{"id":"CVE-2026-99085","title":"sched_ext: Fix CPU hotplug hang when a dying CPU's tasks sit in the BPF scheduler","summary":"In the Linux kernel, the following vulnerability has been resolved:\n\nsched_ext: Fix CPU hotplug hang when a dying CPU's tasks sit in the BPF scheduler\n\nA CPU going down has to empty its own rq. Its hotplug thread waits in\nsched_cpu_wait_…","severity":"none","vendor":"Linux","product":"Linux","affected":["Linux >= f0e1a0643a59bf1f922fa209cec86a170b784f3f < f79cd7a2df3768323fbfb9b0f96adff08fb7eb4c","Linux >= f0e1a0643a59bf1f922fa209cec86a170b784f3f < 638ee6bca19750934ea13f9c325f46fd8ae60064","Linux >= f0e1a0643a59bf1f922fa209cec86a170b784f3f < 173ff9d03ec6373c3158b6e630af33a9b36fd216","Linux >= f0e1a0643a59bf1f922fa209cec86a170b784f3f < d35a535d3e3e71f92a051d94e415f43612c4edfc","Linux 6.12"],"published":"2026-10-23","updated":"2026-10-23","sourceUpdated":"2026-10-23T12:45:03.465Z","source":"CVEORG","sourceUrl":"https://www.cve.org/CVERecord?id=CVE-2026-99085","references":[{"url":"https://git.kernel.org/stable/c/f79cd7a2df3768323fbfb9b0f96adff08fb7eb4c"},{"url":"https://git.kernel.org/stable/c/638ee6bca19750934ea13f9c325f46fd8ae60064"},{"url":"https://git.kernel.org/stable/c/173ff9d03ec6373c3158b6e630af33a9b36fd216"},{"url":"https://git.kernel.org/stable/c/d35a535d3e3e71f92a051d94e415f43612c4edfc"}],"tags":["cve.org"],"ingestedAt":"2026-10-23T13:24:55.509Z","slug":"CVE-2026-99085","body":"## Overview\n\nIn the Linux kernel, the following vulnerability has been resolved:\n\nsched_ext: Fix CPU hotplug hang when a dying CPU's tasks sit in the BPF scheduler\n\nA CPU going down has to empty its own rq. Its hotplug thread waits in\nsched_cpu_wait_empty() until nothing else is left, and the only thing that\nwakes it is balance_push(), which runs from __schedule() on the dying CPU\nand pushes off the migratable tasks the CPU picks.\n\nsched_ext breaks this. A task sitting on a user DSQ or held by the BPF\nscheduler is still counted on its rq, but an inactive CPU no longer calls\nops.dispatch() and can't pull the task back. The dying CPU goes idle with\nthe task still counted, and one of two things happens:\n\n- Another CPU consumes the task. The rq empties without a __schedule() on\n  the dying CPU, so the hotplug thread is never woken and cpu_down() hangs\n  holding cpu_hotplug_lock.\n- The task is affine only to the dying CPU. Nothing can move it, and the\n  offline stalls until the watchdog ejects the BPF scheduler.\n\nWhen the rq goes offline, re-enqueue every task on it that isn't already on\nthe local DSQ. An enqueue on an offline rq lands on the local DSQ, so the\ndying CPU itself picks the tasks and balance_push() pushes them off, the\nsame as for the other sched classes. From then on no sched_ext path on\nanother CPU can pull a task off the rq.\n\nops.dispatch() currently stops as soon as the CPU goes inactive, and CPU\nhotplug then waits for an RCU grace period before taking the rq offline. A\nBPF-held task affine only to the dying CPU and preempted inside an RCU\nread-side critical section would block that grace period, and the rq would\nnever go offline. Test SCX_RQ_ONLINE directly so that ops.dispatch() keeps\nrunning until the rq goes offline. Only the dying CPU's own hotplug thread\nclears the flag during teardown, and task_can_run_on_remote_rq() still keeps\nother rqs' tasks off the CPU.\n\n## Affected\n\n- `Linux >= f0e1a0643a59bf1f922fa209cec86a170b784f3f < f79cd7a2df3768323fbfb9b0f96adff08fb7eb4c`\n- `Linux >= f0e1a0643a59bf1f922fa209cec86a170b784f3f < 638ee6bca19750934ea13f9c325f46fd8ae60064`\n- `Linux >= f0e1a0643a59bf1f922fa209cec86a170b784f3f < 173ff9d03ec6373c3158b6e630af33a9b36fd216`\n- `Linux >= f0e1a0643a59bf1f922fa209cec86a170b784f3f < d35a535d3e3e71f92a051d94e415f43612c4edfc`\n- `Linux 6.12`\n\n## Remediation\n\nRefer to the linked advisories for vendor-supplied fixes and affected version ranges.","depth":"sunlit","depthScore":3,"depthScoreParts":{"impact":2.8,"likelihood":0,"exploitation":0,"ransomware":0},"changes":[]}