freebsd-skq/contrib/unbound/daemon/worker.c

1453 lines
45 KiB
C
Raw Normal View History

2012-07-04 14:24:26 +00:00
/*
* 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.
2012-07-04 14:24:26 +00:00
*/
/**
* \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 "libunbound/context.h"
#include "libunbound/libworker.h"
#include "ldns/sbuffer.h"
2012-07-04 14:24:26 +00:00
#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 */
/**
* 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
#ifdef UNBOUND_ALLOC_STATS
/** measure memory leakage */
static void
debug_memleak(size_t accounted, size_t heap,
size_t total_alloc, size_t total_free)
{
static int init = 0;
static size_t base_heap, base_accounted, base_alloc, base_free;
size_t base_af, cur_af, grow_af, grow_acc;
if(!init) {
init = 1;
base_heap = heap;
base_accounted = accounted;
base_alloc = total_alloc;
base_free = total_free;
}
base_af = base_alloc - base_free;
cur_af = total_alloc - total_free;
grow_af = cur_af - base_af;
grow_acc = accounted - base_accounted;
log_info("Leakage: %d leaked. growth: %u use, %u acc, %u heap",
(int)(grow_af - grow_acc), (unsigned)grow_af,
(unsigned)grow_acc, (unsigned)(heap - base_heap));
}
/** give debug heap size indication */
static void
debug_total_mem(size_t calctotal)
{
#ifdef HAVE_SBRK
extern void* unbound_start_brk;
extern size_t unbound_mem_alloc, unbound_mem_freed;
void* cur = sbrk(0);
int total = cur-unbound_start_brk;
log_info("Total heap memory estimate: %u total-alloc: %u "
"total-free: %u", (unsigned)total,
(unsigned)unbound_mem_alloc, (unsigned)unbound_mem_freed);
debug_memleak(calctotal, (size_t)total,
unbound_mem_alloc, unbound_mem_freed);
#else
(void)calctotal;
#endif /* HAVE_SBRK */
}
#endif /* UNBOUND_ALLOC_STATS */
/** 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
/* debug func in validator module */
size_t total, front, back, mesh, msg, rrset, infra, ac, superac;
size_t me, iter, val, anch;
int i;
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);
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)
2012-07-04 14:24:26 +00:00
+ 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;
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);
debug_total_mem(total);
#else /* no UNBOUND_ALLOC_STATS */
size_t val = 0;
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);
}
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 /* 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)) !=
2012-07-04 14:24:26 +00:00
LDNS_PACKET_QUERY
|| LDNS_QDCOUNT(sldns_buffer_begin(c->buffer)) > 1) {
2012-07-04 14:24:26 +00:00
/* 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)) !=
2012-07-04 14:24:26 +00:00
LDNS_PACKET_QUERY
|| LDNS_QDCOUNT(sldns_buffer_begin(c->buffer)) > 1) {
2012-07-04 14:24:26 +00:00
/* 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;
}
/** 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)
2012-07-04 14:24:26 +00:00
{
if(sldns_buffer_limit(pkt) < LDNS_HEADER_SIZE) {
2012-07-04 14:24:26 +00:00
verbose(VERB_QUERY, "request too short, discarded");
return -1;
}
if(sldns_buffer_limit(pkt) > NORMAL_UDP_SIZE &&
2012-07-04 14:24:26 +00:00
worker->daemon->cfg->harden_large_queries) {
verbose(VERB_QUERY, "request too large, discarded");
return -1;
}
if(LDNS_QR_WIRE(sldns_buffer_begin(pkt))) {
2012-07-04 14:24:26 +00:00
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));
2012-07-04 14:24:26 +00:00
verbose(VERB_QUERY, "request bad, has TC bit on");
return LDNS_RCODE_FORMERR;
