CVE-2022-21728

Out of bounds read in Tensorflow

Severity Score

8.1

*CVSS v3.1

Exploit Likelihood

*EPSS

Affected Versions

*CPE

Public Exploits

*Multiple Sources

Exploited in Wild

*KEV

Decision

*SSVC

Descriptions

Tensorflow is an Open Source Machine Learning Framework. The implementation of shape inference for `ReverseSequence` does not fully validate the value of `batch_dim` and can result in a heap OOB read. There is a check to make sure the value of `batch_dim` does not go over the rank of the input, but there is no check for negative values. Negative dimensions are allowed in some cases to mimic Python's negative indexing (i.e., indexing from the end of the array), however if the value is too negative then the implementation of `Dim` would access elements before the start of an array. The fix will be included in TensorFlow 2.8.0. We will also cherrypick this commit on TensorFlow 2.7.1, TensorFlow 2.6.3, and TensorFlow 2.5.3, as these are also affected and still in supported range.

Tensorflow es un marco de aprendizaje automático de código abierto. La implementación de la inferencia de forma para "ReverseSequence" no comprueba completamente el valor de "batch_dim" y puede resultar en una lectura OOB del montón. Se presenta una comprobación para asegurarse de que el valor de "batch_dim" no supera el rango de la entrada, pero no hay ninguna comprobación para los valores negativos. Las dimensiones negativas son permitidas en algunos casos para imitar la indexación negativa de Python (es decir, la indexación desde el final del array), sin embargo si el valor es demasiado negativo entonces la implementación de "Dim" accedería a elementos antes del inicio de un array. La corrección será incluida en TensorFlow versión 2.8.0. También seleccionaremos este commit en TensorFlow versión 2.7.1, TensorFlow versión 2.6.3, y TensorFlow versión 2.5.3, ya que estos también están afectados y aún están en el rango admitido

*Credits: N/A

Attack Vector

Network

Attack Complexity

Low

Privileges Required

Low

User Interaction

None

Scope

Unchanged

Confidentiality

High

Integrity

None

Availability

High

Attack Vector

Network

Attack Complexity

Low

Authentication

Single

Confidentiality

Partial

Integrity

None

Availability

Partial

* Common Vulnerability Scoring System

SSVC

Decision:-

Exploitation

Automatable

Tech. Impact

* Organization's Worst-case Scenario

Timeline

2021-11-16 CVE Reserved
2022-02-03 CVE Published
2022-04-29 First Exploit
2024-08-03 CVE Updated
2024-10-06 EPSS Updated
---------- Exploited in Wild
---------- KEV Due Date

CWE

CWE-125: Out-of-bounds Read

CAPEC

References (5)

URL	Tag	Source

URL	Date	SRC
https://github.com/mwina/CVE-2022-21728-test	2022-04-29
https://github.com/tensorflow/tensorflow/blob/5100e359aef5c8021f2e71c7b986420b85ce7b3d/tensorflow/core/framework/shape_inference.h#L415-L428	2024-08-03
https://github.com/tensorflow/tensorflow/blob/5100e359aef5c8021f2e71c7b986420b85ce7b3d/tensorflow/core/ops/array_ops.cc#L1636-L1671	2024-08-03

URL	Date	SRC
https://github.com/tensorflow/tensorflow/commit/37c01fb5e25c3d80213060460196406c43d31995	2022-02-08
https://github.com/tensorflow/tensorflow/security/advisories/GHSA-6gmv-pjp9-p8w8	2022-02-08

URL	Date	SRC

Affected Vendors, Products, and Versions

Vendor		Product				Version		Other		Status
Vendor	Product	Version	Other	Status	<-- -->	Vendor	Product	Version	Other	Status
Google Search vendor "Google"		Tensorflow Search vendor "Google" for product "Tensorflow"				<= 2.5.2 Search vendor "Google" for product "Tensorflow" and version " <= 2.5.2"		-		Affected
Google Search vendor "Google"		Tensorflow Search vendor "Google" for product "Tensorflow"				>= 2.6.0 <= 2.6.2 Search vendor "Google" for product "Tensorflow" and version " >= 2.6.0 <= 2.6.2"		-		Affected
Google Search vendor "Google"		Tensorflow Search vendor "Google" for product "Tensorflow"				2.7.0 Search vendor "Google" for product "Tensorflow" and version "2.7.0"		-		Affected