Skip to main content
Vulnerability Database/CVE-2026-98706

CVE-2026-98706: Linux Kernel Privilege Escalation Vulnerability

CVE-2026-98706 is a privilege escalation flaw in the Linux kernel mac80211 subsystem affecting station teardown processes. This vulnerability could allow unauthorized elevation of privileges through improper handling of power-save delivery work. This article covers technical details, affected versions, impact assessment, and mitigation strategies.

Published:

Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.

Sources: National Vulnerability Database, CISA KEV, Exploit-DB

CVE-2026-98706 Overview

CVE-2026-98706 is a race condition in the Linux kernel's mac80211 wireless subsystem. The flaw occurs during access point (AP) or mesh station teardown, where power-save (PS) delivery work can run after the station has already been removed from the driver. A delayed sta_deliver_ps_frames() worker that passed the sta->dead check can resume after teardown, updating num_sta_ps and power-save queues after cleanup has purged them. The regression traces back to commit d34ba2168a3c ("mac80211: don't delay station destruction"), which changed the ordering guarantees previously provided by the shared workqueue.

Critical Impact

A worker thread can touch driver-managed state and TXQs after station removal, leading to inconsistent power-save accounting and potential kernel instability in Wi-Fi AP or mesh deployments.

Affected Products

  • Linux kernel mac80211 subsystem (wireless stack)
  • Linux-based systems operating as Wi-Fi access points
  • Linux-based mesh networking nodes using mac80211

Discovery Timeline

  • 2026-10-18 - CVE-2026-98706 published to NVD
  • 2026-10-21 - Last updated in NVD database

Technical Details for CVE-2026-98706

Vulnerability Analysis

The vulnerability lives in the station teardown path of mac80211. When an AP or mesh station is removed, the kernel schedules a delivery worker (drv_deliver_wk) to flush any buffered power-save frames. The worker checks sta->dead once at entry, then proceeds with full delivery. If the worker is preempted after the check, teardown can concurrently set sta->dead, remove the station from the driver, and purge power-save queues. When the worker resumes, it still executes the delivery sequence against state that no longer exists.

This mismatch allows the driver to be notified and TXQs to be woken after station removal. It also permits updates to num_sta_ps and power-save queues after cleanup has already drained them, corrupting PS accounting.

Root Cause

The existing cancel_work_sync() call sits near the end of __cleanup_single_sta(), after the final driver state transition and after PS accounting and queue purging. By that point, the state the worker should respect has already changed. Before commit d34ba2168a3c, cleanup was deferred through the same ordered workqueue that handled delivery, which implicitly serialized the two paths. Removing that deferral exposed the race.

Attack Vector

The condition triggers during normal station removal in AP or mesh mode and does not require authenticated input from a remote attacker. However, repeated station association and disassociation events, including those driven by hostile Wi-Fi clients, can increase the likelihood of hitting the window. The fix cancels the delivery worker after setting sta->dead and before moving the driver state, then retains a second cancellation at the start of __cleanup_single_sta() to catch work requeued through ieee80211_sta_block_awake() and to cover insertion failures that reach cleanup without setting sta->dead.

See the upstream patches for implementation details: Kernel Commit 097460f, Kernel Commit 67a9e80, Kernel Commit 772bd89, Kernel Commit 843d3a1, and Kernel Commit cbff71f.

Detection Methods for CVE-2026-98706

Indicators of Compromise

  • Kernel warnings or WARN_ON traces originating from mac80211 station teardown paths.
  • Unexpected changes in num_sta_ps counters following station disassociation events.
  • TXQ wake events logged against stations that have already been removed from the driver.

Detection Strategies

  • Audit running kernel versions across Linux hosts acting as Wi-Fi APs or mesh nodes and compare against the fixed stable branches referenced in the upstream commits.
  • Enable mac80211 debug tracing and monitor for station delivery work executing after sta_remove driver callbacks.
  • Correlate dmesg output with hostapd or wpa_supplicant association logs to catch races during high disassociation churn.

Monitoring Recommendations

  • Centralize kernel logs from Linux-based wireless infrastructure and alert on mac80211 oops, soft lockups, or scheduler stalls.
  • Track Wi-Fi station lifecycle events to identify abnormal association and disassociation storms that increase race exposure.
  • Monitor system health metrics such as kernel panics and unexpected reboots on devices running affected kernels.

How to Mitigate CVE-2026-98706

Immediate Actions Required

  • Inventory Linux kernel versions on systems operating as Wi-Fi APs, mesh nodes, or any role that activates the mac80211 AP/mesh code paths.
  • Apply the stable kernel updates containing the referenced commits as soon as distribution packages are available.
  • Prioritize patching on infrastructure exposed to untrusted wireless clients where association churn is high.

Patch Information

The fix is distributed across the stable branches via commits 097460f, 67a9e80, 772bd89, 843d3a1, and cbff71f. The patch cancels drv_deliver_wk after setting sta->dead and before moving the driver state, and keeps the existing cancellation at the start of __cleanup_single_sta() to drain any work requeued through ieee80211_sta_block_awake(). Consult your Linux distribution's security advisories for the specific backported kernel package versions.

Workarounds

  • No upstream workaround is documented; applying the kernel patch is the supported remediation.
  • Where patching is delayed, restrict wireless client access to trusted devices to reduce association churn on affected APs.
  • Consider temporarily disabling AP or mesh mode on affected hosts if the role is not essential until patched kernels are deployed.
bash
# Verify running kernel version and plan upgrade
uname -r
# Example upgrade on Debian/Ubuntu-based systems
sudo apt update && sudo apt install --only-upgrade linux-image-$(uname -r | sed 's/.*-//')
sudo reboot

Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.

Sources: National Vulnerability Database, CISA KEV, Exploit-DB

Experience the Most Advanced Cybersecurity Platform

See how the world’s most intelligent, autonomous cybersecurity platform can protect your organization today and into the future.