{ "schema_version": "1.6.1", "id": "RPI-CVE-2026-31557", "modified": "2026-10-09T11:39:44Z", "upstream": [ "CVE-2026-31557" ], "published": "2026-04-24T14:37:34Z", "summary": "nvmet: move async event work off nvmet-wq", "details": "In the Linux kernel, the following vulnerability has been resolved:\n\nnvmet: move async event work off nvmet-wq\n\nFor target nvmet_ctrl_free() flushes ctrl->async_event_work.\nIf nvmet_ctrl_free() runs on nvmet-wq, the flush re-enters workqueue\ncompletion for the same worker:-\n\nA. Async event work queued on nvmet-wq (prior to disconnect):\n nvmet_execute_async_event()\n queue_work(nvmet_wq, &ctrl->async_event_work)\n\n nvmet_add_async_event()\n queue_work(nvmet_wq, &ctrl->async_event_work)\n\nB. Full pre-work chain (RDMA CM path):\n nvmet_rdma_cm_handler()\n nvmet_rdma_queue_disconnect()\n __nvmet_rdma_queue_disconnect()\n queue_work(nvmet_wq, &queue->release_work)\n process_one_work()\n lock((wq_completion)nvmet-wq) <--------- 1st\n nvmet_rdma_release_queue_work()\n\nC. Recursive path (same worker):\n nvmet_rdma_release_queue_work()\n nvmet_rdma_free_queue()\n nvmet_sq_destroy()\n nvmet_ctrl_put()\n nvmet_ctrl_free()\n flush_work(&ctrl->async_event_work)\n __flush_work()\n touch_wq_lockdep_map()\n lock((wq_completion)nvmet-wq) <--------- 2nd\n\nLockdep splat:\n\n ============================================\n WARNING: possible recursive locking detected\n 6.19.0-rc3nvme+ #14 Tainted: G N\n --------------------------------------------\n kworker/u192:42/44933 is trying to acquire lock:\n ffff888118a00948 ((wq_completion)nvmet-wq){+.+.}-{0:0}, at: touch_wq_lockdep_map+0x26/0x90\n\n but task is already holding lock:\n ffff888118a00948 ((wq_completion)nvmet-wq){+.+.}-{0:0}, at: process_one_work+0x53e/0x660\n\n 3 locks held by kworker/u192:42/44933:\n #0: ffff888118a00948 ((wq_completion)nvmet-wq){+.+.}-{0:0}, at: process_one_work+0x53e/0x660\n #1: ffffc9000e6cbe28 ((work_completion)(&queue->release_work)){+.+.}-{0:0}, at: process_one_work+0x1c5/0x660\n #2: ffffffff82d4db60 (rcu_read_lock){....}-{1:3}, at: __flush_work+0x62/0x530\n\n Workqueue: nvmet-wq nvmet_rdma_release_queue_work [nvmet_rdma]\n Call Trace:\n __flush_work+0x268/0x530\n nvmet_ctrl_free+0x140/0x310 [nvmet]\n nvmet_cq_put+0x74/0x90 [nvmet]\n nvmet_rdma_free_queue+0x23/0xe0 [nvmet_rdma]\n nvmet_rdma_release_queue_work+0x19/0x50 [nvmet_rdma]\n process_one_work+0x206/0x660\n worker_thread+0x184/0x320\n kthread+0x10c/0x240\n ret_from_fork+0x319/0x390\n\nMove async event work to a dedicated nvmet-aen-wq to avoid reentrant\nflush on nvmet-wq.", "severity": [ { "type": "CVSS_V3", "score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H" } ], "affected": [ { "package": { "ecosystem": "Raspberry Pi OS:12", "name": "linux" }, "versions": [ "1:6.1.47-1+rpt4" ], "ecosystem_specific": { "binaries": [ { "binary_name": "linux-image-6.1.0-rpi3-rpi-2712", "binary_version": "1:6.1.47-1+rpt4" }, { "binary_name": "linux-image-6.1.0-rpi3-rpi-v6", "binary_version": "1:6.1.47-1+rpt4" }, { "binary_name": "linux-image-6.1.0-rpi3-rpi-v7", "binary_version": "1:6.1.47-1+rpt4" }, { "binary_name": "linux-image-6.1.0-rpi3-rpi-v7l", "binary_version": "1:6.1.47-1+rpt4" }, { "binary_name": "linux-image-6.1.0-rpi3-rpi-v8", "binary_version": "1:6.1.47-1+rpt4" } ] }, "database_specific": { "per_version": { "1:6.1.47-1+rpt4": { "linux_commit": "655fc658a15ae7a6f37103754adb39ba52a9a14e", "upstream_version": "6.1.47", "assessment": "commit ancestry (strak rule)", "upstream_fixed": null, "compiled_filter": false } } } } ], "references": [ { "type": "ADVISORY", "url": "https://www.cve.org/CVERecord?id=CVE-2026-31557" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/49c7c50ee6325a084216e94395e067ecde8088fa" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/ca111c9d8d6c9d5735878d933a1716c4be86c2d1" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/25ceffc1dabec3b93f458b437aae26f4da293f87" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/2922e3507f6d5caa7f1d07f145e186fc6f317a4e" } ], "database_specific": { "source": "https://git.kernel.org/pub/scm/linux/security/vulns.git", "upstream_modified": "2026-08-05T13:16:28Z", "severity": { "source": "kernel CNA", "cvss_v3_score": 7.5, "cvss_v3_severity": "High", "scenario": "AV:N - The vulnerable async event and teardown path is reachable through NVMe-oF target networking, including nvmet-tcp, by connecting to the target, issuing an admin Async Event Request, and disconnecting.\nAC:L - The attacker controls both relevant actions: queuing async event work and triggering connection teardown. No uncontrollable memory layout or victim timing condition is required beyond repeating the network sequence.\nPR:N - The NVMe discovery controller is reachable with the standard discovery NQN, allows hosts by design, supports async event requests, and skips DH-HMAC-CHAP authentication. In an exposed nvmet-tcp deployment this is reachable without credentials.\nUI:N - No victim user action is needed after the target service is exposed. The remote peer can trigger the sequence by protocol messages and disconnecting.\nS:U - The impact is confined to the kernel/NVMe target host that processes the malicious connection. It does not cross a VM, IOMMU, or separate security authority boundary.\nC:N - The bug is a workqueue/flush recursion issue and does not expose kernel memory or target data. There is no identified read primitive.\nI:N - The issue does not corrupt memory or provide a write/control-flow primitive. It affects teardown/workqueue progress rather than data integrity.\nA:H - The recursive workqueue flush can trigger a kernel lockdep warning and can lead to a panic or teardown hang in hardened/debug configurations. It can be triggered repeatedly over the network, so availability impact is scored high." } } }