CVE-2020-15193

Description

In Tensorflow before versions 2.2.1 and 2.3.1, the implementation of `dlpack.to_dlpack` can be made to use uninitialized memory resulting in further memory corruption. This is because the pybind11 glue code assumes that the argument is a tensor. However, there is nothing stopping users from passing in a Python object instead of a tensor. The uninitialized memory address is due to a `reinterpret_cast` Since the `PyObject` is a Python object, not a TensorFlow Tensor, the cast to `EagerTensor` fails. The issue is patched in commit 22e07fb204386768e5bcbea563641ea11f96ceb8 and is released in TensorFlow versions 2.2.1, or 2.3.1.

Weakness classification

• security-advisories@github.com: CWE-908
• nvd@nist.gov: CWE-908

Primary references

• security-advisories@github.com — Mailing List, Third Party Advisory
• security-advisories@github.com — Patch, Third Party Advisory
• security-advisories@github.com — Third Party Advisory
• security-advisories@github.com — Exploit, Third Party Advisory
• af854a3a-2127-422b-91ae-364da2661108 — Mailing List, Third Party Advisory
• af854a3a-2127-422b-91ae-364da2661108 — Patch, Third Party Advisory
• af854a3a-2127-422b-91ae-364da2661108 — Third Party Advisory
• af854a3a-2127-422b-91ae-364da2661108 — Exploit, Third Party Advisory