# CVE-2021-41216: TensorFlow heap buffer overflow in Transpose shape inference

Source: https://junglewise.ai/threats/cve-2021-41216-tensorflow-heap-buffer-overflow-in-transpose-shape-inference  
Publisher: Junglewise Threat Intelligence, https://junglewise.ai/threats  
Published: 2021-11-05. Updated: 2026-09-26  
License: CC BY 4.0 with a link back

CVE-2021-41216 · Severity: low · CVSS 3.1 · Published 2021-11-05

TensorFlow is an open-source machine learning library used for building and training neural networks. The Transpose operation's shape inference function contains a heap buffer overflow vulnerability that can be triggered by providing negative index values in the permutation parameter. An attacker who can supply malicious inputs to a TensorFlow model could cause a crash (denial of service) or potentially achieve code execution.

Technologies: [Google TensorFlow](https://junglewise.ai/threats/technologies/tensorflow), [tensorflow (PyPI)](https://junglewise.ai/threats/technologies/pypi-tensorflow), [tensorflow-cpu (PyPI)](https://junglewise.ai/threats/technologies/tensorflow-cpu), [tensorflow-gpu (PyPI)](https://junglewise.ai/threats/technologies/tensorflow-gpu). Vendors: [Google](https://junglewise.ai/threats/vendors/google), [PyPI](https://junglewise.ai/threats/vendors/pypi).

## Quick answers

**Is CVE-2021-41216 exploited in the wild?** No. Junglewise Threat Intelligence has no report of CVE-2021-41216 being exploited in the wild.

**How severe is CVE-2021-41216?** Low severity, with a CVSS score of 3.1, published 5 November 2021.

**Which products and versions are affected?** Google TensorFlow 0.12.0 through 2.4.3, 2.5.0-2.5.1, 2.6.0; tensorflow (PyPI) GIT:introduced=0,fixed=c79ba87153ee343401dbe9d1954d7f79e521eb14|ECOSYSTEM:introduced=0,fixed=2.4.4,introduced=2.5.0,fixed=2.5.2,introduced=2.6.0,fixed=2.6.1,introduced=2.7.0rc0,fixed=2.7.0; tensorflow-cpu (PyPI) GIT:introduced=0,fixed=c79ba87153ee343401dbe9d1954d7f79e521eb14|ECOSYSTEM:introduced=0,fixed=2.4.4,introduced=2.5.0,fixed=2.5.2,introduced=2.6.0,fixed=2.6.1,introduced=2.7.0rc0,fixed=2.7.0; tensorflow-gpu (PyPI) GIT:introduced=0,fixed=c79ba87153ee343401dbe9d1954d7f79e521eb14|ECOSYSTEM:introduced=0,fixed=2.4.4,introduced=2.5.0,fixed=2.5.2,introduced=2.6.0,fixed=2.6.1,introduced=2.7.0rc0,fixed=2.7.0.

**Is there a fix for CVE-2021-41216?** Yes. tensorflow (PyPI): upgrade to 2.7.0 or later; tensorflow-cpu (PyPI): upgrade to 2.7.0 or later; tensorflow-gpu (PyPI): upgrade to 2.7.0 or later. Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.

**How is CVE-2021-41216 attacked?** It is attacked with local access.

**Am I affected by CVE-2021-41216?** Check whether you run Google TensorFlow, tensorflow (PyPI), tensorflow-cpu (PyPI) or one of the others listed at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

## How to fix

- tensorflow (PyPI): upgrade to 2.7.0 or later

  ```sh
  pip install "tensorflow>=2.7.0"
  ```

- tensorflow-cpu (PyPI): upgrade to 2.7.0 or later

  ```sh
  pip install "tensorflow-cpu>=2.7.0"
  ```

- tensorflow-gpu (PyPI): upgrade to 2.7.0 or later

  ```sh
  pip install "tensorflow-gpu>=2.7.0"
  ```


## Technical details

The vulnerability is a heap buffer overflow (CWE-787, CWE-120) in the shape inference function of TensorFlow's Transpose operation. The root cause is insufficient validation: the function checks if permutation indices are greater than or equal to the rank but fails to validate against negative indices before using them. When a negative perm value is passed (e.g., perm=[-10]), the Dim() function accepts -1 as a special value for unknown dimensions, but the code does not properly validate other negative values, leading to out-of-bounds memory access. The vulnerability is triggered by calling tf.raw_ops.Transpose() with a negative perm argument and requires local access or the ability to supply inputs to a running TensorFlow model. The fix validates all indices in perm and was patched in commit c79ba87153ee343401dbe9d1954d7f79e521eb14.

## Affected products

- Google TensorFlow 0.12.0 through 2.4.3, 2.5.0-2.5.1, 2.6.0
- Google TensorFlow CPU 1.15.0 through 2.4.3, 2.5.0-2.5.1, 2.6.0
- Google TensorFlow GPU 0.12.0 through 2.4.3, 2.5.0-2.5.1, 2.6.0

## Identifiers

- CVE-2021-41216
- GHSA-3ff2-r28g-w7h9
- PYSEC-2021-408
- PYSEC-2021-625
- PYSEC-2021-823

## Timeline

- 2021-11-04: disclosed: Advisory published
- 2021-11-10: patched: Patches released for TensorFlow 2.4.4, 2.5.2, 2.6.1, and included in 2.7.0

## References

- [https://github.com/tensorflow/tensorflow/security/advisories/GHSA-3ff2-r28g-w7h9](https://github.com/tensorflow/tensorflow/security/advisories/GHSA-3ff2-r28g-w7h9)

## Related threats

- [PYSEC-2026-3331 - TensorFlow has segfault in array_ops.upper_bound](https://junglewise.ai/threats/cve-2023-33976-tensorflow-has-segfault-in-array-ops-upper-bound)
- [PYSEC-2026-3182 - TensorFlow Denial of Service vulnerability](https://junglewise.ai/threats/cve-2023-25661-tensorflow-denial-of-service-vulnerability)
- [PYSEC-2026-3300 - TensorFlow vulnerable to Out-of-Bounds Read in DynamicStitch](https://junglewise.ai/threats/cve-2023-25659-tensorflow-vulnerable-to-out-of-bounds-read-in-dynamicstitch)
- [PYSEC-2026-3362 - TensorFlow vulnerable to seg fault in `tf.raw_ops.Print`](https://junglewise.ai/threats/cve-2023-25660-tensorflow-vulnerable-to-seg-fault-in-tf-raw-ops-print)
- [PYSEC-2026-3293 - TensorFlow vulnerable to integer overflow in EditDistance](https://junglewise.ai/threats/cve-2023-25662-tensorflow-vulnerable-to-integer-overflow-in-editdistance)