CVE-2026-99798: io_uring: defer foreign io-wq enqueues to task_work

Published Nov 20, 2026
·
Updated

In the Linux kernel, the following vulnerability has been resolved:

iouring: defer foreign io-wq enqueues to taskwork

An IOPOLL ring without DEFERTASKRUN can be reaped by a task other than the one that submitted the I/O. If a polled request is marked REQFREISSUE, iofreebatchlist() calls ioqueueiowq() with the original tctx. Ordinary task exit can tear down tctx->iowq after draining only the tracked requests. The live request pins the task, but not its io-wq, so the foreign task can race iowqputandexit() and enqueue through a freed or NULL queue.

Forward foreign userspace enqueues to the owning task before accessing tctx->iowq. Use a dedicated taskwork callback so reissues still go straight to io-wq, and fail them if taskwork is running in cancellation mode. Preserve the existing PFKTHREAD cancellation path. This also covers the other callers of ioqueueiowq() without changing the worker-owned enqueue path.

Affected Software

6 affected componentsFixes available
linux/kernel<6.18.61
6.18.61
linux/kernel>=6.19.0<7.3.2
7.3.2
debian/linux*
debian/linux*
debian/linux<7.3.3-1
7.3.3-1
debian/linux<=6.1.176-1, <=6.1.192-2, <=6.12.107-1, <=6.12.117-1
7.3.4-17.3.5-1

Remediation

Recommended actions to resolve this vulnerability, in priority order.

  1. Upgrade

    Upgrade linux/Kernel to a version that resolves this vulnerability.

    Fixed in 6.18.61
  2. Upgrade

    Upgrade linux/Kernel to a version that resolves this vulnerability.

    Fixed in 7.3.2
  3. Upgrade

    Upgrade debian/linux to a version that resolves this vulnerability.

    Fixed in 7.3.3-1
  4. Upgrade

    Upgrade debian/linux to a version that resolves this vulnerability.

    Fixed in 7.3.4-1Fixed in 7.3.5-1

Frequently Asked Questions

1

What conditions are required to trigger this issue?

The issue requires an io_uring IOPOLL ring without DEFER_TASKRUN, where a task other than the I/O submitter reaps a polled request marked REQ_F_REISSUE. It also requires a race with teardown of the submitting task's io-wq during task exit.

2

Which systems are exposed?

Systems using the Linux kernel io_uring implementation are relevant when they use the affected IOPOLL and cross-task reaping scenario. The provided information does not state whether this configuration is enabled or used by default.

Contact

SecAlerts Pty Ltd.
132 Wickham Terrace
Fortitude Valley,
QLD 4006, Australia
info@secalerts.co
By using SecAlerts services, you agree to our services end-user license agreement. This website is safeguarded by reCAPTCHA and governed by the Google Privacy Policy and Terms of Service. All names, logos, and brands of products are owned by their respective owners, and any usage of these names, logos, and brands for identification purposes only does not imply endorsement. If you possess any content that requires removal, please get in touch with us.
© 2026 SecAlerts Pty Ltd.
ABN: 70 645 966 203, ACN: 645 966 203