Integer overflow in TensorFlow - CVE-2022-29203

 

Integer overflow in TensorFlow - CVE-2022-29203

Published: May 17, 2022 / Updated: May 2, 2026


Vulnerability identifier: #VU129076
CSH Severity: Medium
CVSS v4: 8.7 [CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N]
CVE-ID: CVE-2022-29203
CWE-ID: CWE-190
Exploitation vector: Remote access
Exploit availability: No public exploit available

Vulnerability details

The vulnerability allows a remote attacker to cause a denial of service.

The vulnerability exists due to integer overflow in tf.raw_ops.SpaceToBatchND when processing crafted input tensors. A remote attacker can supply a specially crafted block_shape value to trigger an assertion failure and cause a denial of service.

The overflowed result is used to allocate the output tensor.


Affected software

TensorFlow

How to mitigate CVE-2022-29203

Install security update from vendor's website.

TensorFlow - addressed in versions 2.6.4, 2.7.2, 2.8.1, 2.9.0

External References

Related Security Bulletins