head 1.1; branch 1.1.1; access ; symbols bind-2-20-29:1.1.1.1 ISC:1.1.1; locks ; strict; comment @# @; 1.1 date 2026.09.17.17.44.59; author christos; state Exp; branches 1.1.1.1; next ; commitid 6QdcGjuUmo5e60WG; 1.1.1.1 date 2026.09.17.17.44.59; author christos; state Exp; branches ; next ; commitid 6QdcGjuUmo5e60WG; desc @@ 1.1 log @Initial revision @ text @#!/usr/bin/env perl # Copyright (C) Internet Systems Consortium, Inc. ("ISC") # # SPDX-License-Identifier: MPL-2.0 # # This Source Code Form is subject to the terms of the Mozilla Public # License, v. 2.0. If a copy of the MPL was not distributed with this # file, you can obtain one at https://mozilla.org/MPL/2.0/. # # See the COPYRIGHT file distributed with this work for additional # information regarding copyright ownership. # test the update-policy external protocol require 5.6.0; use IO::File; use IO::Socket::UNIX; use Getopt::Long; my $path; my $typeallowed = "A"; my $pidfile = "authsock.pid"; my $timeout = 0; GetOptions("path=s" => \$path, "type=s" => \$typeallowed, "pidfile=s" => \$pidfile, "timeout=i" => \$timeout); STDOUT->autoflush(1); if (!defined($path)) { print("Usage: authsock.pl --path= --type=type --pidfile=pidfile\n"); exit(1); } unlink($path); my $server = IO::Socket::UNIX->new(Local => $path, Type => SOCK_STREAM, Listen => 8) or die "unable to create socket $path"; chmod 0777, $path; # setup our pidfile open(my $pid,">",$pidfile) or die "unable to open pidfile $pidfile"; print $pid "$$\n"; close($pid); if ($timeout != 0) { # die after the given timeout alarm($timeout); } while (my $client = $server->accept()) { $client->recv(my $buf, 8, 0); my ($version, $req_len) = unpack('N N', $buf); if ($version != 1 || $req_len < 17) { printf("Badly formatted request\n"); $client->send(pack('N', 2)); next; } $client->recv(my $buf, $req_len - 8, 0); my ($signer, $name, $addr, $type, $key, $key_data) = unpack('Z* Z* Z* Z* Z* N/a', $buf); if ($req_len != length($buf)+8) { printf("Length mismatch %u %u\n", $req_len, length($buf)+8); $client->send(pack('N', 2)); next; } printf("version=%u signer=%s name=%s addr=%s type=%s key=%s key_data_len=%u\n", $version, $signer, $name, $addr, $type, $key, length($key_data)); my $result; if ($typeallowed eq $type) { $result = 1; printf("allowed type %s == %s\n", $type, $typeallowed); } else { printf("disallowed type %s != %s\n", $type, $typeallowed); $result = 0; } $reply = pack('N', $result); $client->send($reply); } @ 1.1.1.1 log @Import bind-9.20.29 (previous was 9.20.27) BIND 9.20.29 Security Fixes [CVE-2026-19668] Prevent excessive CPU use validating crafted DNSSEC responses. a0a61dba9e A malicious authoritative server could serve a securely delegated zone whose DS and DNSKEY records carry many distinct key tags but no valid match, forcing a validating resolver into excessive key-tag matching and high CPU use for every query. BIND now bounds this work with the per-query validation limit (max-validations-per-fetch). [GL #5349] [CVE-2026-19033] Require a TSIG on every message of incoming zone transfers. 9404cd2b8c BIND 9 used to accept TSIG-signed zone transfers in which some messages were unsigned, and processed those messages before the next signature could vouch for them. It now requires a TSIG on every message of an incoming AXFR or IXFR; all modern nameserver already sign every message, so no change is expected in practice. [GL #6062] [CVE-2026-77119] Prevent a DNSSEC downgrade of secure delegations via unrelated NSEC3. 3bed9c8e9e A validating resolver could be tricked into treating a secure delegation as unsigned and accepting forged answers for names beneath it, if an attacker could inject responses to its queries. Such forged proofs are now rejected. [GL #6234] [CVE-2026-19941] Prevent forged DNSSEC-validated NXDOMAIN responses. a36bf58daf A validating resolver could accept a signed NSEC record from an unrelated zone as proof that a wildcard did not exist. An on-path attacker or malicious forwarder controlling a signed zone could therefore forge an authenticated NXDOMAIN response for a name that should resolve through a wildcard. BIND now requires the wildcard-denial and name-nonexistence proofs to be signed by the same zone. [GL #6253] [CVE-2026-19666] DNS64 with break-dnssec could cause an assertion failure. 4cec4965c4 When a "dns64" statement is configured with "break-dnssec yes" and its "exclude" list matches some but not all of the addresses in an AAAA RRset, named removes the excluded addresses from the answer instead of synthesizing new ones. If the answer being filtered had been cached together with a proof that the queried name does not exist -- which is what a wildcard match produces -- named terminated with an assertion failure. Only recursive resolvers are affected, and only when "break-dnssec yes" is in use; the answer has to come from the cache, so a server that is only authoritative cannot reach this. [GL #6301] [CVE-2026-19667] Reject negative cache records that do not fit in a dns_rdata_t. dbf08c8581 A single crafted response from a server could make a resolver cache a malformed negative entry and then terminate with an assertion failure when reading it back. Only recursive resolvers are affected, on a default configuration. [GL #6302] [CVE-2026-19662] Prevent resolver crash with cached DNSSEC proofs. c884cc1ba0 Under certain timing conditions, concurrent recursive queries could cause named to crash when cached DNSSEC NOQNAME proof data was replaced while still in use. Cached proof data is now retained until all queries using it have completed. [GL #6333] [CVE-2026-75029] Discard repeated SOA, CNAME, and DNAME records when parsing DNS messages. 