Description

TIER IV Nebula through 1.2.0 contains an out-of-bounds read vulnerability in the Vlp32Decoder::unpack() function that allows unauthenticated remote attackers to cause the decoder to read past the end of a received UDP buffer into adjacent heap memory by sending a short UDP datagram. Attackers can send a malformed datagram to the Velodyne UDP sensor port, which lacks sender-address restrictions present in other drivers, causing fabricated points derived from heap memory contents to be silently published into downstream PointCloud2 messages consumed by Autoware nodes.

Executive Summary

Technical Details

Mitigations

References

Additional Metadata

{
  "score": "7.5",
  "pubDate": "2026-08-17T19:16:42.713Z",
  "pubdate": "2026-08-17T19:16:42.713Z",
  "executiveSummary": "TIER IV Nebula through version 1.2.0 suffers from an out-of-bounds read vulnerability located within the Vlp32Decoder::unpack() function. This flaw permits unauthenticated remote attackers to force the decoder to read past the allocated bounds of a received UDP buffer into adjacent heap memory. Exploitation is achieved by transmitting a short, malformed UDP datagram directly to the Velodyne UDP sensor port. Because this port lacks the sender-address validation mechanisms present in other drivers, the system processes the input and silently incorporates fabricated points derived from unauthorized heap memory contents into downstream PointCloud2 messages utilized by Autoware nodes. The risk implications include potential exposure of sensitive heap data, compromise of perception pipeline data integrity, and downstream autonomous driving behavioral anomalies. Attack requirements are minimal, needing only network connectivity to the target UDP port and the transmission of a crafted short datagram without requiring prior authentication or elevated privileges.",
  "technicalDetails": "The vulnerability resides in the Vlp32Decoder::unpack() function of TIER IV Nebula through version 1.2.0, which processes incoming Velodyne LiDAR UDP data streams. The root cause stems from insufficient bounds checking when parsing incoming packet data lengths within the decoder logic. When an unauthenticated remote attacker transmits a short UDP datagram to the Velodyne UDP sensor port, the unpacking routine attempts to parse expected data structures based on fixed protocol offsets rather than validating the actual payload size of the received buffer.\nNetwork exposure is direct, as the Velodyne UDP sensor port listens for incoming datagrams and lacks sender-address restrictions that are typically implemented in alternative drivers to enforce strict packet origin validation. Consequently, any network-adjacent or remote attacker capable of routing UDP traffic to the target sensor port can initiate the attack flow without providing credentials or possessing prior privileges.\nThe exploitation method involves the payload behavior of sending a deliberately truncated or malformed UDP packet. As Vlp32Decoder::unpack() iterates through the parsing logic assuming a standard packet layout, the lack of input length verification causes the pointer arithmetic to advance beyond the termination boundary of the allocated UDP receive buffer. This results in an out-of-bounds read operation targeting adjacent heap memory regions.\nThe post-exploitation impact and data flow continue as the decoder extracts these arbitrary bytes from adjacent heap memory and translates them into spatial coordinate structures. These fabricated points are then silently published into downstream PointCloud2 messages. Because Autoware perception and localization nodes consume these PointCloud2 messages directly, the injection of memory-derived artifacts corrupts the spatial environment model, potentially leading to incorrect obstacle perception, localization degradation, or safety hazards in autonomous vehicle operations."
}