CVE-2021-29569: TensorFlow - Out-of-Bounds Read in tf.raw_ops.MaxPoolGradWithArgmax
Summary
CVE-2021-29569: Heap out of bounds read in RequantizationRange. 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 · PyPI
- tensorflow · PyPI
- tensorflow-cpu · PyPI
- tensorflow-gpu · PyPI
The same CVE-2021-29569 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 end-to-end open source platform for machine learning. It provides a comprehensive ecosystem for building and deploying machine learning models. TensorFlow includes various operations and functions to perform tasks such as data preprocessing, model training, and inference. However, a vulnerability has been discovered in the implementation of tf.raw_ops.MaxPoolGradWithArgmax function, which can lead to out-of-bounds reads in heap allocated data.
Vulnerability Detail
The vulnerability (CVE-2021-29569) arises from the assumption made in the implementation of tf.raw_ops.MaxPoolGradWithArgmax that the input_min and input_max tensors have at least one element. If these tensors are empty, accessing the 0th element results in a read outside the bounds of the allocated data. An attacker can exploit this by providing specially crafted inputs, potentially leading to information disclosure or arbitrary code execution. The vulnerability affects TensorFlow versions 0 and 2.1.4 to 2.4.2.
How To Fix
Fixed versions by release line
The minimum fixed version depends on the release line you run.
-
< 2.1.42.1.4 -
>= 2.2.0, < 2.2.32.2.3 -
>= 2.3.0, < 2.3.32.3.3 -
>= 2.4.0, < 2.4.22.4.2
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 temporary workaround, users can avoid passing empty tensors (input_min and input_max) to the tf.raw_ops.MaxPoolGradWithArgmax function. By ensuring that these tensors have at least one element, the vulnerability can be mitigated. Here is an example of how to check and handle empty tensors in Python:
if tf.size(input_min) > 0 and tf.size(input_max) > 0:
# Call tf.raw_ops.MaxPoolGradWithArgmax
else:
# Handle empty tensors
Conclusion
The vulnerability in TensorFlow's tf.raw_ops.MaxPoolGradWithArgmax function can result in out-of-bounds reads in heap allocated data. It is crucial to update to the fixed versions of TensorFlow once they are released. Regularly checking for updates and promptly applying them is recommended to maintain the security of TensorFlow deployments. For further support on vulnerability remediation, please contact DevNack.com.
See every vulnerability recorded for tensorflow →
FAQ
TensorFlow is an open source platform for machine learning that provides a comprehensive ecosystem for building and deploying machine learning models.
The vulnerability can potentially lead to information disclosure or arbitrary code execution if an attacker supplies specially crafted inputs.
The vulnerability affects TensorFlow versions 0 and 2.1.4 to 2.4.2.
The vulnerability will be fixed in TensorFlow 2.5.0. Patches will also be backported to TensorFlow 2.4.2, 2.3.3, 2.2.3, and 2.1.4. It is recommended to update to the fixed versions as soon as they are available.
Yes, as a temporary workaround, users can ensure that the input_min and input_max tensors passed to tf.raw_ops.MaxPoolGradWithArgmax have at least one element.
References
Related vulnerabilities
Recent vulnerabilities recorded for tensorflow.
