# CVE-2026-99252: Linux kernel io_uring denial of service via preemption window

Source: https://junglewise.ai/threats/cve-2026-99252-in-the-linux-kernel-the-following-vulnerability-has-been-resolved  
Publisher: Junglewise Threat Intelligence, https://junglewise.ai/threats  
Published: 2026-11-03. Updated: 2026-11-03  
License: CC BY 4.0 with a link back

CVE-2026-99252 · Severity: info · Published 2026-11-03

A scheduling vulnerability in the Linux kernel's io_uring subsystem can cause a consumer task to spin indefinitely while holding a critical lock, preventing other work from progressing. An attacker with the ability to queue task work (such as via io-wq workers or IORING_OP_MSG_RING messages) can trigger this condition, leading to a local denial of service.

Technologies: [Linux kernel](https://junglewise.ai/threats/technologies/x-linux-kernel).

## Quick answers

**Is CVE-2026-99252 exploited in the wild?** No. Junglewise Threat Intelligence has no report of CVE-2026-99252 being exploited in the wild.

**How severe is CVE-2026-99252?** Informational severity, published 3 November 2026.

**Which products and versions are affected?** Linux kernel io_uring subsystem.

**Is there a fix for CVE-2026-99252?** No fixed version is recorded yet. Ensure proper prioritization and CPU affinity policies to prevent high-priority consumer tasks from running on the same CPU as lower-priority io_uring producers when possible. If you run it, Junglewise drafts a mitigation that fits your setup and applies it once you approve, with senior security engineers behind it.

**How is CVE-2026-99252 attacked?** It is attacked with local access, in an operating system.

**Am I affected by CVE-2026-99252?** Check whether you run Linux kernel at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

## How to fix


Ensure proper prioritization and CPU affinity policies to prevent high-priority consumer tasks from running on the same CPU as lower-priority io_uring producers when possible.

## Technical details

The vulnerability exists in the mpscq_push() function, which has a race between making a node the queue tail and linking it to its predecessor. A higher-priority consumer task (SCHED_FIFO) on the same CPU can repeatedly call cond_resched() and mpscq_pop() while a preempted lower-priority producer (SCHED_OTHER) is mid-push, causing the consumer to spin with the uring_lock held. The fix disables preemption during the push operation to close this window.

## Affected products

- Linux kernel io_uring subsystem

## Timeline

- 2026-11-03: disclosed

## Related threats

- [Linux kernel mmap bounds check vulnerability in mman.c](https://junglewise.ai/threats/cve-2026-58813-in-mmap-of-mman-c-there-is-a-possible-way-to-trigger-an-out-of)
- [Linux kernel TIPC double fetch information disclosure](https://junglewise.ai/threats/cve-2026-0193-in-handle-rx-msg-of-tipc-virtio-dev-there-is-a-possible-double)
- [Linux kernel xen-netback use-after-free in mcast_addr_list_free](https://junglewise.ai/threats/cve-2026-100081-in-the-linux-kernel-the-following-vulnerability-has-been)
- [Linux kernel printk kthread boot hang in SMP systems](https://junglewise.ai/threats/cve-2026-99250-in-the-linux-kernel-the-following-vulnerability-has-been-resolved)
- [Linux kernel io_uring memory corruption via list poisoning](https://junglewise.ai/threats/cve-2026-99230-in-the-linux-kernel-the-following-vulnerability-has-been-resolved)