CVE-2026-99799: io_uring/sqpoll: don't wait for the thread to exit
In the Linux kernel, the following vulnerability has been resolved:
iouring/sqpoll: don't wait for the thread to exit
iosqthreadstop() wakes the SQPOLL thread and waits for it to exit. The thread belongs to the thread group that created the ring, so a SIGSTOP or the cgroup freezer holds it in getsignal(), and the wake up doesn't change that. ioringexitwork() then sleeps in that wait until someone continues or kills the process. iouexit has 64 slots, so 64 closed SQPOLL rings of a stopped process keep every other ring from being freed, for everyone.
Don't wait. Set the stop bit, wake the thread, and let it exit when it gets to run again. The sqd can now outlive its last ring, so give it a second refcount with one reference for the rings and one for the thread, and free it with the last one. The thread can't reach a freed ring, rings leave ctxlist before dropping their reference, and the io-wq hash is refcounted.
Affected Software
Remediation
Recommended actions to resolve this vulnerability, in priority order.
- Upgrade
Upgrade
linux/Kernelto a version that resolves this vulnerability.Fixed in 6.12.117 - Upgrade
Upgrade
linux/Kernelto a version that resolves this vulnerability.Fixed in 6.18.61 - Upgrade
Upgrade
linux/Kernelto a version that resolves this vulnerability.Fixed in 7.3.2 - Upgrade
Upgrade
debian/linuxto a version that resolves this vulnerability.Fixed in 6.12.117-1 - Upgrade
Upgrade
debian/linuxto a version that resolves this vulnerability.Fixed in 7.3.3-1 - Upgrade
Upgrade
debian/linuxto a version that resolves this vulnerability.Fixed in 6.12.117-1Fixed in 7.3.4-1Fixed in 7.3.5-1
Frequently Asked Questions
Who is realistically exposed to this issue?
Systems using io_uring rings configured for SQPOLL are exposed. The impact is broader than the process that created the ring: enough affected closed rings from a stopped process can prevent other io_uring rings from being freed system-wide.
What conditions are needed to trigger the resource blockage?
A process must create SQPOLL rings and then be stopped with SIGSTOP or held by the cgroup freezer while its SQPOLL thread is in get_signal(). Closing 64 such rings can fill the 64 io_uring exit-work slots and block freeing of other rings until the stopped process is continued or killed.
What can be done if the fix cannot be applied immediately?
Avoid using SQPOLL for rings created by processes that may be stopped or frozen. If the condition has already occurred, continue or kill the stopped or frozen process so its SQPOLL thread can run and exit.