[CVE-2026-98858][MODERATE 7.0] io_uring: fix task_work add use-after-free with SQPOLL [ Upstream

[CVE-2026-98858][MODERATE 7.0] io_uring: fix task_work add use-after-free with SQPOLL [ Upstream

[AL-KERNEL]

CVE: CVE-2026-98858
Priority: MODERATE 7.0
AL-KERNEL base severity: MODERATE
KPANIC flag: YES
Patch: io_uring: fix task_work add use-after-free with SQPOLL [ Upstream
Commit: 42b3332efacfd2140be4cd0dedad4fb96b8ae90e
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=42b3332efacfd2140be4cd0dedad4fb96b8ae90e
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98858
Analysis date: Mon, 19 Oct 2026 05:56:11 -0400
ActionableScore: 6
ActionableScore lower bound: 5
Actionable bucket: Actionable Moderate / Strong Important candidate requiring manual review
Manual review required: YES

Summary:
A local io_uring SQPOLL race allows ring teardown to free active ctx and tctx objects while task work addition still accesses them, producing confirmed slab use-after-free reads and an 8-byte post-free write with kernel crash and unresolved local exploitation potential.

======================================================================
ABOUT THIS REPORT
======================================================================

The original Linux kernel CVE announcement for CVE-2026-98858 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98858

The original announcement does not normally provide a security severity
estimate, CVSS assessment, or enough information to determine whether the
reported kernel bug represents a practically relevant security issue.

This report was generated by AL-KERNEL, an AI-assisted Linux kernel
vulnerability analysis system developed by Alexander Larkin. It combines
an autonomous classifier with LLM-assisted technical analysis and a
separate ActionableScore mechanism.

The purpose of this report is to prioritize Linux kernel CVEs before
manual review, identify cases that require prompt investigation, and
support automatic closure of issues that are unlikely to have meaningful
security impact.

Published priority for this report: MODERATE 7.0
Manual review required: YES

A detailed explanation of the methodology and priority rules is included
at the end of this message.

======================================================================
AL-KERNEL CLASSIFICATION RESULT
======================================================================

CVE-2026-98858	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-98858 CHECK WITH IMPACT FROM ORIG NN MODERATE | AEGIS IMPACT MODERATE_KPANIC AEGIS_AS=4 LEGACY=MODERATE_KPANIC RESULT=SAME (final reconciliation agrees; DECISION: both MODERATE_KPANIC)	Maybe valid. Check manually. Hints by AL-KERNEL: The best (paranoid) CVSS is 'AV:L/AC:H/PR:L/UI:N/S:U/C:H/I:H/A:H';*CWE-416;CWE-362;*CWE-672;Other CVSS 'AV:L/AC:H/PR:L/UI:N/S:U/C:L/I:L/A:H';BEST CVSS score: '7';DESCR 'A race in the io_uring SQPOLL task work path allows the SQPOLL thread to complete the last request and release the ring while io_req_normal_work_add is still accessing the associated ctx and tctx objects. This results in confirmed slab use-after-free reads and an 8 byte write through queue_work_on, creating a memory corruption condition that can cause a kernel crash and may potentially support local privilege escalation if freed memory is reused in a controllable way. For the CVSS the PR:L is selected because a local user or process must create and operate an io_uring SQPOLL ring, while full administrative privileges are not normally required on current configurations. AC:H is selected because successful triggering depends on winning a race between task work publication, SQPOLL completion, and ring teardown. The issue is not directly network reachable. Impact is at least a local denial of service via kernel crash and in the worst case may affect confidentiality and integrity due to the confirmed use-after-free write. YES REQUIRES MANUAL CHECK. Automated impact decision path. The original SimpleNet classifier selected MODERATE_KPANIC. The primary LLM estimated paranoid CVSS 7 and the production ActionableScore returned 6 with recommendation Actionable Moderate / Strong Important candidate requiring manual review. The production rules then evaluated CVSS thresholds and ActionableScore thresholds. Recorded rule transitions were CVSS and initial triage stage MODERATE_KPANIC to IMPORTANT because the CVSS score, vector or bug class threshold rule was applied. KP and FD specialist checks stage IMPORTANT to MODERATE_KPANIC because FD false positive review allowed a downgrade. The legacy KP specialist produced a review artifact and its decision was processed by the applicable KP rules. The legacy FD or DH downgrade specialist produced a review artifact and its decision was processed by the applicable downgrade and preservation rules. Exact triggered rule markers were INCREASED_TO_HIGH_BASED_ON_GUESSCVSS and DECREASED_TO_MODERATE7_BASED_ON_FALSEPOSCHECKOFDH. The independent weaker Aegis shadow evaluated CVSS 4.2 and ActionableScore 4 with primitive class CLASS_B and manual review REQUIRED. Its independently reconciled impact was MODERATE_KPANIC. The shadow agreed with the final production impact and aggregation did not change that impact. The final reported impact is MODERATE_KPANIC.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Actionable Moderate / Strong Important candidate requiring manual review (with actual score 6)	MAYBE	READ DANGER INIT UAF IO_URING KPANIC  KPANIC INCREASED_TO_HIGH_BASED_ON_GUESSCVSS DECREASED_TO_MODERATE7_BASED_ON_FALSEPOSCHECKOFDH	-	-	checked

