CVE-2022-35970: Segfault in `QuantizedInstanceNorm` in TensorFlow
TensorFlow is an open source platform for machine learning. If QuantizedInstanceNorm is given xmin or xmax 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.
Affected Software
Remediation
Event History
Frequently Asked Questions
What is the severity of CVE-2022-35970?
CVE-2022-35970 has a severity rating that includes a potential denial of service due to a segmentation fault.
How do I fix CVE-2022-35970?
To mitigate CVE-2022-35970, upgrade to a patched version of TensorFlow above the affected releases.
Which versions of TensorFlow are affected by CVE-2022-35970?
CVE-2022-35970 affects TensorFlow versions 2.7.0 to 2.7.2, 2.8.0 to 2.8.1, and 2.9.0 to 2.9.1, along with specific 2.10 release candidates.
What component of TensorFlow is vulnerable in CVE-2022-35970?
The vulnerability in CVE-2022-35970 is related to the 'QuantizedInstanceNorm' function when given nonzero rank tensors for 'x_min' or 'x_max'.
What impact does CVE-2022-35970 have on applications using TensorFlow?
CVE-2022-35970 can lead to a segmentation fault, potentially allowing an attacker to perform a denial of service on applications using the affected TensorFlow versions.