GHSA-xf83-q765-xm6m `CHECK` fail in `TensorListScatter` and `TensorListScatterV2` in eager mode

Not scored Published 21 Nov 2022

A CHECK failure in TensorFlow's TensorListScatter operations allows attackers to cause a denial of service via crafted non-scalar inputs.

Exploited in the wild: Not known
Login needed: Not needed
Attackable: Not stated
Victim action needed: Not needed
Fix: Available

What is GHSA-xf83-q765-xm6m?

GHSA-xf83-q765-xm6m is a denial of service vulnerability in tensorflow, published on 21 Nov 2022.

Official description

Impact

Another instance of CVE-2022-35991, where TensorListScatter and TensorListScatterV2 crash via non scalar inputs inelement_shape, was found in eager mode and fixed.

import tensorflow as tf
arg_0=tf.random.uniform(shape=(2, 2, 2), dtype=tf.float16, maxval=None)
arg_1=tf.random.uniform(shape=(2, 2, 2), dtype=tf.int32, maxval=65536)
arg_2=tf.random.uniform(shape=(2, 2, 2), dtype=tf.int32, maxval=65536)
arg_3=&#39;&#39;
tf.raw_ops.TensorListScatter(tensor=arg_0, indices=arg_1, 
element_shape=arg_2, name=arg_3)

Patches

We have patched the issue in GitHub commit bf9932fc907aff0e9e8cccf769e8b00d30fd81a1.

The fix will be included in TensorFlow 2.11. We will also cherrypick this commit on TensorFlow 2.10.1, 2.9.3, and TensorFlow 2.8.4, 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 Pattarakrit Rattankul

Which versions are affected by GHSA-xf83-q765-xm6m?

18 open source packages ship an affected version, including tensorflow, tensorflow-cpu and tensorflow-gpu.

Package	Affected	Fixed in
`tensorflow` PyPI	All versions	`2.8.4`
`tensorflow` PyPI	All versions	`2.8.4`
`tensorflow` PyPI	From `2.10.0`	`2.10.1`
`tensorflow` PyPI	From `2.10.0`	`2.10.1`
`tensorflow` PyPI	From `2.9.0`	`2.9.3`
`tensorflow` PyPI	From `2.9.0`	`2.9.3`

12 more packages

`tensorflow-cpu` PyPI	All versions	`2.8.4`
`tensorflow-cpu` PyPI	All versions	`2.8.4`
`tensorflow-cpu` PyPI	From `2.10.0`	`2.10.1`
`tensorflow-cpu` PyPI	From `2.10.0`	`2.10.1`
`tensorflow-cpu` PyPI	From `2.9.0`	`2.9.3`
`tensorflow-cpu` PyPI	From `2.9.0`	`2.9.3`
`tensorflow-gpu` PyPI	All versions	`2.8.4`
`tensorflow-gpu` PyPI	All versions	`2.8.4`
`tensorflow-gpu` PyPI	From `2.10.0`	`2.10.1`
`tensorflow-gpu` PyPI	From `2.10.0`	`2.10.1`
`tensorflow-gpu` PyPI	From `2.9.0`	`2.9.3`
`tensorflow-gpu` PyPI	From `2.9.0`	`2.9.3`

How do I fix GHSA-xf83-q765-xm6m?

Upgrade tensorflow to a fixed version for your release line: 2.8.4, 2.10.1 or 2.9.3.

Upgrade, then put limits on request size, rate and nesting depth so one client cannot exhaust the service. More about Denial of service.

Is GHSA-xf83-q765-xm6m being exploited?

There is no known exploitation of GHSA-xf83-q765-xm6m in the wild and no public exploit code.