Skip to content
OpenAdvisories
Low · 3.3 CVE-2026-63632 EUVD-2026-60944GHSA-p893-rvq9-2xf9PYSEC-2026-3587

ONNX: Heap-Buffer-Overflow READ in Gemm Version Converter Adapter via Undersized Input Shape

Published Aug 18, 2026 · Updated Aug 18, 2026 · Assigned by GitHub_M

Description

Open Neural Network Exchange (ONNX) is an open standard for machine learning interoperability. From 1.3.0 until 1.22.0, onnx.version_converter.convert_version() can perform an out-of-bounds read in Gemm_7_6::adapt_gemm_7_6() in onnx/version_converter/adapters/gemm_7_6.h when a Gemm node has input tensors with fewer than two dimensions because B_shape[1], A_shape[0], or A_shape[1] is accessed without a rank check, potentially causing a process crash during an opset 7 to 6 downgrade. This issue is fixed in version 1.22.0.

Affected products

VendorProductAffectedFixed
onnx onnx >= 1.3.0, < 1.22.0 —

Fixed in Linux distributions

Distributions backport fixes, so the version that fixes this vulnerability differs per release — upgrade to at least the version shown. From the distributions' security trackers and advisories.

ReleasePackageFixed inAdvisory
Debian 14 onnx No fix yet —
Debian 13 onnx No fix yet —
Debian 12 onnx No fix yet —
ReleasePackageFixed inAdvisory
Ubuntu 26.10 onnx No fix yet —
Ubuntu 26.04 LTS onnx No fix yet —
Ubuntu 22.04 LTS onnx No fix yet —
Show 1 more (older releases, extended support, other products)
ReleasePackageFixed inAdvisory
Ubuntu Pro 24.04 LTS onnx No fix yet —

Azure Linux

ReleasePackageFixed inAdvisory
Azure Linux 3 pytorch 2.2.2-17 —

Affected packages

EcosystemPackageAffectedFixedRecord
PyPI onnx >= 1.3.0, < 1.22.0 1.22.0 GHSA-p893-rvq9-2xf9

References