CVE-2021-29574: TensorFlow - Null Pointer Dereference in MaxPool3DGradGrad

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 includes various operations and functions for performing mathematical computations and data manipulation. One of these operations is tf.raw_ops.MaxPool3DGradGrad, which is responsible for computing the gradient of the gradient of a 3D max pooling operation. This operation is used in many machine learning models and applications.

Vulnerability Detail

The implementation of tf.raw_ops.MaxPool3DGradGrad in TensorFlow exhibits a null pointer dereference vulnerability. The vulnerability arises when the operation is supplied with empty tensors as inputs. The implementation fails to validate the emptiness of the input tensors, resulting in a null pointer dereference when accessing the elements of the tensors. An attacker can exploit this vulnerability by providing empty tensors as inputs to the MaxPool3DGradGrad operation, leading to undefined behavior and potential crashes or arbitrary code execution. This vulnerability affects versions of TensorFlow up to 2.4.2.

pip install --upgrade "tensorflow==2.21.0"

Workaround

As a workaround, users can ensure that the input tensors provided to the MaxPool3DGradGrad operation are not empty. This can be done by validating the shape and size of the tensors before passing them to the operation. Here is an example of how to check if a tensor is empty in TensorFlow:

import tensorflow as tf

def is_tensor_empty(tensor):
    return tf.reduce_all(tf.equal(tf.shape(tensor), tf.constant([0])))

Conclusion

In conclusion, the MaxPool3DGradGrad operation in TensorFlow is vulnerable to a null pointer dereference when supplied with empty tensors as inputs. This vulnerability has been addressed in TensorFlow 2.5.0 and patches have been provided for earlier affected versions. Users are advised to upgrade or apply the patches to mitigate the risk. For further support on vulnerability remediation, please contact DevNack.com.

See every vulnerability recorded for tensorflow →

FAQ