CVE Vulnerabilities


Access of Resource Using Incompatible Type ('Type Confusion')

Published: Feb 04, 2022 | Modified: Jul 13, 2023
CVSS 3.x
CVSS 2.x

Tensorflow is an Open Source Machine Learning Framework. A malicious user can cause a denial of service by altering a SavedModel such that any binary op would trigger CHECK failures. This occurs when the protobuf part corresponding to the tensor arguments is modified such that the dtype no longer matches the dtype expected by the op. In that case, calling the templated binary operator for the binary op would receive corrupted data, due to the type confusion involved. If Tin and Tout dont match the type of data in out and input_* tensors then flat<*> would interpret it wrongly. In most cases, this would be a silent failure, but we have noticed scenarios where this results in a CHECK crash, hence a denial of service. 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.


The product allocates or initializes a resource such as a pointer, object, or variable using one type, but it later accesses that resource using a type that is incompatible with the original type.

Affected Software

Name Vendor Start Version End Version
Tensorflow Google * 2.5.2 (including)
Tensorflow Google 2.6.0 (including) 2.6.2 (including)
Tensorflow Google 2.7.0 (including) 2.7.0 (including)

Extended Description

When the product accesses the resource using an incompatible type, this could trigger logical errors because the resource does not have expected properties. In languages without memory safety, such as C and C++, type confusion can lead to out-of-bounds memory access. While this weakness is frequently associated with unions when parsing data with many different embedded object types in C, it can be present in any application that can interpret the same variable or memory location in multiple ways. This weakness is not unique to C and C++. For example, errors in PHP applications can be triggered by providing array parameters when scalars are expected, or vice versa. Languages such as Perl, which perform automatic conversion of a variable of one type when it is accessed as if it were another type, can also contain these issues.