CVE-2021-29614: Interpreter crash from `tf.io.decode_raw` in TensorFlow
Summary
CVE-2021-29614: Interpreter crash from tf.io.decode_raw. 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-2021-29614 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 developed by Google. It provides a comprehensive ecosystem of tools, libraries, and resources for building and deploying machine learning models. The tf.io.decode_raw function in TensorFlow is used to decode raw binary data into a tensor. It is commonly used in data preprocessing pipelines for tasks such as reading image or audio files. However, a vulnerability has been discovered in the implementation of tf.io.decode_raw that can lead to incorrect results and crashes in the Python interpreter.
Vulnerability Detail
The vulnerability in tf.io.decode_raw arises when using the fixed_length parameter in combination with wider datatypes. The implementation of the padded version of the function contains a bug related to pointer arithmetic rules. The code mistakenly advances the output data pointer by fixed_length * sizeof(T) bytes, even when only fixed_length bytes are copied from the input. This results in parts of the input not being properly decoded into the output tensor. Additionally, the incorrect pointer advance can lead to out-of-bounds writes, potentially causing more severe attacks. An attacker could exploit this vulnerability to crash the interpreter or write to periodically placed locations in memory.
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 avoid using the fixed_length parameter in tf.io.decode_raw when working with wider datatypes. Instead, the tf.io.decode_raw function can be used without specifying the fixed_length argument, allowing TensorFlow to automatically determine the length based on the datatype. This can help prevent the interpreter crash and incorrect decoding. Here is an example:
import tensorflow as tf
data = tf.constant(["1","2","3","4"])
output = tf.io.decode_raw(data, tf.uint16)
Conclusion
The vulnerability in tf.io.decode_raw can have significant implications for TensorFlow users, potentially leading to incorrect results, crashes, and even more severe attacks. It is crucial to update to the patched versions of TensorFlow and regularly apply security updates to mitigate this vulnerability. For further support on vulnerability remediation, please contact DevNack.com.
See every vulnerability recorded for tensorflow →
FAQ
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.
tf.io.decode_raw is a function in TensorFlow used to decode raw binary data into a tensor. It is commonly used in data preprocessing pipelines for tasks such as reading image or audio files.
The vulnerability in tf.io.decode_raw can lead to incorrect results and crashes in the Python interpreter. It can also result in out-of-bounds writes, potentially enabling more severe attacks.
To fix this vulnerability, it is recommended to upgrade to one of the patched versions of TensorFlow: TensorFlow 2.5.0, TensorFlow 2.4.2, TensorFlow 2.3.3, TensorFlow 2.2.3, or TensorFlow 2.1.4. Regularly checking for and applying security updates is also advisable.
Yes, a workaround is to avoid using the fixed_length parameter in tf.io.decode_raw when working with wider datatypes. Instead, let TensorFlow automatically determine the length based on the datatype.
References
Related vulnerabilities
Recent vulnerabilities recorded for tensorflow.
