Last updated 25 August 2025

HTTP/2 Rapid reset attack The HTTP/2 protocol allows clients to indicate to the server that a previous stream should be canceled by sending a RSTSTREAM frame. The protocol does not require the client and server to coordinate the cancellation in any way, the client may do it unilaterally. The client may also assume that the cancellation will take effect immediately when the server receives the RSTSTREAM frame, before any other data from that TCP connection is processed.

Abuse of this feature is called a Rapid Reset attack because it relies on the ability for an endpoint to send a RSTSTREAM frame immediately after sending a request frame, which makes the other endpoint start working and then rapidly resets the request. The request is canceled, but leaves the HTTP/2 connection open.

The HTTP/2 Rapid Reset attack built on this capability is simple: The client opens a large number of streams at once as in the standard HTTP/2 attack, but rather than waiting for a response to each request stream from the server or proxy, the client cancels each request immediately.

The ability to reset streams immediately allows each connection to have an indefinite number of requests in flight. By explicitly canceling the requests, the attacker never exceeds the limit on the number of concurrent open streams. The number of in-flight requests is no longer dependent on the round-trip time (RTT), but only on the available network bandwidth.

In a typical HTTP/2 server implementation, the server will still have to do significant amounts of work for canceled requests, such as allocating new stream data structures, parsing the query and doing header decompression, and mapping the URL to a resource. For reverse proxy implementations, the request may be proxied to the backend server before the RSTSTREAM frame is processed. The client on the other hand paid almost no costs for sending the requests. This creates an exploitable cost asymmetry between the server and the client.

Multiple software artifacts implementing HTTP/2 are affected. This advisory was originally ingested from the swift-nio-http2 repo advisory and their original conent follows.

swift-nio-http2 specific advisory swift-nio-http2 is vulnerable to a denial-of-service vulnerability in which a malicious client can create and then reset a large number of HTTP/2 streams in a short period of time. This causes swift-nio-http2 to commit to a large amount of expensive work which it then throws away, including creating entirely new Channels to serve the traffic. This can easily overwhelm an EventLoop and prevent it from making forward progress.

swift-nio-http2 1.28 contains a remediation for this issue that applies reset counter using a sliding window. This constrains the number of stream resets that may occur in a given window of time. Clients violating this limit will have their connections torn down. This allows clients to continue to cancel streams for legitimate reasons, while constraining malicious actors.

It was discovered that the jboss init script performed unsafe file handling which could result in local privilege escalation.

An XML deserialization vulnerability was discovered in slf4j's EventData, which accepts an XML serialized string and can lead to arbitrary code execution.

A vulnerability which allows for a potential privilege escalation was found in the Hibernate Validator. If a security manager is present and HV itself is allowed to access private members reflectively as per the SM's configuration, that'll allow calling code without that permission to get hold of private state. The attack vector is to declare a constraint on a private member using XML, validate an invalid instance of that type and access the private member value via ConstraintViolation#getInvalidValue().

dom4j could allow a remote attacker to execute arbitrary code on the system, caused by improper input validation in multiple methods. By sending a specially-crafted XML content, an attacker could exploit this vulnerability to execute arbitrary code on the system.

Google Guava versions 11.0 through 24.1 are vulnerable to unbounded memory allocation in the AtomicDoubleArray class (when serialized with Java serialization) and Compound Ordering class (when serialized with GWT serialization). An attacker could exploit applications that use Guava and deserialize untrusted data to cause a denial of service.

External References:

https://github.com/google/guava/wiki/CVE-2018-10237 https://groups.google.com/forum/#!topic/guava-announce/xqWALw4W1vs/discussion

Upstream Patch:

https://github.com/google/guava/commit/7ec8718f1e6e2814dabaa4b9f96b6b33a813101c

An improper handing of overflow in the UTF-8 decoder with supplementary characters can lead to an infinite loop in the decoder causing a Denial of Service. Versions Affected: Apache Tomcat 9.0.0.M9 to 9.0.7, 8.5.0 to 8.5.30, 8.0.0.RC1 to 8.0.51, and 7.0.28 to 7.0.86.

A vulnerability was found in the way RemoteMessageChannel, introduced in jboss-remoting versions 3.3.10, reads from an empty buffer. An attacker could use this flaw to cause denial of service via high CPU caused by an infinite loop.

It was found that Picketlink implementation replaces special strings for obtaining attribute values with system property values in SAML messages while parsing. An attacker can misuse this to determine values of system properties at the attacked system by formatting the SAML request ID field to the chosen system property name of his liking, obtaining the property value in "InResponseTo" field in the response.

Upstream bug (for Keycloak):

https://issues.jboss.org/browse/KEYCLOAK-4160

It was identified that web auditing, as provided by Red Hat JBoss Enterprise Application Platform 6, logged request parameters in plain text. This may include passwords used for authentication mechanisms such as BASIC and FORMAUTH. A local attacker, with access to audit logs, could compromise application/server credentials.

