CVE-2021-29589

LOW

TensorFlow < 2.1.4 - Denial of Service via Division by Zero in GatherNd Operator

Title source: llm
STIX 2.1

Description

TensorFlow is an end-to-end open source platform for machine learning. The reference implementation of the `GatherNd` TFLite operator is vulnerable to a division by zero error(https://github.com/tensorflow/tensorflow/blob/0d45ea1ca641b21b73bcf9c00e0179cda284e7e7/tensorflow/lite/kernels/internal/reference/reference_ops.h#L966). An attacker can craft a model such that `params` input would be an empty tensor. In turn, `params_shape.Dims(.)` would be zero, in at least one dimension. The fix will be included in TensorFlow 2.5.0. We will also cherrypick this commit on TensorFlow 2.4.2, TensorFlow 2.3.3, TensorFlow 2.2.3 and TensorFlow 2.1.4, as these are also affected and still in supported range.

References (2)

Core 2
Core References
Exploit, Patch, Third Party Advisory x_refsource_confirm
https://github.com/tensorflow/tensorflow/security/advisories/GHSA-3w67-q784-6w7c

Scores

CVSS v3 2.5
EPSS 0.0020
EPSS Percentile 10.0%
Attack Vector LOCAL
CVSS:3.1/AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:L

Details

CWE
CWE-369
Status published
Products (4)
google/tensorflow < 2.1.4
pypi/tensorflow 0 - 2.1.4PyPI
pypi/tensorflow-cpu 0 - 2.1.4PyPI
pypi/tensorflow-gpu 0 - 2.1.4PyPI
Published May 14, 2021
Tracked Since Feb 18, 2026