HighVulnerability

CVE-2021-29591: TensorFlow Lite infinite loop and stack overflow via crafted model graphs

Source: NVD(opens in a new tab)
Identifier: CVE-2021-29591
Published: May 14, 2021
Record updated: October 10, 2026

Affected: tensorflow < 2.1.4, fixed in 2.1.4, released June 15, 2021, 31 days after disclosure
tensorflow >= 2.2.0, < 2.2.3, fixed in 2.2.3, released June 15, 2021, 31 days after disclosure
tensorflow >= 2.3.0, < 2.3.3, fixed in 2.3.3, released June 15, 2021, 31 days after disclosure
and 9 more
Known exploitation: Not listed in the CISA Known Exploited Vulnerabilities catalog at the last check.
EPSS: 0.3%

Summary

TensorFlow's TFlite runtime does not check that graphs are free of loops between nodes. An attacker can craft a model that causes an infinite loop during evaluation, which can become a stack overflow through too many recursive calls. The `While` implementation can be tricked into a case where the body and loop subgraphs are the same, so evaluating one calls `Eval` on the other until stack space is exhausted.

Mitigation

The fix will be included in TensorFlow 2.5.0. The commit will also be cherry-picked to TensorFlow 2.4.2, 2.3.3, 2.2.3 and 2.1.4.

Details

Scores and exploitation

CVSS score: 7.3(high)
Patch available: Yes
EPSS: 0.3%Estimated probability that this CVE is exploited in the next 30 days, from the FIRST Exploit Prediction Scoring System. Read on November 4, 2026.

Weaknesses and techniques

Weakness (CWE): CWE-835
CWE-674
CWE-835

Affected

Packages: tensorflow@< 2.1.4 (fixed: 2.1.4)
tensorflow@>= 2.2.0, < 2.2.3 (fixed: 2.2.3)
tensorflow@>= 2.3.0, < 2.3.3 (fixed: 2.3.3)
tensorflow@>= 2.4.0, < 2.4.2 (fixed: 2.4.2)
tensorflow-cpu@< 2.1.4 (fixed: 2.1.4)
tensorflow-cpu@>= 2.2.0, < 2.2.3 (fixed: 2.2.3)
tensorflow-cpu@>= 2.3.0, < 2.3.3 (fixed: 2.3.3)
tensorflow-cpu@>= 2.4.0, < 2.4.2 (fixed: 2.4.2)
tensorflow-gpu@< 2.1.4 (fixed: 2.1.4)
tensorflow-gpu@>= 2.2.0, < 2.2.3 (fixed: 2.2.3)
tensorflow-gpu@>= 2.3.0, < 2.3.3 (fixed: 2.3.3)
tensorflow-gpu@>= 2.4.0, < 2.4.2 (fixed: 2.4.2)
Vendors: Google

Classification

Attack type: Denial of service
Attack sophistication: Trivial
AI component targeted: Inference
Impact: Availability
Labels: security

How this record was produced

Summaries, mitigations and labels on this page are generated automatically from the source. The linked source is authoritative. The source gives this advisory a description and no title, so the title above is a headline generated from the source text.

Original source: https://nvd.nist.gov/vuln/detail/CVE-2021-29591

First tracked: February 15, 2026 at 08:39 PM ET, from NVD/CVE Database

Classifier: claude-haiku-5-5, prompt v4

Classifier confidence: 95% (self-assessed)

Affected packages in the Exposure Registry

Matched by package name and ecosystem. Each entry shows whether the package delegates to a language model and how many tracked packages depend on it.

tensorflowPyPINo LLM dependency detected · 26 tracked dependents
tensorflow-gpuPyPINo LLM dependency detected · 2 tracked dependents

Summary

Mitigation

Affected packages in the Exposure Registry

Related items