control_cmd, worker)) {
log_err("could not create control compt.");
worker_delete(worker);
return 0;
}
worker->stat_timer = comm_timer_create(worker->base,
worker_stat_timer_cb, worker);
if(!worker->stat_timer) {
log_err("could not create statistics timer");
}
/* we use the msg_buffer_size as a good estimate for what the
* user wants for memory usage sizes */
worker->scratchpad = regional_create_custom(cfg->msg_buffer_size);
if(!worker->scratchpad) {
log_err("malloc failure");
worker_delete(worker);
return 0;
}
server_stats_init(&worker->stats, cfg);
alloc_init(&worker->alloc, &worker->daemon->superalloc,
worker->thread_num);
alloc_set_id_cleanup(&worker->alloc, &worker_alloc_cleanup, worker);
worker->env = *worker->daemon->env;
comm_base_timept(worker->base, &worker->env.now, &worker->env.now_tv);
if(worker->thread_num == 0)
log_set_time(worker->env.now);
worker->env.worker = worker;
worker->env.send_query = &worker_send_query;
worker->env.alloc = &worker->alloc;
worker->env.rnd = worker->rndstate;
worker->env.scratch = worker->scratchpad;
worker->env.mesh = mesh_create(&worker->daemon->mods, &worker->env);
worker->env.detach_subs = &mesh_detach_subs;
worker->env.attach_sub = &mesh_attach_sub;
worker->env.kill_sub = &mesh_state_delete;
worker->env.detect_cycle = &mesh_detect_cycle;
worker->env.scratch_buffer = sldns_buffer_new(cfg->msg_buffer_size);
if(!(worker->env.fwds = forwards_create()) ||
!forwards_apply_cfg(worker->env.fwds, cfg)) {
log_err("Could not set forward zones");
worker_delete(worker);
return 0;
}
if(!(worker->env.hints = hints_create()) ||
!hints_apply_cfg(worker->env.hints, cfg)) {
log_err("Could not set root or stub hints");
worker_delete(worker);
return 0;
}
/* one probe timer per process -- if we have 5011 anchors */
if(autr_get_num_anchors(worker->env.anchors) > 0
#ifndef THREADS_DISABLED
&& worker->thread_num == 0
#endif
) {
struct timeval tv;
tv.tv_sec = 0;
tv.tv_usec = 0;
worker->env.probe_timer = comm_timer_create(worker->base,
worker_probe_timer_cb, worker);
if(!worker->env.probe_timer) {
log_err("could not create 5011-probe timer");
} else {
/* let timer fire, then it can reset itself */
comm_timer_set(worker->env.probe_timer, &tv);
}
}
if(!worker->env.mesh || !worker->env.scratch_buffer) {
worker_delete(worker);
return 0;
}
worker_mem_report(worker, NULL);
/* if statistics enabled start timer */
if(worker->env.cfg->stat_interval > 0) {
verbose(VERB_ALGO, "set statistics interval %d secs",
worker->env.cfg->stat_interval);
worker_restart_timer(worker);
}
return 1;
}
void
worker_work(struct worker* worker)
{
comm_base_dispatch(worker->base);
}
void
worker_delete(struct worker* worker)
{
if(!worker)
return;
if(worker->env.mesh && verbosity >= VERB_OPS) {
server_stats_log(&worker->stats, worker, worker->thread_num);
mesh_stats(worker->env.mesh, "mesh has");
worker_mem_report(worker, NULL);
}
outside_network_quit_prepare(worker->back);
mesh_delete(worker->env.mesh);
sldns_buffer_free(worker->env.scratch_buffer);
forwards_delete(worker->env.fwds);
hints_delete(worker->env.hints);
listen_delete(worker->front);
outside_network_delete(worker->back);
comm_signal_delete(worker->comsig);
tube_delete(worker->cmd);
comm_timer_delete(worker->stat_timer);
comm_timer_delete(worker->env.probe_timer);
free(worker->ports);
if(worker->thread_num == 0) {
log_set_time(NULL);
#ifdef UB_ON_WINDOWS
wsvc_desetup_worker(worker);
#endif /* UB_ON_WINDOWS */
}
comm_base_delete(worker->base);
ub_randfree(worker->rndstate);
alloc_clear(&worker->alloc);
