CVE-2021-41222: TensorFlow GPU - Segfault in SplitV Implementation

Sponsored Habilume Habit tracker with a companion that grows with you Complete habits, earn Lume Points, and watch your Lume grow as you stay consistent. App Store Google Play

Background

TensorFlow is an open source platform for machine learning. It provides a comprehensive ecosystem of tools, libraries, and resources for developing and deploying machine learning models. TensorFlow GPU is a variant of TensorFlow that leverages GPU acceleration for faster computation. It is widely used in various domains, including research, industry, and academia.

Vulnerability Detail

The vulnerability (CVE-2021-41222) exists in the implementation of the SplitV function in TensorFlow GPU. When negative arguments are supplied to SplitV, a segmentation fault (segfault) can be triggered. This occurs when the size_splits parameter contains more than one value and at least one of those values is negative. An attacker can exploit this vulnerability to cause a denial-of-service (DoS) condition or potentially execute arbitrary code on the affected system. The vulnerability affects multiple versions of TensorFlow GPU, including 2.4.4, 2.5.2, 2.6.1, and 2.7.0rc0.

pip install --upgrade "tensorflow-gpu==2.12.0"

Workaround

As a workaround, users can avoid supplying negative arguments to the SplitV function in TensorFlow GPU. It is recommended to validate and sanitize input values before passing them to the function. Here is an example of how to perform input validation in Python:

size_splits = [10, -5, 8]

# Validate input
if any(split < 0 for split in size_splits):
    raise ValueError('Negative values not allowed in size_splits')

# Call SplitV function
result = tf.split(size_splits=size_splits)

Conclusion

In conclusion, the vulnerability (CVE-2021-41222) in TensorFlow GPU's SplitV function can be exploited by supplying negative arguments, leading to a segfault and potential security risks. It is crucial to update to the latest version of TensorFlow GPU (2.7.0) or apply the respective fixes for earlier versions (2.6.1, 2.5.2, and 2.4.4) to mitigate this vulnerability. Regularly monitoring and updating software dependencies is essential to ensure the security and stability of machine learning systems.

See every vulnerability recorded for tensorflow-gpu →

FAQ