A class implementing the Serializable interface is free to implement the “readObject(java.io.ObjectInputStream in)” method however it chooses. This readObject method is used during the deserialization process, when constructing a java object from a serialized byte stream. It is possible to implement the method in such a way that can result in java code being executed during the deserialization of an object of this class (gadget class).

The JMS specification outlines a getObject() method on the javax.jms.ObjectMessage class. The Apache Artemis implementation of this method allows deserialization of objects, from untrusted input. There are several places where Apache Artemis uses this getObject() method. In the JMS Core client, the Artemis broker and the Artemis REST component. These Artemis components may therefore be vulnerable to a remote code execution attack. Successful exploitations of this vulnerability rely on these "gadget classes" being present on the Artemis classpath and the sender of the untrusted input being authenticated and authorized to send messages to the Artemis broker.

An issue was found in the Artemis and HornetQ .When 3rd party tool send multicast messages to Artemis/HornetQ which is configured with UDP Disovery and JGroups Discovery, Artemis/HornetQ creates a huge byte array at receiving the unexpected multicast message. So, it results in heap memory exhaustion, full GC and OutOfMemoryError in the worst case.

A flaw was found in the way the DES/3DES cipher was used as part of the TLS/SSL protocol. A man-in-the-middle attacker could use this flaw to recover some plaintext data by capturing large amounts of encrypted traffic between TLS/SSL server and client if the communication used a DES/3DES based ciphersuite.

A flaw was found in the Java logging library Apache Log4j in version 1.x . This allows a remote attacker to execute code on the server if the deployed application is configured to use JMSAppender.

In 1.x you will find that there are two places where lookups are done - that is JMSAppender.java:207 and JMSAppender.java:222 - if you set TopicBindingName or TopicConnectionFactoryBindingName to something that JNDI can handle - for example "ldap://host:port/a" JNDI will do exactly the same thing it does for 2.x - so 1.x is vulnerable, just attack vector is "safer" as it depends on configuration rather than user input

This flaw in Log4j 2.x is tracked via CVE-2021-44228

A flaw was found in JBoss Enterprise Application Platform. When role-based authorization is used for Enterprise Java Beans (EJB) access, the system does not correctly call the necessary authorization modules. This prevents Java Authorization Contract for Containers (JACC) permissions from being applied, allowing remote attackers to gain unauthorized access to EJBs.

A flaw was found in JBoss Enterprise Application Platform. The processInvocation function within the org.jboss.as.ejb3.security.AuthorizationInterceptor component incorrectly authorizes all requests when no roles are defined for an Enterprise Java Beans (EJB) method invocation. This allows attackers to bypass intended access restrictions for EJB methods, leading to unauthorized access to sensitive functionalities.

apache. Multiple issues were addressed by updating to version 2.4.27.

A denial of service flaw was found in OpenSSL 0.9.8, 1.0.1, 1.0.2 through 1.0.2h, and 1.1.0 in the way the TLS/SSL protocol defined processing of ALERT packets during a connection handshake. A remote attacker could use this flaw to make a TLS/SSL server consume an excessive amount of CPU and fail to accept connections from other clients.

It was found that a flaw in commons-collection library allowed remote code execution wherever deserialization occurs. While JBoss doesnt expose the JMXInvokerServlet by default, other interfaces where deserialization occur might be vulnerable.

Note: classes directly referenced by this flaw: InvokerTransformer, InstantiateFactory, and InstantiateTransformer

External References:

http://foxglovesecurity.com/2015/11/06/what-do-weblogic-websphere-jboss-jenkins-opennms-and-your-application-have-in-common-this-vulnerability/ https://access.redhat.com/solutions/2045023

A deserialization flaw was discovered in the jackson-databind in versions before 2.8.10 and 2.9.1, which could allow an unauthenticated user to perform code execution by sending the maliciously crafted input to the readValue method of the ObjectMapper. This issue extends the previous flaw CVE-2017-7525 by blacklisting more classes that could be used maliciously.

A flaw was found in xnio. A file descriptor leak caused by growing amounts of NIO Selector file, handled between garbage collection cycles, may allow the attacker to cause a denial of service. The highest threat from this vulnerability is to system availability.

A deserialization flaw in jackson-databind was found allowing code execution when given maliocusly crafted input to readValue method of ObjectMapper.

A flaw was discovered in wildfly versions up to 16.0.0.Final that would allow local users who are able to execute init.d script to terminate arbitrary processes on the system. An attacker could exploit this by modifying the PID file in /var/run/jboss-eap/ allowing the init.d script to terminate any process as root.

A deserialization flaw was discovered in the jackson-databind which could allow an unauthenticated user to perform code execution by sending maliciously crafted input to the readValue method of ObjectMapper. This issue extends upon the previous flaws CVE-2017-7525 and CVE-2017-15095 by blacklisting more classes that could be used maliciously.