`CHECK`-failures in binary ops in Tensorflow

Summary

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` don't 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.

Timeline

1. Feb 4, 2022, 10:32 PM UTCCVE published
2. Apr 22, 2025, 3:51 PM UTCCISA SSVC assessment

   Exploitation: poc · Automatable: no · Technical impact: partial

3. Apr 22, 2025, 6:26 PM UTCRecord last updated

Affected versions

• tensorflow tensorflow

  From 2.7.0 · fixed in 2.7.1; From 2.6.0 · fixed in 2.6.3; fixed in 2.5.3

• tensorflowPyPI

  fixed in 2.5.3; From 2.6.0 · fixed in 2.6.3; From 2.7.0 · fixed in 2.7.1

• tensorflow-gpuPyPI

  fixed in 2.5.3; From 2.6.0 · fixed in 2.6.3

• tensorflow-cpuPyPI

  fixed in 2.5.3; From 2.6.0 · fixed in 2.6.3

• tensorflowPyPI

  fixed in 2.5.3

• tensorflowPyPI

  From 2.6.0 · fixed in 2.6.3

• tensorflowPyPI

  From 2.7.0 · fixed in 2.7.1

• tensorflow-cpuPyPI

  fixed in 2.5.3

• tensorflow-cpuPyPI

  From 2.6.0 · fixed in 2.6.3

• tensorflow-cpuPyPI

  From 2.7.0 · fixed in 2.7.1

• tensorflow-gpuPyPI

  fixed in 2.5.3

• tensorflow-gpuPyPI

  From 2.6.0 · fixed in 2.6.3

• tensorflow-gpuPyPI

  From 2.7.0 · fixed in 2.7.1

References 3