* [CVE-2026-93154][MODERATE 7.0] RDMA/irdma: Add refcounting to user ring MRs [ Upstream
@ 2026-09-18 7:32 21% AL-KERNEL
0 siblings, 0 replies; 1+ results
From: AL-KERNEL @ 2026-09-18 7:32 UTC (permalink / raw)
To: kernel-cve
CVE: CVE-2026-93154
Priority: MODERATE 7.0
AL-KERNEL base severity: MODERATE
KPANIC flag: YES
Patch: RDMA/irdma: Add refcounting to user ring MRs [ Upstream
Commit: 07974c267f603a76aead348bca97deee95efdcc9
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=07974c267f603a76aead348bca97deee95efdcc9
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-93154
Analysis date: Fri, 18 Sep 2026 03:32:47 -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-2026-93154 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-93154
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-2026-93154 MODERATE https://www.kernelcve.org/list/?q=CVE-2026-93154 CHECK WITH IMPACT FROM ORIG NN LOW Maybe valid. Check manually. Hints by AL-KERNEL: The best (paranoid) CVSS is 'AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H';*CWE-416;CWE-911;*CWE-664;Other CVSS 'AV:L/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:H';BEST CVSS score: '7.8';DESCR 'The irdma driver did not pin the lifetime of user MRs that back QP/CQ/SRQ ring objects. A local process with access to RDMA uverbs could associate an MR with a user ring and then deregister that MR while the QP CQ or SRQ object still kept references to the backing PBL or MR state. This creates a use-after-free candidate rather than a simple resource leak, because later destroy resize or queue handling paths may touch memory whose owning MR lifetime has ended. For the CVSS the PR:L is used because reliable triggering requires local code execution with access to the RDMA uverbs device, but it does not require full administrator privileges in systems where RDMA access is delegated to users or service accounts. The issue is not directly network reachable because a remote peer cannot deregister the local MR by itself. Impact is at least local denial of service via kernel crash and in the paranoid case may include confidentiality and integrity impact due to possible memory corruption or privilege escalation.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Actionable Moderate / Strong Important candidate requiring manual review (with actual score 6) YES NETWORK HARDWARE LINUS KPANIC MEMORY KPANIC INCREASED_TO_HIGH_BASED_ON_GUESSCVSS DECREASED_TO_MODERATE7_BASED_ON_FALSEPOSCHECKOFKP NO NO checked
======================================================================
ACTIONABLESCORE ANALYSIS
======================================================================
ActionableScore=6
ActionableScoreLower=5
PrimitiveClass=CLASS_B
RealUAF=YES
HostAdminRequired=NO
NamespaceScopedPrivilege=NO
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:
* +2 Real UAF / active lifetime violation. The commit explicitly says userspace must be prevented from deregistering MRs that back QP/CQ/SRQ rings, which directly implies that ring objects can retain references to MR/PBL state after the MR lifetime is ended.
* +1 Local below-host-admin reachability. The affected objects are userspace RDMA verbs objects, and the attacker model is a local process with access to the RDMA uverbs device rather than host administrator.
* +1 Reliable kernel crash / strong DoS. A stale MR/PBL backing an active queue object can realistically lead to kernel fault or device/driver failure when the QP/CQ/SRQ object is later used, resized, or destroyed.
* +1 Attacker-influenced active-object lifetime. The same userspace actor can create or associate the MR-backed user ring object and then deregister the MR, directly influencing the vulnerable lifetime transition.
Paranoid additional signal:
* +1 Structured stale-object exploitation leverage. The stale objects are active RDMA queue backing structures, not passive metadata, and later queue handling may consume MR/PBL state in nontrivial memory/DMA-related processing. This is unresolved exploitation leverage, but not enough to classify as CLASS_A.
Important non-awarded signals:
* +0 Strong LPE / strong exploitation leverage not awarded. Controlled reclaim, refcount takeover, arbitrary write, callback control, or demonstrated privilege escalation is not established.
* +0 Concrete repeated release/refcount signal not awarded. The bug is missing lifetime pinning of MRs associated with rings, not a visible second release/put on the exact same dangling reference.
* +0 Confidentiality and integrity not independently awarded in ActionableScore. They are plausible in a paranoid CVSS sense, but no concrete leak, arbitrary write, or protected object modification is demonstrated.
* +0 Remote reachability not awarded. A remote RDMA peer cannot deregister the local MR by itself based on the supplied evidence.
* +0 Extra availability not awarded. The plausible crash/DoS is already counted once.
* No negative privilege penalty applied. Host-admin-equivalent privilege is not positively established.
