Out-of-bounds read in TensorFlow - CVE-2020-26267

 

Out-of-bounds read in TensorFlow - CVE-2020-26267

Published: December 11, 2020 / Updated: January 7, 2021


Vulnerability identifier: #VU49325
CSH Severity: Low
CVSS v4: 2 [CVSS:4.0/AV:L/AC:L/AT:P/PR:L/UI:N/VC:L/VI:N/VA:N/SC:N/SI:N/SA:N]
CVE-ID: CVE-2020-26267
CWE-ID: CWE-125
Exploitation vector: Local access
Exploit availability: No public exploit available

Vulnerability details

The vulnerability allows a local user to gain access to potentially sensitive information.

The vulnerability exists due to a boundary condition within the the tf.raw_ops.DataFormatVecPermute API when processing the src_format and dst_format attributes. The code assumes that these two arguments define a permutation of NHWC. A local user can run a specially crafted program to trigger out-of-bounds read error and read contents of memory on the system.


Affected software

TensorFlow

How to mitigate CVE-2020-26267

Install updates from vendor's website.

TensorFlow - addressed in versions 1.15.5, 2.0.4, 2.1.3, 2.2.2, 2.3.2, 2.4.0

External References

Related Security Bulletins