Use of uninitialized resource in TensorFlow - CVE-2022-23573
Published: May 10, 2022
Vulnerability identifier: #VU62917
CSH Severity: Medium
CVSS v4.0: CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N/E:U/U:Green
CVE-ID: CVE-2022-23573
CWE-ID: CWE-908
Exploitation vector: Remote access
Exploit availability:
No public exploit available
Vendor: TensorFlow
Affected software:
TensorFlow
TensorFlow
Detailed vulnerability description
The vulnerability allows a remote user to compromise the target system.
The vulnerability exists due to implementation of `AssignOp` that can result in copying uninitialized data to a new tensor. A remote use can pass specially crafted data to the application, trigger uninitialized usage of resources and execute arbitrary code on the system.
How to mitigate CVE-2022-23573
Install updates from vendor's website.
Sources
- https://github.com/tensorflow/tensorflow/security/advisories/GHSA-q85f-69q7-55h2
- https://github.com/tensorflow/tensorflow/blob/a1320ec1eac186da1d03f033109191f715b2b130/tensorflow/core/kernels/assign_op.h#L30-L143
- https://github.com/tensorflow/tensorflow/commit/ef1d027be116f25e25bb94a60da491c2cf55bd0b