CVE-2022-23594
If an attacker changes the `SavedModel` format on disk to invalidate these assumptions and the `GraphDef` is then converted to MLIR-based IR then they can cause a crash in the Python interpreter.
Does this matter?
Lower severity and a low EPSS score (0.14%). Track it; it rarely justifies an emergency change on its own.
Description
Tensorflow is an Open Source Machine Learning Framework. The TFG dialect of TensorFlow (MLIR) makes several assumptions about the incoming `GraphDef` before converting it to the MLIR-based dialect. If an attacker changes the `SavedModel` format on disk to invalidate these assumptions and the `GraphDef` is then converted to MLIR-based IR then they can cause a crash in the Python interpreter. Under certain scenarios, heap OOB read/writes are possible. These issues have been discovered via fuzzing and it is possible that more weaknesses exist. We will patch them as they are discovered.
- CVSS 3.1
- 5.5 MEDIUMCVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H
- EPSS
- 0.14% probability · 4th percentile
- CISA KEV
- Not listed
- Weakness
- CWE-125, CWE-787
- Affected
- google/tensorflow
- Source
- security-advisories@github.com
References
- https://github.com/tensorflow/tensorflow/security/advisories/GHSA-9x52-887g-fhc2Third Party Advisory
- https://github.com/tensorflow/tensorflow/tree/274df9b02330b790aa8de1cee164b70f72b9b244/tensorflow/core/ir/importexportThird Party Advisory
- https://github.com/tensorflow/tensorflow/security/advisories/GHSA-9x52-887g-fhc2Third Party Advisory
- https://github.com/tensorflow/tensorflow/tree/274df9b02330b790aa8de1cee164b70f72b9b244/tensorflow/core/ir/importexportThird Party Advisory
Source: NVD record, EPSS from FIRST.org, KEV from CISA. Refreshed daily.