PT-2026-42819 · Arm · Armnn

Michele Spagnuolo · Published 2026-05-22 · Updated 2026-05-26 · CVE-2026-42627

Positive Technologies (AI Summary)

CVSS v3.1

6.2

Medium

Base vector

AV:L/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H

Name of the Vulnerable Software and Affected Versions Arm ArmNN versions prior to 2026-03-28

Description An integer overflow exists in the TensorShape::GetNumElements() function within armnn/Tensor.cpp. This occurs when tensor dimensions are multiplied using 32-bit unsigned arithmetic without overflow detection, leading the GetNumBytes() function to return an understated allocation size. A crafted TFLite model file can exploit this to bypass buffer size validation, triggering a heap-based buffer over-read during model optimization. Specifically, the BatchToSpaceNdLayer reads beyond the allocated buffer during the Optimize()->InferOutputShapes() process.

Recommendations Update to a version released after 2026-03-27.

Fix

Integer Overflow

Weakness Enumeration

CWE-190

Related Identifiers

CVE-2026-42627

Affected Products

Armnn

References · 9

• https://osv.dev/vulnerability/UBUNTU-CVE-2026-42627 · Vendor Advisory
• https://nvd.nist.gov/vuln/detail/CVE-2026-42627 · Security Note
• https://security-tracker.debian.org/tracker/CVE-2026-42627 · Vendor Advisory
• https://ubuntu.com/security/CVE-2026-42627 · Vendor Advisory
• https://cve.org/CVERecord?id=CVE-2026-42627 · Security Note
• https://osv.dev/vulnerability/DEBIAN-CVE-2026-42627 · Vendor Advisory
• https://github.com/ARM-software/armnn/blob/main/src/armnn/Tensor.cpp · Note
• https://github.com/ARM-software/armnn/blob/main/src/armnnTfLiteParser/TfLiteParser.cpp · Note
• https://t.me/cveNotify/164717 · Telegram Post