Learn about CVE-2021-29618, a vulnerability in TensorFlow where passing a complex argument to `tf.transpose` with `conjugate=True` causes a crash. Impact rated LOW. Find out affected versions and mitigation steps.
TensorFlow is an open-source machine learning platform. A vulnerability exists in passing a complex argument to
tf.transpose
simultaneously with conjugate=True
, leading to a crash. The affected versions are TensorFlow < 2.1.4, >= 2.2.0, < 2.2.3, >= 2.3.0, < 2.3.3, and >= 2.4.0, < 2.4.2. The impact is rated as LOW.
Understanding CVE-2021-29618
This section provides insights into the vulnerability found in TensorFlow.
What is CVE-2021-29618?
CVE-2021-29618 involves a vulnerability in TensorFlow where passing specific arguments together results in a crash.
The Impact of CVE-2021-29618
The impact of this vulnerability is considered LOW, affecting versions within specific ranges of TensorFlow.
Technical Details of CVE-2021-29618
Below are the technical details related to this vulnerability.
Vulnerability Description
The vulnerability arises from passing a complex argument to
tf.transpose
concurrently with conjugate=True
.
Affected Systems and Versions
The affected versions are TensorFlow < 2.1.4, >= 2.2.0, < 2.2.3, >= 2.3.0, < 2.3.3, and >= 2.4.0, < 2.4.2.
Exploitation Mechanism
Successful exploitation occurs when the specific argument conditions are met concurrently in TensorFlow.
Mitigation and Prevention
To address CVE-2021-29618, consider the following mitigation strategies.
Immediate Steps to Take
Users should update TensorFlow to versions where the fix has been included, such as TensorFlow 2.5.0 or the cherry-picked commits in TensorFlow 2.4.2, 2.3.3, 2.2.3, and 2.1.4.
Long-Term Security Practices
Implement secure coding practices and stay informed about TensorFlow's security updates to prevent future vulnerabilities.
Patching and Updates
Regularly update TensorFlow to the latest versions where the vulnerability has been fixed to ensure system security.