Certain DNSSEC aspects of the DNS protocol (in RFC 4033, 4034, 4035, 6840, and related RFCs) allow remote attackers to cause a denial of service (CPU consumption) via one or more DNSSEC responses, aka the "KeyTrap" issue. One of the concerns is that, when there is a zone with many DNSKEY and RRSIG records, the protocol specification implies that an algorithm must evaluate all combinations of DNSKEY and RRSIG records.

The DNS message parsing code in `named` includes a section whose computational complexity is overly high. It does not cause problems for typical DNS traffic, but crafted queries and responses may cause excessive CPU load on the affected `named` instance by exploiting this flaw. This issue affects both authoritative servers and recursive resolvers. This issue affects BIND 9 versions 9.0.0 through 9.16.45, 9.18.0 through 9.18.21, 9.19.0 through 9.19.19, 9.9.3-S1 through 9.11.37-S1, 9.16.8-S1 through 9.16.45-S1, and 9.18.11-S1 through 9.18.21-S1.

The code that processes control channel messages sent to `named` calls certain functions recursively during packet parsing. Recursion depth is only limited by the maximum accepted packet size; depending on the environment, this may cause the packet-parsing code to run out of available stack memory, causing `named` to terminate unexpectedly. Since each incoming control channel message is fully parsed before its contents are authenticated, exploiting this flaw does not require the attacker to hold a valid RNDC key; only network access to the control channel's configured TCP port is necessary. This issue affects BIND 9 versions 9.2.0 through 9.16.43, 9.18.0 through 9.18.18, 9.19.0 through 9.19.16, 9.9.3-S1 through 9.16.43-S1, and 9.18.0-S1 through 9.18.18-S1.

By flooding the target resolver with queries exploiting this flaw an attacker can significantly impair the resolver's performance, effectively denying legitimate clients access to the DNS resolution service.

In BIND 9.5.0 -> 9.11.29, 9.12.0 -> 9.16.13, and versions BIND 9.11.3-S1 -> 9.11.29-S1 and 9.16.8-S1 -> 9.16.13-S1 of BIND Supported Preview Edition, as well as release versions 9.17.0 -> 9.17.1 of the BIND 9.17 development branch, BIND servers are vulnerable if they are running an affected version and are configured to use GSS-TSIG features. In a configuration which uses BIND's default settings the vulnerable code path is not exposed, but a server can be rendered vulnerable by explicitly setting values for the tkey-gssapi-keytab or tkey-gssapi-credential configuration options. Although the default configuration is not vulnerable, GSS-TSIG is frequently used in networks where BIND is integrated with Samba, as well as in mixed-server environments that combine BIND servers with Active Directory domain controllers. For servers that meet these conditions, the ISC SPNEGO implementation is vulnerable to various attacks, depending on the CPU architecture for which BIND was built: For named binaries compiled for 64-bit platforms, this flaw can be used to trigger a buffer over-read, leading to a server crash. For named binaries compiled for 32-bit platforms, this flaw can be used to trigger a server crash due to a buffer overflow and possibly also to achieve remote code execution. We have determined that standard SPNEGO implementations are available in the MIT and Heimdal Kerberos libraries, which support a broad range of operating systems, rendering the ISC implementation unnecessary and obsolete. Therefore, to reduce the attack surface for BIND users, we will be removing the ISC SPNEGO implementation in the April releases of BIND 9.11 and 9.16 (it had already been dropped from BIND 9.17). We would not normally remove something from a stable ESV (Extended Support Version) of BIND, but since system libraries can replace the ISC SPNEGO implementation, we have made an exception in this case for reasons of stability and security.

In BIND 9.0.0 -> 9.11.29, 9.12.0 -> 9.16.13, and versions BIND 9.9.3-S1 -> 9.11.29-S1 and 9.16.8-S1 -> 9.16.13-S1 of BIND Supported Preview Edition, as well as release versions 9.17.0 -> 9.17.11 of the BIND 9.17 development branch, when a vulnerable version of named receives a query for a record triggering the flaw described above, the named process will terminate due to a failed assertion check. The vulnerability affects all currently maintained BIND 9 branches (9.11, 9.11-S, 9.16, 9.16-S, 9.17) as well as all other versions of BIND 9.

In BIND 9.0.0 -> 9.11.21, 9.12.0 -> 9.16.5, 9.17.0 -> 9.17.3, also affects 9.9.3-S1 -> 9.11.21-S1 of the BIND 9 Supported Preview Edition, An attacker on the network path for a TSIG-signed request, or operating the server receiving the TSIG-signed request, could send a truncated response to that request, triggering an assertion failure, causing the server to exit. Alternately, an off-path attacker would have to correctly guess when a TSIG-signed request was sent, along with other characteristics of the packet and message, and spoof a truncated response to trigger an assertion failure, causing the server to exit.