* No race penalty applied. This is primarily a lifetime/ownership bug, not a narrow timing race.
Repeated release/refcount checklist:
* First release can end lifetime: yes in the broader sense that userspace MR deregistration can end the MR lifetime.
* Exact same reference survives to a second release/refcount mutation: not established.
* Second operation is a real lifetime/refcount mutation: not established.
* +1 Concrete repeated release/refcount awarded: no.
* Paranoid repeated-refcount leverage +1 awarded: no.
* FLOOR 5 applied: not needed, raw conservative score already equals 5.
3. Reachability analysis
The likely attacker is a local process that can open and use the RDMA uverbs interface for the Intel irdma device. Such access is below host-admin in many RDMA deployments, although it may depend on device node permissions, cgroup/device policy, or service account delegation.
The complete trigger is not network-only. A remote RDMA peer may interact with queues after setup, but the key vulnerable operation is local userspace deregistration of the MR that backs QP/CQ/SRQ rings.
No namespace-scoped capability chain is established. The issue is not about CAP_NET_ADMIN in a network namespace or an attacker-created namespace.
Prior setup requires the irdma driver and compatible RDMA hardware/device exposure. This is hardware-specific but not enough by itself for a rare-hardware penalty, because RDMA user verbs are the normal intended interface of the subsystem.
No special hard race is required by the supplied patch. The vulnerable sequence is an ownership/lifetime sequence: associate MR-backed ring state, deregister MR, then later continue using or destroying the queue object.
4. Primitive and severity interpretation
The exact primitive is a kernel lifetime/ownership violation for user MRs backing active RDMA QP/CQ/SRQ ring objects. The missing refcount allows userspace to end the MR lifetime while kernel queue objects still keep associated PBL/MR references.
This is an active structured object case, not fixed stale storage. The stale state is tied to RDMA queue resources and memory registration state that can participate in later queue management, resize, destruction, or device-facing operations.
The demonstrated consequence is at least local DoS through stale kernel object use. The speculative but technically meaningful concern is memory corruption or privilege escalation if stale MR/PBL state can be reused or consumed in an attacker-influenced way.
PrimitiveClass is CLASS_B. A real UAF/lifetime primitive exists, but controlled reclaim, arbitrary kernel read/write, callback hijack, or demonstrated LPE is not established.
ImportantCandidate is YES because this is not a simple crash-only NULL dereference or passive leak. It is an active-object lifetime bug in RDMA user memory registration and queue backing state, with unresolved exploitation leverage.
AutoDowngradeAllowed is NO because automatically closing this as ordinary Moderate would lose a concrete user-reachable kernel lifetime primitive involving MR/PBL ownership and active queue objects.
Likely operational severity is actionable Moderate at minimum, with Strong Important review warranted if surrounding code confirms stale MR/PBL reuse in DMA, queue traversal, or attacker-influenced memory handling.
5. One-sentence report phrase
A local process with RDMA uverbs access could deregister an MR that still backs QP/CQ/SRQ ring state in the irdma driver, creating a stale MR/PBL lifetime condition that can cause kernel DoS and requires manual review for possible memory corruption impact.
6. Manual review recommendation
MANUAL CHECK REQUIRED
Reason: this is a real active-object lifetime bug in RDMA MR/PBL state reachable from userspace, and the patch-supported primitive is stronger than a simple crash-only validation failure even though controlled exploitation is not demonstrated.
7. Call-site confidence
Call-site confidence: medium
======================================================================
UPSTREAM PATCH SUMMARY
======================================================================
Patch: RDMA/irdma: Add refcounting to user ring MRs [ Upstream
Commit: 07974c267f603a76aead348bca97deee95efdcc9
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=07974c267f603a76aead348bca97deee95efdcc9
Changed files:
drivers/infiniband/hw/irdma/utils.c
drivers/infiniband/hw/irdma/verbs.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=07974c267f603a76aead348bca97deee95efdcc9
======================================================================
DETAILED REPORT METHODOLOGY
======================================================================
The original Linux kernel CVE announcement for CVE-2026-93154 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-93154
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:L/I:L/A:H
The Best / paranoid CVSS vector: AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H
The Best / paranoid CVSS score: 7.8
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 [email protected] (and both
send reply to CVE record itself too and see "reply" button below for howto reply).
^ permalink raw reply [relevance 21%]
Results 1-1 of 1 | reverse | sort options + mbox downloads above
-- links below jump to the message on this page --
2026-09-18 7:32 21% [CVE-2026-93154][MODERATE 7.0] RDMA/irdma: Add refcounting to user ring MRs [ Upstream AL-KERNEL
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox