[CVE-2026-90099][MODERATE 7.0] net/sched: account classifier filter allocations to memcg [ Upstream

[CVE-2026-90099][MODERATE 7.0] net/sched: account classifier filter allocations to memcg [ Upstream

[AL-KERNEL]

CVE: CVE-2026-90099
Priority: MODERATE 7.0
AL-KERNEL base severity: MODERATE
KPANIC flag: YES
Patch: net/sched: account classifier filter allocations to memcg [ Upstream
Commit: 18c68c8f3d3ff0a4def1391021b439c4333a6c71
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=18c68c8f3d3ff0a4def1391021b439c4333a6c71
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-90099
Analysis date: Thu, 17 Sep 2026 19:43:07 -0400
ActionableScore: 6
ActionableScore lower bound: 5
Actionable bucket: Actionable Moderate / Strong Important candidate requiring manual review
Manual review required: YES

Summary:
No one-sentence report phrase was found.

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

The original Linux kernel CVE announcement for CVE-2026-90099 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-90099

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-2026-90099	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-90099 CHECK WITH IMPACT FROM ORIG NN LOW	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-770;**CWE-400;*CWE-416;Other CVSS 'AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H';BEST CVSS score: '7';DESCR 'Traffic control classifier allocation paths used GFP_KERNEL instead of GFP_KERNEL_ACCOUNT, so a local user with access to a user and network namespace could create many tc filters and pin kernel memory outside memcg charging. This can bypass cgroup memory limits and grow system slab usage until the host suffers memory pressure or denial of service. For the CVSS the PR:L is used because reliable triggering requires local code execution with tc filter creation capability in a network namespace or CAP_NET_ADMIN, not remote packet delivery. The issue is not directly network reachable and is triggered through the tc control plane. In addition to resource exhaustion, cls_basic had an error path where an object inserted into an IDR could be freed without idr_remove after a percpu allocation failure, leaving a dangling pointer and making the issue a UAF candidate under memory pressure.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Actionable Moderate / Strong Important candidate requiring manual review (with actual score 6)	YES	REMOTE BPF ERRORPATH NETWORK HARDWARE LINUS  KPANIC INCREASED_FROM_LOW_BASED_ON_REQUIREMANUALCHECK KPANIC INCREASED_TO_HIGH_BASED_ON_ACTIONABLESCOREHIGHERTHAN5 DECREASED_TO_MODERATE7_BASED_ON_FALSEPOSCHECKOFKP	NO	NO	checked

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

ActionableScore=6
ActionableScoreLower=5
PrimitiveClass=CLASS_B
RealUAF=YES
HostAdminRequired=NO
NamespaceScopedPrivilege=YES
ImportantCandidate=YES
ManualReview=REQUIRED
AutoDowngradeAllowed=NO

1. ActionableScore

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

2. Signal breakdown

Conservative positive signals:

* +1 Local unprivileged trigger. The supplied context explicitly says an unprivileged user in a fresh user and network namespace can reproduce the issue with tc filters.
* +1 Cross-boundary isolation or policy bypass. The main bug bypasses memcg charging and allows kernel slab memory to be pinned outside cgroup memory limits.
* +1 Resource exhaustion with host impact. The supplied context says system slab grows much faster than memory.current, creating realistic host memory pressure or DoS from a constrained workload.
* +2 Real UAF / lifetime violation. In cls_basic, fnew is inserted into the IDR before alloc_percpu. On allocation failure the error path kfree()s fnew without idr_remove(), leaving a dangling pointer in the IDR.

Conservative negative signals:

* No -2 host-admin penalty. Host-admin-equivalent privilege is not required by the supplied reproduction conditions.
* No -1 race penalty. This is not a timing race. It is an accounting bypass plus an ordered error-path lifetime bug.
* No niche subsystem penalty. tc classifiers are optional, but the supplied evidence shows attacker-instantiable objects in namespaces.

Paranoid additional signal:

* +1 Structured stale-object concern. The freed cls_basic filter object remains reachable through an IDR handle, which is an active kernel lookup structure. Later handle-based classifier operations may consume the stale object as structured state. Controlled reclaim or arbitrary read/write is not demonstrated, so this remains Class B rather than Class A.

Important non-awarded signals:

* Controlled reclaim is not demonstrated.
* Arbitrary write is not demonstrated.
* Confidentiality and integrity are not awarded as direct impacts.
* Remote reachability is not established. This is tc control-plane abuse, not packet-driven network input.
* Concrete repeated release/refcount was not awarded. The patch shows kfree followed by a dangling IDR reference, not a second put/free/refcount mutation on the exact same reference.
* Weak LPE is not separately awarded because the unresolved concern is already represented by the structured stale-object Class B signal.
* Extra availability is not awarded beyond the resource-exhaustion DoS signal.

