CVE-2021-41204: Segfault while copying constant resource tensor
Impact During TensorFlow's Grappler optimizer phase, constant folding might attempt to deep copy a resource tensor. This results in a segfault, as these tensors are supposed to not change.
Patches We have patched the issue in GitHub commit 7731e8dfbe4a56773be5dc94d631611211156659.
The fix will be included in TensorFlow 2.7.0. We will also cherrypick this commit on TensorFlow 2.6.1, TensorFlow 2.5.2, and TensorFlow 2.4.4, 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
Other sources
TensorFlow is an open source platform for machine learning. In affected versions during TensorFlow's Grappler optimizer phase, constant folding might attempt to deep copy a resource tensor. This results in a segfault, as these tensors are supposed to not change. The fix will be included in TensorFlow 2.7.0. We will also cherrypick this commit on TensorFlow 2.6.1, TensorFlow 2.5.2, and TensorFlow 2.4.4, as these are also affected and still in supported range.
Affected Software
Event History
Frequently Asked Questions
What is the severity of CVE-2021-41204?
The severity of CVE-2021-41204 is classified as high due to the potential for a segmentation fault during the runtime of TensorFlow's Grappler optimizer.
How do I fix CVE-2021-41204?
To fix CVE-2021-41204, upgrade TensorFlow to version 2.4.4 or later, 2.5.2, or 2.6.1.
Which software versions are affected by CVE-2021-41204?
CVE-2021-41204 affects TensorFlow versions up to 2.4.4 and versions between 2.5.0 to 2.5.2, and 2.6.0 to 2.6.1.
What types of tensors are involved in CVE-2021-41204?
CVE-2021-41204 involves resource tensors, which are not meant to change and can lead to a segmentation fault if mishandled.
Is CVE-2021-41204 related to constant folding in TensorFlow?
Yes, CVE-2021-41204 is related to constant folding within TensorFlow's Grappler optimizer.