[CVE-2026-93287][MODERATE REGULAR] i2c: smbus: reject oversized block transfers in the common path

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

CVE: CVE-2026-93287
Priority: MODERATE REGULAR
AL-KERNEL base severity: MODERATE
KPANIC flag: NO
Patch: i2c: smbus: reject oversized block transfers in the common path
Commit: f09edffba97d37b1fdff5f818a9a45f8a98ac171
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=f09edffba97d37b1fdff5f818a9a45f8a98ac171
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-93287
Analysis date: Thu, 24 Sep 2026 20:44:55 -0400
ActionableScore: 3
ActionableScore lower bound: 2
Actionable bucket: Borderline / manual review recommended
Manual review required: YES

Summary:
No one-sentence report phrase was found.

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

The original Linux kernel CVE announcement for CVE-2026-93287 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-93287

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-2026-93287	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-93287 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-787;CWE-125;*CWE-20;Other CVSS 'AV:L/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:H';BEST CVSS score: '7.8';DESCR 'An out of bounds access in the common SMBus transfer path occurs because caller controlled block lengths were not validated before SMBus tracepoints and native smbus_xfer callbacks were executed. A local process with access to an i2c-dev device can submit an oversized block write through an ioctl. The smbus_write tracepoint can then copy up to 255 bytes into a 34 byte event buffer. This creates an out of bounds read and a trace event buffer overflow which may crash the kernel or corrupt kernel memory. A related native callback path in i2c-stub can also access data block entries beyond the fixed 34 byte SMBus buffer. For the CVSS the PR:L value is used because access to /dev/i2c-X may be delegated to a non administrative local user through device permissions, an i2c group, a service account, or a container policy. The issue is not directly network reachable because a local i2c-dev ioctl caller is required. Impact is at least local denial of service through kernel memory corruption. In the worst case the out of bounds write primitive may affect confidentiality and integrity and should be reviewed manually.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Borderline / manual review recommended (with actual score 3)	YES	READ KASAN OOB USB HARDWARE LINUS  INCREASED_TO_MODERATE7_BASED_ON_GUESSCVSS DECREASED_TO_MODERATEREG_BASED_ON_ACTIONABLESCORELESSTHAN4	NO	NO	checked

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

ActionableScore=3
ActionableScoreLower=2
PrimitiveClass=CLASS_B
RealUAF=NO
HostAdminRequired=NO
NamespaceScopedPrivilege=UNKNOWN
ImportantCandidate=NO
ManualReview=RECOMMENDED
AutoDowngradeAllowed=YES

1. ActionableScore

* Raw conservative score: 2
* Conservative score: 2
* Raw paranoid score: 3
* Paranoid score: 3
* No floor or cap applied.
* Final recommended bucket: **Borderline / manual review recommended**::

2. Signal breakdown

* +1 Constrained kernel buffer overwrite. The SMBus write tracepoint copies `data->block[0] + 1` bytes into a fixed 34 byte trace-event buffer. A caller-controlled length of `0xfe` causes a 255 byte copy, exceeding the destination by 221 bytes.
* +1 Local below-host-admin trigger. The vulnerable request is reachable through `i2cdev_ioctl_smbus()` and `/dev/i2c-X`. Access is governed by device-node DAC permissions and can be granted to a non-administrative local account or `i2c` group member. No network path is involved.
* +1 Paranoid only. When the vulnerable tracepoint is enabled, the large fixed overflow can plausibly corrupt trace-ring event memory and cause a kernel fault or broader trace subsystem instability. This is availability impact only and is not counted as confidentiality or integrity impact.

Important non-awarded signals:

* Controlled reclaim is not demonstrated.
* Arbitrary or attacker-directed kernel write is not demonstrated. The destination is a fixed trace-event buffer, and the bytes beyond the 34 byte SMBus union originate from adjacent kernel stack memory rather than a fully attacker-controlled payload.
* Confidentiality is not awarded. The out-of-bounds source read may copy adjacent stack bytes into a trace record, but no attacker-readable disclosure path is established. Access to tracing data is normally separately permission-gated.
* Integrity is not awarded. A bounded overwrite exists, but the supplied evidence does not establish control of an exploitable target, active object replacement, callback corruption, or a privileged state modification.
* Remote reachability is not established. The trigger requires a local `i2c-dev` ioctl caller.
* No UAF, refcount misuse, repeated release, or lifetime mutation is present.
* No extra availability point is awarded in the conservative score because KASAN confirms the invalid memory access, but the supplied evidence does not demonstrate a production kernel panic or reliable host-wide crash.
* The `i2c-stub` path is a separate out-of-bounds array access in a test-oriented native SMBus implementation. It does not independently establish broad exposure across all physical I2C drivers.

