CVE-2021-29547: TensorFlow - Out-of-Bounds Access in QuantizedBatchNormWithGlobalNormalization

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 is widely used in various domains, including research, industry, and academia. It offers support for both CPU and GPU computations, making it suitable for a wide range of applications. However, a vulnerability has been discovered in TensorFlow that can lead to a denial of service (DoS) attack.

Vulnerability Detail

The vulnerability (CVE-2021-29547) in TensorFlow exists in the tf.raw_ops.QuantizedBatchNormWithGlobalNormalization function. An attacker can trigger a segmentation fault and cause a DoS by accessing data outside of the bounds. The implementation assumes that the inputs are not empty, but if any of the inputs is empty, accessing the element at index 0 will result in accessing data outside of the bounds. This vulnerability can be exploited by an attacker to crash the application or cause it to become unresponsive, leading to a DoS condition. The impact of this vulnerability can be severe, as it can disrupt critical machine learning processes and affect the availability of TensorFlow-based applications.

pip install --upgrade "tensorflow==2.21.0"

Workaround

As a temporary workaround, users can check if any of the inputs to tf.raw_ops.QuantizedBatchNormWithGlobalNormalization are empty before calling the function. If any input is found to be empty, the function should not be called to avoid triggering the vulnerability. Here is an example of how to check for empty inputs in Python:

if input_tensor.size == 0:
    # Do not call QuantizedBatchNormWithGlobalNormalization
else:
    # Call QuantizedBatchNormWithGlobalNormalization

This workaround ensures that the vulnerable function is not called with empty inputs, preventing the out-of-bounds access.

Conclusion

The vulnerability (CVE-2021-29547) in TensorFlow can be exploited by an attacker to cause a denial of service (DoS) by accessing data outside of the bounds. It is crucial for users to update their TensorFlow installations to the latest patched version as soon as it becomes available. In the meantime, a temporary workaround is available to check for empty inputs before calling the vulnerable function. For further support on vulnerability remediation, please contact DevNack.com.

See every vulnerability recorded for tensorflow →

FAQ