
Cloud Vulnerability DB
A community-led vulnerabilities database
CVE-2026-99254 is a Denial of Service vulnerability in the Linux kernel's io_uring subsystem, specifically in the task_work handling during the exit path. The flaw causes excessive callback queuing that can pin a CPU and prevent a task from completing cancellation or other critical operations for extended periods. It affects Linux kernel versions starting from 5.12, with fixes backported to stable branches at 6.1.191, 6.6.161, 6.12.115, 6.18.60, 7.2.14, and 7.3.2. Disclosed on November 3, 2026, it carries a CVSS v3.1 base score of 5.5 (Medium) (GitHub Advisory, Microsoft).
The root cause is a logic flaw in io_tctx_exit_cb(): when io_ring_exit_work() attempts to remove task context (tctx) nodes via task_work callbacks, the callback re-queues itself immediately if the target task is concurrently executing io_uring_cancel_generic(), which defers node removal to its own completion. This creates a feedback loop where up to 64 work items continuously re-queue callbacks to a single task, preventing task_work_run() from ever finding an empty list and returning. No privilege is required to trigger this condition — a local user can craft a scenario with multiple io_uring ring pairs that poll each other, then call exit_group() or execve(), generating hundreds of thousands of callbacks within seconds. The fix introduces a delay before re-queuing the callback when the task is detected to be in cancellation, breaking the tight loop (GitHub Advisory).
Successful exploitation results in a sustained Denial of Service: the affected task monopolizes a CPU core and is unable to complete its exit or exec operation, potentially for minutes or longer. In a demonstrated scenario, a task with 60 pairs of mutually polling rings executed 280,000 callbacks in 5 seconds, 20 seconds after exit_group(). An exiting task cannot be killed during this period, and a live task executing execve() can similarly become stuck, impacting system availability and responsiveness. There is no confidentiality or integrity impact (GitHub Advisory).
There is no public proof-of-concept exploit and no evidence of in-the-wild exploitation at this time. The vulnerability requires only local access with low privileges — no special permissions are needed to use io_uring. The EPSS score is approximately 0.197% (9th percentile), indicating a low near-term exploitation probability. The vulnerability is not listed in the CISA Known Exploited Vulnerabilities catalog (GitHub Advisory).
exit_group() (or execve() for a live-task variant) to initiate the kernel's io_uring exit path (io_ring_exit_work()).io_tctx_exit_cb() detects the task is in io_uring_cancel_generic() and immediately re-queues itself, causing tens of thousands of callbacks to flood the task's work queue within seconds.io_uring_cancel_generic() or task_work_run() for an abnormally long time (minutes) after calling exit_group() or execve(); the process cannot be killed.WARN_ON_ONCE(time_after(jiffies, timeout)) messages from io_ring_exit_work() appearing in dmesg or kernel logs, indicating the exit work exceeded the 5-minute timeout.perf or /proc tracing during a process exit sequence (GitHub Advisory).Apply the Linux kernel patches backported to the following stable versions: 6.1.191, 6.6.161, 6.12.115, 6.18.60, 7.2.14, and 7.3.2. Patches are also available via the kernel stable git tree (commits a22fb44e988b, 2c75ebae8a7f, 61c761fc21e0, 14f56e9c0b8f, abdc57c7ebab, f334e79a11e0, 22e643e61b9b). As an interim workaround, restrict or disable io_uring access for untrusted local users using seccomp policies or kernel boot parameters (e.g., io_uring_disabled=1 if available on the kernel version). Microsoft has also issued a patch for the Azure Linux 3 kernel (azl3_kernel_6.6.159.1-1) (GitHub Advisory, Microsoft).
The vulnerability was covered in the context of Microsoft's November 2026 Patch Tuesday, which addressed 1,235 flaws including this Linux kernel issue affecting Azure Linux. Debian has accepted the fix into both unstable and oldstable-security branches. The Yocto Project security mailing list also flagged the issue for embedded Linux maintainers. No notable independent researcher commentary or significant social media discussion specific to this CVE has been identified beyond standard vulnerability tracking coverage.
Fix availability across major Linux distributions and their releases.
bookworm
linux-6.12: 6.12.115-2~deb12u1
sid
linux: 7.2.14-1
trixie
linux: 6.12.115-1
bionic
linux
bionic (esm-infra)
linux
bionic (fips-updates)
linux-fips
bionic (fips)
linux-fips
devel
linux
focal
linux
focal (esm-infra)
linux-hwe-5.15
focal (fips-updates)
linux-fips
Source: This report was generated using AI
Free Vulnerability Assessment
Evaluate your cloud security practices across 9 security domains to benchmark your risk level and identify gaps in your defenses.
Get a personalized demo
"Best User Experience I have ever seen, provides full visibility to cloud workloads."
"Wiz provides a single pane of glass to see what is going on in our cloud environments."
"We know that if Wiz identifies something as critical, it actually is."