---
id: "CVE-2026-114408"
kind: "vulnerability"
aliases: ["CVE-2026-114408"]
title: "Eclipse ThreadX NetX Duo: forged ICMPv6 messages crash network stack"
severity: "high"
cvss: 8.7
epss: null
exploited: false
status: "open"
category: "runtime"
products: []
ecosystems: []
tags: ["embedded", "iot", "ipv6", "netx-duo", "runtime", "threadx"]
not_affected_if: ["ipv6-disabled", "non-netx-stack", "pmtu-discovery-disabled"]
published_at: "2026-11-02T16:17:27+00:00"
first_seen_at: "2026-11-02T16:24:04+00:00"
updated_at: "2026-11-02T16:26:18+00:00"
sources: [{"source": "nvd", "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-114408", "first_seen_at": "2026-11-02T16:24:04+00:00"}]
---

# Eclipse ThreadX NetX Duo: forged ICMPv6 messages crash network stack

NetX Duo's IPv6 path MTU handling does not check whether it added a destination entry, so a flood of forged ICMPv6 Packet Too Big messages can leave a NULL pointer dereference. On targets that fault on address 0, the IP helper thread crashes and all network processing stops.

_Summary, check and fix written by a model from the sources; labelled as such._

## How to check

- Check whether the device runs Eclipse ThreadX NetX Duo in versions `6.4.5.202504` through `6.5.2.202603a`.
- Check whether IPv6 path MTU discovery is enabled in the build configuration.

## Fix

- Eclipse ThreadX NetX Duo: unknown (the sources do not name a fixed version)

## Timeline

- 2026-11-02T16:24:04+00:00 nvd https://nvd.nist.gov/vuln/detail/CVE-2026-114408

## Documents

- [advisory] Vendor advisory: github.com https://github.com/eclipse-threadx/netxduo/security/advisories/GHSA-xjgr-2wmf-f78j

## From the sources

In Eclipse ThreadX NetX Duo versions 6.4.5.202504 through 6.5.2.202603a, built with IPv6 path MTU discovery, the ICMPv6 Packet Too Big handler uses the destination-table entry it adds without checking that the add succeeded. A sender on the network can fill the destination table with Packet Too Big messages quoting new destinations, after which the next message makes the IP helper thread dereference a NULL pointer, crashing it on targets that fault on address 0 and stopping all network processing.

---

_Free to read at a human pace. An API with keys and an MCP server are in preparation; early access: hello@infrastack.ai_