/* udns_resolver.c resolver stuff (main module) Copyright (C) 2005 Michael Tokarev This file is part of UDNS library, an async DNS stub resolver. This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library, in file named COPYING.LGPL; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */ #ifdef HAVE_CONFIG_H # include "config.h" #endif #if defined(_WINDOWS) || defined(WINDOWS) # include /* includes */ # include /* needed for struct in6_addr */ #else # include # include # include # include # include # include # ifdef HAVE_POLL # include # else # ifdef HAVE_SYS_SELECT_H # include # endif # endif # ifdef HAVE_TIMES # include # endif # define closesocket(sock) close(sock) #endif /* !WINDOWS */ #include #include #include #include #include #include #include "udns.h" #ifndef EAFNOSUPPORT # define EAFNOSUPPORT EINVAL #endif #ifndef MSG_DONTWAIT # define MSG_DONTWAIT 0 #endif struct dns_qlist { struct dns_query *head, *tail; }; struct dns_query { struct dns_query *dnsq_next; /* double-linked list */ struct dns_query *dnsq_prev; unsigned dnsq_origdnl0; /* original query DN len w/o last 0 */ unsigned dnsq_flags; /* control flags for this query */ unsigned dnsq_servi; /* index of next server to try */ unsigned dnsq_servwait; /* bitmask: servers left to wait */ unsigned dnsq_servskip; /* bitmask: servers to skip */ unsigned dnsq_servnEDNS0; /* bitmask: servers refusing EDNS0 */ unsigned dnsq_try; /* number of tries made so far */ dnscc_t *dnsq_nxtsrch; /* next search pointer @dnsc_srchbuf */ time_t dnsq_deadline; /* when current try will expire */ dns_parse_fn *dnsq_parse; /* parse: raw => application */ dns_query_fn *dnsq_cbck; /* the callback to call when done */ void *dnsq_cbdata; /* user data for the callback */ #ifndef NDEBUG struct dns_ctx *dnsq_ctx; /* the resolver context */ #endif /* char fields at the end to avoid padding */ dnsc_t dnsq_id[2]; /* query ID */ dnsc_t dnsq_typcls[4]; /* requested RR type+class */ dnsc_t dnsq_dn[DNS_MAXDN+DNS_DNPAD]; /* the query DN +alignment */ }; /* working with dns_query lists */ static __inline void qlist_init(struct dns_qlist *list) { list->head = list->tail = NULL; } static __inline void qlist_remove(struct dns_qlist *list, struct dns_query *q) { if (q->dnsq_prev) q->dnsq_prev->dnsq_next = q->dnsq_next; else list->head = q->dnsq_next; if (q->dnsq_next) q->dnsq_next->dnsq_prev = q->dnsq_prev; else list->tail = q->dnsq_prev; } static __inline void qlist_add_head(struct dns_qlist *list, struct dns_query *q) { q->dnsq_next = list->head; if (list->head) list->head->dnsq_prev = q; else list->tail = q; list->head = q; q->dnsq_prev = NULL; } static __inline void qlist_insert_after(struct dns_qlist *list, struct dns_query *q, struct dns_query *prev) { if ((q->dnsq_prev = prev) != NULL) { if ((q->dnsq_next = prev->dnsq_next) != NULL) q->dnsq_next->dnsq_prev = q; else list->tail = q; prev->dnsq_next = q; } else qlist_add_head(list, q); } union sockaddr_ns { struct sockaddr sa; struct sockaddr_in sin; #ifdef HAVE_IPv6 struct sockaddr_in6 sin6; #endif }; #define sin_eq(a,b) \ ((a).sin_port == (b).sin_port && \ (a).sin_addr.s_addr == (b).sin_addr.s_addr) #define sin6_eq(a,b) \ ((a).sin6_port == (b).sin6_port && \ memcmp(&(a).sin6_addr, &(b).sin6_addr, sizeof(struct in6_addr)) == 0) struct dns_ctx { /* resolver context */ /* settings */ unsigned dnsc_flags; /* various flags */ unsigned dnsc_timeout; /* timeout (base value) for queries */ unsigned dnsc_ntries; /* number of retries */ unsigned dnsc_ndots; /* ndots to assume absolute name */ unsigned dnsc_port; /* default port (DNS_PORT) */ unsigned dnsc_udpbuf; /* size of UDP buffer */ /* array of nameserver addresses */ union sockaddr_ns dnsc_serv[DNS_MAXSERV]; unsigned dnsc_nserv; /* number of nameservers */ unsigned dnsc_salen; /* length of socket addresses */ dnsc_t dnsc_srchbuf[1024]; /* buffer for searchlist */ dnsc_t *dnsc_srchend; /* current end of srchbuf */ dns_utm_fn *dnsc_utmfn; /* register/cancel timer events */ void *dnsc_utmctx; /* user timer context for utmfn() */ time_t dnsc_utmexp; /* when user timer expires */ dns_dbgfn *dnsc_udbgfn; /* debugging function */ /* dynamic data */ struct udns_jranctx dnsc_jran; /* random number generator state */ unsigned dnsc_nextid; /* next queue ID to use if !0 */ int dnsc_udpsock; /* UDP socket */ struct dns_qlist dnsc_qactive; /* active list sorted by deadline */ int dnsc_nactive; /* number entries in dnsc_qactive */ dnsc_t *dnsc_pbuf; /* packet buffer (udpbuf size) */ int dnsc_qstatus; /* last query status value */ }; static const struct { const char *name; enum dns_opt opt; unsigned offset; unsigned min, max; } dns_opts[] = { #define opt(name,opt,field,min,max) \ {name,opt,offsetof(struct dns_ctx,field),min,max} opt("retrans", DNS_OPT_TIMEOUT, dnsc_timeout, 1,300), opt("timeout", DNS_OPT_TIMEOUT, dnsc_timeout, 1,300), opt("retry", DNS_OPT_NTRIES, dnsc_ntries, 1,50), opt("attempts", DNS_OPT_NTRIES, dnsc_ntries, 1,50), opt("ndots", DNS_OPT_NDOTS, dnsc_ndots, 0,1000), opt("port", DNS_OPT_PORT, dnsc_port, 1,0xffff), opt("udpbuf", DNS_OPT_UDPSIZE, dnsc_udpbuf, DNS_MAXPACKET,65536), #undef opt }; #define dns_ctxopt(ctx,idx) (*((unsigned*)(((char*)ctx)+dns_opts[idx].offset))) #define ISSPACE(x) (x == ' ' || x == '\t' || x == '\r' || x == '\n') struct dns_ctx dns_defctx; #define SETCTX(ctx) if (!ctx) ctx = &dns_defctx #define SETCTXINITED(ctx) SETCTX(ctx); assert(CTXINITED(ctx)) #define CTXINITED(ctx) (ctx->dnsc_flags & DNS_INITED) #define SETCTXFRESH(ctx) SETCTXINITED(ctx); assert(!CTXOPEN(ctx)) #define SETCTXINACTIVE(ctx) \ SETCTXINITED(ctx); assert(!ctx->dnsc_nactive) #define SETCTXOPEN(ctx) SETCTXINITED(ctx); assert(CTXOPEN(ctx)) #define CTXOPEN(ctx) (ctx->dnsc_udpsock >= 0) #if defined(NDEBUG) || !defined(DEBUG) #define dns_assert_ctx(ctx) #else static void dns_assert_ctx(const struct dns_ctx *ctx) { int nactive = 0; const struct dns_query *q; for(q = ctx->dnsc_qactive.head; q; q = q->dnsq_next) { assert(q->dnsq_ctx == ctx); assert(q == (q->dnsq_next ? q->dnsq_next->dnsq_prev : ctx->dnsc_qactive.tail)); assert(q == (q->dnsq_prev ? q->dnsq_prev->dnsq_next : ctx->dnsc_qactive.head)); ++nactive; } assert(nactive == ctx->dnsc_nactive); } #endif enum { DNS_INTERNAL = 0xffff, /* internal flags mask */ DNS_INITED = 0x0001, /* the context is initialized */ DNS_ASIS_DONE = 0x0002, /* search: skip the last as-is query */ DNS_SEEN_NODATA = 0x0004, /* search: NODATA has been received */ }; int dns_add_serv(struct dns_ctx *ctx, const char *serv) { union sockaddr_ns *sns; SETCTXFRESH(ctx); if (!serv) return (ctx->dnsc_nserv = 0); if (ctx->dnsc_nserv >= DNS_MAXSERV) return errno = ENFILE, -1; sns = &ctx->dnsc_serv[ctx->dnsc_nserv]; memset(sns, 0, sizeof(*sns)); if (dns_pton(AF_INET, serv, &sns->sin.sin_addr) > 0) { sns->sin.sin_family = AF_INET; return ++ctx->dnsc_nserv; } #ifdef HAVE_IPv6 if (dns_pton(AF_INET6, serv, &sns->sin6.sin6_addr) > 0) { sns->sin6.sin6_family = AF_INET6; return ++ctx->dnsc_nserv; } #endif errno = EINVAL; return -1; } int dns_add_serv_s(struct dns_ctx *ctx, const struct sockaddr *sa) { SETCTXFRESH(ctx); if (!sa) return (ctx->dnsc_nserv = 0); if (ctx->dnsc_nserv >= DNS_MAXSERV) return errno = ENFILE, -1; #ifdef HAVE_IPv6 else if (sa->sa_family == AF_INET6) ctx->dnsc_serv[ctx->dnsc_nserv].sin6 = *(struct sockaddr_in6*)sa; #endif else if (sa->sa_family == AF_INET) ctx->dnsc_serv[ctx->dnsc_nserv].sin = *(struct sockaddr_in*)sa; else return errno = EAFNOSUPPORT, -1; return ++ctx->dnsc_nserv; } int dns_set_opts(struct dns_ctx *ctx, const char *opts) { unsigned i, v; int err = 0; SETCTXINACTIVE(ctx); for(;;) { while(ISSPACE(*opts)) ++opts; if (!*opts) break; for(i = 0; ; ++i) { if (i >= sizeof(dns_opts)/sizeof(dns_opts[0])) { ++err; break; } v = strlen(dns_opts[i].name); if (strncmp(dns_opts[i].name, opts, v) != 0 || (opts[v] != ':' && opts[v] != '=')) continue; opts += v + 1; v = 0; if (*opts < '0' || *opts > '9') { ++err; break; } do v = v * 10 + (*opts++ - '0'); while (*opts >= '0' && *opts <= '9'); if (v < dns_opts[i].min) v = dns_opts[i].min; if (v > dns_opts[i].max) v = dns_opts[i].max; dns_ctxopt(ctx, i) = v; break; } while(*opts && !ISSPACE(*opts)) ++opts; } return err; } int dns_set_opt(struct dns_ctx *ctx, enum dns_opt opt, int val) { int prev; unsigned i; SETCTXINACTIVE(ctx); for(i = 0; i < sizeof(dns_opts)/sizeof(dns_opts[0]); ++i) { if (dns_opts[i].opt != opt) continue; prev = dns_ctxopt(ctx, i); if (val >= 0) { unsigned v = val; if (v < dns_opts[i].min || v > dns_opts[i].max) { errno = EINVAL; return -1; } dns_ctxopt(ctx, i) = v; } return prev; } if (opt == DNS_OPT_FLAGS) { prev = ctx->dnsc_flags & ~DNS_INTERNAL; if (val >= 0) ctx->dnsc_flags = (ctx->dnsc_flags & DNS_INTERNAL) | (val & ~DNS_INTERNAL); return prev; } errno = ENOSYS; return -1; } int dns_add_srch(struct dns_ctx *ctx, const char *srch) { int dnl; SETCTXINACTIVE(ctx); if (!srch) { memset(ctx->dnsc_srchbuf, 0, sizeof(ctx->dnsc_srchbuf)); ctx->dnsc_srchend = ctx->dnsc_srchbuf; return 0; } dnl = sizeof(ctx->dnsc_srchbuf) - (ctx->dnsc_srchend - ctx->dnsc_srchbuf) - 1; dnl = dns_sptodn(srch, ctx->dnsc_srchend, dnl); if (dnl > 0) ctx->dnsc_srchend += dnl; ctx->dnsc_srchend[0] = '\0'; /* we ensure the list is always ends at . */ if (dnl > 0) return 0; errno = EINVAL; return -1; } static void dns_drop_utm(struct dns_ctx *ctx) { if (ctx->dnsc_utmfn) ctx->dnsc_utmfn(NULL, -1, ctx->dnsc_utmctx); ctx->dnsc_utmctx = NULL; ctx->dnsc_utmexp = -1; } static void _dns_request_utm(struct dns_ctx *ctx, time_t now) { struct dns_query *q; time_t deadline; int timeout; q = ctx->dnsc_qactive.head; if (!q) deadline = -1, timeout = -1; else if (!now || q->dnsq_deadline <= now) deadline = 0, timeout = 0; else deadline = q->dnsq_deadline, timeout = (int)(deadline - now); if (ctx->dnsc_utmexp == deadline) return; ctx->dnsc_utmfn(ctx, timeout, ctx->dnsc_utmctx); ctx->dnsc_utmexp = deadline; } static __inline void dns_request_utm(struct dns_ctx *ctx, time_t now) { if (ctx->dnsc_utmfn) _dns_request_utm(ctx, now); } void dns_set_dbgfn(struct dns_ctx *ctx, dns_dbgfn *dbgfn) { SETCTXINITED(ctx); ctx->dnsc_udbgfn = dbgfn; } void dns_set_tmcbck(struct dns_ctx *ctx, dns_utm_fn *fn, void *data) { SETCTXINITED(ctx); dns_drop_utm(ctx); ctx->dnsc_utmfn = fn; ctx->dnsc_utmctx = data; if (CTXOPEN(ctx)) dns_request_utm(ctx, 0); } static unsigned dns_nonrandom_32(void) { #if defined(_WINDOWS) || defined(WINDOWS) FILETIME ft; GetSystemTimeAsFileTime(&ft); return ft.dwLowDateTime; #else struct timeval tv; gettimeofday(&tv, NULL); return tv.tv_usec; #endif } /* This is historic deprecated API */ UDNS_API unsigned dns_random16(void); unsigned dns_random16(void) { unsigned x = dns_nonrandom_32(); return (x ^ (x >> 16)) & 0xffff; } static void dns_init_rng(struct dns_ctx *ctx) { udns_jraninit(&ctx->dnsc_jran, dns_nonrandom_32()); ctx->dnsc_nextid = 0; } void dns_close(struct dns_ctx *ctx) { struct dns_query *q, *p; SETCTX(ctx); if (CTXINITED(ctx)) { if (ctx->dnsc_udpsock >= 0) closesocket(ctx->dnsc_udpsock); ctx->dnsc_udpsock = -1; if (ctx->dnsc_pbuf) free(ctx->dnsc_pbuf); ctx->dnsc_pbuf = NULL; q = ctx->dnsc_qactive.head; while((p = q) != NULL) { q = q->dnsq_next; free(p); } qlist_init(&ctx->dnsc_qactive); ctx->dnsc_nactive = 0; dns_drop_utm(ctx); } } void dns_reset(struct dns_ctx *ctx) { SETCTX(ctx); dns_close(ctx); memset(ctx, 0, sizeof(*ctx)); ctx->dnsc_timeout = 4; ctx->dnsc_ntries = 3; ctx->dnsc_ndots = 1; ctx->dnsc_udpbuf = DNS_EDNS0PACKET; ctx->dnsc_port = DNS_PORT; ctx->dnsc_udpsock = -1; ctx->dnsc_srchend = ctx->dnsc_srchbuf; qlist_init(&ctx->dnsc_qactive); dns_init_rng(ctx); ctx->dnsc_flags = DNS_INITED; } struct dns_ctx *dns_new(const struct dns_ctx *copy) { struct dns_ctx *ctx; SETCTXINITED(copy); dns_assert_ctx(copy); ctx = malloc(sizeof(*ctx)); if (!ctx) return NULL; *ctx = *copy; ctx->dnsc_udpsock = -1; qlist_init(&ctx->dnsc_qactive); ctx->dnsc_nactive = 0; ctx->dnsc_pbuf = NULL; ctx->dnsc_qstatus = 0; ctx->dnsc_srchend = ctx->dnsc_srchbuf + (copy->dnsc_srchend - copy->dnsc_srchbuf); ctx->dnsc_utmfn = NULL; ctx->dnsc_utmctx = NULL; dns_init_rng(ctx); return ctx; } void dns_free(struct dns_ctx *ctx) { assert(ctx != NULL && ctx != &dns_defctx); dns_reset(ctx); free(ctx); } int dns_open(struct dns_ctx *ctx) { int sock; unsigned i; int port; union sockaddr_ns *sns; #ifdef HAVE_IPv6 unsigned have_inet6 = 0; #endif SETCTXINITED(ctx); assert(!CTXOPEN(ctx)); port = htons((unsigned short)ctx->dnsc_port); /* ensure we have at least one server */ if (!ctx->dnsc_nserv) { sns = ctx->dnsc_serv; sns->sin.sin_family = AF_INET; sns->sin.sin_addr.s_addr = htonl(INADDR_LOOPBACK); ctx->dnsc_nserv = 1; } for (i = 0; i < ctx->dnsc_nserv; ++i) { sns = &ctx->dnsc_serv[i]; /* set port for each sockaddr */ #ifdef HAVE_IPv6 if (sns->sa.sa_family == AF_INET6) { if (!sns->sin6.sin6_port) sns->sin6.sin6_port = (unsigned short)port; ++have_inet6; } else #endif { assert(sns->sa.sa_family == AF_INET); if (!sns->sin.sin_port) sns->sin.sin_port = (unsigned short)port; } } #ifdef HAVE_IPv6 if (have_inet6 && have_inet6 < ctx->dnsc_nserv) { /* convert all IPv4 addresses to IPv6 V4MAPPED */ struct sockaddr_in6 sin6; memset(&sin6, 0, sizeof(sin6)); sin6.sin6_family = AF_INET6; /* V4MAPPED: ::ffff:1.2.3.4 */ sin6.sin6_addr.s6_addr[10] = 0xff; sin6.sin6_addr.s6_addr[11] = 0xff; for(i = 0; i < ctx->dnsc_nserv; ++i) { sns = &ctx->dnsc_serv[i]; if (sns->sa.sa_family == AF_INET) { sin6.sin6_port = sns->sin.sin_port; memcpy(sin6.sin6_addr.s6_addr + 4*3, &sns->sin.sin_addr, 4); sns->sin6 = sin6; } } } ctx->dnsc_salen = have_inet6 ? sizeof(struct sockaddr_in6) : sizeof(struct sockaddr_in); if (have_inet6) sock = socket(PF_INET6, SOCK_DGRAM, IPPROTO_UDP); else sock = socket(PF_INET, SOCK_DGRAM, IPPROTO_UDP); #else /* !HAVE_IPv6 */ sock = socket(PF_INET, SOCK_DGRAM, IPPROTO_UDP); ctx->dnsc_salen = sizeof(struct sockaddr_in); #endif /* HAVE_IPv6 */ if (sock < 0) { ctx->dnsc_qstatus = DNS_E_TEMPFAIL; return -1; } #if defined(_WINDOWS) || defined(WINDOWS) { unsigned long on = 1; if (ioctlsocket(sock, FIONBIO, &on) == SOCKET_ERROR) { closesocket(sock); ctx->dnsc_qstatus = DNS_E_TEMPFAIL; return -1; } } #else /* !WINDOWS */ if (fcntl(sock, F_SETFL, fcntl(sock, F_GETFL) | O_NONBLOCK) < 0 || fcntl(sock, F_SETFD, FD_CLOEXEC) < 0) { closesocket(sock); ctx->dnsc_qstatus = DNS_E_TEMPFAIL; return -1; } #endif /* WINDOWS */ /* allocate the packet buffer */ if ((ctx->dnsc_pbuf = malloc(ctx->dnsc_udpbuf)) == NULL) { closesocket(sock); ctx->dnsc_qstatus = DNS_E_NOMEM; errno = ENOMEM; return -1; } ctx->dnsc_udpsock = sock; dns_request_utm(ctx, 0); return sock; } int dns_sock(const struct dns_ctx *ctx) { SETCTXINITED(ctx); return ctx->dnsc_udpsock; } int dns_active(const struct dns_ctx *ctx) { SETCTXINITED(ctx); dns_assert_ctx(ctx); return ctx->dnsc_nactive; } int dns_status(const struct dns_ctx *ctx) { SETCTX(ctx); return ctx->dnsc_qstatus; } void dns_setstatus(struct dns_ctx *ctx, int status) { SETCTX(ctx); ctx->dnsc_qstatus = status; } /* End the query: disconnect it from the active list, free it, * and return the result to the caller. */ static void dns_end_query(struct dns_ctx *ctx, struct dns_query *q, int status, void *result) { dns_query_fn *cbck = q->dnsq_cbck; void *cbdata = q->dnsq_cbdata; ctx->dnsc_qstatus = status; assert((status < 0 && result == 0) || (status >= 0 && result != 0)); assert(cbck != 0); /*XXX callback may be NULL */ assert(ctx->dnsc_nactive > 0); --ctx->dnsc_nactive; qlist_remove(&ctx->dnsc_qactive, q); /* force the query to be unconnected */ /*memset(q, 0, sizeof(*q));*/ #ifndef NDEBUG q->dnsq_ctx = NULL; #endif free(q); cbck(ctx, result, cbdata); } #define DNS_DBG(ctx, code, sa, slen, pkt, plen) \ do { \ if (ctx->dnsc_udbgfn) \ ctx->dnsc_udbgfn(code, (sa), slen, pkt, plen, 0, 0); \ } while(0) #define DNS_DBGQ(ctx, q, code, sa, slen, pkt, plen) \ do { \ if (ctx->dnsc_udbgfn) \ ctx->dnsc_udbgfn(code, (sa), slen, pkt, plen, q, q->dnsq_cbdata); \ } while(0) static void dns_newid(struct dns_ctx *ctx, struct dns_query *q) { /* this is how we choose an identifier for a new query (qID). * For now, it's just sequential number, incremented for every query, and * thus obviously trivial to guess. * There are two choices: * a) use sequential numbers. It is plain insecure. In DNS, there are two * places where random numbers are (or can) be used to increase security: * random qID and random source port number. Without this randomness * (udns uses fixed port for all queries), or when the randomness is weak, * it's trivial to spoof query replies. With randomness however, it * becomes a bit more difficult task. Too bad we only have 16 bits for * our security, as qID is only two bytes. It isn't a security per se, * to rely on those 16 bits - an attacker can just flood us with fake * replies with all possible qIDs (only 65536 of them), and in this case, * even if we'll use true random qIDs, we'll be in trouble (not protected * against spoofing). Yes, this is only possible on a high-speed network * (probably on the LAN only, since usually a border router for a LAN * protects internal machines from packets with spoofed local addresses * from outside, and usually a nameserver resides on LAN), but it's * still very well possible to send us fake replies. * In other words: there's nothing a DNS (stub) resolver can do against * spoofing attacks, unless DNSSEC is in use, which helps here alot. * Too bad that DNSSEC isn't widespread, so relying on it isn't an * option in almost all cases... * b) use random qID, based on some random-number generation mechanism. * This way, we increase our protection a bit (see above - it's very weak * still), but we also increase risk of qID reuse and matching late replies * that comes to queries we've sent before against new queries. There are * some more corner cases around that, as well - for example, normally, * udns tries to find the query for a given reply by qID, *and* by * verifying that the query DN and other parameters are also the same * (so if the new query is against another domain name, old reply will * be ignored automatically). But certain types of replies which we now * handle - for example, FORMERR reply from servers which refuses to * process EDNS0-enabled packets - comes without all the query parameters * but the qID - so we're forced to use qID only when determining which * query the given reply corresponds to. This makes us even more * vulnerable to spoofing attacks, because an attacker don't even need to * know which queries we perform to spoof the replies - he only needs to * flood us with fake FORMERR "replies". * * That all to say: using sequential (or any other trivially guessable) * numbers for qIDs is insecure, but the whole thing is inherently insecure * as well, and this "extra weakness" that comes from weak qID choosing * algorithm