CVE-2020-15197: TensorFlow SparseCountSparseOutput Rank Validation Vulnerability
Summary
CVE-2020-15197: Denial of Service in Tensorflow. Affects tensorflow versions >= 2.3.0, < 2.3.1. Fixed in 2.3.1. How to fix ↓
Also tracked for
- tensorflow · PyPI
- tensorflow-cpu · PyPI
- tensorflow-gpu · PyPI
- tensorflow · PyPI
The same CVE-2020-15197 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
Habilume
Habit tracker with a companion that grows with you
Complete habits, earn Lume Points, and watch your Lume grow as you stay consistent.
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. The SparseCountSparseOutput function in TensorFlow is used to count the number of nonzero elements in a sparse tensor. However, prior to version 2.3.1, this implementation does not validate the input arguments, specifically the 'indices' tensor, to ensure it has a rank of 2. This can lead to a CHECK assertion failure and a crash if malicious users pass tensors of different ranks. This vulnerability can be exploited to cause denial of service in serving installations where users have control over the input sparse tensor components.
Vulnerability Detail
The vulnerability in TensorFlow's SparseCountSparseOutput function arises from the lack of validation for the rank of the 'indices' tensor. This tensor is assumed to be a matrix and its elements are accessed accordingly. However, if a tensor of a different rank is provided, a CHECK assertion failure occurs, resulting in a crash. Malicious users can exploit this vulnerability to cause a denial of service in serving installations. The issue has been patched in commit 3cbb917b4714766030b28eba9fb41bb97ce9ee02 and is addressed in TensorFlow version 2.3.1.
How To Fix
Minimum fixed version
2.3.1
This is the first release that contains the fix.
Tip — Not sure which version your app runs? Vulert can match this fix to it. Scan free →
Workaround
As a workaround, users can manually validate the rank of the 'indices' tensor before passing it to the SparseCountSparseOutput function. This can be done by checking the shape of the tensor and ensuring it has a rank of 2. If the tensor has a different rank, appropriate actions can be taken to handle the situation and prevent a crash.
Conclusion
The TensorFlow SparseCountSparseOutput Rank Validation Vulnerability (CVE-2020-15197) allows malicious users to cause a denial of service in serving installations. It can be mitigated by updating TensorFlow to version 2.3.1 or later and regularly updating dependencies. For further support on vulnerability remediation, contact DevNack.com.
See every vulnerability recorded for tensorflow →
FAQ
TensorFlow is an open-source machine learning framework developed by Google. It provides tools and resources for building and deploying machine learning models.
The SparseCountSparseOutput function in TensorFlow is used to count the number of nonzero elements in a sparse tensor.
This vulnerability can be exploited by malicious users to cause a denial of service in serving installations where they have control over the input sparse tensor components.
To fix this vulnerability, update TensorFlow to version 2.3.1 or later. Regularly update all dependencies and libraries used in the application.
Yes, a workaround is to manually validate the rank of the 'indices' tensor before passing it to the SparseCountSparseOutput function.
References
Related vulnerabilities
Recent vulnerabilities recorded for tensorflow.
Habilume
Habit tracker with a companion that grows with you
Complete habits, earn Lume Points, and watch your Lume grow as you stay consistent.
