During the execution of scx_enable() in the Linux kernel, the thread's scheduling class is switched from the fair class to the ext class. Because the fair class has a higher scheduling priority, workloads that saturate fair-class tasks can prevent the enable thread from running, causing the system to hang indefinitely. The issue arose when the protection mechanism changed from preempt_disable() to scx_bypass(), which does not guard against this starvation. The fix moves the enable work to a dedicated real-time (SCHED_FIFO) kernel thread that cannot be starved by either fair or ext class tasks, eliminating the deadlock path.

All Linux kernel implementations that contain the unpatched scx_enable() logic are affected, regardless of distribution. Both the default scheduler and any extended (sched_ext) implementations that rely on scx_enable() may experience the starvation problem. The vulnerability does not specify a particular kernel release, so any kernel built before the patch commit is potentially vulnerable.

The CVSS score is not disclosed in the available data, and no EPSS value is provided, indicating that the exploitation likelihood has not been quantified. The vulnerability is not listed in CISA's KEV catalog, suggesting no known widespread exploitation at this time. The prerequisite for exploitation is the saturation of fair-class workloads, which may be engineered or opportunistic; therefore, the exploitation difficulty could be low to moderate in environments with high fair-class load, but the impact of a single successful event is severe, leading to a permanent system halt until reboot.