adds almost nothing to the underlying problem. * * It CAN NOT be made secure. Period. That's it. * Unless we choose to implement DNSSEC, which is a whole different story. * Forcing TCP mode makes it better, but who uses TCP for DNS anyway? * (and it's hardly possible because of huge impact on the recursive * nameservers). * * Note that ALL stub resolvers (again, unless they implement and enforce * DNSSEC) suffers from this same problem. * * Here, I use a pseudo-random number generator for qIDs, instead of a * simpler sequential IDs. This is _not_ more secure than sequential * ID, but some found random IDs more enjoyeable for some reason. So * here it goes. */ /* Use random number and check if it's unique. * If it's not, try again up to 5 times. */ unsigned loop; dnsc_t c0, c1; for(loop = 0; loop < 5; ++loop) { const struct dns_query *c; if (!ctx->dnsc_nextid) ctx->dnsc_nextid = udns_jranval(&ctx->dnsc_jran); c0 = ctx->dnsc_nextid & 0xff; c1 = (ctx->dnsc_nextid >> 8) & 0xff; ctx->dnsc_nextid >>= 16; for(c = ctx->dnsc_qactive.head; c; c = c->dnsq_next) if (c->dnsq_id[0] == c0 && c->dnsq_id[1] == c1) break; /* found such entry, try again */ if (!c) break; } q->dnsq_id[0] = c0; q->dnsq_id[1] = c1; /* reset all parameters relevant for previous query lifetime */ q->dnsq_try = 0; q->dnsq_servi = 0; /*XXX probably should keep dnsq_servnEDNS0 bits? * See also comments in dns_ioevent() about FORMERR case */ q->dnsq_servwait = q->dnsq_servskip = q->dnsq_servnEDNS0 = 0; } /* Find next search suffix and fills in q->dnsq_dn. * Return 0 if no more to try. */ static int dns_next_srch(struct dns_ctx *ctx, struct dns_query *q) { unsigned dnl; for(;;) { if (q->dnsq_nxtsrch > ctx->dnsc_srchend) return 0; dnl = dns_dnlen(q->dnsq_nxtsrch); if (dnl + q->dnsq_origdnl0 <= DNS_MAXDN && (*q->dnsq_nxtsrch || !(q->dnsq_flags & DNS_ASIS_DONE))) break; q->dnsq_nxtsrch += dnl; } memcpy(q->dnsq_dn + q->dnsq_origdnl0, q->dnsq_nxtsrch, dnl); if (!*q->dnsq_nxtsrch) q->dnsq_flags |= DNS_ASIS_DONE; q->dnsq_nxtsrch += dnl; dns_newid(ctx, q); /* new ID for new qDN */ return 1; } /* find the server to try for current iteration. * Note that current dnsq_servi may point to a server we should skip -- * in that case advance to the next server. * Return true if found, false if all tried. */ static int dns_find_serv(const struct dns_ctx *ctx, struct dns_query *q) { while(q->dnsq_servi < ctx->dnsc_nserv) { if (!(q->dnsq_servskip & (1 << q->dnsq_servi))) return 1; ++q->dnsq_servi; } return 0; } /* format and send the query to a given server. * In case of network problem (sendto() fails), return -1, * else return 0. */ static int dns_send_this(struct dns_ctx *ctx, struct dns_query *q, unsigned servi, time_t now) { unsigned qlen; unsigned tries; { /* format the query buffer */ dnsc_t *p = ctx->dnsc_pbuf; memset(p, 0, DNS_HSIZE); if (!(q->dnsq_flags & DNS_NORD)) p[DNS_H_F1] |= DNS_HF1_RD; if (q->dnsq_flags & DNS_AAONLY) p[DNS_H_F1] |= DNS_HF1_AA; if (q->dnsq_flags & DNS_SET_CD) p[DNS_H_F2] |= DNS_HF2_CD; p[DNS_H_QDCNT2] = 1; memcpy(p + DNS_H_QID, q->dnsq_id, 2); p = dns_payload(p); /* copy query dn */ p += dns_dntodn(q->dnsq_dn, p, DNS_MAXDN); /* query type and class */ memcpy(p, q->dnsq_typcls, 4); p += 4; /* add EDNS0 record. DO flag requires it */ if (q->dnsq_flags & DNS_SET_DO || (ctx->dnsc_udpbuf > DNS_MAXPACKET && !(q->dnsq_servnEDNS0 & (1 << servi)))) { *p++ = 0; /* empty (root) DN */ p = dns_put16(p, DNS_T_OPT); p = dns_put16(p, ctx->dnsc_udpbuf); /* EDNS0 RCODE & VERSION; rest of the TTL field; RDLEN */ memset(p, 0, 2+2+2); if (q->dnsq_flags & DNS_SET_DO) p[2] |= DNS_EF1_DO; p += 2+2+2; ctx->dnsc_pbuf[DNS_H_ARCNT2] = 1; } qlen = p - ctx->dnsc_pbuf; assert(qlen <= ctx->dnsc_udpbuf); } /* send the query */ tries = 10; while (sendto(ctx->dnsc_udpsock, (void*)ctx->dnsc_pbuf, qlen, 0, &ctx->dnsc_serv[servi].sa, ctx->dnsc_salen) < 0) { /*XXX just ignore the sendto() error for now and try again. * In the future, it may be possible to retrieve the error code * and find which operation/query failed. *XXX try the next server too? (if ENETUNREACH is returned immediately) */ if (--tries) continue; /* if we can't send the query, fail it. */ dns_end_query(ctx, q, DNS_E_TEMPFAIL, 0); return -1; } DNS_DBGQ(ctx, q, 1, &ctx->dnsc_serv[servi].sa, sizeof(union sockaddr_ns), ctx->dnsc_pbuf, qlen); q->dnsq_servwait |= 1 << servi; /* expect reply from this ns */ q->dnsq_deadline = now + (dns_find_serv(ctx, q) ? 