Integer overflow in TensorFlow - CVE-2022-35963

 

Integer overflow in TensorFlow - CVE-2022-35963

Published: September 15, 2022 / Updated: May 3, 2026


Vulnerability identifier: #VU129124
CSH Severity: Low
CVSS v4: 6.8 [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]
CVE-ID: CVE-2022-35963
CWE-ID: CWE-190
Exploitation vector: Local access
Exploit availability: No public exploit available

Vulnerability details

The vulnerability allows a local user to cause a denial of service.

The vulnerability exists due to integer overflow in FractionalAvgPoolGrad when processing a crafted orig_input_tensor_shape input. A local user can supply a crafted orig_input_tensor_shape value to cause a denial of service.


Affected software

TensorFlow

How to mitigate CVE-2022-35963

Install security update from vendor's website.

TensorFlow - addressed in versions 2.7.4, 2.8.3, 2.9.2, 2.10.0

External References

Related Security Bulletins