{
  "schema_version": "1.6.1",
  "id": "RPI-CVE-2026-68108",
  "modified": "2026-10-09T11:39:44Z",
  "upstream": [
    "CVE-2026-68108"
  ],
  "published": "2026-08-10T12:08:00Z",
  "summary": "drm/amdgpu/vce: fix integer overflow in image size",
  "details": "In the Linux kernel, the following vulnerability has been resolved:\n\ndrm/amdgpu/vce: fix integer overflow in image size\n\nFix a security vulnerability where malicious VCE command streams\nwith oversized dimensions (e.g. 65536\u00d765536) cause 32-bit integer\noverflow, wrapping the calculated buffer size to 0. This bypasses\nvalidation and allows GPU firmware to perform out-of-bound memory\naccess.\n\nThe fix uses 64-bit arithmetic to detect overflow and rejects\ninvalid dimensions before they reach the hardware.\n\nV2: remove redundant check\nV3: modify max height value\nV4: remove size64\n\n(cherry picked from commit cbe408dba581755ad1279a487ec786d8927d778d)",
  "severity": [
    {
      "type": "CVSS_V3",
      "score": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:C/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": "6.1.183",
            "compiled_filter": false
          }
        }
      }
    }
  ],
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://www.cve.org/CVERecord?id=CVE-2026-68108"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/893db20383800cfe92e638705984eebb13bc81a5"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/a07430abd556de3707adfcadcc60db3fa64e4b2b"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/a6d7065b91a14790980ce6f4960db0ca8c3c9940"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/7eebef042c12dfe0568593ee6a8926d16505925e"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/00c311a13d225266800c712f2b7db2711c6897de"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/186bfdc4e26d019b2e7570cb121964a1d89b2e5b"
    }
  ],
  "database_specific": {
    "source": "https://git.kernel.org/pub/scm/linux/security/vulns.git",
    "upstream_modified": "2026-08-19T16:47:12Z",
    "severity": {
      "source": "kernel CNA",
      "cvss_v3_score": 8.8,
      "cvss_v3_severity": "High",
      "scenario": "AV:L - The flaw is reached only through the AMDGPU_CS DRM ioctl on a local render node (/dev/dri/renderD*); there is no network, adjacent-radio, or physical-device path to the VCE command-stream parser.\nAC:L - On affected AMD GPUs using physical-mode VCE parsing, an attacker fully controls the IB and can set create-command dimensions (e.g., 65536x65536) to deterministically overflow the 32-bit size to zero without races or uncontrollable layout conditions.\nPR:L - Exploitation requires only unprivileged local access to the amdgpu DRM render node; AMDGPU_CS is gated by DRM_AUTH|DRM_RENDER_ALLOW with no root, DRM-master, or non-namespace capability requirement.\nUI:N - No victim interaction is needed beyond the attacker submitting crafted VCE command buffers through their own render-node context; no third party must open files or take other actions.\nS:C - Bypassing size-based BO relocation/validation lets VCE firmware DMA outside the intended buffer boundaries, crossing the GPU memory-isolation boundary that kernel validation is meant to enforce on GART/VRAM mappings.\nC:H - With validation defeated, VCE firmware can read arbitrarily beyond undersized encode/context/MV buffers, exposing other GPU-accessible memory including GTT-mapped system memory and adjacent VRAM allocations.\nI:H - The same out-of-bounds VCE firmware accesses enable arbitrary writes to GPU-mapped memory beyond validated buffers, providing memory-corruption primitives suitable for privilege escalation or control-flow influence.\nA:H - Oversized VCE operations against undersized buffers can trigger GPU page faults, engine hangs, GPU resets, or broader system disruption that the attacker can repeat via further command submissions."
    }
  }
}