{
  "id": "CVE-2026-80975",
  "url": "https://poc.griffinguard.io/cve/CVE-2026-80975",
  "published_at": "2026-09-11T19:42:37.086Z",
  "title": "mfd: qnap-mcu: keep the reply buffer alive past a command timeout",
  "vendor": "Linux",
  "product": "Linux",
  "cvss": {
    "score": 7.8,
    "source": "cna",
    "vector": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H"
  },
  "kev": {
    "in_kev": false,
    "added_at": null,
    "days_after_publication": null
  },
  "first_poc": null,
  "poc_repo_count": 0,
  "exploit_count": 0,
  "suspicious_repo_count": 0,
  "edge_device": null,
  "description": "In the Linux kernel, the following vulnerability has been resolved:\n\nmfd: qnap-mcu: keep the reply buffer alive past a command timeout\n\nqnap_mcu_exec() publishes an on-stack buffer to the receive path:\n\n\tunsigned char rx[QNAP_MCU_RX_BUFFER_SIZE];\n\t...\n\treply-\u003edata = rx;\n\treply-\u003elength = length;\n\nand qnap_mcu_receive_buf() writes into it from the serdev receive path,\nwhich runs out of flush_to_ldisc() and is not serialized against\nqnap_mcu_exec() at all. bus_lock cannot cover it, because qnap_mcu_exec()\nholds that mutex across wait_for_completion_timeout().\n\nOn a timeout qnap_mcu_exec() returns with reply-\u003edata still pointing at\nits own frame. A reply that arrives late, or an unsolicited message from\nthe MCU, is then written into a stack frame that has been left, corrupting\nwhatever runs next on that stack. The same applies when qnap_mcu_write()\nfails, since that path returns without touching the reply state either.\n\nMove the receive buffer into struct qnap_mcu. It is 37 bytes and the\nstructure is devm_kzalloc()ed, so it lives as long as the driver, and a\nlate write lands in memory that is still valid and is reinitialized by the\nnext command. bus_lock keeps commands from sharing it.\n\nThis deliberately does not clear reply-\u003edata or reply-\u003elength on the\ntimeout path. Doing so races with qnap_mcu_receive_buf(), which reads both\nafter its\n\n\tif (!reply-\u003elength)\n\t\treturn size;\n\ncheck: clearing reply-\u003edata gives a NULL dereference, and clearing\nreply-\u003elength alone removes the reply-\u003ereceived == reply-\u003elength exit\ncondition, so the copy loop runs until the uart chunk is consumed and\noverruns the buffer. Leaving both set keeps the write bounded by\nreply-\u003elength, which qnap_mcu_exec() has already checked against\nsizeof(mcu-\u003erx).",
  "cvss_by_source": {
    "cna": 7.8
  },
  "repos": [],
  "remediation": {
    "affected": [
      {
        "vendor": "Linux",
        "product": "Linux",
        "ranges": [
          "\u003e= 998f70d1806bb718a7565f350283e4a79c8cbb4b, \u003c 0b6680e306397097a97767447368221fac753809",
          "\u003e= 998f70d1806bb718a7565f350283e4a79c8cbb4b, \u003c 8391ee06d08845a0a165b5fe679ba326915166b2",
          "\u003e= 998f70d1806bb718a7565f350283e4a79c8cbb4b, \u003c 47504742cea7878ebd1bf1491bbed923df6b90b1"
        ],
        "fixed_in": [
          "0b6680e306397097a97767447368221fac753809",
          "8391ee06d08845a0a165b5fe679ba326915166b2",
          "47504742cea7878ebd1bf1491bbed923df6b90b1"
        ]
      },
      {
        "vendor": "Linux",
        "product": "Linux",
        "ranges": [
          "6.14"
        ],
        "fixed_in": [
          "0",
          "6.18.51",
          "7.2.4",
          "7.3"
        ]
      }
    ],
    "upgrade": [
      "Upgrade Linux to 0b6680e306397097a97767447368221fac753809 or 8391ee06d08845a0a165b5fe679ba326915166b2 or 47504742cea7878ebd1bf1491bbed923df6b90b1 or later.",
      "Upgrade Linux to 0 or 6.18.51 or 7.2.4 or 7.3 or later."
    ]
  },
  "exploit_refs": [],
  "hunt_profile": [],
  "rules": []
}
