CVE-2020-15190
Segfault in Tensorflow
In short
A flaw in TensorFlow's Switch operation causes the program to crash when processing certain tensor outputs. The eager runtime tries to access an undefined tensor (nullptr) instead of checking if it exists first, leading to a segmentation fault.
Technical detail
The tf.raw_ops.Switch operation outputs two tensors conditionally based on a boolean input, but the eager runtime traverses both outputs without validation. When one tensor is undefined (nullptr), binding a reference to it triggers undefined behavior and segmentation fault. The vulnerability requires direct use of the vulnerable operation and affects TensorFlow versions before 1.15.4, 2.0.3, 2.1.2, 2.2.1, and 2.3.1.
Summary generated and translated by AI from the official description.
In Tensorflow before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, the `tf.raw_ops.Switch` operation takes as input a tensor and a boolean and outputs two tensors. Depending on the boolean value, one of the tensors is exactly the input tensor whereas the other one should be an empty tensor. However, the eager runtime traverses all tensors in the output. Since only one of the tensors is defined, the other one is `nullptr`, hence we are binding a reference to `nullptr`. This is undefined behavior and reported as an error if compiling with `-fsanitize=null`. In this case, this results in a segmentation fault The issue is patched in commit da8558533d925694483d2c136a9220d6d49d843c, and is released in TensorFlow versions 1.15.4, 2.0.3, 2.1.2, 2.2.1, or 2.3.1.
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:L
Affected products
tensorflow · tensorflowWant to know if your infrastructure is exposed to this?
Talk to TrueHacking →References
http://lists.opensuse.org/opensuse-security-announce/2020-10/msg00065.htmlhttps://github.com/tensorflow/tensorflow/commit/da8558533d925694483d2c136a9220d6d49d843chttps://github.com/tensorflow/tensorflow/releases/tag/v2.3.1https://github.com/tensorflow/tensorflow/security/advisories/GHSA-4g9f-63rx-5cw4