0d630758c2 A DNS message could carry the same SOA, CNAME, or DNAME record many times, and named kept every copy while parsing it. With name compression those copies took up far more memory internally than in the message itself, and every later processing step had to handle all of them. named now keeps the first copy of such a record and discards identical repeats. [GL #6335] [CVE-2026-77692] Fix an unauthenticated crash on HTTPS using SIG(0) 5a24401c5c A specifically crafted HTTPS query using SIG(0) as authentication could crash named if the client closes the connection before named actually verifies the signature. This is now fixed. [GL #6343] [CVE-2026-81736] Cached HTTPS/SVCB aliases could exhaust resolver CPU. 20bbb1639a A recursive resolver that had cached a large set of interlinked HTTPS or SVCB records in alias form could be driven to do an excessive amount of work assembling a single response, because it followed every cached alias target when building the additional section. A client permitted to use recursion, together with an attacker-controlled zone used to plant the records, could repeat small queries to consume enough CPU to delay or deny service to other clients. The amount of additional processing done for one query is now bounded. [GL #6347] [CVE-2026-76163] Prevent TKEY queries from terminating named without global options. 7645138538 named could terminate unexpectedly when a remote client sent a TKEY query if the configuration did not include a global options statement. This has been fixed. ISC thanks Owais Lone (thesecguy) for reporting the issue. [GL #6357] [CVE-2026-78301] Out-of-zone records in a zone database could be served as authoritative. 72a10c3a0b When a zone database contained records for names outside the zone -- such as a delegation above the zone apex, left behind by a secondary that had accepted out-of-zone data from its primary -- the server could treat them as authoritative and answer queries for names inside the zone with that out-of-zone data instead of the zone's own. A server that was also a resolver could follow such a delegation and cache the answers of the server it named, affecting names outside the configured zone. Zone database lookups are now confined to names at or below the zone's origin. ISC would like to thank Henrique Pereira for reporting the issue. [GL #6361] [CVE-2026-80274] Crash on wildcard answers carrying both NSEC and NSEC3 proofs. 0e44451b1a When a wildcard answer arrived with both NSEC and NSEC3 records at the name proving that the queried name does not exist, the resolver could pick different records when caching the answer and when retrieving the proof, depending on the order in which the authoritative server sent them. This could terminate named with an assertion failure, fail the query with SERVFAIL, or serve a denial record other than the one that had been verified. The resolver now caches and serves the same denial record it accepted when the answer was received. ISC would like to thank hythyt for reporting the issue. [GL #6369] [CVE-2026-81563] Following HTTPS/SVCB aliases could leak resolver cache memory. 3162df369e When a recursive server answered a query for an HTTPS or SVCB record in alias form and the alias target had more than 13 records, the target records were pinned in the cache permanently instead of being released once the answer was sent. A remote party who could make the server follow such aliases to a steady stream of fresh names could grow the cache beyond the configured max-cache-size until the server was unable to resolve unrelated names. The records are now released correctly. ISC would like to thank Samy Medjahed/Ap4sh for reporting the issue. [GL #6374] New Features Add an agent skill for the isc_job/isc_async/isc_work APIs. fe32990b06 Documents when to use isc_job_run(), isc_async_run() or isc_work_enqueue(), and the contract each one imposes. No functional change. [GL !12561] Removed Features Remove unused closest encloser proof caching. abd8b5bfd8 BIND used to cache an NSEC3 closest encloser proof alongside positive wildcard answers so that a resolver could re-send it when answering from its cache. That stopped being used in BIND 9.9 (2011), when positive wildcard responses were changed to omit that NSEC3 record -- RFC 5155 requires only the next closer name proof -- and the closest encloser came to be derived during validation instead. The caching code has been unreachable ever since, so this removes it with no change in behaviour. [GL #5803] [GL !12660] Feature Changes Reject oversized and malformed DNSKEY records up front. 