[CVE-2026-43128][IMPORTANT] RDMA/umem: Fix double dma_buf_unpin in failure path [ Upstream

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

CVE: CVE-2026-43128
Priority: IMPORTANT
AL-KERNEL base severity: IMPORTANT
KPANIC flag: YES
Patch: RDMA/umem: Fix double dma_buf_unpin in failure path [ Upstream
Commit: 70542b69abff34d24b11ae0bb200cc7a766d18df
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=70542b69abff34d24b11ae0bb200cc7a766d18df
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-43128
Analysis date: Wed, 27 May 2026 17:03:23 -0400
ActionableScore: 7
ActionableScore lower bound: 5
Actionable bucket: Strong Important candidate / Actionable Moderate at minimum
Manual review required: YES

Summary:
A local process with RDMA uverbs and dma-buf access can trigger a failure path where a pinned dma-buf is unpinned twice, causing RDMA DMA memory lifetime corruption with DoS impact and plausible but unproven privilege-escalation concern.

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

The original Linux kernel CVE announcement for CVE-2026-43128 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-43128

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: IMPORTANT
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-2026-43128	IMPORTANT	CHECK	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-911;*CWE-672;Other CVSS 'AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:H';BEST CVSS score: '7';DESCR 'RDMA umem dma-buf registration can hit a double dma_buf_unpin when ib_umem_dmabuf_map_pages() fails after the dma-buf was marked pinned. The first unpin happens in the immediate error path and a second unpin can happen later through ib_umem_release() and revoke because the pinned state is still recorded. For the CVSS the PR:L is selected because practical triggering requires local access to RDMA uverbs and a dma-buf file descriptor, which can be available to a local user, container workload, or graphics/RDMA service without full host root. The issue is not network reachable through RDMA traffic itself and normally requires triggering a map_pages failure path, so AC:H is appropriate. Impact is at least local denial of service through pin lifetime imbalance, WARN, crash, or incorrect unmap-unpin ordering. In the paranoid view, privilege escalation cannot be excluded because double unpin is a real lifetime corruption pattern around shared dma-buf and DMA memory state, although the patch does not demonstrate a reliable arbitrary write primitive.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Strong Important candidate / Actionable Moderate at minimum (with actual score 7)	YES	SIMPLEFIX NETWORK LINUS KPANIC MEMORY 	YES	YES	checked

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

ActionableScore=7
ActionableScoreLower=5

## 1. ActionableScore

* Conservative score: 5
* Paranoid score: 7
* Final recommended bucket: **Strong Important candidate / Actionable Moderate at minimum**::

## 2. Signal breakdown

Conservative signals:

* Local unprivileged trigger: +1. The path can be reached by a local process using RDMA uverbs plus a dma-buf fd, depending on device permissions and deployment.
* Generic memory corruption / strong lifetime primitive class: +2. This is a double `dma_buf_unpin()` condition, similar to double release / refcount lifetime corruption.
* Real lifetime corruption: +1. The `pinned` state remains set after the first unpin, so later release/revoke can unpin the same dma-buf attachment again.
* Reliable kernel crash / strong DoS: +1. A double unpin can cause refcount imbalance, WARN/Oops, invalid cleanup ordering, or device/DMA memory management failure.
* Availability impact realistic: +1. Failure is in RDMA memory registration cleanup and can plausibly crash or destabilize the kernel/device context.
* Privileged kernel/device-management memory corruption path: +1. The bug is in RDMA + dma-buf memory management, a trusted DMA-adjacent subsystem.
* Hard or special failure path required: -1. Triggering requires `ib_umem_dmabuf_map_pages()` to fail after the dma-buf was pinned.
* Rare/configuration-dependent exposure: -1. Requires RDMA umem dma-buf support, suitable hardware/driver stack, and accessible dma-buf/RDMA interfaces.

Conservative total: 5

Paranoid additional interpretation:

* Weak LPE concern: +1. Double unpin of shared dma-buf/DMA memory state may lead to premature unpin or stale lifetime assumptions, but no reliable overwrite primitive is shown.
* Limited C/I concern from lifetime corruption: +1. In the worst defensible case, premature unpin of shared DMA-backed memory could affect confidentiality or integrity, but this remains weaker than a demonstrated arbitrary write.

Paranoid total: 7

Not counted:

* No remote/network-triggerable score. RDMA traffic from a peer does not directly hit this local memory-registration failure path.
* No Dirty-Pipe-like/page-cache score. The patch does not show writes into shared file-backed cache or COW bypass.
* No strong LPE score. The patch does not demonstrate UAF reclaim, controlled overwrite, reliable object confusion, or arbitrary write.

## 3. Reachability analysis

The bug is reachable through local RDMA memory registration using a dma-buf backed umem object. A practical attacker needs access to RDMA uverbs and a dma-buf fd, which may be available to local users, GPU/RDMA workloads, containers with delegated device access, or service accounts on accelerator systems. It is not directly reachable from remote RDMA packets.

Namespaces and containers matter because `/dev/infiniband/*` and dma-buf capable devices are often selectively delegated. If such devices are exposed into a container, this should not be treated as full host-root-only. If the deployment restricts RDMA and dma-buf access to fully privileged host root, practical severity is lower.

Realistic exploitation also requires forcing or naturally hitting the `map_pages` failure path, for example through reservation timeout or resource/fence failure. That makes exploitability harder than a straight-line UAF, but the double-unpin lifetime bug is still security-relevant.

## 4. Severity interpretation

This is at least Actionable Moderate because it is a real double-unpin lifetime bug in DMA-adjacent RDMA memory management. The realistic demonstrated impact is local DoS or kernel/device instability through invalid cleanup ordering and pin refcount imbalance.

The Important-class concern is theoretical but defensible: premature unpinning of shared dma-buf memory may create stale lifetime assumptions around DMA mappings. However, the patch does not show a controlled overwrite or reliable privilege-escalation primitive, so this should be treated as Strong Important candidate rather than confirmed Important.

## 5. One-sentence report phrase

A local process with RDMA uverbs and dma-buf access can trigger a failure path where a pinned dma-buf is unpinned twice, causing RDMA DMA memory lifetime corruption with DoS impact and plausible but unproven privilege-escalation concern.

## 6. Manual review recommendation

MANUAL CHECK REQUIRED

Reason: this is a double-unpin lifetime bug in RDMA dma-buf memory management, and delegated RDMA/GPU/container access can make PR:L realistic even though exploitation depends on a special failure path.

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

Patch: RDMA/umem: Fix double dma_buf_unpin in failure path [ Upstream
Commit: 70542b69abff34d24b11ae0bb200cc7a766d18df
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=70542b69abff34d24b11ae0bb200cc7a766d18df

Changed files:
  drivers/infiniband/core/umem_dmabuf.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=70542b69abff34d24b11ae0bb200cc7a766d18df

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

The original Linux kernel CVE announcement for CVE-2026-43128 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-43128

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: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: 7

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: IMPORTANT
KPANIC detected for this report: YES
Published priority for this report (same as in Subject): IMPORTANT

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).