Cloudanix VulnDB

CVE-2026-99090

None

CVSS

—

EPSS

0.2%

Published

Oct 26, 2026

CISA due date

—

Cloudanix analysis

Schedule a fix SSVC: Track

A timing bug in AMD graphics drivers could cause warning messages during graphics card recovery, making systems slightly less stable until fixed.

Exploitation
Local Privilege
Attacker needs
Local User
Complexity
Moderate
Impact
Other
Component
OS kernel
Class
Race Condition
Remediation
Patch
Effort
Low
Patch risk
Low

What it actually does

Race condition in amdgpu vblank counter callback allows concurrent display register and scanout reads without the reset_domain semaphore during GPU reset, triggering lockdep_assert_held and TAINT_WARN.

Why it matters to the business

GPU recovery on AMD systems can produce kernel warnings or tainted flags, reducing reliability and complicating diagnostics in graphics workloads.

How an attack unfolds

An attacker with local access triggers repeated GPU resets while display commit paths invoke the vblank callback, causing unprotected hardware reads that violate the reset semaphore contract.

How to fix it

Apply the upstream fix that acquires the reset semaphore for read before hardware access and falls back to cached software vblank count when reset is active.

How to confirm the fix worked

Boot with lockdep enabled, trigger GPU resets under display load, and confirm absence of the amdgpu_device_skip_hw_access lockdep warning and TAINT_WARN.

Typical attack path

  1. Obtain local code execution or trigger display activity
  2. Force or wait for GPU reset to acquire reset semaphore for write
  3. Invoke vblank counter or scanout position read concurrently
  4. Trigger unprotected hardware access and lockdep warning

How to detect it

Review kernel logs for amdgpu_device_skip_hw_access lockdep failures and TAINT_WARN entries around GPU reset timestamps; correlate with display commit activity.

  • amdgpu_device_skip_hw_access
  • TAINT_WARN
  • reset_domain->sem

What your security team should do

  • Prioritize kernel updates containing commit 51a4359f7431a4be5e42803c8bdaf7e0e5be2934 on AMD GPU systems
  • Monitor dmesg for TAINT_WARN and lockdep assertions during reset events
  • Test graphics recovery paths in staging environments before production rollout

The record

In the Linux kernel, the following vulnerability has been resolved: drm/amdgpu: serialize vblank counter reads against GPU reset The vblank counter callback reads display registers and scanout position without holding reset_domain->sem. During GPU recovery, display commit work can call the callback while reset owns the semaphore for writing. The hardware access then triggers the lockdep_assert_held() in amdgpu_device_skip_hw_access() and sets TAINT_WARN when lockdep is enabled. Try to acquire the reset semaphore for reading before accessing the hardware, and hold it across the counter and scanout-position reads. If reset is already in progress, return DRM's cached software vblank count instead. This avoids blocking behind reset and avoids accessing display hardware while it is being reset. The cached count is less accurate than the hardware count, but is used only when reset prevents a safe hardware read. This addresses the reset-time warning from this vblank path; it does not suppress or fix unrelated lockdep warnings. (cherry picked from commit 51a4359f7431a4be5e42803c8bdaf7e0e5be2934)

Affected packages

Ecosystem Package Introduced Fixed