[CVE-2026-93104][MODERATE 7.0] RDMA/rvt: Return NULL after port allocation failure [ Upstream

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

CVE: CVE-2026-93104
Priority: MODERATE 7.0
AL-KERNEL base severity: MODERATE
KPANIC flag: YES
Patch: RDMA/rvt: Return NULL after port allocation failure [ Upstream
Commit: b1e1b25723af44328524c7b54433e4bf0ab9930c
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=b1e1b25723af44328524c7b54433e4bf0ab9930c
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-93104
Analysis date: Thu, 17 Sep 2026 23:31:06 -0400
ActionableScore: 5
ActionableScore lower bound: unknown
Actionable bucket: Actionable Moderate / Strong Important candidate requiring manual review
Manual review required: YES

Summary:
`rvt_alloc_device() can return a freed RDMA device pointer after port allocation failure, causing callers to dereference released kernel memory and creating a local use after free with at least kernel crash impact.`

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

The original Linux kernel CVE announcement for CVE-2026-93104 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-93104

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-2026-93104	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-93104 CHECK WITH IMPACT FROM ORIG NN IMPORTANT	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-672;CWE-754;Other CVSS 'AV:L/AC:H/PR:H/UI:N/S:U/C:N/I:N/A:H';BEST CVSS score: '7';DESCR 'rvt_alloc_device can deallocate the IB device after a port array allocation failure but still return the released pointer to callers. Callers treat any non NULL result as a valid rvt_dev_info object and may dereference freed memory, which creates an explicit use after free in the RDMA rdmavt initialization path. For the CVSS the PR:L in the paranoid score reflects a local user influence scenario where an attacker can create memory pressure or otherwise race driver initialization in an environment where RDMA device setup is performed by system services. The issue is not network reachable because it is triggered through local driver allocation and initialization failure handling rather than RDMA packet processing. Impact is at least denial of service via kernel crash and in the worst case may allow confidentiality or integrity impact due to use after free memory corruption, so manual review is recommended.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Actionable Moderate / Strong Important candidate requiring manual review (with actual score 5)	YES	UAF NETWORK HARDWARE LINUS KPANIC  KPANIC DECREASED_TO_MODERATE70_BASED_ON_ACTIONABLESCORELOWERTHAN6	NO	NO	checked

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

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

## 1. ActionableScore

* Raw conservative score: 3
* Conservative score: 5
* Raw paranoid score: 3
* Paranoid score: 5
* Applied conservative floor/cap: Real UAF FLOOR 5 applied because a freed active RDMA device allocation is returned as a valid object to callers, not merely consumed in a local cleanup fault path.
* Applied paranoid floor/cap: Same FLOOR 5 applied. No extra paranoid point was added because controlled reclaim, repeated refcount mutation, arbitrary write, or demonstrated LPE is not established.
* Final recommended bucket: **Actionable Moderate / Strong Important candidate requiring manual review**::

## 2. Signal breakdown

Positive signals:

* +2 Real UAF / lifetime violation: `rvt_alloc_device()` calls `ib_dealloc_device(&rdi->ibdev)` after port allocation failure but then returns `rdi`, a pointer to released allocation.
* +1 Reliable kernel crash / strong DoS: commit states callers treat any non-NULL return as valid and dereference it, resulting in use-after-free and likely kernel fault.
* FLOOR 5: the stale pointer escapes an allocation helper and is returned to external callers as a valid active `rvt_dev_info` object. This is broader than a purely local destructor-only stale dereference.

Non-awarded signals:

* +0 Remote reachability: not established. The bug is in local RDMA device allocation/init error handling, not packet processing.
* +0 Local unprivileged trigger: not established. A local user may influence memory pressure, but the complete trigger path through RDMA device initialization is not proven below host admin.
* +0 Concrete repeated release/refcount: not awarded. The visible bug is release then stale return/dereference, not a concrete second `put` or refcount mutation on the same dangling reference.
* +0 Structured stale-object leverage: not separately awarded. The object is active and structured, but attacker-controlled replacement or field consumption after reclaim is not shown.
* +0 Weak LPE: not separately awarded. UAF alone is not enough.
* +0 Strong exploitation leverage: no controlled reclaim, callback control, arbitrary write, refcount takeover, or demonstrated privilege escalation.
* +0 Confidentiality/Integrity: not awarded because no disclosure or unauthorized modification primitive is established.
* +0 Extra Availability: not awarded because the crash is already counted once.
* No negative privilege penalty: host-admin-equivalent privilege is plausible but not positively established by the supplied evidence.
* No race/timing penalty: this is an ordered allocation-failure error path, not a narrow race.
* No rare-hardware penalty: RDMA/rvt is specialized, but the supplied evidence does not prove a rare physical-only prerequisite sufficient for a scoring penalty.

