CVE-2026-99658: mm/mmu_notifier: finish non-blocking interval invalidation on error
In the Linux kernel, the following vulnerability has been resolved:
mm/mmunotifier: finish non-blocking interval invalidation on error
When an ordinary MMU notifier rejects a non-blocking invalidation, the interval invalidation started beforehand is left unfinished: the caller skips mmunotifierinvalidaterangeend(). Subsequent interval notifier operations can then wait forever.
The only in-tree trigger is nouveau SVM, which uses both notifier types and rejects non-blocking invalidations. Reaping via the OOM reaper or processmrelease() can therefore leave the GPU fault worker stuck, holding a reference that prevents the dying process's address space and memory from being freed.
End the interval invalidation when an ordinary notifier fails a non-blocking invalidation, like mnitreeinvalidate() already does when an interval notifier fails. Make that handling clearer in the documentation of mmunotifierinvalidaterangestartnonblock().
Fix generated by an LLM and sent by Josef Bacik in private.
[ david: simplify description, update function doc ]
Affected Software
Remediation
Recommended actions to resolve this vulnerability, in priority order.
- Upgrade
Upgrade
linux/Kernelto a version that resolves this vulnerability.Fixed in 6.12.118 - Upgrade
Upgrade
linux/Kernelto a version that resolves this vulnerability.Fixed in 6.18.64 - Upgrade
Upgrade
linux/Kernelto a version that resolves this vulnerability.Fixed in 7.3.5 - Upgrade
Upgrade
debian/linuxto a version that resolves this vulnerability.Fixed in 7.3.5-1 - Compensating control
When an ordinary MMU notifier rejects a non-blocking invalidation, call mmu_notifier_invalidate_range_end() so the interval invalidation is completed.
Event History
Frequently Asked Questions
Which systems are realistically exposed to this issue?
The only in-tree trigger identified is nouveau SVM. Systems using the Nouveau driver’s shared virtual memory support, where both ordinary and interval MMU notifiers are used, are the relevant exposure case.
What condition is needed to trigger the stuck state?
An ordinary MMU notifier must reject a non-blocking invalidation after an interval invalidation has started. During OOM reaping or process_mrelease(), this can leave the GPU fault worker waiting indefinitely while it retains a reference to the dying process’s address space.
What is the practical impact if the issue occurs?
The GPU fault worker can become stuck, retaining a reference that prevents the affected process address space and its memory from being freed. This can interfere with cleanup of a dying process.