[CVE-2026-90417][MODERATE 7.0] RDMA/cxgb4: Fix dereg_skb leak and double free in write_tpt_entry() [ Upstream

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

CVE: CVE-2026-90417
Priority: MODERATE 7.0
AL-KERNEL base severity: MODERATE
KPANIC flag: YES
Patch: RDMA/cxgb4: Fix dereg_skb leak and double free in write_tpt_entry() [ Upstream
Commit: 6b15fc9bc117fb6597fa1c2cc30284aafe71101c
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=6b15fc9bc117fb6597fa1c2cc30284aafe71101c
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-90417
Analysis date: Thu, 17 Sep 2026 19:13:37 -0400
ActionableScore: 6
ActionableScore lower bound: 5
Actionable bucket: Actionable Moderate / Strong Important candidate requiring manual review
Manual review required: YES

Summary:
RDMA/cxgb4 has an skb ownership bug where `dereg_mem()` can consume `mhp->dereg_skb` and a later error path can free the same skb again, creating a local double-free memory corruption condition that requires manual Important-vs-Moderate review.

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

The original Linux kernel CVE announcement for CVE-2026-90417 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-90417

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-2026-90417	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-90417 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-415;*CWE-401;CWE-703;CWE-415;Other CVSS 'AV:L/AC:H/PR:L/UI:N/S:U/C:L/I:L/A:H';BEST CVSS score: '7';DESCR 'A double free and leak issue exists in RDMA/cxgb4 memory registration cleanup. write_tpt_entry can return before handing the caller supplied skb to the transmit path when the device is in fatal error state or when internal allocation fails. One caller can then lose the dereg_skb reference and leak memory, while c4iw_get_dma_mr can later free the same skb again after dereg_mem has already consumed it. For the CVSS the PR:L is used because a local user with access to the RDMA device or verbs interface may be able to exercise memory registration and cleanup paths without full root privileges. The issue is not directly network reachable and normally depends on local RDMA device access plus an error or low resource condition. Impact is at least local denial of service and memory corruption via skb double free, with possible confidentiality or integrity impact in a worst case kernel heap corruption scenario.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Actionable Moderate / Strong Important candidate requiring manual review (with actual score 6)	YES	INIT SIMPLEFIX LEAK NETWORK SKB DMAorINTERRUPT HARDWARE LINUS KPANIC MEMORY  KPANIC INCREASED_TO_HIGH_BASED_ON_GUESSCVSS KPANIC LOWERED_FROM_HIGH_BASED_ON_NULLPTR KPANIC INCREASED_TO_HIGH_BASED_ON_ACTIONABLESCOREHIGHERTHAN5 DECREASED_TO_MODERATE7_BASED_ON_FALSEPOSCHECKOFDH	NO	NO	checked

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

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

## 1. ActionableScore

* Raw conservative score: 5
* Conservative score: 5
* Raw paranoid score: 6
* Paranoid score: 6
* Final recommended bucket: **Actionable Moderate / Strong Important candidate requiring manual review**::

No floor or cap changed the conservative score. No floor or cap changed the paranoid score.

## 2. Signal breakdown

Positive signals:

* +2 Real UAF / lifetime violation: `dereg_mem()` consumes `mhp->dereg_skb` through the offload send path, and the old error path then reached `kfree_skb(mhp->dereg_skb)` on the same reference.
* +1 Concrete repeated release/refcount operation: the first consume/free can end the skb lifetime, the exact same `mhp->dereg_skb` reference remains reachable, and the later `kfree_skb()` is a concrete second lifetime-mutating operation.
* +1 Concrete persistent kernel resource/reference leak: in another caller, `write_tpt_entry()` could return before handing the preallocated skb to transmit, while `c4iw_dereg_mr()` ignored the error and leaked `mhp->dereg_skb`.
* +1 Reliable kernel crash / strong DoS: skb double free is a real kernel memory lifetime corruption condition and can plausibly produce allocator corruption, refcount corruption, panic, or fatal fault.
* Paranoid +1 repeated-refcount leverage: the second `kfree_skb()` can act on storage whose original skb lifetime may already have ended, leaving unresolved post-free lifetime corruption leverage.

Negative signals:

* No host-admin penalty: host-admin-equivalent privilege is not positively established by the supplied evidence.
* No race/timing penalty: this is an ordered error/unwind ownership bug, not a narrow timing race.
* No remote reachability bonus: the issue is in a local RDMA driver path and is not directly packet-triggerable.
* No rare-hardware penalty applied: cxgb4 RDMA hardware is required, but driver-specific hardware alone is not enough for a scoring penalty under the supplied rules.

