---
id: "CVE-2026-99042"
kind: "vulnerability"
aliases: ["CVE-2026-99042"]
title: "Linux kernel: io_uring connect filter can be bypassed with 24-byte IPv6 address"
severity: "unknown"
cvss: null
epss: 0.0021
exploited: false
status: "open"
category: "infrastructure"
products: []
ecosystems: []
tags: ["bpf", "infrastructure", "io_uring", "linux-kernel", "self-hosted"]
not_affected_if: ["default-config", "no-io-uring-connect-filter"]
published_at: "2026-10-23T13:17:27+00:00"
first_seen_at: "2026-10-23T21:57:33+00:00"
updated_at: "2026-10-25T12:38:50+00:00"
sources: [{"source": "nvd", "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-99042", "first_seen_at": "2026-10-23T21:57:33+00:00"}, {"source": "epss", "url": "https://api.first.org/data/v1/epss?cve=CVE-2026-99042", "first_seen_at": "2026-10-25T12:38:50+00:00"}]
---

# Linux kernel: io_uring connect filter can be bypassed with 24-byte IPv6 address

In the Linux kernel's io_uring connect path, the BPF filter reads port and address fields only when the IPv6 sockaddr is 28 bytes, so a 24-byte sockaddr_in6 makes the filter see zero values. An attacker can therefore connect to a denied IPv6 port or address that the filter was meant to block.

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

## How to check

- Run `uname -r` and compare with the fixed kernel versions listed by the source (unknown in the sources provided).
- Check whether io_uring connect filters (BPF) are used in your environment; the sources do not say how to detect this otherwise.

## Fix

- Update the Linux kernel to a release containing the upstream commit that uses SIN6_LEN_RFC2133 as the threshold (fixed version: unknown in the sources).
- The sources report that 7.2 has the same code and 7.3-rc6 was tested; the stable fixed version is unknown.

## Timeline

- 2026-10-23T21:57:33+00:00 nvd https://nvd.nist.gov/vuln/detail/CVE-2026-99042
- 2026-10-25T12:38:50+00:00 epss https://api.first.org/data/v1/epss?cve=CVE-2026-99042

## From the sources

In the Linux kernel, the following vulnerability has been resolved:

io_uring/net: populate CONNECT filter fields for 24-byte sockaddr_in6

io_connect_bpf_populate() fills connect.port and connect.v6_addr for
AF_INET6 only when addr_len >= sizeof(struct sockaddr_in6), i.e. 28
bytes.

The IPv6 connect paths accept SIN6_LEN_RFC2133 (24 bytes), which
already holds sin6_port and sin6_addr. io_connect() passes addr_len to
the socket layer unchanged.

At addr_len 24 the filter sees family AF_INET6 with port and address
zero, while the kernel connects to the real destination. A filter that
denies an IPv6 port or address is bypassed. Denying connect.port == P
rejects [::1]:P with -EACCES at addr_len 28 but connects at addr_len 24.

Use SIN6_LEN_RFC2133 as the threshold. sin6_scope_id stays zero when
absent.

The threshold mismatch was found by an AI coding assistant (Claude Code)
comparing this hook against the net/ipv6 length checks on my
instructions; it also drafted the reproducer and this patch. I verified
the bypass and the fix in a VM on 7.3-rc6 and that 7.2 has the same
code.

---

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