Linux Kernel GPU Suspension Interrupt Synchronization Vulnerability

A vulnerability in the Linux kernel's handling of GPU power management can lead to kernel crashes. This issue arises because the runtime power management suspend callback does not wait for interrupt request (IRQ) handlers to finish on other CPU cores before suspending the GPU. As a result, an IRQ handler could be active while the GPU is suspended, causing a crash when the system tries to access GPU registers. This vulnerability affects the Linux kernel versions prior to 6.16, specifically in the GPU driver for Imagination Technologies' Rogue architecture.

Impact

Failing to properly synchronize interrupts before suspending the GPU can lead to kernel panics. On platforms like the Texas Instruments AM68 SK, this mismanagement can cause asynchronous SError interrupts, which are critical hardware errors, leading to a complete system halt.

Reproduction

The vulnerability can be reproduced by triggering a GPU power off sequence through the runtime power management suspend callback while an IRQ handler is still processing on a different CPU core. This can be simulated by generating safety events that invoke the IRQ handler, then initiating the power off sequence before the handler has completed its work.

Remediation

Users can update to the latest stable version of the Linux kernel, where this vulnerability has been addressed. Instructions for downloading the latest version can be found on the official Linux kernel website.