AVID-2026-R0790

Description

Division by zero in TFLite’s implementation of TransposeConv (CVE-2021-29588)

Details

TensorFlow is an end-to-end open source platform for machine learning. The optimized implementation of the TransposeConv TFLite operator is vulnerable to a division by zero error. An attacker can craft a model such that stride_{h,w} values are 0. Code calling this function must validate these arguments. 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-2021-29588 describes a division-by-zero vulnerability in TensorFlow Lite’s TransposeConv operator. This is a software vulnerability in a widely used AI framework/component (TensorFlow/TFLite) that is directly involved in building, deploying, and running AI models. It affects AI software stacks and their runtimes, i.e., the supply chain for general-purpose AI systems. The vulnerability is exploitable by crafting a model argument (stride values) and results in a crash/denial of service, meeting the security vulnerability criterion. The report provides sufficient evidence (references, affected versions, and fix plan).

References

• NVD entry
• https://github.com/tensorflow/tensorflow/security/advisories/GHSA-vfr4-x8j2-3rf9
• https://github.com/tensorflow/tensorflow/commit/801c1c6be5324219689c98e1bd3e0ca365ee834d

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