{"id":"CVE-2025-39886","title":"bpf: Tell memcg to use allow_spinning=false path in bpf_timer_init()","summary":"In the Linux kernel, the following vulnerability has been resolved:\n\nbpf: Tell memcg to use allow_spinning=false path in bpf_timer_init()\n\nCurrently, calling bpf_map_kmalloc_node() from __bpf_async_init() can\ncause various locking issues…","severity":"medium","cvss":5.5,"cvssVector":"CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H","cvssSource":"adp","vendor":"Linux","product":"Linux","affected":["Linux >= b00628b1c7d595ae5b544e059c27b1f5828314b4 < 3e42431a705be8417ac5b696daa1dd43b3d27169","Linux >= b00628b1c7d595ae5b544e059c27b1f5828314b4 < 449682e76f32601f211816d3e2100bed87e67a4c","Linux >= b00628b1c7d595ae5b544e059c27b1f5828314b4 < cd1fd26bb13473c1734e3026b2b97025a0a4087b","Linux >= b00628b1c7d595ae5b544e059c27b1f5828314b4 < ac70cd446f83ccb25532b343919ab86eacdcd06a","Linux >= b00628b1c7d595ae5b544e059c27b1f5828314b4 < 6d78b4473cdb08b74662355a9e8510bde09c511e","Linux 5.15"],"ssvc":{"exploitation":"none","automatable":"no","technicalImpact":"partial","timestamp":"2026-01-14T19:26:19.815693Z"},"published":"2025-09-23","updated":"2026-10-30","sourceUpdated":"2026-10-30T12:22:35.281Z","source":"CVEORG","sourceUrl":"https://www.cve.org/CVERecord?id=CVE-2025-39886","references":[{"url":"https://git.kernel.org/stable/c/3e42431a705be8417ac5b696daa1dd43b3d27169"},{"url":"https://git.kernel.org/stable/c/449682e76f32601f211816d3e2100bed87e67a4c"},{"url":"https://git.kernel.org/stable/c/cd1fd26bb13473c1734e3026b2b97025a0a4087b"},{"url":"https://git.kernel.org/stable/c/ac70cd446f83ccb25532b343919ab86eacdcd06a"},{"url":"https://git.kernel.org/stable/c/6d78b4473cdb08b74662355a9e8510bde09c511e"}],"tags":["cve.org"],"epss":0.00151,"epssPercentile":0.04299,"ingestedAt":"2026-11-03T21:55:50.622Z","slug":"CVE-2025-39886","body":"## Overview\n\nIn the Linux kernel, the following vulnerability has been resolved:\n\nbpf: Tell memcg to use allow_spinning=false path in bpf_timer_init()\n\nCurrently, calling bpf_map_kmalloc_node() from __bpf_async_init() can\ncause various locking issues; see the following stack trace (edited for\nstyle) as one example:\n\n...\n [10.011566]  do_raw_spin_lock.cold\n [10.011570]  try_to_wake_up             (5) double-acquiring the same\n [10.011575]  kick_pool                      rq_lock, causing a hardlockup\n [10.011579]  __queue_work\n [10.011582]  queue_work_on\n [10.011585]  kernfs_notify\n [10.011589]  cgroup_file_notify\n [10.011593]  try_charge_memcg           (4) memcg accounting raises an\n [10.011597]  obj_cgroup_charge_pages        MEMCG_MAX event\n [10.011599]  obj_cgroup_charge_account\n [10.011600]  __memcg_slab_post_alloc_hook\n [10.011603]  __kmalloc_node_noprof\n...\n [10.011611]  bpf_map_kmalloc_node\n [10.011612]  __bpf_async_init\n [10.011615]  bpf_timer_init             (3) BPF calls bpf_timer_init()\n [10.011617]  bpf_prog_xxxxxxxxxxxxxxxx_fcg_runnable\n [10.011619]  bpf__sched_ext_ops_runnable\n [10.011620]  enqueue_task_scx           (2) BPF runs with rq_lock held\n [10.011622]  enqueue_task\n [10.011626]  ttwu_do_activate\n [10.011629]  sched_ttwu_pending         (1) grabs rq_lock\n...\n\nThe above was reproduced on bpf-next (b338cf849ec8) by modifying\n./tools/sched_ext/scx_flatcg.bpf.c to call bpf_timer_init() during\nops.runnable(), and hacking the memcg accounting code a bit to make\na bpf_timer_init() call more likely to raise an MEMCG_MAX event.\n\nWe have also run into other similar variants (both internally and on\nbpf-next), including double-acquiring cgroup_file_kn_lock, the same\nworker_pool::lock, etc.\n\nAs suggested by Shakeel, fix this by using __GFP_HIGH instead of\nGFP_ATOMIC in __bpf_async_init(), so that e.g. if try_charge_memcg()\nraises an MEMCG_MAX event, we call __memcg_memory_event() with\n@allow_spinning=false and avoid calling cgroup_file_notify() there.\n\nDepends on mm patch\n\"memcg: skip cgroup_file_notify if spinning is not allowed\":\nhttps://lore.kernel.org/bpf/20250905201606.66198-1-shakeel.butt@linux.dev/\n\nv0 approach s/bpf_map_kmalloc_node/bpf_mem_alloc/\nhttps://lore.kernel.org/bpf/20250905061919.439648-1-yepeilin@google.com/\nv1 approach:\nhttps://lore.kernel.org/bpf/20250905234547.862249-1-yepeilin@google.com/\n\n## Affected\n\n- `Linux >= b00628b1c7d595ae5b544e059c27b1f5828314b4 < 3e42431a705be8417ac5b696daa1dd43b3d27169`\n- `Linux >= b00628b1c7d595ae5b544e059c27b1f5828314b4 < 449682e76f32601f211816d3e2100bed87e67a4c`\n- `Linux >= b00628b1c7d595ae5b544e059c27b1f5828314b4 < cd1fd26bb13473c1734e3026b2b97025a0a4087b`\n- `Linux >= b00628b1c7d595ae5b544e059c27b1f5828314b4 < ac70cd446f83ccb25532b343919ab86eacdcd06a`\n- `Linux >= b00628b1c7d595ae5b544e059c27b1f5828314b4 < 6d78b4473cdb08b74662355a9e8510bde09c511e`\n- `Linux 5.15`\n\n## Remediation\n\nRefer to the linked advisories for vendor-supplied fixes and affected version ranges.","depth":"sunlit","depthScore":30,"depthScoreParts":{"impact":30.3,"likelihood":0,"exploitation":0,"ransomware":0},"changes":[]}