Heap-based buffer overflow in TensorFlow - CVE-2021-29536
Published: May 13, 2021 / Updated: May 4, 2026
Vulnerability identifier: #VU129354
CSH Severity: Medium
CVSSv4.0: CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:N/VA:L/SC:N/SI:N/SA:N/E:U/U:Green
CVE-ID: CVE-2021-29536
CWE-ID: CWE-122
Exploitation vector: Remote access
Exploit availability:
No public exploit available
Vulnerable software:
TensorFlow
TensorFlow
Software vendor:
TensorFlow
TensorFlow
Description
The vulnerability allows a remote attacker to cause a denial of service.
The vulnerability exists due to a heap-based buffer overflow in QuantizedReshape when processing invalid quantization threshold tensors. A remote attacker can pass empty input_min or input_max tensors to cause a denial of service.
The issue occurs because the implementation assumes the threshold arguments are valid scalars and accesses element 0 even when the tensors are empty.
Remediation
Install security update from vendor's website.