Segfault in `Requantize` in TensorFlow
tensorflow
Affected versions
| Product | Affected | Fixed in |
|---|---|---|
| tensorflow PyPI |
|
|
| tensorflow-cpu PyPI |
|
tensorflow
| Product | Affected | Fixed in |
|---|---|---|
| tensorflow PyPI |
|
|
| tensorflow-cpu PyPI |
|
≥ 2.8.0 · < 2.8.1≥ 2.9.0 · < 2.9.1
|
| tensorflow-gpu PyPI |
|
|
TensorFlow is an open source platform for machine learning. If `Requantize` is given `input_min`, `input_max`, `requested_output_min`, `requested_output_max` tensors of a nonzero rank, it results in a segfault that can be used to trigger a denial of service attack. We have patched the issue in GitHub commit 785d67a78a1d533759fcd2f5e8d6ef778de849e0. The fix will be included in TensorFlow 2.10.0. We will also cherrypick this commit on TensorFlow 2.9.1, TensorFlow 2.8.1, and TensorFlow 2.7.2, as these are also affected and still in supported range. There are no known workarounds for this issue.
From the CVE record · GitHub_M
TensorFlow’s Requantize operation can segfault when its input_min, input_max, requested_output_min, and requested_output_max tensors have nonzero rank. An attacker who can cause Requantize to process such tensors can trigger a denial of service; the input does not specify whether network access, authentication, or user interaction is required.
Engineers running the PyPI tensorflow, tensorflow-cpu, or tensorflow-gpu packages should check for versions below 2.7.2, from 2.8.0 to below 2.8.1, or from 2.9.0 to below 2.9.1.