---
id: "CVE-2026-98850"
kind: "vulnerability"
aliases: ["CVE-2026-98850"]
title: "Linux kernel: packet TX ring overflow lets a CAP_NET_RAW user panic the host"
severity: "unknown"
cvss: null
epss: null
exploited: false
status: "open"
category: "infrastructure"
products: []
ecosystems: []
tags: ["infrastructure", "kernel", "linux", "self-hosted"]
not_affected_if: ["no-cap-net-raw-users"]
published_at: "2026-10-18T16:17:39+00:00"
first_seen_at: "2026-10-23T21:57:17+00:00"
updated_at: "2026-10-25T05:13:04+00:00"
sources: [{"source": "nvd", "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-98850", "first_seen_at": "2026-10-23T21:57:17+00:00"}]
---

# Linux kernel: packet TX ring overflow lets a CAP_NET_RAW user panic the host

A flaw in the Linux kernel's AF_PACKET TX ring code lets the signed byte counter wrap negative when a process keeps sending frames, which trips a kernel BUG. A local user holding CAP_NET_RAW can crash the host, and on systems set to panic on oops that takes the whole machine down.

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

## How to check

- unknown

## Fix

- Update the Linux kernel to a release that includes the upstream commit "net/packet: prevent TX_RING byte count overflow" (version unknown from the sources)

## Timeline

- 2026-10-23T21:57:17+00:00 nvd https://nvd.nist.gov/vuln/detail/CVE-2026-98850

## From the sources

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

net/packet: prevent TX_RING byte count overflow

tpacket_snd() drains every SEND_REQUEST frame in a TX ring and adds
each valid packet length to the signed int len_sum. Userspace can
recycle completed frames concurrently, so a single blocking send call
is not bounded by the ring size.

After enough successful transmissions len_sum wraps into the negative
range. In particular, 65536 65535-byte frames followed by a 65007-byte
frame produce 0xfffffdef, or -EIOCBQUEUED. sock_sendmsg_nosec() treats
that as an impossible return and triggers a BUG. On systems configured
to panic on oops, a CAP_NET_RAW holder can panic the host kernel.

Stop before the next frame would make the byte count exceed INT_MAX.
The positive short count leaves that frame pending for a later send
call.

A two-thread TPACKET_V2 reproducer triggered the BUG on an arm64 build
of Linux 7.3-rc4. With this change, the same reproducer returned the
expected positive short count without an oops.

---

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