Linux Kernel AMD GPU Driver SDMA ECC IRQ Handling Vulnerability

A vulnerability in the Linux kernel AMD GPU driver related to the SDMA (Synchronous Data Memory Access) version 4.0 has been addressed. The issue arose because the driver improperly managed the ECC (Error-Correcting Code) interrupt request (IRQ) for certain ASICs (Application-Specific Integrated Circuits). Specifically, the driver would disable the ECC IRQ without checking if SDMA ECC was enabled, leading to warnings during the suspend cycle on affected chips. The vulnerability was present in the Linux kernel's stable releases prior to the application of the patch that corrected this behavior.

Impact

The vulnerability could cause unnecessary warnings during the suspend cycle on certain AMD chips with SDMA IP version 4.0, where SDMA ECC is not enabled.

Reproduction

The vulnerability can be reproduced on AMD chips that have SDMA IP version 4.0 and where SDMA ECC is not enabled. When the system is suspended, the driver will incorrectly disable the ECC IRQ, causing a warning to be generated.

Remediation

Users can upgrade to the latest version of the Linux kernel where this vulnerability has been addressed. The specific commit that resolves this issue is available in the Linux kernel stable tree.