CVE-2026-99268: Linux kernel io_uring poll notification race condition

CVE-2026-99268 · Severity: info · Published 2026-11-03

The Linux kernel's io_uring subsystem uses an asynchronous I/O interface that applications rely on to manage high-performance networking and disk operations. This issue was a synchronization bug where applications using deferred task-run mode could miss wake-up notifications when I/O work became available, potentially causing delays in processing completed I/O operations. The fix ensures applications waiting on the ring are properly notified when new work arrives.

Technologies: Linux Kernel. Vendors: Linux.

Quick answers

Is CVE-2026-99268 exploited in the wild?
No. Junglewise Threat Intelligence has no report of CVE-2026-99268 being exploited in the wild.
How severe is CVE-2026-99268?
Informational severity, published 3 November 2026.
Which products and versions are affected?
Linux Kernel Versions from 5.1 forward with IORING_SETUP_DEFER_TASKRUN support.
Is there a fix for CVE-2026-99268?
A patch was published on 2026-10-30: Fix committed to stable kernel branches Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.
How is CVE-2026-99268 attacked?
It is attacked with local access, in an operating system.
Am I affected by CVE-2026-99268?
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.

Technical details

With IORING_SETUP_DEFER_TASKRUN, work completions are queued as local task work and only become completion queue events when the submitter enters the kernel. Applications waiting via poll() or epoll on the ring file descriptor could miss notifications when work was added after they had checked but before they entered the wait, because the poll_wq wake-up was only issued on CQE commitment (done by the waiting task itself). The fix wakes poll_wq when the first entry is added to the work list, using RCU callbacks to avoid lock recursion when called from wake handlers.

Affected products

  • Linux Linux kernel Versions from 5.1 forward with IORING_SETUP_DEFER_TASKRUN support

Timeline

  • 2026-11-03: disclosed
  • 2026-10-30: patched: Fix committed to stable kernel branches

References

This page as Markdown. Free to reuse under CC BY 4.0 with a link back.