GHSA-mw6j-hh29-h379

`CHECK` failure in depthwise ops via overflows

No scoreFix available

GHSA-mw6j-hh29-h379 is a vulnerability with no CVSS score yet in tensorflow, published May 25, 2022. No source lists public exploit code. Fixed in tensorflow 2.6.4.

What does this mean?

Severity: UnratedNo CVSS score yet
Exploit chance: UnscoredNo EPSS score yet
Exploitation: None knownNot on CISA’s or ENISA’s lists
Fixed in: 2.6.4tensorflow +14 more

Description

Impact

The implementation of depthwise ops in TensorFlow is vulnerable to a denial of service via CHECK-failure (assertion failure) caused by overflowing the number of elements in a tensor:

import tensorflow as tf

input = tf.constant(1, shape=[1, 4, 4, 3], dtype=tf.float32)
filter_sizes = tf.constant(1879048192, shape=[13], dtype=tf.int32)
out_backprop = tf.constant(1, shape=[1, 4, 4, 3], dtype=tf.float32)
tf.raw_ops.DepthwiseConv2dNativeBackpropFilter(
    input=input, filter_sizes=filter_sizes, out_backprop=out_backprop, strides=[1, 1, 1, 1], padding="SAME")

This is another instance of TFSA-2021-198 (CVE-2021-41197).

Patches

We have patched the issue in GitHub commit 3796cc4fcd93ae55812a457abc96dcd55fbb854b.

The fix will be included in TensorFlow 2.9.0. We will also cherrypick this commit on TensorFlow 2.8.1, TensorFlow 2.7.2, and TensorFlow 2.6.4, 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 Neophytos Christou from Secure Systems Lab at Brown University.

Affected packages

Affected packages
Package	Affected	Fixed
`tensorflow` PyPI	before 2.6.4	2.6.4
`tensorflow` PyPI	2.7.0 up to 2.7.2	2.7.2
`tensorflow` PyPI	2.8.0 up to 2.8.1	2.8.1
`tensorflow-cpu` PyPI	before 2.6.4	2.6.4
`tensorflow-cpu` PyPI	2.7.0 up to 2.7.2	2.7.2
`tensorflow-cpu` PyPI	2.8.0 up to 2.8.1	2.8.1
`tensorflow-gpu` PyPI	before 2.6.4	2.6.4
`tensorflow-gpu` PyPI	2.7.0 up to 2.7.2	2.7.2
`tensorflow-gpu` PyPI	2.8.0 up to 2.8.1	2.8.1
`tensorflow` PyPI	before 2.6.4, plus listed versions 0.12.00.12.11.0.01.0.11.1.0 Show 45 moreShow less 1.10.01.10.11.11.01.12.01.12.21.12.31.13.11.13.21.14.01.15.01.15.21.15.31.15.41.15.51.2.01.2.11.3.01.4.01.4.11.5.01.5.11.6.01.7.01.7.11.8.01.9.02.0.02.0.12.0.22.0.32.0.42.1.02.1.12.1.22.1.32.1.42.2.02.2.12.2.22.2.32.3.02.3.12.3.22.3.32.3.4	2.6.4

Show 5 moreShow less

Affected packages
Package	Affected	Fixed
`tensorflow` PyPI	2.7.0 up to 2.7.2, plus listed versions 2.7.1	2.7.2
`tensorflow-cpu` PyPI	before 2.6.4, plus listed versions 1.15.02.1.02.1.12.1.22.1.3 Show 23 moreShow less 2.1.42.2.02.2.12.2.22.2.32.3.02.3.12.3.22.3.32.3.42.4.02.4.12.4.22.4.32.4.42.5.02.5.12.5.22.5.32.6.02.6.12.6.22.6.3	2.6.4
`tensorflow-cpu` PyPI	2.7.0 up to 2.7.2, plus listed versions 2.7.1	2.7.2
`tensorflow-gpu` PyPI	before 2.6.4, plus listed versions 0.12.00.12.11.0.01.0.11.1.0 Show 45 moreShow less 1.10.01.10.11.11.01.12.01.12.21.12.31.13.11.13.21.14.01.15.01.15.21.15.31.15.41.15.51.2.01.2.11.3.01.4.01.4.11.5.01.5.11.6.01.7.01.7.11.8.01.9.02.0.02.0.12.0.22.0.32.0.42.1.02.1.12.1.22.1.32.1.42.2.02.2.12.2.22.2.32.3.02.3.12.3.22.3.32.3.4	2.6.4
`tensorflow-gpu` PyPI	2.7.0 up to 2.7.2, plus listed versions 2.7.1	2.7.2

What each source reports

GitHub AdvisoriesNo score

`CHECK` failure in depthwise ops via overflows

Impact

The implementation of depthwise ops in TensorFlow is vulnerable to a denial of service via CHECK-failure (assertion failure) caused by overflowing the number of elements in a tensor:

import tensorflow as tf

input = tf.constant(1, shape=[1, 4, 4, 3], dtype=tf.float32)
filter_sizes = tf.constant(1879048192, shape=[13], dtype=tf.int32)
out_backprop = tf.constant(1, shape=[1, 4, 4, 3], dtype=tf.float32)
tf.raw_ops.DepthwiseConv2dNativeBackpropFilter(
    input=input, filter_sizes=filter_sizes, out_backprop=out_backprop, strides=[1, 1, 1, 1], padding="SAME")

This is another instance of TFSA-2021-198 (CVE-2021-41197).

Patches

We have patched the issue in GitHub commit 3796cc4fcd93ae55812a457abc96dcd55fbb854b.

The fix will be included in TensorFlow 2.9.0. We will also cherrypick this commit on TensorFlow 2.8.1, TensorFlow 2.7.2, and TensorFlow 2.6.4, 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 Neophytos Christou from Secure Systems Lab at Brown University.

Record: GHSA-mw6j-hh29-h379
Weakness: CWE-190
Updated: 2026-11-02

OSVNo score

`CHECK` failure in depthwise ops via overflows

Impact

The implementation of depthwise ops in TensorFlow is vulnerable to a denial of service via CHECK-failure (assertion failure) caused by overflowing the number of elements in a tensor:

import tensorflow as tf

input = tf.constant(1, shape=[1, 4, 4, 3], dtype=tf.float32)
filter_sizes = tf.constant(1879048192, shape=[13], dtype=tf.int32)
out_backprop = tf.constant(1, shape=[1, 4, 4, 3], dtype=tf.float32)
tf.raw_ops.DepthwiseConv2dNativeBackpropFilter(
    input=input, filter_sizes=filter_sizes, out_backprop=out_backprop, strides=[1, 1, 1, 1], padding="SAME")

This is another instance of TFSA-2021-198 (CVE-2021-41197).

Patches

We have patched the issue in GitHub commit 3796cc4fcd93ae55812a457abc96dcd55fbb854b.

The fix will be included in TensorFlow 2.9.0. We will also cherrypick this commit on TensorFlow 2.8.1, TensorFlow 2.7.2, and TensorFlow 2.6.4, 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 Neophytos Christou from Secure Systems Lab at Brown University.

Record: GHSA-mw6j-hh29-h379
Weakness: CWE-190
Updated: 2026-11-02

History

History
Recorded	Changed	Sources
Nov 2, 2026	first indexed	GitHub Advisories, OSV

Dates are when Every CVE recorded each change. Upstream dates may differ.

References

This page doesn't list everything

Left out here: 29 versions. The JSON record has them.