#VU30549 Out-of-bounds write in TensorFlow - CVE-2019-16778

Vulnerability identifier: #VU30549

Vulnerability risk: High

CVSSv4.0: 8.1 [CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N/E:U/U:Amber]

CVE-ID: CVE-2019-16778

CWE-ID: CWE-787

Exploitation vector: Network

Exploit availability: No

Vulnerable software:
TensorFlow
Server applications / Other server solutions

Vendor: TensorFlow

Description

The vulnerability allows a remote non-authenticated attacker to execute arbitrary code.

In TensorFlow before 1.15, a heap buffer overflow in UnsortedSegmentSum can be produced when the Index template argument is int32. In this case data_size and num_segments fields are truncated from int64 to int32 and can produce negative numbers, resulting in accessing out of bounds heap memory. This is unlikely to be exploitable and was detected and fixed internally in TensorFlow 1.15 and 2.0.

Mitigation
Install update from vendor's website.

Vulnerable software versions

TensorFlow: 1.0.0 - 1.0.1, 1.1.0 - 1.14.0, 1.2.0 - 1.2.1, 1.3.0 - 1.3.1, 1.4.0 - 1.4.1, 1.5.0 - 1.5.1, 1.6.0, 1.7.0 - 1.7.1, 1.8.0, 1.9.0

---

External links
https://github.com/tensorflow/tensorflow/blob/master/tensorflow/security/advisory/tfsa-2019-002.md
https://github.com/tensorflow/tensorflow/commit/db4f9717c41bccc3ce10099ab61996b246099892
https://github.com/tensorflow/tensorflow/security/advisories/GHSA-844w-j86r-4x2j

---

Q & A

Can this vulnerability be exploited remotely?

Yes. This vulnerability can be exploited by a remote non-authenticated attacker via the Internet.

Is there known malware, which exploits this vulnerability?

No. We are not aware of malware exploiting this vulnerability.