CVE-2021-29517: TensorFlow GPU Division by Zero Vulnerability

Background

TensorFlow is an end-to-end 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 specific package that enables GPU acceleration for TensorFlow computations. It allows users to leverage the power of GPUs to accelerate training and inference processes in machine learning models.

Vulnerability Detail

A division by zero vulnerability exists in the Conv3D implementation of TensorFlow GPU. The vulnerability arises due to a modulo operation that is performed based on user-controlled input. If the fifth element of the 'filter' tensor is set to zero, a division by zero occurs. Additionally, if the shape of the two tensors is invalid, it can trigger an Eigen assertion, leading to a program crash. This vulnerability can be exploited by a malicious user to cause a denial-of-service (DoS) attack or potentially execute arbitrary code on the affected system.

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

Workaround

As a temporary workaround, users can ensure that the 'filter' tensor does not contain a zero as the fifth element. Additionally, it is important to validate the shapes of the tensors used in the Conv3D operation to avoid triggering the Eigen assertion. Here is an example code snippet that demonstrates the workaround:

import tensorflow as tf

# Ensure 'filter' tensor does not have zero as the fifth element
filter = tf.constant([1, 2, 3, 4, 1], dtype=tf.float32)

# Validate tensor shapes
if filter.shape[0] != expected_shape[0] or filter.shape[1] != expected_shape[1] or filter.shape[2] != expected_shape[2]:
    raise ValueError('Invalid tensor shape')

# Perform Conv3D operation
output = tf.nn.conv3d(input, filter, strides, padding)

Note: This workaround should only be used until the package is updated to a fixed version.

Conclusion

The division by zero vulnerability in TensorFlow GPU's Conv3D implementation can lead to a program crash or potential code execution. It is crucial to update the TensorFlow GPU package to the fixed versions or apply the provided workaround to mitigate the risk. For further assistance or support in vulnerability remediation, please contact DevNack.com.

See every vulnerability recorded for tensorflow-gpu →

FAQ