Cloudanix VulnDB

CVE-2026-99084

None

CVSS

—

EPSS

0.1%

Published

Oct 23, 2026

CISA due date

—

Cloudanix analysis

Schedule a fix SSVC: Track

A flaw in Linux directory handling lets one process hold up many others that share the same special empty directory, potentially freezing operations across containers and system tasks.

Exploitation
Local Privilege
Attacker needs
Low Privilege
Complexity
Low
Impact
Denial Of Service
Component
OS kernel
Class
Denial of service
Remediation
Patch
Effort
Medium
Patch risk
Low

What it actually does

A local reader can acquire and hold the directory i_rwsem shared lock indefinitely across iterate_shared() on an immutable directory (nullfs root inode) by backing the getdents buffer with a faulting FUSE mapping, causing all subsequent exclusive lockers (mkdir, create, lookup, mount) to queue behind it on the shared inode.

Why it matters to the business

One process in an empty mount namespace can stall directory operations, namespace creation, and kernel-thread activity for all other users of the same root inode, producing broad denial-of-service in container and namespace-heavy environments.

How an attack unfolds

An attacker creates a FUSE mount whose server deliberately delays page faults, issues getdents on the nullfs root directory using a buffer backed by that mapping, holds the shared lock while the server sleeps, and thereby blocks any exclusive operation queued behind the reader.

How to fix it

Apply the kernel change that introduces the FOP_IMMUTABLE flag, sets it on nullfs directory file_operations, and makes iterate_dir() skip i_rwsem acquisition for directories marked immutable.

How to confirm the fix worked

Confirm the FOP_IMMUTABLE bit is present and honored in the nullfs directory file_operations, then verify that concurrent readdir on an immutable directory no longer serializes writers.

Typical attack path

  1. Locate a shared immutable directory inode (nullfs root of an empty mount namespace)
  2. Back a userspace buffer with a deliberately slow FUSE mapping
  3. Call getdents() to acquire the shared i_rwsem across iterate_shared()
  4. Force a fault so the FUSE server sleeps while holding the lock
  5. Issue any exclusive operation (mkdir, create, mount, lookup miss) from another context to observe indefinite blocking

How to detect it

Inspect kernel lockdep or trace events for extended i_rwsem contention on directory inodes; correlate with getdents syscalls whose buffers reside on FUSE mounts and with stalled namespace or mount operations in affected processes.

What your security team should do

  • Expedite deployment of the kernel patch that adds FOP_IMMUTABLE handling
  • Audit workloads that rely on empty mount namespaces or nullfs-style roots
  • Add monitoring for prolonged directory lock contention in kernel traces
  • Consider restricting unprivileged FUSE mounts until the fix is applied

The record

In the Linux kernel, the following vulnerability has been resolved: readdir: take no inode lock on an immutable directory iterate_dir() takes the directory's i_rwsem shared and holds it across ->iterate_shared(). For a directory that never has an entry and is never removed the lock keeps nothing still, it only orders every reader and every writer of that inode behind each other. For the directory of a nullfs instance that matters. The instance of the initial mount namespace is the root of every empty mount namespace and the private instance is the root of every kernel thread, so one inode is shared across users who have nothing else in common. And a reader can hold the lock for as long as it likes: back the getdents() buffer with a mapping of a file on a FUSE mount of your own, let the copy of "." and ".." fault and let the server wait. Queue an exclusive taker behind it, a mkdir() in that directory goes through start_dirop() before the read-only mount is reported, and from then on every lookup that misses the dcache in that directory, every create and every mount on it waits until the server answers. One user of an empty mount namespace stalls all the others. Add FOP_IMMUTABLE for the file operations of a directory that never changes and is never removed and let iterate_dir() skip the lock for it. The flag never changes for a file, ->f_pos is protected by f_pos_lock since directories are FMODE_ATOMIC_POS, IS_DEADDIR can't be set on such a directory and neither touch_atime() nor fsnotify take i_rwsem. Set it on the nullfs directory. The placeholder directories of libfs never have an entry either but their owners remove them, so they keep the lock.

Affected packages

Ecosystem Package Introduced Fixed