}
if(LDNS_OPCODE_WIRE(sldns_buffer_begin(pkt)) != LDNS_PACKET_QUERY) {
2012-07-04 14:24:26 +00:00
verbose(VERB_QUERY, "request unknown opcode %d",
LDNS_OPCODE_WIRE(sldns_buffer_begin(pkt)));
2012-07-04 14:24:26 +00:00
return LDNS_RCODE_NOTIMPL;
}
if(LDNS_QDCOUNT(sldns_buffer_begin(pkt)) != 1) {
2012-07-04 14:24:26 +00:00
verbose(VERB_QUERY, "request wrong nr qd=%d",
LDNS_QDCOUNT(sldns_buffer_begin(pkt)));
2012-07-04 14:24:26 +00:00
return LDNS_RCODE_FORMERR;
}
if(LDNS_ANCOUNT(sldns_buffer_begin(pkt)) != 0) {
2012-07-04 14:24:26 +00:00
verbose(VERB_QUERY, "request wrong nr an=%d",
LDNS_ANCOUNT(sldns_buffer_begin(pkt)));
2012-07-04 14:24:26 +00:00
return LDNS_RCODE_FORMERR;
}
if(LDNS_NSCOUNT(sldns_buffer_begin(pkt)) != 0) {
2012-07-04 14:24:26 +00:00
verbose(VERB_QUERY, "request wrong nr ns=%d",
LDNS_NSCOUNT(sldns_buffer_begin(pkt)));
2012-07-04 14:24:26 +00:00
return LDNS_RCODE_FORMERR;
}
if(LDNS_ARCOUNT(sldns_buffer_begin(pkt)) > 1) {
2012-07-04 14:24:26 +00:00
verbose(VERB_QUERY, "request wrong nr ar=%d",
LDNS_ARCOUNT(sldns_buffer_begin(pkt)));
2012-07-04 14:24:26 +00:00
return 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);
2012-07-04 14:24:26 +00:00
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;
2012-07-04 14:24:26 +00:00
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 */
regional_free_all(worker->scratchpad);
return 0;
}
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 */
regional_free_all(worker->scratchpad);
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;
error_encode(repinfo->c->buffer, LDNS_RCODE_SERVFAIL,
&msg->qinfo, id, flags, edns);
regional_free_all(worker->scratchpad);
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;
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)) {
error_encode(repinfo->c->buffer, LDNS_RCODE_SERVFAIL,
&msg->qinfo, id, flags, edns);
}
regional_free_all(worker->scratchpad);
if(worker->stats.extended) {
if(secure) worker->stats.ans_secure++;
server_stats_insrcode(&worker->stats, repinfo->c->buffer);
}
return 1;
}
/** answer query from the cache */
static int
answer_from_cache(struct worker* worker, struct query_info* qinfo,
struct reply_info* rep, uint16_t id, uint16_t flags,
struct comm_reply* repinfo, struct edns_data* edns)
{
time_t timenow = *worker->env.now;
2012-07-04 14:24:26 +00:00
uint16_t udpsize = edns->udp_size;
int secure;
int must_validate = (!(flags&BIT_CD) || worker->env.cfg->ignore_cd)
&& worker->env.need_to_validate;
/* 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(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);
regional_free_all(worker->scratchpad);
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;
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);
regional_free_all(worker->scratchpad);
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(!reply_info_answer_encode(qinfo, rep, id, flags,
repinfo->c->buffer, timenow, 1, worker->scratchpad,
udpsize, edns, (int)(edns->bits & EDNS_DO), secure)) {
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);
regional_free_all(worker->scratchpad);
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 */
static void
reply_and_prefetch(struct worker* worker, struct query_info* qinfo,
uint16_t flags, struct comm_reply* repinfo, time_t leeway)
2012-07-04 14:24:26 +00:00
{
/* first send answer to client to keep its latency
* as small as a cachereply */
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).
* @param edns: edns reply information.