6c22109924 Oversized RSA key material in a DNSKEY record was only rejected after it had been converted, allocating memory proportional to the record size. Such records are now rejected before conversion, as are Ed25519 and Ed448 keys with trailing bytes that were previously silently ignored. [GL #4537] [GL !12560] Bug Fixes Prevent a crash when using both dns64 and filter-a. bce5d10d18 An assertion failure was possible when using both dns64 and the filter-a plugin simultaneously; this has been fixed. [GL #5979] [GL !12663] Fix update-policy grant external address passing. b1e955c326 Only TCP client addresses are supposed to be passed to an external handler for the associated update-policy rule, but UDP client addresses were also being passed. This could have caused the external handler to return a result it otherwise wouldn't. This has been fixed. [GL #6061] [GL !12555] Missing required NSEC3 for delegation not detected. e84ed2e9d7 A missing required NSEC3 record for an insecure delegation in a non OPTOUT range was not being detected. This has been fixed. [GL #6063] [GL !12611] Tighten EUI48 and EUI48 text parsing. ff50f2cdf1 Malformed EUI48 and EUI64 records could be accepted. This has been fixed. [GL #6082] [GL !12521] GeoIP ACL state can be stale or wrong after reload. 63baf425b3 named caches GeoIP information after looking it up, but the cached information was not invalidated when the GeoIP database was reloaded, so it could continue to be used. We now invalidate existing cached GeoIP information as part of the reloading process. [GL #6083] [GL !12662] Honor DNSSEC policy key tag ranges. b82e5834b7 When a DNSSEC policy configured a non-default tag-range, dnssec-keygen and dnssec-ksr could accept generated keys outside that range. Both tools now honor the configured minimum and maximum key tags. [GL #6091] [GL !12549] Fix double free in mdig when EDNS options are specified. af5bd0b0ff When the default_query is cloned the EDNS options need to be cloned rather than the pointer copied. The old behaviour results in a double free of the options. This has been fixed. [GL #6095] [GL !12661] Fix a crash when an IXFR falls back to AXFR with updates still pending. e34062bc7e When a secondary zone received an incremental transfer (IXFR) and the primary then caused named to fall back to a full transfer (AXFR) while some of the already-received incremental changes were still waiting to be applied, named could later crash when that transfer finished. The pending changes are now discarded correctly before the AXFR retry. [GL #6114] [GL !12624] Fix DS requests to parental agents over TLS. 55830d30f6 TLS configuration for parental agents was being ignored when sending DS requests. This has been fixed. [GL #6135] [GL !12613] Fix a crash when resolving names below a cached DNAME. b94e940f52 A recursive resolver could crash when it answered a query for a name beneath a cached DNAME while that same DNAME record was concurrently refreshed or evicted from the cache. [GL #6182] [GL !12593] Rndc-confgen -q (quiet) option is documented but doesn't work. 7e4a7ca1a7 The command line parsing in rndc-confgen was broken so rndc-confgen -q did not work. This has been fixed. [GL #6187] [GL !12575] Enforce query ACLs for redirect zones and searched DLZs. bc69876b2e Queries answered from redirect zones or searched DLZ databases did not consistently honor allow-query and allow-query-on, potentially exposing restricted DNS data to excluded clients or through excluded listening addresses. These ACLs are now enforced before redirect or DLZ data is returned. [GL #6251], #6252 [GL !12646] Check "asnum" validity in GeoIP ACLs. 28c2bfdc7b We now check the validity of autonomous system (AS) numbers when parsing GeoIP ACLs that use asnum elements at configuration time. asnum values start with an optional case-insensitive "AS" prefix, followed only by decimal digits, with no spaces or other extraneous characters. The value represented cannot exceed 2^32. [GL #6255] [GL !12511] Prevent crashes while reporting DNSSEC signing statistics. c190514f0a Servers with zone-statistics full could terminate while reporting DNSSEC signing statistics for a zone tracking adding more than four signing keys. [GL #6256] [GL !12674] Fix various nits in the netmgr code. c28cdad51b The MR consists of couple of small fixes and uncaught errors in the Network Manager. [GL #6257] [GL !12576] Fix a crash on remote-servers lists that reference themselves. aaae614f9d Since 9.21.16 and 9.20.17, a remote-servers, primaries, masters, or parental-agents list that referenced itself, directly or through another list, made named crash on startup or reconfiguration. Such references are again skipped and the remaining entries in the list are used, as in earlier versions. [GL #6287] [GL !12604] A record from outside a response policy zone could stop named. d135513b37 A response policy zone transferred from a primary can contain a record whose name lies outside the zone. Such a record could stop named, both when it arrived and again at every startup afterwards, because a secondary keeps it in its own copy of the zone. Records like this are now rejected and logged; previously one could also silently create a policy entry for an unrelated name. [GL #6304] [GL !12543] "rndc flushtree ." failed to flush the cache. 