[CVE-2026-31765][MODERATE REGULAR] drm/amdgpu: Change AMDGPU_VA_RESERVED_TRAP_SIZE to 64KB

[CVE-2026-31765][MODERATE REGULAR] drm/amdgpu: Change AMDGPU_VA_RESERVED_TRAP_SIZE to 64KB

[AL-KERNEL]

CVE: CVE-2026-31765
Priority: MODERATE REGULAR
AL-KERNEL base severity: MODERATE
KPANIC flag: NO
Patch: drm/amdgpu: Change AMDGPU_VA_RESERVED_TRAP_SIZE to 64KB
Commit: 6b2614a0ff05a2d2836311425091c8feca6f0c21
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=6b2614a0ff05a2d2836311425091c8feca6f0c21
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-31765
Analysis date: Sat, 01 Aug 2026 13:22:37 -0400
ActionableScore: 3
ActionableScore lower bound: 2
Actionable bucket: Borderline manual-review Moderate / not Important
Manual review required: NO

Summary:
AMDGPU/KFD on 64K PAGE_SIZE systems can crash during local ROCm/KFD VM initialization because the reserved trap VA area is smaller than the computed CWSR TBA/TMA allocation, allowing a local GPU user to trigger a kernel DoS without evidence of privilege escalation.

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

The original Linux kernel CVE announcement for CVE-2026-31765 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-31765

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: NO

A detailed explanation of the methodology and priority rules is included
at the end of this message.

======================================================================
AL-KERNEL CLASSIFICATION RESULT
======================================================================

CVE-2026-31765	MODERATE	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:N/I:N/A:H';CWE-131;CWE-682;**CWE-476;Other CVSS 'AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:H';BEST CVSS score: '5.5';DESCR 'AMDGPU/KFD can crash on 64K PAGE_SIZE systems because the reserved trap VA area was only 8KB while the CWSR TBA/TMA allocation was derived from 2 * PAGE_SIZE and could become 128KB. A local user with access to the KFD or ROCm GPU device path can trigger the issue during VM acquisition or GPU runtime initialization, for example by running rocminfo or related compute tests. For the CVSS the PR:L is used because reliable triggering does not require full administrator privileges when the user already has normal access to the GPU device node. The issue is not network reachable and requires local code execution on a system with the affected AMDGPU/KFD configuration. Impact is denial of service via kernel crash. No concrete UAF, OOB write, arbitrary write, or information leak primitive is visible from the patch context.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Borderline manual-review Moderate / not Important (with actual score 3) 	MAYBE	OOB NULLPTR INIT KERNEL_PANIC_PLUS_UAF HARDWARE TEST MEMORY  DECREASED_TO_MODERATE_REGULAR_FROM_MODERATE7_BASED_ON_GUESSCVSS_AND_CIANNH	-	-	checked

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

ActionableScore=3
ActionableScoreLower=2

## 1. ActionableScore

* Conservative score: 2
* Paranoid score: 3
* Final recommended bucket: **Borderline manual-review Moderate / not Important**

## 2. Signal breakdown

Conservative scoring:

* Local unprivileged trigger: +1
  A normal local user with access to the KFD or ROCm GPU device path can trigger the crash by running rocminfo or similar GPU compute initialization paths.

* Reliable kernel crash / strong DoS: +1
  The commit includes a concrete Oops / NULL pointer dereference during kfd_ioctl_acquire_vm.

* Rare AND configuration-dependent exposure: -1
  The issue depends on AMDGPU/KFD with 64K PAGE_SIZE systems. This is not a broadly default x86 desktop/server configuration.

* Availability impact realistic: +1
  The crash is host-kernel level and appears reproducible on affected systems through ordinary ROCm/KFD initialization.

Paranoid delta:

* Do not subtract rare configuration dependency in the paranoid interpretation: +1 relative to conservative
  If the evaluated product commonly ships 64K PAGE_SIZE kernels with AMDGPU/KFD enabled, the rarity penalty should not be applied.

Not counted:

* No generic memory corruption bonus. The visible primitive is NULL / invalid pointer dereference, not UAF, OOB write, double free, type confusion, or arbitrary write.
* No privilege escalation bonus. The patch and trace do not show attacker-controlled reclaim, overwrite, callback control, or a concrete LPE path.
* No confidentiality or integrity impact. There is no visible information leak or write primitive.
* No network reachability. This is a local ioctl / GPU runtime initialization path.

Call-site confidence: high. The trace shows the ioctl path through KFD VM acquisition and the patch directly adjusts the mismatched reserved VA and allocation sizing constants.

## 3. Reachability analysis

The bug is triggerable by a local user who can access the AMDGPU/KFD compute device interface, commonly via render, video, or ROCm-related device permissions. It does not normally require full root or CAP_SYS_ADMIN if the user already has access to the GPU device node.

Namespaces and containers may matter if GPU device nodes are passed into containers. In that case, a containerized or reduced-privilege user with GPU access may still be able to trigger the host kernel crash.

The path is not network reachable. It requires local code execution and an affected AMDGPU/KFD setup on a 64K PAGE_SIZE kernel. The practical exposure is therefore narrower than a common local syscall bug, but realistic on systems using ROCm or AMD GPU compute with this page-size configuration.

## 4. Severity interpretation

This behaves like an ordinary-to-borderline Moderate local DoS issue, not an Important-class memory corruption vulnerability. The demonstrated impact is a kernel crash caused by a size mismatch between the reserved trap VA area and the KFD CWSR TBA/TMA allocation size on 64K page systems.