*/
static void
chaos_replystr(sldns_buffer* pkt, const char* str, struct edns_data* edns)
2012-07-04 14:24:26 +00:00
{
size_t len = strlen(str);
unsigned int rd = LDNS_RD_WIRE(sldns_buffer_begin(pkt));
unsigned int cd = LDNS_CD_WIRE(sldns_buffer_begin(pkt));
2012-07-04 14:24:26 +00:00
if(len>255) len=255; /* cap size of TXT record */
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, 1); /* ancount */
sldns_buffer_write_u16(pkt, 0); /* nscount */
sldns_buffer_write_u16(pkt, 0); /* arcount */
2012-07-04 14:24:26 +00:00
(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 */
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, len);
sldns_buffer_flip(pkt);
2012-07-04 14:24:26 +00:00
edns->edns_version = EDNS_ADVERTISED_VERSION;
edns->udp_size = EDNS_ADVERTISED_SIZE;
edns->bits &= EDNS_DO;
attach_edns_record(pkt, edns);
}
/**
* 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)
2012-07-04 14:24:26 +00:00
{
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_replystr(pkt, buf, edns);
} else {
log_err("gethostname: %s", strerror(errno));
chaos_replystr(pkt, "no hostname", edns);
}
}
else chaos_replystr(pkt, cfg->identity, edns);
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_replystr(pkt, PACKAGE_STRING, edns);
else chaos_replystr(pkt, cfg->version, edns);
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);
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);
}
2012-07-04 14:24:26 +00:00
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_t h;
struct lruhash_entry* e;
struct query_info qinfo;
struct edns_data edns;
enum acl_access acl;
int rc = 0;
2012-07-04 14:24:26 +00:00
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_DNSTAP
if(worker->dtenv.log_client_query_messages)
dt_msg_send_client_query(&worker->dtenv, &repinfo->addr, c->type,
c->buffer);
#endif
2012-07-04 14:24:26 +00:00
acl = acl_list_lookup(worker->daemon->acl, &repinfo->addr,
repinfo->addrlen);
if((ret=deny_refuse_all(c, acl, worker, repinfo)) != -1)
{
if(ret == 1)
goto send_reply;
return ret;
2012-07-04 14:24:26 +00:00
}
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);
2012-07-04 14:24:26 +00:00
return 1;
}
comm_point_drop_reply(repinfo);
return 0;
}
worker->stats.num_queries++;
/* 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);
sldns_buffer_rewind(c->buffer);
LDNS_QR_SET(sldns_buffer_begin(c->buffer));
LDNS_RCODE_SET(sldns_buffer_begin(c->buffer),
2012-07-04 14:24:26 +00:00
LDNS_RCODE_FORMERR);
server_stats_insrcode(&worker->stats, c->buffer);
goto send_reply;
2012-07-04 14:24:26 +00:00
}
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),
2012-07-04 14:24:26 +00:00
LDNS_RCODE_REFUSED);
if(worker->stats.extended) {
worker->stats.qtype[qinfo.qtype]++;
server_stats_insrcode(&worker->stats, c->buffer);
}
goto send_reply;
2012-07-04 14:24:26 +00:00
}
if((ret=parse_edns_from_pkt(c->buffer, &edns)) != 0) {
verbose(VERB_ALGO, "worker parse edns: formerror.");
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), ret);
2012-07-04 14:24:26 +00:00
server_stats_insrcode(&worker->stats, c->buffer);
goto send_reply;
2012-07-04 14:24:26 +00:00
}
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;
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);
2012-07-04 14:24:26 +00:00
attach_edns_record(c->buffer, &edns);
goto send_reply;
2012-07-04 14:24:26 +00:00
}
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) {
2012-07-04 14:24:26 +00:00
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),
2012-07-04 14:24:26 +00:00
LDNS_RCODE_SERVFAIL);
sldns_buffer_set_position(c->buffer, LDNS_HEADER_SIZE);
sldns_buffer_write_at(c->buffer, 4,
2012-07-04 14:24:26 +00:00
(uint8_t*)"\0\0\0\0\0\0\0\0", 8);
sldns_buffer_flip(c->buffer);
goto send_reply;
2012-07-04 14:24:26 +00:00
}
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);
goto send_reply;
2012-07-04 14:24:26 +00:00
}
if(local_zones_answer(worker->daemon->local_zones, &qinfo, &edns,
c->buffer, worker->scratchpad, repinfo)) {
2012-07-04 14:24:26 +00:00
regional_free_all(worker->scratchpad);
if(sldns_buffer_limit(c->buffer) == 0) {
2012-07-04 14:24:26 +00:00
comm_point_drop_reply(repinfo);
return 0;
}
server_stats_insrcode(&worker->stats, c->buffer);
goto send_reply;
2012-07-04 14:24:26 +00:00
}
/* 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)
{
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))) &&
2012-07-04 14:24:26 +00:00
acl != acl_allow_snoop ) {
sldns_buffer_set_limit(c->buffer, LDNS_HEADER_SIZE);
