Browse / AI Framework and Platform Flaws / GHSA-43q8-3fv7-pr5x
Improper Validation of Integrity Check Value in TensorFlow
- Class
- AI Framework and Platform Flaws (filed by a model that read the source)
- Technique
- Memory corruption
- Affected
- tensorflow-cpu, tensorflow (PyPI)
- CVSS
- CVSS:3.1/AV:L/AC:H/PR:L/UI:N/S:U/C:H/I:H/A:H
How it works, in short
Sparse split does not check that the axis argument is scalar, so crafted arguments cause out-of-bounds reads and segfaults. (Our summary, written by a model. The source text is below.)
Details
Impact
The implementation of tf.sparse.split does not fully validate the input arguments. Hence, a malicious user can trigger a denial of service via a segfault or a heap OOB read:
import tensorflow as tf
data = tf.random.uniform([1, 32, 32], dtype=tf.float32)
axis = [1, 2]
x = tf.sparse.from_dense(data)
result = tf.sparse.split(x,3, axis=axis)
The code assumes axis is a scalar. This is another instance of TFSA-2021-190 (CVE-2021-41206).
Patches
We have patched the issue in GitHub commit 61bf91e768173b001d56923600b40d9a95a04ad5 (merging #53695).
The fix will be included in TensorFlow 2.8.0. We will also cherrypick this commit on TensorFlow 2.7.1, TensorFlow 2.6.3, and TensorFlow 2.5.3, as these are also affected and still in supported range.
For more information
Please consult our security guide for more information regarding the security model and how to contact us with issues and questions.
Attribution
This vulnerability has been reported externally via a GitHub issue.
References
- https://github.com/tensorflow/tensorflow/security/advisories/GHSA-43q8-3fv7-pr5x
- https://github.com/tensorflow/tensorflow/security/advisories/GHSA-pgcq-h79j-2f69
- https://github.com/tensorflow/tensorflow/pull/53695
- https://github.com/tensorflow/tensorflow/commit/61bf91e768173b001d56923600b40d9a95a04ad5
- https://github.com/tensorflow/tensorflow
Source: GitHub Advisory Database, used under CC BY 4.0.