96e8b585ed rndc flushtree flushes cache data below a specified name. If the name specified is the DNS root, it should fully empty the cache, the same as rndc flush. However, there was a bug causing the command, in that case, to have no effect on the cache at all; this has been fixed. [GL #6308] [GL !12582] Invalid key-store configuration could abort the DNSSEC tools. 1d796ab072 Invalid configured key-stores named "key-directory" in configuration files could abort the DNSSEC tools. This has been fixed. [GL #6313] [GL !12653] NSEC signature set could bypass the secure-delegation check. c966177f6c When proving that a delegation is insecure, the validator bounded an NSEC record's authority by the signer of whichever RRSIG happened to come first in the record's signature set, rather than the signature that actually verified. A grandparent NSEC padded with an extra, unverifiable signature could therefore pass the check that keeps such proofs from reaching below a signed child zone. The validator now requires every signature on the NSEC to name the same signer and refuses proofs whose signature set is malformed or larger than max-validations-per-fetch allows. [GL #6321] Fix a possible nsupdate issue when using GSS-TSIG. 4ddcab2d3c The nsupdate process could terminate unexpectedly when using the GSS-TSIG mode executed with the nsupdate -g option. This has been fixed. [GL #6325] [GL !12588] Fix isccc_alist_define error paths. af1349552a If there is an out of memory error in isccc_alist_define a memory leak (the sexpr holding the key name) or a double free (value) could occur. This has been fixed. [GL #6329] [GL !12636] Check for empty 'endpoints' list. 23f58af443 Configuring an http block with endpoints {}; previously caused a crash in named. This is now rejected earlier by the configuration check. [GL #6330] [GL !12552] Named could crash with a single-element geoip sortlist. 0e996a4d3b If named was configured with a single-element sortlist containing a geoip ACL element, any matching query triggered an assertion failure. This has been fixed. [GL #6342] [GL !12583] Prevent out-of-bailiwick CNAMEs from evicting cached records. cdedd4acd5 A recursive resolver could remove valid cached records when a DNS response contained an out-of-bailiwick CNAME with the same owner name. Out-of-bailiwick data is now discarded before it can modify the cache. [GL #6345] [GL !12651] Restore periodic cleanup of stale resolver address data. 356f4013f8 Stale resolver address data could remain cached until memory pressure or an explicit flush. Correct the cleanup interval so it is removed periodically. [GL #6346] [GL !12589] Fix named-checkconf/named crash with malformed key name. 9f218f6aaf When a primary/remote-server key name was malformed, named-checkconf and named were both crashing (after warning about the invalid key name). This is now fixed. [GL #6362] [GL !12639] Fix -Wformat-truncation warning in totext_in_wks() f97c2bea40 BIND 9 failed to build with GCC 16 at -O3: rendering a WKS record as text triggered a -Wformat-truncation error, which is fatal in developer builds. The port number is now printed with a 16-bit format specifier, so the compiler can see it always fits the output buffer. [GL !12542] Fix off-by-one errors caused by magic hardcoded values. 726c6cb795 Fix off-by-one comparinson errors: "named -p http=" dropped the first digit of the given port (for example, "http=8080" selected port 80) and now uses the port as given, and "named-rrchecker -C" compared only part of the "CLASS" prefix when filtering generic class names, which was harmless in practice but is now corrected. [GL !12616] Hmac_verify() now accepts truncated HMACs only when requested. c81b111496 The hmac_verify() function incorrectly compares only up to 'sig->length' bytes, but the signature and its length should not be trusted, e.g. in case if it comes from a user query. Don't accept signatures which length isn't equal to the expected calculated HMAC length unless it is explicitly requested by the caller, e.g. for truncated TSIG [1] support. [1] https://datatracker.ietf.org/doc/html/rfc8945#name-tsig-truncation -policy [GL !12629] Prevent resolver crashes while processing DNS over TCP. 81b3b6d89f Recursive resolvers could terminate with an assertion failure while processing DNS responses over TCP under sustained traffic. The failure was observed on resolvers configured globally with forward only; the same transport path is also used by iterative resolution. This has been fixed. [GL !12537] @ text @@