3. Reachability analysis

The primary trigger is a local process issuing the SMBus ioctl through an accessible `/dev/i2c-X` device node. The caller supplies the block length in `data->block[0]` for SMBus write-related protocols, including `I2C_SMBUS_I2C_BLOCK_DATA`, `I2C_SMBUS_BLOCK_PROC_CALL`, and SMBus block writes.

Host-admin-equivalent privilege is not inherently required by the ioctl itself. Access is controlled by the permissions on the relevant `i2c-dev` node, which may be restricted to root but may also be delegated to a local non-administrative user, service account, or group. Therefore, `HostAdminRequired=NO` reflects that the complete operation can be reachable below host-admin privilege when device-node access is delegated.

The complete trigger is not proven namespace-scoped. I2C adapters and their `/dev/i2c-X` nodes commonly correspond to host-owned hardware or host-created devices. A container or network namespace alone does not provide the required adapter access. Therefore, `NamespaceScopedPrivilege=UNKNOWN`.

The tracepoint overflow requires the SMBus trace event to be enabled. Enabling kernel tracing commonly requires prior privileged configuration. The `i2c-stub` path also generally requires the test driver to be built, loaded, and configured. These are relevant deployment restrictions, but they are prior setup conditions and do not eliminate the risk where an administrator has already enabled tracing or exposed an I2C device to a less-privileged local actor.

No race, special timing condition, remote packet delivery, firmware compromise, or unusual physical hardware interaction is required after the vulnerable local ioctl path is reachable.

4. Primitive and severity interpretation

The patch establishes two concrete out-of-bounds access classes.

First, the SMBus tracepoint path performs an out-of-bounds read from the fixed SMBus data union and an out-of-bounds write into the fixed-size trace-event buffer. The caller controls the requested transfer length, but does not control the bytes read beyond the 34 byte SMBus union. The write target is also constrained to the trace event allocation. This is therefore a constrained kernel buffer overwrite rather than a demonstrated arbitrary write or controlled active-object corruption.

Second, the native `i2c-stub` callback can read `data->block[1+i]` beyond the 34 byte union when an oversized I2C block length is supplied. The supplied evidence establishes an invalid array access but does not establish a write, a user-visible information leak, or use of attacker-shaped replacement objects.

This is `CLASS_B` because a real kernel out-of-bounds write exists in the tracepoint path, but the overwrite target and payload are materially constrained. It is not `CLASS_A` because no controlled reclaim, target selection, arbitrary write, callback manipulation, credential modification, disclosure path, or demonstrated privilege escalation is established.

The demonstrated practical consequence is kernel memory-safety failure, potentially including a kernel warning, KASAN failure, trace-buffer corruption, or denial of service if the corrupted trace state is later consumed. Confidentiality and integrity consequences remain speculative because the supplied patch and reproducer do not establish an attacker-readable leak or attacker-controlled corruption of security-sensitive state.

`ImportantCandidate=NO` is appropriate because the concrete overwrite is bounded and fixed-target, and meaningful exploitation leverage beyond denial of service is not established. `AutoDowngradeAllowed=YES` is appropriate because the issue remains a constrained `CLASS_B` corruption candidate without demonstrated controlled overwrite contents, target selection, active-object reuse, security-boundary bypass, or lifetime corruption. Manual review is still recommended because the tracepoint path contains a real kernel buffer overflow with a large overflow size.

Likely operational severity is low to moderate in typical systems because the vulnerable tracepoint or `i2c-stub` path must be enabled and the attacker needs local access to an I2C device node. It should not be treated as a confirmed local privilege escalation based on the supplied evidence alone.

5. One-sentence report phrase

A local user with access to an i2c-dev device can submit an oversized SMBus block transfer that reaches an unchecked tracepoint memcpy or native SMBus callback, causing an out-of-bounds kernel memory access and possible denial of service.

