The macros that TensorFlow uses for writing assertions (e.g., CHECK LT, CHECK GT, etc.) have an incorrect logic when comparing size t and int values. Due to type conversion rules, several of the macros would trigger incorrectly.

We have patched the issue in GitHub commit b917181c29b50cb83399ba41f4d938dc369109a1 (merging GitHub PR #55730).

The fix will be included in TensorFlow 2.9.0. We will also cherrypick this commit on TensorFlow 2.8.1, TensorFlow 2.7.2, and TensorFlow 2.6.4, as these are also affected and still in supported range.

Please consult our security guide for more information regarding the security model and how to contact us with issues and questions.

This vulnerability has been reported externally via a GitHub issue.