Use of uninitialized resource in TensorFlow - CVE-2020-26266
Published: December 11, 2020 / Updated: January 7, 2021
Vulnerability identifier: #VU49324
CSH Severity: Low
CVSS v4: 2.4 [CVSS:4.0/AV:L/AC:L/AT:N/PR:L/UI:A/VC:L/VI:L/VA:N/SC:N/SI:N/SA:N]
CVE-ID: CVE-2020-26266
CWE-ID: CWE-908
Exploitation vector: Local access
Exploit availability:
No public exploit available
Vulnerability details
The vulnerability allows a local user to bypass certain security restrictions.
The vulnerability exists due to usage of uninitialized resources in Eigen. Under certain cases a saved model can trigger use of uninitialized values during code execution. This is caused by having tensor buffers be filled with the default value of the type but forgetting to default initialize the quantized floating point types in Eigen.
Affected software
TensorFlow
How to mitigate CVE-2020-26266
Install updates from vendor's website.
TensorFlow - addressed in versions 1.15.5, 2.0.4, 2.1.3, 2.2.2, 2.3.2, 2.4.0
External References
- https://github.com/tensorflow/tensorflow/commit/ace0c15a22f7f054abcc1f53eabbcb0a1239a9e2
- https://github.com/tensorflow/tensorflow/security/advisories/GHSA-qhxx-j73r-qpm2
- https://github.com/tensorflow/tensorflow/releases/tag/v2.3.2
- https://github.com/tensorflow/tensorflow/releases/tag/v2.2.2
- https://github.com/tensorflow/tensorflow/releases/tag/v2.1.3
- https://github.com/tensorflow/tensorflow/releases/tag/v2.0.4
- https://github.com/tensorflow/tensorflow/releases/tag/v1.15.5