[CVE-2026-90103][MODERATE REGULAR] NFSv4.2: fix LAYOUTSTATS send buffer exhaustion [ Upstream

[AL-KERNEL <alkernel@kernelcve.org>]

CVE: CVE-2026-90103
Priority: MODERATE REGULAR
AL-KERNEL base severity: MODERATE
KPANIC flag: NO
Patch: NFSv4.2: fix LAYOUTSTATS send buffer exhaustion [ Upstream
Commit: 842ac26615424577f24f0486b9d3bf97f28fad9c
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=842ac26615424577f24f0486b9d3bf97f28fad9c
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-90103
Analysis date: Thu, 17 Sep 2026 13:52:35 -0400
ActionableScore: 3
ActionableScore lower bound: unknown
Actionable bucket: Borderline / manual review recommended
Manual review required: YES

Summary:
A malicious or compromised NFS server can exhaust an NFSv4.2 flexfiles LAYOUTSTATS XDR reservation and leave a client mirror lock permanently held, causing affected NFS operations to stall.

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

The original Linux kernel CVE announcement for CVE-2026-90103 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-90103

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 REGULAR
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-2026-90103	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-90103 CHECK WITH IMPACT FROM ORIG NN LOW	Maybe valid. Check manually. Hints by AL-KERNEL: The best (paranoid) CVSS is 'AV:A/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H';CWE-131;CWE-667;CWE-755;Other CVSS 'AV:A/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:L';BEST CVSS score: '6.5';DESCR 'An NFSv4.2 flexfiles client can reserve too little XDR send buffer space for a LAYOUTSTATS record because the fixed 256 byte budget does not cover the maximum server supplied filehandle and network address together with all encoded fields. A malicious or compromised NFS server can control the relevant values through LAYOUTGET and GETDEVICEINFO and can cause xdr_reserve_space to return NULL while the client is encoding layout statistics. The affected calls execute while the mirror lock is held, and the failure path can leave that lock permanently held, causing NFS operations to stall and resulting in denial of service. For the CVSS the PR:N value is used because the attacker does not need an account or privileges on the client after the client has connected to the controlled NFS server. AV:A is used because exploitation requires access through an established NFS relationship and a suitable NFSv4.2 flexfiles configuration rather than arbitrary Internet traffic. The demonstrated impact is availability loss. There is no concrete UAF, OOB write, information leak, or arbitrary write primitive in the supplied code, so confidentiality and integrity remain unaffected.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Borderline / manual review recommended (with actual score 3)  SKIP CVE-2026-90103 UNKNOWN SKIP No affected files built, so skip this CVE NO  - - unknown 	MAYBE	OOB HARDWARE LINUS  DECREASED_TO_MODERATEREG_BASED_ON_ACTIONABLESCORELESSTHAN4	-	-	checked

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

ActionableScore=3
PrimitiveClass=CLASS_C
RealUAF=NO
HostAdminRequired=NO
NamespaceScopedPrivilege=UNKNOWN
ImportantCandidate=NO
ManualReview=RECOMMENDED
AutoDowngradeAllowed=YES

## 1. ActionableScore

* Raw conservative score: 3
* Conservative score: 3
* Raw paranoid score: 3
* Paranoid score: 3
* No floor or cap was applied.
* Final recommended bucket: **Borderline / manual review recommended**::

## 2. Signal breakdown

Positive signals:

* **+1 Memory layout invariant restoration:** The 256-byte reservation is inconsistent with the maximum encoded LAYOUTSTATS record. Raising it to 384 bytes restores consistency between the reservation and possible encoded length.
* **+1 Concrete persistent kernel resource state leak:** `xdr_reserve_space()` can return `NULL` while `dss_info->mirror->lock` is held, and the failure path leaves that lock permanently held.
* **+1 Strong subsystem DoS:** The permanently held lock can indefinitely stall affected NFS client operations.

Negative signals:

* None applied. The specialized NFSv4.2 flexfiles configuration is not, by itself, sufficient for a rare-subsystem penalty.
* No host-administrator penalty applies because the evaluated attacker is the malicious or compromised NFS server, not the administrator who previously mounted the filesystem.