Repeated-release check:

* First release can end lifetime: yes, `ib_dealloc_device()` releases/deallocates the IB device object.
* Exact same reference survives: yes, `rdi` remains in scope and was returned before the fix.
* Second operation is real lifetime/refcount mutation: no visible second release/refcount mutation is shown.
* Concrete repeated release/refcount +1 awarded: no.
* Paranoid repeated-refcount leverage +1 awarded: no.
* FLOOR 5 applied: yes, due to returned active stale object escaping to callers, not due to repeated release/refcount.

## 3. Reachability analysis

The vulnerable path is reached when `rvt_alloc_device()` succeeds in allocating the main RDMA device object but fails to allocate the port array. The caller then receives a non-NULL pointer that actually refers to released storage.

The complete attacker model is unresolved. Device allocation and RDMA driver initialization commonly occur in privileged probe/module/device-management paths, but the supplied evidence does not positively prove that host-admin-equivalent privilege is required for the complete vulnerable sequence. Conversely, below-host-admin reachability is also not proven. A local user might influence allocation failure through memory pressure, but that alone does not establish control over the complete RDMA initialization path.

Namespace-scoped privilege is not established. No evidence shows that the complete RDMA/rvt allocation and initialization path can be performed inside attacker-created namespaces.

No direct network trigger is established. The issue is not in RDMA packet receive/transmit processing. It is an allocation/init failure bug.

No race is required. The relevant sequence is allocation failure followed by returning a released pointer.

## 4. Primitive and severity interpretation

The exact primitive is a real use-after-free caused by returning a pointer after `ib_dealloc_device()` has released the underlying IB device allocation.

The stale object is an active structured RDMA device object, not fixed stale storage such as a static DMA buffer. The stale use is not merely a local cleanup helper fault because the freed pointer escapes the allocator helper and is interpreted by callers as a successful allocation result.

This is CLASS_B. It is a real kernel lifetime/corruption primitive requiring exploitation-leverage review, but CLASS_A is not justified because the patch does not establish controlled reclaim, arbitrary read/write, function pointer control, refcount takeover, or a demonstrated privilege escalation chain.

The demonstrated consequence is kernel crash/DoS from dereferencing freed memory. The speculative consequence is broader memory corruption or privilege escalation if the freed object can be reclaimed or if caller-side initialization writes into attacker-influenced replacement storage, but that is not established from the supplied patch.

ImportantCandidate=YES because a freed active kernel object is returned to callers as valid and the downstream caller behavior is not fully visible. This should not be treated as an automatically closable crash-only bug.

AutoDowngradeAllowed=NO because automatic downgrade would discard a concrete unresolved active-object UAF primitive. This does not mean the issue is proven Important, but it does require manual review.

Operational severity is Actionable Moderate, with possible Important consideration depending on call-site behavior and practical reachability.

## 5. One-sentence report phrase

`rvt_alloc_device() can return a freed RDMA device pointer after port allocation failure, causing callers to dereference released kernel memory and creating a local use after free with at least kernel crash impact.`

## 6. Manual review recommendation

MANUAL CHECK REQUIRED

Reason: the patch fixes an explicit use-after-free where a released active RDMA device object escapes to callers as a valid allocation result. Controlled exploitation is not demonstrated, but the lifetime primitive is concrete and should not be automatically downgraded.

## 7. Call-site confidence

Call-site confidence: medium

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

Patch: RDMA/rvt: Return NULL after port allocation failure [ Upstream
Commit: b1e1b25723af44328524c7b54433e4bf0ab9930c
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=b1e1b25723af44328524c7b54433e4bf0ab9930c

Changed files:
  drivers/infiniband/sw/rdmavt/vt.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=b1e1b25723af44328524c7b54433e4bf0ab9930c

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

The original Linux kernel CVE announcement for CVE-2026-93104 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-93104

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:H/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: unknown
  Paranoid ActionableScore: 5

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).