CVE-2025-62603

Fast DDS is a C++ implementation of the DDS (Data Distribution Service) standard of the OMG (Object Management Group ). ParticipantGenericMessage is the DDS Security control-message container that carries not only the handshake but also on going security-control traffic after the handshake, such as crypto-token exchange, rekeying, re-authentication, and token delivery for newly appearing endpoints. On receive, the CDR parser is invoked first and deserializes the message_data (i .e., the DataHolderSeq) via the readParticipantGenericMessage → readDataHolderSeq path. The DataHolderSeq is parsed sequentially: a sequence count (uint32), and for each DataHolder the class_id string (e.g. DDS:Auth:PKI-DH:1.0+Req), string properties (a sequence of key/value pairs), and binary properties (a name plus an octet-vector). The parser operat es at a stateless level and does not know higher-layer state (for example, whether the handshake has already completed), s o it fully unfolds the structure before distinguishing legitimate from malformed traffic. Because RTPS permits duplicates, delays, and retransmissions, a receiver must perform at least minimal structural parsing to check identity and sequence n umbers before discarding or processing a message; the current implementation, however, does not peek only at a minimal header and instead parses the entire DataHolderSeq. As a result, prior to versions 3.4.1, 3.3.1, and 2.6.11, this parsi ng behavior can trigger an out-of-memory condition and remotely terminate the process. Versions 3.4.1, 3.3.1, and 2.6.11 p atch the issue.

Weakness

The product reads data past the end, or before the beginning, of the intended buffer.

Affected Software

Potential Mitigations

• Assume all input is malicious. Use an “accept known good” input validation strategy, i.e., use a list of acceptable inputs that strictly conform to specifications. Reject any input that does not strictly conform to specifications, or transform it into something that does.
• When performing input validation, consider all potentially relevant properties, including length, type of input, the full range of acceptable values, missing or extra inputs, syntax, consistency across related fields, and conformance to business rules. As an example of business rule logic, “boat” may be syntactically valid because it only contains alphanumeric characters, but it is not valid if the input is only expected to contain colors such as “red” or “blue.”
• Do not rely exclusively on looking for malicious or malformed inputs. This is likely to miss at least one undesirable input, especially if the code’s environment changes. This can give attackers enough room to bypass the intended validation. However, denylists can be useful for detecting potential attacks or determining which inputs are so malformed that they should be rejected outright.
• To reduce the likelihood of introducing an out-of-bounds read, ensure that you validate and ensure correct calculations for any length argument, buffer size calculation, or offset. Be especially careful of relying on a sentinel (i.e. special character such as NUL) in untrusted inputs.

Related Attack Patterns

• https://cwe.mitre.org/data/definitions/540.html

References

• https://github.com/eProsima/Fast-DDS/commit/354218514d32beac963ff5c306f1cf159ee37c5f
• https://github.com/eProsima/Fast-DDS/commit/a726e6a5daba660418d1f7c05b6f203c17747d2b
• https://github.com/eProsima/Fast-DDS/commit/ced3b6f92d928af1eae77d5fe889878128ad421a
• https://security-tracker.debian.org/tracker/CVE-2025-62603