Repeated release/refcount checklist:

* First release can end lifetime: yes, kfree(fnew) ends the object lifetime.
* Exact same reference survives: the object pointer remains indirectly reachable via IDR.
* Second operation is a real lifetime/refcount mutation: not established.
* Concrete repeated release/refcount +1 awarded: no.
* Paranoid repeated-refcount leverage +1 awarded: no.
* FLOOR 5 applied: no explicit floor needed. Conservative raw score is already 5.

3. Reachability analysis

The supplied context establishes below-host-admin reachability: an unprivileged user can create a fresh user and network namespace with unshare -Urn, obtain the relevant namespace-scoped network administration capability, and create tc filters on a namespace-local device such as lo.

The COMPLETE resource-exhaustion trigger is namespace-scoped according to the supplied reproduction conditions: CONFIG_USER_NS, CONFIG_NET_NS, CONFIG_NET_SCHED, CONFIG_NET_CLS_*, CONFIG_NET_CLS_ACT, and CONFIG_MEMCG are required, but no host root operation is described for the attacker after namespace creation.

For cls_basic, the attacker can also influence the allocation-failure path by operating under memcg memory.max after GFP_KERNEL_ACCOUNT is used. The commit explicitly says this makes the previously dead path attacker-reachable and leaves a dangling IDR pointer if idr_remove is missing.

No direct network trigger is supported. Packets are not the trigger. The trigger is local tc/rtnetlink control-plane configuration. No special hardware is required. No hard race is required.

4. Primitive and severity interpretation

The primary demonstrated issue is a memcg accounting bypass for tc classifier allocations. This is a resource-isolation violation that can allow a namespace-contained workload to pin kernel memory outside its memory cgroup accounting and pressure the host.

The secondary issue is stronger: cls_basic can publish fnew in an IDR and then free it on alloc_percpu failure without removing the IDR entry. That creates a real stale pointer to a freed active structured object. The stale storage is not just fixed async buffer storage. It is an object reachable via a kernel lookup table and associated with classifier handle operations.

This is not proven Class A because the patch does not demonstrate controlled reclaim, arbitrary kernel read/write, callback control, refcount takeover, credential manipulation, or a complete LPE chain. It is Class B because an active-object lifetime violation and dangling registry pointer are directly established.

Demonstrated consequences are memcg bypass and DoS through resource exhaustion. The unresolved consequence is whether the dangling IDR pointer can be turned into useful post-free structured-object corruption beyond crash or handle poisoning. That unresolved lifetime primitive makes Important-vs-Moderate review necessary.

ImportantCandidate=YES because a below-host-admin attacker can reach a concrete active-object dangling pointer in addition to a memcg isolation bypass. AutoDowngradeAllowed=NO because automatically closing this as ordinary resource exhaustion would lose the unresolved Class B stale-object primitive.

Operationally, this should be treated as Actionable Moderate at minimum, with manual review for possible Important classification.

5. One-sentence report phrase

A local unprivileged user with user and network namespace access can create many tc classifier filters to pin kernel memory outside memcg accounting, and cls_basic also contains an attacker-reachable error path that can leave a freed filter object reachable through an IDR handle.

6. Manual review recommendation

MANUAL CHECK REQUIRED

Reason: the memcg bypass alone is actionable resource-isolation DoS, and the cls_basic path establishes a real dangling IDR pointer to freed classifier state with unresolved post-free exploitation leverage.

7. Call-site confidence

Call-site confidence: high

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

Patch: net/sched: account classifier filter allocations to memcg [ Upstream
Commit: 18c68c8f3d3ff0a4def1391021b439c4333a6c71
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=18c68c8f3d3ff0a4def1391021b439c4333a6c71

Changed files:
  net/sched/cls_api.c
  net/sched/cls_basic.c
  net/sched/cls_bpf.c
  net/sched/cls_cgroup.c
  net/sched/cls_flow.c
  net/sched/cls_flower.c
  net/sched/cls_fw.c
  net/sched/cls_matchall.c
  net/sched/cls_route.c
  net/sched/cls_u32.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=18c68c8f3d3ff0a4def1391021b439c4333a6c71

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

The original Linux kernel CVE announcement for CVE-2026-90099 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-90099

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:L/PR:L/UI:N/S:U/C:N/I:N/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).