CVE-2021-29566: Heap OOB access in tensorflow

Background

TensorFlow is an open-source machine learning framework developed by Google. 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. However, a vulnerability has been discovered in TensorFlow that could potentially allow an attacker to write outside the bounds of heap allocated arrays.

Vulnerability Detail

The vulnerability, identified as CVE-2021-29566, specifically affects the Dilation2DBackpropInput function in TensorFlow. By passing invalid arguments to this function, an attacker can trigger a heap out-of-bounds (OOB) access. The issue arises due to a lack of validation before writing to the output array. While the loop indices for the out_backprop array are guaranteed to be within range, there are no similar guarantees for the in_backprop array. This vulnerability could lead to arbitrary memory corruption and potentially enable an attacker to execute arbitrary code.

pip install --upgrade "tensorflow==2.21.0"

Workaround

As a workaround, users can avoid passing invalid arguments to the Dilation2DBackpropInput function. It is important to ensure that the input arguments are valid and within the expected range. Here is an example of how to use the function safely:

import tensorflow as tf

input_tensor = tf.constant([1.1] * 81, shape=[3, 3, 3, 3], dtype=tf.float32)
filter = tf.constant([], shape=[0, 0, 3], dtype=tf.float32)
out_backprop = tf.constant([1.1] * 1062, shape=[3, 2, 59, 3], dtype=tf.float32)

try:
    tf.raw_ops.Dilation2DBackpropInput(
        input=input_tensor, filter=filter, out_backprop=out_backprop,
        strides=[1, 40, 1, 1], rates=[1, 56, 56, 1], padding='VALID'
    )
except Exception as e:
    print('Error:', e)

By handling any exceptions raised by the function, users can prevent the vulnerability from being exploited.

Conclusion

The vulnerability in TensorFlow has been addressed by the TensorFlow team. Users are advised to update to TensorFlow 2.5.0 or apply the relevant patches for earlier versions. It is important to ensure that valid arguments are passed to the Dilation2DBackpropInput function to prevent the vulnerability from being exploited. For further support on vulnerability remediation, please contact DevNack.com.

See every vulnerability recorded for tensorflow →

FAQ