CVE-2021-29587: TensorFlow SpaceToDepth TFLite Operator 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. The SpaceToDepth TFLite operator is used to rearrange data by spatially partitioning it into blocks and then concatenating them. This operator is an integral part of TensorFlow's deep learning framework and is widely used in various applications.

Vulnerability Detail

The vulnerability (CVE-2021-29587) exists in the Prepare step of the SpaceToDepth TFLite operator in TensorFlow. The operator fails to check for a zero value before performing a division operation. An attacker can exploit this by crafting a malicious model that sets the block_size parameter to zero. When the model is executed, it triggers a division by zero error, leading to a potential denial-of-service (DoS) condition or other unexpected behavior. The vulnerability affects TensorFlow versions 2.1.4 and below, 2.2.3 and below, 2.3.3 and below, and 2.4.2 and below.

pip install --upgrade "tensorflow==2.21.0"

Workaround

As a temporary workaround, users can manually check and ensure that the block_size parameter used in the SpaceToDepth TFLite operator is not set to zero. This can be done by inspecting the model code or configuration files. If a zero value is found, it should be modified to a non-zero value to mitigate the vulnerability. Here is an example of how the block_size parameter can be set correctly:

import tensorflow as tf

block_size = 2

# Create the SpaceToDepth layer
space_to_depth_layer = tf.keras.layers.Lambda(lambda x: tf.nn.space_to_depth(x, block_size))

Conclusion

The TensorFlow SpaceToDepth TFLite Operator Division by Zero Vulnerability (CVE-2021-29587) poses a risk to TensorFlow applications using affected versions. It allows an attacker to craft a malicious model that triggers a division by zero error, potentially leading to a denial-of-service (DoS) condition or other unexpected behavior. Users are strongly advised to upgrade their TensorFlow installations to version 2.5.0 or later to mitigate this vulnerability. Alternatively, users on older versions can upgrade to the respective fixed versions. For further support and assistance in vulnerability remediation, please contact DevNack.com.

See every vulnerability recorded for tensorflow →

FAQ