{
  "id": "CVE-2026-90246",
  "url": "https://poc.griffinguard.io/cve/CVE-2026-90246",
  "published_at": "2026-09-17T16:07:48.524Z",
  "title": "apparmor: fix integer overflow in verify_tags() bounds check",
  "vendor": "Linux",
  "product": "Linux",
  "cvss": {
    "score": 7.1,
    "source": "cna",
    "vector": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:N/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\napparmor: fix integer overflow in verify_tags() bounds check\n\nverify_tags() validates the tagset table unpacked from a policy blob.\nFor each set it reads a count and checks that advancing the index by\nthat count stays inside sets.table[]:\n\n\tu32 cnt = tags-\u003esets.table[i];\n\n\tif (i+cnt \u003e= tags-\u003esets.size) {\n\ni, cnt and sets.size are all u32, so i+cnt is evaluated modulo 2^32.\nsets.table[] is filled by unpack_tagsets() with aa_unpack_u32(), so\nevery entry is a raw unbounded 32-bit word taken from the policy blob,\nand verify_tags() is the function that is supposed to validate it.  A\ncount close to U32_MAX makes the sum wrap to a small value, the guard\npasses, and the inner loop then walks sets.table[++i] past the end of\nthe kcalloc(size, sizeof(u32)) allocation.\n\nNote that sets.size is bounded by 65535, because unpack_tagsets() reads\nit with aa_unpack_array() as a u16, so the wrap cannot be reached by\ngrowing the table; it is reached purely through the attacker-supplied\ncount.\n\nWith sets.size = 2 and sets.table = { 0, 0xffffffff }:\n\n  i = 0: cnt = 0, guard 0 + 0 \u003e= 2 is false, inner loop does not run\n  i = 1: cnt = 0xffffffff, guard (1 + 0xffffffff) mod 2^32 == 0 \u003e= 2 is\n         false, so the guard is bypassed and the inner loop reads\n         sets.table[2] -- one element past a two element allocation\n\nThe walk continues until an out-of-bounds value happens to be \u003e=\nhdrs.size or the access faults, so a crafted policy yields an\nout-of-bounds read on the policy load path\n(aa_replace_profiles -\u003e aa_unpack -\u003e unpack_policydb -\u003e unpack_tags -\u003e\nverify_tags).  unpack_tags() runs before the perms and DFA tables are\nunpacked, so no other table needs to be well formed to reach it.\n\nPolicy load is gated by aa_may_manage_policy(), which checks\nCAP_MAC_ADMIN relative to the subject's own user namespace rather than\nthe init user namespace, so with the default\nunprivileged_userns_apparmor_policy=1 the path is reachable from an\nunprivileged task in a matched-level nested namespace, not only by a\nglobally privileged one.\n\nPerform the addition in u64 so that it cannot wrap, restoring the\nintended i + cnt \u003c sets.size guarantee.",
  "cvss_by_source": {
    "cna": 7.1
  },
  "repos": [],
  "remediation": {
    "affected": [
      {
        "vendor": "Linux",
        "product": "Linux",
        "ranges": [
          "\u003e= 3d28e2397af7a89ac3de33c686ed404cda59b5d5, \u003c ef79a405f83993f0fda5efde371d98433f7d7a47",
          "\u003e= 3d28e2397af7a89ac3de33c686ed404cda59b5d5, \u003c 465946d3c560c0137e6a180ba054dddc4d049f5a"
        ],
        "fixed_in": [
          "ef79a405f83993f0fda5efde371d98433f7d7a47",
          "465946d3c560c0137e6a180ba054dddc4d049f5a"
        ]
      },
      {
        "vendor": "Linux",
        "product": "Linux",
        "ranges": [
          "7.0"
        ],
        "fixed_in": [
          "0",
          "7.2.6",
          "7.3"
        ]
      }
    ],
    "upgrade": [
      "Upgrade Linux to ef79a405f83993f0fda5efde371d98433f7d7a47 or 465946d3c560c0137e6a180ba054dddc4d049f5a or later.",
      "Upgrade Linux to 0 or 7.2.6 or 7.3 or later."
    ]
  },
  "exploit_refs": [],
  "hunt_profile": [],
  "rules": []
}