Material non-awarded signals:

* **Direct remote reachability +2 not awarded:** Exploitation requires an established NFS relationship and prior client-side mounting/configuration. Arbitrary unsolicited network traffic cannot reach this path.
* No UAF, stale-object use, repeated release, refcount mutation, OOB write, or controlled kernel-memory corruption is established.
* The phrase about driving the encoder past the send buffer does not establish an actual overwrite because `xdr_reserve_space()` detects exhaustion and returns `NULL`.
* No confidentiality or integrity points were awarded.
* No extra Availability point was awarded because the persistent lock and resulting NFS stall are the same demonstrated DoS consequence.
* No broad/default-subsystem point was awarded because the vulnerable path specifically requires NFSv4.2 pNFS flexfiles LAYOUTSTATS processing.
* No privilege escalation, controlled reclaim, arbitrary write, authorization bypass, or practical non-DoS exploit chain is demonstrated.

## 3. Reachability analysis

The triggering actor is a malicious or compromised NFS server selected by the client. The server controls the filehandle and network address returned through `LAYOUTGET` and `GETDEVICEINFO`, allowing it to produce an oversized flexfiles LAYOUTSTATS record.

The server-side attacker requires no account, local process, capability, or host-administrator privilege on the affected client. Therefore, `HostAdminRequired=NO`. However, a separate privileged actor normally must have mounted or configured the client to use that NFS server, so prior privileged setup exists.

The attack is not generally Internet-wide. It requires:

* an established NFS client-server relationship;
* NFSv4.2 pNFS flexfiles use;
* execution of the LAYOUTSTATS encoding path;
* attacker control of, or compromise of, the selected NFS server.

No race or narrow timing window is required. No special hardware is involved. Namespace-scoped privilege is not established or materially relevant to the server-originated trigger.

## 4. Primitive and severity interpretation

The exact primitive is an undersized XDR reservation followed by failed reservation handling while a mutex remains held. This creates a persistent lock state and a denial-of-service condition.

This is not an active structured-object UAF or fixed stale-storage UAF. There is no concrete repeated release or refcount operation. No object lifetime ends while another path retains a stale reference.

The issue is `CLASS_C` because the demonstrated behavior is an invalid error state and permanent lock retention without an independently useful memory-corruption or security-boundary primitive.

The demonstrated consequence is stalled NFS client processing. A broader system-wide stall is possible only if important workloads depend on the affected NFS path, but that is an environmental consequence rather than a new kernel exploitation primitive.

`ImportantCandidate=NO` because no corruption, ownership violation, authorization bypass, disclosure, or privilege-escalation leverage is established. `AutoDowngradeAllowed=YES` because the remaining primitive is a constrained availability failure, although manual review is recommended to determine the operational importance of the affected NFS deployment.

## 5. One-sentence report phrase

A malicious or compromised NFS server can exhaust an NFSv4.2 flexfiles LAYOUTSTATS XDR reservation and leave a client mirror lock permanently held, causing affected NFS operations to stall.

## 6. Manual review recommendation

MANUAL CHECK RECOMMENDED

The issue is DoS-only but consumes server-controlled data over an established network trust relationship, and the operational scope of the permanently held lock should be confirmed for the affected deployment.

## 7. Call-site confidence

Call-site confidence: medium

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

Patch: NFSv4.2: fix LAYOUTSTATS send buffer exhaustion [ Upstream
Commit: 842ac26615424577f24f0486b9d3bf97f28fad9c
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=842ac26615424577f24f0486b9d3bf97f28fad9c

Changed files:
  include/linux/nfs_xdr.h

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=842ac26615424577f24f0486b9d3bf97f28fad9c

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

The original Linux kernel CVE announcement for CVE-2026-90103 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-90103

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:A/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:L
  The Best / paranoid CVSS vector: AV:A/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H
  The Best / paranoid CVSS score: 6.5

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: unknown
  Paranoid ActionableScore: 3

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: NO
Published priority for this report (same as in Subject): MODERATE REGULAR

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).