SB2020102153 - Multiple vulnerabilities in TensorFlow

CSH
CYBERSECURITY HELP
Sign In REGISTER

Main Vulnerability Database SB2020102153

SB2020102153 - Multiple vulnerabilities in TensorFlow

Published: October 21, 2020 Updated: May 4, 2026

Security Bulletin ID SB2020102153
CSH Severity
Medium
Patch available
YES
Number of vulnerabilities 3
Exploitation vector Remote access
Highest impact Denial of service

Breakdown by Severity

Medium 67% Low 33%
  • Low
  • Medium
  • High
  • Critical

Description

This security bulletin contains information about 3 vulnerabilities.


1) Out-of-bounds read (CVE-ID: CVE-2020-26269)

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

The vulnerability exists due to an out-of-bounds read in filesystem glob matching in GetMatchingPaths when parsing crafted filesystem path patterns. A local user can invoke glob matching on a crafted path to cause a denial of service.

The issue occurs because directory index assumptions in the parallel implementation are not verified under certain scenarios.


2) Input validation error (CVE-ID: CVE-2020-15266)

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

The vulnerability exists due to improper input validation in tf.image.crop_and_resize when processing a boxes argument with very large values. A remote attacker can supply a specially crafted boxes argument to cause a denial of service.


3) Out-of-bounds read (CVE-ID: CVE-2020-15265)

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

The vulnerability exists due to an out-of-bounds read in tf.quantization.quantize_and_dequantize when processing an invalid axis value. A remote attacker can pass a specially crafted axis value to cause a denial of service.

In normal builds, the dimension check is compiled out, which can lead to a segmentation fault.


Remediation

Install update from vendor's website.

References