CVE-2021-29535: TensorFlow-GPU Heap Buffer Overflow in QuantizedMul

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 variant of TensorFlow that leverages the power of GPUs for accelerated computation. It allows users to train and run machine learning models on GPUs, which can significantly speed up the training process. However, TensorFlow-GPU is vulnerable to a heap buffer overflow in the QuantizedMul operation, which can be exploited by an attacker to execute arbitrary code or cause a denial of service.

Vulnerability Detail

The vulnerability in TensorFlow-GPU arises from a flaw in the implementation of the QuantizedMul operation. When invalid thresholds for quantization are passed as arguments, the implementation assumes that the arguments are valid scalars and tries to access the numeric value directly. However, if any of these tensors is empty, accessing the element at position 0 results in a heap buffer overflow. This can be exploited by an attacker to overwrite adjacent memory and potentially execute arbitrary code or crash the application. The vulnerability affects versions of TensorFlow-GPU up to 2.4.3.

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

Workaround

As a workaround, users can avoid passing empty tensors as arguments to the QuantizedMul operation. By ensuring that the thresholds for quantization are valid and non-empty, the vulnerability can be mitigated. Here is an example of how to check for empty tensors before invoking the QuantizedMul operation:

import tensorflow as tf

def safe_quantized_mul(a, b, c, d):
    if tf.size(a) == 0 or tf.size(b) == 0 or tf.size(c) == 0 or tf.size(d) == 0:
        raise ValueError('Invalid thresholds for quantization')
    else:
        return tf.quantized_mul(a, b, c, d)

Conclusion

The heap buffer overflow vulnerability in TensorFlow-GPU's QuantizedMul operation can be mitigated by upgrading to the fixed versions or applying the backported fixes. Users should ensure that their TensorFlow-GPU installations are up to date to protect against potential exploitation. For further support on vulnerability remediation, please contact DevNack.com.

See every vulnerability recorded for tensorflow-gpu →

FAQ