Division by zero in TensorFlow - CVE-2021-29517
Published: May 13, 2021 / Updated: May 4, 2026
Vulnerability identifier: #VU129375
CSH Severity: Low
CVSSv4.0: CVSS:4.0/AV:L/AC:L/AT:N/PR:L/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N/E:U/U:Clear
CVE-ID: CVE-2021-29517
CWE-ID: CWE-369
Exploitation vector: Local access
Exploit availability:
No public exploit available
Vulnerable software:
TensorFlow
TensorFlow
Software vendor:
TensorFlow
TensorFlow
Description
The vulnerability allows a local user to cause a denial of service.
The vulnerability exists due to division by zero in the Conv3D implementation when processing crafted tensor shapes. A local user can supply tensors with a zero-valued fifth filter dimension to cause a denial of service.
Invalid tensor shapes can also trigger an Eigen assertion.
Remediation
Install security update from vendor's website.