======================================================================
ACTIONABLESCORE ANALYSIS
======================================================================

ActionableScore=6
ActionableScoreLower=5
PrimitiveClass=CLASS_B
RealUAF=YES
HostAdminRequired=NO
NamespaceScopedPrivilege=NO
ImportantCandidate=YES
ManualReview=REQUIRED
AutoDowngradeAllowed=NO
RequiresHostUID0=NO
AuthorizationDecisionBypass=NO
ProtectedResourceImpact=NONE
B4PendingReview=NO
B4PendingConfirmed=NO

1. ActionableScore

* Raw conservative score: 5
* Conservative score: 5
* Raw paranoid score: 6
* Paranoid score: 6
* No floor or cap changed either score.
* Final recommended bucket: **Actionable Moderate / Strong Important candidate requiring manual review**::

2. Signal breakdown

Conservative signals:

* +2 Real UAF. `mpscq_push()` publishes the request, after which SQPOLL can complete it and permit `ctx` and `tctx` to be freed while `io_req_normal_work_add()` still uses them.
* +1 Local unprivileged trigger. Current io_uring SQPOLL operation does not intrinsically require host UID 0 or host-admin capabilities, subject to deployment restrictions.
* +1 High-control kernel data-plane API. The caller controls ring creation, request submission, request lifetime, and repeated SQPOLL activity relevant to the vulnerable transition.
* +1 Strong DoS. KASAN confirms post-free access, including an 8-byte write in `queue_work_on()`. On a production kernel this can produce an Oops or kernel crash.
* Total: 5.

Paranoid-only signal:

* +1 Unresolved post-free write leverage. The stale operation reaches `queue_work_on()` and writes into already freed slab storage. The destination and value are not shown to be attacker-directed, but this is materially stronger than an immediate read-only stale-pointer fault.
* Total: 6.

Not awarded:

* No controlled reclaim or attacker-shaped replacement object is demonstrated.
* No arbitrary or directed kernel write is demonstrated.
* No callback or function-pointer control is established.
* No confidentiality, protected-state integrity, or privilege-escalation consequence is proven.
* No extra Availability point is awarded because it would double-count the same crash.
* No race penalty is applied. The input does not establish that the race is one-shot, externally uncontrollable, or prohibitively difficult to repeat.
* No structured stale-object leverage point is separately awarded because controlled replacement or meaningful consumption of attacker-shaped replacement fields is not established.
* No concrete repeated release/refcount point is awarded. The patch does not show the same dangling reference receiving two distinct lifetime-ending operations.
* The first retirement can free `ctx` and `tctx`, but the later vulnerable operation is an ordinary stale access and workqueue write, not a second `put`, `free`, or refcount mutation on the exact reference.
* Paranoid repeated-refcount leverage is therefore withheld.
* FLOOR 5 for repeated release is not applicable. The conservative raw score independently reaches 5.
* NEW52 B5-R does not apply because no retry-accumulated heap OOB write exists.
* NEW53 B3-PENDING does not apply because no shared or file-backed storage write is involved.
* V32, V33, and V34 do not apply because this is not checksum-metadata corruption and no networking trust boundary exists.
* B4-PENDING does not apply because there is no authorization decision or DENY-to-ALLOW transition.

