Attack ChainAIDerived
Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain
-
AccessGain local low-privileged access
-
DeliveryAllocate AMDXDNA_BO_SHARE command-chain buffer
-
ExploitRace command_count to zero in shared memory
-
ExecutionTrigger __counted_by bounds-check failure
-
ImpactKernel panic / host crash
Vulnerability Assessment
| Weakness |
CWE
OWASP Top 10:2025
A06:2025 - Insecure Design
ASVS 5.0
V15 - Secure Coding and Architecture Concurrency control, locking
MITRE ATT&CK
1 technique across 1 tactic: Privilege Escalation The technique list with ATT&CK links and the kill-chain view is part of the Basic plan and up. |
| Exploitation Conditions | Requires a local authenticated user (PR:L) on a system that has AMD XDNA NPU hardware with the accel/amdxdna kernel driver loaded. The attacker must be able to allocate an AMDXDNA_BO_SHARE shared buffer object holding a struct amdxdna_cmd_chain and submit a command chain, then concurrently modify the userspace-controlled command_count field in that shared memory (a race) so the __counted_by(command_count) bounds check fails. Limiting factors: no remote vector (AV:L), requires the specific AMD NPU hardware and driver to be present and accessible, requires local access rather than network reachability, and yields only a crash - no code execution, data disclosure, or integrity impact. |
| Impact | This is a genuine but low-urgency issue. All signals agree on a modest, local denial-of-service risk rather than a top priority. The CVSS vector (CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H) confirms the flaw is only reachable by an already-authenticated local user (PR:L), with no confidentiality or integrity impact and availability-only consequences (kernel panic). EPSS is very low at 0.15% (4th percentile), consistent with the local-only vector and the narrow hardware requirement. There is no CISA KEV listing and no public exploit identified at time of analysis, so this is neither actively exploited nor weaponized. The one notable data conflict is the input tag 'Information Disclosure,' which contradicts the CVSS impact metrics (C:N/I:N/A:H) - this is a race-condition denial-of-service bug, not an information-disclosure issue, and the tag appears mislabeled. Overall this is a correctly-rated MEDIUM (5.5) local DoS, not a high-CVSS-low-real-risk inflation nor an underrated threat; its real-world reach is constrained by the need for both local access and AMD XDNA NPU hardware with the amdxdna driver loaded. |
| Exploit ScenarioDerived | A local, low-privileged user on a system with an AMD XDNA NPU and the amdxdna accelerator driver allocates an AMDXDNA_BO_SHARE buffer object containing an amdxdna_cmd_chain structure, then concurrently rewrites the userspace-controlled command_count field (e.g., to zero) while the kernel processes the command chain. The __counted_by bounds check derived from that attacker-mutable field fails and triggers a kernel panic, crashing the host. No public exploit is identified at time of analysis, and exploitation requires local access plus the specific AMD NPU hardware/driver. |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv1 – Oct 21, 2026
Kernel panic in the Linux amdxdna accelerator driver (the DRM accel driver for AMD XDNA NPUs) lets a local user crash the system by racing a modification of the userspace-controlled command_count field of a struct amdxdna_cmd_chain that is stored in shared AMDXDNA_BO_SHARE memory, so the compiler-generated __counted_by bounds check fails. Only systems with AMD XDNA NPU hardware, the accel/amdxdna driver loaded, and an attacker with local access to the accelerator device node are affected; the impact is denial of service via crash, with no confidentiality or integrity impact. No public exploit code was identified at time of analysis, EPSS is 0.15 percent (4th percentile), and vendor fixes are available in Linux 6.18.53, 7.2.7 and 7.3-rc2.
RemediationAI
Vendor-released patch: update to Linux 6.18.53, 7.2.7 or 7.3-rc2 (or a later release) that contains the amdxdna fixes at commits 93fa3e925b15b0ded0a549fe7f12bfbb1c4e171a, 52f3e086760a9a3e02a46a10b57caffd73b1c204 and b3709d354545e70388177500761f92d906c4dfd6; distribution kernels should be updated when the respective backports ship. Where patching cannot be done immediately, restrict access to the AMD XDNA accelerator device nodes (typically /dev/accel/accel*) to trusted local users via ownership and mode or a device cgroup, which shrinks the attack surface but does not eliminate the race for those still permitted. On hosts that do not need NPU acceleration, blacklisting or not loading the amdxdna module removes the exposure entirely, at the cost of disabling the NPU/accelerator functionality and any workloads depending on it. Advisory details are available through https://nvd.nist.gov/vuln/detail/CVE-2026-98146 and the referenced kernel.org stable commits.
