Skip to content
StratQuad

Threats

Linux kernel cgroup workqueue deadlock during controller unmount

CVE-2025-39953Medium

The Linux kernel's cgroup subsystem can deadlock when mounting and unmounting perf_event and net_prio controllers simultaneously. When a cgroup root is destroyed whilst a CSS (cgroup subsystem state) is going offline, both operations queue work to the same single-threaded workqueue, causing the root destruction to block waiting for the CSS offline to complete, whilst the CSS offline waits behind the root destruction. An attacker with the ability to mount and unmount cgroup controllers can trigger this deadlock, causing a hung task.

Plain English summary of the NVD and CISA text, checked by machine against both.

Scores and dates

CVSS 3.1
5.5
MEDIUM
EPSS score
0.00
Probability of exploitation in 30 days
EPSS percentile
4th
Share of scored CVEs at or below this one

CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H

Daily EPSS score, 3 Nov to 14 Nov, source FIRST

Detail

Affected versions

  • linux linux kernel: from 4.6, before 5.4.300
  • linux linux kernel: from 5.5, before 5.10.245
  • linux linux kernel: from 5.11, before 5.15.194
  • linux linux kernel: from 5.16, before 6.1.154
  • linux linux kernel: from 6.2, before 6.6.108
  • linux linux kernel: from 6.7, before 6.12.49
  • linux linux kernel: from 6.13, before 6.16.9
  • linux linux kernel: 6.17

Reports naming this CVE1 in total