CVE-2021-41221: Access to invalid memory during shape inference in `Cudnn*` ops

Severity: medium (CVSS 4.6 v2.0)
Exploited in the wild: not listed in CISA KEV
EPSS: 0.2% probability of exploitation in the next 30 days
Public exploit or detection template: yes
Published: 2021-11-05 · updated 2026-06-17
Weakness: CWE-120

Description

TensorFlow is an open source platform for machine learning. In affected versions the shape inference code for the `Cudnn*` operations in TensorFlow can be tricked into accessing invalid memory, via a heap buffer overflow. This occurs because the ranks of the `input`, `input_h` and `input_c` parameters are not validated, but code assumes they have certain values. The fix will be included in TensorFlow 2.7.0. We will also cherrypick this commit on TensorFlow 2.6.1, TensorFlow 2.5.2, and TensorFlow 2.4.4, as these are also affected and still in supported range.

Affected products

tensorflow: from 2.4.0 before 2.4.4; from 2.5.0 before 2.5.2; from 2.6.0 before 2.6.1; 2.7.0-rc0, 2.7.0-rc1 — fixed in 2.4.4, 2.5.2, 2.6.1

How to fix

Upgrade to a fixed version: tensorflow 2.4.4, 2.5.2, 2.6.1. Check the vendor advisory for mitigations.

References

Get alerted when CVE-2021-41221 or the next vulnerability affects your stack