PT-2021-21805 · Google · Tensorflow

Published

2021-08-12

·

Updated

2024-03-06

·

CVE-2021-37686

CVSS v4.0

6.8

Medium

VectorAV:L/AC:L/AT:N/PR:L/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N
Name of the Vulnerable Software and Affected Versions TensorFlow version 2.6.0
Description The strided slice implementation in TFLite has a logic bug which can allow an attacker to trigger an infinite loop. This arises from newly introduced support for ellipsis in axis definition. An attacker can craft a model such that ellipsis end idx is smaller than i (e.g., always negative). In this case, the inner loop does not increase i and the continue statement causes execution to skip over the preincrement at the end of the outer loop.
Recommendations For version 2.6.0, the issue has been patched in GitHub commit dfa22b348b70bb89d6d6ec0ff53973bacb4f4695, and the fix will be included in subsequent versions. As a temporary workaround, consider disabling the affected strided slice implementation in TFLite until a patch is available. Restrict access to the vulnerable TFLite module to minimize the risk of exploitation. Avoid using the ellipsis end idx variable in the affected TFLite implementation until the issue is resolved.

Fix

Infinite Loop

Weakness Enumeration

Related Identifiers

BIT-TENSORFLOW-2021-37686
CVE-2021-37686
GHSA-MHHC-Q96P-MFM9
OPENSUSE-SU-2022:10014-1
OPENSUSE-SU-2024:12116-1
PYSEC-2021-308
PYSEC-2021-599
PYSEC-2021-797

Affected Products

Tensorflow