{
  "schema_version": "1.6.1",
  "id": "RPI-CVE-2026-63848",
  "modified": "2026-10-09T11:39:44Z",
  "upstream": [
    "CVE-2026-63848"
  ],
  "published": "2026-07-19T14:05:10Z",
  "summary": "drm/amdgpu/jpeg: set no_user_fence for JPEG v2.0 ring",
  "details": "In the Linux kernel, the following vulnerability has been resolved:\n\ndrm/amdgpu/jpeg: set no_user_fence for JPEG v2.0 ring\n\nJPEG rings do not support 64-bit user fence writes, reject CS\nsubmissions with user fences.\n\n(cherry picked from commit 96179da0c6b059eb31706a0abe8dd6381c533143)",
  "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-2026-63848"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/f675801889b265634aefd30aa4503fc2b9e6ce1c"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/2e216c2ff159b2eb1da6e9c716d727efc73c64b5"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/b41248d1c18384835f6532e68592ee07605da283"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/41c4f3f68a343d62bd352a95ace93a11c4ad92ed"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/e5f612dc91650561fe2b5b76dd6d2898ec9ad480"
    }
  ],
  "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 reached only through DRM_IOCTL_AMDGPU_CS on an AMDGPU render node (/dev/dri/renderD*), requiring a local ioctl to submit a command buffer with AMDGPU_CHUNK_ID_FENCE to the JPEG v2.0 decode ring (AMDGPU_HW_IP_VCN_JPEG).\nAC:L - On affected VCN 2.0.x hardware with JPEG v2.0 (e.g. Navi/Vega20), an attacker fully controls CS submission parameters and can reliably trigger the unsupported 64-bit user fence path without races or conditions beyond attacker control.\nPR:L - DRM_IOCTL_AMDGPU_CS is gated by DRM_AUTH|DRM_RENDER_ALLOW, so any unprivileged local user with render-node access (typically render/video group membership) can reach the vulnerable path without root or init-namespace capabilities.\nUI:N - Exploitation requires only programmatic ioctl submission by the attacker; no victim interaction such as opening files, mounting filesystems, or clicking prompts is needed.\nS:U - The flaw enables GPU-initiated memory corruption within the attacker's own AMDGPU context/VM via the render node, which is a standard local kernel/GPU-driver privilege boundary rather than a cross-authority escape such as VM guest-to-host or IOMMU bypass.\nC:H - Submitting a 64-bit user fence to a JPEG ring that only supports 32-bit fence writes causes the hardware to perform misdirected or out-of-bounds GPU memory writes, providing a memory-corruption primitive that can be leveraged for information disclosure.\nI:H - The unsupported fence write gives the attacker a GPU-initiated arbitrary-write primitive at attacker-chosen GPU virtual addresses within their VM mappings, which is memory corruption exploitable for further integrity compromise or code execution.\nA:H - Incorrect fence handling on the JPEG decode ring can hang or fault the GPU, trigger driver error-recovery paths, and cause system-wide graphics disruption or kernel oops, constituting high availability impact."
    }
  }
}