{
  "schema_version": "1.6.1",
  "id": "RPI-CVE-2026-63853",
  "modified": "2026-10-09T11:39:44Z",
  "upstream": [
    "CVE-2026-63853"
  ],
  "published": "2026-07-19T14:05:10Z",
  "summary": "drm/amdgpu/vcn: set no_user_fence for VCN v4.0 enc ring",
  "details": "In the Linux kernel, the following vulnerability has been resolved:\n\ndrm/amdgpu/vcn: set no_user_fence for VCN v4.0 enc ring\n\nVCN encoder and decoder rings do not support 64-bit user fence writes,\nreject CS submissions with user fences.\n\n(cherry picked from commit fd852c048b46f9825e904a4f3f4538fe9d8827d9)",
  "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-2026-63853"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/1286b6872de0aee1feeeaa6dbac86369806de9a5"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/6bdd2ed6458d35c368fbe9550a4d7f342abd3a92"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/51f694221047c84fa185be98210eb2c354ffb8c6"
    }
  ],
  "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 - Exploitation requires submitting a crafted AMDGPU command stream through DRM_IOCTL_AMDGPU_CS on a local /dev/dri/renderD* device; there is no network, adjacent-radio, or physical-access path to the vulnerable VCN v4.0 unified encoder ring handling.\nAC:L - An attacker with render-node access can deterministically submit a CS containing an AMDGPU_CHUNK_ID_FENCE chunk targeting the VCN encoder ring; no race, special timing, or victim state beyond having a VCN 4.0-capable AMDGPU is required.\nPR:L - The path is reachable by any local user who can open the DRM render node (typically membership in the render/video group on desktops, workstations, and cloud GPU VMs); it does not require real root in the init namespace.\nUI:N - Triggering the bug is fully programmatic via ioctl-based CS submission and does not require any victim to open files, mount filesystems, or perform other interactive actions.\nS:U - Impact is confined to the AMDGPU/VCN driver and GPU resources of the submitting security context; it does not cross a VM, container, or IOMMU security boundary to affect a different authority.\nC:H - Without the no_user_fence guard, amdgpu_ib_commit emits a 64-bit user fence (AMDGPU_FENCE_FLAG_64BIT) on a ring whose emit_fence path only supports 32-bit fence writes, causing the VCN hardware to perform an unvalidated GPU DMA write that can corrupt adjacent GPU-mapped memory and be leveraged for information disclosure.\nI:H - The same rogue GPU-initiated fence write is a memory-corruption primitive that can modify GPU-accessible buffers (including shared dma-buf mappings) outside the intended 8-byte fence field, enabling integrity violations and potential further exploitation.\nA:H - Malformed fence/trap commands on the VCN unified encoder ring can hang or fault the VCN block and force GPU/driver reset, denying graphics and compute availability to all users of the device on shared workstations and servers."
    }
  }
}