CVE-2021-37670: TensorFlow-GPU Heap Overflow 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 variant of TensorFlow that leverages GPU acceleration for faster computation. It is widely used in various industries for deep learning tasks.

Vulnerability Detail

The vulnerability (CVE-2021-37670) exists in TensorFlow-GPU versions 2.3.0 to 2.4.3 and 2.5.0 to 2.5.0. It allows an attacker to read data from outside the bounds of heap allocated memory by exploiting a flaw in the tf.raw_ops.UpperBound and tf.raw_ops.LowerBound functions. By sending specially crafted illegal arguments, an attacker can bypass input validation and access sensitive information or cause a denial of service. The issue arises due to the lack of rank validation for the sorted_input argument in the implementation of these functions.

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

Workaround

As a temporary workaround, users can apply input validation checks to the sorted_input argument in their code before passing it to the tf.raw_ops.UpperBound and tf.raw_ops.LowerBound functions. Here is an example of how to perform the validation:

import tensorflow as tf

def safe_upper_bound(sorted_input):
    if tf.rank(sorted_input) != 1:
        raise ValueError('Invalid rank for sorted_input')
    return tf.raw_ops.UpperBound(sorted_input)

Conclusion

The TensorFlow-GPU heap overflow vulnerability (CVE-2021-37670) allows an attacker to read data from outside the bounds of heap allocated memory. It is crucial to update to TensorFlow-GPU version 2.6.0 or later to mitigate this vulnerability. Alternatively, implement input validation checks for the sorted_input argument in your code. For further assistance or support in vulnerability remediation, please contact DevNack.com.

See every vulnerability recorded for tensorflow-gpu →

FAQ