PT-2021-21768 · Google · Tensorflow
Mihaimaruseac
·
Published
2021-08-12
·
Updated
2024-03-06
·
CVE-2021-37652
CVSS v4.0
8.5
High
| Vector | AV:L/AC:L/AT:N/PR:L/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N |
Name of the Vulnerable Software and Affected Versions
TensorFlow versions prior to 2.6.0
TensorFlow versions 2.5.1 and earlier
TensorFlow versions 2.4.3 and earlier
TensorFlow versions 2.3.4 and earlier
Description
The implementation for
tf.raw ops.BoostedTreesCreateEnsemble can result in a use after free error if an attacker supplies specially crafted arguments. This occurs due to a reference counted resource being decremented when initialization fails, but later refactoring changed the resource to a smart pointer, leading to a double-free process where members of the resource object are accessed after being freed.Recommendations
For versions prior to 2.6.0, update to TensorFlow 2.6.0 or later.
For versions 2.5.1 and earlier, cherrypick the commit 5ecec9c6fbdbc6be03295685190a45e7eee726ab to fix the issue.
For versions 2.4.3 and earlier, cherrypick the commit 5ecec9c6fbdbc6be03295685190a45e7eee726ab to fix the issue.
For versions 2.3.4 and earlier, cherrypick the commit 5ecec9c6fbdbc6be03295685190a45e7eee726ab to fix the issue.
As a temporary workaround, consider avoiding the use of
tf.raw ops.BoostedTreesCreateEnsemble until a patch is available.Fix
Double Free
Use After Free
Found an issue in the description? Have something to add? Feel free to write us 👾
Related Identifiers
Affected Products
Tensorflow