Security Advisory

CVE-2025-62799: Fast DDS DATA_FRAG Heap Buffer Overflow

Last updated:
  • CVE-2025-62799

**CVE-2025-62799** is a heap buffer overflow in eProsima Fast DDS, published by NVD on 2026-02-03: one malformed RTPS DATA_FRAG packet from an unauthenticated sender writes past the payload buffer and crashes the subscriber. NVD scores it 9.8. The CVE record names 3.4.1, 3.3.1 and 2.6.11, and eProsima also ships the fix in 2.14.6 and 3.2.4.

Vulnerability summary

CVSS base score

9.8

Critical severity

EPSS score (FIRST)

0.5%

Probability of exploitation in the wild in the next 30 days

CISA KEV catalog

No

Not listed

Published

Vulnerability ID: CVE-2025-62799 · Source: NVD

Affected components

ComponentCPEVersion range
eProsima Fast DDScpe:2.3:a:eprosima:fast_dds:*:*:*:*:*:*:*:*Up to 3.4.0 (GHSA-cg8v-g888-pjh6); fixed in 3.4.1, 3.3.1 and 2.6.11 (CVE record) and in 3.2.4 and 2.14.6 (advisory and release notes); NVD's CPE ranges omit 2.7.0 to 2.14.5

What does this mean for affected devices?

NVD's description: "An unauthenticated sender can transmit a single malformed RTPS DATA_FRAG packet where fragmentSize and sampleSize are crafted to violate internal assumptions. Due to a 4-byte alignment step during fragment metadata initialization, the code writes past the end of the allocated payload buffer, causing immediate crash (DoS) and potentially enabling memory corruption (RCE risk)." NVD's Primary score is 9.8, vector CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H, with CWE-787. The CNA's CVSS v4.0 score on the record is 7.2 (attack requirements present, exploit maturity unreported) and it lists CWE-122 (NVD).

eProsima's advisory GHSA-cg8v-g888-pjh6 walks through it. The receive path does not check fragmentSize against sampleSize; with sampleSize 50003 and fragmentSize 50002 the second fragment's offset aligns from 50002 to 50004, and a 4-byte write lands past the 50003-byte buffer. The reporter reproduced it on Ubuntu 22.04 with ROS 2 Humble and Fast DDS 2.6.10 against a RELIABLE subscriber with no RTPS security, which the advisory calls the "typical default". It names the exposed population as "any Fast-DDS RELIABLE subscriber listening on UDP (e.g., ROS 2 nodes in robotics/industrial systems)".

The fixed releases are 3.4.1 (2025-11-20), 2.6.11 (2025-12-18), 3.3.1 (2026-01-29), 2.14.6 (2026-03-03) and 3.2.4 (2026-04-09); each release note lists CVE-2025-62799. The CVE record names only the first three, and NVD's CPE ranges follow it: before 2.6.11, 3.0.0 to before 3.3.1, and 3.4.0. No range covers 2.7.0 to 2.14.5, although the advisory marks everything up to 3.4.0 as affected, so a CPE-only scanner passes a vulnerable 2.14 build and flags a fixed 3.2.4 build. That answers the 2.14 question in our DDS guide's worked example: 2.14.6 carries the fix. Debian's security tracker lists fastdds as vulnerable in bookworm (2.9.1), trixie (3.1.2) and sid (3.3.0), with the fix tracked as unfixed in unstable (bug 1121097).

Fast DDS is the DDS library inside many ROS 2 robots, controllers and gateways, and eProsima's platform support table lists QNX 7.1 on arm64 and arm32 besides Linux and Android. As of 2026-11-16 no device vendor advisory names affected products, and CISA's CSAF repository has no ICS advisory for this CVE, so the version in each firmware image decides exposure.

CVE-2025-62799 is not in the CISA KEV catalog as of 2026-11-16 (catalog version 2026.11.13). FIRST EPSS gives it 0.0054, the 46.0th percentile, in the 2026-11-16 dataset. CISA's SSVC entry on the NVD record reads exploitation "none", automatable "no", technical impact "total" (CISA KEV, FIRST EPSS).