Affected VersionsENISA
Affected Packages
Source: OSV (Open Source Vulnerabilities)
Affects your stack
Affects 1 package in euvd
Checking this CVE against your own stacks and servers is part of the Basic plan and up.
Technical ContextAI · show details
The flaw is in the accel/amdxdna driver's command-chain handling. struct amdxdna_cmd_chain ends in a flexible array member annotated with __counted_by(command_count), a kernel/compiler feature that instruments the structure so that the flexible-array bounds sanitizer derives the array length from the count field at access time. Because an amdxdna_cmd_chain instance is placed in AMDXDNA_BO_SHARE shared buffer memory that is mapped into userspace, command_count is not a trusted kernel value: a userspace process can rewrite it concurrently while the kernel validates and walks the command chain. If the count is driven to zero, the generated check on the flexible array access fails and the sanitizer raises a trap, producing a kernel panic (denial of service). The fix removes the __counted_by annotation so the driver no longer uses the userspace-controlled field as the authoritative flexible-array bound. CWE is not assigned in the provided data, but the root-cause class is consistent with a race condition combined with reliance on an attacker-controlled length value (CWE-362/CWE-129 style).
Know when Amd gets its next vulnerability
- Watch products and get an email when something new is published
- Check this CVE against your own stack and servers
- Exploit and PoC references, threat intelligence from 22 sources
No credit card · 7-day full trial
Severity by source
AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H
Local user with device access can race a shared-memory field to crash the kernel, so AV:L/PR:L/AC:L with availability-only impact (A:H) and no C/I effect.
Primary rating from NVD.
CVSS VectorNVD
Lifecycle Timeline
4DescriptionNVD
In the Linux kernel, the following vulnerability has been resolved:
accel/amdxdna: Remove __counted_by from struct amdxdna_cmd_chain
struct amdxdna_cmd_chain contains a flexible array annotated with __counted_by(command_count). Since the structure is stored in shared AMDXDNA_BO_SHARE memory, userspace can modify command_count concurrently. If command_count is changed to zero, the bounds check generated from __counted_by may fail and trigger a kernel panic.
Remove __counted_by to avoid relying on the userspace-controlled command_count for the flexible array bounds check.
Threat Intelligence Overview
How this decision was reached
The engine reads each signal below and applies a fixed list of rules from top to bottom. The first rule that matches sets the recommended action – like a statement, every line is shown.
What every action level means
SSVC AssessmentENISA
Priority Score
Attack Technique
Race conditions occur when multiple concurrent requests exploit the brief time gap between checking a condition and acting upon it (TOCTOU - Time of Check, Time of Use).
On BIG-IP versions 16.0.x before 16.0.1.1, 15.1.x before 15.1.2.1, 14.1.x before 14.1.4, 13.1.x before 13.1.3.6, and 12.
Local privilege escalation and information disclosure in the Linux kernel on AMD Zen2 CPUs allows low-privileged users t
An exploitable code execution vulnerability exists in the Shader functionality of AMD Radeon DirectX 11 Driver atidxx64.
An exploitable code execution vulnerability exists in the Shader functionality of AMD Radeon DirectX 11 Driver atidxx64.
An exploitable code execution vulnerability exists in the Shader functionality of AMD Radeon DirectX 11 Driver atidxx64.
An exploitable memory corruption vulnerability exists in AMD atidxx64.dll 26.20.15019.19000 graphics driver. Rated criti
plays_service.exe in the plays.tv service before 1.27.7.0, as distributed in AMD driver-installation packages and Gaming
A flaw was found in the KVM's AMD code for supporting SVM nested virtualization. Rated high severity (CVSS 8.8), this vu
AMD ATI atillk64.sys 5.11.9.0 allows low-privileged users to interact directly with physical memory by calling one of se
An exploitable out-of-bounds read vulnerability exists in AMD ATIDXX64.DLL driver, version 26.20.13001.29010. Rated high
AMD fglrx-driver before 15.9 allows local users to gain privileges via a symlink attack. Rated high severity (CVSS 7.8),
AMD fglrx-driver before 15.7 allows local users to gain privileges via a symlink attack. Rated high severity (CVSS 7.8),
EUVD-2026-86977
GHSA-hq4v-3xxp-hfhg
NVD
CVE.org
EPSS
Exploit-DB