{
  "schema_version": "1.6.1",
  "id": "RPI-CVE-2025-39945",
  "modified": "2026-10-09T11:39:44Z",
  "upstream": [
    "CVE-2025-39945"
  ],
  "published": "2025-10-04T07:33:57Z",
  "summary": "cnic: Fix use-after-free bugs in cnic_delete_task",
  "details": "In the Linux kernel, the following vulnerability has been resolved:\n\ncnic: Fix use-after-free bugs in cnic_delete_task\n\nThe original code uses cancel_delayed_work() in cnic_cm_stop_bnx2x_hw(),\nwhich does not guarantee that the delayed work item 'delete_task' has\nfully completed if it was already running. Additionally, the delayed work\nitem is cyclic, the flush_workqueue() in cnic_cm_stop_bnx2x_hw() only\nblocks and waits for work items that were already queued to the\nworkqueue prior to its invocation. Any work items submitted after\nflush_workqueue() is called are not included in the set of tasks that the\nflush operation awaits. This means that after the cyclic work items have\nfinished executing, a delayed work item may still exist in the workqueue.\nThis leads to use-after-free scenarios where the cnic_dev is deallocated\nby cnic_free_dev(), while delete_task remains active and attempt to\ndereference cnic_dev in cnic_delete_task().\n\nA typical race condition is illustrated below:\n\nCPU 0 (cleanup)              | CPU 1 (delayed work callback)\ncnic_netdev_event()          |\n  cnic_stop_hw()             | cnic_delete_task()\n    cnic_cm_stop_bnx2x_hw()  | ...\n      cancel_delayed_work()  | /* the queue_delayed_work()\n      flush_workqueue()      |    executes after flush_workqueue()*/\n                             | queue_delayed_work()\n  cnic_free_dev(dev)//free   | cnic_delete_task() //new instance\n                             |   dev = cp->dev; //use\n\nReplace cancel_delayed_work() with cancel_delayed_work_sync() to ensure\nthat the cyclic delayed work item is properly canceled and that any\nongoing execution of the work item completes before the cnic_dev is\ndeallocated. Furthermore, since cancel_delayed_work_sync() uses\n__flush_work(work, true) to synchronously wait for any currently\nexecuting instance of the work item to finish, the flush_workqueue()\nbecomes redundant and should be removed.\n\nThis bug was identified through static analysis. To reproduce the issue\nand validate the fix, I simulated the cnic PCI device in QEMU and\nintroduced intentional delays \u2014 such as inserting calls to ssleep()\nwithin the cnic_delete_task() function \u2014 to increase the likelihood\nof triggering the bug.",
  "severity": [
    {
      "type": "CVSS_V3",
      "score": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/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": "6.1.154",
            "compiled_filter": false
          }
        }
      }
    }
  ],
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://www.cve.org/CVERecord?id=CVE-2025-39945"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/fde6e73189f40ebcf0633aed2b68e731c25f3aa3"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/7b6a5b0a6b392263c3767fc945b311ea04b34bbd"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/0405055930264ea8fd26f4131466fa7652e5e47d"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/e1fcd4a9c09feac0902a65615e866dbf22616125"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/8eeb2091e72d75df8ceaa2172638d61b4cf8929a"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/6e33a7eed587062ca8161ad1f4584882a860d697"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/0627e1481676669cae2df0d85b5ff13e7d24c390"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/cfa7d9b1e3a8604afc84e9e51d789c29574fb216"
    }
  ],
  "database_specific": {
    "source": "https://git.kernel.org/pub/scm/linux/security/vulns.git",
    "upstream_modified": "2026-08-05T12:19:11Z",
    "severity": {
      "source": "kernel CNA",
      "cvss_v3_score": 7.8,
      "cvss_v3_severity": "High",
      "scenario": "AV:L - The vulnerable teardown path is reached only through local administrative operations on the netdev (link down, MTU/ring change, driver unbind, netns move) and local iSCSI/FCoE offload management; no remotely supplied data reaches `cnic_cm_stop_bnx2x_hw()` or `cnic_delete_task()`.\nAC:L - The attacker controls both sides of the race \u2014 issuing iSCSI/FCoE connection destroys keeps `delete_task` in its 10 ms self-requeue cycle for the entire 2-second grace window while the interface-down is triggered, giving hundreds of attempts per try, and holding the cnic UIO chardev open stretches `cnic_stop_hw()`'s wait loop to 1.5 s to widen the window further.\nPR:L - In the delegated-NIC container deployment, a bnx2x port placed into a network namespace lets a user-namespace-root (unprivileged host user) with `CAP_NET_ADMIN` over that netns drive `NETDEV_GOING_DOWN` and, via netns teardown/device migration, the `NETDEV_UNREGISTER` that reaches `cnic_free_dev()` \u2014 so no real host root is needed for the free side.\nUI:N - The attacker performs both the connection-destroy activity and the interface teardown; no victim action or cooperation is required at any point.\nS:U - The use-after-free corrupts kernel slab memory and is exploited within the kernel's own security authority, with no crossing of a hypervisor, IOMMU, or sandbox boundary.\nC:H - The freed `cnic_dev`/`cnic_local` chunk and `cp->ctx_tbl` can be reclaimed by attacker-sprayed data, and the surviving work reads function pointers, DMA addresses, and ULP handles out of that freed memory, giving a path to disclose arbitrary kernel memory.\nI:H - `cnic_delete_task()` performs `test_and_clear_bit()` writes into the freed `ctx_tbl`, `kfree()`s pointers read from freed memory (arbitrary-free primitive), and makes an indirect call through `cp->ethdev->drv_ctl` located in the freed allocation \u2014 a control-flow hijack primitive after heap grooming.\nA:H - The stale `delayed_work` timer is armed on `kfree()`d memory and the callback dereferences freed structures, so even an unsuccessful exploitation attempt reliably produces slab corruption, an oops, or a kernel panic."
    }
  }
}