Infinite loop in TensorFlow Lite
tensorflow
Affected versions
| Product | Affected | Fixed in |
|---|---|---|
| tensorflow PyPI |
|
|
| tensorflow-cpu PyPI |
tensorflow
| Product | Affected | Fixed in |
|---|---|---|
| tensorflow PyPI |
|
|
| tensorflow-cpu PyPI |
≥ 2.6.0rc0 · < 2.6.0rc2≥ 2.3.0 · < 2.3.4≥ 2.4.0 · < 2.4.3
|
| tensorflow-gpu PyPI |
|
|
TensorFlow is an end-to-end open source platform for machine learning. In affected versions the strided slice implementation in TFLite has a logic bug which can allow an attacker to trigger an infinite loop. This arises from newly introduced support for [ellipsis in axis definition](https://github.com/tensorflow/tensorflow/blob/149562d49faa709ea80df1d99fc41d005b81082a/tensorflow/lite/kernels/strided_slice.cc#L103-L122). An attacker can craft a model such that `ellipsis_end_idx` is smaller than `i` (e.g., always negative). In this case, the inner loop does not increase `i` and the `continue` statement causes execution to skip over the preincrement at the end of the outer loop. We have patched the issue in GitHub commit dfa22b348b70bb89d6d6ec0ff53973bacb4f4695. TensorFlow 2.6.0 is the only affected version.
From the CVE record · GitHub_M
TFLite’s strided slice implementation has a logic bug that can cause an infinite loop. An attacker can trigger it by crafting a model that an affected TensorFlow application processes; the input does not specify whether authentication is required.
Check installations of the PyPI packages tensorflow, tensorflow-cpu, and tensorflow-gpu in versions >=2.6.0rc0 <2.6.0rc2, >=2.3.0 <2.3.4, or >=2.4.0 <2.4.3.