The Linux kernel io_uring/io-wq subsystem had a flaw where io_worker_handle_work() took a snapshot of the IO_WQ_BIT_EXIT flag before processing a chain of linked work items but did not re-check this flag for each linked item. If the exit bit was set after the first linked item started, subsequent linked items could incorrectly continue execution by reusing stale state, skipping cancellation logic. The patch moves the exit bit check deeper inside the work processing loop to ensure all linked work items respect the exit condition. This vulnerability was identified through syzbot testing on devices with slow linked work items.

The vulnerability could allow linked work items to continue running after the kernel has initiated exit procedures for the io_uring work queue. This could potentially lead to unexpected behavior or resource usage during shutdown sequences. No known exploits are reported in the wild. The impact is limited to the internal kernel work queue processing and does not indicate direct privilege escalation or remote code execution based on the provided data.

A fix has been committed to the Linux kernel source code addressing this issue by improving the exit bit checking logic for linked work items. Users should apply the official kernel update containing commit 10dc95939817 or later to remediate this vulnerability. Since this is a kernel-level fix, updating the kernel to a patched version is the recommended mitigation.