regional_destroy(worker->scratchpad);
free(worker);
}
struct outbound_entry*
worker_send_query(struct query_info* qinfo, uint16_t flags, int dnssec,
int want_dnssec, int nocaps, struct sockaddr_storage* addr,
socklen_t addrlen, uint8_t* zone, size_t zonelen, int ssl_upstream,
struct module_qstate* q)
{
struct worker* worker = q->env->worker;
struct outbound_entry* e = (struct outbound_entry*)regional_alloc(
q->region, sizeof(*e));
if(!e)
return NULL;
e->qstate = q;
e->qsent = outnet_serviced_query(worker->back, qinfo, flags, dnssec,
want_dnssec, nocaps, q->env->cfg->tcp_upstream,
ssl_upstream, addr, addrlen, zone, zonelen, q,
worker_handle_service_reply, e, worker->back->udp_buff, q->env);
if(!e->qsent) {
return NULL;
}
return e;
}
void
worker_alloc_cleanup(void* arg)
{
struct worker* worker = (struct worker*)arg;
slabhash_clear(&worker->env.rrset_cache->table);
slabhash_clear(worker->env.msg_cache);
}
void worker_stats_clear(struct worker* worker)
{
server_stats_init(&worker->stats, worker->env.cfg);
mesh_stats_clear(worker->env.mesh);
worker->back->unwanted_replies = 0;
worker->back->num_tcp_outgoing = 0;
}
void worker_start_accept(void* arg)
{
struct worker* worker = (struct worker*)arg;
listen_start_accept(worker->front);
if(worker->thread_num == 0)
daemon_remote_start_accept(worker->daemon->rc);
}
void worker_stop_accept(void* arg)
{
struct worker* worker = (struct worker*)arg;
listen_stop_accept(worker->front);
if(worker->thread_num == 0)
daemon_remote_stop_accept(worker->daemon->rc);
}
/* --- fake callbacks for fptr_wlist to work --- */
struct outbound_entry* libworker_send_query(
struct query_info* ATTR_UNUSED(qinfo),
uint16_t ATTR_UNUSED(flags), int ATTR_UNUSED(dnssec),
int ATTR_UNUSED(want_dnssec), int ATTR_UNUSED(nocaps),
struct sockaddr_storage* ATTR_UNUSED(addr), socklen_t ATTR_UNUSED(addrlen),
uint8_t* ATTR_UNUSED(zone), size_t ATTR_UNUSED(zonelen),
int ATTR_UNUSED(ssl_upstream), struct module_qstate* ATTR_UNUSED(q))
{
log_assert(0);
return 0;
}
int libworker_handle_reply(struct comm_point* ATTR_UNUSED(c),
void* ATTR_UNUSED(arg), int ATTR_UNUSED(error),
struct comm_reply* ATTR_UNUSED(reply_info))
{
log_assert(0);
return 0;
}
int libworker_handle_service_reply(struct comm_point* ATTR_UNUSED(c),
void* ATTR_UNUSED(arg), int ATTR_UNUSED(error),
struct comm_reply* ATTR_UNUSED(reply_info))
{
log_assert(0);
return 0;
}
void libworker_handle_control_cmd(struct tube* ATTR_UNUSED(tube),
uint8_t* ATTR_UNUSED(buffer), size_t ATTR_UNUSED(len),
int ATTR_UNUSED(error), void* ATTR_UNUSED(arg))
{
log_assert(0);
}
void libworker_fg_done_cb(void* ATTR_UNUSED(arg), int ATTR_UNUSED(rcode),
sldns_buffer* ATTR_UNUSED(buf), enum sec_status ATTR_UNUSED(s),
char* ATTR_UNUSED(why_bogus))
{
log_assert(0);
}
void libworker_bg_done_cb(void* ATTR_UNUSED(arg), int ATTR_UNUSED(rcode),
sldns_buffer* ATTR_UNUSED(buf), enum sec_status ATTR_UNUSED(s),
char* ATTR_UNUSED(why_bogus))
{
log_assert(0);
}
void libworker_event_done_cb(void* ATTR_UNUSED(arg), int ATTR_UNUSED(rcode),
sldns_buffer* ATTR_UNUSED(buf), enum sec_status ATTR_UNUSED(s),
char* ATTR_UNUSED(why_bogus))
{
log_assert(0);
}
int context_query_cmp(const void* ATTR_UNUSED(a), const void* ATTR_UNUSED(b))
{
log_assert(0);
return 0;
}
int order_lock_cmp(const void* ATTR_UNUSED(e1), const void* ATTR_UNUSED(e2))
{
log_assert(0);
return 0;
}
int codeline_cmp(const void* ATTR_UNUSED(a), const void* ATTR_UNUSED(b))
{
log_assert(0);
return 0;
}
Reference in New Issue View Git Blame Copy Permalink