CVE-2022-21726: Out of Bounds Read 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 industries and research fields for tasks such as image recognition, natural language processing, and recommendation systems. However, a vulnerability has been discovered in TensorFlow that could potentially lead to out-of-bounds read access, posing a security risk to applications utilizing this framework.

Vulnerability Detail

The vulnerability, identified as CVE-2022-21726, originates from an implementation flaw in the Dequantize function of TensorFlow. This function is responsible for converting quantized tensors to floating-point tensors. The issue arises from inadequate validation of the axis parameter, which specifies the axis along which the dequantization operation is performed. By providing a value outside the valid range, an attacker can trigger heap out-of-bounds accesses, potentially leading to information disclosure or arbitrary code execution. The vulnerability has a CVSS score of 8.1, indicating a high severity level.

pip install --upgrade "tensorflow==2.21.0"

Workaround

As a temporary workaround, users can avoid the vulnerability by ensuring that the axis parameter passed to the Dequantize function is within the valid range. The valid range for axis is from -1 (default value) to the number of dimensions of the input tensor. By verifying and limiting the axis value, the risk of triggering the out-of-bounds read can be mitigated. Here is an example of how to use a valid axis value:

import tensorflow as tf

@tf.function
def test():
  y = tf.raw_ops.Dequantize(
    input=tf.constant([1,1],dtype=tf.qint32),
    min_range=[1.0],
    max_range=[10.0],
    mode='MIN_COMBINED',
    narrow_range=False,
    axis=-1,
    dtype=tf.bfloat16)
  return y

test()

Conclusion

In conclusion, the vulnerability (CVE-2022-21726) in TensorFlow poses a significant security risk due to its potential for out-of-bounds read access. Users are strongly advised to update their TensorFlow installation to one of the patched versions (2.5.3, 2.6.3, 2.7.1, or 2.8.0) to mitigate this vulnerability. Additionally, users can employ the provided workaround to limit the axis parameter and reduce the risk of exploitation. For further support on vulnerability remediation, please contact DevNack.com.

See every vulnerability recorded for tensorflow →

FAQ