Impact
The implementation of AddManySparseToTensorsMap is vulnerable to an integer overflow which results in a CHECK-fail when building new TensorShape objects (so, an assert failure based denial of service):
import tensorflow as tf
import numpy as np
tf.raw_ops.AddManySparseToTensorsMap(
sparse_indices=[(0,0),(0,1),(0,2),(4,3),(5,0),(5,1)],
sparse_values=[1,1,1,1,1,1],
sparse_shape=[2**32,2**32],
container='',
shared_name='',
name=None)We are missing some validation on the shapes of the input tensors as well as directly constructing a large TensorShape with user-provided dimensions. The latter is an instance of TFSA-2021-198 (CVE-2021-41197) and is easily fixed by replacing a call to TensorShape constructor with a call to BuildTensorShape static helper factory.
Patches
We have patched the issue in GitHub commits b51b82fe65ebace4475e3c54eb089c18a4403f1c and a68f68061e263a88321c104a6c911fe5598050a8.
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 by Faysal Hossain Shezan from University of Virginia.
References
Impact
The implementation of
AddManySparseToTensorsMapis vulnerable to an integer overflow which results in aCHECK-fail when building newTensorShapeobjects (so, an assert failure based denial of service):We are missing some validation on the shapes of the input tensors as well as directly constructing a large
TensorShapewith user-provided dimensions. The latter is an instance of TFSA-2021-198 (CVE-2021-41197) and is easily fixed by replacing a call toTensorShapeconstructor with a call toBuildTensorShapestatic helper factory.Patches
We have patched the issue in GitHub commits b51b82fe65ebace4475e3c54eb089c18a4403f1c and a68f68061e263a88321c104a6c911fe5598050a8.
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 by Faysal Hossain Shezan from University of Virginia.
References