AVID-2026-R0949

Description

Memory exhaustion in Tensorflow (CVE-2022-21733)

Details

Tensorflow is an Open Source Machine Learning Framework. The implementation of StringNGrams can be used to trigger a denial of service attack by causing an out of memory condition after an integer overflow. We are missing a validation on pad_witdh and that result in computing a negative value for ngram_width which is later used to allocate parts of the output. The fix will be included in TensorFlow 2.8.0. We will also cherrypick this commit on TensorFlow 2.7.1, TensorFlow 2.6.3, and TensorFlow 2.5.3, as these are also affected and still in supported range.

Reason for inclusion in AVID: CVE-2022-21733 describes a memory exhaustion/DoS vulnerability in TensorFlow’s StringNGrams operation caused by an integer overflow, affecting multiple TF versions. This is a software vulnerability in an AI framework used to develop, train, and serve AI systems, i.e., a component of the general-purpose AI stack. The report provides CVE details, impact, and references, satisfying evidence requirements.

References

• NVD entry
• https://github.com/tensorflow/tensorflow/security/advisories/GHSA-98j8-c9q4-r38g
• https://github.com/tensorflow/tensorflow/commit/f68fdab93fb7f4ddb4eb438c8fe052753c9413e8
• https://github.com/tensorflow/tensorflow/blob/5100e359aef5c8021f2e71c7b986420b85ce7b3d/tensorflow/core/kernels/string_ngrams_op.cc#L29-L161

Affected or Relevant Artifacts

• Developer: n/a
• Deployer: n/a
• Artifact Details:

Impact

AVID Taxonomy Categorization

• Risk domains: Security
• SEP subcategories: S0100: Software Vulnerability
• Lifecycle stages: L06: Deployment

CVSS

Other information

• Report Type: Advisory
• Credits:
• Date Reported: 2022-02-03
• Version: 0.3.3
• AVID Entry