CVE-2026-99130

In the Linux kernel, the following vulnerability has been resolved:

io uring/cancel: don't spin on local work the task can't run

io uring cancel generic() refuses to sleep while any ring on the exiting task's tctx has local work pending. It assumes the task is that ring's submitter and that the next io uring try cancel requests() pass will run the work. This does not hold for a ring that was still IORING SETUP R DISABLED when the task got its tctx node from io uring create() or io ringfd register(). The submitter is set later by IORING REGISTER ENABLE RINGS, possibly from another task, and only the submitter may run local work on an IORING SETUP DEFER TASKRUN ring. To the exiting task the work stays pending forever, the loop never reaches schedule(), and the task trips the WARN ON ONCE() once and then spins in do exit().

An unprivileged user can set this up at will by creating a disabled ring, having a child enable it and stop, posting to the ring with IORING OP MSG RING and exiting with a request in flight that only the child can complete. The exiting thread burns a CPU until the child runs or dies, and with panic on warn set the WARN panics the kernel.

Fix by skipping the sleep only when io allowed defer tw run() allows the task to run the work. This is reasonable since work it cannot run cannot complete any of its requests either, so nothing is lost by sleeping. Drop the WARN ON ONCE() since handing a disabled ring to another thread is a supported scenario, not a bug state.