CVE-2021-29556: Division by 0 in TensorFlow Reverse
Summary
CVE-2021-29556: Division by 0 in Reverse. Affects tensorflow across 4 version ranges.
< 2.1.4
>= 2.2.0, < 2.2.3
>= 2.3.0, < 2.3.3
>= 2.4.0, < 2.4.2
Fixed in 4 releases; the minimum fixed version depends on the release line you run. How to fix ↓
Also tracked for
- tensorflow · Bitnami
- tensorflow-cpu · PyPI
- tensorflow-gpu · PyPI
The same CVE-2021-29556 is tracked separately for each ecosystem and vendor. Its fix information is the same; the package and affected versions differ.
Check if this vulnerability affects your application
Upload a manifest or SBOM and we'll show whether you're affected, plus every other risky package in the same scan.
No signup required · Results in about 60 seconds
Background
TensorFlow is an open-source machine learning framework widely used for developing and training various deep learning models. It provides a comprehensive set of tools and libraries for building and deploying machine learning applications. TensorFlow is available for multiple programming languages, including Python, and is known for its flexibility and scalability. However, a vulnerability has been discovered in the Reverse function of TensorFlow, which can lead to a denial of service attack.
Vulnerability Detail
The vulnerability, identified as CVE-2021-29556, exists in the Reverse function of TensorFlow. This function is used to reverse the order of elements along specified dimensions in a tensor. The implementation of this function performs a division based on the first dimension of the tensor argument. An attacker can exploit this vulnerability by providing a tensor with a first dimension of 0, causing a division by zero error. This error triggers a floating-point exception (FPE) runtime error, resulting in a denial of service. The impact of this vulnerability is limited to the availability of the affected TensorFlow application.
How To Fix
Latest published version: 2.21.0 as of 22 Oct 2026
pip install --upgrade "tensorflow==2.21.0"Upgrading to it may include breaking changes; the versions below are the minimum fixed release for this vulnerability.
If you can't move to the latest
The minimum fixed version depends on the release line you run.
-
< 2.1.42.1.4pip install --upgrade "tensorflow==2.1.4" -
>= 2.2.0, < 2.2.32.2.3pip install --upgrade "tensorflow==2.2.3" -
>= 2.3.0, < 2.3.32.3.3pip install --upgrade "tensorflow==2.3.3" -
>= 2.4.0, < 2.4.22.4.2pip install --upgrade "tensorflow==2.4.2"
Tip — Not sure which release line you're on? Vulert can find the fix for the version you are using. Scan free →
Workaround
As a workaround, users can ensure that the tensor provided to the Reverse function has a non-zero first dimension. This can be achieved by checking the shape of the tensor before passing it to the function and handling the case where the first dimension is zero separately. Here is an example of how to implement this workaround in Python:
import tensorflow as tf
tensor_input = tf.constant([], shape=[0, 1, 1], dtype=tf.int32)
dims = tf.constant([False, True, False], shape=[3], dtype=tf.bool)
if tensor_input.shape[0] == 0:
# Handle the case where the first dimension is zero
# Perform necessary operations or raise an exception
else:
tf.raw_ops.Reverse(tensor=tensor_input, dims=dims)
Conclusion
In conclusion, the Division by 0 vulnerability in TensorFlow's Reverse function can lead to a denial of service attack. Users are advised to update their TensorFlow installation to one of the patched versions to mitigate the risk. A workaround is also available by ensuring that the tensor provided to the Reverse function has a non-zero first dimension. For further support on vulnerability remediation, please contact DevNack.com.
See every vulnerability recorded for tensorflow →
FAQ
The impact of this vulnerability is limited to the availability of the affected TensorFlow application. An attacker can trigger a denial of service by exploiting the vulnerability, causing the application to become unresponsive.
To fix this vulnerability, users should update their TensorFlow installation to one of the patched versions: TensorFlow 2.1.4, TensorFlow 2.2.3, TensorFlow 2.3.3, TensorFlow 2.4.2, or TensorFlow 2.5.0 (upcoming release). This can be done by upgrading the TensorFlow package using the package manager or by manually installing the updated version from the official TensorFlow repository.
Yes, a workaround is available for this vulnerability. Users can ensure that the tensor provided to the Reverse function has a non-zero first dimension. This can be achieved by checking the shape of the tensor before passing it to the function and handling the case where the first dimension is zero separately.
No, this vulnerability cannot be exploited remotely. An attacker needs to have access to the affected TensorFlow application to exploit the vulnerability and trigger a denial of service.
This vulnerability has been discovered by Ying Wang and Yakun Zhang of Baidu X-Team.
References
Related vulnerabilities
Recent vulnerabilities recorded for tensorflow.
