CVE-2022-21728
Description
Tensorflow is an Open Source Machine Learning Framework. The implementation of shape inference for `ReverseSequence` does not fully validate the value of `batch_dim` and can result in a heap OOB read. There is a check to make sure the value of `batch_dim` does not go over the rank of the input, but there is no check for negative values. Negative dimensions are allowed in some cases to mimic Python's negative indexing (i.e., indexing from the end of the array), however if the value is too negative then the implementation of `Dim` would access elements before the start of an array. 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.
How to fix
- tensorflow2.5.3, 2.6.3, 2.7.1 or later
- tensorflow-cpu2.5.3, 2.6.3, 2.7.1 or later
- tensorflow-gpu2.5.3, 2.6.3, 2.7.1 or later
Affected products
| Product | Affected versions | Fixed in |
|---|---|---|
| google tensorflow | v ≤ 2.5.2, 2.6.0 ≤ v ≤ 2.6.2, 2.7.0 | — |
| tensorflowPyPI | v < 2.5.3, 2.6.0 ≤ v < 2.6.3, 2.7.0 ≤ v < 2.7.1 | 2.5.3, 2.6.3, 2.7.1 |
| tensorflow-cpuPyPI | v < 2.5.3, 2.6.0 ≤ v < 2.6.3, 2.7.0 ≤ v < 2.7.1 | 2.5.3, 2.6.3, 2.7.1 |
| tensorflow-gpuPyPI | v < 2.5.3, 2.6.0 ≤ v < 2.6.3, 2.7.0 ≤ v < 2.7.1 | 2.5.3, 2.6.3, 2.7.1 |
Check your version
Choose the product and enter the version you run to see whether it falls in the affected range.
Attack Surface Management
See the assets tied to your domains, the technologies they run and known vulnerabilities in one list.