CVE-2026-98866: Linux kernel use-after-free in scheduler cache group access

CVE-2026-98866 · Severity: info · Published 2026-10-18

The Linux kernel's scheduler cache-aware load balancing feature contained a use-after-free vulnerability when processes switched memory contexts. An attacker could trigger this by rapidly forking or executing processes, potentially causing kernel memory corruption or a denial of service. The fix decouples scheduler cache lookups from the process memory structure to eliminate the race condition.

Technologies: Linux Kernel. Vendors: Linux.

Quick answers

Is CVE-2026-98866 exploited in the wild?
No. Junglewise Threat Intelligence has no report of CVE-2026-98866 being exploited in the wild.
How severe is CVE-2026-98866?
Informational severity, published 18 October 2026.
Which products and versions are affected?
Linux Kernel CONFIG_SCHED_CACHE enabled; fixed in 7.2.x and later stable series.
Is there a fix for CVE-2026-98866?
Yes: fixed in Linux 7.2.x stable series and later; upstream fix commit b636fef85bda7d1bab9c0a45067ab1508d79d946. Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.
How is CVE-2026-98866 attacked?
It is attacked with local access, in an operating system.
Am I affected by CVE-2026-98866?
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

  • Fixed in: Linux 7.2.x stable series and later; upstream fix commit b636fef85bda7d1bab9c0a45067ab1508d79d946

Technical details

A use-after-free vulnerability exists in the scheduler's cache group accounting when a task's memory management structure (mm_struct) is switched during execution. The scheduler code previously accessed the cache group via mm->sched_cache_grp, but this reference could become invalid when the mm_struct was reassigned in exec_mmap() or similar operations. The fix adds a per-task sched_cache_grp pointer with independent reference counting, eliminating the dependency on mm_struct lifetime.

Affected products

  • Linux Linux kernel CONFIG_SCHED_CACHE enabled; fixed in 7.2.x and later stable series

Timeline

  • 2026-10-18: disclosed: Advisory published
  • 2026-09-21: patched: Upstream fix in linux-next

References

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