Theoretical memory corruption potential is weak because an allocation/reservation mismatch exists in GPU VM setup, but the actual observed failure is an invalid pointer / NULL dereference during mutex handling. There is no evidence of controlled overwrite, stale object reuse, page-cache corruption, UAF reclaim, or privilege escalation.

## 5. One-sentence report phrase

AMDGPU/KFD on 64K PAGE_SIZE systems can crash during local ROCm/KFD VM initialization because the reserved trap VA area is smaller than the computed CWSR TBA/TMA allocation, allowing a local GPU user to trigger a kernel DoS without evidence of privilege escalation.

## 6. Manual review recommendation

MANUAL CHECK RECOMMENDED

Reason: the issue is a local low-privilege kernel crash path in a GPU ioctl/runtime initialization flow, and device-node access may be delegated to non-root users or containers. However, no concrete memory corruption or LPE primitive is visible, so it does not require Important-class escalation based on the provided patch alone.

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

Patch: drm/amdgpu: Change AMDGPU_VA_RESERVED_TRAP_SIZE to 64KB
Commit: 6b2614a0ff05a2d2836311425091c8feca6f0c21
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=6b2614a0ff05a2d2836311425091c8feca6f0c21

Commit description:

Currently, AMDGPU_VA_RESERVED_TRAP_SIZE is hardcoded to 8KB, while
KFD_CWSR_TBA_TMA_SIZE is defined as 2 * PAGE_SIZE. On systems with
4K pages, both values match (8KB), so allocation and reserved space
are consistent.

However, on 64K page-size systems, KFD_CWSR_TBA_TMA_SIZE becomes 128KB,
while the reserved trap area remains 8KB. This mismatch causes the
kernel to crash when running rocminfo or rccl unit tests.

Kernel attempted to read user page (2) - exploit attempt? (uid: 1001)
BUG: Kernel NULL pointer dereference on read at 0x00000002
Faulting instruction address: 0xc0000000002c8a64
Oops: Kernel access of bad area, sig: 11 [#1]
LE PAGE_SIZE=64K MMU=Radix SMP NR_CPUS=2048 NUMA pSeries
CPU: 34 UID: 1001 PID: 9379 Comm: rocminfo Tainted: G E
6.19.0-rc4-amdgpu-00320-gf23176405700 #56 VOLUNTARY
Tainted: [E]=UNSIGNED_MODULE
Hardware name: IBM,9105-42A POWER10 (architected) 0x800200 0xf000006
of:IBM,FW1060.30 (ML1060_896) hv:phyp pSeries
NIP:  c0000000002c8a64 LR: c00000000125dbc8 CTR: c00000000125e730
REGS: c0000001e0957580 TRAP: 0300 Tainted: G E
MSR:  8000000000009033 <SF,EE,ME,IR,DR,RI,LE> CR: 24008268
XER: 00000036
CFAR: c00000000125dbc4 DAR: 0000000000000002 DSISR: 40000000
IRQMASK: 1
GPR00: c00000000125d908 c0000001e0957820 c0000000016e8100
c00000013d814540
GPR04: 0000000000000002 c00000013d814550 0000000000000045
0000000000000000
GPR08: c00000013444d000 c00000013d814538 c00000013d814538
0000000084002268
GPR12: c00000000125e730 c000007e2ffd5f00 ffffffffffffffff
0000000000020000
GPR16: 0000000000000000 0000000000000002 c00000015f653000
0000000000000000
GPR20: c000000138662400 c00000013d814540 0000000000000000
c00000013d814500
GPR24: 0000000000000000 0000000000000002 c0000001e0957888
c0000001e0957878
GPR28: c00000013d814548 0000000000000000 c00000013d814540
c0000001e0957888
NIP [c0000000002c8a64] __mutex_add_waiter+0x24/0xc0
LR [c00000000125dbc8] __mutex_lock.constprop.0+0x318/0xd00
Call Trace:
0xc0000001e0957890 (unreliable)
__mutex_lock.constprop.0+0x58/0xd00
amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu+0x6fc/0xb60 [amdgpu]
kfd_process_alloc_gpuvm+0x54/0x1f0 [amdgpu]
kfd_process_device_init_cwsr_dgpu+0xa4/0x1a0 [amdgpu]
kfd_process_device_init_vm+0xd8/0x2e0 [amdgpu]
kfd_ioctl_acquire_vm+0xd0/0x130 [amdgpu]
kfd_ioctl+0x514/0x670 [amdgpu]
sys_ioctl+0x134/0x180
system_call_exception+0x114/0x300
system_call_vectored_common+0x15c/0x2ec

This patch changes AMDGPU_VA_RESERVED_TRAP_SIZE to 64 KB and
KFD_CWSR_TBA_TMA_SIZE to the AMD GPU page size. This means we reserve
64 KB for the trap in the address space, but only allocate 8 KB within
it. With this approach, the allocation size never exceeds the reserved
area.

Fixes: 34a1de0 ("drm/amdkfd: Relocate TBA/TMA to opposite side of VM hole")
Reviewed-by: Christian König <christian.koenig@amd.com>
Suggested-by: Felix Kuehling <felix.kuehling@amd.com>
Suggested-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Donet Tom <donettom@linux.ibm.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 31b8de5 )
Cc: stable@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Changed files:
  drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
  drivers/gpu/drm/amd/amdkfd/kfd_priv.h

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=6b2614a0ff05a2d2836311425091c8feca6f0c21

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

The original Linux kernel CVE announcement for CVE-2026-31765 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-31765

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:L/PR:L/UI:N/S:U/C:N/I:N/A:H
  The Best / paranoid CVSS score: 5.5

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).