sldns_buffer_write_at(c->buffer, 4,
2012-07-04 14:24:26 +00:00
(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),
2012-07-04 14:24:26 +00:00
LDNS_RCODE_REFUSED);
sldns_buffer_flip(c->buffer);
2012-07-04 14:24:26 +00:00
server_stats_insrcode(&worker->stats, c->buffer);
log_addr(VERB_ALGO, "refused nonrec (cache snoop) query from",
&repinfo->addr, repinfo->addrlen);
goto send_reply;
2012-07-04 14:24:26 +00:00
}
h = query_info_hash(&qinfo, sldns_buffer_read_u16_at(c->buffer, 2));
2012-07-04 14:24:26 +00:00
if((e=slabhash_lookup(worker->env.msg_cache, h, &qinfo, 0))) {
/* answer from cache - we have acquired a readlock on it */
if(answer_from_cache(worker, &qinfo,
(struct reply_info*)e->data,
*(uint16_t*)(void *)sldns_buffer_begin(c->buffer),
sldns_buffer_read_u16_at(c->buffer, 2), repinfo,
2012-07-04 14:24:26 +00:00
&edns)) {
/* prefetch it if the prefetch TTL expired */
if(worker->env.cfg->prefetch && *worker->env.now >=
((struct reply_info*)e->data)->prefetch_ttl) {
time_t leeway = ((struct reply_info*)e->
2012-07-04 14:24:26 +00:00
data)->ttl - *worker->env.now;
lock_rw_unlock(&e->lock);
reply_and_prefetch(worker, &qinfo,
sldns_buffer_read_u16_at(c->buffer, 2),
2012-07-04 14:24:26 +00:00
repinfo, leeway);
rc = 0;
goto send_reply_rc;
2012-07-04 14:24:26 +00:00
}
lock_rw_unlock(&e->lock);
goto send_reply;
2012-07-04 14:24:26 +00:00
}
verbose(VERB_ALGO, "answer from the cache failed");
lock_rw_unlock(&e->lock);
}
if(!LDNS_RD_WIRE(sldns_buffer_begin(c->buffer))) {
2012-07-04 14:24:26 +00:00
if(answer_norec_from_cache(worker, &qinfo,
*(uint16_t*)(void *)sldns_buffer_begin(c->buffer),
sldns_buffer_read_u16_at(c->buffer, 2), repinfo,
2012-07-04 14:24:26 +00:00
&edns)) {
goto send_reply;
2012-07-04 14:24:26 +00:00
}
verbose(VERB_ALGO, "answer norec from cache -- "
"need to validate or not primed");
}
sldns_buffer_rewind(c->buffer);
2012-07-04 14:24:26 +00:00
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,
sldns_buffer_read_u16_at(c->buffer, 2),
&edns, repinfo, *(uint16_t*)(void *)sldns_buffer_begin(c->buffer));
2012-07-04 14:24:26 +00:00
worker_mem_report(worker, NULL);
return 0;
send_reply:
rc = 1;
send_reply_rc:
#ifdef USE_DNSTAP
if(worker->dtenv.log_client_response_messages)
dt_msg_send_client_response(&worker->dtenv, &repinfo->addr,
c->type, c->buffer);
#endif
return rc;
2012-07-04 14:24:26 +00:00
}
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. */
2012-07-04 14:24:26 +00:00
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 */
2012-07-04 14:24:26 +00:00
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);
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
2012-07-04 14:24:26 +00:00
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
2012-07-04 14:24:26 +00:00
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);
2012-07-04 14:24:26 +00:00
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, &worker_alloc_cleanup, worker,
cfg->do_udp, worker->daemon->connect_sslctx, cfg->delay_close,
dtenv);
2012-07-04 14:24:26 +00:00
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);
2012-07-04 14:24:26 +00:00
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);
2012-07-04 14:24:26 +00:00
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(uint8_t* qname, size_t qnamelen, uint16_t qtype,
uint16_t qclass, uint16_t flags, int dnssec, int want_dnssec,
int nocaps, struct sockaddr_storage* addr, socklen_t addrlen,
uint8_t* zone, size_t zonelen, struct module_qstate* q)
2012-07-04 14:24:26 +00:00
{
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, qname,
qnamelen, qtype, qclass, flags, dnssec, want_dnssec, nocaps,
2012-07-04 14:24:26 +00:00
q->env->cfg->tcp_upstream, q->env->cfg->ssl_upstream, addr,
addrlen, zone, zonelen, worker_handle_service_reply, e,
2013-04-05 09:06:26 +00:00
worker->back->udp_buff);
2012-07-04 14:24:26 +00:00
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;
2012-07-04 14:24:26 +00:00
}
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(uint8_t* ATTR_UNUSED(qname),
size_t ATTR_UNUSED(qnamelen), uint16_t ATTR_UNUSED(qtype),
uint16_t ATTR_UNUSED(qclass), 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), struct module_qstate* ATTR_UNUSED(q))
2012-07-04 14:24:26 +00:00
{
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),
2012-07-04 14:24:26 +00:00
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),
2012-07-04 14:24:26 +00:00
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;
}