Using a specially-crafted message, an attacker may potentially cause a BIND server to reach an inconsistent state if the attacker knows (or successfully guesses) the name of a TSIG key used by the server. Since BIND, by default, configures a local session key even on servers whose configuration does not otherwise make use of it, almost all current BIND servers are vulnerable. In releases of BIND dating from March 2018 and after, an assertion check in tsig.c detects this inconsistent state and deliberately exits. Prior to the introduction of the check the server would continue operating in an inconsistent state, with potentially harmful results.

A malicious actor who intentionally exploits this lack of effective limitation on the number of fetches performed when processing referrals can, through the use of specially crafted referrals, cause a recursing server to issue a very large number of fetches in an attempt to process the referral. This has at least two potential effects: The performance of the recursing server can potentially be degraded by the additional work required to perform these fetches, and The attacker can exploit this behavior to use the recursing server as a reflector in a reflection attack with a high amplification factor.

To provide fine-grained controls over the ability to use Dynamic DNS (DDNS) to update records in a zone, BIND 9 provides a feature called update-policy. Various rules can be configured to limit the types of updates that can be performed by a client, depending on the key used when sending the update request. Unfortunately, some rule types were not initially documented, and when documentation for them was added to the Administrator Reference Manual (ARM) in change #3112, the language that was added to the ARM at that time incorrectly described the behavior of two rule types, krb5-subdomain and ms-subdomain. This incorrect documentation could mislead operators into believing that policies they had configured were more restrictive than they actually were. This affects BIND versions prior to BIND 9.11.5 and BIND 9.12.3.

named in ISC BIND 9.x before 9.9.9-P5, 9.10.x before 9.10.4-P5, and 9.11.x before 9.11.0-P2 allows remote attackers to cause a denial of service (assertion failure and daemon exit) via a crafted DS resource record in an answer.

named in ISC BIND 9.x before 9.9.9-P5, 9.10.x before 9.10.4-P5, and 9.11.x before 9.11.0-P2 allows remote attackers to cause a denial of service (assertion failure and daemon exit) via a malformed response to an RTYPE ANY query.

named in ISC BIND 9.x before 9.9.9-P4, 9.10.x before 9.10.4-P4, and 9.11.x before 9.11.0-P1 allows remote attackers to cause a denial of service (assertion failure and daemon exit) via a DNAME record in the answer section of a response to a recursive query, related to db.c and resolver.c.

ISC BIND 9.1.0 through 9.8.4-P2 and 9.9.0 through 9.9.2-P2 allows remote attackers to cause a denial of service (assertion failure and daemon exit) via malformed options data in an OPT resource record.

ISC BIND 9.x before 9.9.9-P2, 9.10.x before 9.10.4-P2, and 9.11.x before 9.11.0b2, when lwresd or the named lwres option is enabled, allows remote attackers to cause a denial of service (daemon crash) via a long request that uses the lightweight resolver protocol.

ISC BIND through 9.9.9-P1, 9.10.x through 9.10.4-P1, and 9.11.x through 9.11.0b1 allows primary DNS servers to cause a denial of service (secondary DNS server crash) via a large AXFR response, and possibly allows IXFR servers to cause a denial of service (IXFR client crash) via a large IXFR response and allows remote authenticated users to cause a denial of service (primary DNS server crash) via a large UPDATE message.

named in ISC BIND 9.x before 9.9.8-P4 and 9.10.x before 9.10.3-P4 allows remote attackers to cause a denial of service (assertion failure and daemon exit) via a crafted signature record for a DNAME record, related to db.c and resolver.c.

named in ISC BIND 9.x before 9.9.8-P4 and 9.10.x before 9.10.3-P4 does not properly handle DNAME records when parsing fetch reply messages, which allows remote attackers to cause a denial of service (assertion failure and daemon exit) via a malformed packet to the rndc (aka control channel) interface, related to alist.c and sexpr.c.

buffer.c in named in ISC BIND 9.10.x before 9.10.3-P3, when debug logging is enabled, allows remote attackers to cause a denial of service (REQUIRE assertion failure and daemon exit, or daemon crash) or possibly have unspecified other impact via (1) OPT data or (2) an ECS option.

apl_42.c in ISC BIND 9.x before 9.9.8-P3, 9.9.x, and 9.10.x before 9.10.3-P3 allows remote authenticated users to cause a denial of service (INSIST assertion failure and daemon exit) via a malformed Address Prefix List (APL) record.