In the Linux kernel, the following vulnerability has been resolved: writeback: report a Tasks-RCU quiescent state per cgwb drain pass… (CVE-2026-98536)

The vulnerability in the Linux kernel's writeback subsystem arises from the cleanup_offline_cgwbs_workfn() function draining a dying cgroup writeback domain by repeatedly calling cleanup_offline_cgwb() with a cond_resched() call between passes. On CONFIG_PREEMPTION kernels with PREEMPT_DYNAMIC full or lazy modes (such as arm64 and x86 architectures), cond_resched() is effectively disabled, causing the drain loop to never report a Tasks-RCU quiescent state. This results in long-running drain operations that block synchronize_rcu_tasks() calls, leading to tasks stuck in uninterruptible sleep (D state) and triggering the hung task detector, which can panic the system. The fix adds a call to cond_resched_tasks_rcu_qs() within the drain loop to ensure progress and prevent stalls.

The vulnerability can cause prolonged kernel worker tasks during cgroup writeback draining, which blocks synchronize_rcu_tasks() calls. This blocking can lead to numerous tasks stuck in uninterruptible sleep, triggering the kernel's hung task detector and causing system panics. This affects system stability and reliability, particularly on arm64 and x86 architectures running kernels with PREEMPT_DYNAMIC full or lazy preemption modes.

A fix has been implemented in the Linux kernel to address this issue by adding calls to cond_resched_tasks_rcu_qs() in the drain loop. Users should update to the fixed kernel versions once available. Since this is a kernel-level fix, applying the official kernel update from the vendor or distribution is the recommended remediation. Patch status is not explicitly confirmed in the provided data; users should consult their Linux distribution's security advisories for the official fix availability.