{
  "schema_version": "1.6.1",
  "id": "RPI-CVE-2026-53009",
  "modified": "2026-10-09T11:39:44Z",
  "upstream": [
    "CVE-2026-53009"
  ],
  "published": "2026-06-24T16:31:02Z",
  "summary": "ice: fix double-free of tx_buf skb",
  "details": "In the Linux kernel, the following vulnerability has been resolved:\n\nice: fix double-free of tx_buf skb\n\nIf ice_tso() or ice_tx_csum() fail, the error path in\nice_xmit_frame_ring() frees the skb, but the 'first' tx_buf still points\nto it and is marked as valid (ICE_TX_BUF_SKB).\n'next_to_use' remains unchanged, so the potential problem will\nlikely fix itself when the next packet is transmitted and the tx_buf\ngets overwritten. But if there is no next packet and the interface is\nbrought down instead, ice_clean_tx_ring() -> ice_unmap_and_free_tx_buf()\nwill find the tx_buf and free the skb for the second time.\n\nThe fix is to reset the tx_buf type to ICE_TX_BUF_EMPTY in the error\npath, so that ice_unmap_and_free_tx_buf().\nMove the initialization of 'first' up, to ensure it's already valid in\ncase we hit the linearization error path.\n\nThe bug was spotted by AI while I had it looking for something else.\nIt also proposed an initial version of the patch.\n\nI reproduced the bug and tested the fix by adding code to inject\nfailures, on a build with KASAN.\n\nI looked for similar bugs in related Intel drivers and did not find any.",
  "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": null,
            "compiled_filter": false
          }
        }
      }
    }
  ],
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://www.cve.org/CVERecord?id=CVE-2026-53009"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/4c08fc2119ef0281cfa2cee007acf0a251be55f2"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/1a303baa715e6b78d6a406aaf335f87ff35acfcd"
    }
  ],
  "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": 7.8,
      "cvss_v3_severity": "High",
      "scenario": "AV:L - The vulnerable code is in the ice driver's local transmit path reached through socket/netdev transmission, not by parsing received network packets directly. A remote peer may induce ordinary replies, but the malformed/offload-error SKB state needed for this bug is locally generated.\nAC:L - There is no race; once an attacker can create an SKB that makes ice_tso() or ice_tx_csum() fail, the stale tx_buf state is deterministic. In a controlled netdev namespace or delegated ice/SF deployment, the attacker can also quiesce/bring down the interface before the next packet overwrites the slot.\nPR:L - The strongest reasonable path requires local ability to generate crafted/offloaded transmit SKBs and control the affected netdev lifecycle, which can be namespace-level CAP_NET_ADMIN over a delegated ice netdev rather than host init-namespace root. It is therefore more than unauthenticated remote access but not necessarily high host administrator privilege.\nUI:N - No victim user action is required in the attacker-controlled local scenario; the attacker can send the triggering packet and then cause cleanup. The bug is completed by driver cleanup, not by a separate user opening or mounting content.\nS:U - The corruption occurs within the host kernel/driver memory management context and affects the same security authority. This is not a KVM guest-to-host, IOMMU, or separate-scope boundary bypass.\nC:H - This is a double-free of an skb pointer retained in the TX ring, a heap memory corruption condition equivalent to a use-after-free primitive. Such corruption can plausibly be shaped into kernel memory disclosure.\nI:H - A double-free of a kernel heap object can corrupt allocator metadata or a reallocated object and can plausibly be leveraged for arbitrary write or control-flow impact. Under the required conservative scoring rule, this warrants high integrity impact.\nA:H - Even without full exploitation, the second dev_kfree_skb_any() on an already freed skb can trigger allocator/KASAN/slab corruption detection or a kernel crash. Repeated triggering can therefore cause high availability impact."
    }
  }
}