3. Reachability analysis

The evaluated attacker is an ordinary local process able to create and operate an io_uring instance with `IORING_SETUP_SQPOLL`. The process can submit requests and repeatedly exercise request completion and ring teardown behavior.

Modern SQPOLL use is not intrinsically restricted to host UID 0 or a host-admin-equivalent capability. Therefore `HostAdminRequired=NO` and `RequiresHostUID0=NO`. This assessment can change operationally if the product disables io_uring, restricts it to a configured group through `io_uring_disabled`, or applies sandbox and service policy restrictions.

No namespace-scoped capability is needed for the intrinsic interface, so `NamespaceScopedPrivilege=NO`. No prior privileged object provisioning, special hardware, network access, or unusual boot mode is required by the supplied path.

The trigger is a concurrency race between publication through `mpscq_push()`, SQPOLL completion of the final request, ring exit, and continued execution in `io_req_normal_work_add()`. Repeated attempts appear practical because the attacker can create rings and submit requests repeatedly, but the supplied evidence does not quantify trigger reliability.

4. Primitive and severity interpretation

The exact primitive is a slab use-after-free involving active structured io_uring state. SQPOLL can retire the published request and initiate teardown of the last ring reference. `io_ring_exit_work()` then frees `ctx` and stops the thread with its `tctx`, while the producer remains in `io_req_normal_work_add()` and continues accessing those objects.

This is not fixed stale storage and is not limited to a shallow destructor read. The stale path includes a confirmed 8-byte write through `queue_work_on()`. It is therefore CLASS_B active-object lifetime corruption rather than CLASS_C crash-only invalid state.

It is not CLASS_A because the patch does not establish controlled reclaim, selected overwrite destination, attacker-controlled payload, type confusion, callback influence, arbitrary read or write, or demonstrated local privilege escalation.

No concrete repeated release/refcount sequence exists. The retirement path may free the objects, but the later operation accesses and writes through stale state rather than performing a second lifetime mutation on the same reference.

The demonstrated consequence is use-after-free memory access and potential kernel crash. The technically defensible unresolved consequence is that the post-free write could affect reused slab storage and potentially provide exploitation leverage. Whether the attacker can shape reuse or turn the write into privilege escalation requires manual source and allocator analysis.

`ImportantCandidate=YES` because this is a below-host-admin active-object UAF with a confirmed post-free write, not merely a poisoned-pointer read or immediate NULL dereference. `AutoDowngradeAllowed=NO` because automatically reducing it to ordinary DoS would discard unresolved leverage from the concrete stale write. This does not prove Important severity or local privilege escalation.

5. One-sentence report phrase

A local io_uring SQPOLL race allows ring teardown to free active ctx and tctx objects while task work addition still accesses them, producing confirmed slab use-after-free reads and an 8-byte post-free write with kernel crash and unresolved local exploitation potential.

6. Manual review recommendation

MANUAL CHECK REQUIRED

The confirmed write into freed slab storage requires review of the `queue_work_on()` target, slab reuse opportunities, attacker control over ring and request timing, and whether reclaimed `tctx` or related workqueue state can provide useful corruption beyond DoS.

7. Call-site confidence

Call-site confidence: high

======================================================================
UPSTREAM PATCH SUMMARY
======================================================================

Patch: io_uring: fix task_work add use-after-free with SQPOLL [ Upstream
Commit: 42b3332efacfd2140be4cd0dedad4fb96b8ae90e
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=42b3332efacfd2140be4cd0dedad4fb96b8ae90e

Changed files:
  io_uring/io_uring.c

Diff excerpt:

Not included in this email. See the upstream URL for the full patch.

Full patch:
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=42b3332efacfd2140be4cd0dedad4fb96b8ae90e

======================================================================
DETAILED REPORT METHODOLOGY
======================================================================

The original Linux kernel CVE announcement for CVE-2026-98858 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98858

The original CVE announcement normally does not include a security-level
estimate. In particular, it may not contain a CVSS assessment, an impact
level, or enough information to determine whether the reported bug is a
practically relevant security issue. One purpose of this parallel CVE list
is to provide that missing technical and prioritization information.

