CVE-2021-29612: TensorFlow Heap Buffer Overflow in BandedTriangularSolve

Background

TensorFlow is an end-to-end open source platform for machine learning. It provides a comprehensive ecosystem of tools, libraries, and resources for building and deploying machine learning models. TensorFlow is widely used in various domains, including research, industry, and academia. The vulnerability described here affects the TensorFlow package, specifically the tf.raw_ops.BandedTriangularSolve function.

Vulnerability Detail

The vulnerability in TensorFlow arises from a heap buffer overflow in the Eigen implementation of tf.raw_ops.BandedTriangularSolve. The implementation fails to properly validate the input tensors, allowing an attacker to trigger the overflow by providing empty tensors. Additionally, the use of the OP_REQUIRES macro in the implementation does not effectively handle errors, as callers of helper functions do not check the status before continuing. This vulnerability can be exploited by an attacker to execute arbitrary code or cause a denial of service.

pip install --upgrade "tensorflow-cpu==2.21.0"

Workaround

As a temporary workaround, users can ensure that the input tensors provided to tf.raw_ops.BandedTriangularSolve are not empty. This can be done by adding appropriate checks in the calling code to validate the input tensors before passing them to the vulnerable function. Here is an example of how to perform the check in Python:

import tensorflow as tf

# Assuming `input_tensor1` and `input_tensor2` are the input tensors
if input_tensor1.shape.num_elements() == 0 or input_tensor2.shape.num_elements() == 0:
    raise ValueError('Input tensors must not be empty')

# Call the vulnerable function
output_tensor = tf.raw_ops.BandedTriangularSolve(input_tensor1, input_tensor2)

It is important to note that this workaround only prevents the vulnerability if the input tensors are empty. Upgrading to the fixed versions of TensorFlow is still recommended.

Conclusion

In conclusion, the vulnerability in TensorFlow's tf.raw_ops.BandedTriangularSolve function allows for a heap buffer overflow, which can be exploited by an attacker to execute arbitrary code or cause a denial of service. The TensorFlow team has addressed this vulnerability in the latest versions, and users are advised to update their installations to mitigate the risk. For further support on vulnerability remediation, please contact DevNack.com.

See every vulnerability recorded for tensorflow-cpu →

FAQ