CVE-2022-21733: Memory Exhaustion in TensorFlow

Background

TensorFlow is an open-source machine learning framework widely used for developing and training deep learning models. It provides a comprehensive ecosystem of tools, libraries, and resources for building and deploying machine learning applications. TensorFlow supports various platforms and is available for Python, C++, and other programming languages. However, a vulnerability has been discovered in TensorFlow that could lead to memory exhaustion.

Vulnerability Detail

The vulnerability in TensorFlow arises from an integer overflow in the implementation of the StringNGrams function. By providing specific input parameters, an attacker can trigger a denial of service (DoS) attack, causing an out-of-memory (OOM) condition. The issue occurs due to a lack of validation on the pad_width parameter, resulting in the computation of a negative value for ngram_width. This negative value is later used to allocate parts of the output, leading to memory exhaustion. An attacker can exploit this vulnerability to disrupt the availability of TensorFlow applications.

pip install --upgrade "tensorflow==2.21.0"

Workaround

As a temporary workaround, users can avoid the vulnerability by ensuring that the pad_width parameter passed to the StringNGrams function is a non-negative value. By validating and sanitizing the input, the risk of triggering the memory exhaustion vulnerability can be mitigated. Here is an example of how to use the StringNGrams function with a validated pad_width parameter:

import tensorflow as tf

tf.raw_ops.StringNGrams(
  data=['123456'],
  data_splits=[0,1],
  separator='a'*15,
  ngram_widths=[],
  left_pad='',
  right_pad='',
  pad_width=max(0, -5), 
  preserve_short_sequences=True)

By using max(0, -5) as the pad_width value, the negative input will be converted to zero, preventing the vulnerability from being triggered.

Conclusion

The memory exhaustion vulnerability in TensorFlow can be mitigated by updating to the patched versions or by implementing the provided workaround. It is crucial for TensorFlow users to promptly address this vulnerability to ensure the security and availability of their machine learning applications. For further support and assistance in vulnerability remediation, please contact DevNack.com.

See every vulnerability recorded for tensorflow →

FAQ