The original goal of the AL-KERNEL project was to prioritize Linux kernel
CVE analysis automatically before manual review. The system can also help
identify non-security issues that may be suitable for automatic closure.

This report was generated by AL-KERNEL, an AI-assisted Linux kernel
vulnerability analysis system developed by Alexander Larkin.

The first analysis stage combines an autonomous classifier with additional
LLM-based analysis. The autonomous classifier runs locally on a CPU and is
based on a backpropagation neural network. Together, these mechanisms
produce a technical vulnerability description, identify likely weakness
types, estimate CVSS severity, and provide input for ActionableScore.

Two CVSS estimates are retained because incomplete kernel vulnerability
information often permits more than one defensible interpretation:

  Conservative CVSS vector: AV:L/AC:H/PR:L/UI:N/S:U/C:L/I:L/A:H
  The Best / paranoid CVSS vector: AV:L/AC:H/PR:L/UI:N/S:U/C:H/I:H/A:H
  The Best / paranoid CVSS score: 7

The conservative vector represents a lower-impact interpretation.
The Best/paranoid vector intentionally represents a plausible upper-bound
interpretation and should not automatically be treated as demonstrated
real-world impact.

CVSS may also need to be adjusted for a particular Linux deployment,
because actual reachability, privileges, enabled kernel configuration,
hardware, namespaces, exposed device nodes, and other environmental
conditions can differ significantly between systems.

A separate ActionableScore mechanism evaluates practical remediation
urgency. Its analysis may include reachability, attack prerequisites,
subsystem exposure, memory-corruption characteristics, denial-of-service
reliability, and possible confidentiality, integrity, or
privilege-escalation impact.

  Conservative ActionableScore: 5
  Paranoid ActionableScore: 6

The final base severity is taken directly from the second tab-separated
field of the AL-KERNEL classification result. ActionableScore does not
replace or independently override that final AL-KERNEL decision, and
if ActionableScore adjusted impact level of ALKERNEL, then you would see
self-readable flags above like INCREASED_TO_HIGH_BASED_ON_ACTIONABLESCOREHIGHEREQTHAN7.

For an AL-KERNEL result of MODERATE, this report uses the following
additional presentation split:

  ActionableScore below 5   -> MODERATE REGULAR
  ActionableScore 5 or more -> MODERATE 7.0

The distinction between MODERATE REGULAR and MODERATE 7.0 makes it
possible to identify Moderate issues that should receive manual analysis
and fixes before lower-priority MODERATE REGULAR issues. In many cases,
MODERATE REGULAR fixes may wait for a later rebase or routine update.

There is one override in which MODERATE REGULAR becomes MODERATE 7.0
even when the ActionableScore is below 5. When the AL-KERNEL result
contains the KPANIC flag, a MODERATE result is always presented as
MODERATE 7.0. The KPANIC flag selected with few regexps without
usage of AI at all, so it helps to detect cases when Kernel Crash happens
and similar (to filter False-Negative results from the LLM usage).

KPANIC indicates that a reliable kernel crash, kernel panic, or similarly
serious kernel availability impact was identified by the classification
workflow.

AL-KERNEL base severity for this report: MODERATE
KPANIC detected for this report: YES
Published priority for this report (same as in Subject): MODERATE 7.0

These results are intended to support engineering triage. They are
machine-generated estimates, and cases marked for manual review should
be validated by a human security engineer before final disposition.
For more info read docs linked from here: https://kernelcve.org/
(and you can submit you own patch there to generate such a report
for non-existant CVE-id yet).

Note that in many cases this AI tool selects higher severity, than
real is (means you can expect Importants instead of Moderate 7.0 or
Moderates 7.0 instead of regular Moderates). If you see such cases,
please use reply email interface to add additional manual analyses
info to this particular CVE.
And please, please, let me know when you see Lows instead of Importants
or Important instead of Low (because particular for such cases I
need to tune this AI tool to make it better for this one and next similar).
My contact email for such notifications is alexanjelausa@gmail.com (and both
send reply to CVE record itself too and see "reply" button below for howto reply).