The Linux kernel vulnerability CVE-2026-64292 involved veventq memory allocation inside a spinlock in the iommufd subsystem. Because the queue depth is user-controlled, this could lead to exhaustion of atomic memory reserves, causing kernel instability or denial of service. The fix moves the allocation outside the spinlock and uses GFP_NOWAIT to fail fast under memory pressure without consuming atomic reserves or triggering direct reclaim in IRQ context. On failure, the kernel queues a lost_events_header to notify user space and returns -ENOMEM to indicate kernel memory pressure. This behavior is distinct from normal queue overflow handling, which returns success but signals event loss. A subsequent patch will cap the veventq depth to further mitigate risk.

This vulnerability could allow user space to exhaust atomic memory reserves in the kernel by requesting large vevent queue depths, potentially leading to denial of service or kernel instability. The issue affects kernel memory management under specific conditions related to iommufd vevent queues. There is no indication of privilege escalation or code execution. No known exploits are reported in the wild.

A fix has been implemented that moves veventq memory allocation outside the spinlock and uses GFP_NOWAIT to prevent atomic memory exhaustion. This fix returns appropriate error codes on allocation failure and signals event loss to user space. Users should update to a Linux kernel version that includes this fix. Patch status is not explicitly confirmed in the provided data; users should consult the official Linux kernel advisories or vendor updates for the exact fixed versions and apply them accordingly.