Read the Fast DDS version from the firmware SBOM rather than trusting a CPE match, and move to the fixed release on your branch. The packet is parsed before any DDS Security check, so a secured domain still needs the patch; our DDS security guide lists this CVE with the other post-2021 DDS flaws, and the memory corruption fix guide covers length checks in packet parsers. Related Fast DDS advisories: CVE-2024-28231, CVE-2023-50257 and CVE-2021-38425.

How to fix this class of issueThis advisory's weakness class is Memory corruption. Our practical guide covers how to detect, fix, verify and disclose it: How to fix memory corruption in embedded firmware.

Evidence

  • Single malformed DATA_FRAG from unauthenticated sender; 4-byte alignment writes past payload buffer; fixed 3.4.1, 3.3.1, 2.6.11; NVD Primary 9.8 (AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H), CWE-787; CNA CVSS v4.0 7.2 (AT:P, E:U), CWE-122; CPE <2.6.11, 3.0.0-<3.3.1, 3.4.0; CISA ADP SSVC none/no/total. — NVD
  • sampleSize 50003 / fragmentSize 50002, offset 50002 aligned to 50004 beyond 50003-byte buffer; tested Ubuntu 22.04, ROS 2 Humble, Fast DDS 2.6.10, RELIABLE subscriber, no RTPS security (typical default); affected <=3.4.0; patched 2.6.11, 2.14.6, 3.2.4, 3.3.1, 3.4.1. — eProsima GHSA-cg8v-g888-pjh6
  • Release dates: 3.4.1 2025-11-20, 2.6.11 2025-12-18, 3.3.1 2026-01-29, 2.14.6 2026-03-03, 3.2.4 2026-04-09; each body lists CVE-2025-62799. — eProsima Fast DDS releases
  • bookworm 2.9.1+ds-1+deb12u2, trixie 3.1.2+ds-1 and sid 3.3.0+ds-3 vulnerable; unstable unfixed, bug 1121097; fixed by commit d6dd58f (v3.4.1). — Debian security tracker
  • Platform table lists QNX 7.1 and arm32/arm64 targets. — Fast DDS PLATFORM_SUPPORT.md
  • No CISA ICS advisory in the cisagov/CSAF repository mentions CVE-2025-62799 (checked 2026-11-13). — cisagov/CSAF
  • Not in the CISA KEV catalog as of 2026-11-16 (catalog version 2026.11.13). — CISA KEV
  • EPSS 0.0054 (46.0th percentile), dataset 2026-11-16. — FIRST EPSS

Related advisories

Sources

  1. NVD
  2. eProsima Fast DDS security advisory GHSA-cg8v-g888-pjh6
  3. eProsima Fast DDS v3.4.1 release notes
  4. eProsima Fast DDS v2.14.6 release notes
  5. eProsima Fast DDS v3.2.4 release notes
  6. eProsima Fast DDS commit d6dd58f (fix)
  7. Debian security tracker: CVE-2025-62799
  8. eProsima Fast DDS platform support
  9. CISA KEV
  10. FIRST EPSS

Frequently asked questions

Is Fast DDS 2.14 affected by CVE-2025-62799?

Yes, up to 2.14.5. eProsima's advisory marks every version up to 3.4.0 as affected and lists 2.14.6, released 2026-03-03, as a fixed release, even though the CVE record names only 3.4.1, 3.3.1 and 2.6.11 (per GHSA-cg8v-g888-pjh6 and the 2.14.6 release notes).

Does DDS Security stop CVE-2025-62799?

The reporter tested a subscriber with no RTPS security, and NVD describes the sender as unauthenticated. The DATA_FRAG is parsed in the receive path, so patch rather than rely on configuration (per NVD and GHSA-cg8v-g888-pjh6).

Why does NVD score CVE-2025-62799 at 9.8?

NVD rates a network attack with no privileges and high impact on confidentiality, integrity and availability. The CNA's CVSS v4.0 score is 7.2 because it marks attack requirements present (per NVD).

Is CVE-2025-62799 exploited in the wild?

It is not in the CISA KEV catalog as of 2026-11-16 (catalog version 2026.11.13). FIRST EPSS gives it 0.0054, the 46.0th percentile, in the 2026-11-16 dataset (per CISA KEV and FIRST EPSS).