CVE-2021-29559: TensorFlow-gpu - Out-of-bounds Heap Access in tf.raw_ops.UnicodeEncode

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 GPU-enabled version of TensorFlow that leverages the power of GPUs for accelerated computation. It is widely used in various industries and research domains for training and inference of deep learning models.

Vulnerability Detail

The vulnerability identified as CVE-2021-29559 affects TensorFlow-gpu versions prior to 2.5.0. It resides in the tf.raw_ops.UnicodeEncode function, specifically in the handling of input values and splits for sparse tensors. The implementation assumes that the input_value/input_splits pair specifies a valid sparse tensor, leading to an out-of-bounds access of the heap-allocated array. An attacker can exploit this vulnerability to access data outside the bounds of the array, potentially resulting in information disclosure or arbitrary code execution. The issue has been addressed in TensorFlow 2.5.0, and patches have been backported to versions 2.4.2, 2.3.3, 2.2.3, and 2.1.4.

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

Workaround

As a temporary workaround, you can consider implementing input validation and sanitization in your code to ensure that only valid input values and splits are passed to the tf.raw_ops.UnicodeEncode function. By validating the input and ensuring it adheres to the expected format, you can mitigate the risk of triggering the out-of-bounds heap access. Here's an example of how you can perform input validation in Python:

def validate_input(input_value, input_splits):
    # Perform validation checks on input_value and input_splits
    # Return True if the input is valid, False otherwise
    pass

input_value = ...
input_splits = ...

if validate_input(input_value, input_splits):
    # Call tf.raw_ops.UnicodeEncode with the validated input
    tf.raw_ops.UnicodeEncode(input_value, input_splits)
else:
    # Handle the case when the input is invalid
    ...

Conclusion

In conclusion, the vulnerability CVE-2021-29559 affects TensorFlow-gpu versions prior to 2.5.0 and allows for out-of-bounds heap access in the tf.raw_ops.UnicodeEncode function. It is crucial to upgrade to the latest version or apply the provided patches to mitigate the risk of exploitation. If you have any further questions or need assistance with vulnerability remediation, please contact DevNack.com.

See every vulnerability recorded for tensorflow-gpu →

FAQ