1 : ctx->dnsc_timeout << q->dnsq_try); /* move the query to the proper place, according to the new deadline */ qlist_remove(&ctx->dnsc_qactive, q); { /* insert from the tail */ struct dns_query *p; for(p = ctx->dnsc_qactive.tail; p; p = p->dnsq_prev) if (p->dnsq_deadline <= q->dnsq_deadline) break; qlist_insert_after(&ctx->dnsc_qactive, q, p); } return 0; } /* send the query out using next available server * and add it to the active list, or, if no servers available, * end it. */ static void dns_send(struct dns_ctx *ctx, struct dns_query *q, time_t now) { /* if we can't send the query, return TEMPFAIL even when searching: * we can't be sure whenever the name we tried to search exists or not, * so don't continue searching, or we may find the wrong name. */ if (!dns_find_serv(ctx, q)) { /* no more servers in this iteration. Try the next cycle */ q->dnsq_servi = 0; /* reset */ q->dnsq_try++; /* next try */ if (q->dnsq_try >= ctx->dnsc_ntries || !dns_find_serv(ctx, q)) { /* no more servers and tries, fail the query */ /* return TEMPFAIL even when searching: no more tries for this * searchlist, and no single definitive reply (handled in dns_ioevent() * in NOERROR or NXDOMAIN cases) => all nameservers failed to process * current search list element, so we don't know whenever the name exists. */ dns_end_query(ctx, q, DNS_E_TEMPFAIL, 0); return; } } dns_send_this(ctx, q, q->dnsq_servi++, now); } static void dns_dummy_cb(struct dns_ctx *ctx, void *result, void *data) { if (result) free(result); data = ctx = 0; /* used */ } /* The (only, main, real) query submission routine. * Allocate new query structure, initialize it, check validity of * parameters, and add it to the head of the active list, without * trying to send it (to be picked up on next event). * Error return (without calling the callback routine) - * no memory or wrong parameters. *XXX The `no memory' case probably should go to the callback anyway... */ struct dns_query * dns_submit_dn(struct dns_ctx *ctx, dnscc_t *dn, int qcls, int qtyp, int flags, dns_parse_fn *parse, dns_query_fn *cbck, void *data) { struct dns_query *q; SETCTXOPEN(ctx); dns_assert_ctx(ctx); q = calloc(sizeof(*q), 1); if (!q) { ctx->dnsc_qstatus = DNS_E_NOMEM; return NULL; } #ifndef NDEBUG q->dnsq_ctx = ctx; #endif q->dnsq_parse = parse; q->dnsq_cbck = cbck ? cbck : dns_dummy_cb; q->dnsq_cbdata = data; q->dnsq_origdnl0 = dns_dntodn(dn, q->dnsq_dn, sizeof(q->dnsq_dn)); assert(q->dnsq_origdnl0 > 0); --q->dnsq_origdnl0; /* w/o the trailing 0 */ dns_put16(q->dnsq_typcls+0, qtyp); dns_put16(q->dnsq_typcls+2, qcls); q->dnsq_flags = (flags | ctx->dnsc_flags) & ~DNS_INTERNAL; if (flags & DNS_NOSRCH || dns_dnlabels(q->dnsq_dn) > ctx->dnsc_ndots) { q->dnsq_nxtsrch = flags & DNS_NOSRCH ? ctx->dnsc_srchend /* end of the search list if no search requested */ : ctx->dnsc_srchbuf /* beginning of the list, but try as-is first */; q->dnsq_flags |= DNS_ASIS_DONE; dns_newid(ctx, q); } else { q->dnsq_nxtsrch = ctx->dnsc_srchbuf; dns_next_srch(ctx, q); } /* q->dnsq_deadline is set to 0 (calloc above): the new query is * "already expired" when first inserted into queue, so it's safe * to insert it into the head of the list. Next call to dns_timeouts() * will actually send it. */ qlist_add_head(&ctx->dnsc_qactive, q); ++ctx->dnsc_nactive; dns_request_utm(ctx, 0); return q; } struct dns_query * dns_submit_p(struct dns_ctx *ctx, const char *name, int qcls, int qtyp, int flags, dns_parse_fn *parse, dns_query_fn *cbck, void *data) { int isabs; SETCTXOPEN(ctx); if (dns_ptodn(name, 0, ctx->dnsc_pbuf, DNS_MAXDN, &isabs) <= 0) { ctx->dnsc_qstatus = DNS_E_BADQUERY; return NULL; } if (isabs) flags |= DNS_NOSRCH; return dns_submit_dn(ctx, ctx->dnsc_pbuf, qcls, qtyp, flags, parse, cbck, data); } /* process readable fd condition. * To be usable in edge-triggered environment, the routine * should consume all input so it should loop over. * Note it isn't really necessary to loop here, because * an application may perform the loop just fine by it's own, * but in this case we should return some sensitive result, * to indicate when to stop calling and error conditions. * Note also we may encounter all sorts of recvfrom() * errors which aren't fatal, and at the same time we may * loop forever if an error IS fatal. */ void dns_ioevent(struct dns_ctx *ctx, time_t now) { int r; unsigned servi; struct dns_query *q; dnsc_t *pbuf; dnscc_t *pend, *pcur; void *result; union sockaddr_ns sns; socklen_t slen; SETCTX(ctx); if (!CTXOPEN(ctx)) return; dns_assert_ctx(ctx); pbuf = ctx->dnsc_pbuf; if (!now) now = time(NULL); again: /* receive the reply */ slen = sizeof(sns); r = recvfrom(ctx->dnsc_udpsock, (void*)pbuf, ctx->dnsc_udpbuf, MSG_DONTWAIT, &sns.sa, &slen); if (r < 0) { /*XXX just ignore recvfrom() errors for now. * in the future it may be possible to determine which * query failed and requeue it. * Note there may be various error conditions, triggered * by both local problems and remote