Non-awarded signals:

* Controlled reclaim is not demonstrated.
* Arbitrary or directed kernel write is not demonstrated.
* Confidentiality and integrity points are not awarded directly. The double free supports unresolved memory corruption review, but not a proven C/I consequence.
* Structured stale-object leverage is not separately awarded because the key additional fact is already counted as concrete repeated release/refcount.
* Weak LPE is not separately awarded for the same repeated-free fact.
* Extra Availability is not awarded because the same crash/DoS is already counted once.
* CLASS_A is not assigned because there is no demonstrated controlled reclaim, refcount takeover, callback control, arbitrary write, or LPE.

Repeated-release checks:

* First release can end lifetime: yes. `dereg_mem()` reaches a path where `c4iw_ofld_send()` consumes the skb on success and error.
* Exact same reference survives: yes. `mhp->dereg_skb` remains available to the old `err_free_skb` path.
* Second operation mutates lifetime/refcount: yes. `kfree_skb()` is a real skb lifetime release.
* Concrete repeated release/refcount +1 awarded: yes.
* Paranoid repeated-refcount leverage +1 awarded: yes, because the second free itself performs lifetime mutation through potentially freed storage.
* FLOOR 5 applied: not needed, because raw conservative score is already 5.

## 3. Reachability analysis

The plausible attacker is a local actor able to access the RDMA/cxgb4 verbs or kernel RDMA interface paths that create and tear down memory registrations. The complete below-host-admin trigger is plausible in systems where RDMA device nodes are delegated to non-root users, but the supplied evidence does not prove the full privilege boundary for this specific cxgb4 path. Therefore `HostAdminRequired=UNKNOWN`, not `NO`.

The issue is not directly network reachable. Remote RDMA traffic does not by itself establish control over the local memory registration error/unwind path. A cxgb4 RDMA-capable device and loaded driver are prerequisites. The trigger also depends on fatal device state or allocation-failure/error paths, but this is not a race and should not be scored as hard timing.

Namespace-scoped privilege is unresolved. The supplied context does not establish that all required RDMA device access, memory registration, deregistration, and error-path triggering can be completed inside attacker-created namespaces.

## 4. Primitive and severity interpretation

The exact primitive is an skb ownership/lifetime bug with both a leak and a double free. The double free is stronger than an ordinary stale cleanup dereference because the second operation is not merely reading a stale pointer. It performs a real lifetime-mutating `kfree_skb()` on the same skb reference after an earlier path may already have consumed and freed it.

This is an active kernel lifetime corruption issue, but not a proven strong exploitation primitive. The demonstrated practical consequence is local kernel crash or memory corruption leading to DoS. The speculative consequence is stronger exploitation from heap lifetime corruption, but controlled reclaim, arbitrary write, callback influence, and privilege escalation are not established.

PrimitiveClass is `CLASS_B`. ImportantCandidate is `YES` because a concrete repeated lifetime mutation through a potentially dangling reference leaves meaningful Important-vs-Moderate exploitation leverage unresolved. AutoDowngradeAllowed is `NO` because automatically treating this as ordinary crash-only cleanup would discard the double-free lifetime primitive.

Likely operational severity is Actionable Moderate at minimum, with Strong Important candidate review warranted if RDMA access is available to non-admin users or if further analysis shows usable skb reclaim or allocator manipulation.

## 5. One-sentence report phrase

RDMA/cxgb4 has an skb ownership bug where `dereg_mem()` can consume `mhp->dereg_skb` and a later error path can free the same skb again, creating a local double-free memory corruption condition that requires manual Important-vs-Moderate review.

## 6. Manual review recommendation

MANUAL CHECK REQUIRED

Reason: the patch establishes a concrete skb double free with a second lifetime-mutating operation on the same potentially freed reference, and exploitation leverage beyond crash is unresolved.

## 7. Call-site confidence

Call-site confidence: high

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

Patch: RDMA/cxgb4: Fix dereg_skb leak and double free in write_tpt_entry() [ Upstream
Commit: 6b15fc9bc117fb6597fa1c2cc30284aafe71101c
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=6b15fc9bc117fb6597fa1c2cc30284aafe71101c

Changed files:
  drivers/infiniband/hw/cxgb4/mem.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=6b15fc9bc117fb6597fa1c2cc30284aafe71101c

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

The original Linux kernel CVE announcement for CVE-2026-90417 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-90417

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).