6. Manual review recommendation

MANUAL CHECK RECOMMENDED

The tracepoint path contains a concrete large out-of-bounds write, but the supplied evidence does not establish attacker control over the overflow payload beyond the valid SMBus buffer, control over the overwrite target, or a disclosure or privilege-escalation chain.

7. Call-site confidence

Call-site confidence: high

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

Patch: i2c: smbus: reject oversized block transfers in the common path
Commit: f09edffba97d37b1fdff5f818a9a45f8a98ac171
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=f09edffba97d37b1fdff5f818a9a45f8a98ac171

Commit description:

The SMBus block transfer length data->block[0] is validated in
i2c_smbus_xfer_emulated() but that check runs too late for tracepoints
and is skipped entirely when the adapter provides a native smbus_xfer
implementation. This allows user-controlled oversized block lengths to
reach tracepoint memcpy calls and driver callbacks unchecked.

Add an early validation in __i2c_smbus_xfer() that rejects block
transfers whose caller-supplied length is zero or exceeds
I2C_SMBUS_BLOCK_MAX before any tracepoint fires or driver callback
runs. data->block[0] is filled in by the device on SMBus block reads,
so the check is scoped to operations where the length is actually
supplied by the caller. This is consistent with the existing -EINVAL
convention in the emulated path and protects all downstream consumers
at once: the smbus_write tracepoint, all native smbus_xfer driver
implementations, and the emulated path.

Two distinct bugs are fixed by this change:

Bug 1: smbus_write tracepoint OOB (include/trace/events/smbus.h)
  trace_smbus_write() fires before any validation and copies
  data->block[0]+1 bytes into a 34-byte event buffer. With
  block[0]=0xfe the tracepoint copies 255 bytes, overflowing by 221.

 BUG: KASAN: stack-out-of-bounds in trace_event_raw_event_smbus_write+0x27c/0x530
 Read of size 255 at addr ffff88800d98fcf8 by task poc_smbus/91
 Call Trace:
  <TASK>
  __asan_memcpy+0x23/0x80
  trace_event_raw_event_smbus_write+0x27c/0x530
  __i2c_smbus_xfer+0x43a/0xa40
  i2c_smbus_xfer+0x19e/0x340
  i2cdev_ioctl_smbus+0x38f/0x7f0
  i2cdev_ioctl+0x35e/0x680
  __x64_sys_ioctl+0x147/0x1e0
  do_syscall_64+0xcf/0x15a0
  entry_SYSCALL_64_after_hwframe+0x76/0x7e
  </TASK>

Bug 2: i2c-stub I2C_SMBUS_I2C_BLOCK_DATA OOB (drivers/i2c/i2c-stub.c)
  stub_xfer() implements .smbus_xfer directly and only clamps
  block[0] against 256-command, not I2C_SMBUS_BLOCK_MAX. With
  block[0]=0xff and command=0 the loop accesses block[1+i] for
  i up to 254, far past the 34-byte union.

 UBSAN: array-index-out-of-bounds in drivers/i2c/i2c-stub.c:223:44
 index 34 is out of range for type '__u8 [34]'
 Call Trace:
  <TASK>
  __ubsan_handle_out_of_bounds+0xd7/0x120
  stub_xfer+0x1971/0x198f [i2c_stub]
  __i2c_smbus_xfer+0x306/0xa40
  i2c_smbus_xfer+0x19e/0x340
  i2cdev_ioctl_smbus+0x38f/0x7f0
  i2cdev_ioctl+0x35e/0x680
  __x64_sys_ioctl+0x147/0x1e0
  do_syscall_64+0xcf/0x15a0
  entry_SYSCALL_64_after_hwframe+0x76/0x7e
  </TASK>

Both traces reproduced on v7.0-rc6+i2c/for-current with KASAN+UBSAN.

Fixes: 8a32599 ("i2c: Add message transfer tracepoints for SMBUS [ver #2]")
Fixes: 4710317 ("i2c-stub: Implement I2C block support")
Reported-by: Xiang Mei <xmei5@asu.edu>
Signed-off-by: Weiming Shi <bestswngs@gmail.com>
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Changed files:
  include/trace/events/smbus.h
  drivers/i2c/i2c-stub.c
  drivers/i2c/i2c-core-smbus.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=f09edffba97d37b1fdff5f818a9a45f8a98ac171

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

The original Linux kernel CVE announcement for CVE-2026-93287 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-93287

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: 2
  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).