problems. It isn't * quite trivial to determine whenever an error is local * or remote. On local errors, we should stop, while * remote errors should be ignored (for now anyway). */ #if defined(_WINDOWS) || defined(WINDOWS) if (WSAGetLastError() == WSAEWOULDBLOCK) #else if (errno == EAGAIN) #endif { dns_request_utm(ctx, now); return; } goto again; } pend = pbuf + r; pcur = dns_payload(pbuf); /* check reply header */ if (pcur > pend || dns_numqd(pbuf) > 1 || dns_opcode(pbuf) != 0) { DNS_DBG(ctx, -1/*bad reply*/, &sns.sa, slen, pbuf, r); goto again; } /* find the matching query, by qID */ for (q = ctx->dnsc_qactive.head; ; q = q->dnsq_next) { if (!q) { /* no more requests: old reply? */ DNS_DBG(ctx, -5/*no matching query*/, &sns.sa, slen, pbuf, r); goto again; } if (pbuf[DNS_H_QID1] == q->dnsq_id[0] && pbuf[DNS_H_QID2] == q->dnsq_id[1]) break; } /* if we have numqd, compare with our query qDN */ if (dns_numqd(pbuf)) { /* decode the qDN */ dnsc_t dn[DNS_MAXDN]; if (dns_getdn(pbuf, &pcur, pend, dn, sizeof(dn)) < 0 || pcur + 4 > pend) { DNS_DBG(ctx, -1/*bad reply*/, &sns.sa, slen, pbuf, r); goto again; } if (!dns_dnequal(dn, q->dnsq_dn) || memcmp(pcur, q->dnsq_typcls, 4) != 0) { /* not this query */ DNS_DBG(ctx, -5/*no matching query*/, &sns.sa, slen, pbuf, r); goto again; } /* here, query match, and pcur points past qDN in query section in pbuf */ } /* if no numqd, we only allow FORMERR rcode */ else if (dns_rcode(pbuf) != DNS_R_FORMERR) { /* treat it as bad reply if !FORMERR */ DNS_DBG(ctx, -1/*bad reply*/, &sns.sa, slen, pbuf, r); goto again; } else { /* else it's FORMERR, handled below */ } /* find server */ #ifdef HAVE_IPv6 if (sns.sa.sa_family == AF_INET6 && slen >= sizeof(sns.sin6)) { for(servi = 0; servi < ctx->dnsc_nserv; ++servi) if (sin6_eq(ctx->dnsc_serv[servi].sin6, sns.sin6)) break; } else #endif if (sns.sa.sa_family == AF_INET && slen >= sizeof(sns.sin)) { for(servi = 0; servi < ctx->dnsc_nserv; ++servi) if (sin_eq(ctx->dnsc_serv[servi].sin, sns.sin)) break; } else servi = ctx->dnsc_nserv; /* check if we expect reply from this server. * Note we can receive reply from first try if we're already at next */ if (!(q->dnsq_servwait & (1 << servi))) { /* if ever asked this NS */ DNS_DBG(ctx, -2/*wrong server*/, &sns.sa, slen, pbuf, r); goto again; } /* we got (some) reply for our query */ DNS_DBGQ(ctx, q, 0, &sns.sa, slen, pbuf, r); q->dnsq_servwait &= ~(1 << servi); /* don't expect reply from this serv */ /* process the RCODE */ switch(dns_rcode(pbuf)) { case DNS_R_NOERROR: if (dns_tc(pbuf)) { /* possible truncation. We can't deal with it. */ /*XXX for now, treat TC bit the same as SERVFAIL. * It is possible to: * a) try to decode the reply - may be ANSWER section is ok; * b) check if server understands EDNS0, and if it is, and * answer still don't fit, end query. */ break; } if (!dns_numan(pbuf)) { /* no data of requested type */ if (dns_next_srch(ctx, q)) { /* if we're searching, try next searchlist element, * but remember NODATA reply. */ q->dnsq_flags |= DNS_SEEN_NODATA; dns_send(ctx, q, now); } else /* else - nothing to search any more - finish the query. * It will be NODATA since we've seen a NODATA reply. */ dns_end_query(ctx, q, DNS_E_NODATA, 0); } /* we've got a positive reply here */ else if (q->dnsq_parse) { /* if we have parsing routine, call it and return whatever it returned */ /* don't try to re-search if NODATA here. For example, * if we asked for A but only received CNAME. Unless we'll * someday do recursive queries. And that's problematic too, since * we may be dealing with specific AA-only nameservers for a given * domain, but CNAME points elsewhere... */ r = q->dnsq_parse(q->dnsq_dn, pbuf, pcur, pend, &result); dns_end_query(ctx, q, r, r < 0 ? NULL : result); } /* else just malloc+copy the raw DNS reply */ else if ((result = malloc(r)) == NULL) dns_end_query(ctx, q, DNS_E_NOMEM, NULL); else { memcpy(result, pbuf, r); dns_end_query(ctx, q, r, result); } goto again; case DNS_R_NXDOMAIN: /* Non-existing domain. */ if (dns_next_srch(ctx, q)) /* more search entries exists, try them. */ dns_send(ctx, q, now); else /* nothing to search anymore. End the query, returning either NODATA * if we've seen it before, or NXDOMAIN if not. */ dns_end_query(ctx, q, q->dnsq_flags & DNS_SEEN_NODATA ? DNS_E_NODATA : DNS_E_NXDOMAIN, 0); goto again; case DNS_R_FORMERR: case DNS_R_NOTIMPL: /* for FORMERR and NOTIMPL rcodes, if we tried EDNS0-enabled query, * try w/o EDNS0. */ if (ctx->dnsc_udpbuf > DNS_MAXPACKET && !