CVE-2021-37640: Integer division by 0 in sparse reshaping in TensorFlow
Impact The implementation of tf.rawops.SparseReshape can be made to trigger an integral division by 0 exception:
python import tensorflow as tf
tf.rawops.SparseReshape( inputindices = np.ones((1,3)), inputshape = np.array([1,1,0]), newshape = np.array([1,0])) The implementation calls the reshaping functor whenever there is at least an index in the input but does not check that shape of the input or the target shape have both a non-zero number of elements.
The reshape functor blindly divides by the dimensions of the target shape. Hence, if this is not checked, code will result in a division by 0. Patches We have patched the issue in GitHub commit 4923de56ec94fff7770df259ab7f2288a74feb41.
The fix will be included in TensorFlow 2.6.0. We will also cherrypick this commit on TensorFlow 2.5.1 as this is the other affected version.
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 by members of the Aivul Team from Qihoo 360.
Other sources
TensorFlow is an end-to-end open source platform for machine learning. In affected versions the implementation of tf.rawops.SparseReshape can be made to trigger an integral division by 0 exception. The implementation calls the reshaping functor whenever there is at least an index in the input but does not check that shape of the input or the target shape have both a non-zero number of elements. The reshape functor blindly divides by the dimensions of the target shape. Hence, if this is not checked, code will result in a division by 0. We have patched the issue in GitHub commit 4923de56ec94fff7770df259ab7f2288a74feb41. The fix will be included in TensorFlow 2.6.0. We will also cherrypick this commit on TensorFlow 2.5.1 as this is the other affected version.
Affected Software
Remediation
Recommended actions to resolve this vulnerability, in priority order.
- Upgrade
Upgrade
pip/tensorflow-gputo a version that resolves this vulnerability.Fixed in 2.5.1 - Upgrade
Upgrade
pip/tensorflow-gputo a version that resolves this vulnerability.Fixed in 2.4.3 - Upgrade
Upgrade
pip/tensorflow-gputo a version that resolves this vulnerability.Fixed in 2.3.4 - Upgrade
Upgrade
pip/tensorflow-cputo a version that resolves this vulnerability.Fixed in 2.5.1 - Upgrade
Upgrade
pip/tensorflow-cputo a version that resolves this vulnerability.Fixed in 2.4.3 - Upgrade
Upgrade
pip/tensorflow-cputo a version that resolves this vulnerability.Fixed in 2.3.4 - Upgrade
Upgrade
pip/tensorflowto a version that resolves this vulnerability.Fixed in 2.5.1 - Upgrade
Upgrade
pip/tensorflowto a version that resolves this vulnerability.Fixed in 2.4.3 - Upgrade
Upgrade
pip/tensorflowto a version that resolves this vulnerability.Fixed in 2.3.4 - Upgrade
Upgrade
tensorflowto a version that resolves this vulnerability.Fixed in 2.6.0Patch 4923de56ec94fff7770df259ab7f2288a74feb41 - Upgrade
Upgrade
tensorflowto a version that resolves this vulnerability.Fixed in 2.5.1Patch 4923de56ec94fff7770df259ab7f2288a74feb41 - Compensating control
If you cannot apply the patch/upgrade immediately, avoid calling `tf.raw_ops.SparseReshape` with target shapes that have a non-zero number of elements mismatch that can lead to an integral division by 0 (e.g., cases involving a zero in the target shape dimensions as described in the example with `input_shape = np.array([1,1,0])` and `new_shape = np.array([1,0])`).
Event History
Frequently Asked Questions
What is the severity of CVE-2021-37640?
CVE-2021-37640 has been classified with a medium severity level due to the potential for integral division by zero exceptions.
How do I fix CVE-2021-37640?
To fix CVE-2021-37640, upgrade TensorFlow to version 2.5.1 or later if you are using versionss 2.3.0 to 2.4.3.
What versions are affected by CVE-2021-37640?
CVE-2021-37640 affects TensorFlow versions 2.3.0 to 2.3.4, 2.4.0 to 2.4.3, and specific versions 2.5.0 and several release candidates.
What is the cause of CVE-2021-37640?
CVE-2021-37640 is caused by the implementation of the tf.raw_ops.SparseReshape function triggering an integral division by zero exception.
Are any packages affected by CVE-2021-37640?
Yes, the TensorFlow packages tensorflow, tensorflow-gpu, and tensorflow-cpu in the affected versions have vulnerabilities due to CVE-2021-37640.