AVID-2026-R0754

Description

Division by 0 in QuantizedBiasAdd (CVE-2021-29546)

Details

TensorFlow is an end-to-end open source platform for machine learning. An attacker can trigger an integer division by zero undefined behavior in tf.raw_ops.QuantizedBiasAdd. This is because the implementation of the Eigen kernel(https://github.com/tensorflow/tensorflow/blob/61bca8bd5ba8a68b2d97435ddfafcdf2b85672cd/tensorflow/core/kernels/quantization_utils.h#L812-L849) does a division by the number of elements of the smaller input (based on shape) without checking that this is not zero. The fix will be included in TensorFlow 2.5.0. We will also cherrypick this commit on TensorFlow 2.4.2, TensorFlow 2.3.3, TensorFlow 2.2.3 and TensorFlow 2.1.4, as these are also affected and still in supported range.

Reason for inclusion in AVID: CVE-2021-29546 describes a software vulnerability in TensorFlow (an AI framework), specifically a divide-by-zero issue in QuantizedBiasAdd. This directly affects AI/ML software stacks and pipelines, is a component-level issue in the AI supply chain (TensorFlow versions used to build/deploy ML systems), and is identified as a security vulnerability with documented evidence (description, CVSS, affected versions, and references). Therefore it fits AVID criteria for AI-related software supply-chain vulnerabilities and should be kept.

References

• NVD entry
• https://github.com/tensorflow/tensorflow/security/advisories/GHSA-m34j-p8rj-wjxq
• https://github.com/tensorflow/tensorflow/commit/67784700869470d65d5f2ef20aeb5e97c31673cb

Affected or Relevant Artifacts

• Developer: tensorflow
• Deployer: tensorflow
• Artifact Details:

Impact

AVID Taxonomy Categorization

• Risk domains: Security
• SEP subcategories: S0100: Software Vulnerability
• Lifecycle stages: L06: Deployment

CVSS

CWE

Other information

• Report Type: Advisory
• Credits:
• Date Reported: 2021-05-14
• Version: 0.3.3
• AVID Entry