{"id":"CVE-2026-99065","published":"2026-10-23T13:17:30.367","lastModified":"2026-10-23T13:17:30.367","description":"In the Linux kernel, the following vulnerability has been resolved:\n\ncgroup/cpuset: Handle cpu hotplug race in guarantee_active_cpus()\n\nWith commit 2125c0034c5d (\"cgroup/cpuset: Make cpuset hotplug processing\nsynchronous\"), the cpuset hotplug operation becomes synchronous. That\ncommit also removes the code that handles race between cpuset_hotplug_work\nand cpu hotplug notifier with the assumption that race is now gone.\nLater commit 7a0aabd9ce69 (\"cgroup/cpuset: Always use cpu_active_mask\")\nupdates the cpuset code to always use cpu_active_mask instead of\ncpu_online_mask in some places including guarantee_online_cpus() which\nis renamed to guarantee_active_cpus() in that commit.\n\nIn the case of CPU offline operation, cpu_active_mask is updated\nfirst in sched_cpu_deactivate() to remove the offline CPU before\ncpuset_handle_hotplug() is called to update the effective_cpus of the\naffected cpusets. The cpu_online_mask is updated after that near the end\nof the offline operation to remove the offline CPU. As a result, the race\ncomes back and the top cpuset may not have any active CPU leading to NULL\npointer dereference during the race window when guarantee_active_cpus()\nis called after cpu_active_mask is updated to remove the CPU to be torn\ndown but before cpuset_handle_hotplug() is able to properly update the\neffective_cpus of the top cpuset.\n\nFix this by adding back the NULL cs check to avoid this problem.","cvssScore":null,"cvssSeverity":null,"cvssVector":null,"cwes":[],"vendors":[],"products":[],"references":[{"url":"https://git.kernel.org/stable/c/23b1ab15ca91f2cffaf149d9bb0fafeb700c7db2","tags":[]},{"url":"https://git.kernel.org/stable/c/8d88d3221204fbb1b6dc2c0f2f2181fe49126639","tags":[]},{"url":"https://git.kernel.org/stable/c/a21c255159bfe0034740292e10d2e258feeeb10d","tags":[]}],"exploitRefs":[],"hasPoc":false,"ai":null}