PT-2021-18297 · Google · Tensorflow
Yakun Zhang
+1
·
Published
2021-05-14
·
Updated
2024-03-06
·
CVE-2021-29546
CVSS v3.1
7.8
High
| Vector | AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H |
Name of the Vulnerable Software and Affected Versions
TensorFlow versions prior to 2.5.0
TensorFlow versions 2.1.4 through 2.4.2
Description
An attacker can trigger an integer division by zero undefined behavior in
tf.raw ops.QuantizedBiasAdd. This is because the implementation of the Eigen kernel does a division by the number of elements of the smaller input without checking that this is not zero. The issue can be triggered with a specific input, such as an empty tensor, which causes the division by zero.Recommendations
For TensorFlow versions prior to 2.5.0, update to version 2.5.0 or later.
For TensorFlow version 2.1.4, update to version 2.1.4 with the cherrypicked commit.
For TensorFlow version 2.2.3, update to version 2.2.3 with the cherrypicked commit.
For TensorFlow version 2.3.3, update to version 2.3.3 with the cherrypicked commit.
For TensorFlow version 2.4.2, update to version 2.4.2 with the cherrypicked commit.
As a temporary workaround, consider avoiding the use of
tf.raw ops.QuantizedBiasAdd until a patch is available.Exploit
Fix
Divide By Zero
Found an issue in the description? Have something to add? Feel free to write us 👾
Weakness Enumeration
Related Identifiers
Affected Products
Tensorflow