CVE-2021-41221
Description
TensorFlow is an open source platform for machine learning. In affected versions the shape inference code for the `Cudnn*` operations in TensorFlow can be tricked into accessing invalid memory, via a heap buffer overflow. This occurs because the ranks of the `input`, `input_h` and `input_c` parameters are not validated, but code assumes they have certain values. The fix will be included in TensorFlow 2.7.0. We will also cherrypick this commit on TensorFlow 2.6.1, TensorFlow 2.5.2, and TensorFlow 2.4.4, as these are also affected and still in supported range.
How to fix
- google tensorflow2.4.4, 2.5.2, 2.6.1 or later
- tensorflow2.6.1, 2.5.2, 2.4.4, 2.7.0 or later
- tensorflow-cpu2.6.1, 2.5.2, 2.4.4, 2.7.0 or later
- tensorflow-gpu2.6.1, 2.5.2, 2.4.4, 2.7.0 or later
Affected products
| Product | Affected versions | Fixed in |
|---|---|---|
| google tensorflow | 2.4.0 ≤ v < 2.4.4, 2.5.0 ≤ v < 2.5.2, 2.6.0 ≤ v < 2.6.1, 2.7.0 | 2.4.4, 2.5.2, 2.6.1 |
| tensorflowPyPI | 2.6.0 ≤ v < 2.6.1, 2.5.0 ≤ v < 2.5.2, v < 2.4.4, 2.7.0rc0 ≤ v < 2.7.0 | 2.6.1, 2.5.2, 2.4.4, 2.7.0 |
| tensorflow-cpuPyPI | 2.6.0 ≤ v < 2.6.1, 2.5.0 ≤ v < 2.5.2, v < 2.4.4, 2.7.0rc0 ≤ v < 2.7.0 | 2.6.1, 2.5.2, 2.4.4, 2.7.0 |
| tensorflow-gpuPyPI | 2.6.0 ≤ v < 2.6.1, 2.5.0 ≤ v < 2.5.2, v < 2.4.4, 2.7.0rc0 ≤ v < 2.7.0 | 2.6.1, 2.5.2, 2.4.4, 2.7.0 |
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.