The bug in io_uring's io‑wq causes a worker thread to lose a wake‑signal when it exits, so a user‑space process that submits an asynchronous FIFO open never wakes and waits indefinitely. The result is that the submitting thread remains blocked, potentially exhausting kernel resources and degrading system stability. This is a denial‑of‑service condition that can be triggered by a normal FIFO open without a writer.

The flaw exists in the Linux kernel’s io_uring implementation. No specific kernel version is listed, meaning all kernels that contain the current io_uring code are potentially affected. Systems running any version of the Linux kernel that has not yet received the upstream fix should be considered vulnerable.

The vulnerability has no published EPSS score and is not listed in the CISA KEV catalog, but the missing wake signal introduces a race condition that is exploitable with user‑level access. An attacker who can submit an io_uring request that opens a FIFO without a partner can cause the target process to stall and, if repeated, may lead to resource exhaustion. Because the fault occurs during thread cleanup, its impact is limited to the affected process, but repeated exploitation could degrade overall system performance.