INCIDENT MAP

Browse / AI Framework and Platform Flaws / CVE-2023-25662

TensorFlow vulnerable to integer overflow in EditDistance

CVE-2023-25662 ยท 2023-03-24 high

Class
AI Framework and Platform Flaws (filed by a model that read the advisory)
Technique
Denial of service
Affected
tensorflow-cpu, tensorflow (PyPI)
CVSS
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H
EPSS
0.4% chance of exploitation in 30 days

How it works, in short

An empty hypothesis shape causes an unsigned integer underflow in EditDistance shape handling, which makes the operation hang. (Our summary, written by a model. The source text is below.)

Details

Impact

TFversion 2.11.0 //tensorflow/core/ops/array_ops.cc:1067 const Tensor* hypothesis_shape_t = c->input_tensor(2); std::vector<DimensionHandle> dims(hypothesis_shape_t->NumElements() - 1); for (int i = 0; i < dims.size(); ++i) { dims[i] = c->MakeDim(std::max(h_values(i), t_values(i))); }

if hypothesis_shape_t is empty, hypothesis_shape_t->NumElements() - 1 will be integer overflow, and the it will deadlock

import tensorflow as tf
para={
    'hypothesis_indices': [[]],
    'hypothesis_values': ['tmp/'],
    'hypothesis_shape': [],
    'truth_indices': [[]],
    'truth_values': [''],
    'truth_shape': [],
    'normalize': False
    }
tf.raw_ops.EditDistance(**para)

Patches

We have patched the issue in GitHub commit 08b8e18643d6dcde00890733b270ff8d9960c56c.

The fix will be included in TensorFlow 2.12.0. We will also cherrypick this commit on TensorFlow 2.11.1

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 r3pwnx

References

Source: GitHub Advisory Database, used under CC BY 4.0.