(q->dnsq_servnEDNS0 & (1 << servi))) { /* we always trying EDNS0 first if enabled, and retry a given query * if not available. Maybe it's better to remember inavailability of * EDNS0 in ctx as a per-NS flag, and never try again for this NS. * For long-running applications.. maybe they will change the nameserver * while we're running? :) Also, since FORMERR is the only rcode we * allow to be header-only, and in this case the only check we do to * find a query it belongs to is qID (not qDN+qCLS+qTYP), it's much * easier to spoof and to force us to perform non-EDNS0 queries only... */ q->dnsq_servnEDNS0 |= 1 << servi; dns_send_this(ctx, q, servi, now); goto again; } /* else we handle it the same as SERVFAIL etc */ case DNS_R_SERVFAIL: case DNS_R_REFUSED: /* for these rcodes, advance this request * to the next server and reschedule */ default: /* unknown rcode? hmmm... */ break; } /* here, we received unexpected reply */ q->dnsq_servskip |= (1 << servi); /* don't retry this server */ /* we don't expect replies from this server anymore. * But there may be other servers. Some may be still processing our * query, and some may be left to try. * We just ignore this reply and wait a bit more if some NSes haven't * replied yet (dnsq_servwait != 0), and let the situation to be handled * on next event processing. Timeout for this query is set correctly, * if not taking into account the one-second difference - we can try * next server in the same iteration sooner. */ /* try next server */ if (!q->dnsq_servwait) { /* next retry: maybe some other servers will reply next time. * dns_send() will end the query for us if no more servers to try. * Note we can't continue with the next searchlist element here: * we don't know if the current qdn exists or not, there's no definitive * answer yet (which is seen in cases above). *XXX standard resolver also tries as-is query in case all nameservers * failed to process our query and if not tried before. We don't do it. */ dns_send(ctx, q, now); } else { /* else don't do anything - not all servers replied yet */ } goto again; } /* handle all timeouts */ int dns_timeouts(struct dns_ctx *ctx, int maxwait, time_t now) { /* this is a hot routine */ struct dns_query *q; SETCTX(ctx); dns_assert_ctx(ctx); /* Pick up first entry from query list. * If its deadline has passed, (re)send it * (dns_send() will move it next in the list). * If not, this is the query which determines the closest deadline. */ q = ctx->dnsc_qactive.head; if (!q) return maxwait; if (!now) now = time(NULL); do { if (q->dnsq_deadline > now) { /* first non-expired query */ int w = (int)(q->dnsq_deadline - now); if (maxwait < 0 || maxwait > w) maxwait = w; break; } else { /* process expired deadline */ dns_send(ctx, q, now); } } while((q = ctx->dnsc_qactive.head) != NULL); dns_request_utm(ctx, now); /* update timer with new deadline */ return maxwait; } struct dns_resolve_data { int dnsrd_done; void *dnsrd_result; }; static void dns_resolve_cb(struct dns_ctx *ctx, void *result, void *data) { struct dns_resolve_data *d = data; d->dnsrd_result = result; d->dnsrd_done = 1; ctx = ctx; } void *dns_resolve(struct dns_ctx *ctx, struct dns_query *q) { time_t now; struct dns_resolve_data d; int n; SETCTXOPEN(ctx); if (!q) return NULL; assert(ctx == q->dnsq_ctx); dns_assert_ctx(ctx); /* do not allow re-resolving syncronous queries */ assert(q->dnsq_cbck != dns_resolve_cb && "can't resolve syncronous query"); if (q->dnsq_cbck == dns_resolve_cb) { ctx->dnsc_qstatus = DNS_E_BADQUERY; return NULL; } q->dnsq_cbck = dns_resolve_cb; q->dnsq_cbdata = &d; d.dnsrd_done = 0; now = time(NULL); while(!d.dnsrd_done && (n = dns_timeouts(ctx, -1, now)) >= 0) { #ifdef HAVE_POLL struct pollfd pfd; pfd.fd = ctx->dnsc_udpsock; pfd.events = POLLIN; n = poll(&pfd, 1, n * 1000); #else fd_set rfd; struct timeval tv; FD_ZERO(&rfd); FD_SET(ctx->dnsc_udpsock, &rfd); tv.tv_sec = n; tv.tv_usec = 0; n = select(ctx->dnsc_udpsock + 1, &rfd, NULL, NULL, &tv); #endif now = time(NULL); if (n > 0) dns_ioevent(ctx, now); } return d.dnsrd_result; } void *dns_resolve_dn(struct dns_ctx *ctx, dnscc_t *dn, int qcls, int qtyp, int flags, dns_parse_fn *parse) { return dns_resolve(ctx, dns_submit_dn(ctx, dn, qcls, qtyp, flags, parse, NULL, NULL)); } void *dns_resolve_p(struct dns_ctx *ctx, const char *name, int qcls, int qtyp, int flags, dns_parse_fn *parse) { return dns_resolve(ctx, dns_submit_p(ctx, name, qcls, qtyp, flags, parse, NULL, NULL)); } int dns_cancel(struct dns_ctx *ctx, struct dns_query *q) { SETCTX(ctx); dns_assert_ctx(ctx); assert(q->dnsq_ctx == ctx); /* do not allow cancelling syncronous queries */ assert(q->dnsq_cbck != dns_resolve_cb && "can't cancel syncronous query"); if (q->dnsq_cbck == dns_resolve_cb) return (ctx->dnsc_qstatus = DNS_E_BADQUERY); qlist_remove(&ctx->dnsc_qactive, q); --ctx->dnsc_nactive; dns_request_utm(ctx, 0); return 0; }