In the Linux kernel, the following vulnerability has been resolved: openvswitch: fix soft lockup in the netlink flow dump A production compute node… (CVE-2026-99025)

The Linux kernel's openvswitch netlink flow dump function (ovs_flow_cmd_dump) called ovs_flow_stats_get for each flow, releasing a spinlock with spin_unlock_bh multiple times per CPU. Each release triggered local_bh_enable, which ran pending softirq work in the dumping thread's context. On busy CPUs, the softirq backlog refilled as fast as it drained, causing the dumping thread to act as the CPU's softirq engine without yielding, resulting in a soft lockup. The fix was to hold bottom halves off across the entire callback, preventing nested spin_unlock_bh calls from draining softirqs prematurely and allowing proper CPU scheduling under CONFIG_PREEMPT.

This vulnerability could cause a soft lockup and kernel panic on production compute nodes handling many datapath flows, leading to system instability or downtime. It affects the Linux kernel's openvswitch component during netlink flow dumps, potentially impacting network function and reliability on affected systems.

A fix has been implemented in the Linux kernel to hold bottom halves off across the entire callback in ovs_flow_cmd_dump, preventing the soft lockup condition. Users should update to a Linux kernel version that includes this fix. Patch status is not explicitly confirmed in the provided data; check the vendor advisory or Linux kernel release notes for the exact fixed versions and update accordingly.