CVE-2026-23213
PCIe MMIO Access During AMD SMU Reset Causes System Hang
Publication date: 2026-02-18
Last updated on: 2026-03-18
Assigner: kernel.org
Description
Description
CVSS Scores
EPSS Scores
| Probability: | |
| Percentile: |
Meta Information
Affected Vendors & Products
| Vendor | Product | Version / Range |
|---|---|---|
| linux | linux_kernel | 6.19 |
| linux | linux_kernel | 6.19 |
| linux | linux_kernel | 6.19 |
| linux | linux_kernel | 6.19 |
| linux | linux_kernel | From 6.13 (inc) to 6.18.10 (exc) |
| linux | linux_kernel | From 5.9 (inc) to 6.12.70 (exc) |
Helpful Resources
Exploitability
| CWE ID | Description |
|---|---|
| CWE-UNKNOWN |
Attack-Flow Graph
AI Powered Q&A
Can you explain this vulnerability to me?
This vulnerability exists in the Linux kernel's AMD DRM driver during the SMU Mode 1 reset process. When the ASIC undergoes a Mode 1 reset, it becomes temporarily inaccessible via PCIe. If any component tries to access MMIO registers during this reset window, it can cause uncompleted PCIe transactions.
These uncompleted transactions may lead to Non-Maskable Interrupt (NMI) panics or system hangs, causing instability or crashes.
The fix involves setting a flag (`no_hw_access`) immediately after triggering the reset to prevent any hardware register access while the device is offline, along with a memory barrier to ensure this flag update is visible to all processor cores.
How can this vulnerability impact me? :
This vulnerability can cause system instability by triggering NMI panics or system hangs during the reset process of the AMD GPU hardware.
Such instability can lead to unexpected system crashes or freezes, potentially resulting in data loss or disruption of services running on the affected system.
How does this vulnerability affect compliance with common standards and regulations (like GDPR, HIPAA)?:
I don't know
How can this vulnerability be detected on my network or system? Can you suggest some commands?
I don't know
What immediate steps should I take to mitigate this vulnerability?
To mitigate this vulnerability, ensure that the Linux kernel includes the fix which disables MMIO access during the SMU Mode 1 reset.
Specifically, after triggering the Mode 1 reset on the AMD DRM driver, the 'no_hw_access' flag should be set to true immediately. This prevents any MMIO register accesses during the reset window, avoiding uncompleted PCIe transactions that can cause NMI panics or system hangs.
Additionally, a memory barrier (smp_mb()) is used to ensure the flag update is visible to all cores before the driver enters a sleep or wait state.